"Medicine, Faculty of"@en . "Population and Public Health (SPPH), School of"@en . "DSpace"@en . "UBCV"@en . "Wood, Evan"@en . "2009-11-13T14:12:24Z"@en . "2003"@en . "Doctor of Philosophy - PhD"@en . "University of British Columbia"@en . "Objectives: To identify the proportion of HIV-related deaths that occur among\r\npersons who have never accessed antiretroviral therapy; model the potential public\r\nhealth impact of improved access to antiretroviral therapy; determine the role of\r\nsocio-economic status on access and as a predictor of HIV disease progression;\r\nevaluate the impact of adherence to HAART on survival rates; and to determine\r\nrates of therapy discontinuation and virological response among HIV-infected\r\ninjection drug users.\r\nMethods: British Columbia's HIV/AIDS Drug Treatment Program provides\r\nantiretroviral therapy to all eligible HIV positive persons in British Columbia free of\r\ncharge. Data from the Centre as well as mortality data from Vital Statistics, socioeconomic\r\ndata from Statistics Canada, and prescription refill data from pharmacies\r\nwere used to evaluate and model mortality patterns and antiretroviral therapy use\r\namong HIV-infected individuals.\r\nResults: Overall, 32.7% of persons who have died of HIV/AIDS during the period\r\n1995 - 2001 died without accessing antiretroviral therapy. Several markers of socioeconomic\r\nstatus were associated with non-receipt of therapy, and socio-economic\r\nstatus was associated with the receipt of sub-optimal therapy among those who\r\ninitiated treatment. Models developed also indicate that improved uptake of\r\nantiretroviral therapy among injection drug users could have a substantial impact on\r\nlife expectancy and mortality in this setting. Among patients who initiated triple\r\ntherapy, rates of sub-optimal adherence are common, and are associated with\r\nmarkedly higher rates of mortality. In particular, patients with a history of injection\r\ndrug use may be at risk of treatment cessation and poor virologic response.\r\nConclusions: Limited access to antiretroviral therapy, particularly among certain\r\nsub-populations, has contributed on to ongoing AIDS mortality. Among persons\r\nwho initiate therapy, the problem of limited access has been compounded by suboptimal\r\nadherence and prescription patterns, as well as high rates of treatment\r\ndiscontinuation. These data suggest that novel strategies are required to improve\r\naccess and adherence to antiretroviral therapy despite a universal healthcare system\r\nthat provides antiretrovirals and HIV/AIDS care free of charge."@en . "https://circle.library.ubc.ca/rest/handle/2429/14923?expand=metadata"@en . "9969466 bytes"@en . "application/pdf"@en . "T H E IMPACT OF ACCESS A N D A D H E R E N C E O N MORTALITY F R O M H I V DISEASE IN T H E ERA OF M O D E R N ANTIRETROVIRAL THERAPY b y E V A N W O O D B.Sc. University o f Victoria, 1997 A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF D O C T O R OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES Department o f Health Care and Epidemiology; Faculty o f Medicine We accept this thesis as conforming to the required standard Robe r t T^ f c l o gg Mark V f ' T y n d a l f Michael V . O'Sh^ug^messy Montaner T H E UNIVERSITY O F BRITISH C O L U M B I A March, 2003 \u00C2\u00A9 E v a n Wood, 2003 In p r e s e n t i n g th is thes is in partial fu l f i lment o f t h e requ i rements fo r an a d v a n c e d d e g r e e at the Univers i ty o f Brit ish C o l u m b i a , I agree that t h e Library shall m a k e it f ree ly avai lable f o r re fe rence a n d study . I fur ther agree that p e r m i s s i o n f o r e x t e n s i v e c o p y i n g of this thesis fo r scho lar ly p u r p o s e s may b e g ranted by the h e a d o f m y d e p a r t m e n t o r b y his o r h e r representat ives . It is u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n of this thesis fo r f inancial ga in shall no t b e a l l o w e d w i t h o u t m y w r i t t e n p e r m i s s i o n . D e p a r t m e n t of W&*~\YWc^ sg_ e ) 6^)\c)eAAA i a (o T h e Un ivers i ty o f British C o l u m b i a V a n c o u v e r , C a n a d a D a t e D E - 6 (2/88) ABSTRACT Objectives: To identify the proportion of HIV-related deaths that occur among persons who have never accessed antiretroviral therapy; model the potential public health impact of improved access to antiretroviral therapy; determine the role of socio-economic status on access and as a predictor of HIV disease progression; evaluate the impact of adherence to H A A R T on survival rates; and to determine rates of therapy discontinuation and virological response among HIV-infected injection drug users. Methods: British Columbia's H I V / A I D S D r u g Treatment Program provides antiretroviral therapy to all eligible HIV positive persons in British Columbia free of charge. Data from the Centre as well as mortality data from Vital Statistics, socio-economic data from Statistics Canada, and prescription refill data from pharmacies were used to evaluate and model mortality patterns and antiretroviral therapy use among HIV-infected individuals. Results: Overall, 32.7% of persons who have died of H I V / A I D S during the period 1995 - 2001 died without accessing antiretroviral therapy. Several markers of socio-economic status were associated with non-receipt of therapy, and socio-economic status was associated with the receipt of sub-optimal therapy among those who initiated treatment. Models developed also indicate that improved uptake of antiretroviral therapy among injection drug users could have a substantial impact on life expectancy and mortality in this setting. Among patients who initiated triple therapy, rates of sub-optimal adherence are common, and are associated with markedly higher rates of mortality. In particular, patients with a history of injection drug use may be at risk of treatment cessation and poor virologic response. Conclusions: Limited access to antiretroviral therapy, particularly among certain sub-populations, has contributed on to ongoing AIDS mortality. Among persons who initiate therapy, the problem of limited access has been compounded by sub-optimal adherence and prescription patterns, as well as high rates of treatment discontinuation. These data suggest that novel strategies are required to improve access and adherence to antiretroviral therapy despite a universal healthcare system that provides antiretrovirals and H I V / A I D S care free of charge. n TABLE OF CONTENTS ABSTRACT i i T A B L E OF CONTENTS < i i i LIST OF TABLES v i LIST OF FIGURES v i i i ACKNOWLEDGEMENTS X CHAPTER 1: INTRODUCTION 1 1.1 H I V Epidemiology and Treatment 1 1.2 Current Issues 3 1.3 Outstanding Questions and Study Justification 4 1.4 Study Objectives and Thesis Organization 5 1.5 Summary.. 7 1.6 References 8 CHAPTER 2: B A C K G R O U N D . 12 2.1 F o r w a r d : \"Access to A n t i r e t r o v i r a l Therapy among M a r g i n a l i z e d Populations i n the Developed W o r l d \" 12 2.1 Introduction 13 2.2 Populations at Risk of Poor Access 15 2.3 Access to and Discontinuation of H I V Treatment 18 2.4 Barriers to the use of Antiretroviral Therapy :....20 2.5 Strategies for Improving Access and Retention in H I V Treatment 22 2.6 Summary 26 2.6 References... 28 CHAPTER 3: STUDY SETTING, OVERVIEW OF M E T H O D S , A N D LIMITATIONS 40 3.1 HTV Epidemiology and Treatment i n British Columbia 40 3.2 Methodological Approaches and Data Sources.... '. 42 i . Socio-economic Status 42 i i . Mortality Data .\" 43 i i i . Demographic Projection Models 44 iv. Treatment Discontinuation and Adherence Estimates 45 3.3 Statistical Methods 46 3.4 Strengths and Limitations 48 3.5 Summary : 49 3.6 References 51 CHAPTER 4: ACCESS TO ANTIRETROVIRAL THERAPY PRIOR TO D E A T H 54 4.1 Forward:\"Race, Gender, Socio-economic Status, and Access to HIV-1 Treatment among Persons who have Died i n the Era of Antiretroviral Therapy\" 54. 4.2 Abstract 55 4.3 Introduction 56 4.4 Methods ..57 4.5 Results 61 4.6 Discussion 63 4.7 Summary 65 4.8 References 66 CHAPTER 5: M O D E L L I N G T H E IMPACT OF IMPROVED U P T A K E OF T R E A T M E N T 1 76 5.1 Forward:\"Antiretroviral Medication Use A m o n g Injection Drug Users: Two Potential Futures\" 76 5.2 Abstract '. 77 5.3 Introduction 78 5.4 Methods '. 79 5.5 Results ,83 5.6 Discussion 85 5.7 Summary 88 5.8 References 90 CHAPTER 6: M O D E L L I N G T H E IMPACT OF IMPROVED U P T A K E OF T R E A T M E N T II 99 6.1 Forward:\"Extent to Which Low-Level use of Antiretroviral Treatment Could Curb the AIDS Epidemic in sub-Saharan Africa\" 99 6.2 Abstract .' '. 100 6.3 Introduction 101 6.4 Methods 102 6.5 Results 105 6.6 Discussion 108 6.7 Summary 114 6.8 References 115 6.9 Correspondence 126 CHAPTER 7: SOCIO-ECONOMIC STATUS A N D ACCESS TO TRIPLE THERAPY 130 7.1 Forward:\"Socioeconomic Status, Access to Triple Therapy, and Survival from HIV-disease Since 1996\" 130 7.2 Abstract '. 131 7.3 Introduction 132 7.4 Methods 133 7.5 Results 137 7.6 Discussion 141 7.7 Summary : 144 7.8 References 145 CHAPTER 8: A D H E R E N C E TO THERAPY A N D O U T C O M E S FROM H A A R T -. 156 8.1 Forward:\"The Impact of Adherence and CD4 Cell Count on Survival after the Initiation of HAART\" 156 8.2 Abstract '. 157 8.3 Introduction 158 8.4 Methods 159 8.5 Results 163 8.6 Discussion 167 8.7 Summary 169 8.8 References 170 CHAPTER 9: O U T C O M E S F R O M H A A R T A M O N G H I V - I N F E C T E D I D U 183 9.1 Forward:\"Highly Active Antiretroviral Therapy and the HIV-infected Injection Drug User: The State of Care in a Universal Healthcare System\" 183 9.2 Abstract '. 184 9.3 Introduction 185 9.4 Methods , 186 9.5 Results 189 9.6 Discussion 191 9.7 Summary '. 193 9.8 References .\u00E2\u0080\u00A2 195 C H A P T E R 10: SUMMARY, R E C O M M E N D A T I O N S , F U R T H E R RESEARCH, A N D CONCLUSIONS 204 10.1 Summary of Findings 204 10.2 Unique Contribution, Impact, and Implications 209 i v 10.3 Recomendations .....213 10.4 Future Research 215 10.5 Conclusions 217 . 10.6 References 218 APPENDIX 1: STATEMENT OF AUTHORSHIP 224 APPENDIX 2: CERTIFICATE OF ETHICAL APPROVAL .....226 V LIST OF T A B L E S T A B L E 4.1 U N I V A R I A T E A N A L Y S E S O F S O C I O - D E M O G R A P H I C A N D N E I G H B O R H O O D - B A S E D C H A R A C T E R I S T I C S W I T H R E G A R D T O E V E R ACCESSING HIV T R E A T M E N T PRIOR T O D E A T H 71 T A B L E 4.2 LOGISTIC REGRESSION ANALYSIS OF FACTORS ASSOCIATED WITH EVER RECEIVING ANTIRETROVIRAL T R E A T M E N T PRIOR T O D E A T H 72 T A B L E 4.3 U N I V A R I A T E A N D M U L T I V A R I A T E C O X P R O P O R T I O N A L H A Z A R D S . R E G R E S S I O N O F F A C T O R S A S S O C I A T E D W I T H T I M E T O ANTIRETROVIRAL T H E R A P Y DISCONTINUATION > 3 M O N T H S 73 T A B L E 5.1 P O T E N T I A L LIFE E X P E C T A N C I E S IN T H E D O W N T O W N EASTSIDE D U R I N G T H E YEAR 2006, G I V E N T H R E E LEVELS OF HIV P R E V A L E N C E 94 T A B L E 6.1 M O D E L PARAMETERS, D A T A SOURCES, A N D V A L U E S U S E D IN M O D E L S 120 T A B L E 6.2 C U M U L A T I V E N E W AIDS CASES A N D HIV-POSITIVE BIRTHS O V E R T H E STUDY PERIOD, A N D LIFE E X P E C T A N C Y IN T H E YEAR 2005 121 T A B L E 6.3 M O D E L L E D T O T A L COSTS, PROPORTION O F PER P E R S O N H E A L T H -C A R E EXPENDITURE, A N D COST PER LIFE-YEAR G A I N E D REQUIRED T O F I N A N C E ANTIRETROVIRAL USE, BY SCENARIO 122 T A B L E 7.1 U N I V A R I A T E A N A L Y S E S C O M P A R I N G S O C I O - D E M O G R A P H I C A N D CLINICAL CHARACTERISTICS O F PARTICIPANTS PRESCRIBED D O U B L E VERSUS TRIPLE C O M B I N A T I O N ANTIRETROVIRAL T H E R A P Y 150 T A B L E 7.2 M U L T I V A R I A T E A N A L Y S I S OF F A C T O R S A S S O C I A T E D W I T H BEING PRESCRIBED TRIPLE C O M B I N A T I O N THERAPY AS T H E INITIAL REGIMEN 151 T A B L E 7.3: U N I V A R I A T E A N D M U L T I V A R I A T E A N A L Y S I S O F T H E B A S E L I N E FACTORS A S S O C I A T E D WITH D E A T H A M O N G 1,408 ANTIRETROVIRAL N A I V E PERSONS FIRST PRESCRIBED D O U B L E OR TRIPLE C O M B I N A T I O N A N T I R E T R O V I R A L T H E R A P Y B E T W E E N A U G U S T 1, 1996 A N D D E C E M B E R 31,1999 , 152 V I T A B L E 7.4 A D J U S T E D RELATIVE H A Z A R D FOR T H E RISK OF D E A T H STRATIFIED BY S O C I O - E C O N O M I C S T R A T A A N D TRIPLE T H E R A P Y I N T H E INITIAL REGIMEN 153 T A B L E 8.1 U N I V A R I A T E A N D M U L T I V A R I A T E ANALYSIS OF FACTORS ASSOCIATED W I T H D E A T H A M O N G 1416 PERSONS FIRST PRESCRIBED TRIPLE C O M B I N A T I O N ANTIRETROVIRAL T H E R A P Y B E T W E E N A U G U S T 1, 1996 A N D JULY 31,2000 175 T A B L E 8.2 U N I V A R I A T E A N D MULTIVARIATE ANALYSIS OF FACTORS ASSOCIATED W I T H D E A T H A M O N G 1053 PERSONS FIRST PRESCRIBED TRIPLE C O M B I N A T I O N A N T I R E T R O V I R A L T H E R A P Y B E T W E E N JULY 1, 1997 A N D JULY 1,2000 ; 176 T A B L E 8.3(A) A D J U S T E D R E L A T I V E H A Z A R D OF M O R T A L I T Y FOR BASELINE C D 4 C E L L C O U N T A N D PHYSICIAN EXPERIENCE CATEGORIES , 177 T A B L E 8.3(B) A D J U S T E D R E L A T I V E H A Z A R D O F M O R T A L I T Y FOR BASELINE C D 4 C E L L C O U N T A N D A D H E R E N C E CATEGORIES 178 T A B L E 9.1 U N I V A R I A T E A N A L Y S E S O F B A S E L I N E C H A R A C T E R I S T I C S A M O N G T R E A T M E N T N A I V E PATIENTS INITIATING HAART STRATIFIED BY HISTORY OF INJECTION D R U G USE 199 T A B L E 9.2(A) C R U D E A N D A D J U S T E D R E L A T I V E H A Z A R D S (RH) OF T R E A T M E N T DISCONTINUATION 200 T A B L E 9.2(B) C R U D E A N D A D J U S T E D R E L A T I V E H A Z A R D S (RH) O F HIV R N A SUPPRESSION ..200 T A B L E 9.2(c) C R U D E A N D A D J U S T E D R E L A T I V E H A Z A R D S (RH) OF HIV R N A R E B O U N D 200 v i i LIST OF FIGURES FIGURE 4.1 K A P L A N - M E I E R P R O D U C T LIMIT E S T I M A T E S O F C U M U L A T I V E T H E R A P Y D I S C O N T I N U A T I O N _>. 3 M O N T H S A M O N G H I V - I N F E C T E D SUBJECTS W H O L A T E R D I E D B E T W E E N J A N U A R Y 1, 1995 A N D D E C E M B E R 31, 2001 STRATIFIED BY A), ETHNICITY; B) G E N D E R ; A N D C) S O C I O - E C O N O M I C S T A T U S . N O T E , T H E C U R V E S D O N O T D R O P U N T I L 1 M O N T H A F T E R T H E INITIATION OF T H E R A P Y SINCE A L L I N D I V I D U A L S A R E INITIALLY D I S P E N S E D 1 M O N T H S W O R T H O F T R E A T M E N T 75 FIGURE 5.1 T H E C H A N G E IN O V E R A L L LIFE EXPECTANCIES A T BIRTH PROJECTED IN SCENARIOS O N E A N D T W O WITH A 7% P R E V A L E N C E . ART REFERS T O T H E P E R C E N T A G E OF H IV -POSITIVE I N D I V I D U A L S A C T I V E L Y RECEIVING ANTIRETROVIRAL T H E R A P Y 96 FIGURE 5.2 A N N U A L A I D S D E A T H S PROJECTED IN SCENARIOS O N E A N D T W O W I T H A 7% P R E V A L E N C E . A R T REFERS T O T H E P E R C E N T A G E O F H I V - P O S I T I V E INDIVIDUALS A C T I V E L Y RECEIVING ANTIRETROVIRAL T H E R A P Y 97 FIGURE 5.3 T H E A N N U A L A I D S D E A T H RATE PROJECTED IN SCENARIOS O N E A N D T W O WITH A 7% P R E V A L E N C E . A R T REFERS T O T H E P E R C E N T A G E OF . H I V - P O S I T I V E INDIVIDUALS A C T I V E L Y RECEIVING A N T I R E T R O V I R A L T H E R A P Y . : 98 FIGURE 6.1 A N N U A L H I V - P O S I T I V E BIRTHS PROJECTED IN SCENARIOS Z E R O T O T H R E E , WITH T H E ASSUMPTION OF 40% R E D U C T I O N 124 FIGURE 6.2 A N N U A L LIFE E X P E C T A N C Y PROJECTED IN SCENARIOS ZERO T O F O U R 125 FIGURE 7.1 A N A L Y S I S O F 1,408 A N T I R E T R O V I R A L N A I V E PARTICIPANTS W H O INITIATED=THERAPY WITH EITHER A D O U B L E OR TRIPLE C O M B I N A T I O N REGIMEN STRATIFIED BY L O W A N D H I G H S O C I O - E C O N O M I C STATUS 155 viii FIGURE 8.1 FIGURE 8.2 FIGURE 8.3 FIGURE 9.1 K A P L A N - M E I E R P R O D U C T L I M I T E S T I M A T E S O F C U M U L A T I V E PROGRESSION T O D E A T H A M O N G 1,416 H I V - I N F E C T E D SUBJECTS W H O S T A R T E D N A I V E O N A N T I R E T R O V I R A L T H E R A P Y B E T W E E N A U G U S T 1,1996 A N D JULY 1, 2000, STRATIFIED BY C D 4 C E L L C O U N T GROUPINGS (< 50,50 - 1 9 9 , > 200) .180 K A P L A N - M E I E R P R O D U C T L I M I T E S T I M A T E S O F C U M U L A T I V E PROGRESSION T O D E A T H A M O N G 1,414 H I V - I N F E C T E D SUBJECTS W H O S T A R T E D N A I V E O N A N T I R E T R O V I R A L T H E R A P Y B E T W E E N A U G U S T 1, 1996 A N D JULY 1, 2000, STRATIFIED BY C D 4 C E L L C O U N T GROUPINGS (< 50, 50 - 199, > 200), A N D H T V - E X P E R I E N C E D ( P A N E L A ) A N D N O N - E X P E R I E N C E D ( P A N E L B) CATEGORIES .181 K A P L A N - M E I E R P R O D U C T L I M I T E S T I M A T E S O F C U M U L A T I V E PROGRESSION T O D E A T H A M O N G 1,416 H I V - I N F E C T E D SUBJECTS W H O S T A R T E D N A I V E O N A N T I R E T R O V I R A L T H E R A P Y B E T W E E N A U G U S T 1,1996 A N D JULY 1, 2000, STRATIFIED BY C D 4 C E L L C O U N T GROUPINGS (< 50, 50 - 199, > 200) A N D A D H E R E N T ( P A N E L A ) A N D N O N - A D H E R E N T ( P A N E L B) CATEGORIES. ,182 T I M E T O T R E A T M E N T D I S C O N T I N U A T I O N A M O N G P A T I E N T S INITIATING H A A R T STRATIFIED BY HISTORY OF INJECTION D R U G USE. N O T E , T H E C U R V E D O E S N O T D R O P U N T I L 1 M O N T H A F T E R T H E INITIATION OF T H E R A P Y SINCE A L L INDIVIDUALS A R E INITIALLY DISPENSED 1 M O N T H S W O R T H OF T R E A T M E N T .202 FIGURE 9.2 T I M E T O H I V R N A SUPPRESSION ( P A N E L A ) A N D H I V R N A R E B O U N D ( P A N E L B) A M O N G P A T I E N T S INITIATING H A A R T STRATIFIED BY HISTORY OF INJECTION D R U G USE .203 I X A C K N O W L E D G E M E N T S I would like to express my sincere gratitude to my committee members: Drs Robert Hogg, Mark Tyndall , Michael O'Shaughnessy, and Julio Montaner for their generosity and wisdom. Over the last several years, their support consistently went well beyond what is expected of a graduate supervisor, and for that I will always be in their debt. I would also like to extend my regards to my fellow researchers and statisticians at the BC Centre for Excellence in H I V / A I D S and the Department of Healthcare and Epidemiology for all their advice and mentoring over the last several years. In particular, Dr Martin Schechter, Paula Braitstein, Benita Yip, Keith Chan, and Kevin Craib. I would also like to acknowledge the financial contributions of the BC Health Research Foundation, the Canadian Health Services Research Foundation, the George F Elliot Foundation, the National Health Research Development Program, the Canadian Institutes of Health Research, and the BC Centre for Excellence in H I V / A I D S who all provided funding for me during the course of my PhD. Portions of this dissertation were recognized by the Canadian Association for HIV Research, when I was awarded their Young Investigator Award in the summer of 2000. I am also proud to note that one of the manuscripts presented here was also cited by Boehringer Ingelheim in July, 2000 as impetus for their donation of free antiretrovirals to developing countries for the prevention of mother to child transmission (see Appendix 3). Finally, since most of the manuscripts presented here have now been externally reviewed, I would like to pass along my gratitude to the anonymous peer-reviewers who have provided critical feedback for much of this work prior to its publication. CHAPTER 1: INTRODUCTION 1.1 H I V E P I D E M I O L O G Y A N D T R E A T M E N T O n June 5, 1981 the U.S. Centers for Disease Control's Mortality and Morbidity Weekly Report published the first account of a cluster of peculiar deaths from Pneumocycstis carinii pneumonia and Kaposi's sarcoma among otherwise healthy gay men in California.1 The deaths were attributed to what was subsequently referred to as \"gay cancer\" and \"gay-related immune deficiency\" or GRID. Later this same year, cases of the unexplained disease begun to appear in illicit drug users.2 In 1983, the virus responsible for the illness was isolated and would become known as the human immunodeficiency virus or HrV.3 HIV is a retrovirus that causes chronic infection in humans. The virus replicates in CD4 lymphocytes resulting in their destruction and eventually immunodeficiency and death in untreated individuals.^ The immunodeficiency is characterized by progressive loss of immune function and eventually leads to Acquired Immunodeficiency Syndrome (AIDS). AIDS is defined by the occurrence of an AIDS defining illnesses in accordance with the Centers for Disease Control and the World Health Organization, or by a CD4 cell count less that 200 cells/mm 3 in the presence of positive HIV serology in the U.S.5 Soon after HIV was identified, the search for drugs to fight the virus was started. Drugs that have been developed to date have primarily aimed to inhibit the virus at various stages of its replication. The first major milestone in the 1 treatment of HIV was made in 1987 with the approval of zidovudine or A Z T , a nucleoside reverse transcriptase inhibitor (NRTI) that was demonstrated to improve survival among patients with advanced disease.^ Since the early use of A Z T , additional NRTIs and other classes of therapeutic agents, including non-nucleoside reverse transcriptase inhibitors (NNRTI) and protease inhibitors (PI), were developed and shown to be efficacious through clinical trials/'^ Since this time, it has been demonstrated that triple combination therapy with three antiretroviral agents is superior to mono or dual therapy,9/10 and in 1997 triple combination therapy, commonly referred to as highly active antiretroviral therapy or H A A R T , became the standard of care.H For the most part, these drugs are only available in the developed world, and treatment remains out of reach for the majority of H I V infected persons. 12,13 This is particularly concerning with regards treatments to prevent mother to child transmission of HIV.14 The arsenal of drugs currently available for the treatment of HIV has resulted in major strides in the medical management of HIV disease. Newer antiretroviral regimens have proven effective in decreasing plasma HIV R N A and improving CD4 cell counts, which has resulted in a substantial alteration of the natural history of HIV infection, and dramatic improvements in HIV-related morbidity and mortality. 10,15-17 In fact, in areas of the world where antiretrovirals are widely available, HIV disease progression and AIDS deaths have become so rare that HIV infection is increasingly being viewed as a chronic 2 and manageable illness. 13/19 Despite these remarkable achievements, several major challenges exist in the treatment of HIV disease. 1.2 CURRENT ISSUES Current therapeutic strategies aim to reduce the circulating HIV R N A to as low a level as possible for as long as possible, however, the eradication of HIV from the individual is not possible with presently available therapeutic agents 2 0 A s such, persons on therapy must take a daily regimen of at least three antiretroviral drugs, and follow a scheduled dosing that often involves coordination of dietary intake. 2 ! Previous studies have consistently shown that very high levels of adherence are required to achieve successful HIV R N A suppression and to prevent virological rebound. 2 2 ' 2 3 T / n e challenges of following this strict routine are further complicated by the common side effects of antiretrovirals that include nausea, diarrhea, heart disease, and in more serious cases, morphologic changes and life threatening toxitities. 2^/ 2^ In many instances, the challenges of adhering to a complex regimen of antiretroviral therapy are compounded by the challenges associated with delivering medical care to marginalized and difficult to reach populations. 2^ As is the case with other diseases that may be more common among marginalized populations, such as tuberculosis, 2^ access to appropriate healthcare for HIV infected populations has become a growing problem.28\"30 3 1.3 O U T S T A N D I N G QUESTIONS A N D STUDY JUSTIFICATION Previous studies have shown that women, ethnic minorities, persons of lower socio-economic status, and injection drug users may have lower access to antiretroviral therapy even i n settings where patients are el igible for antiretrovirals and A I D S care free of charge.28-30 However, these studies have been l imited i n that they have been conducted among populations of persons still l iv ing w i t h H I V , and have not examined patients actually dying of A I D S without accessing therapy. For this reason, it is presently not k n o w n to what extent apparent disparities i n access to antiretrovirals w i l l improve as H I V -disease becomes more advanced among persons who appear to have lower access. In addition, researchers and policy makers have little understanding of the potential impact that improved access to therapy could have on the health of communities where H I V prevalence is high among populations, such as injection drug users, where access to antiretroviral therapy is k n o w n to be poor. Most notably, few researchers have examined the potential changes i n life expectancy and AIDS mortality i n these high prevalence settings. Similarly, among persons w h o do access antiretroviral treatment, the impact Of socio-economic status and transmission group on disease progression, adherence, and access to triple therapy have not been examined. For instance, although lower socio-economic status was associated wi th higher rates of disease 4 progression in the p r e - H A A R T era, we do not presently know if socio-economic status is a relevant prognostic marker in the era of HAART.31 A final issue that is closely linked to questions regarding access, is the role that adherence plays in HIV disease progression. While many studies have shown how adherence is closely associated with virological response, 2 2 few studies have demonstrated an association between adherence and survival. Furthermore, adherence studies are often limited in that they are restricted to the proportion of patients who fill out adherence surveys or to the proportion of patients who consistently return for clinic visits. 2^ As a result, the role that treatment discontinuation may play in response to therapy has not been well explored. 1.4 S T U D Y O B J E C T I V E S A N D THESIS O R G A N I Z A T I O N The overall aim of this thesis is to address many of these outstanding questions from a population health perspective in a way that is relevant to both policy-makers and clinicians. This thesis will explore 5 objectives: 1. To identify the proportion of HIV-related deaths that occur among persons who have never accessed antiretroviral therapy despite the presence of a universal healthcare system that delivers A I D S care and all antiretroviral therapy free of charge. In Chapter 4,1 examine those characteristics such as age, gender, socio-economic status, and ethnicity as potentially associated with access to antiretrovirals prior to death. In addition, since it is also presently unknown what happens to many people after they discontinue antiretroviral therapy, as a sub-objective, I will evaluate factors associated with therapy discontinuation prior to death among the proportion of persons who had accessed H A A R T . It is my hypothesis that a substantial proportion of AIDS mortality in the province is among persons who have never accessed therapy, and that specific 5 populations, such as persons of lower socio-economic status may be at elevated risk of poor access. 2. To model the potential public health impact of improved access to antiretroviral therapy among injection drug users. In Chapters 5 and 6, demographic modeling will be used to derive estimates of the potential impact of improved access to antretrovirals on population-level health indicators such as life expectancy and AIDS mortality. The modeling procedure developed to produce these estimates in Chapter 5 will also be used to derive estimates in a lesser-developed country in Chapter 6. Here, not only will the impact of improved access to H A A R T on life expectancy be modeled, but also the number of HIV-positive births that could be avoided through the provision of short course antiretroviral prophylaxis will be estimated. It is my hypothesis that reduced mortality due to improved uptake of antiretrovirals could have a substantial demographic impact. 3. To determine the role of socio-economic status as a predictor of disease progression among persons initiating antiretroviral therapy since 1996. Although the aim will be to determine whether socio-economic status is associated with mortality after the initiation of H A A R T , an in-depth analysis of prescribing patterns will also evaluate if socio-economic status is associated with access to triple therapy. Specifically, if socio-economic status, independent of clinical characteristics, is associated with the use of double or triple therapy during the time that these combinations remained standard clinical practice.32 it is my hypothesis that socio-economic status is not a relevant predictor of mortality in the era of H A A R T . 4. To evaluate the impact of adherence to H A A R T on survival rates among patients in British Columbia while stratifying by CD4 cell count. Given the high numbers of patients that are presently not receiving therapy, and the side effects associated with antiretrovirals, a critical clinical question surrounds the optimal time to initiate H A A R T . 19 This study will therefore strive to evaluate the combined role of CD4 cell count and adherence on rates, of mortality among persons initiating H A A R T in the province. It is my hypothesis that adherence is among the strongest determinants of survival independent of CD4 cell count. 5. To develop a clearer clinical picture of the virological response rates to antiretroviral therapy, and rates of therapy discontinuation among HIV-infected injection drug users in comparison to non-injection drug users. Since the epidemiology of HIV in British Columbia is such that large numbers of IDU will be expected to be coming forward for treatment only in the next several years,^^ this study aims to evaluate outcomes from therapy among those HIV-infected IDU that have presented for treatment since 1996. It is my hypothesis that patients infected through injection 6 drug use have higher rates of treatment discontinuation and poorer virological outcomes on H A A R T . This thesis is divided into 10 chapters. This first chapter provides some preliminary background and context, and outlines the study's justification and objectives. Chapter 2 is a review paper that provides a summary and analysis of the research to date in the area of access to antiretrovirals. Chapter 3 provides an overview of the study setting, the methods employed, and the limitations of my approach. Chapters 4 through 9 are based on the six research papers that have been published or submitted for publication. Finally, Chapter 10 provides a discussion of the findings, outlines the unique contribution of the work, makes recommendations, and outlines areas for future research. 1.5 SUMMARY In summary, HIV disease is a relatively recent phenomenon that has wide-ranging public health implications. Although great strides have been made in the clinical management of HIV disease, major challenges have emerged with respect to the delivery of antiretroviral therapy. Among the greatest concerns are the challenges that exist with regards to the high degree of adherence required and the difficulties experienced in complying with these regimens. Additional challenges stem from the delivery of HIV care and treatment to populations that have been traditionally difficult to reach with healthcare services. 7 1.6 REFERENCES 1. Pneumocycstis Pneumonia - Los Angeles. MMWR Morb Mortal Wkly Rep 1981;30:250-2. 2. Current Trends Update on Acquired Immune Deficiency Syndrome (AIDS) -United States. MMWR Morb Mortal Wkly Rep 1982;31(37):507-508. 3. Barre-Sinoussi F, Chermann JC, Rey F, et al. Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science 1983;220(4599):868-71. 4. Popovic M , Flomenberg N , Volkman DJ, et al. Alteration of T-cell functions by infection with H T L V - I or H T L V - I I . Science 1984;226(4673):459-62. 5. Proposed 'World Health Organization staging system for HIV infection and disease': preliminary testing by an international collaborative cross-sectional study. The W H O International Collaborating Group for the Study of the W H O Staging System. AIDS 1993;7(5):711-8. 6. Vella S, Giuliano M , Pezzotti P, et al. Survival of zidovudine-treated patients with AIDS compared with that of contemporary untreated patients. Italian Zidovudine Evaluation Group. JAMA 1992;267(9):1232-6. 7. Collier A C , Coombs RW, Schoenfeld D A , et al. Treatment of Human Immunodeficiency Virus Infection with Saquinavir, Zidovudine, and Zalcitabine. N Engl J Med 1996;334(16):1011-8. 8. Moyle G . The emerging roles of non-nucleoside reverse transcriptase inhibitors in antiretroviral therapy. Drugs 2001;61(l):19-26. 9. Hogg RS, Rhone SA, Yip B, et al. Antiviral effect of double and triple drug combinations amongst H I V - infected adults: lessons from the 8 implementation of viral load-driven antiretroviral therapy. AIDS 1998;12(3):279-84. 10. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. ' 11. Carpenter C C , Fischl M A , Hammer S M , et al. Antiretroviral therapy for HIV infection in 1997. Updated recommendations of the International AIDS Society-USA panel. JAMA 1997;277(24):1962-9. 12. Hogg RS, Anis A , Weber A E , O'Shaughnessy M V , Schechter M T . Triple-combination antiretroviral therapy in sub-Saharan Africa [letter] [see comments]. Lancet 1997;350(9088):1406. 13. Hogg RS, Weber A E , Craib KJ, et al. One world, one hope: the cost of providing antiretroviral therapy to all nations. AIDS 1998;12(16):2203-9. 14. Guay L A , Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapihe compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: H I V N E T 012 randomised trial [see comments]. Lancet 1999;354(9181):795-802. 15. Wood E, Yip B, Hogg RS, et al. Full suppression of viral load is needed to achieve an optimal CD4 cell count response among patients on triple drug antiretroviral therapy. AIDS 2000;14(13):1955-60. 16. Wood E, Hogg RS, Yip B, et al. \"Discordant\" increases in CD4 cell count relative to plasma viral load in a closely followed cohort of patients initiating antiretroviral therapy. / Acquir Immune Defic Syndr 2002;30(2):159-66. 17. Hogg RS, O'Shaughnessy M V , Gataric N , et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997;349(9061):1294. 9 18. Hogg RS, Yip B, Chan KT, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 19. Pomerantz RJ. Initiating antiretroviral therapy during HIV infection: confusion and clarity. JAMA 2001;286(20):2597-2599. 20. Carpenter C C , Cooper D A , Fischl M A , et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000;283(3):381-90. 21. Yeni P G , Hammer S M , Carpenter C C , et al. Antiretroviral treatment for adult H I V infection in 2002: updated recommendations of the International AIDS Society-USA Panel. JAMA 2002;288(2):222-35. 22. Low-Beer S, Yip B, O'Shaughnessy M V , Hogg RS, Montaner JS. Adherence to triple therapy and viral load response. / Acquir Immune Defic Syndr 2000;23(4):360-l. 23. Paterson D L , Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000;133(l):21-30. 24. Heath K V , H o g g RS, C h a n K, et al. Lipodystrophy-associated morphological, cholesterol, and triglyceride abnormalities in a population-based H I V / A I D S Treatment Program. AIDS 2001;15:231-239. 25. Cote H C , Brumme Z L , Craib KJ, et al. Changes in mitochondrial D N A as a marker of nucleoside toxicity in HIV-infected patients. N. Engl J Med 2002;346(ll):811-20. 26. Celentano D D , Vlahov D , Cohn S, Shadle V M , Obasanjo O, Moore RD. Self-reported antiretroviral therapy in injection drug users. JAMA 1998;280(6):544-6. 10 27. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999;48(12):1751-8. 28. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 29. Moore RD, Stanton D , Gopalan R, Chaisson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994;330(ll):763-8. 30. Karon JM, Fleming PL, Steketee RW, De Cock K M . HIV in the United States at the turn of the century: an epidemic in transition. Am J Public Health 2001;91(7):1060-8. 31. Hogg RS, Strathdee SA, Craib KJ, O'Shaughnessy M V , Montaner JS, Schechter M T . Lower socioeconomic status and shorter survival following HIV infection [see comments]. Lancet 1994;344(8930):1120-4. 32. Carpenter C C , Fischl M A , Hammer S M , et al. Antiretroviral therapy for H I V infection in 1996. Recommendations of an international panel. International AIDS Society-USA. JAMA 1996;276(2):146-54. 33. Strathdee SA, Patrick D M , Currie SL, et al. Needle exchange is not enough: lessons from the Vancouver injecting drug use study. AIDS 1997;ll(8):F59-65. 1 1 CHAPTER 2 BACKGROUND 2.1 F O R W A R D This literature review was invited by the editors of AIDS as a review paper for their annual supplement. The review was conducted by Evan Wood who also prepared the first draft, and then invited a number of experts to provide feedback and be co-authors on the manuscript. It is currently under review as: Wood E, Montaner JS, Bangsberg D , Tyndall M , Strathdee SA, . Bozzette S, O'Shaughnessy M V , H o g g RS. Access to antiretroviral therapy among marginalized populations in the developed world. AIDS: in press. Please note that, while this review is intended to provide an introduction to the thesis, it has been updated to reflect the current state of knowledge, contains some unpublished data, and in several instances refers to the manuscripts which make up this thesis. 1 2 CHAPTER 2: ACCESS TO ANTIRETROVIRAL THERAPY A M O N G MARGINALIZED POPULATIONS IN THE DEVELOPED WORLD 2.1 I N T R O D U C T I O N Since its introduction in the mid-1990's, the benefits of antiretroviral therapy for the management of HIV disease have been well established.I'2 New antiretroviral regimens have proven effective in decreasing H I V plasma viral load, improving CD4 cell counts, and have substantially altered the natural history of HIV infection.3/4 As a result, substantial improvements in HIV-related morbidity and mortality have been documented among persons receiving appropriate antiretroviral regimens, and in many areas of the world, HIV infection is increasingly being viewed as a chronic and manageable i l l n e s s . ^ Nevertheless, the clinical management of HIV disease continues to present major challenges. Treatment of HIV disease with the regimens that are presently available aims to prevent progression to AIDS or death by reducing plasma HIV R N A to as low a level as possible for as long as possible.^/8 The eradication of HIV from the individual is not considered possible with presently available therapeutic agents. As such, persons undergoing treatment for HIV disease must take a daily regimen of at least three antiretroviral drugs (i.e., highly active antiretroviral therapy, or H A A R T ) , and follow a scheduled dosing protocol that often involves coordination of dietary intake.^ To date, a great deal of effort has been expended in evaluating patient, physician, and healthcare delivery-related factors that may limit or enhance daily adherence to antiretroviral therapy.10-15 13 A related challenge, which is the focus of this review, involves the lesser-studied problem of access to antiretroviral therapy. While issues surrounding access to antiretroviral therapy are most pressing in the developing world, where HIV treatment remains prohibitively expensive,! 6 emerging data suggest that high levels of ongoing H I V / A I D S mortality persist in the developed world, in a large part due to limited use of H A A R T . 1 ^ While a proportion of these deaths are likely due to sub-optimal adherence to daily antiretroviral therapy,! ^ there is growing evidence that a high proportion of the ongoing AIDS mortality in the . developed world is due to poor access to therapy among disadvantaged populations. Limited access to H A A R T may be of particular concern for specific populations, and may involve treatment discontinuation and /or intermittent use of antiretrovirals, as well as premature mortality among those who never accessed HIV treatment prior to AIDS diagnosis or death. There is evidence that both of these concerns are common in developed world settings, even in countries where H I V / A I D S patients are eligible for antiretroviral therapy and medical care free of charge. 19,20 The following review will outline the evidence that suggests that limited retention in treatment and poor access to antiretroviral therapy are contributing to ongoing AIDS mortality in developed world settings. In addition, we will review the research to date on barriers to access and retention in HIV treatment, as well as strategies that may improve access to antiretrovirals among at-risk populations. 1 4 2.2 P O P U L A T I O N S A T R I S K O F P O O R A C C E S S T O H A A R T There is growing evidence that emerging disparities i n HIV-related mortality are due to problems w i t h access to antiretrovirals and retention i n treatment among specific sub-populations. These populations include groups that have traditionally been at risk of inferior access to health care such as persons of lower socio-economic status, ethnic minorities, and injection drug users (IDU).21-23 Limited access may stem from issues ranging from physician reluctance to prescribe H A A R T based on presumed inability to adhere to therapy to cultural barriers that may result in individuals refusing H A A R T when it is offered. A s w i l l be discussed below, among persons wi th H I V infection, female gender has also been consistently associated wi th poorer access to antiretroviral therapy. The role of socio-economic status i n H I V disease progression has been the focus of much study since it was first found to be an independent predictor of H I V disease progression prior to the advent of HAART.24 ,25 This f inding was consistent w i t h what has been observed for many other diseases,26 and subsequent studies suggested that differences may be due to poor access to H I V / A I D S care among lower income patients.27,28 A recent Canadian study has demonstrated substantially elevated mortal i ty after the ini t iat ion of antiretroviral therapy among lower income HIV-infected patients, although the difference was shown to be due to the higher prescription of dual therapy i n lower income individuals.29 In this study, the prescription of sub-optimal dual therapy persisted after adjustment for baseline clinical characteristics including 15 plasma H I V R N A and C D 4 cell count. Other studies have similarly demonstrated that persons of lower socio-economic status are less likely to be prescribed antiretroviral therapies.30,31 Ethnic minorities and females have also been shown to have poorer access to antiretroviral therapy. These concerns were first reported prior to the advent of H A A R T . In one U.S. study, it was demonstrated that AIDS mortality and opportunistic infections were elevated among African American patients, and that the differences were attributable to more advanced H I V disease at presentation for zidovudine therapy and less frequent use of PCP prophylaxis among this population.28 A subsequent study, demonstrated that African Americans were significantly less likely to have been prescribed antiretroviral therapy or PCP prophylaxis at the time of being referred to an HIV clinic.^ 2 With regard to gender, another U.S. study demonstrated that 58% of women delayed entry into specialized HIV care for > 3 months following an H I V diagnosis, and that upon entry 65% were symptomatic and 40% were severely immunocompromised.33 These ethnic and gender disparities have persisted since the introduction and widespread use of H A A R T . For instance, a large representative sample of HIV-infected adults from across the U.S. demonstrated that, while problems with access to HIV care have diminished over time, as late as 1998 fewer African Americans and women had started taking antiretroviral medication after adjustment for CD4 cell count.34 Similarly, another large national probability sample of H I V infected adults from the U.S. found that women, African 1 6 Americans, and the least educated were less likely to access H A A R T .30 These findings are consistent with the results from studies evaluating access at the state and city level.35-37 While social and cultural factors may explain some of the association between socio-economic status, race, gender, and poor access to therapy, there is little doubt that much of the association is due to the fact that in comparison to HIV-infected individuals with higher access to therapy, in particular gay and bisexual males,38 HIV-infected illicit drug users may be more likely to be women, ethnic minorities, and have lower incomes. In addition, those infected through sexual contact with IDU and/or through the sex-trade may be more likely to reside in lower income areas, to be female, and to be ethnic minorities. With regards injection drug use, numerous studies have shown drug use to be a major barrier to accessing antiretroviral therapy.30 For instance, a Swiss study demonstrated that active injection drug users outside of a drug treatment program, and those who acquired HIV infection through injection drug use had a significantly higher risk of inadequate treatment.39 A n Italian study that stratified patients by HIV exposure category found that IDU began pre-AIDS antiretroviral therapy significantly later than homosexual men and heterosexuals, and that the risk of disease progression was elevated among IDU within a cohort primarily using zidovudine.40 A subsequent Italian study found that patients with a history of injection drug use were significantly less likely to be prescribed protease inhibitors among a population eligible for antiretroviral therapy.41 Similarly, a French study demonstrated that despite regular access to 1 7 AIDS specialized hospital care, continued drug use was a major barrier to being prescribed antiretroviral treatment. 4 2 Similar findings have been reported from the U.S. where it has been found that HIV-infected IDU who were not receiving antiretrovirals have tended to be active drug users.36,43 Among injection drug users in Canada, younger age and female gender has been associated with being less likely to have ever received any antiretroviral therapy after adjustment for HIV R N A and CD4 cell c o u n t 4 4 2.3 A C C E S S T O A N D D I S C O N T I N U A T I O N O F H I V T R E A T M E N T Since most studies have been conducted among cohorts still living with HIV, it is presently not known to what extent apparent disparities in access to antiretrovirals will improve as HIV-disease becomes more advanced among those with lower access.!'7 For instance, there is evidence that differences in access between populations may diminish to some extent over time, 4^ and that some populations, particularly injection drug users, may initiate therapy at lower CD4 cell c o u n t s . 3 1 ' 4 6 ' 4 7 Nevertheless, there is emerging evidence to suggest that these disparities often persist until death. For instance, while the proportion of new HIV infections among African Americans has risen during the epidemic in the United States, African Americans did not account for greater than 50% of newly reported HIV infections until 1995, at which time 50.9% of the year's reported H I V infections were among African Americans. 4 ^ Therefore, assuming equivalent access to antiretrovirals by race, one might expect that the proportion 18 of new AIDS cases and AIDS deaths among African Americans would remain below 50% for several years to come. However, as of 2001, 76.3% of incident AIDS cases in the United States were among African Americans and 52% of AIDS deaths were among African Americans, suggesting that marked access disparities persist until AIDS and death.49 This conclusion is supported by research showing that while the numbers of AIDS deaths and AIDS diagnoses decreased dramatically in the U.S. with the advent of H A A R T , the proportional decreases in mortality were smallest among African Americans.20 These findings are not unique to the United States. For instance, in the province of British Columbia, Canada, where HIV antiretroviral therapy is available free of charge, one third of deaths attributable to HIV-infection during the period 1995 to 2001 occurred amongst persons who never accessed any HIV treatment.19 Preliminary analysis of these data suggest that both females and ethnic minorities are more likely to die without ever receiving antiretroviral treatment.19 In addition to complete lack of access to antiretroviral therapy prior to death, treatment discontinuation and 'intermittent' use of antiretrovirals have also become a growing concern. Studies have commonly identified both side effects of antiretroviral therapy and history of injection drug use to be associated with therapy discontinuation.50-52 A n acknowledged limitation of these studies stems from the fact that they are commonly based on selected samples of patients who return to clinics and/or fill out surveys, and population-level estimates of treatment discontinuation are less commonly available. However, evidence of 1 9 the high prevalence and negative impact of poor retention in treatment comes from population-based studies that have sought to derive population-level estimates of patient adherence by using prescription refill compliance as a surrogate. Several recent studies have applied this approach to estimate the impact of exposure to therapy on disease progression, and have demonstrated that 'adherence/ defined based on prescription refills, is strongly and independently associated with survival.53,54 p o r instance, among the 1422 treatment naive patients who initiated H A A R T in British Columbia, Canada since July 1996, 355 (25%) picked up medications < 75% of the time during their first year on therapy. Among those who picked up antiretrovirals < 75% of the time, there was nearly a three-fold increased risk of death compared to those who picked up antiretrovirals at least 75% of the time. Although further evaluation of these data are necessary, among the 193 (13.6%) deaths observed among this cohort of patients as of March 31 2002, 27 (14%) were among individuals who discontinued therapy within one month of initially starting H A A R T and did not reinitiate therapy again prior to death (BC Centre for Excellence in H I V / A I D S unpublished data). 2.4 BARRIERS T O T H E U S E O F A N T I R E T R O V I R A L T H E R A P Y The provision of healthcare services to marginalized populations has historically been compromised by an array of complex social dynamics. This is particularly the case for those addicted to drugs and alcohol, the mentally ill, and the homeless.55 Among the greatest challenges for the increasing population infected with HIV through injection drug use stems from the fact that, in most 20 countries around the world, the most commonly applied approach to reducing drug use-related problems is to impose criminal sanctions on those who use illicit drugs. This is demonstrated by a recent report which found that 20% to 26% of all people living with HIV in the United States in 1997 passed through a correctional facility that year.5 6 As a result of enforcement efforts in the community, those addicted to illicit drugs are often driven physically and. socially into environments where they are extremely difficult to reach for the purpose of providing medical services.57-60 A m o n g patients who are incarcerated, transition between prison and the community is often associated with interruptions in care and treatment. In addition to barriers stemming from criminal sanctions for illicit drug users/ a commonly reported barrier to accessing antiretrovirals may be physicians who are skeptical about injection drug users' ability to adhere to therapy, and it is likely that concerns regarding possible transmission of antiretroviral resistant virus has influenced prescribing decisions. 6 1 This is supported by previous surveys of physicians showing that physician's judgment of patient adherence is critical for the prescription of HAART.39,42 physicians have noted that patient homelessness, heavy alcohol use, injection drug use, and prior psychiatric hospitalization have all contributed to reluctance to initiate H A A R T . 6 2 These findings raise concerns for several reasons. First, a recent study demonstrated that a substantial proportion of homeless and marginally housed individuals managed high adherence to antiretroviral therapy including protease 2 1 inhibitors, and that resistance to Pis were rare among those that were non-adherent. 6 3 Furthermore, ethical analyses have suggested that physicians should not indefinitely withhold H A A R T from patients who are presumed to be poorly adherent.55/64,65 This argument is strengthened by the studies that have consistently demonstrated that providers may be poor judges of adherence. 6 6\" 6^ Other service delivery barriers to accessing antiretroviral therapy may range from geographic to cultural barriers. For instance, an earlier review suggested that geographic location may influence access to H A A R T , with more accessible locations possibly being associated with better access. 3 1 This is consistent with other studies that have found missing clinic visits to be a strong predictor of virological failure. 6^ Finally, the consistent associations between ethnic minority status and poor access suggest that cultural differences, as a result of language differences, perceived or real stigma, or other social barriers contribute to limited access.30,32,35-37 Finally, even when physicians are willing to prescribe antiretroviral therapy and no other barriers are in place, illicit drug use, mental illness, and homelessness may all present major barriers to accessing and retention in an HIV treatment program. 3 6'4 3/55,70 While each of these issues presents unique challenges, there are promising strategies that may help to address these concerns. 22 2.5 S T R A T E G I E S F O R I M P R O V I N G A C C E S S A N D R E T E N T I O N While interventions to improve daily adherence to antiretroviral therapy have received a growing amount of attention, 7 1 \" 7 3 limited data exist with regard to strategies that may optimize access and maximize retention in HIV treatment among populations that have traditionally had poor access to antiretroviral therapy. Nevertheless, through what is known about the significant barriers that presently exist, and what has been learned from experience with other chronic diseases, several strategies deserve further consideration. A m o n g the best sources of evidence for strategies that may improve access and retention in H I V / A I D S treatment come from interventions that have successfully delivered healthcare to other marginalized populations. Among adherence interventions, the highest rates of treatment retention and adherence to therapies have been observed with patients enrolled in directly observed therapy programs for the treatment of tuberculosis. 7 4 ' 7 ^ In one study, 77% of injection drug users took 100% of prescribed isoniazid therapy to prevent tuberculosis, a markedly higher adherence rate in comparison to those randomized to self-administered therapy regardless of counseling and education, although both groups were successfully retained in the study up to 6 months. 7^ However, doubts have been raised about the feasibility of directly observed therapy for HIV infection for several reasons, not the least of which is that it may require many years of supervised therapy. In addition, challenges presented by extending directly observed therapy to the treatment of HIV include the potentially negative impact on individual freedoms, as well as 23 concerns that patients may be reluctant to disclose toxicities and side effects out of fear of interrupting scheduled reimbursement / 6 ' ^ Furthermore, for the few studies of directly observed therapy for H I V infection that have been reported,^8'79 ^ n a s been noted that it is difficult to determine what effect direct supervision or other aspects of assisted adherence programs have had the largest effect.^ Nevertheless, the ability to retain patients in treatment and ensure high adherence suggests that the directly observed therapy model deserves further consideration. With regards to physician-patient relationships, while physician reluctance to prescribe antiretrovirals has been shown to be a barrier to access, it is important to note that certain physician-related characteristics have also been strongly associated with better access, adherence, and retention in treatment. For instance, satisfaction with one's physician has been associated with higher levels of adherence,^^ and willingness to initiate H A A R T has been associated with patient trust in their physician.80 In addition, among injection drug users, being treated by a physician with greater experience treating H I V disease was associated with accessing antiretroviral therapy.44,81 These findings demonstrate both that the physician-patient relationship is critical,^ 2 and that care for populations at risk of poor access is best delivered by highly experienced physicians, especially considering that patients at risk of poor access may face additional challenges such as H C V infection, mental illness, and addiction. In addition, studies have consistently shown that the provision of drug treatment services can enhance adherence and retention in HIV treatment for 24 illicit drug users. A m o n g the most well evaluated strategies has been the provision of methadone maintenance therapy (MMT) with antiretroviral therapy, which has been demonstrated to improve outcomes from H A A R T among HIV-infected IDU.83,84 As such, efforts should be made to remove barriers to the use of M M T among opiate dependent HIV-infected patients who are willing to initiate substitution therapy with M M T . ^ 5 However, there is sufficient evidence to demonstrate that wider access to methadone alone will not be sufficient to close that gap for many opiate dependent patients. First, many studies of patients successfully treated with antiretrovirals and M M T are often based on selected clinic-based samples, and both observational and randomized studies have demonstrated the high rates of loss to follow-up among patients initiating MMT.86,87 Although still controversial in many settings, heroin prescription programs appear to have substantial potential to improve retention and treatment outcomes among intractable opiate addicts.88,89 The randomized trial that has been proposed to take place in three Canadian cities does not exclude HIV-infected opiate addicts, and it will be interesting to see if differences emerge between the control arm (MMT) and prescribed heroin with regard to antiretroviral access and retention in HIV treatment.90 The growing number of H I V infected individuals who are using methamphetamine or cocaine may be even more difficult to reach with healthcare services, and present even greater challenges with regards to the delivery of H A A R T . 9 1 These challenges stem from the effects of the drugs themselves, which often lead to chaotic behavior, as well as the higher frequency 2 5 of injections and lack of pharmacologic replacement as in the case of heroin/methadone.92 For these individuals, efforts should be made by the physician to help stabilize them when possible,55/82 a n c \ there is also evidence that meeting users on their own turf has substantial potential. For instance, a novel project initiated in San Francisco runs out of a storefront office in the heart of a high HIV prevalence inner city neighbor hood.93 The program offers both medical and social services, several of which have been shown to improve access to antiretrovirals, including HIV specialist nurses, access to medical services without appointment, and an onsite pharmacist.30,43,94 Interestingly, the project also offers harm reduction services including needle exchange, which may help to attract eligible patients into their other programs including drug treatment and medical care.95 Among HIV infected individuals retained in the program, 64% achieved HIV R N A < 500 c/ mL, and only 2 patients had shown an increase in HIV R N A relative to their pre-program levels.93 Programs that seek to improve access to healthcare by reducing socio-cultural and geographic barriers are an important area for further study.96 2.6 S U M M A R Y We now have a great deal of evidence demonstrating that lower income populations, women, ethnic minorities, and illicit drug users are at risk of poor access to antiretroviral therapy and higher rates of treatment discontinuation. While the majority of these studies have been conducted among living cohorts, 26 data is emerging to suggest that limited access is contributing to the ongoing H I V / A I D S mortality ratesin the developed world. Fears regarding the transmission of antiretroviral resistant HIV among the homeless, mentally i l l , and injection drug users have largely been u n f o u n d e d . 5 5 ' 6 1 ^ 6 3 ' 9 7 While this concern deserves continued monitoring, in many instances it is likely that resistance is not observed because a proportion of these patients will cease antiretroviral therapy outright or will be insufficiently adherent for resistance to develop . 6 3 These issues have several implications. First, interventions to improve retention among those who initiate therapy are urgently required, and the barriers and strategies described above deserve immediate attention. Second, among individuals who have not accessed therapy, strategies to improve contact with HIV care providers are needed. When contact is made, guidelines for physicians must be based on available evidence. 6 6 \" 6 ^ This evidence suggests that physicians should seek to address modifiable barriers to adherence and retention in HIV treatment prior to the start of therapy among patients not requiring immediate treatment. 5 5 ' 6 5 In addition, it suggests that no patient should be denied the opportunity to initiate H A A R T regardless of perceived or real barriers to optimal adherence including continued illicit drug use. Given the extent of the public health crisis, and since the full benefits of H A A R T are not compromised when patients are successfully retained in treatment, regardless of ethnicity, gender, and history of injection drug use, 5 2 '98 efforts to address social, cultural, and medical barriers are an urgent priority. 27 2.6 R E F E R E N C E S 1. Palella FJ, Jr., Delaney K M , Moorman A C , et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators [see comments]. New England Journal of Medicine 1998;338(13):853-60. 2. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 3. Wood E, Yip B, Hogg RS, et al. Full suppression of viral load is needed to achieve an optimal CD4 cell count response among patients on triple drug antiretroviral therapy. AIDS 2000;14(13):1955-60. 4. Moore R D , Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS 1999;13(14):1933-42. 5. Ledergerber B, Egger M , Opravil M , et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss H I V Cohort Study. Lancet 1999;353(9156):863-8. 6. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 7. Yeni PG, Hammer S M , Carpenter C C , et al. Antiretroviral treatment for adult HIV infection in 2002: updated recommendations of the International AIDS Society-USA Panel. JAMA 2002;288(2):222-35. 8. Chun TW, Fauci AS. Latent reservoirs of HIV: obstacles to the eradication of virus. Proc Natl Acad Sci USA 1999;96(20):10958-61. 28 9. Ickovics JR, Meade CS. Adherence to H A A R T among patients with HIV: breakthroughs and barriers. AIDS Care 2002;14(3):309-18. 10. Bangsberg DR, Bronstone A , Hofmann R. A computer-based assessment detects regimen misunderstandings and nonadherence for patients on HIV antiretroviral therapy. AIDS Care 2002;14(1):3-15. 11. Catz SL, Kelly JA, Bogart L M , Benotsch E G , McAuliffe T L . Patterns, correlates, and barriers to medication adherence among persons prescribed new treatments for HTV disease. Health Psychol 2000;19(2):124-33. 12. Chesney M A , Ickovics JR, Chambers DB, et al. Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: the A A C T G adherence instruments. Patient Care Committee & Adherence Working Group of the Outcomes Committee of the Adult AIDS Clinical Trials Group (AACTG). AIDS Care 2000;12(3):255-66. 13. Gifford A L , Bormann JE, Shively MJ, Wright BC, Richman D D , Bozzette SA. Predictors of self-reported adherence and plasma HIV concentrations in patients on multidrug antiretroviral regimens. / Acquir Immune Defic Syndr 2000;23(5):386-95. 14. Roberts KJ. Physician-patient relationships, patient satisfaction, and antiretroviral medication Adherence among HIV-infected adults attending a public health clinic. AIDS Patient Care STDS 2002;16(l):43-50. 15. Murphy D A , Roberts KJ, Martin DJ, Marelich W, Hoffman D . Barriers to antiretroviral adherence among HIV-infected adults. AIDS Patient Care STDS 2000;14(l):47-58. 16. Wood E, Braitstein P, Montaner JS, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa [see comments]. Lancet 2000;355(9221):2095-100. 29 17. Wood E, Schechter M T , Tyndall M W , Montaner JS, O'Shaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000;14(9):1229-35. 18. Paterson D L , Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000;133(l):21-30. 19. Wood E, Montaner JS, Tyndall M W , Schechter M T , O'Shaughnessy M V , Hogg RS. Treatment access and H I V / A I D S deaths in the era of 'Chronic' infection, in a setting that delivers anitretrovirals and AIDS care free of charge. XIV International AIDS Conference. Barcelona, Spain July 7-12, 2002. [TuPeC4753]. 2002. 20. Karon JM, Fleming PL, Steketee RW, De Cock K M . HIV in the United States at the turn of the century: an epidemic in transition. Am J Public Health 2001;91(7):1060-8. 21. Fein O. The influence of social class on health status: American and British research on health inequalities. / Gen Intern Med 1995;10(10):577-86. 22. Montgomery LE , Kiely JL, Pappas G. The effects of poverty, race, and family structure on US children's health: data from the NHIS, 1978 through 1980 and 1989 through 1991. Am J Public Health 1996;86(10):1401-5. 23. Pappas G , Queen S, Hadden W, Fisher G . The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986. N Engl J Med 1993;329(2):103-9. 24. Schechter M T , Hogg RS, Aylward B, Craib KJ, Le T N , Montaner JS. Higher socioeconomic status is associated with slower progression of HIV infection independent of access to health care. / Clin Epidemiol 1994;47(l):59-67. 30 25. Hogg RS, Strathdee SA, Craib KJ, O'Shaughnessy M V , Montaner JS, Schechter M T . Lower socioeconomic status and shorter survival following HIV infection [see comments]. Lancet 1994;344(8930):1120-4. 26. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999;48(12):1751-8. 27. Chaisson RE, Keruly JC, Moore R D . Race, sex, drug use, and progression of human immunodeficiency virus disease. N Engl J Med 1995;333(12):751-6. 28. Easterbrook PJ, Keruly JC, Creagh-Kirk T, Richman D D , Chaisson RE, Moore R D . Racial and ethnic differences in outcome in zidovudine-treated patients with advanced HIV disease. Zidovudine Epidemiology Study Group. JAMA 1991;266(19):2713-8. 29. Wood E, Montaner JS, Chan K, et al. Socioeconomic status, access to triple therpay, and survival from HIV-disease since 1996. AIDS 2002;16(15):2065-2072. 30. Andersen R, Bozzette S, Shapiro M , et al. Access of vulnerable groups to antiretroviral therapy among persons in care for HIV disease in the United States. H C S U S Consortium. HIV Cost and Services Utilization Study. Health Sew Res 2000;35(2):389-416. 31. McKinney M M , Marconi K M . Delivering H I V services to vulnerable populations: a review of C A R E Act-funded research. Public Health Rep 2002;117(2):99-113. 32. Moore RD, Stanton D , Gopalan R, Chaisson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994;330(ll):763-8. 3 1 33. Ickovics JR, Forsyth B, Ethier K A , Harris P, Rodin J. Delayed entry into health care for women with HIV disease. AIDS Patient Care STDS 1996;10(l):21-4. 34. Shapiro M F , Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States: results from the H I V Cost and Services Utilization Study. JAMA 1999;281(24):2305-15. 35. Anderson K H , Mitchell JM. Differential access in the receipt of antiretroviral drugs for the treatment of AIDS and its implications for survival. Arch Intern Med 2000;160(20):3114-20. 36. Cook JA, Cohen M H , Grey D , et al. Use of highly active antiretroviral therapy in a cohort of HIV-seropositive women. Am J Public Health 2002;92(l):82-7. 37. H s u L C , Vittinghoff E, Katz M H , Schwarcz SK. Predictors of use of highly active antiretroviral therapy (HAART) among persons with AIDS in San Francisco, 1996-1999. J Acquir Immune Defic Syndr 2001;28(4):345-50. 38. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver's West End. Can J Public Health 2000;91(2):125-8. 39. Bassetti S, Battegay M , Furrer H , et al. Why is highly active antiretroviral therapy ( H A A R T ) not prescribed or discontinued? Swiss H I V Cohort Study. / Acquir Immune Defic Syndr 1999;21(2):114-9. 40. Dorrucci M , Pezzotti P, Phillips A N , Alliegro M B , Rezza G . Antiretroviral treatment and progression to AIDS in HIV seroconverters from different risk groups. HIV Italian Seroconversion Study. AIDS 1997;ll(4):461-7. 41. Murri R, Fantoni M , Del Borgo C, et al. Intravenous drug use, relationship with providers, and stage of HIV disease influence the prescription rates of protease inhibitors. / Acquir Immune Defic Syndr 1999;22(5):461-6. 3 2 42. Carrieri M P , Moatti JP, Vlahov D , Obadia Y, Reynaud-Maurupt C , Chesney M . Access to antiretroviral treatment among French HIV infected injection drug users: the influence of continued drug use. M A N I F 2000 Study Group. / Epidemiol Community Health 1999;53(l):4-8. 43. Celentano D D , Vlahov D , Cohn S, Shadle V M , Obasanjo O, Moore R D . Self-reported antiretroviral therapy in injection drug users. JAMA 1998;280(6):544-6. 44. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 45. Fairfield K M , Libman H , Davis RB, Eisenberg D M . Delays in protease inhibitor use in clinical practice. / Gen Intern Med 1999;14(7):395-401. 46. Kahn JG, Zhang X, Cross LT, Palacio H , Birkhead GS, Morin SF. Access to and use of HIV antiretroviral therapy: variation by race/ethnicity in two public insurance programs in the U.S. Public Health Rep 2002;117(3):252-62. 47. Celentano D D , Galai N , Sethi A K , et al. Time to initiating highly active antiretroviral therapy among HIV-infected injection drug users. Aids 2001;15(13):1707-15. 48. Centers for Disease Control and Prevention H I V / A I D S Surveillance Report 1995;7 (no. 2).. 49. Centers for Disease Control and Prevention H I V / A I D S Surveillance Report 2001;13 (no. 2).. 50. Ahdieh Grant L, Silverberg MJ, Palacio H , et al. Discontinuation of potent antiretroviral therapy: predictive value of and impact on CD4 cell counts and HIV R N A levels. AIDS 2001;15(16):2101-8. 3 3 51. d'Arminio Monforte A , Lepri A C , Rezza G , et al. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients. I . C O . N . A . Study G r o u p . Italian Cohort of Antiretroviral -Naive Patients. AIDS 2000;14(5):499-507. 52. Palepu A , Yip B, Miller C , et al. Factors associated with the response to antiretroviral therapy among HIV-infected patients with and without a history of injection drug use. AIDS 2001;15:423-424. 53. Hogg RS, Heath K V , Bangsberg D , et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up. AIDS 2002;16(7):1051-1058. 54. Wood E, Hogg RS, Yip B, Harrigan RP, O'Shaughnessy M V , Montaner JS. The impact of baseline CD4 T-cell count and adherence on rates of disease progression among HIV-infected patients initiating triple drug antiretroviral therapy. AIDS 2002;in press. 55. Bangsberg D , Tulsky JP, Hecht F M , Moss A R . Protease inhibitors in the homeless. JAMA 1997;278(l):63-5. 56. Hammett T M , Harmon M P , Rhodes W. The burden of infectious disease among inmates of and releasees from US correctional facilities, 1997. Am J Public Health 2002;92(ll):1789-94. 57. Buchanan D , Khoshnood K, Stopka T, Shaw S, Santelices C , Singer M . Ethical dilemmas created by the criminalization of status behaviors: case examples from ethnographic field research with injection drug users. Health Educ Behav 2002;29(l):30-42. 58. Grund JP, Blanken P, Adriaans N F , Kaplan C D , Barendregt C , Meeuwsen M . Reaching the unreached: targeting hidden IDU populations with clean needles via known user groups. J Psychoactive Drugs 1992;24(l):41-7. 34 59. Bluthenthal, R.N. , Lorvick, J., Krai, A . H . , et al., \"Collateral damage in the war on drugs: HIV risk behaviours among injection drug users,\" International Journal of Drug Policy, v. 10, (1999), pp. 25-38. 60. Bluthenthal R N , Krai A H , Lorvick J, Watters JK. Impact of law enforcement on syringe exchange programs: a look at Oakland and San Francisco. Med Anthropol 1997;18(l):61-83. 61. Wainberg M A , Friedland G . Public health implications of antiretroviral therapy and HIV drug resistance. JAMA 1998;279(24):1977-83. 62. Bogart L M , Kelly JA, Catz SL, Sosman JM. Impact of medical and nonmedical factors on physician decision making for H I V / A I D S antiretroviral treatment. / Acquir Immune Defic Syndr 2000;23(5):396-404. 63. Bangsberg DR, Hecht F M , Charlebois E D , et al. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS 2000;14(4):357-66. 64. Sollitto S, Mehlman M , Youngner S, Lederman M M . Should physicians withhold highly active antiretroviral therapies from HIV-AIDS patients who are thought to be poorly adherent to treatment? Aids 2001;15(2):153-9. 65. Bangsberg DR, Moss A . When should we delay highly active antiretroviral therapy? / Gen Intern Med 1999;14(7):446-8. 66. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 67. Miller L G , Liu H , Hays RD, et al. H o w well do clinicians estimate patients' adherence to combination antiretroviral therapy? / Gen Intern Med 2002;17(1):1-11. 3 5 68. Gross R, Bilker WB, Friedman H M , Coyne JC, Strom BL. Provider inaccuracy in assessing adherence and outcomes with newly initiated antiretroviral therapy. AIDS 2002;16(13):1835-7. 69. Lucas G M , Chaisson RE, Moore RD. Highly active antiretroviral therapy in a large urban clinic: risk factors for virologic failure and adverse drug reactions. Ann Intern Med 1999;131(2):81-7. 70. Gallego Deike L, Gordillo Alvarez-Valdes M V . Mental disorders in patients infected with the human immunodeficiency virus. An Med Interna 2001;18(ll):597-604. 71. Safren SA, Otto M W , Worth JL, et al. Two strategies to increase adherence to H I V antiretroviral medication: life-steps and medication monitoring. Behav Res Ther 2001;39(10):1151-62. 72. Murphy D A , L u M C , Martin D, Hoffman D , Marelich W D . Results of a pilot intervention trial to improve antiretroviral adherence among HIV-positive patients. / Assoc Nurses AIDS Care 2002;13(6):57-69. 73. Broadhead RS, Heckathorn D D , Altice FL, et al. Increasing drug users' adherence to H I V treatment: results of a peer-driven intervention feasibility study. Soc Sci Med 2002;55(2):235-46. 74. Frieden TR, Fujiwara PI, Washko R M , Hamburg M A . Tuberculosis in New York City-turning the tide. N Engl J Med 1995;333(4):229-33. 75. Chaisson RE, Barnes G L , Hackman J, et al. A randomized, controlled trial of interventions to improve adherence to isoniazid therapy to prevent tuberculosis in injection drug users. Am J Med 2001;110(8):610-5. 76. Bangsberg DR, Mundy L M , Tulsky JP. Expanding directly observed therapy: tuberculosis to human immunodeficiency virus. Am J Med 2001;110(8):664-6. 36 77. Liechty C, Bangsberg D . Doubts about D O T : Antiretroviral therapy for resource poor settings. AIDS (In Press) 2002. 78. Mitty JA, Stone V E , Sands M , Macalino G , Flanigan T. Directly observed therapy for the treatment of people with human immunodeficiency virus infection: a work in progress. Clin Infect Dis 2002;34(7):984-90. 79. Fischl M , Rodriguez A , Scerpella E, Mondroig R, Thompson L, Rechtine D. Directly observed therapy on outcomes in H I V clinical trials. 7th Conference on Retroviruses and Opportunistic Infections (CROI). San Francisco, 2000.. 80. Altice FL, Mostashari F, Friedland G H . Trust and the acceptance of and adherence to antiretroviral therapy. / Acquir Immune Defic Syndr 2001;28(l):47-58. 81. Gardner LI, Holmberg SD, Moore J, et al. Use of highly active antiretroviral therapy in HIV-infected women: impact of HIV specialist care. / Acquir Immune Defic Syndr 2002;29(l):69-75. 82. McCally M , Haines A , Fein O, Addington W, Lawrence RS, Cassel C K . Poverty and ill health: physicians can, and should, make a difference. Ann Intern Med 1998;129(9):726-33. 83. Antela A , Casado JL, Gonzalez MJ, et al. Influence of a methadone maintenance programme on the improved outcome of a cohort on injecting drug users with advanced HIV disease. AIDS 1997;ll(ll):1405-6. 84. Moreno A , Perez-Elias MJ, Casado JL, et al. Long-term outcomes of protease inhibitor-based therapy in antiretroviral treatment-naive HIV-infected injection drug users on methadone maintenance programmes. AIDS 2001;15(8):1068-70. 37 85. Wenger L D , Rosenbaum M . Drug treatment on demand-not. / Psychoactive Drugs 1994;26(1):1-11. 86. Johnson RE, Chutuape M A , Strain E C , Walsh SL, Stitzer M L , Bigelow G E . A comparison of levomethadyl acetate, buprenorphine, and methadone for opioid dependence. N Engl J Med 2000;343(18):1290-7. 87. Mino A , Page D , Dumont P, Broers B. Treatment failure and methadone dose in a public methadone maintenance treatment programme in Geneva. Drug Alcohol Depend 1998;50(3):233-9. 88. Perneger T V , Giner F, del Rio M , Mino A . Randomised trial of heroin maintenance programme for addicts who fail in conventional drug treatments. BMJ 1998;317(7150):13-8. 89. Rehm J, Gschwend P, Steffen T, Gutzwiller F, Dobler-Mikola A , Uchtenhagen A . Feasibility, safety, and efficacy of injectable heroin prescription for refractory opioid addicts: a follow-up study. Lancet 2001;358(9291):1417-23. 90. Brissette S. Medical prescription of heroin\u00E2\u0080\u0094a review. Can HIV AIDS Policy Law Rev 2001;6(1-2):1,92-8. 91. Tyndall M W , Craib KJ, Currie S, L i K, O'Shaughnessy M V , Schechter M T . Impact of HIV infection on mortality in a cohort of injection drug users. / Acquir Immune Defic Syndr 2001;28(4):351-7. 92. Tyndall M W , Spittal P M , Laliberte NJ, L i K, O'Shaughnessy M V , Schechter M T . Intensive injection cocaine use as a primary risk factor of H I V seroconversion among polydrug users in Vancouver. AIDS 2002;In Press. 93. Bamberger JD, Unick J, Klein P, Fraser M , Chesney M , Katz M H . Helping the urban poor stay with antiretroviral HIV drug therapy. Am J Public Health 2000;90(5):699-701. 38 94. Sherer R, Stieglitz K, Narra J, et al. HIV multidisciplinary teams work: support services improve access to and retention in HIV primary care. AIDS Care 2002;14(l):S31-44. 95. Riley E D , W u A W , Junge B, Marx M , Strathdee SA, Vlahov D . Health services utilization by injection drug users participating in a needle exchange program. Am J Drug Alcohol Abuse 2002;28(3):497-511. 96. Wood E, Kerr T, Spittal P M , et al. The potential public health and community impacts of safer injecting facilities: Evidence from a cohort of injection drug users. JAIDS (In Press) 2003. 97. Alexander CS, Dong W, Schechter M T , et al. Prevalence of primary HIV drug resistance among seroconverters during an explosive outbreak of HIV infection among injecting drug users. Aids 1999;13(8):981-5. 98. Jensen-Fangel S, Pedersen L, Pedersen C , et al. The effect of race/ethnicity on the outcome of highly active antiretroviral therapy for human immunodeficiency virus type 1-infected patients. Clin Infect Dis 2002;35(12):1541-8. 3 9 CHAPTER 3: STUDY SETTING, OVERVIEW OF M E T H O D S , A N D LIMITATIONS 3.1 H I V E P I D E M I O L O G Y A N D T R E A T M E N T IN BRITISH C O L U M B I A British Columbia, most notably Vancouver's West End, has long been home to a large population of gay and bisexual males,! and in the province cases of Pneumocycstis carinii pneumonia and Kaposi's sarcoma were first observed among gay males in this area soon after the initial reports from the United States.2 Among injection drug users (IDU), extensive transmission of HIV began in the early 1990's,3 and the first major spike in IDU presenting for H I V treatment was reported in 1995.4 Since that time, access to antiretrovirals among IDU has been a major concern in the province. 5 In British Columbia, all anti-HIV medications have been centrally distributed at no cost to eligible HIV-infected individuals since 1986, and in October 1992 antiretroviral provision became the responsibility of the British Columbia Centre for Excellence in H I V / A I D S ' Drug Treatment Program (the Centre). In the province, all H I V / A I D S care is provided free of charge through a universal healthcare system. The Centre distributes antiretroviral medications based on specific guidelines generated by a Therapeutic Guidelines Committee. In 1992, the guidelines recommended dual nucleoside antiretroviral therapy for people with 3 CD4 cell counts below 350 cells/mm . In December 1995 dual therapy was made 3 available to everyone with a CD4 cell count below 500 cells/mm . The Centre 4 0 adopted plasma viral load driven antiretroviral therapy guidelines in June 1996, consistent with those put forward by the International AIDS Society - U S A . 6 In British Columbia, zidovudine has been available since 1986. The other four nucleoside analogues were made available over a period of four years: didanosine and zalcitabine in 1992, stavudine in 1993, and lamivudine in 1994. Saquinavir, indinavir, and ritonavir have been available since 1996, and nelfinavir has been available since 1998. Delavirdine and nevirapine were made' available in Canada in 1998 and efavirenz became available a year later in 1999. The Centre's D r u g Treatment Program remains the only source of free antiretroviral medications in the province of British Columbia and < 1% of HIV-infected British Columbians purchase antiretroviral therapy outside the program. 5 Physicians enrolling an HIV-positive individual into the Centre's Drug Treatment Program must complete a drug request enrollment form. The enrollment form acts as a legal prescription and provides information on the HIV-positive applicant's address and enrolling physician, past HIV-specific drug history, CD4 cell counts, and current drug requests. Approved prescriptions are renewed every two months. At the time of the initial refill, each participant is asked to complete an enrollment survey and informed consent form, and their physician is asked to complete a clinical staging form. Thereafter, participant surveys and clinical staging forms are completed annually. 4 1 3.2 M E T H O D O L O G I C A L APPROACHES A N D D A T A SOURCES Each manuscript presented in chapters 4 through 9 applied similar statistical techniques to fulfil the objectives laid out in Chapter 1, and the methods employed in each of these studies is described in detail in each of the chapters. However, several methodological approaches used in this thesis augmented the existing tools used by researchers at the BC Centre for Excellence in H I V / A I D S , and these methods will be described in detail below. 3.2 (i): SOCIO-ECONOMIC STATUS Among the unique contributions of this thesis that I initiated at the Centre was to employ geographic information systems tools to augment the existing research activities at the Centre. For instance, a primary variable of interest in several of the sub-studies in this thesis was socio-economic status. However, socio-economic status is not routinely and universally collected as part of the Centre's follow-up of HIV-infected patients. Therefore, in order to derive a measure of socioeconomic status for each HIV-infected individual in the province who was eligible for the studies presented in chapters 4 and 7, socio-economic status was assessed through the use of census data. As has been done by others, 7 individuals were linked to census-based socio-economic data by examining each patient's home residence postal code at the time of their death (Chapter 4), or at their time of enrolment to the treatment program (Chapter 7). Statistics Canada strives to collect socio-demographic data for all Canadians every five years, and then makes public data that have been compiled for census geographic areas, such as census tracts, and 42 higher levels of agglomeration, such as census sub-divisions. Statistics Canada defines census tracts as small geographic units representing urban or rural neighborhood-like areas. The Vancouver Census Metropolitan Area, for instance, consists of 285 Census Tracts. In the studies described below, the census geographic area in which each treatment program participant's home residence was located was identified using Statistics Canada's Postal Code Conversion File. In short, this file links postal codes to Census data. This file enabled the median income of each neighborhood in the province to be linked to each HIV-infected individual as described in Chapters 4 and 7. For example, if an individual resided in a neighborhood with a median annual income of $20,000 that individual was assigned this value as an estimate of socio-economic status. In addition, in Chapter 5, the Statistics Canada data was also used to derive estimates of the population stratified by age and gender that reside in Vancouver's Downtown Eastside. 3.2 (ii): MORTALITY D A T A A s noted above, the BC Centre for Excellence is the only source of free antiretroviral therapy in the province, and previous studies have shown that virtually all individuals on treatment access therapy through the program.^ A l l deaths from H I V / A I D S in the province were determined through a linkage with the province's death registry housed at the British Columbia Ministry of Health's Vital Statistics Agency in Victoria. 4 3 Chapter 4 is based upon a linkage I conducted with Vital Statistics. Specifically, I obtained a file from Vital Statistics, who queried their database to extract records for all deaths that contained an ICD-9 (1995-99) or ICD-10 (2000-1) code indicating that HIV met the criteria of an underlying or associated cause of death during the years 1995 through 2001. Vital Statistics defines a death as 'HIV underlying' if the attending physician and/or coroner determine that the death was directly due to HIV infection. Conversely, Vital Statistics defines a death as 'HIV associated' if the attending physician and /or coroner determine that HIV infection was a condition which contributed to the individual's underlying cause of death. Persons who died of accidental causes, such as illicit drug overdoses, suicides, and accidents are not routinely tested for HIV in the province, and unless an individual is experiencing morbidity due to confirmed HIV infection their death should not be coded as HIV associated.^ I was able to determine the proportion of persons dying without ever accessing antiretrovirals by linking the file I obtained back to the Centre's H I V / A I D S Treatment Program. This linkage was ethically approved (see Appendix 2). In effect, for the purposes of this thesis, I assembled the cohort of HIV infected individuals who have died since 1995, and then using methods described above, I evaluated the impact of socio-economic status on access to therapy as well as its association with treatment discontinuation. 3.2 (iii): DEMOGRAPHIC PROJECTION M O D E L S In Chapters 5 and 6, the potential impact of H I V infection and antiretroviral use on life expectancy and AIDS deaths was estimated using 4 4 standard demographic population projection techniques. In addition, in Chapter 6, the use of antiretroviral prophylaxis for the prevention of mother-to-child transmission of HIV was also evaluated. While the populations modeled were different in each case, the methods employed were similar in both chapters, and were based on a methodology that employed DemProj\u00E2\u0084\u00A2 demographic modeling software's AIDS impact model. Specifically, like any traditional population projection, the models required information on a number of demographic parameters including population size, fertility, migration, and mortality rates. In addition, estimates of HIV prevalence by age and gender. Finally, the average time from HIV infection to AIDS and from AIDS to death was also estimated based on previous studies.9/10 This technique had been previously described.H/12 A more detailed description of the specific methods and the sources of data for each population projection are provided in Chapters 5 and 6. 3.2 (iv): T R E A T M E N T DISCONTINUATION A N D ADHERENCE ESTIMATES In Chapters 4, 7, 8, and 9 the impact of H I V / A I D S Treatment Program participants' exposure to antiretrovirals was estimated in several different ways using prescription refill compliance. 1 3 Specifically, in Chapters 4 and 9 the issue of treatment discontinuation was of particular interest, and based on the findings of earlier work, 1 4 treatment discontinuation was defined as the first of a > 3 month period without receiving antiretroviral therapy. 4 5 In Chapter, 7 adherence was identified as a potential source of confounding. Therefore, to adjust for this concern, prescription refill compliance was treated as a continuous variable based on the ratio of time that medication dispensed would last as a proportion of follow-up time. This calculation was restricted to each patient's first year on therapy to avoid reverse causation that could occur among patients who cease antiretroviral therapy after they have become too sick to take medication. 1^ Finally, in Chapter 8 patients were non-adherent if they received antiretroviral medications < 75% of the time during the first year of follow-up, since we sought to be conservative in our estimates, and this level of adherence is well below what is presently recommended by therapeutic guidelines. The decision to use this dichotomy was also based on an earlier analysis, which found that picking up less than 75% of prescriptions during the first year of follow-up was predictive of shorter survival . 1 4 3.3 S T A T I S T I C A L M E T H O D S Standard statistical techniques were used in each study, with the exception of Chapters 5 and 6, in which demographic population projection methods were employed. In Chapters 4, 7, 8, 9, both cross sectional and proportional hazards analyses were used. For the cross sectional analyses, categorical explanatory variables were analyzed using Pearson's Chi-square test and continuous variables were analyzed using the Wilcoxon rank sum test. Fisher's exact test was used for 2X2 contingency tables in which any of the 46 expected cell frequencies was less than 5. To adjust for potential confounders and derive adjusted estimates of effect, logistic models were then fit. Unless otherwise specified, the a priori determined model building protocol was to adjust for all variables associated with the outcome in univariate analyses. For the longitudinal analyses that evaluated the time to an event, such as death or treatment cessation, cumulative event rates or the product limit estimate was derived using Kaplan-Meier methods, and survival curves were compared between groups with the log-rank test. In addition, Cox proportional hazards regression was used to calculate univariate and adjusted relative hazards and 95% confidence intervals (CIs). 1 6 In each case, the assumption of proportional hazards was validated by inspection of log (-log [survival function]) estimates against log time plots. In addition, in Chapters 7 and 8 fixed models were fit using indicator variables to represent patients with differing clinical or behavioral characteristics. For instance, indicator variables were used to in order to derive adjusted relative hazards of mortality among adherent and non-adherent patients in each CD4 cell count strata. In this way, the role of adherence in disease progression could be estimated among patients with similar baseline CD4 cell counts while adjusting for relevant covariates. A l l statistical analyses were performed using SAS software version 6.0 (SAS, Cary, NC) . A l l tests of significance were two-sided, with a p-value of less than 0.05 indicating that an association was statistically significant. A more 47 detailed description of the specific statistical methods employed in each manuscript can be found in the relevant chapters. 3.4 S T R E N G T H S A N D L I M I T A T I O N S Each study that makes up this thesis has unique strengths and limitations that will be addressed in detail in each chapter. Nevertheless, several of these issues should be examined at the outset. First, it is important to stress that the data arose in a setting where all H I V / A I D S care, antiretrovirals, and laboratory monitoring are available free of charge, and where previous studies have shown that virtually all patients acquire antiretrovirals through a centralized source.^ Second, the centralized death registry enabled complete population-level data on H I V / A I D S deaths for the entire province.!- 7 As such, the interpretability of the findings presented here are not compromised by selection factors that may bias other cohort studies. Finally, since a complete prospective record of antiretroviral dispensation was maintained, it was possible to determine precisely each individual's level of treatment receipt. Second, although previous studies have suggested that ecological measures may result in an underestimation of the true relationship between socio-economic status and health outcomes,1^ it is important to stress that there was potential for misclassification because an ecological measure of socioeconomic status was used in Chapters 4 and 7. This was an unavoidable limitation and where appropriate in the manuscripts this concern was stressed. 48 Third, it is also critical to note that when developing population projections a number of parameters, such as future HIV prevalence, must be estimated, and as a result output values from the models are also only estimates. Therefore it should also be stressed that the models and scenarios that were developed in Chapters 5 and 6 are only hypothetical and a number of interventions could result in markedly different outcomes. Finally, previous studies have demonstrated that measuring daily adherence or other measures of exposure to antiretroviral therapy can be fraught with difficulties that may over or underestimate a patient's actual adherence to treatment.19-22 Although using refill compliance as a surrogate for adherence has been previously validated, 1 - 3 there is likely a strong conservative bias operating in instances where refill compliance was used to estimate adherence because patients may have been less than optimally adherent to daily treatment despite consistent refill of prescriptions. The specific implications of this concern are addressed in each chapter where relevant. 3.5 S U M M A R Y In summary, the methods employed in this thesis are diverse and rely on a number of data sources and required a number of linkages. Through modeling exercises and the statistical evaluation of quantitative data these studies strive to demonstrate the likely public health impacts that non-adherence and limited access to antiretrovirals will have as the HIV epidemic evolves if present levels of antiretroviral use continue. While each of the approaches applied in the various 4 9 sub-studies has individual strengths and weaknesses, together the combined studies provide insight into both what has happened with regards to access and adherence since the widespread use of antiretroviral therapy. The unique contribution that these studies have made will be described in detail in Chapter 10. The next chapter will begin with an evaluation of the prevalence of not receiving antiretroviral treatment among persons who have died since 1995, and an examination of what factors are associated with receipt of antiretrovirals prior to death in the province of British Columbia. 50 3.6 REFERENCES 1. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver's West End. Can J Public Health 2000;91(2):125-8. 2. Schechter M T , Boyko WJ, Douglas B, et al. The Vancouver Lymphadenopathy-AIDS Study: 6. HIV seroconversion in a cohort of homosexual men. CMA} 1986;135(12):1355-60. 3. Strathdee SA, Patrick D M , Currie SL, et al. Needle exchange is not enough: lessons from the Vancouver injecting drug use study. AIDS 1997;11(8):F59-65. 4. Montaner JS, Hogg RS, Craib KJ, Strathdee SA, O'Shaughnessy M V , Schechter M T . IDU numbers and AIDS. BC Medical Journal 1995;37(4):239. 5. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 6. Carpenter C C , Fischl M A , Hammer S M , et al. Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. International AIDS Society-USA. JAMA 1996;276(2):146-54. 7. Katz M H , Hsu L, Lingo M , Woelffer G , Schwarcz SK. Impact of socioeconomic status on survival with AIDS. Am J Epidemiol 1998;148(3):282-91. 8. Miller C L , Chan KJ, Palepu A , et al. Socio-Demographic Profile and HIV and Hepatitis C Prevalence Among Persons who Died of a Drug Overdose. Addict Res and Theory 2001;9(5):459-470. 9. Bacchetti P, Moss A R . Incubation period of AIDS in San Francisco. Nature 1989;338(6212):251-3. 5 1 10. Chevret S, Costagliola D , Lefrere JJ, Valleron AJ. A new approach to estimating AIDS incubation times: results in homosexual infected men. / Epidemiol Community Health 1992;46(6):582-6. 11. Stover J. Modelling the demographic impact of AIDS. / Health Popul Nutr 2002;20(2):102-3. 12. Stover J, Way P. Projecting the impact of AIDS on mortality. AIDS 1998;12 Suppl 1529-39. 13. Steiner JF, Prochazka A V . The assessment of refill compliance using pharmacy records: methods, validity, and applications. / Clin Epidemiol 1997;50(1):105-16. 14. Hogg RS, Heath K V , Bangsberg D , et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up. AIDS 2002;16(7):1051-1058. 15. Schechter M T , Hogg RS, Aylward B, Craib KJ, Le T N , Montaner JS. Higher socioeconomic status is associated with slower progression of H I V infection independent of access to health care. / Clin Epidemiol 1994;47(l):59-67. 16. Cox DR. Regression models and life tables. / Royal Stat Soc B 1972:187-202. 17. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 18. Hyndman JC, Holman C D , Hockey RL, Donovan RJ, Corti B, Rivera J. Misclassification of social disadvantage based on geographical areas: comparison of postcode and collector's district analyses. Int J Epidemiol 1995;24(l):165-76. 5 2 19. Liu H , Golin C E , Miller L G , et al. A comparison study of multiple measures of adherence to HIV protease inhibitors. Ann Intern Med 2001;134(10):968-77. 20. Duran S, Solas C , Spire B, et al. 'Do HIV-infected injecting drug users over-report adherence to highly active antiretroviral therapy?' A comparison between patients' self-reports and serum protease inhibitor concentrations in the French Manif 2000 cohort study. AIDS 2001;15(8):1075-7. 21. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 22. Wendel CS, Mohler MJ, Kroesen K, Ampel N M , Gifford A L , Coons SJ. Barriers to use of electronic adherence monitoring in an HIV clinic. Ann Pharmacother 2001;35(9):1010-5. 5 3 CHAPTER 4 ACCESS TO ANTIRETROVIRAL THERAPY PRIOR TO DEATH 4.1 FORWARD This chapter is presently under review as: Wood E, Montaner JS, Tyndall M , Schechter M T , O'Shaughnessy M V , Hogg RS. Race, Gender, Socio-economic Status, and Access to HIV-1 Treatment among persons who have died in the era of Antiretroviral Therapy. 54 4.2 ABSTRACT Background: Marked declines in HIV-related mortality as a result of modern antiretrovirals (ARVs) were first documented in the mid-1990s. Since this time, little work has been done to evaluate population-level H I V / A I D S mortality and access to antiretroviral therapy. Methods: We evaluated all HIV-related deaths in the province of British Columbia, Canada over the period January 1, 1995 to December 31, 2001 and sought to compare persons who had accessed A R V s prior to death to those who had died without ever accessing HIV treatment by fitting a logistic model. As a sub-analysis, we evaluated the time to a > 3 month therapy discontinuation among those individuals who accessed antiretrovirals prior to death using Kaplan-Meier methods and Cox proportional hazards regression. Results: Overall, 1241 deaths were attributed to HlV-infection during the study period. Of these, 406 (32.7%) deaths occurred amongst persons who never accessed any HIV treatment. In adjusted analyses, Aboriginal ethnicity (Adjusted Odds Ratio [AOR]: 0.61 [95% CI: 0.42 - 0.88]), female gender (AOR: 0.68 [95% CI: 0.47 - 0.98]) and lower median income (AOR: 0.70 [95% CI: 0.54 - 0.92] per $10,000 decrease) were negatively associated with receiving HIV treatment prior to death. Among the 835 (67.3%) individuals who accessed therapy, similar socio-demographic factors were associated with higher rates of therapy discontinuation prior to death. Conclusions: Overall, one third of HIV-related deaths occurred among untreated individuals, despite the presence of a universal healthcare system that delivers all A R V s and AIDS care free of charge. Furthermore, Aboriginal ethnicity, female gender, and lower socio-economic status were all independently associated with lower rates of A R V use prior to death, and similar results were observed when rates of therapy discontinuation were evaluated among the patients who accessed therapy prior to death. The data demonstrate high levels of ongoing H T V / A I D S mortality due to problems with the delivery of antiretrovirals, and indicate that specific sub-populations may be particularly affected by this concern. 5 5 4.3 I N T R O D U C T I O N Since its introduction in the mid-1990's, the benefits of antiretroviral therapy for the treatment of HIV disease have been well established.1 New antiretroviral regimens have proven effective in decreasing plasma viral load and improving CD4 cell counts, and marked improvements in HIV-related morbidity and mortality have been documented among persons on antiretroviral therapy. 2' 3 Although the side-effects of HIV treatment present challenges,4' 5 disease progression and AIDS deaths have become so rare among persons on antiretrovirals that, in many areas of the world, HIV infection is increasingly being viewed as a chronic and manageable illness. 6' 7 We have previously reported that the benefits of antiretroviral therapy are measurable at a population-level,3 and identified a significant improvement in the life expectancy of gay and bisexual males in the city of Vancouver's gay neighborhood.& These findings have emerged amidst growing concerns regarding inequitable access to antiretroviral therapy among those infected during the later stages of the HIV epidemic in North America, such as illicit drug users, ethnic minorities, and persons of lower socio-economic status.^ - 1 1 However, the majority of previous studies have been limited in that they have been conducted among populations of persons still living with H l V . i O ' 12-16 p o r this reason, it is presently not known to what extent apparent disparities in access to antiretrovirals will improve as HTV-disease becomes more advanced among persons who may have lower access.I7 Of particular interest, may be settings in which all medical care and antiretroviral therapy are provided free of charge through a universal healthcare system since in these settings financial considerations may be less relevant than in settings where all or some of these services must be paid for by the patient. Therefore, the present study was conducted to evaluate access to antiretrovirals among persons who have died of 56 H I V / A I D S in the province of British Columbia since 1995, and to determine factors associated with receiving antiretrovirals prior to death. 4.4 M E T H O D S In the province of British Columbia, Canada all anti-HIV medications have been centrally distributed at no cost to eligible HIV-infected individuals through the British Columbia Centre for Excellence in H I V / A I D S Drug Treatment Program since 1992. The Treatment Program remains the only source of free antiretroviral therapy in the province and pharmaceutical sales suggest that <1% of HIV-infected British Columbians purchase antiretroviral therapy outside the program.^-7 The Centre's H I V / A I D S Drug Treatment program has received ethical approval from the University of British Columbia Ethics Review Committee, and the program conforms to the province's Freedom of Information and Protection of Privacy Act. The Centre distributes antiretroviral medications based on specific guidelines generated by a Therapeutic Guidelines Committee. Since 1992, the guidelines have made available dual nucleoside antiretroviral therapy for people with a C D 4 + cell counts below 350 cells /mm 3 . In December 1995 dual therapy was made available to everyone with a C D 4 + cell count below 500 cells/mm3. In June 1996 the Centre adopted plasma viral load driven antiretroviral therapy guidelines, consistent with those put forward by the International AIDS Society - USA.18 m British Columbia, Zidovudine has been available since 1986. The other four nucleoside analogues were made available over a period of four years: didanosine and zalcitabine in 1992, stavudine in 1993, and lamivudine in 1994. Saquinavir, indinavir, and ritonavir have been available since 1996, and nelfinavir has been available since 1998. Delavirdine and nevirapine were made available in Canada in 1998 and efavirenz became available a year later in 1999. 57 For the purposes of the present study, a file was acquired from the British Columbia Ministry of Health's Vital Statistics Agency, who queried their database to extract records for all deaths that contained an ICD-9 (1995-99) or ICD-10 (2000-1) code indicating that HIV met the criteria of an underlying or associated cause of death. Vital Statistics defines a death as 'HIV underlying' if the attending physician and/or coroner determine that the death was directly due to HIV infection. Alternatively, Vital Statistics defines a death as 'HIV associated' if the attending physician and/or coroner determine that HIV infection was only a condition which contributed to the individual's underlying cause of death. Persons who died of accidental causes, such as illicit drug overdoses, suicides, and accidents are not routinely tested for HIV in the province, and unless an individual is experiencing morbidity due to confirmed HIV infection their death should not be coded as HIV associated.19 Nevertheless, to assess potential confounding that could stem from deaths among persons who were not eligible for antiretroviral treatment, we evaluated the number of cases that non-accidental causes of death (overdose, suicide, and accidents) occurred in the data set. For all deaths, Vital Statistics compiles several demographic characteristics as well as the postal code of the home residence of the deceased. Demographic variables compiled included: age at time of death, Aboriginal ethnicity, and gender. Similar to the experience of African Americans in the U.S . , 9 Aboriginal Canadians have been shown to be disproportionately affected by the HIV epidemic. 2^ Aboriginal Canadians also experience elevated mortality from causes of death related to poverty in comparison to other ethnic groups. 2 ! We therefore decided to evaluate the impact of this variable on treatment access in the present study. Since only limited socio-demographic data.were available from the death record, as we and others have done previously,! 1/ 2 2 we used census-based characteristics as a surrogate for socio-economic variables. First, a profile of the province was acquired 5 8 from Statistics Canada, which divided up the province into census tracts and census subdivisions and provided summary socio-demographic data for each census geographic area.23 Second, a postal code conversion file was used to link each study subject's home residence postal code to the corresponding neighborhood-based data. This linkage enabled neighborhood median income and urban/rural status to be assigned to each individual. In addition, we determined the census tract codes for Vancouver's Downtown Eastside, the neighborhood that is well recognized as the epicenter of the province's HIV epidemic among injection drug users (IDU),17 and examined the impact of this variable on treatment access. We evaluated all deaths among HIV-infected persons that occurred during the period January 1,1995 to December 31, 2001. January 1, 1995 was selected as the start of the study period since this is when we first observed marked declines in HTV/AIDS deaths due to the increased use of stavudine and lamivudine in the province.3/ 24 Among all subjects, we determined which individuals had ever received antiretrovirals from the Treatment Program through a confidential record linkage between Vital Statistics and the BC Centre for Excellence in HTV/AIDS. For the primary analysis, we conservatively assumed that an individual had accessed HIV treatment if they had ever filled > 1 prescription of antiretroviral medication from the Treatment Program prior to death. Although we restricted our analyses to persons who had died between 1995 and 2001, to be conservative, persons who received antiretrovirals at any time between 1992 and the end of 2001 were defined as having accessed treatment. Univariate and multivariate statistical techniques were used to evaluate associations between study subject's characteristics and treatment access prior to death. Variables considered in these analyses included: neighborhood-based socio-economic variables as described above, gender, ethnicity (Aboriginal versus other), cause of death (HlV-underlying versus HlV-associated), and age. Categorical explanatory variables 59 were analyzed using Pearson's Chi-square test and continuous variables were analyzed using the Wilcoxon rank sum test. For the primary analysis of treatment access (ever vs never), a logistic regression model was then fit by including all variables that were statistically significant in univariate analyses. In addition, to control for potential confounding, we a priori decided to adjust the final multivariate model for age, year of death, cause of death, and residence in the IDU HIV epicenter, regardless of whether these variables achieved statistical significance in univariate analyses. We assumed that filling > 1 prescription of antiretroviral medication may be an imprecise measure of treatment access, and therefore conducted a sub-analyses to evaluate treatment access prior to death. Since we have previously found that receiving antiretrovirals < 75% of the time during their first year on therapy is associated with mortality,^ we evaluated the proportion of patients who received antiretrovirals > 75% of the time during their first year on therapy and defined these patients as being consistently treated. However, since we had access to longitudinal data, we also evaluated the time to the first > 3 month stoppage of antiretroviral therapy prior to death (not restricted to the first year). Here, cumulative therapy discontinuation rates were estimated using Kaplan-Meier methods. As we have done previously, for the Kaplan-Meier analysis, median income was dichotomized into those patients above and below the Canadian low-income cut-off of $14,147 C N D . H / 26 Patients who continued to pick up therapy until within 3 months of their date of death, were right censored at their date of death and coded as non-events. Cox proportional hazards regression was then used to calculate univariate and adjusted relative hazards and 95% confidence intervals (CIs), using the same model building protocol described above. Here, however, we also adjusted for year of therapy initiation to adjust for changes in drug availability during the study period. The assumption of proportional hazards was validated by inspection of log (-log [survival function]) estimates against log time plots. 60 A l l tests of significance were two sided with a p-value of < 0.05 indicating that an association was statistically significant. A l l statistical analyses were performed using SAS software version 6.0 (SAS, Cary, N C ) . 4.5 RESULTS Overall , 1241 deaths among persons > 18 years had an ICD-9 (1995-1999) or ICD-10 (2000-2001) code indicating that H I V was either the underlying or an associated cause of death during the study period. Of these, 145 (11.7%) causes of death were deemed by the attending physician and/or coroner as HIV-associated, among w h o m 60 (41.4%) died without accessing treatment. A m o n g this group, the most common underlying causes of death were: 22 (15.2%) due to various malignancies; 18 (12.4%) due to accidental causes (overdose, accidents, and suicides); 15 (10.3%) due to liver disease; and 11 (7.6%) due to various cardiac conditions. A m o n g the 18 (12.4%) accidental deaths, 12 (66.7%) individuals accessed therapy prior to death. Due to the low number of accidental deaths, and the high proportion that had accessed therapy, we accepted the physician and/or coroner's judgement that H I V infection had contributed to the underlying cause of death and kept all individuals in the primary analysis. The remaining 1096 (88.3%) individuals died directly of an HlV-under ly ing cause (all AIDS related), among w h o m 346 (31.5%) died without accessing treatment. Overall , 19 (1.5%) individuals did not have a postal code that could be linked to census data, due to homelessness or otherwise, and were therefore excluded from all multivariate analyses. These 19 individuals, were more likely to be Aboriginal (53% vs 12%; p < 0.05) and female (32% vs 12%; p < 0.05). Table 4.1 shows the socio-demographic characteristics of the overall study population and the results of univariate comparisons of those that did (n = 835 [67.3%]) and did not (n = 406 [32.7%]) receive antiretrovirals prior to death. A s shown here, 6 1 lower median income (OR = 0.63 per $10,000 decrease), female gender (OR = 0.66), Aboriginal ethnicity (OR = 0.54), and residence in the IDU HIV epicentre (OR = 0.48), were negatively associated with accessing antiretrovirals prior to death. Age, urban residence, and year of death were not statistically associated with receiving antiretrovirals prior to death. Table 4.2 shows the results of the multivariate analysis of factors associated with receiving any antiretroviral treatment prior to death. As shown here, Aboriginal ethnicity (Adjusted Odds Ratio [AOR] = 0.61 [95% CI: 0.42 - 0.88]; p = 0.008), female gender (AOR = 0.68 [95% CI: 0.47 - 0.98]; p = 0.040), and lower median income (AOR = 0.70 [95% CI: 0.54 - 0.92]; p = 0.010 per $10,000 decrease) were all negatively associated with receiving antiretrovirals prior to death in adjusted analyses. Overall, among the 835 individuals who received treatment prior to death, only 380 (45.5%) received antiretroviral medication > 75% of the time during their first year on therapy and 455 (54.5%) received antiretrovirals < 75% of the time. Among the 82 Aboriginal persons who accessed therapy, only 23 (28.1%) received antiretrovirals > 75% of the time, and among the 90 women who accessed treatment only 32 (35.6%) received antiretrovirals > 75% of the time. Figure 4.1 shows the cumulative therapy discontinuation rates (> 3 months) among the patients who accessed therapy stratified by: (a) ethnicity, (b) gender, and (c) socio-economic status. As shown here, Aboriginal persons, females, and persons in the lower income strata had statistically elevated rates of therapy discontinuation (all log rank p < 0.01). The curves do not drop until 1 month after the initiation of therapy since all individuals are initially dispensed 1 months worth of antiretroviral treatment. 6 2 Table 4.3 shows the univariate and adjusted relative hazards of therapy discontinuation for those who accessed therapy prior to death. As shown here, in adjusted analyses, females (Adjusted relative hazard [ARH]: 1.33 [95% CI: 1.00 - 1.77]; p = 0.049) and persons of lower socio-economic status (ARH: 1.33 [95% CI: 1.16 - 1.72] per $10,000 decrease; p < 0.001) experienced statistically elevated rates of therapy discontinuation. In the adjusted analysis, ethnicity did not achieve statistical significance. 4.6 DISCUSSION Even in a setting where all healthcare is provided free of charge by the state, a high H I V / A I D S death rates persist due to the lack of, or only marginal, access to antiretrovirals. In the present study, one third of HIV-related deaths occurred among untreated individuals, and those that died without ever receiving HIV treatment were more likely to be Aboriginal, female, and to reside in a neighborhood with a lower median income. Among those that accessed treatment, less than half received consistent treatment prior to death and similar socio-demographic characteristics were associated with elevated rates of treatment discontinuation prior to death. In North America, Aboriginal persons (including American Indians and Alaska Natives) may be at elevated risk of HIV infection. 2 7 , 28 p o r m i s reason it is particularly worrisome that the present study identified strong associations between Aboriginal ethnicity and not receiving antiretroviral therapy. These data demonstrate that culturally sensitive interventions aimed at improving access to antiretrovirals among HIV-infected Aboriginal persons is an urgent priority. Although previous studies of access to antiretrovirals have tended to be conducted among living populations, these findings are also consistent with the experience of other HIV-positive ethnic minorities in other settings. 10,13,14 63 In addition, we found associations between lower income and female gender, and not receiving antiretrovirals, as well as receiving inconsistent treatment, prior to death. These findings are also consistent with previous studies that have found lower socio-economic status and female gender to be associated with worse access to healthcare 10/12,29,30, a n c i interventions aimed at improving access to antiretrovirals among females and lower income persons must also be made a priority.10/13, 31 There are several features of our study that should be highlighted. First of all, it is important to stress that the data arose in a setting where all H I V / A I D S care, antiretrovirals, and laboratory monitoring are available free of charge, and where previous studies have shown that virtually all patients acquire antiretrovirals through a centralized source.12 Second, the centralized death registry enabled complete population-level data on H I V / A I D S deaths for the entire province. Finally, since a complete prospective record of antiretroviral dispensation was maintained, it was possible to determine precisely each individual's level of treatment prior to death. For these reasons, our results are not influenced by selection factors that may compromise the interpretation of other cohort studies. Conversely, although previous studies have suggested that ecological measures may result in an underestimation of the true relationship between socio-economic status and health outcomes H / 32r w e should stress that there was potential for misclassification because an ecological measure of socioeconomic status was used. Similarly, although we adjusted our analyses for residence in the neighborhood where the majority of IDU are concentrated, a major limitation of data from death registries is that they do not contain data regarding behaviors such as the use of illicit drugs. A multitude of studies have found addiction to be a barrier to accessing adequate HIV treatment 12/ 33/ 34^ a n c j w e r e this data available it is likely that our models would have been affected. This is because illicit drug users tend to reside in lower income areas, and because in comparison to other risk groups, 64 females are more likely to have been infected through illicit drug use or through sexual contact with illicit drug users.3^/ 36 Nevertheless, it should be noted that such an adjustment, if possible, would not change the underlying fact that women and persons of lower-socio-economic status have markedly worse access to antiretrovirals and elevated therapy discontinuation rates in our setting. 4.7 S U M M A R Y In summary, we found that one third of HIV-related deaths occurred among untreated individuals, and that Aboriginal ethnicity, female gender, and lower socio-economic status were all associated with dying without treatment. Among those who accessed treatment prior to death, less than half received consistent treatment, and similar socio-demographic factors were associated with elevated rates of therapy discontinuation. Given that these data were derived in a universal healthcare setting where all H I V / A I D S care and antiretrovirals are available free of charge, it is likely that similar problems exist in many other settings in the developed world. To prevent ongoing levels of H I V / A I D S mortality, novel strategies will be required to improve access to antiretrovirals among populations with poor access to care. 65 4.8 R E F E R E N C E S 1. Yeni PG, Hammer SM, Carpenter C C , et al. Antiretroviral treatment for adult HIV infection in 2002: updated recommendations of the International AIDS Society-USA Panel. J A M A 2002; 288:222-35. 2. Moore RD, Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS 1999; 13:1933-42. 3. Hogg RS, O'Shaughnessy M V , Gataric N , et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997; 349:1294. 4. Carr A , Samaras K, Thorisdottir A , Kaufmann GR, Chisholm DJ, Cooper D A . Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodystrophy, hyperlipidaemia, and diabetes mellitus: a cohort study [see comments]. Lancet 1999; 353:2093-9. 5. Heath K V , Hogg RS, Chan K, et al. Lipodystrophy-associated morphological, cholesterol, and triglyceride abnormalities in a population-based H I V / A I D S Treatment Program. AIDS 2001; 15:231-239. 6. Ledergerber B, Egger M , Opravil M , et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss HIV Cohort Study. Lancet 1999; 353:863-8. 7. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. J A M A 2001; 286:2568-2577. 8. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver's West End. Can J Public Health 2000; 91:125-8. 66 9. Karon JM, Fleming PL, Steketee RW, De Cock K M . HIV in the United States at the turn of the century: an epidemic in transition. A m J Public Health 2001; 91:1060-8. 10. Anderson K H , Mitchell JM. Differential access in the receipt of antiretroviral drugs for the treatment of AIDS and its implications for survival. Arch Intern Med 2000; 160:3114-20. 11. Wood E, Montaner JS, Chan K, et al. Socioeconomic status, access to triple therpay, and survival from HIV-disease since 1996. AIDS 2002; 16:2065-2072. 12. Strathdee SA, Palepu A , Cornelisse PG, et al. Barriers to use of free antiretroviral therapy in injection drug users. J A M A 1998; 280:547-9. 13. Andersen R, Bozzette S, Shapiro M , et al. Access of vulnerable groups to antiretroviral therapy among persons in care for HIV disease in the United States. HCSUS Consortium. HIV Cost and Services Utilization Study. Health Serv Res 2000; 35:389-416. 14. Moore RD, Stanton D , Gopalan R, Chaisson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994; 330:763-8. 15. Cook JA, Cohen M H , Grey D , et al. Use of highly active antiretroviral therapy in a cohort of HIV-seropositive women. A m J Public Health 2002; 92:82-7. 16. H s u L C , Vittinghoff E, Katz M H , Schwarcz SK. Predictors of use of highly active antiretroviral therapy (HAART) among persons with AIDS in San Francisco, 1996-1999. J Acquir Immune Defic Syndr 2001; 28:345-50. 17. Wood E, Schechter M T , Tyndall M W , Montaner JS, O'Shaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000; 14:1229-35. 67 18. Carpenter C Q Fischl 'MA, Hammer SM, et al. Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. International AIDS Society-USA. J A M A 1996; 276:146-54. 19. Miller C L , Chan KJ, Palepu A , et al. Socio-Demographic Profile and HIV and Hepatitis C Prevalence Among Persons who Died of a Drug Overdose. Addict Res and Theory 2001; 9:459-470. 20. Health Canada. Bureau of H I V / A I D S , Centre for Infectious Disease, Prevention and Control, Health Canada. H I V / A I D S among Aboriginal persons in Canada remains a pressing issue. HIV/AIDS Epi Update 2001. 21. Anand SS, Yusuf S, Jacobs R, et al. Risk factors, atherosclerosis, and cardiovascular disease among Aboriginal people in Canada: the Study of Health Assessment and Risk Evaluation in Aboriginal Peoples (SHARE-AP). Lancet 2001; 358:1147-53. 22. Katz M H , Hsu L, Lingo M , Woelffer G , Schwarcz SK. Impact of socioeconomic status on survival with AIDS. A m J Epidemiol 1998; 148:282-91. 23. Statistics Canada. 1996 Population Census of Canada; Summary Area Profile Data by Census Tract and Census Subdivision. 24. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. J A M A 1998; 279:450-4. 25. Hogg RS, Heath K V , Bangsberg D, et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up. AIDS 2002; 16:1051-1058. 26. Statistics Canada Income Statistics Division. Low Income Cutoffs from 1991 to 2000. Catalogue no. 75F0002M1E. 68 27. H I V / A I D S among American Indians and Alaskan Natives\u00E2\u0080\u0094United States, 1981-1997. M M W R Morb Mortal Wkly Rep 1998; 47:154-60. 28. Heath K V , Cornelisse P G , Strathdee SA, et al. HIV-associated risk factors among young Canadian Aboriginal and non-Aboriginal men who have sex with men. Int J STD AIDS 1999; 10:582-7. 29. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999; 48:1751-8. 30. Shapiro M F , Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. J A M A 1999; 281:2305-15. 31. Bangsberg DR, Mundy L M , Tulsky JP. Expanding directly observed therapy: tuberculosis to human immunodeficiency virus. A m J Med 2001; 110:664-6. 32. Hyndman JC, Holman C D , Hockey RL, Donovan RJ, Corti B, Rivera J. Misclassification of social disadvantage based on geographical areas: comparison of postcode and collector's district analyses. Int J Epidemiol 1995; 24:165-76. 33. Celentano D D , Vlahov D, Cohn S, Shadle V M , Obasanjo O, Moore RD. Self-reported antiretroviral therapy in injection drug users. J A M A 1998; 280:544-6. 34. Palepu A , Yip B, Miller C , et al. Factors associated with the response to antiretroviral therapy among HIV-infected patients with and without a history of injection drug use. AIDS 2001; 15:423-424. 35. Edlin BR, Irwin K L , Faruque S, et al. Intersecting epidemics\u00E2\u0080\u0094crack cocaine use and HIV infection among inner-city young adults. Multicenter Crack Cocaine and HIV Infection Study Team. N Engl J Med 1994; 331:1422-7. 69 Spittal P M , Craib KJ, Wood E, et al. Risk factors for elevated HIV incidence rates among female injection drug users in Vancouver. C M A J 2002; 166:894-9. 7 0 Table 4.1 Univariate analyses of socio-demographic and neighborhood-based characteristics with regard to ever accessing HIV treatment prior to death. Accessed Treatment No Yes Unadjusted Characteristic n, (%) n, (%) Odds Ratio (95% C.I.) p-value Median Income* Median $18.1 $19.4 Inter-quartile Range (13.6-21.0) (15.5-22.0) 0.63 (0.50-0.79) <0.001 Gender Male 343(84.5) 745(89.2) Female 63 (15.5) 90 (10.8) 0.66 (0.47 - 0.93) 0.017 Ethnicity Other 338 (83.3) 753 (90.2) Aboriginal 68(16.8) 82(9.8) 0.54 (0.38-0.77) 0.001 Cause of Death HIV-associated 60 (14.8) 85 (10.2) HIV-underlying 346(85.2) 750(89.8) 1.53 (1.07-2.18) 0.018 Age Median 41 41 Inter-quartile Range (34-48) (36-47) 1.00 (0.89-1.12) 0.525 IDU HIV Epicentre N o 364(89.7) 791 (94.7) Yes 42(10.3) 44(5.3) 0.48 (0.31-0.75) 0.001 Urban N o 45(11.3) 75(9.1) Yes 354(88.7) 748(90.9) 1.27 (0.86-1.87) 0.233 Death Year Median 1997 1997 Inter-quartile Range (1995-1999) (1996-1999) 1.03 (0.98-1.09) 0.298 *Income in thousands $ C N D . Note: O R = odds ratio; C.I. = Confidence interval; I D U = injection drug use. Odds ratios for continuous variables are as follows: Age: per 10 years older; Median Income: per $10,000 decrease. 7 1 Table 4.2 Logistic regression analysis* of factors associated with ever receiving antiretroviral treatment prior to death. Variable Adjusted Odds Ratio 95% Confidence Interval p-value Aboriginal Ethnicity (Yes versus No) 0.61 (0.42 - 0.88) 0.008 Gender (Female versus Male) 0.68 (0.47 - 0.98) 0.040 Median Income (per $10,000 decrease) 0.70 (0.54 - 0.92) 0.010 * Model was also adjusted for year of death, residence in the IDU HIV epicenter, age, and cause of death. 72 Table 4.3 Univariate and multivariate Cox proportional hazards regression of factors associated time to antiretroviral therapy discontinuation. Unadjusted Adjusted Relative Hazard (RH) Relative Hazard (RH) R H (95% CI) Risk Ratio (95% CI) Ethnicity Aboriginal vs Other 1.56 (1.18-2.06) 1.19 (0.87-1.61) Gender Female vs Male 1.41 (1.09-1.85) 1.33 (1.00-1.77) Median Income per $10,000 decrease 1.55 (1.31-1.84) 1.41 (1.16-1.72) *Model was adjusted for year of death, year of therapy initiation, age, cause of death, and residence in the I D U H I V epicenter. 7 3 L E G E N D F O R F I G U R E Figure 4.1 Kaplan-Meier product limit estimates of cumulative therapy discontinuation > 3 months among HIV-infected subjects who later died between January 1, 1995 and December 31, 2001 stratified by a) ethnicity; b) gender; and c) socio-economic status. Note, the curves do not drop until 1 month after the initiation of therapy since all individuals are initially dispensed 1 months worth of treatment. 7 4 Figure 4.1(a): Ethnicity CL CO I\u00E2\u0080\u0094 CD o c d c CC E cd 100 y 90-80 -70 -60-50-40 30 20 -\ 10 0 \u00E2\u0080\u00A2i , Other Aboriginal Log-Rank p = 0.001 1 1 1 6 12 18 Time from Start of ARVs (Months) - 1 24 Figure 4.1(b): Gender >\u00C2\u00BB c l co s\u00E2\u0080\u0094 cd JZ h-c o c d co E CD 100 T 90 -80 -70 -60 -50 -40 -30 -20 -10 -0--Male Female Log-Rank p = 0.008 1 1 1 6 12 18 Time from Start of ARVs (Months) i 24 Figure 4.1(c): Socio-economic Status Q. CO \ CD o d) CO E CD CC 100 90-80-70-60-50-40-30-20-10-> $14 'T47 < $14,147 1 1 1 6 12 18 Time from Start of ARVs (Months) \u00E2\u0080\u0094 i 24 Log-Rank p < 0.001 7 5 C H A P T E R 5 M O D E L L I N G T H E I M P A C T O F I M P R O V E D U P T A K E O F T R E A T M E N T I 5.1 F O R W A R D This chapter was published in AIDS in 2000 as: Wood E, Schechter M T , Tyndall M , O'Shaughnessy M V , Montaner JSG, Hogg RS. Antiretroviral medication use among injection drug users: Two potential futures. AIDS 2000; 14:1229-1236. 7 6 5.2 A B S T R A C T Objective: To model the potential impact of HIV infection rates and antiretroviral medication use on life expectancy and mortality in the Downtown Eastside of Vancouver, British Columbia from 1999 to 2006. Methods: Population projections were made to estimate the population of the Downtown Eastside in the year 2006.Two scenarios were modeled to predict the impact of HIV infection and antiretroviral use on mortality and life expectancy. The use of antiretroviral therapy was estimated to be 80% in this first scenario and 20% in the second. The HIV prevalence by age and sex, and by year infected was estimated using data from the Vancouver Injection Drug User Study (VIDUS). Results: If the level of antiretroviral therapy use among HIV positive persons was 80% at baseline, then we estimate that the life expectancy at birth in the year 2006 will be 60.8 years for men and 72.8 years for women, and 172 AIDS deaths will occur between 1999 and 2006. In contrast, if the present level of antiretroviral medication use persists, the life expectancy at birth in the year 2006 will be 56.9 years for males and 68.6 years for females, and 503 AIDS deaths will occur between 1999 and 2006. Conclusion: Our analysis suggests that if the low levels of antiretroviral therapy use persist, life expectancy in Vancouver's Downtown Eastside will soon be on par with many of the world's least developed countries. Our findings highlight the large health status decline that can be expected in many inner city neighborhoods if low levels of antiretroviral use persist. Although reasonable coverage targets for injection drug users (IDU) have not been established, expanded access to and use of antiretrovirals is urgently needed to avert a drastic decline in health status. 77 5.3 INTRODUCTION Like many infectious diseases, there is a great deal of social and geographic variation in the incidence of and mortality from HlV/AIDS.1 '2 In Canada the highest incidence of HIV occurs in the larger urban centres, and evidence suggests that the high rates of mortality attributed to this cause are concentrated within central districts and among specific population sub-groups. 3 ' 4 As a result, certain neighborhoods have been hardest hit by the H I V / A I D S epidemic. This is true of the Downtown Eastside of Vancouver, British Columbia which since the mid-1990's has experienced an explosive HIV epidemic among injection drug users (IDU).5 Vancouver's Downtown Eastside is the most impoverished urban neighborhood in Canada. 6 The area suffers from a host of problems including high levels of illicit injection drug use. As a result, the 18 percent annual HIV incidence among injection drug users observed in this neighborhood in 1997 is among the highest ever documented in the developed world . 7 A cohort study among this population indicates that the bulk of HIV infections occurred in the mid-1990's, and the majority of infected individuals remain AIDS-free.5/7 As is the case in many other low-income communities that have experienced injection drug use-associated HIV epidemics, the use of antiretroviral therapy remains very low.8/9 Little work has been done to forecast how the health status of this neighborhood may be affected as the epidemic evolves. In recent years, antiretroviral therapies have substantially altered the natural history of HIV infection.10'11 New anti-HIV regimens have led to significant declines in AIDS related morbidity and mortality among persons on antiretroviral therapy in British Columbia. 12,13 Beyond the individual level, these benefits have had a significant impact on the health status of certain neighborhoods in Canada, such as Vancouver's 78 West End. This area is recognized as British Columbia's largest Gay community, and the substantial increases in life expectancy that have been documented among West End Gay and Bisexual males have been attributed to the high level of antiretroviral medication use. 14 In contrast, previous studies of antiretroviral therapy use in the Downtown Eastside have indicated that only a fraction of eligible HIV positive injection drug users are accessing antiretroviral therapy.8 Similar concerns have been reported in many other settings 9 ' !^ if m i s situation persists, the survival benefits and decreases in morbidity that have been attributed to antiretroviral medication use among gay men, will not accrue in inner city neighborhoods such as the Downtown Eastside where the majority of HIV infected individuals are IDU. The present study was undertaken to model the potential impacts of H I V / A I D S on the health status of men and women who reside in Vancouver's Downtown Eastside during the year 2006. We have constructed two scenarios in which the health status is dependent on the level of antiretroviral medication use. Specifically, in the first scenario, we modeled the hypothetical life expectancy if the proportion of individuals on antiretroviral therapy in the Downtown Eastside were to increase to the level of use presently occurring among HIV-positive persons in Vancouver's West End. In the second scenario, we calculated the Downtown Eastside's hypothetical health status if the present level of antiretroviral medication use persists until 2006. 5.4 M E T H O D S In this modeling exercise, we projected the potential impact of HIV on the male and female AIDS mortality rate and life expectancy in the Downtown Eastside over the period 1999 to 2006. 2006 was chosen as the endpoint of the projection because this date is approximately 10 years after the majority of HIV infections occurred. 7 Fertility, life expectancy, population, and migration data were input into DemProjTM 7 9 demographic modeling software to project the impact of HIV infection on the life expectancy in the Downtown Eastside in the year 2006.16 Two hypothetical scenarios were created using DemProj's AIDS impact model.16 in the first scenario, we modeled the potential mortality and life expectancies in the year 2006 if the proportion of HIV positive persons on antiretroviral therapy was 80% over the 8 year period. This is the level of up-take that has been estimated, through the Province's H I V / A I D S Treatment Program's records and surveys of West End residents, to occur presently among HIV-positive individuals in Vancouver's West E n d . l 7 In the second scenario we modeled the potential mortality and life expectancies for the Downtown Eastside in the year 2006 if the present level of antiretroviral therapy use among HIV-positive injection drug users persists over the study period. The values assigned to model parameters in each of these two scenarios were based, where possible, on empirical data. Like any traditional population projection, our models required information on a number of demographic parameters including population size, fertility, migration, and mortality rates.18 Population estimates and migration data for the Census Tracts that compose the Downtown Eastside were acquired from the 1996 Census. 6 Baseline gender specific life expectancy data was acquired from the British Columbia Division of Vital Statistics. The level of non-HIV-related deaths was kept constant at baseline levels throughout the entire study period. In addition, age specific fertility data was acquired from the British Columbia Ministry of Health. In addition to these demographic parameters, we required information on four important epidemiological characteristics of the Downtown Eastside HIV epidemic. First, the total number of injection drug users who reside in the Downtown Eastside was estimated to be 5,000.7'19 Although there may be more immigration than emigration of injection drug users into the Downtown Eastside we fixed the number of IDUs residing in the neighborhood at 5,000 for the entire study period, since previous 80 migration analyses of this neighborhood have shown that the net migration flow of HIV positive persons on antiretroviral therapy into and out of this neighborhood is approximately zero.20 Second, we estimated the proportion of injection drug users who are HIV positive by age, gender, and year of infection, by projecting the Vancouver Injection Drug Use Study (VIDUS) cohort that is HIV positive onto the estimated population of Downtown Eastside IDU. The HIV prevalence in the Downtown Eastside was estimated using data from the VIDUS Study, a prospective cohort study that began in May 1996. Study subjects were recruited through self-referral and street outreach. 7 Subjects were eligible if they had injected illicit drugs at least once in the previous month, resided in the greater Vancouver region, and provided written informed consent. At baseline and semi-annually, subjects provided blood samples and completed an interviewer-administered questionnaire. The questionnaire elicits demographic data including age, gender, and place of residence. As the future HIV prevalence of the neighborhood is unknown and may be most affected by public policies, a sensitivity analysis was applied to this parameter. Three models were run in both scenarios one and two. In the first model, migration and mortality were assumed to reduce the prevalence after 1999; in the second model, new infections and mortality were balanced to keep the prevalence constant over the study period; and in the third model, the prevalence was assumed to increase after 1999. Third, we determined the number of HIV positive injection drug users in the Downtown Eastside by age and gender that were on antiretroviral therapy as of June 1999 through a linkage with the BC Centre for Excellence in H T V / A I D S Drug Treatment Program. The Treatment Program remains the only source of free antiretroviral therapy in the province and pharmaceutical sales suggest that less than 1% of HIV-infected British Columbians purchase antiretroviral therapy outside the program.^ In 8 1 order to avoid underestimating in our estimate of current antiretroviral therapy use among injection drug users and instead be liberal in our estimate, we assumed that all individuals who were on antiretroviral therapy who resided in the Downtown Eastside were IDU. Fourth, the average time from HIV infection to AIDS was assumed to differ greatly in scenarios one and two because of the different levels of antiretroviral medication use. In the first scenario we modeled the potential mortality and life expectancies in the year 2006 if the proportion of HIV positive persons on antiretroviral therapy was 80 percent over the 8 year period. In the second scenario we modeled the potential mortality and life expectancies for the Downtown Eastside in the year 2006 if the level of antiretroviral therapy use among HIV positive injection drug users remains at 20%. The median survival time for persons not using antiretroviral therapy was based on previous estimates and set at 10 years.21/22 However, the median survival time for persons on antiretroviral therapy had to be estimated because, as a result of the recent introduction of triple combination therapy, the actual survival benefits afforded remain to be quantified. Population based studies have indicated that during the period 1989 to 1994 in which monotherapy and later double combination therapy were introduced the median survival increased by approximately 2 years in comparison to the pre-therapy era.23 A more recent study, comparing double combination therapy to triple combination therapy using a conservative intent-to-treat analysis suggested that the cumulative death rate 12 months after initiation of therapy was 7.4% for patients initially prescribed double combination therapy and 1.6% for patients prescribed triple combination therapy.24 We conservatively estimated from these data that the median time from HIV infection to death for persons on modern antiretroviral therapy is 17 years. Similarly, in order to be conservative, the rate of perinatal 82 transmission was fixed at 6 percent in both scenarios. Therefore, perinatal transmission and infant AIDS mortality had virtually no impact on our results. 5.5 RESULTS O n the basis of census and vital statistics data, the mid-year population of men and women living in the Downtown Eastside at baseline was estimated to be 12,410 and 7,405 respectively in 1996. The most recent Census indicates that the neighborhood's unemployment rate was 25.8%, the median income was Can$10,200, and the proportion of the population who were Aboriginal was 8.3 percent in 1996.6 Previous analyses of the VIDUS cohort have shown that severely depressed socioeconomic status and over-representation of Aboriginal persons are also characteristics of IDU in this neighborhood.8/25 British Columbia vital statistics estimate that the life expectancy in this population in 1996 was 77.7 years for women and 65.0 for men. Based on the number of estimated injection drug users in this population and the HIV seroprevalence rate in the VIDUS study, we estimated that the adult HIV prevalence in the Downtown Eastside was 7 percent at baseline. This prevalence was kept constant after 1999 in both scenarios. Of the persons estimated to be HIV-positive at baseline, we extrapolated from BC Centre for Excellence in HTV/AIDS data that only 20% were on any antiretroviral therapy. In the first scenario, in which 80% of injection drug users are accessing antiretroviral medication over the study period, the projected life expectancy at birth in the year 2006 will be 60.8 for men and 72.8 for women, if HIV prevalence remains constant at 7% (Table 5.1). As shown in Figure 5.1, in this scenario AIDS deaths result in a reduction in life expectancy of 4.2 years among men and 4.9 years among women over the study period. As shown in Figure 5.2, there will be 46 AIDS deaths per year occurring in this neighborhood by the year 2006. 83 In the second scenario, in which the present level of antiretroviral medication use persists, our results indicate that the life expectancy at birth in the year 2006 will be 56.9 for men and 68.6 for women, if the prevalence remains constant at 7% (Table 5.1). As noted in Table 5.1 and shown in Figure 5.1, in this scenario AIDS deaths result in a reduction in life expectancy of 8.1 years among men and 9.1 years among women over the study period. As shown in Figure 5.2, there will be 93 AIDS deaths per year occurring in this neighborhood by the year 2006. In both scenarios the death rate from all non-HIV-related causes was fixed at baseline levels. The differences in the life expectancies forecast by the two scenarios are due solely to differences in mortality form an AIDS-defining illness or a cause directly associated with HIV infection. Our results indicate that if the present level of antiretroviral use persists and HIV prevalence remains constant at 7%, in scenario two, a total of 503 AIDS deaths can be expected between 1999 and 2006 with the death rate per thousand reaching 5.7 by the year 2006 (Figure 5.3). In contrast, if the level of antiretroviral use increases to the same level of use occurring presently in Vancouver's West End, only 172 AIDS deaths would be expected between 1999 and 2006 with the death rate per thousand reaching 2.7 by 2006 (Figure 5.3). Table 5.1 provides the results of the sensitivity analyses, in which the prevalence estimate was adjusted to reflect decreasing, constant, and increasing prevalence. As shown here, the divergent patterns of life expectancy predicted by scenarios one and two will occur even if the prevalence decreases, but will be worsened by an increase in prevalence. If the prevalence decreases to 5%, an overall life expectancy difference of 3 years is forecast; if the prevalence remains constant at 7%, a 4 year difference in life expectancy is forecast; and if the prevalence increases to 9%, a 5.3 year difference in life expectancy is forecast. 84 5.6 DISCUSSION Given the current clinical efficacy of anti-HIV therapies, Vancouver's Downtown Eastside will experience a more than 4 year decrease in life expectancy if the level of antiretroviral therapy use among HIV positive persons increases to the level of access occurring presently in the West End. This demonstrates that the management of HIV disease has a way to go before AIDS mortality is not negatively impacting health status in many areas, regardless of the level of antiretroviral therapy use. Alarmingly, if the present level of antiretroviral medication use persists over the study period, the life expectancy at birth in the year 2006 will drop by over 8 years for men and over 9 years for women. The life expectancies projected in this scenario, if occurring presently, would put the Downtown Eastside below all but the very least developed nations in the w o r l d . 2 6 It is important to put into context the over 4 year difference in life expectancy projected by the two scenarios. Previous studies have demonstrated that all motor vehicle accidents are responsible for 0.5 years of life expectancy lost among Canadian men. 3 Similarly, all lung cancer deaths are responsible for a 0.9 years of life expectancy lost among Canadian men and breast cancer for 0.5 years of life expectancy lost among all Canadian w o m e n . 2 7 Viewed in this light, the 4 year difference in potential life expectancies forecast by scenarios one and two represents an enormous difference in health status. Increasing antiretroviral therapy use appears to be the only intervention with the potential to prevent a life expectancy decline of over 8 years for men and over 9 years for women in the Downtown Eastside. Similar declines in life expectancy due to H I V / A I D S are not unprecedented. In Thailand AIDS mortality has resulted in a 2 year decline in life expectancy.2^ in Nigeria AIDS mortality has reduced life expectancy by 4 years. 2^ It has been projected that by 85 the year 2010, 8 to 31 years of life expectancy will have been lost in those countries most affected by H I V / A I D S in Sub-Saharan Africa .28 These projections were made under the assumption that antiretroviral therapy is out of reach for virtually all infected individuals, however, and survival times and incidence rates reflect zero antiretroviral therapy use. Unfortunately, despite the fact that antiretroviral therapy in British Columbia is provided at no cost through a universal health care system, Treatment Program records indicate that the present level of antiretroviral therapy use in the Downtown Eastside is below 20%. There are a number of potential benefits attributable to the use of antiretroviral therapy beyond the documented survival benefits. For example, numerous studies have shown that antiretroviral therapy is extremely effective for decreasing plasma viral load and suppressing viral replication.29,30 As a result, investigators have suggested that persons on antiretroviral therapy may be less infectious than individuals not on therapy.31 As such, increasing access to and use of antiretrovirals in the Downtown Eastside has the potential to reduce HIV incidence rates among those that continue to engage in high-risk activities. In addition, the testing and counseling that accompany HIV treatment also has the potential to reduce risk behavior. In addition, from a fiscal standpoint numerous studies have shown that the costs of providing antiretroviral therapy may be more than recovered by the reduced medical service utilization of persons on therapy.32,33 The costs of providing medical care for HIV positive injection drug users have been shown to be extremely high.34,35 These findings suggest that providing antiretroviral medications as part of a comprehensive intervention program could help to reduce future health care costs. The present level of antiretroviral medication use in the West End was used to represent an achievable level under quite favorable conditions. However, there are vast differences between the HIV positive populations, as well as the community 86 characteristics of the West End and the Downtown Eastside. In the West End there are a number of services that have contributed to the use of antiretroviral therapy including, a large tertiary AIDS care centre, community-based AIDS service organizations, treatment information resources, as well as treatment-related peer and social support. In contrast, despite the efforts of many dedicated professionals, current service provision for persons with HIV in the Downtown Eastside may be inadequate, since the majority of persons with HIV that reside in this neighborhood must meet the combined challenges of drug addiction and HIV infection. Increasing the up-take of antiretroviral therapy among injection drug users and other marginalized populations will be challenging due to numerous social and medical obstacles. Inadequate housing, poor access to medical services, and migration patterns make adherence to complicated antiretroviral combinations difficult. Competing medical problems, including mental illness, further complicates antiretroviral use. Although efforts to increase access to antiretroviral medications should be a priority, reasonable coverage targets have not been established. Increasing the use of antiretroviral therapy in the Downtown Eastside will require the addition of a number of client tailored resources, and could most easily be achieved as part of a comprehensive preventative intervention/harm reduction strategy. In this modeling exercise we have made a number of assumptions that may have introduced certain biases. The scenarios that we have developed are hypothetical and a number of interventions could change our results. For example, our estimate of the survival benefits afforded by modern antiretroviral therapy was very conservative, and given the pace at which anti-HIV therapy is improving it is possible that new therapies will be able to extend the lives of persons on therapy well beyond our estimates. A n y advances will only widen the health status gap between those individuals who are and are not accessing therapy. In addition, it is possible that certain interventions could 87 result in substantial reductions in the high mortality rates from non-HIV/AIDS deaths, which we fixed at baseline levels in both models. Many of these deaths are from avoidable causes such as overdoses,2^ and it should not be the goal of health care planners to maintain the present life expectancy, which for males at baseline was on par with many under-developed countries.28,36 Furthermore, in virtually all instances we have been cautious in our estimates and we feel that most biases that may have entered our analyses are conservative. For example, we calculated the 28.4 percent HIV prevalence in VIDUS based on all individuals who have returned for a follow-up visit. Kaplan Meier estimates which account for those individuals who were lost to follow-up, suggest that the actual HIV prevalence may be as high as 40%. 5.7 S U M M A R Y This study demonstrates the huge decrease in life expectancy that can be expected in the Downtown Eastside if the low level of anti-HIV medication use among injection drug users persists. Our analyses indicate that if the level of antiretroviral therapy use among HIV positive persons increased to the level that is occurring presently in the province's gay neighborhood, the life expectancy at birth in the year 2006 will be 60.8 years for males and 72.8 years for females. In this scenario, 172 AIDS deaths can be expected between 1999 and 2006. In contrast, if the present level of antiretroviral medication use persists, the life expectancy at birth in the year 2006 will be 56.9 for males and 68.6 for females. In this scenario 503 AIDS deaths can be expected between 1999 and 2006. The life expectancies projected in this scenario, if occurring presently, would put the Downtown Eastside below all but the very least developed nations, during a time that life expectancy is rising in the developed world. 2 ^ These values were calculated under the assumption of a constant prevalence level. Sensitivity analyses, which considered the effect of increasing prevalence, indicate that a prevalence increase will only widen the divergent health status patterns. Our analyses 88 support the growing body of work which suggests that certain preventative interventions and harm reduction strategies are urgently needed to avert a human catastrophe. Unfortunately, one of the potential futures outlined in this paper appears to be the path down which we are going, and our results present a challenge to health policy makers. 8 9 5.8 R E F E R E N C E S 1. Schechter M T , Hogg RS, Aylward B, Craib KJ, Le T N , Montaner JS. Higher socioeconomic status is associated with slower progression of HIV infection independent of access to health care. / Clin Epidemiol 1994;47(l):59-67. 2. Wood E, Yip B, Gataric N , et al. Determinants of geographic mobility among participants in a population-based H I V / A I D S drug treatment program. Health Place 2000;6(l):33-40. 3. Hogg RS, Heath K V , Strathdee SA, Montaner JS, O'Shaughnessy M V , Schechter MT. H I V / A I D S mortality in Canada: evidence of gender, regional and local area differentials. AIDS 1996;10(8):889-94. 4. Wood E, Chan K, Montaner JS, et al. The end of the line: has rapid transit contributed to the spatial diffusion of HIV in one of Canada's largest metropolitan areas? [In Process Citation]. Soc Sci Med 2000;51(5):741-8. 5. Patrick D M , Strathdee SA, Archibald CP, et al. Determinants of HIV seroconversion in injection drug users during a period of rising prevalence in Vancouver. International Journal ofSTD & AIDS 1997;8(7):437-45. 6. Statistics Canada. 1996 Population Census of Canada. 7. Strathdee SA, Patrick D M , Currie SL, et al. Needle exchange is not enough: lessons from the Vancouver injecting drug use study. AIDS 1997;ll(8):F59-65. 8. Strathdee SA, Palepu A , Cornelisse PG, et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 9. Andersen R, Bozzette S, Shapiro M , et al. Access of vulnerable groups to antiretroviral therapy among persons in care for HIV disease in the United States. H C S U S Consortium. HIV Cost and Services Utilization Study. Health Serv Res 2000;35(2):389-416. 9 0 10. Jacobson M A , French M . Altered natural history of AIDS-related opportunistic infections in the era of potent combination antiretroviral therapy. AIDS 1998;12(Suppl A):S157-63. 11. Moore RD, Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS 1999;13(14):1933-42. 12. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 13. Hogg RS, O'Shaughnessy M V , Gataric N , et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997;349(9061):1294. 14. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver's West End. Can J Public Health 2000;91(2):125-8. 15. Carrieri MP, Moatti JP, Vlahov D, Obadia Y, Reynaud-Maurupt C, Chesney M . Access to antiretroviral treatment among French HIV infected injection drug users: the influence of continued drug use. M A N I F 2000 Study Group. / Epidemiol Community Health 1999;53(l):4-8. 16. The Futures Group International. DemProj Version 4.0. Spectrum Policy Project, 1997. 17. Low-Beer S, Bartholomew K, Weber A E , et al. A demographic and health profile of gay and bisexual men in a large Canadian urban setting. AIDS Care 2002;14(l):lll-5. 18. Stover J, Way P. Projecting the impact of AIDS on mortality. AIDS 1998;12 Suppl l:S29-39. 19. Health Canada (1988). Consortium to characterize injection drug users in Canada (Montreal, Toronto, and Vancouver). 9 1 20. Wood E, Yip B, Gataric N , et al. The impact of HIV care and support on the mobility of persons taking anti-HlV therapy. Can J Infect Dis 1999;10(Suppl B):48B. 21. Bacchetti P, Moss AR. Incubation period of AIDS in San Francisco. Nature 1989;338(6212):251-3. 22. Chevret S, Costagliola D, Lefrere JJ, Valleron AJ. A new approach to estimating AIDS incubation times: results in homosexual infected men. / Epidemiol Community Health 1992;46(6):582-6. 23. Veugelers PJ, Cornelisse PG, Craib KJ, et al. Models of survival in HIV infection and their use in the quantification of treatment benefits. Am J Epidemiol 1998;148(5):487-96. 24. Hogg RS, Rhone SA, Yip B, et al. Antiviral effect of double and triple drug combinations amongst HIV- infected adults: lessons from the implementation of viral load-driven antiretroviral therapy. AIDS 1998;12(3):279-84. 25. Tyndall M W , Craib KJ, Currie S, Li K, O'Shaughnessy M V , Schechter M T . Impact of HIV infection on mortality in a cohort of injection drug users. / Acquir Immune Defic Syndr 2001;28(4):351-7. 26. United Nations Populations Division (1996). World Populations Prospects, the 1996 Revision. 27. Hogg RS, Whitehead J, Ricketts M , et al. Patterns of geographic mobility of persons with AIDS in Canada from time of AIDS index diagnosis to death. Clin Invest Med 1997;20(2):77-83. 28. U.S. Agency for International Development (1988). World Population Profile. 29. Wood E, Yip B, Hogg RS, et al. Full suppression of viral load is needed to achieve an optimal CD4 cell count response among patients on triple drug antiretroviral therapy. AIDS 2000;14(13):1955-60. 92 30. Wood E, Hogg RS, Yip B, et al. \"Discordant\" increases in CD4 cell count relative to plasma viral load in a closely followed cohort of patients initiating antiretroviral therapy. / Acquir Immune Defic Syndr 2002;30(2):159-66. 31. Quinn T C , Wawer MJ, Sewankambo N , et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group [see comments]: New England Journal of Medicine 2000;342(13):921-9. 32. Anis A H , Hogg RS, Wang X H , et al. Modelling the potential economic impact of viral load-driven triple drug combination antiretroviral therapy. Pharmacoeconomics 1998;13(6):697-705. 33. Cook J, Dasbach E, Coplan P, et al. Modeling the long-term outcomes and costs of HIV antiretroviral therapy using HIV R N A levels: application to a dinical trial. AIDS Res Hum Retroviruses 1999;15(6):499-508. 34. Palepu A , Strathdee SA, Hogg RS, et al. The social determinants of emergency department and hospital use by injection drug users in Canada. Journal of Urban Health 1999;76(4):409-18. 35. Palepu A , Tyndall M W , Leon H , et al. Hospital utilization and costs in a cohort of injection drug users. C M A / 2001;165(4):415-20. 36. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999;48(12):1751-8. 93 Table 5.1 Potential life expectancies in the Downtown Eastside during the year 2006, given three levels of HIV prevalence. Life Expectancy Life Expectancy Life Expectancy Cumulative Males Females Overall AIDS Deaths 5% Prevalence 80% ART* use 60.8 72.8 65.4 171 20% ART* use 57.6 69.9 62.4 497 7% Prevalence 80% ART* use 60.8 72.8 65.4 172 20% ART* use 56.9 68.6 61.4 503 9% Prevalence 80% ART* use 60.4 72.0 64.9 185 20% ART* use 55.6 65.8 59.6 556 *Percent actively receiving antiretroviral therapy. 9 4 L E G E N D F O R F I G U R E S Figure 5.1 The change in overall life expectancies at birth projected in scenarios one and two with a 7% prevalence. A R T refers to the percentage of HIV-positive individuals actively receiving antiretroviral therapy. Figure 5.2 Annual AIDS deaths projected in scenarios one and two with a 7% prevalence. ART refers to the percentage of HIV-positive individuals actively receiving antiretroviral therapy. Figure 5.3 The annual AIDS death rate projected in scenarios one and two with a 7% prevalence. ART refers to the percentage of HIV-positive individuals actively receiving antiretroviral therapy. 9 5 Figure 5.1 Figure 5.2 ioo - i 1999 2000 2001 2002 2003 2004 2005 2006 9 7 Figure 5.3 D T 1 I 1 I 1 I 1 1 1 1 1 I 1 I 1999 2000 2001 2002 2003 2004 2005 2006 Year 9 8 C H A P T E R 6 M O D E L L I N G T H E I M P A C T O F I M P R O V E D U P T A K E O F T R E A T M E N T II 6.1 F O R W A R D This chapter was fast tracked for publication in the Lancet as: Wood E, Braitstein P, Schechter M T , Tyndall M , Montaner JSG O'Shaughnessy M V , Hogg RS. Extent to which low-level antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa. Lancet 2000; 355: 2095-2100. It was followed by the a letter of correspondence: Wood E, Braitstein P, Montaner JS, O'Shaughnessy M V , Hogg RS. H I V / A I D S Epidemiology. Lancet 2000;356(9238):1357-8. 99 6.2 A B S T R A C T Objective: Despite growing international pressure to provide HIV-1 treatment to less-developed countries, potential demographic and epidemiological impacts have yet to be characterized. We modeled the future impact of antiretroviral use in South Africa from 2000 to 2005. Methods: We produced a population projection model that assumed zero antiretroviral use to estimate the future demographic impacts of the HIV-1 epidemic. We also constructed four antiretroviral-adjusted scenarios to estimate the potential effect of antiretroviral use. We modeled total drug cost, cost per life-year gained, and the proportion of per-person health-care expenditure required to finance antiretroviral treatment in each scenario. Results: With no antiretroviral use between 2000 and 2005, there will be about 276,000 cumulative HIV-1-positive births, 2,302,000 cumulative new AIDS cases, and the life expectancy at birth will be 46.6 years by 2005. By contrast, 110,000 HIV-l-positive births could be prevented by short-course antiretroviral prophylaxis, as well as a decline of up to 1 year of life expectancy. The direct drug costs of universal coverage for this intervention would be US$54 million\u00E2\u0080\u0094less than 0.001% of the per-person health-care expenditure. In comparison, triple-combination treatment for 25% of the HIV-l-positive population could prevent a 3.1-year decline in life expectancy and more than 430,000 incident AIDS cases. The drug costs of this intervention would, however, be more than $19 billion at present prices, and would require 12.5% of the country's per-person health-care expenditure. Conclusions: Although there are barriers to widespread HIV-1 treatment, in particular the use of H A A R T , limited use of antiretrovirals could have an immediate and substantial impact on South Africa's AIDS epidemic. 100 6.3 I N T R O D U C T I O N Antiretroviral treatment suppresses HIV-1 replication, and significantly alters the natural history of the infection.!' 2 In more-developed countries, widespread use of antiretroviral regimens have substantially reduced AIDS-related morbidity and mortality, and the use of antiretroviral prophylaxis among HIV-l-positive pregnant women has almost eliminated perinatal transmission.34 Trials done in Africa among breastfeeding women have shown the effectiveness of short-course antiretroviral regimens for preventing vertical transmission of HIV-1.5-8 Of all the advances, this finding may be of greatest relevance to less-developed countries, where the cost of antiretrovirals have made triple drug treatment inaccessible to most of those infected.9 The most devastating HIV-1 and AIDS epidemic is in sub-Saharan Africa. Already, AIDS mortality has had a severe demographic impact, and population projections suggest that AIDS mortality will lead to further declines in the region's health status. 10 These models have, however, been calculated under the assumption that there will be zero antiretroviral use. It has been suggested that HIV-1 treatment must be provided to help curb Africa's AIDS epidemic, and there has been mounting pressure locally and internationally to make antiretrovirals accessible.H'12 However, the potential demographic and epidemiological benefits of antiretroviral provision have yet to be characterized for the region. We undertook this study to model the potential demographic and epidemiological impact of antiretroviral use on South Africa's AIDS epidemic from 2000 to 2005 for different regimens, as well as effects on health-care costs. 101 6.4 M E T H O D S We created models to project future HIV-l-positive births, new AIDS cases, and life expectancy in the Republic of South Africa from 2000 to 2005. The values assigned to model parameters in each scenario were based, where possible, on empirical data. We developed the models with Spectrum software (version 4.0) by standard demographic techniques described previously.! 3 Briefly, baseline population numbers, stratified by age and sex, were acquired from the US Census Bureau's international database.^4 We acquired total fertility-rate data from the US Census Bureau's World Population Profile. 15 The age distribution of fertility was estimated from the United Nations sub-Saharan Africa model fertility table. We estimated the current and future mortality rates from life expectancy data and model life tables.15 These parameters were constant in all scenarios. In addition, we used Spectrum's AIDS Impact model to adjust the population projections for current and projected HTV-l-associated mortality. In the first model, scenario zero, annual HIV-l-positive births, new AIDS cases, and life expectancy were estimated with the assumption of negligible antiretroviral use. In this model, the perinatal transmission rate was set at 30%, based on a consensus which suggested that the absolute rate of mother-to-child transmission among breastfeeding women ranges from 25-35%.l 4 The duration from HIV-1 infection to AIDS and death among adults and infants was based on previous estimates for sub-Saharan Africa.I 6 We assumed that adult elapsed time from HIV-1 infection to AIDS was normally distributed, with a median of 8.0 years (SD 3.5), as was the duration from AIDS to death, with a median of 1 year.16 50% of HIV-l-positive infants, infected at birth or during breastfeeding, were 102 assumed to develop AIDS after 2 years, with 95% developing AIDS by 6 years after infection.16 Future HIV-l-prevalence estimates were derived from the US Census Bureau's World Population Profile.14 Despite the potential of antiretrovirals to reduce horizontal transmission of HIV-1, prevalence was unaffected by their use in all models. 17,18 Similarly, in all models we kept the non-HIV-l-related mortality constant at baseline levels for the whole study period. We modeled scenarios in which the HlV-l-specific parameters were antiretroviral adjusted to estimate the impact of low-level antiretroviral use. First, the findings of trials of orally administered short-course antiretroviral regimens to prevent vertical transmission of HIV-1 were used to estimate the reductions in vertical transmission that could be achieved through various degrees of antiretroviral prophylaxis.5-8 Those trials have reported reductions in HIV-1 transmission from 37% to 52%.5~8 To allow for reduced efficacy because of poor adherence, subsequent HIV-1 infection attributable to breastfeeding, or both, as well as increased preventive effectiveness because of improved adherence, early weaning from breast milk and formula substitution, or both, sensitivity analyses were applied to this parameter. We assessed the impact of prophylaxis use, given a reduction in vertical transmission among mothers who access the intervention of 40% with a range of estimates from 30% to 50%. Second, we modeled the epidemiological and demographic impacts of the use of triple-combination treatment among HFV-l-positive non-pregnant adults. Although the survival benefits of antiretroviral treatment have yet to be quantified, an analysis has suggested life expectancy could increase by 7 years among injecting drug users.l^ 1 0 3 We applied a sensitivity analysis to this parameter and assessed the impact of treatment, assuming that the lives of adults who access the intervention would be extended by a median of 6 years (range 5-7). Four antiretroviral-adjusted scenarios were modeled. In scenario one, we assumed that 25% of all HIV-l-positive pregnant women and infants would receive antiretroviral prophylaxis over the whole study period to prevent vertical transmission. In scenario two, 75% would use antiretroviral prophylaxis. In scenario three, 100% prophylaxis would be provided to all pregnant women irrespective of HIV-1 serostatus. Although concerns have been raised about universal treatment without testing or counseling,^ we modeled this scenario since it has been argued that inability to provide voluntary testing and counseling may translate into no treatment.2'-' In the fourth scenario, we modeled the impact of triple-combination antiretroviral treatment of 25% of non-pregnant HIV-l-positive adults. To assess the demographic impact of this intervention, the perinatal transmission rate was kept at 30% in this scenario. For each scenario we calculated HIV-l-positive births, incident AIDS cases, and life expectancy over the study period. We calculated HIV-l-positive births based on a reduction of perinatal transmission from 30% to 50%. Other indices were robust to this parameter, and reported values are based on the assumption of a 40% reduction in perinatal transmission. We did an economic analysis to estimate the direct drug costs of antiretroviral provision in all scenarios, by methods described previously.^ Briefly, model-derived total antiretroviral-drug costs were determined by multiplying the average cost of treatment by the estimated number of pregnant women, neonates, and non-pregnant 104 adults eligible for treatment in the projected year of observation. Estimates of drug costs reflect the UN-brokered preferential pricing strategy.^l By use of Monte Carlo simulation methodology, we did sensitivity analyses by varying the key model parameters. For each Monte Carlo trial, 10,000 iterations were run. For each iteration in a trial, a random number was generated for the cost of short-course antiretroviral prophylaxis, the cost of triple-combination treatment, and the number of individuals eligible for treatment. Each random number conformed to the preset probability distribution that was used to describe the potential uncertainty in each model parameter. We assumed the cost of short-course prophylaxis had a right-skewed Weibull distribution with a median of US$8 and a 95th percentile of $100 to represent the range of drug costs in reported trials. This probability distribution favored the low-cost single-dose H I V N E T 012 nevirapine regimen/ but allowed for the higher-cost multidose Zidovudine and lamivudine regimen used in the PETRA study.8 We used prevalence and fertility estimates to calculate the number of HIV-l-positive pregnant women over the study period. This estimate was assumed to be normally distributed with SD of 20,000. Based on reported annual percentages of people on antiretroviral therapy in Canada and the USA, we assumed that 25% of HIV-l-positive South Africans would access antiretroviral treatment per year. 9 Annual triple-drug cost was estimated to be $2900 per person, and we assumed the annual cost of triple-combination treatment was normally distributed with an SD of $200. This cost represents the high end of the range ($8) of reported estimates for the daily cost of triple-combination treatment with a 105 protease inhibitor, according to the U N preferential pricing strategy.2J- Similarly, we assumed the total number of HIV-l-infected adults in South Africa over the study period was normally distributed and with an SD of 1,000,000. Al l dollar figures are expressed in year 2000 US$ currency. Proportion of per-person health-care expenditure was calculated by dividing the per-person cost of antiretrovirals by the per-person health-care expenditure. 2 2 i n addition, we calculated the cost per life-year gained in each scenario, which was estimated by dividing the total costs of antiretrovirals by the number of life-years gained over the 5-year population projection period. Life-years were calculated by comparing scenarios one to four with scenario zero. Model inputs are summarized in Table 6.1. 6.5 RESULTS A n estimated 22,186,700 women and 21,975,000 men live in the Republic of South. Africa in the year 2000; it is estimated that around 12.0% are infected with HIV-1 and that 16.0% will be infected by 2005.14,15 i f m e situation of negligible antiretroviral use (scenario zero) persists, there will be 2,302,000 incident AIDS cases and 276,000 H I V - l -positive births between 2000 and 2005, and the life expectancy at birth will be 46.6 years by 2005. The life expectancy projected in this scenario is similar to the US Census Bureau and the U N projections for 2005.15/16 Estimates of the total number of HIV-l-positive births, new AIDS cases, and life expectancy at birth for each scenario are presented in Table 6.2. In scenario one (25% prophylaxis in pregnant women), 230,000 HIV-l-positive pregnant women and neonates would use antiretroviral prophylaxis over the study period, the cumulative 106 number of HIV-l-positive births would be 248,000 (range 241,000-255,000), there would be 2,271,000 cumulative AIDS cases, and the life expectancy at birth would be 46.8 years by 2005. In scenario two (75%), 689,000 HIV-l-positive pregnant women and neonates would use antiretroviral prophylaxis over the study period, the cumulative number of HIV-l-positive births would be 193,000 (172,000-213,000), there would be 2,208,000 cumulative AIDS cases, and the life expectancy at birth would be 47.2 by 2005. If antiretroviral prophylaxis were provided to all of the 6,500,000 women projected to give birth over the study period (scenario three), the cumulative number of H I V - l -positive births would be 166,000 (138,000-193,000), there would be 2,177,000 cumulative AIDS cases, and the life expectancy at birth would be 47.5 years by 2005. In scenario four (triple-combination treatment in 25% of HIV-l-positive adults), there would be 1 871,000 (1,583,000-2,190,000) AIDS cases over the study period and the life expectancy would be 49.7 years (47.4-51.9) at birth in 2005. Scenario four would have the most substantial epidemiological impact over the study period, and would prevent 431,000 new AIDS cases. The estimate of at least 140,000 AIDS cases prevented by the use of antiretroviral prophylaxis modeled in scenario three would, however, be because of prevented HIV-1 infections, and would not represent a delay in the onset of AIDS, as would occur in scenario four. Unless a more effective short-course prophylaxis is developed, annual H I V - l -positive births will increase over the study period, irrespective of antiretroviral use, because of the increasing HIV-1 prevalence and limited efficacy of the regimens we have modeled (Figure 6.1). Steep declines in HIV-l-positive births in each year would, however, result from increasing prophylaxis use, and the HIV-l-positive births in 2000 107 in scenario zero would still exceed the HIV-l-positive births in 2005 in scenario three (Figure 6.1). The use of triple-combination treatment would have the most striking demographic impact, resulting in a life-expectancy gain of 3.1 years that would be sustained for the entire study period. In comparison, short-course antiretroviral prophylaxis could improve life expectancy by 1 year over the study period (Figure 6.2). Although increasing prevalence will result in declining life expectancy in all scenarios of prophylactic treatment, the life expectancy in 2005 modeled in scenario three would still be higher than that for the 2000 in scenario zero. Non-HIV-l-related mortality was fixed at baseline rates in all models, and changes in life expectancy would be due solely to reductions in AIDS deaths. For each scenario, the direct drug costs, proportion of per-person health-care expenditure, and cost per life-year gained for the whole study period would be: scenario one $1,838,000, less than 0.001%, and $19; scenario two, $5,514,000, less than 0.001%, and $19; scenario three, $52,160,000, less than 0.001%, and $133; and scenario four, $19 billion, 12.5%, and $15,000 (Table 6.3). Since the costs and benefits of targeted prophylaxis are directly related, irrespective of the degree of coverage, the cost per life-year gained is the same in scenarios one and two. 6.6 DISCUSSION Short-course antiretroviral prophylaxis that would reduce perinatal transmission by 40% could prevent up to 110,000 cumulative infant HTV-l-infections by 2005. This reduction was seen in the universal prophylaxis model, but targeted provision to as few 108 as 25% of HIV-l-positive women could prevent 28 000 HIV-l-infections in infants. The direct drug cost for antiretroviral prophylaxis would be less than 0.001% of the per-person health expenditure. The triple-combination treatment for 25% of HIV-l-positive adults could increase life expectancy by 3.1 years and prevent more than 430 000 AIDS cases over the study period, but with greatly disproportionate costs and expenditure. Clearly, the cost of antiretrovirals will have to be further reduced before antiretroviral treatment becomes cost-effective in this context. The 1-year life expectancy gained in scenario three represents a substantial demographic effect. In Canada, for example, all lung-cancer deaths lead to 0.9 years of life expectancy lost among men, and breast cancer to 0.5 years among all women.23 Provision of short-course antiretrovirals could, therefore, have a demographic impact similar to that from elimination of all lung-cancer deaths in Canada, after adjustment for competing risks. Although model four suggests that 25% treatment coverage could gain more than 3 years of life-expectancy, the scale of drug provision is not comparable, since fertility and prevalence estimates suggest that 919,000 HIV-l-positive women will become pregnant between 2000 and 2005, whereas coverage with 25% antiretroviral treatment would represent more than 6.5 million person-years of daily combination treatment over the study period. Short-course antiretrovirals can effectively reduce rates of vertical transmission, 7^ D U t the regimens tested in less-developed countries are of an inferior standard to those used in more-developed countries and have less effect on vertical-transmission rates. Although we did not assess cost effectiveness, short-course 109 regimens were shown to be cost effective in Africa before the preferential pricing strategy announcement . 2 ^ 2 6 Feasibility must, however, be taken into account, as well as cost effectiveness, in health-policy decisions. Although primary-health-care clinics and hospitals in urban and rural South Africa may have the capacity to do rapid testing and administer short-course treatment, 2 5 further investment and education might be required to reach the number of pregnant women who make antenatal-clinic visits modeled in scenario two. Only around 50% of HIV-l-positive women return to get their test results, which makes reaching the women most in need difficult 2 7 Furthermore, it is doubtful whether the human resources and infrastructure are in place to provide the amount of triple-combination treatment that we modeled in scenario four. In addition to the medical staff and laboratory equipment that would be required to ensure the success of this program, additional infrastructure will be required to ensure confidentiality, security, and adequate drug distribution. Given the limited resources and competing health concerns in sub-Saharan Africa, priority must be given to interventions that can be most easily administered and have the greatest cost effectiveness. Our analyses show that, despite price reductions, interventions other than triple-combination treatment will probably be more cost effective, such as treatment of symptomatic sexually transmitted diseases among prostitutes to prevent horizontal HIV-1 transmission, and public-health interventions such as expanded immunization coverage.2^ Currently, the most cost effective, preventatively effective, and most easily administered antiretroviral agent is probably single-dose nevirapine/ A report of fatal 1 1 0 side-effects has, however, cast doubt on the safety of nevirapine prophylaxis in South A f r i c a , 2 9 but extensive surveillance of this drug in more-developed and less-developed countries has confirmed the overall safety profile seen in the original clinical trials,30 and safety data from about 700 mother-infant pairs has shown no important adverse reactions with single-dose nevirapine7/31,32 Although confirmation of the H I V N E T 012 findings is necessary, nevirapine may reduce transmission rates by more than the estimates we have modeled, since the reported 47% reduction was compared with the 14-dose intrapartum-postpartum Zidovudine group. The additional life expectancy associated with triple-combination treatment is not yet known. Two studies that outlined potential future scenarios attempted to model the potential epidemiological and demographic impacts of antiretroviral treatment. In the first study, which assessed the use of antiretrovirals among homosexual and bisexual men, antiretroviral treatment was estimated to provide an additional 6-24 years to an individual's life expectancy after infection, and the researchers concluded that the use of antiretrovirals could substantially lower the incidence of infection .33 i n the second study, which assessed antiretroviral use among injecting drug users, modern antiretrovirals were estimated to increase the median duration from HIV-1 infection to death by 7 years. 13 In these two analyses, even if HIV-1 prevalence increased, antiretrovirals could substantially reduce the burden of morbidity and mortality. The conclusions of the two studies were not attributable to reduced perinatal transmission. In HIV-1 endemic areas of more-developed countries, such as Vancouver, Canada, and San Francisco, USA, combination antiretroviral treatment is used by 111 around 50% of HIV-1-infected people (BC Centre for Excellence in HIV/AIDS, unpublished data). 3 3 Striking improvements have been reported in health status and life expectancy and coincide with the widespread availability of antiretroviral therapy.34,35 Health-status improvements attributed to antiretroviral therapy have occurred despite growing HIV-1 prevalence. By contrast, almost no triple-combination HIV-1 treatment is used in sub-Saharan Africa. Although short-course antiretroviral prophylaxis could prevent thousands of neonatal HIV-1 infections, more complex prophylaxis regimens and the use of triple-combination antiretroviral therapy would lead to greater demographic benefits. Such treatment might, however, cost as much as $10,000 dollars per life-year saved after adjustment for disability, 3 6 and our findings suggest that the drug cost of treating 25% of HIV-l-positive South Africans would be more than $19 billion. Clearly, drug cost remains a key barrier to the implementation of widespread antiretroviral provision. Although price reductions will probably increase the use of antiretrovirals in less-developed countries, major concerns must be addressed\u00E2\u0080\u0094for example, resistant viral strains emerge quickly in non-adherent individuals, which has implications for future treatment options and the transmission of resistant v i r u s . 3 7 The long-term benefits of expanded antiretroviral use in less-developed countries will rely on improvements in health-care infrastructure in addition to the availability of antiretroviral drugs. Several additional benefits may accrue from the use of antiretroviral treatments that were not accounted for in our models. HIV-1 testing and counseling that should accompany the targeted provision of antiretrovirals might provide secondary benefits, such as reduced sexual transmission. HIV-1 viral load is the main predictor of the risk of heterosexual transmission of HIV-1, and transmission is rare among people who 112 have viral loads of less than 1500 copies/mL of HIV-1 RNA.17 Since the benefits of triple-combination treatment are directly attributable to the degree of viral-load suppression they can induce, 3 such regimens could result in reduced horizontal HIV-1 transmission. Our models may, therefore, underestimate the benefits of antiretroviral use. Achievement of the degree of antiretroviral coverage that we have modeled will require investments in health-care infrastructure that we did not model. Non-drug costs associated with the delivery of the interventions we modeled might be as high as drug costs, but will depend on many factors, including one-time infrastructure costs, as well as the degree of centralization of services. When population projections are developed, several parameters, such as future HIV-1 prevalence, must be estimated, and output values are, therefore, estimates. The scenarios we developed were hypothetical and some assumptions could have introduced certain biases. Most importantly, we assumed that HIV-1 prevalence in South Africa would increase substantially over the study period in all models. A widespread educational campaign could lower HIV-1 incidence and invalidate our models. We assumed also that short-course antiretrovirals could lower infant HIV-1 infections by 30-50% among breastfeeding women. These estimates were based on several clinical trials in which overall compliance with the short-course regimen was quite poor . 5 / 6 For this and other reasons noted, this range might be a conservative estimate of the potential reductions in H I V - l transmission that could be attained through the use of short-course antiretrovirals, especially if early weaning and formula substitution are a safe alternative to breastfeeding.38 113 6.7 S U M M A R Y At present, there are major barriers to the widespread provision of antiretroviral therapy in South Africa, including limited health-care infrastructure and drug costs. While drug cost remains a major impediment to the use of H A A R T in this setting, our estimates suggest that the drug cost of antiretroviral prophylaxis is low and that low-level prophylaxis use could have a substantial demographic and epidemiological impact. 114 6.8 REFERENCES 1. Wood E, Yip B, Hogg RS, et al. Full suppression of viral load is needed to achieve an optimal CD4 cell count response among patients on triple drug antiretroviral therapy. AIDS 2000;14(13):1955-60. 2. Phair JP. Determinants of the natural history of human immunodeficiency virus type 1 infection. / Infect Dis 1999;179(Suppl 2):S384-6. 3. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 4. IAS position paper on prevention of HIV 1 mother-to-child transmission. International Aids Society. AIDS 1999;13(ll):5-9. 5. Dabis F, Msellati P, Meda N , et al. 6-month efficacy, tolerance, and acceptability of a short regimen of oral Zidovudine to reduce vertical transmission of HIV in breastfed children in Cote d'lvoire and Burkina Faso: a double-blind placebo-controlled multicentre trial. DITRAME Study Group. Diminution de la Transmission Mere-Enfant. Lancet 1999;353(9155):786-92. 6. Wiktor SZ, Ekpini E, Karon JM, et al. Short-course oral Zidovudine for prevention of mother-to-child transmission of HIV-1 in Abidjan, Cote d'lvoire: a randomised trial. Lancet 1999;353(9155):781-5. 7. Guay L A , Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with Zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: H I V N E T 012 randomised trial [see comments]. Lancet 1999;354(9181):795-802. 8. Saba J. Current status of PETRA study. 2nd Conference on Global Strategies for the Prevention of HIV Transmission from Mothers to Infants 1999, Montreal, Quebec. 115 9. Hogg RS, Weber A E , Craib KJ, et al. One world, one hope: the cost of providing antiretroviral therapy to all nations. AIDS 1998;12(16):2203-9. 10. Stover J, Way P. Projecting the impact of AIDS on mortality. AIDS 1998;12 Suppl l:S29-39. 11. Science, ethics, and the future of research into maternal infant transmission of HIV-1. Perinatal HIV Intervention Research in Developing Countries Workshop participants [see comments]. Lancet 1999;353(9155):832-5. 12. De Cock K M , Fowler M G , Mercier E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice [In Process Citation]. JAMA 2000;283(9):1175-82. 13. Wood E, Schechter M T , Tyndall M W , Montaner JS, OShaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000;14(9):1229-35. 14. U.S. Bureau of the Census: International Database (2000). 15. U.S. Agency for International Development (1988). World Population Profile. 16. Secretariat U N . The Demographic Impact of H I V / A I D S . New York: Population Division Department of Economic and Social Affairs United Nations Secretariat, 1999. 17. Quinn T C , Wawer MJ, Sewankambo N , et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group [see comments]. New England Journal of Medicine 2000;342(13):921-9. 18. Vernazza PL, Gilliam BL, Flepp M , et al. Effect of antiviral treatment on the shedding of HIV-1 in semen [see comments]. Aids 1997;ll(10):1249-54. 116 19. Campa A M , Shor-Posner G, Baum M K . H I V N E T nevirapine trials. Lancet 1999;354:1817. 20. Marseille E, Kahn JG, Guay L, Fowler M G , Jackson JB. H I V N E T nevirapine trials. Lancet 1999;354:1817. 21. UNAIDS Press Release. New public/private sector effort to initiated to accesserate acces to H I V / A I D S care adn treatment in developing countries. May, 2000. 22. World Development Indicators 2000. Washington, D C : World Bank 2000. 23. Hogg RS, Whitehead J, Ricketts M , et al. Patterns of geographic mobility of persons with AIDS in Canada from time of AIDS index diagnosis to death. Clin Invest Med 1997;20(2):77-83. 24. Marseille E, Kahn JG, Mmiro F, et al. Cost effectiveness of single-dose nevirapine regimen for mothers and babies to decrease vertical HIV-1 transmission in sub-Saharan Africa [see comments]. Lancet 1999;354(9181):803-9. 25. Soderlund N , Zwi K, Kinghorn A , Gray G . Prevention of vertical transmission of HIV: analysis of cost effectiveness of options available in South Africa [see comments]. BMJ 1999;318(7199):1650-6. 26. Marseille E, Kahn JG, Saba J. Cost-effectiveness of antiviral drug therapy to reduce mother-to-child HIV transmission in sub-Saharan Africa. AIDS 1998;12(8):939-48. 27. Mclntyre JA, Gray G E . Acceptability of voluntary HIV testing by pregnant women in developing countries: an international survey. Int Conf AIDS 1988;12:406 (abstr 23322). 28. Gilson L, Mkanje R, Grosskurth H , et al. Cost-effectiveness of improved treatment services for sexually transmitted diseases in preventing HIV-1 infection in Mwanza Region, Tanzania. Lancet 1997;350(9094):1805-9. 117 29. Anon. Dorctors slam minister for blaming deaths on drug. Sunday Times of South Africa 2000, April 9. 30. Pollard RB, Robinson P, Dransfield K. Safety profile of nevirapine, a nonnucleoside reverse transcriptase inhibitor for the treatment of human immunodeficiency virus infection. Clin Ther 1998;20(6):1071-92. 31. Musoke P, Guay L A , Bagenda D, et al. A phase I/II study of the safety and pharmacokinetics of nevirapine in HIV-l-infected pregnant Ugandan women and their neonates (HIVNET 006). Aids 1999;13(4):479-86. 32. Mirochnick M , Fenton T, Gagnier P, et al. Pharmacokinetics of nevirapine in human immunodeficiency virus type 1-infected pregnant women and their neonates. Pediatric AIDS Clinical Trials Group Protocol 250 Team. / Infect Dis 1998;178(2):368-74. 33. Blower SM, Gershengorn HB, Grant R M . A tale of two futures: HIV and antiretroviral therapy in San Francisco. Science 2000;287(5453):650-4. 34. Wood E, Low-Beer S, Bartholomew K, et al. Modern antiretroviral therapy improves life expectancy of gay and bisexual males in Vancouver's West End. Can J Public Health 2000;91(2):125-8. 35. Vittinghoff E, Scheer S, O'Malley P, Colfax G, Holmberg SD, Buchbinder SP. Combination antiretroviral therapy and recent declines in AIDS incidence and mortality. / Infect Dis 1999;179(3):717-20. 36. Organization U a W H . Guidance modules on antiretroviral treatments-module 2. Introducing antiretroviral treatments into health systems: economic considerations. Geneva, 1998. 37. Gallant JE. Strategies for long-term success in the treatment of HIV infection. JAMA 2000;283(10):1329-34. 118 38. Nduati R, John G, Mbori-Ngacha D , et al. Effect of breastfeeding and formula feeding on transmission of HIV-1: a randomized clinical trial. JAMA 2000;283(9):1167-74. 119 Table 6.1 Model parameters, data sources, and values used in models. Parameter Source Values Used Demographic specific data Population data US Census Bureau 1 4 Baseline life expectancy United Nations 1 6 Age distribution of non-HIV mortality Coale-Demeny Life Table Total fertility rate HIV-1 prevalence per year Perinatal transmission rate Time H I V infection to AIDS Efficacy of prophylaxis Efficacy of treatment Economic analysis Antiretroviral prophylaxis Triple combination treatment Health-care expend, per person US Census Bureau 1 5 U S Census Bureau 1 5 De Cock et a l 1 2 United Nations 1 6 Dabis, Wiktor, Guay, Saba 6\" 8 W o o d et a l 1 3 U N A I D S press release1 6 U N A I D S press release1 6 World Bank 2 5 Age and sex specific 47.5 years Age specific distribution 3.0 in 2000; 2.7 by 2005 12% in 2000; 16% by 2005 30% 8 years 30-50% reduction 5-7 years US$4-100 US$2,900 US$571 1 2 0 Table 6.2 Cumulative new AIDS cases and HIV-positive births over the study period, and life expectancy in the year 2005. Antiretroviral Medication Use Cumulative HIV+ Cumulative New Projected Life Births* AIDS Cases* Expectancy 2005 Scenario Zero Scenario One 25% prophylactic use Scenario Two 75% prophylactic use Scenario Three Universal prophylactic use Scenario Four 25% Triple Combination Therapy use 276 248 (241-255) 193 (172-213) 166 (138-193) 276 2,302 2,271 2,208 2,177 46.6 46.8 47.2 47.5 1,871 (1583-2190) 49.7 (47.4-51.9) Values are provided in thousands. 121 Table 6.3 Modeled total costs, proportion of per person health-care expenditure, and cost per life-year gained required to finance antretroviral use, by scenario. Scenario Total Drug Cost (95% CI) ($US thousands) Proportion of per Person Health Expenditure Cost per Life Year (95% CI) Gained Scenario One 25% prophylactic use Scenario Two 75% prophylactic use 1,838 (1,094 - 3,141) 5,514 (3,293 - 9,408) Scenario Three Universal prophylactic use 52,160 (31,087 - 88,645) Scenario Four 25% Triple Combination Therapy use $19,074,750 (16,189,525 - 22,033,718) <0.001% <0.001% <0.001% 12.5% $19 (11 - 32) $19 (11 - 32) $133 (79 - 226) $15,000 (12,700 - 17,300) *Each Monte Carlo trial was based on 10,000 iterations. Cost per life year in scenario four rounded to nearest $100. 122 L E G E N D F O R FIGURES Figure 6.1 Annual HIV-positive births projected in scenarios zero to three, with the assumption of 40% reduction. Figure 6.2 Annual life expectancy projected in scenarios zero to four. 123 Figure 6.1 60 n Scenario 0 Scenario 1 Scenario 2 Scenario 3 0% 25% 75% 100% Prophylaxis Use (%) 124 Figure 6.2 54-1 46 H 1 1 1 1 1 ' 1 ' 1 1 1 1 1\u00E2\u0080\u0094 1999 2000 2001 2002 2003 2004 2005 Year 125 T H E L A N C E T \u00E2\u0080\u00A2 V o l 356 \u00E2\u0080\u00A2 October 14, 2000 C O R R E S P O N D E N C E C O M M E N T A R Y Sir\u00E2\u0080\u0094The report of Evan Wood and colleagues (June 17, p 2096)1 illustrates the dangers of taking a narrow medical view of the H I V / A I D S epidemic in Africa. It is inherently obvious that widespread provision of antiretroviral therapy would have an impact on life expectancy, all else being equal (which of course it isn't). A more helpful approach would be to compare the impact of such levels of expenditure on meeting basic health needs and eradicating poverty. Wood and colleagues admit that they do not take into account the costs of the extra human resources and infrastructure required to provide antiretrovirals, but to consider the immediate drug costs alone is totally unrealistic, even for effective prevention of mother-to-child transmission. The investigators should also consider the overall economic cost to a country of the huge economic outflows of expenditure on antiretrovirals, which would represent a further massive drain of wealth from the South to the North. Finally, the title of the paper is very misleading, implying that the focus of the modeling was sub-Saharan Africa as a whole, when it was in fact South Africa alone. South Africa's gross domestic product per person is much higher than that of nearly all other sub-Saharan African countries, and health expenditure is 10-20 times greater. Wood and colleagues' paper runs the risk of reversing the growing realization that the H I V / A I D S epidemic in southern Africa is a broad social, cultural, political, and economic issue, rather than a purely medical one. I urge more 126 circumspection in future before rushing papers such as this into print and giving them such prominence. Charles Todd Delegation of the European Commission in Zimbabwe, PO Box 4252, Harare, Zimbabwe, (e-mail: charles.todd.@delzwe.cec.eu.int) 1. Wood E, Braitstein P, Montaner JSG, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa. Lancet 2000; 355: 2095-100. Authors' reply Sir\u00E2\u0080\u0094We acknowledge that AIDS in Africa is a complex problem, and that a widespread campaign which addresses education, poverty, and social issues is required to substantially curb the epidemic. We believe, however, that to move policy makers frequently requires research that focuses on individual features of a complicated problem. Charles Todd argues that it is obvious that widespread access to triple-combination therapy would improve life expectancy. However, we did not merely intend to quantify the potential benefits of treatment. When it was announced that UNAIDS had brokered a deal to substantially lower the price of antiretrovirals,1 we sought to assess to what extent these reductions would make triple-combination treatment accessible to infected people in southern Africa. Our analyses suggest that, despite the price reductions, drug costs remain prohibitively high. Therefore, we did not take into account the costs of additional infrastructure required to deliver antiretrovirals in the context of 1 2 7 treatment. Second, Todd raises the point that we did not model the costs associated with the voluntary testing and counseling that should accompany a prophylaxis program. Our rationale was two-fold. First, the cost effectiveness of voluntary testing and counselling for mother-to-child transmission has already been shown. Second, in-country expenditures on these programs are likely to have a beneficial effect in addressing the root causes of the epidemic. Therefore, we did not analyse the potential benefits of voluntary testing and counselling or its costs. Last, we referred to sub-Saharan Africa in the title because the benefits of prophylaxis are probably generalizable to all H I V - l -endemic nations in the region. We agree that the expenditures on health care are not the same throughout sub-Saharan Africa, 4 and that the proportion of the health-care budget required to finance antiretroviral procurement is likely to be even higher in other nations. We believe these differences will probably have a negligible effect in terms of the cost of obtaining drugs for prophylaxis, but triple-combination treatment is, therefore, even less realistic in nations with lower health-care expenditure per person. There have been barriers to putting successful research of HIV-1 vertical transmission into policy. One barrier may be the perception that the cost of procuring drugs precludes the implementation of prophylaxis programs. Other people may doubt the benefits of short-course prophylaxis if coverage is incomplete or if subsequent breast-milk-related transmission partly erodes early benefits. We allowed for a high degree of uncertainty and used conservative reductions in transmission, and determined that these perceptions are 128 unfounded. Our findings indicate the affordability of prophylaxis, but not in the widespread use of antiretroviral therapy. Clearly, UNAIDS, the pharmaceutical industry, and donor countries must continue to work together to make antiretrovirals more affordable. *Evan Wood, Paula Braitstein, Julio S G Montaner, Michael V O'Shaughnessy, Robert S Hogg. British Columbia Center for Excellence in HIV AIDS, St Paul's Hospital, and the ^Department of Health Care and Epidemiology, Unversity of British Columbia, Vancouver, British Columbia, Canada, (email: ewood@hivnet.ubc.ca) 1. UNAIDS. New public/private sector effort initiated to accelerate access to HIV-AIDS care and treatment in developing countries. Press Release, May 11; 2000. 2. Marseille E, Kahn JG, Mmiro F, et al. Cost effectiveness of single-dose nevirapine regimen for mothers and babies to decrease vertical HIV-1 transmission in sub-Saharan Africa. Lancet 1999; 354: 803-09. 3. Allen S, Tice J, Van e Perre P, et al. Effects of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa. BMJ 1992; 304:1605-09. 4. Hogg RS, Weber A E , Craib KJ, et al. One world, one hope: the cost of providing antiretroviral therapy to all nations. AIDS 1998; 12: 2203-09. 129 CHAPTER 7 SOCIO-ECONOMIC STATUS AND ACCESS TO TRIPLE THERAPY 7.1 FORWARD This chapter was published in AIDS in 2002 as: Wood E, Montaner, JS, Chan K, Schechter M T , Tyndall M , O'Shaughnessy M V , Hogg RS. Socio-economic status, access to triple therapy, and survival from HIV Disease since 1996. AIDS 2002 Oct 18;16(15):2065-2072. 130 7.2 ABSTRACT Objectives: In the p r e - H A A R T era, socio-economic status was associated with survival from HIV disease. We have explored socio-economic status, access to triple therapy (HAART) , and mortality in the context of a universal healthcare system. Methods: We evaluated 1408 individuals who initiated double or triple therapy over the period August 1, 1996 to December 31, 1999, and who were followed until March 31, 2000. Cumulative HIV-related mortality rates were estimated using Kaplan-Meier methods, and Cox proportional hazards regression. Results: In the overall Cox model, we found that adherence (RR = 0.83; [95% CI: 0.78 - 0.89] per 10% increase), CD4 cell count (RR = 1.53 [95% CI: 1.33 - 1.77] per 100 cell decrease), and lower socio-economic status (RR = 2.19 [95% CI: 1.43 -3.35] high versus low), were associated with HIV-related mortality. However, socio-economic status was not significant among patients prescribed triple therapy in a stratified analysis (RR = 1.43 [95% CI: 0.80 -2.54]), or in a sub-analysis restricted to patients prescribed H A A R T in the initial regimen. When we investigated if inequitable access to H A A R T by socio-economic status could explain the discrepancy, we found that persons in the lower socio-economic strata were less likely to be prescribed H A A R T (OR = 0.76; [95% CI: 0.63 - 0.91]) in multi-variable analyses. Conclusions: In a universal healthcare system, socio-economic status was strongly associated with HIV-related mortality. When we investigated possible explanations for this association, we found that persons of lower socio-economic status were less likely to receive H A A R T after adjustment for clinical characteristics. Our findings highlight the need for monitoring of therapeutic guidelines to ensure equitable access, as treatment strategies are updated. 131 7.3 INTRODUCTION Since its introduction in 1996, the benefits of highly active antiretroviral therapy (HAART) in the management of HIV disease have been well established. New antiretroviral regimens have proven effective in decreasing plasma viral load, improving CD4 cell counts, and have substantially altered the natural history of HIV infection. 1/2 As a result, substantial improvements in HIV-related morbidity and mortality have been documented among persons on antiretroviral therapy.3'^ Prior to the advent of H A A R T , we identified a survival gradient by socioeconomic status among HIV-l-infected gay men.5/6 This finding was particularly interesting due to the fact that that the data were derived in a universal healthcare setting, and could not be explained by differences in traditional prognostic markers such as age or CD4 cell count, or differential use of Zidovudine or Pneumocystis carinii pneumonia (PCP) prophylaxis. Similarly, other studies conducted prior to the advent of H A A R T identified gender and ethnic gradients in survival from HIV disease, as well as disparities in survival between risk groups/ '^ although in some studies the social disparities in survival could be attributed to inequitable access to Zidovudine or P C P prophylaxis.9/10 More recently, it was observed that differences in survival from AIDS by socioeconomic status were emerging in the era of H A A R T , possibly due to inequitable access to antiretroviral treatment. H However, it is presently unknown if socioeconomic disparities in HIV-related mortality exist among persons who are accessing antiretroviral treatment. Therefore, the present study was conducted to determine if socio-economic status was associated with HIV-l-related mortality since the advent and 132 widespread provision of H A A R T among participants in a province-wide treatment program that operates within a universal healthcare setting. 7.4 METHODS In British Columbia anti-HIV medications are centrally distributed at no cost to eligible HIV-infected individuals through the BC Centre for Excellence in H I V / A I D S Drug Treatment Program. The Centre distributes antiretroviral drugs based on specific guidelines generated by the Therapeutic Guidelines Committee. In June 1996 the Centre adopted plasma viral load driven antiretroviral therapy guidelines, consistent with those put forward by the International AIDS Society - USA.12 i n brief, antiretroviral therapy naive individuals with plasma viral load > 100,000 copies/mL were offered triple drug regimens (i.e. two nucleosides plus a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor), while those with plasma viral loads from 5,000 to 100,000 copies/mL were offered dual nucleoside therapy. In July of 1997 the guidelines were revised to recommend triple combination therapy for all antiretroviral naive individuals with plasma viral loads > 5,000 copies/mL or a CD4 cell count below 500 cel ls /mm 3 . Plasma H I V - R N A was measured using the Amplicor HIV-1 M o n i t o r \u00E2\u0084\u00A2 manufactured by Roche Diagnostic Systems (Branchburg, NJ). Participants meeting these criteria are enrolled in the program by their general practitioner on submission of a prescription for antiretroviral or approved antimicrobial medications. The Centre's Drug Treatment Program remains the only source of free antiretroviral medications in the province of British Columbia and < 1% of HIV-infected British Columbians purchase antiretroviral therapy outside the program.13 133 Physicians enrolling an HIV-positive individual into the Centre's Drug Treatment Program must complete a drug request enrollment form. The enrollment form acts as a legal prescription and provides information on the HIV-positive applicant's address and enrolling physician, past HIV-specific drug history, CD4 cell counts, and current drug requests. Approved prescriptions are renewed every two months. At the time of the initial refill, each participant is asked to complete an enrollment survey and informed consent form and their physician is asked to complete a clinical staging form. Thereafter, participant surveys and clinical staging forms are completed annually. In these analyses all HIV-positive men and women aged > 18 in the province of British Columbia who were antiretroviral naive, who had baseline plasma viral load and CD4 cell count data available, and who initiated therapy with either double or triple drug therapy between August 1,1996 and December 31,1999 were eligible. The follow-up period extended to March 31, 2000. For the purposes of analysis we followed the intent to treat principle; thus all eligible subjects were included as they were first dispensed antiretrovirals regardless of whether they later modified their therapeutic regimen. In order to obtain complete socioeconomic data for each participant, socioeconomic status was assessed through an ecological measure. A s has been done by others,^4 we linked participants to census-based socio-economic data by linking each patient's postal code to 1996 Canadian Census income data, and the median income of individuals in each patient's neighborhood was used as marker of socioeconomic status. Each participant's postal code upon enrollment was the value used for the linkage in an effort to minimize the downward drift effect that may occur as patients become sicker over time.5/6 Participants were 134 defined as low socio-economic status if the median income of individuals in their neighborhood was below the Canadian low income cut-off of $14,147 (CND), whereas* those who resided in a neighborhood with an individual median income above this level were defined as high socio-economic status.I5 In exploratory analyses, we evaluated associations between socio-economic status (low versus high) and baseline characteristics, as well as associations between H A A R T as the initial regimen (yes versus no) and baseline characteristics using single variable and multi-variable logistic regression analyses. These analyses were performed to evaluate potential confounders of socio-economic status, and to assess the potential for selection bias that might be attributable to inequitable access to triple combination therapy by socio-demographic characteristics. Variables (described below) were considered in single variable and multi-variable variable logistic regression analyses. Categorical explanatory variables were analyzed using Pearson's Chi-square test and continuous variables were analyzed using the Wilcoxon rank sum test. Since the outcome variable in these analyses was not a rare event we used the binomial distribution in the multi-variable analysis. A l l variables that were significant (p < 0.05) in single variable analyses were considered in the models. The primary endpoint in this analysis was HIV-related death. Deaths occurring during the follow-up period were identified on a continuous basis from physician reports and through record linkages carried out with the British Columbia Division of Vital Statistics. Deaths from non-HIV related causes were censored at time of death, but classified as non-events in this analysis. For the survival analyses, all patients were considered. Cumulative mortality rates were estimated by Kaplan-Meier methods. Survival curves were 135 compared between groups with the log-rank test. Cox proportional hazards regression was then used to calculate single variable and adjusted relative hazards and 95% confidence intervals (CIs) . 1 6 Time since the initiation of therapy was measured in days. The assumption of proportional hazards was validated by inspection of log (-log (survival function)) estimates against log time plots. Participants who were still alive at the end of the study period were right censored as of March 31, 2000. Variables potentially associated with survival were analyzed using single variable and multi-variable techniques. The proportional hazards models were adjusted for all variables that were significant in single variable Cox models, and for variables that were associated with socio-economic status or with receiving triple therapy as the initial regimen. As well, the overall model was adjusted for two or three drugs in the initial regimen. Variables considered included: age (continuous), gender (male versus female), socio-economic status (high versus low), baseline clinical AIDS diagnosis (yes versus no), history of injection drug use (yes versus no), baseline CD4 cell count (continuous), log 1 0 plasma viral load (continuous), adherence (continuous), and physician experience (continuous). Adherence was estimated by calculating the ratio of time that medication dispensed would last as a proportion of follow-up time through a linkage with the province's antiretroviral dispensing pharmacies. This calculation was restricted to each patient's first year on therapy to avoid reverse causation that may occur among patients who cease antiretroviral therapy after they have become too sick to take medication. We have previously demonstrated how this estimate strongly predicts response to therapy among patients in the treatment program. 1 7 / 1 ^ Physician experience was defined as has been done previously, 1 3 136 as the number of patients previously enrolled into the H I V / A I D S treatment program by that patient's physician at the time of the patient's enrollment. 7.5 RESULTS In total, 1408 patients were antiretroviral naive and initiated therapy during the study period, and had baseline plasma viral load and CD4 cell count data available and who were linked to census-based socio-economic data. Participants resided in a total of 357 different census areas. Of the 1408 subjects, 377 (27%) were initially prescribed double combination and 1031 (73%) patients were initially prescribed triple combination with either a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor. As of March 31, 2000, a total of 126 deaths were identified in the overall study population. Of these, 24 were not directly attributed to an HIV-related cause (3 suicides and 21 accidental drug overdoses) and were censored as non-events at the time of death. The remaining 102 deaths gave a crude mortality rate of 7.24%. Using Kaplan-Meier methods, the product limit estimate of the cumulative mortality rate at 12 months was 2.71%. Similarly, among the sub-set of 1031 participants initially prescribed triple therapy, 69 HIV-related deaths were observed for a crude mortality rate of 6.69%. A m o n g this group, the product limit estimate of the cumulative mortality rate at 12 months was 2.77%. Figure 7.1 shows the Kaplan Meier plots of time to death in the 2 socio-economic strata for the entire cohort. As shown here, the product limit estimate (SE) of the cumulative mortality rate at 12 months for the high socio-economic group was 2.1% (0.4%). Conversely, the product limit estimate (SE) of the cumulative mortality rate at 12 months for the low socio-economic group was 137 5.2% (1.5%). Statistically significant differences in survival were noted between those in the high and low socio-economic strata (log rank p <0.001). In single variable Cox regression analyses, we found that the hazard ratio for mortality among patients in the low socio-economic strata was 2.57 (95% CI: 1.71 - 3.89; p < 0.001). After adjustment for demographic variables age and gender, the hazard ratio was 2.41 (95% CI: 1.59 - 3.66; p < 0.001). Analyses were then conducted to investigate factors associated with socio-economic status. In single variable analyses, we found that history of injection drug use (p = 0.001), receiving double therapy (p = 0.001) older age (p = 0.001), and non-adherence (p < 0.001) were associated with patients in the lower socio-economic strata. A l l of these variables remained independently associated with lower socio-economic status, when entered together into multi-variable analyses (all p < 0.01). We found no evidence that male gender, clinical diagnosis of AIDS at baseline, physician experience, CD4 cell count, or plasma viral load were associated with socio-economic status. Second, in order to evaluate the potential for confounding that might be attributable to inequitable access to triple combination therapy by socio-economic status, we evaluated variables associated with initially receiving a triple drug regimen. The results of the single variable analyses among the entire patient population is shown in Table 7.1. As would be expected given the guidelines prior to July 1997, plasma viral load was strongly associated with receiving a triple drug regimen (p < 0.001), as was having a lower CD4 cell count (p < 0.001) and a baseline AIDS diagnosis (p < 0.001). In contrast, being female (p = 0.002), history of injection drug use (p < 0.001), and being in the lower socio-economic strata (p = 0.001) were negatively associated with receiving a triple 138 combination regimen in the single variable analyses. We found no evidence that age or physician experience were associated with receiving a triple drug regimen. The results of the multi-variable analysis of factors associated with receiving triple combination therapy in the initial regimen are shown in Table 7.2. As shown here, after adjustment for baseline CD4 cell count, plasma viral load, persons with a history of injection drug use (AOR = 0.84; [95% CI: 0.71 -0.99]) and persons who were in the low socio-economic strata were significantly less likely (AOR = 0.76; [95% CI: 0.63 - 0.91]) to be prescribed a triple combination regimen. The model was also adjusted for physician experience since we assumed experience treating HIV disease could be a source of residual confounding. These results were unaffected if the analysis was repeated using traditional logistic regression. The single variable and multi-variable analyses of the baseline factors associated with death for the entire cohort are presented in Table 7.3. As shown here, being older (Risk Ratio [RR] = 1.03 per year), having AIDS at baseline (RR = 1.97), being in the lower socio-economic strata (RR = 2.58), having a lower baseline CD4 cell count (RR = 1.50 per 100 cell decrease), and a higher baseline plasma viral load (RR = 1.67 per log 1 0 increase) were associated with death in the single variable analyses. Being more adherent by our definition was negatively associated with death (RR = 0.86 per 10% increase). Gender and a history of injection drug use were not associated with survival. The final multi-variable model was adjusted for variables that were associated with socio-economic status or receiving triple therapy in earlier analyses, and variables that were significantly associated with survival in single variable analyses. The final model 1 3 9 was also adjusted for two or three drugs in the initial regimen. In multi-variable analyses, being more adherent was protective against death (RR = 0.83 [95% CI: 0.78 - 0.89] per 10% increase), whereas being in the lower socio-economic strata (RR = 2.03 (95% CI: 1.32 - 3.12]), and having a lower baseline CD4 cell count (RR = 1.53 [95% CI: 1.32 - 1.77]) per 100 cell decrease) were independently associated with death. Age, baseline AIDS, injection drug use, and baseline viral load were all non-significant. We then evaluated statistical interactions between covariates, and an interaction between socio-economic status and initial triple therapy was found which remained statistically significant after adjustment for socio-economic status, triple therapy in the initial regimen, history of injection drug use, age, adherence, baseline AIDS, CD4 cell count, and plasma viral load (p=0.047). No other interactions were detected and the Risk Ratios shown in Table 7.3 were not significantly affected by the inclusion of the interaction term. To address this concern, we built an additional multi-variable Cox model which stratified patients by the two socio-economic strata (high versus low) and triple therapy in the initial regimen (yes versus no). In this analysis, patients in the high socio-economic strata who received triple therapy as their initial regimen were the reference category. The results of this analysis are shown in Table 7.4. As shown here, patients in the low socio-economic strata who received double therapy were significantly more likely to die (RR = 2.99 [95% CLT.71 - 5.24]). Also shown here, there was no difference in survival between patients in the high and low socio-economic strata among patients prescribed triple therapy as their initial regimen (p = 0.226). Furthermore, when we repeated the analysis shown in Table 7.3, restricting to patients who received triple therapy as their initial 1 4 0 regimen, we found that socio-economic status was not significantly associated (p= 0.199) with survival (data not shown). 7.6 DISCUSSION In the present study we found that persons of lower socio-economic status were substantially less likely to receive triple combination therapy as their initial regimen, even after adjustment for baseline clinical characteristics. In addition, we identified lower-socioeconomic status to be independently associated with mortality amongst patients receiving antiretroviral treatments since 1996. Interestingly, the association between socio-economic status and survival did not persist in analyses that were stratified by double or triple therapy, or when analyses were restricted to persons who received triple combination therapy as their initial regimen. Although lower socioeconomic status was associated with shorter survival in the present study, it would be incorrect to conclude that persons in lower socio-economic strata derive less benefit from H A A R T . Previous studies have demonstrated that persons initially prescribed double therapy have a lesser virologic response and more rapid viral r e b o u n d , ! 9 ' 2 ^ as well as significantly shorter survival than persons who initiate therapy with triple t h e r a p y . 2 ^ ' 2 2 Furthermore, pre-treatment with dual therapy may preclude a worse treatment outcome even among those who later switch to a triple drug regimen because of the more rapid development of resistance 2 3, 2 4 Therefore, it is likely that socioeconomic status was associated with shorter survival in the overall group and not in the sub-group of participants prescribed triple therapy, because of the heightened prescription of double therapy among persons in the lower socio-economic strata. 141 In the present study, persons of lower socioeconomic status and persons with a history of injection drug use were less likely to be prescribed triple therapy. Limited access to H A A R T among persons with a history of injection drug use and persons of lower socio-economic status may reflect the concerns of physicians who have hoped that persons with less stable lifestyles may have better adherence to a less complex regimen. In addition, it is also possible that concerns regarding possible transmission of protease inhibitor or N N R T I resistant virus influenced prescribing decisions. However, a recent study demonstrated that a substantial proportion of homeless and marginally housed individuals had good adherence to antiretroviral therapy including protease inhibitors, and that resistance to Pis were rare among those that were non-adherent. 25 Furthermore, it was recently argued that physicians should not indefinitely withhold H A A R T from patients who are thought to be poorly adherent , 2 6 and studies have shown that providers may be poor judges of adherence. 2 7 Studies conducted well before combination antiretroviral therapy found shorter survival amongst lower socio-economic groups attributable to inequitable access to PCP prophylaxis or Zidovudine mono-therapy.9/28 [ s therefore worrisome that this pattern appears to have repeated itself, especially considering that the data for the present study were derived in a setting that delivers antiretroviral therapy and all medical care free of charge through a universal healthcare system. Although dual therapy is no longer the standard of care for HIV infection,29 in light of the above, our findings should not be viewed irrelevant in the modern era. With the growing number of therapeutic agents and rapid evolution of therapeutic guidelines, it is important that not only are guidelines updated regularly, but also that mechanisms exist to monitor and 142 ensure their successful implementation, so that social disparities in access to effective treatments can be avoided. It must be stressed that persons who have not initiated antiretroviral therapy in the province of British Columbia were not included in the present study. Problems with access to antiretrovirals are most pressing in the developing world,30 but are also of substantial concern in the developed world, where improving access to antiretrovirals among marginalized groups creates challenges even in the context of universal healthcare systems like ours. 13/31,32 In fact, it is likely that social gradients in survival would be even more substantial if the whole H I V infected population in BC was considered because of the uneven access to appropriate H I V / A I D S care in various social strata.33/34 Further, the survival gradient between patients with low and high socio-economic status would also be much greater if all cause-mortality were considered, due to the high number of deaths from illicit drug overdoses observed in the cohort, which previous studies have shown to be associated with lower income neighborhoods in the province.I' 2 We should also stress that because an ecological measure of socioeconomic status was used in our survival analyses, there was potential for misclassification of individuals who live in a neighborhood that is not reflective of their actual socioeconomic status.35 However, similar misclassification concerns surround the use of self-reported annual income,36 especially among persons with a chronic illness, because of spousal or family support, as well as socially desirable reporting.37,38 Furthermore, previous studies have suggested that ecological measures may result in underestimation of the true relationship between socio-economic status and health-related measures.35 143 7.7 SUMMARY In summary, in a population-based setting in which antiretrovirals are delivered free of charge through a universal healthcare system, there is evidence that lower socio-economic status is associated with shorter survival. Our analyses indicate that the association between socio-economic status and survival may be confounded by inequitable access to triple therapy by socio-economic status. This finding suggests that further efforts must be focussed on increasing the awareness, implementation, and monitoring of therapeutic guidelines to ensure that physicians treating HIV positive patients deliver care in accordance with the most recent evidence-based approaches. 144 7.8 REFERENCES 1. Tyndall M W , Craib KJ, Currie S, L i K, O'Shaughnessy M V , Schechter M T . Impact of HIV infection on mortality in a cohort of injection drug users. / Acquir Immune Defic Syndr 2001;28(4):351-7. 2. Miller C L , Chan KJ, Palepu A , Tyndall M W , Hogg RS, O'Shaughnessy M V . Socio-Demographic Profile and HIV and Hepatitis C Prevalence Among Persons who Died of a D r u g Overdose. Addict Res and Theory 2001;9(5):459-470. 3. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 4. Palella FJ, Jr., Delaney K M , Moorman A C , et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators [see comments]. New England Journal of Medicine 1998;338(13):853-60. 5. Schechter M T , Hogg RS, Aylward B, Craib KJ, Le T N , Montaner JS. Higher socioeconomic status is associated with slower progression of H I V infection independent of access to health care. / Clin Epidemiol 1994;47(l):59-67. 6. Hogg RS, Strathdee SA, Craib KJ, O'Shaughnessy M V , Montaner JS, Schechter M T . Lower socioeconomic status and shorter survival following HIV infection [see comments]. Lancet 1994;344(8930):1120-4. 7. Rothenberg R, Woelfel M , Stoneburner R, Milberg J, Parker R, Truman B. Survival with the acquired immunodeficiency syndrome. Experience with 5833 cases in New York City. N Engl J Med 1987;317(21):1297-302. 145 8. Melnick SL, Sherer R, Louis T A , et al. Survival and disease progression according to gender of patients with HIV infection. The Terry Beirn Community Programs for Clinical Research on A I D S . JAMA 1994;272(24):1915-21. 9. Easterbrook PJ, Keruly JC, Creagh-Kirk T, Richman D D , Chaisson RE, Moore R D . Racial and ethnic differences in outcome in zidovudine-treated patients with advanced HIV disease. Zidovudine Epidemiology Study Group. JAMA 1991;266(19):2713-8. 10. Moore RD, Hidalgo J, Sugland BW, Chaisson RE. Zidovudine and the natural history of the acquired immunodeficiency syndrome. N Engl J Med 1991;324(20):1412-6. 11. Rapiti E, Porta D , Forastiere F, Fusco D , Perucci C A . Socioeconomic status and survival of persons with AIDS before and after the introduction of highly active antiretroviral therapy. Lazio A I D S Surveillance Collaborative Group. Epidemiology 2000;11(5):496-501. 12. Carpenter C C , Fischl M A , Hammer S M , et al. Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. International AIDS Society-USA. JAMA 1996;276(2):146-54. 13. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 14. Katz M H , H s u L , Lingo M , Woelffer G , Schwarcz SK. Impact of socioeconomic status on survival with AIDS. Am J Epidemiol 1998;148(3):282-91. 15. Statistics Canada Income Statistics Division. Low Income Cutoffs from 1991 to 2000. Catalogue no. 75F0002M1E. 1 4 6 16. Cox DR. Regression models and life tables (with discussion). / Royal Stat Soc B 1972:187-202. 17. Low-Beer S, Yip B, O'Shaughnessy M V , Hogg RS, Montaner JS. Adherence to triple therapy and viral load response. / Acquir Immune Defic Syndr 2000;23(4):360-l. 18. Palepu A , Yip B, Miller C , et al. Factors associated with the response to antiretroviral therapy among HIV-infected patients with and without a history of injection drug use. AIDS 2001;15:423-424. 19. Opravil M , Cone RW, Fischer M , et al. Effects of early antiretroviral treatment on HIV-1 R N A in blood and lymphoid tissue: a randomized trial of double versus triple therapy. Swiss HIV Cohort Study. / Acquir Immune Defic Syndr 2000;23(l):17-25. 20. Rhone SA, Hogg RS, Yip B, et al. Do dual nucleoside regimens have a role in an era of plasma viral load- driven antiretroviral therapy? / Infect Dis 1998;178(3):662-8. 21. Hogg RS, Rhone SA, Yip B, et al. Antiviral effect of double and triple drug combinations amongst H I V - infected adults: lessons from the implementation of viral load-driven antiretroviral therapy. AIDS 1998;12(3):279-84. 22. Hogg RS, Yip B, Kully C , et al. Improved survival among HIV-infected patients after initiation of triple-drug antiretroviral regimens [see comments]. CMA] 1999;160(5):659-65. 23. Miller V , Phillips A , Rottmann C , et al. Dual resistance to zidovudine and lamivudine in patients treated with zidovudine-lamivudine combination therapy: association with therapy failure. / Infect Dis 1998;177(6):1521-32. 147 24. Deeks SG. Determinants of virological response to antiretroviral therapy: implications for long-term strategies. Clin Infect Dis 2000;30 Suppl 2:S177-84. 25. Bangsberg DR, Hecht F M , Charlebois E D , et al. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS 2000;14(4):357-66. 26. Sollitto S, Mehlman M , Youngner S, Lederman M M . Should physicians withhold highly active antiretroviral therapies from HIV-AIDS patients who are thought to be poorly adherent to treatment? AIDS 2001;15:153-159. 27. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 28. Moore RD, Stanton D , Gopalan R, Chaisson RE. Racial-Differences in the Use of Drug-Therapy For Hiv Disease in an Urban-Community. New England Journal of Medicine 1994;330(ll):763-768. 29. Carpenter C C , Cooper D A , Fischl M A , et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000;283(3):381-90. ' 30. Wood E, Braitstein P, Montaner JS, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa [see comments]. Lancet 2000;355(9221):2095-100. 31. van der Werf MJ, Schinkel J, van Santen G , Vergouwe I, Wix R A , van Ameijden EJ. Highly active antiretroviral therapy among drug users in Amsterdam: self-perceived reasons for not receiving therapy. Aids 1999;13(10):1280-1. 1 4 8 32. Wood E, Schechter M T , Tyndall M W , Montaner JS, O'Shaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000;14(9):1229-35. 33. Shapiro M F , Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. Jama 1999;281(24):2305-15. 34. Anderson K H , Mitchell JM. Differential access in the receipt of antiretroviral drugs for the treatment of AIDS and its implications for survival. Arch Intern Med 2000;160(20):3114-20. 35. Hyndman JC, Holman C D , Hockey RL, Donovan RJ, Corti B, Rivera J. Misclassification of social disadvantage based on geographical areas: comparison of postcode and collector's district analyses. Int J Epidemiol 1995;24(l):165-76. 36. Liberatos P, Link B G , Kelsey JL. The measurement of social class in epidemiology. Epidemiol Rev 1988;10:87-121. 37. Demissie K, Hanley JA, Menzies D , Joseph L, Ernst P. Agreement in measuring socio-economic status: area-based versus individual measures. Chronic Dis Can 2000;21(l):l-7. 38. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999;48(12):1751-8. 149 Table 7.1 Single variable analyses comparing socio-demographic and clinical characteristics of participants prescribed double versus triple combination antiretroviral therapy. Initial Regimen Double Triple Characteristic n=377, (%) n=1031, (%) p-value Baseline HIV-1 RNA (copies/ml) Median 48,000 120,000 <0.001 IQR 19,000 - 92,000 41,000 - 300,000 Baseline CD4+ count (cells X 109/L) Median 330 270 <0.001 IQR 220-440 130-420 Prior AIDS Diagnosis No 347(92.4) 882(85.5) <0.001 Yes 30 (8.0) 149 (14.5) Age Median 36 37 0.110 IQR 32 - 42 32 - 43 Gender Female 73 (19.4) 132(12.8) 0.002 Male 304 (80.6) 899 (87.2) Injection Drug Use N o 264(70.0) 808(78.4) < 0.001 Yes 113 (30.0) 223 (21.6) Median Income High 292(77.5) 874(84.8) 0.001 Low 85 (22.5) 157 (15.2) Physician Experience Median 36 44 0.459 IQR 9-106 5 -137 150 Table 7.2 Multi-variable analysis* of factors associated with being prescribed triple combination therapy as the initial regimen. Adjusted Variable Odds Ratio 95% CI p-value Baseline HIV-1 RNA (copies/mL) (per loglO increase) 1.64 (1.47-1.83) <0.001 Baseline C D 4 cell count (per 100 cell decrease) 1.03 (0.99-1.06) 0.058 Injection Drug Use (Yes versus No) 0.84 (0.71 - 0.99) 0.033 Median Income (<$14,147vs> $14,147) 0.76 (0.63-0.91) 0.003 * Variables that were statistically significant at the 0.05 cut-off in the single variable analysis were considered in the multi-variable model. Model was adjusted for physician experience (non-significant). 1 5 1 Table 7.3: Single variable and multi-variable* analysis of the baseline factors associated with death among 1,408 antiretroviral naive persons first prescribed double or triple combination antiretroviral therapy between August 1,1996 and December 31,1999 were eligible. Variable Risk Ratios (RR) Crude Adjusted* RR ( 9 5 % CI) p-value RR ( 9 5 % CI) p-value Gender (Male versus female) 0.67 (0.41-1.11) 0.117 Age in Years (Continuous) 1.03 (1.00-1.05) 0.019 1.02 (0.99-1.04) 0.069 Prior AIDS Diagnosis (Yes versus no) 1.97 (1.23-3.15) 0.005 1.24 (0.74-2.09) 0.412 Injection Drug Use (Yes versus no) 1.38 (0.91-2.08) 0.132 1.28 (0.84-1.96) 0.250 Adherence (per 10% increase) 0.86 (0.81-0.91) <0.001 0.83 (0.78-0.89) <0.001 Median Income (<$14,147 vs > $14,147) 2.58 (1.71-3.89) <0.001 2.03 (1.32-3.12) 0.001 Baseline C D 4 + count (cells/mm3) per 100 cell decrease 1.50 (1.32-1.70) <0.001 1.53 (1.32-1.77) <0.001 Baseline HIV-1 RNA per log 1 0 increase 1.67 (1.20 - 2.34) 0.002 1.26 (0.89-1.79) 0.237 * Covariates were entered into the model if associated with socio-economic status, associated with receiving triple therapy in the initial regimen, or survival (in single variable analysis). Model was also adjusted for two or three drugs in the intial regimen (non-significant). 152 Table 7.4 Adjusted hazard ratios for the risk of death stratified by socio-economic strata and triple therapy in the initial regimen. Median Income ($CDN) Initial Therapy Adjusted Hazard Ratio (95% CI) p-value > $14,147 Triple 1.00 - R E F \u00E2\u0080\u0094 \u00E2\u0080\u0094REF\u00E2\u0080\u0094 <$14,147 Triple 1.43 (0.80 -2.54) 0.226 > $14,147 Double 0.84 (0.46-1.55) 0.581 <$14,147 Double 2.99 (1.71 - 5.24) <0.001 *Model was adjusted for age, adherence, history of injection drug use, baseline AIDS, CD4 cell count, and plasma viral load. 153 L E G E N D F O R F I G U R E S Figure 7.1 Analysis of 1,408 antiretroviral naive participants who initiated therapy with either a double or triple combination regimen stratified by low and high socio-economic status. 154 Figure 7.1: Probability of Survival Stratified by Socio-economic Status amongst all patients prescribed either double or triple therapy 82-80-78-76-74-| 1 1 1 1 1 1 0 6 12 18 24 30 36 Time From Start of Antiretrovirals (months) Socio-economic Strata fnl; High 1166 1120 1049 926 779 654 479 Low 242 219 206 176 148 119 88 1 5 5 C H A P T E R 8 A D H E R E N C E T O T H E R A P Y A N D O U T C O M E S F R O M H A A R T 8.1 F O R W A R D This chapter was is presently in press with AIDS as: Wood E, Hogg RS, Yip B, Harrigan RP, O'Shaughnessy M V , Montaner JS. Is there a CD4 cell count that precludes a survival response to H A A R T ? : The impact of physician experience and adherence on HIV disease progression after the initiation of therapy. 156 8.2 A B S T R A C T Objective: Therapeutic guidelines advise that 200-350 cells/mm 3 may approximate an irreversible threshold beyond which response to therapy is compromised. We evaluated whether non-immune based factors such as physician experience and adherence could affect survival among HIV infected adults starting H A A R T . Methods: Analysis of 1416 antiretroviral naive patients who initiated triple therapy between August 1, 1996 and July 31, 2000, and were followed until July 31, 2001. Patients whose physicians had previously enrolled > 6 patients were defined as having an experienced physician. Patients who received medications > 75% of the time during the first year of H A A R T were defined as adherent. Cumulative mortality rates and adjusted relative hazards were determined for various CD4 cell count strata. Results: Among patients with < 50 cells/mm 3 the adjusted relative hazard of mortality was 5.07 (95% CI: 2.50 - 10.26) for patients of experienced physicians and was 11.99 (95% CI: 6.33 - 22.74) among patients with inexperienced physicians, in comparison to patients with > 200 cells/mm 3 treated by experienced physicians. Similarly, among patients with < 50 cells/mm 3 , the adjusted relative hazard of mortality was 6.19 (95% CI: 3.03 - 12.65) for adherent patients and was 35.71 (95% CI: 16.17 - 78.85) for non-adherent patients, in comparison to adherent patients with > 200 cells/mm 3 . Conclusion: Survival rates are dramatically improved among patients starting with CD4 counts < 200 cells/mm 3 once adjusted for conservative estimates of physician experience and adherence. These findings have relevance for therapeutic guidelines considering initiating therapy at CD4 cell counts above 200 cells/mm 3 . 157 8.3 I N T R O D U C T I O N The benefits of triple drug highly active antiretroviral therapy (HAART) in the management of HIV disease are well established. Through the suppression of plasma HIV-1 R N A , H A A R T has been shown to improve CD4 cell counts and in turn decrease morbidity and mortality among HIV-infected patients.I'2 However, a critical question surrounds the optimal time to initiate H A A R T . We have recently evaluated rates of disease progression among patients initiating H A A R T , and found that only patients initiating therapy when CD4 cell count had declined below 200 cells/mm 3 were at increased risk of disease progression.^ Although this finding has recently been independently confirmed,^\" 6 therapeutic guidelines have been conservative in their recommendations on when to start therapy.3/7,8 These thresholds have been debated, primarily because of concerns regarding limited follow-up and potential for irreversible immune damage at CD4 cell counts below 350 cells/mm 3 . The British HIV Association Guidelines advise clinicians to consider initiating therapy when CD4 cell count has declined within the range of 200 -350 cells/mm 3 while taking into account the rate of CD4 cell count decline and HIV-1 R N A level, and immediate initiation of therapy if CD4 cell count has declined to 200 cells/mm 3.9 Conversely, the US Public Health Service guidelines currently recommend treatment if CD4 cell count declines below 350 CD4 cells/mm 3 or if plasma viral load is above 55,000 copies/mL.^ More recently, the International AIDS Society-USA guidelines specifically support the use of a higher than 200 CD4 cells/mm 3 threshold with the decision to treat being based on the rate of CD4 decline and the HIV-1 R N A level . 1 0 However, prior to the advent of H A A R T , it was observed that non-immune based factors, such as physician experience, were strongly associated with survival.H 158 Although a role of physician experience in the H A A R T era has been speculated upon,12-14 the impact of physician experience in the H A A R T era has not been demonstrated. In addition, although a wealth of studies have demonstrated an association between adherence to antiretroviral therapy and improved viral suppression,15-17 the relationship between adherence to therapy and AIDS-related mortality in the H A A R T era has not been well-demonstrated. For the present analysis, we hypothesized that if 200 - 350 cells/mm 3 represented an irreversible biological threshold that would limit the response to antiretroviral therapy, adjustment for conservative estimates of non-immune based factors would not have a substantial impact on HIV related mortality among patients initiating therapy with < 200 cells/mm 3 . The present analysis was therefore conducted to evaluate whether survival rates could be affected by conservative estimates of physician experience and patient adherence. 8.4 METHODS The dispensation of antiretroviral medications in the province of British Columbia has been described in detail elsewhere.3/18 In brief, the Centre remains the only free source of antiretroviral medications in the province. The Centre's H I V / A I D S Drug Treatment program has received ethical approval from the University of British Columbia Ethics Review Committee at its St. Paul's Hospital site, and the program conforms with the province's Freedom of Information and Protection of Privacy Act. In June 1996 the Centre adopted plasma viral load driven antiretroviral therapy guidelines, consistent with those put forward by the International AIDS Society -USA.19 In brief, antiretroviral therapy naive individuals with plasma viral load > 100,000 copies/ mL were offered triple drug regimens (i.e. two nucleosides plus a 159 protease inhibitor or a non-nucleoside reverse transcriptase inhibitor), while those with plasma viral loads from 5,000 to 100,000 copies/mL were offered dual nucleoside therapy. The Centre guidelines were revised in July 1997 to recommend triple combination therapy for all antiretroviral naive individuals with plasma HIV-1 R N A levels greater than 5,000 copies/mL or CD4 cell counts below 500 cells/mm3.20 The Centre's guidelines recommend further that plasma HIV-1 R N A levels be monitored at baseline, at 4 weeks after starting antiretroviral therapy and every three months thereafter. Plasma viral loads were measured using the Amplicor HIV-1 Monitor\u00E2\u0084\u00A2, (Roche Diagnostics Branchburg, NJ). At the time of enrollment, physicians enrolling an HIV-infected individual must complete a drug request enrollment form, which acts as a legal prescription and supplies baseline information, including past HIV-specific drug history, CD4 cell count, plasma HIV-1 R N A , current drug requests, and enrolling physician data. The primary care physicians are asked to complete a dinical staging form using the World Health Organization (WHO) clinical staging system. 21 For all program participants, a complete prospective profile of antiretroviral therapy is maintained, including the medications dispensed, the dose, the dispensation dates, and the quantity dispensed. The primary endpoint in this analysis was death. Deaths occurring during the follow-up period were identified on a continuous basis from physician reports and through record linkages carried out with the British Columbia Division of Vital Statistics. Deaths from accidental causes were censored at time of death, and classified as non-events in the primary analysis. However, we also conducted sub-analyses of all cause mortality. Physician experience was defined a priori, based on data derived in the pre-H A A R T era, which found that having a physician who had provided primary care for 160 greater than 5 HIV-infected patients was associated with improved survival Based on these earlier findings and our objective to only adjust for conservative estimates of physician experience, we defined a patient as having an HIV-experienced physician if the patient's enrolling primary care physician had previously enrolled only > 6 patients into the program. Since patients were continuously enrolled in the treatment program during the study period, a 'non-experienced' physician could become 'experienced' by our definition over time, however, the experience level assigned to each patient was fixed and based on the HIV-related experience of their physician at the time of each patient's first follow-up. Patients were conservatively defined as adherent or non-adherent using prescription refill compliance. 2 2 The definition of adherence was based on the ratio of time that medication dispensed would last as a proportion of follow-up time. This calculation was restricted to each patient's first year on therapy to avoid reverse causation that could occur among patients who cease antiretroviral therapy after they have become too sick to take medication. If a person died prior to one year after starting therapy, then the proportion of follow-up time and the proportion of time receiving treatment were used to calculate this ratio. We have previously demonstrated how this estimate strongly predicts virological response, and how it can adjust for the potentially confounding effect of treatment interruption 2 3 - 2 5 . for the purposes of this analysis, patients were defined as non-adherent if they received antiretroviral medications < 75% of the time during the first year of follow-up, since this level of adherence is well below what is presently recommended by therapeutic guidelines. The decision to use this dichotomy was also based on an earlier analysis, which found that picking up less than 75% of prescriptions during the first year of follow-up was predictive of shorter survival . 2 6 161 The primary analysis was restricted to HIV-infected men and women who were antiretroviral naive and were first prescribed triple drug antiretroviral therapy between August 1,1996 and July 31, 2000. Study subjects were initially prescribed triple drug combination therapy with regimens including two nucleoside reverse transcriptase inhibitors and either a protease inhibitor (PI) or a non-nucleoside reverse transcriptase inhibitor (NNRTI) at the discretion of the enrolling physician. For the purposes of this analysis we followed the intent to treat principle, and subjects were included as they were first dispensed antiretrovirals, regardless of whether they later modified their therapeutic regimen. Cumulative mortality rates were estimated using Kaplan-Meier methods. Event-free subjects were right censored as of July 31, 2001. Participants were pooled into three baseline CD4 cell count groups based upon breakpoints identified previously. 3 These groups were < 50, 50 - 199, and > 200 cells/mm 3 respectively. We then further stratified the 3 CD4 cell count groupings into those with HIV-experienced and non-experienced physicians and into those who were adherent and non-adherent as described above. Cox proportional hazards regression was used to calculate univariate and adjusted relative hazards and 95% confidence intervals (CIs). 2 7 The assumption of proportional hazards was validated by inspection of log (-log [survival function]) estimates against log time plots. Variables examined in this analysis included: baseline plasma HIV-1 R N A levels (log 1 0 transformed); baseline CD4 cell count (< 50, 50 - 199, and > 200 cells/mm 3); protease-inhibitor use in the initial regimen (yes vs. no); time of initiation of therapy (after vs. before July 1997); a prior diagnosis of AIDS (yes vs. no); age (continuous); sex; and physician experience and adherence as described above. We also conducted a sub-analysis of those participants that initiated therapy after July 1997, in order to rule out the possibility of residual confounding, that may not be eliminated 162 by simple adjustment with this temporal cut-off. A l l statistical analyses were performed using SAS software version 6.0 (SAS, Cary, NC) . 8.5 RESULTS During the study period, 1,580 antiretroviral naive participants began triple combination therapy with a PI or NNRTI. Of these, 164 (10.4%) were excluded from this analysis for not having both baseline CD4 cell count and plasma HIV-1 R N A level measures available within 6 months prior to the start of antiretroviral therapy. Therefore, the study sample was based on 1416 (90.0%) subjects (1194 [84.3%], men; 222 [15.7%] women). No differences in sex, and AIDS at baseline, and subsequent mortality was observed between the study sample and those excluded. However, persons excluded from this analysis were more likely to be younger (median age 36 vs. 37; p = 0.04) and taking protease inhibitors (78.7% vs. 69.0%; p = 0.01). At baseline, the median CD4 cell count was 270 (IQR: 130 - 420) and median plasma HIV R N A level was 119,500 copies/mL (IQR: 38,000 - 300,000). As of July 31, 2001, 132 patients had died in the study population. Twenty-two of these were not attributed to HFV and were censored as non-events at the time of death. These 22 deaths included 5 suicides and 17 accidental drug overdoses. The remaining 110 deaths gave a crude AIDS-related mortality rate of 7.8%. Figure 8.1 shows the 42 month Kaplan Meier survival estimates for the overall cohort stratified by CD4 cell count using the 3 previously identified CD4 cell count strata 3. The number of patients in each strata are shown on the figures. Statistically significant differences in survival were noted among the 3 CD4 strata (all log rank p <0.05). 163 Figure 8.2 shows the Kaplan Meier plots of time to death in the 3 CD4 cell count groups stratified by physician experience. Among participants whose enrolling primary care physician was experienced (Panel 2A), statistically significant differences in survival were noted between those with > 200 CD4 cells / m m 3 at baseline and the other two CD4 strata (log rank p < 0.05), however, there was no longer a significant difference between the < 50 cells/mm 3 and 50-199 cells/mm 3 CD4 cell count groupings. Among those whose enrolling primary care physician was non-experienced (Panel 2B), statistically significant differences in survival were noted among the 3 CD4 strata (all log rank p <0.05). When comparisons were made across the two physician experience strata, a statistically significant difference was found between the two < 50 cells/mm 3 groupings, with higher mortality found among the patients with less experienced physicians (p = 0.013). Figure 8.3 shows the Kaplan Meier plots of the time to death in the 3 CD4 cell count groups, stratified according to their level of adherence. Among those who were adherent (Panel 3A), statistically significant differences in survival were noted between those with > 200 CD4 cells at baseline and the other two strata (log rank p <0.05), however, the difference between the < 50 cells/mm 3 and 50-199 cells/mm 3 CD4 cell count strata was not statistically significant. Among non-adherent patients (Panel 3B), statistically significant differences in survival were noted between all CD4 cell count strata (all log rank p <0.05). When comparisons were made across the two adherence strata, significant differences were found between all pairs of CD4 cell count groupings with mortality being higher among non-adherent patients (all log rank p <0.05). Of note, among adherent patients with baseline CD4 cell counts > 200 cells/mm 3 , the cumulative mortality rate was only 2.6% after a full 42 months of treatment (Figure 8.2 Panel A), and we found no differences in survival among CD4 strata above 200 (data not shown). The median baseline CD4 cell count was 365 cells/mm 3 (IQR: 270 - 460) 164 among the 14 (2.6%) deaths among adherent patients with baseline CD4 > 200 cells/mm 3 . Table 8.1 shows the analyses of factors associated with the time to death for the overall cohort. In univariate analyses, a clinical diagnosis of AIDS at baseline, older age, physician experience, adherence, and a CD4 cell count at baseline of less than 200 cells/mm 3 and a higher baseline HIV-1 R N A , were associated with mortality. Date of therapy initiation, gender, history of injection drug use, and protease inhibitor use were not statistically significant as predictors of mortality. In the multivariate analyses, in comparison to patients with a baseline CD4 cell count > 200 cells/mm 3 , the relative hazard of death was 3.81 (95% CI, 2.33 - 6.24), and was 7.08 (95% CI, 4.08 - 12.26) for patients with a baseline CD4 cell counts 50-199 cells/mm 3 and < 50 cells/mm 3 respectively in adjusted analyses. Similarly, the relative hazard of death was 0.67 (95% CI, 0.45 - 0.99) amongst participants whose physician was experienced, and the relative hazard of death among patients who were adherent was 0.30 (95% CI, 0.20 - 0.45) in adjusted analyses. These estimates were not significantly affected by adjusting for additional combinations of non-significant variables, although we chose to force the history of injection drug use variable into the model in order to adjust for patient characteristics that may be associated with non-compliance to medical care 28, N 0 statistically significant interactions were found, and no interactions were entered into the final model, because they did not improve the model's fit. In confirmatory analyses, we found that results were consistent when we restricted the analysis considering only the 1053 participants who initiated therapy after July 1, 1997. These results are shown in Table 8.2. We also evaluated if residual confounding due to the imprecision of our adherence estimate might explain the association between physician experience and survival. When we conducted analyses to evaluate potential associations between physician experience and adherence, we 165 found that adherence was not associated with physician experience ([chi-square] p = 0.128). In addition, we also repeated the survival analyses shown in Tables 8.1 considering all cause mortality (including overdoses, suicides, etc). In the latter analysis, physician experience was not independently associated with survival, whereas adherence remained significantly associated with survival in all models. Finally, in order to provide adjusted estimates of the risk of death by CD4 cell count and physician experience categories, we repeated the Cox regression analyses with variables representing each CD4 and physician experience strata shown in Table 8.3(a). Covariates adjusted for are noted on the tables, and the reference category is those patients with > 200 CD4 cells at baseline who were seen by an experienced physician. Shown here, for patients with less than 50 cells / m m 3 at baseline, the adjusted relative hazard of death for patients with experienced physicians was 5.07 (95%CI, 2.50 - 10.26) and was 11.99 (95%CI, 6.33 - 22.74) for patients with non-experienced physicians. We also assumed that residual confounding may persist, despite adjustment for baseline AIDS, if the type and severity of baseline AIDS cases was not evenly distributed between physician experience groups, and repeated the above analysis excluding the 182 patients who had a clinical diagnosis of AIDS prior to initiating treatment. Here, the adjusted relative hazard of death for patients with less than 50 cells/mm 3 was 4.27 (95%CI, 1.75 - 10.43) for patients with experienced physicians and was 15.80 (95%CI, 8.08 - 30.89) for patients with non-experienced physicians. We also repeated the analysis, in order to provide adjusted estimates of the risk of death for the CD4 cell count and adherence categories as shown in Table 8.3(b). Here, the reference category was the group of patients with > 200 CD4 cells at baseline who were adherent. For patients with > 200 cells/mm 3 , the relative hazard of death for non-adherent patients was 3.55 (95%CI, 1.72 - 7.35) in comparison to adherent patients 1 6 6 in the same CD4 group. For patients with 50-199 cells/mm 3 , the relative hazard of death for adherent patients was 4.54 (95%CL 2.37 - 8.70) and was 9.78 (95%CL 4.48 -21.31) for non-adherent patients. For patients with less than 50 cells/mm 3 , the relative hazard of death for adherent patients was 6.19 (95%CI, 3.03 - 12.65) and was 35.71 (95%CL 16.17 - 78.85) for non-adherent patients. 8.6 DISCUSSION Our results show that survival rates following the initiation of antiretroviral therapy are dramatically improved among patients starting with CD4 cell counts below 200 cells/mm 3 once adjusted for conservative estimates of physician experience and patient adherence. These data suggest that a CD4 cell count of 200 - 350 does not represent an irreversible biological threshold beyond which response to therapy is compromised. Conversely, inappropriate care of advanced patients and patient non-adherence may be the strongest determinants of patient survival, instead of when antiretroviral therapy is initiated prior to 200 cells/mm 3 . Since mortality was elevated among all CD4 strata below 200 cells/mm 3 , however, our data do not support delaying the initiation of H A A R T below this level, and close follow-up will be necessary to examine if survival differences emerge among patients with baseline CD4 cell counts > 200 cells/mm 3 . Previous studies have shown that patients are not at immediate risk of death until CD4 cell count drops below 50 cells /mm3,29,30 a n c \ m a t AIDS-related opportunistic illnesses (O.I.s) are rare after the initiation of H A A R T , except among patients with less than 50 cells/mm3.31 However, it has also been shown that physicians who have treated < 6 HIV infected patients are significantly less knowledgeable about prescription of prophylaxis against O.I.s, provision of preventative vaccinations, as well as the treatment of acute O.I.s.32 These findings 167 likely explain why physician experience had the largest impact on highly advanced patients, and why physician experience was found to be independent of adherence. Additional evidence that the association between physician experience and survival is attributable to the management of advanced patients and not patient selection factors comes from the fact that physician experience was no longer significant when all-cause mortality was considered. Nevertheless, it was not the objective of the present study to identify the ideal level of experience for the optimal management of HIV-infected individuals, and further study will be required to identify the physician-related factors, such as the treatment or prevention of opportunistic infections, that may be responsible for our observations. Although there has been debate about the role of physician experience in the H A A R T era,13,33-37 t 0 o u r knowledge, these results are the first to demonstrate an independent effect of physician experience on survival since the widespread use of triple therapy. Our results are also the first to demonstrate statistically significant differences in survival among adherent and non-adherent patients at all levels of CD4 count, and in particular the dramatic impact adherence has on the survival of highly advanced patients. Although the multiple stratifications shown in Table 8.3(b) resulted in moderately large confidence intervals, these findings will contribute to the debate over when therapy is best initiated and will be useful for guidelines considering the management of the large proportion of patients that continue to present for treatment when CD4 cell counts have decreased to below 200 cells/mm3.3/38 Previous studies have demonstrated that measuring daily adherence to antiretroviral therapy can be fraught with difficulties that may over or underestimate a patient's actual exposure to treatment .39-42 Although using refill compliance as a surrogate for adherence has been previously validated, 2 2 ' 2 ^ there is likely a very strong conservative bias operating in our study because patients may have been less 168 than optimally adherent to daily treatment despite receiving > 75 percent medication during the first year on therapy. It has consistently been found that very high daily adherence is required, to achieve an undetectable plasma viral load, to prevent viral rebound, and to prevent the evolution of resistant virus.16,23,43,44 Further study will be required to determine the optimal level of adherence which, as illustrated above, is likely to vary depending on the nature of the endpoint selected (virological, immunological, or clinical) as well as the duration of follow-up. In addition, as we have previously discussed, 2^ a limitation of the present study is that patients were defined as adherent based on their behavior during the first year of therapy and then assigned to adherent strata for an analysis of baseline characteristics. Although we adjusted our analyses for pertinent demographic and clinical characteristics, like all studies of patients treated in observational cohorts, unmeasured differences may exist among study populations and for this reason caution is warranted. However, it was not the objective of the present study to elucidate the level of adherence needed to optimize the outcomes associated with antiretroviral therapy. 8.7 S U M M A R Y In summary, we found that mortality was dramatically altered by very conservative adjustments for physician experience and adherence, suggesting that inappropriate care of' advanced patients and patient non-adherence may be the strongest determinants of patient survival, instead of when therapy is initiated prior 200 cells/mm 3 . Since mortality was elevated among all CD4 strata below 200 cells/mm 3 , our data do not support delaying the initiation of H A A R T to any level below 200 cells/mm 3 . These results should be useful for the development of therapeutic guidelines, and should be helpful to emphasize the role of physician experience and patient adherence in the treatment of HIV infection. 169 8.8 R E F E R E N C E S 1. Hammer SM, Squires K E , Hughes M D , et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med 1997;337(ll):725-33. 2. Montaner JS, Reiss P, Cooper D, et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: the INCAS Trial. Italy, The Netherlands, Canada and Australia Study. JAMA 1998;279(12):930-7. 3. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 4. Sterling TR, Chaisson RE, Moore RD. HIV-1 R N A , CD4 T-lymphocytes, and dinical response to highly active antiretroviral therapy. AIDS 2001;15(17):2251-7. 5. Chen R, Westfall A , Coud G, Chatham A , Acosta E, Raper J, et al. Long-term survival after initiation of antiretroviral therapy. Presented at the Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, Febuary 4 - 8, 2001. Abstract No: 341. 6. Egger M , May M , Chene G , et al. Prognosis of HIV-l-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002;360(9327):119-29. 7. US Public Health Service. Guidelines for the use of Antiretroviral Agents in HIV-infected Adults and Adolescents. Updated February 4, 2002. Available at: http: / / www.hivatis.org. Accessed May 8, 2002. 170 8. Pomerantz RJ. Initiating antiretroviral therapy during HIV infection: confusion and clarity. JAMA 2001;286(20):2597-2599. 9. British HIV Association (BHIVA) guidelines for the treatment of HIV-infected adults with antiretroviral therapy. HIV Med 2001;2(4):276-313. 10. Yeni PG, Hammer SM, Carpenter C C , et al. Antiretroviral treatment for adult HIV infection in 2002: updated recommendations of the International AIDS Society-U S A Panel. JAMA 2002;288(2):222-35. 11. Kitahata M M , Koepsell TD, Deyo RA, Maxwell C L , Dodge WT, Wagner E H . Physicians' experience with the acquired immunodeficiency syndrome as a factor in patients' survival. N Engl J Med 1996;334(ll):701-6. 12. Montaner JS, Mellors JW. Antiretroviral therapy for previously treated patients. N Engl J Med 2001;345(6):452-5. 13. Lewis C E . Management of patients with HIV/AIDS. Who should care? JAMA 1997;278(14):1133-4. 14. Gallant JE. Strategies for long-term success in the treatment of HIV infection. JAMA 2000;283(10):1329-34. 15. Paterson D L , Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000;133(l):21-30. 16. Gross R, Bilker WB, Friedman H M , Strom BL. Effect of adherence to newly initiated antiretroviral therapy on plasma viral load. AIDS 2001;15(16):2109-17. 17. Haubrich R H , Little SJ, Currier JS, et al. The value of patient-reported adherence to antiretroviral therapy in predicting virologic and immunologic response. California Collaborative Treatment Group. AIDS 1999;13(9):1099-107. 171 18. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 19. Carpenter C C , Fischl M A , Hammer SM, et al. Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. International AIDS Society-USA. JAMA 1996;276(2):146-54. 20. Carpenter C C , Fischl M A , Hammer SM, et al. Antiretroviral therapy for HIV infection in 1997. Updated recommendations of the International AIDS Society-U S A panel. JAMA 1997;277(24):1962-9. 21. World Health Organization. Acquired immune deficiency syndrome (AIDS): interim proposal for a W H O staging system for HIV infection and disease. Wkly Epidemol Rec 1990; 65: 221-228. 22. Steiner JF, Prochazka A V . The assessment of refill compliance using pharmacy records: methods, validity, and applications. / Clin Epidemiol 1997;50(1):105-16. 23. Low-Beer S, Yip B, O'Shaughnessy M V , Hogg RS, Montaner JS. Adherence to triple therapy and viral load response. / Acquir Immune Defic Syndr 2000;23(4):360-l. 24. Palepu A , Yip B, Miller C, et al. Factors associated with the response to antiretroviral therapy among HIV-infected patients with and without a history of injection drug use. AIDS 2001;15:423-424. 25. Wood E, Montaner JS, Chan K, et al. Socioeconomic status, access to triple therpay, and survival from HIV-disease since 1996. AIDS 2002;16(15):2065-2072. 26. Hogg RS, Heath K V , Bangsberg D, et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up. AIDS 2002;16(7):1051-1058. 27. Cox DR. Regression models and life tables. / Royal Stat Soc B 1972:187-202. 172 28. Anis A H , Sun H , Guh DP, Palepu A , Schechter M T , O'Shaughnessy M V . Leaving hospital against medical advice among HIV-positive patients. Cmaj 2002;167(6):633-7. 29. Phillips A N , Elford J, Sabin C, Bofill M , Janossy G, Lee C A . Immunodeficiency and the risk of death in HIV infection. JAMA 1992;268(19):2662-6. 30. Ehmann W C , Eyster M E , Wilson SE, Andes W A , Goedert TJ. Relationship of CD4 lymphocyte counts to survival in a cohort of hemophiliacs infected with HIV. Multicenter Hemophilia Cohort Study. / Acquir Immune Defic Syndr 1994;7(10):1095-8. 31. Ledergerber B, Egger M , Erard V , et al. AIDS-related opportunistic illnesses occurring after initiation of potent antiretroviral therapy: the Swiss HIV Cohort Study. JAMA 1999;282(23):2220-6. 32. Montaner JS, Hogg RS, Heath K V , et al. Heterogeneity of physician agreement with recommended therapeutic guidelines for the management of HIV-associated disease. Antivir Ther 1996;l(3):157-66. 33. Zuger A , Sharp V L . 'HIV specialists': the time has come. JAMA 1997;278(14):1131-2. 34. Dooha S. Controversies: the role of HIV specialists. JAMA 1998;279(11):833; discussion 835. 35. Nitta A T . Controversies: the role of HIV specialists. JAMA 1998;279(11):834; discussion 835. 36. Legg JJ, Balano KB, Goldschmidt R H . Controversies: the role of HIV specialists. JAMA 1998;279(ll):833-4; discussion 835. 37. Strelnick A H , Futterman D , Carrascal A , et al. Controversies: the role of HIV specialists. JAMA 1998;279(ll):834-5. 173 38. Gupta SB, Gilbert RL, Brady AR, Livingstone SJ, Evans BG. CD4 cell counts in adults with newly diagnosed HIV infection: results of surveillance in England and Wales, 1990-1998. CD4 Surveillance Scheme Advisory Group. Aids 2000;14(7):853-61. 39. Liu H , Golin C E , Miller L G , et al. A comparison study of multiple measures of adherence to HIV protease inhibitors. Ann Intern Med 2001;134(10):968-77. 40. Duran S, Solas C, Spire B, et al. 'Do HIV-infected injecting drug users over-report adherence to highly active antiretroviral therapy?' A comparison between patients' self-reports and serum protease inhibitor concentrations in the French Manif 2000 cohort study. AIDS 2001;15(8):1075-7. 41. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 42. Wendel CS, Mohler MJ, Kroesen K, Ampel N M , Gifford A L , Coons SJ. Barriers to use of electronic adherence monitoring in an HIV clinic. Ann Pharmacother 2001;35(9):1010-5. 43. Gallego O, de Mendoza C, Perez-Elias MJ, et al. Drug resistance in patients experiencing early virological failure under a triple combination including indinavir. AIDS 2001;15(13):1701-6. 44. Le Moing V , Chene G , Carrieri M P , et al. Predictors of virological rebound in HIV-1-infected patients initiating a protease inhibitor-containing regimen. AIDS 2002;16(l):21-29. 174 Table 8.1 Univariate and multivariate* analysis of factors associated with death among 1416 persons first prescribed triple combination antiretroviral therapy between August 1,1996 and July 31,2000. Relative Hazard Variable Crude Adjusted* RH, (95% CI) RH, (95% CI) Start of 3-drug therapy (after July 1997 vs before) 0.71 (0.47 - 1.05) Gender (Male versus female) 1.37 (0.75- 2.50) Protease Inhibitor Use (Yes versus No) 1.48 (0.91 - 2.42) Injection Drug Use (Yes versus No) 1.03 (0.67- 1.57) 0.96(0.62 -1.49) Prior AIDS Diagnosis (Yes versus No) 2.07(1.33- 3.21) 1.04 (0.64 -1.67) Age (per year) 1.03 (1.01 - 1.05) 1.02 (1.00 -1.04) Physician Experience (> 6 versus < 5) - 0.56(0.38- 0.83) 0.67 (0.45 - 0.99) Adherence (> 75% versus <75%) 0.47 (0.32 - 0.69) 0.30 (0.20 - 0.45) Baseline CD4 count (cells/mm3) >200 50 -199 <50 1.0C 2.16(1.48-3.85 (2.58 -) 3.16) 5.75) 1.00 3.81 (2.33 - 6.24) 7.08 (4.08 -12.26) Baseline HIV-1 RNA (copies/mL) (Per LoglO increase) 1.99 (1.41 - 2.80) 1.32 (0.95 -1.82) *For all patients with non-missing values for all variables (n = 1414). Model was adjusted for variables that were significant in univariate analyses and history of injection drug use. 175 Table 8.2 Univariate and multivariate analysis of factors associated with death among 1053 persons first prescribed triple combination antiretroviral therapy between July 1,1997 and July 1,2000. Relative Hazard (RR) Variable Crude Adjusted* RH, (95% CI) RH, (95% CI) Gender (Male versus female) 1.36 (0.67- 2.75) Protease Inhibitor Use (Yes versus No) 1.38 (0.80- 2.37) Injection Drug Use (Yes versus No) 0.93 (0.52 - 1.67) 0.91 (0.49 -1.66) Prior AIDS Diagnosis (Yes versus No) 2.17 (1.18- 4.00) 0.93 (0.48 -1.78) Age (per year) 1.02 (0.99- 1.05) 1.01 (0.99 -1.04) Physician Experience (> 5versus < 5) 0.47 (0.28 - 0.79) 0.51 (0.31 - 0.86) Adherence (> 75% versus <75%) 0.69 (0.41 - 1.16) 0.43 (0.24 - 0.76) Baseline CD4+ count (cells mm3) >200 50 -199 <50 1.0C 2.16 (1.48 -2.77 (1.53 -I 3.16) 5.02) 1.00 5.49 (2.91 -10.37) 5.86 (2.68 -12.80) Baseline HIV-1 RNA (copies/mL) (Per Log 1 0 increase) 1.98 (1.31 - 2.97) 1.36 (0.92 -2.02) 176 Table 8.3(a) Adjusted relative hazard* of mortality for baseline CD4 cell count and physician experience categories. Number of HIV+ Patients Previously Treated C D 4 Cell Count Adjusted Relative Hazard (95% CI) >6 >200 1.00 \u00E2\u0080\u0094 <5 >200 1.30 (0.58 -2.90) >6 50 -199 3.89 (2.24 - 6.79) <5 50 - 199 4.07 (1.96-8.46) >6 <50 5.07 (2.50 -10.26) <5 <50 11.99 (6.33 - 22.74) *Model was adjusted for adherence, history of injection drug use, baseline plasma viral load, age, and AIDS at baseline. The number of patients in each strata is shown on Figure 8.2. 177 Table 8.3(b) Adjusted relative hazard* of mortality for baseline CD4 cell count and adherence categories. Adherence Level C D 4 Cell Count Adjusted Relative Hazard (95% CI) > 75% >200 1.00 < 75% >200 3.55 (1.72-7.35) > 75% 50 -199 4.54 (2.37-8.70) < 75% 50 - 199 9.78 (4.48 - 21.31) > 75% <50 6.19 (3.03-12.65) < 75% <50 35.71 (16.17-78.85) *Model was adjusted for physician experience, history of injection drug use, baseline plasma viral load, age, and AIDS at baseline. The number of patients in each strata is shown on Figure 8.3. 178 L E G E N D FOR FIGURES Figure 8.1 Kaplan-Meier product limit estimates of cumulative progression to death among 1,416 HIV-infected subjects who started naive on antiretroviral therapy between August 1, 1996 and July 1, 2000, stratified by CD4 cell count groupings (< 50,50 -199, > 200). Figure 8.2 Kaplan-Meier product limit estimates of cumulative progression to death among 1,414 HIV-infected subjects who started naive on antiretroviral therapy between August 1, 1996 and July 1, 2000, stratified by CD4 cell count groupings (< 50, 50 - 199, > 200), and HIV-experienced (Panel A) and non-experienced (Panel B) categories. Figure 8.3 Kaplan-Meier product limit estimates of cumulative progression to death among 1,416 HIV-infected subjects who started naive on antiretroviral therapy between August 1, 1996 and July 1, 2000, stratified by CD4 cell count groupings (< 50, 50 - 199, > 200) and adherent (Panel A) and non-adherent (Panel B) categories. 179 Figure 8.1: Survival Stratified by CD4 Cell Count Groupings 200 cells/mm3 50 - 199 cells/mm3 < 50 cells/mm3 \u00E2\u0080\u0094r-12 I 18 T\" 24 I 30 r~ 36 -1 42 312 101 49 > 200 cells/mm3 891 860 50-200 cells/mm3 355 335 < 50 cells/mm3 170 154 Time From Start of Antiretroviral Therapy (months) 827 748 653 534 428 319 148 275 126 234 96 187 79 144 57 180 Figure 8.2: Survival Stratified by Basline CD4 Cell Count and Physician Experience 1 0 0 (a) > 6 Patients 80 H 78 76 200 cells/mm 3 \u00E2\u0080\u00A2j|_50 - 199 cells/mm 3 < 50 cells/mm 3 0 6 12 18 24 30 36 Time From Start of ARV therapy (months) N's >200 702 674 649 585 505 419 340 251 50-199 265 249 240 204 171 134 99 70 < 50 105 100 97 82 61 49 33 26 \u00E2\u0080\u0094I 42 100 (b) < 5 Patients 98 96 -94 -92-90 -88 -86 84-82-80 -78 -76 -74 -72 -70 -\u00C2\u00A3 (0 68 -O 6 6 ->. 64 -4-1 = 62 -- Q CO 60- | - Q 58-1 O Q. 56 54-52-50-48 -46 -44 -42 -40 -38 -36 -34 -32 -30 200 cells/mm 3 50 - 199 cells/mm 3 1 50 cells/mm 3 0 6 12 18 24 30 Time From Start of ARV therapy (months) r~ T\" I 36 I 42 N's >200 188 185 50-199 89 85 < 50 65 54 177 78 51 162 71 44 147 63 35 114 53 30 87 45 24 60 31 23 Figure 8.3: Survival Stratified by Basline CD4 Cell Count and Adherence > 200 cells/mmJ 100 96 -94-92 -90 -88 -86 -84 -82 -80 -78 -76 74 j 72-j 70 68 -66 -64 -62 60 58 56 -| 54 52 50 48 46 44 42 40 38 36 34 j 32 -| 30 (a) > 75% A D H E R E N T 199 cells/mm 3 50 cells/mm3 \u00E2\u0080\u0094I 1 1 1 1 1 1 0 6 12 18 24 .30 36 42 Time From Start of ARV therapy (months) N's >200 630 623 604 554 427 392 320 243 50-199 286 276 264 231 198 160 123 87 < 50 145 134 130 115 89 73 54 46 100 984 96-| 94 92H 90 88 86 84-| 82 80 78 76 74-72-70-68-66-64-62-60-58-56-54-52-50-48-46 44 42-| 40 38 36 34 32 30 (b) < 75% A D H E R E N T ~~^\u00E2\u0080\u0094i , 200 cells/mm 3 50 - 199 cells/mm 3 L < 50 cells/mm 3 12 18 24 - r -30 \u00E2\u0080\u0094r-36 -1 42 Time From Start of ARV therapy (months) N's >200 50-199 < 50 261 69 25 237 59 20 223 55 18 204 44 11 181 36 7 142 108 69 27 21 14 6 3 3 182 CHAPTER 9 OUTCOMES FROM H A A R T AMONG H I V INFECTED I D U 9.1 FORWARD This chapter is presently in press with the Canadian Medical Association Journal as: Wood E, Montaner, JS, Schechter M T , Tyndall M W , O'Shaughnessy M V , Hogg RS. Highly Active Antiretroviral Therapy and the HIV-infected Injection Drug User: The State of Care in a Universal Healthcare System. 183 9.2 ABSTRACT Objective: The benefits of highly active antiretroviral therapy (HAART) are well documented, although concerns regarding access to treatment among some populations are of growing concern. We evaluated outcomes from H A A R T among HIV-infected patients with and without a history of injection drug use in a setting that delivers antiretrovirals and H I V / A I D S care free of charge. Methods: Populational analysis of all antiretroviral naive HIV-infected men and women who initiated triple therapy between August 1, 1996 and July 31, 2000, and who were followed until March 31, 2002 in a province-wide HIV treatment program. We evaluated time to discontinuation of therapy, HIV-1 R N A suppression, and HIV-1 R N A rebound using Kaplan-Meier methods and Cox proportional hazards regression. We also evaluated if differences in virological response persisted over the study period by fitting logistic models that considered suppression and rebound as a binary response. Results: Overall, 1422 patients initiated H A A R T among whom 359 (25.3%) were IDU. In Cox analyses, history of injection drug use was associated with more rapid discontinuing therapy (Adjusted Relative Hazard [ARH]: 1.40 [95% CI: 1.20 - 1.63]). With regards to clinical outcome, IDU had slower rates of HIV-1 R N A suppression (ARH: 0.72 [95% CI: 0.62 - 0.83]), and among IDU who did achieve virological suppression, faster rates of virological failure were observed (ARH: 1.41 [95% CI: 1.13 - 1.76]) in comparison to non-IDU. Worse virological outcomes persisted throughout follow-up when these endpoints were considered in logistic models. Conclusions: IDU discontinued treatment at elevated rates, and experienced lower rates of virological suppression and higher rates of virological failure. The data suggest that unless substantial changes are made to the H I V / A I D S care delivery systems, in areas where there is high HIV prevalence among IDU, healthcare services may soon require additional resources to deal with high levels of morbidity and mortality among HIV-infected injection drug users. 184 9.3 INTRODUCTION The benefits of highly active antiretroviral therapy ( H A A R T ) in the management of HIV disease are well established. Through the suppression of plasma HIV-1 R N A , H A A R T has been shown to improve CD4 cell counts and in turn decrease morbidity and mortality among HIV-infected patients.I' 2 As a result, triple drug therapy has become the standard of care for HIV infected individuals.^ In North America, these findings have emerged amidst growing concerns regarding inequitable access to antiretroviral therapy among those infected during later stages of the HIV epidemic particularly illicit injection drug users (IDU).4-6 i n British Columbia, increases in the number of HIV-infected injection drug users presenting for antiretoviral treatment were first reported in 1995.7 However, the majority of HIV infections among IDU in the province did not occur until after this time,^ and hence the bulk of HIV-infected IDU will only be starting to require HIV treatment over the next several years.3/9 At present, little work has been done to calculate population-level estimates of the proportion of IDU initiating H A A R T and their outcomes from treatment in settings that deliver H I V / A I D S care and antiretroviral therapy free of charge. Since a growing number of IDU will soon be at risk of HIV-related morbidity and mortality in many settings,4\"6 it is critical to evaluate how the healthcare system has performed in terms of treating IDU who have presented for antiretroviral therapy in the H A A R T era. Therefore, the present study was conducted to characterize all H I V / A I D S patients who have initiated H A A R T in the province of British Columbia since 1996, and to evaluate time to treatment 185 discontinuation, and virological response to treatment among patients with and without a history of injection drug use. 9.4 METHODS The dispensation of antiretroviral medications in the province of British Columbia, Canada has been described in detail elsewhere. 10,11 In brief, the BC Centre for Excellence H I V / A I D S Treatment Program remains the only free source of antiretroviral medications in the province, and < 1% of HIV infected British Columbians receive antiretrovirals from other sources.10 In June 1996 the Centre adopted plasma viral load driven antiretroviral therapy guidelines, consistent with those put forward by the International AIDS Society - U S A 12. Consistent with contemporary practice 13, the Centre guidelines were revised in July 1997 to recommend triple combination therapy for all antiretroviral naive individuals with plasma HIV-1 R N A levels greater than 5,000 copies/mL or CD4 cell counts below 500 cells/mm 3 . A l l HIV-infected men and women in the current study were entered into the Centre's H I V / A I D S D r u g Treatment Program when they were first prescribed antiretroviral agents. Physicians enrolling an HIV-infected individual must complete a drug request enrollment form, which acts as a legal prescription and supplies baseline information, including past HIV-specific drug history, CD4 cell count, plasma HIV-1 R N A , current drug requests, and enrolling physician data. At the time of the initial refill, participants are asked to complete a participant survey, which elicits information on sociodemographic characteristics. The treating physicians are also asked to complete a clinical staging form using the World Health Organization (WHO) clinical staging system.14 In the present study, we evaluated all HIV-infected men and women 186 who were antiretroviral naive and were first prescribed triple drug antiretroviral therapy between August 1, 1996 and Jul 31, 2000. Triple therapy was defined as two nucleoside reverse transcriptase inhibitors and either a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor. As an exploratory analysis we evaluated the baseline demographic and clinical characteristics of I D U in comparison to non-IDU. Categorical explanatory variables described below were analyzed using Pearson's C h i -square test and continuous variables were analyzed using the Wilcoxon rank sum test. We also fit a logistic model by adjusting for all variables that were significant in univariate analyses. In order to evaluate duration on H A A R T after initiating treatment, we evaluated the time to the first treatment discontinuation. Discontinuation was defined as the first day of a > three month period without receiving any antiretrovirals. This outcome was a priori selected based on an earlier analysis in which we found that receiving antiretrovirals < 75% of the time during the first year on therapy was associated with elevated mortality in the cohort 1 6 . However, since we assumed that many patients would restart therapy and that virological outcomes are more clinically relevant we performed several additional analyses as described below. First, we also evaluated time to virological suppression. A s previously,! ' 7 ' 1^ suppression was defined as the first of two consecutive plasma HIV R N A measures < 500 copies/mL. Event-free patients were right censored at the date of last HIV R N A measure as of March 31, 2002 or at the date of last HIV R N A measure if prior to this date. 187 We also evaluated time to HIV-1 R N A rebound after initial virological suppression below 500 copies/mL at least once. As previously,!^ HIV R N A rebound was defined as the first of two consecutive plasma HIV-1 R N A levels above 500 copies/mL. To be conservative, the rebound event was assumed to occur at the mid-point between the last HIV R N A measure < 500 copies/mL and the first of the two consecutive measures > 500 copies/mL. In order to be consistent with previous studies we only evaluated patients who became suppressed prior to 32 weeks after first initiating therapy.^ Suppressed patients who did not rebound were right censored as of March 31, 2001 or at the last HIV R N A measure if follow-up ended prior to this date. For the three measures of treatment outcome (discontinuation, suppression, and rebound), cumulative event rates were estimated using Kaplan-Meier methods. Cox regression was then used to calculate univariate and adjusted relative hazards and 95% confidence intervals (CIs).!9 The assumption of proportional hazards was validated by inspection of log (-log [survival function]) estimates against log time plots. In addition, we assumed that differential rates of virological suppression and rebound would be of limited clinical significance if differences diminished over time, and as a sub-analyses we also evaluated the adjusted odds of ever achieving virological suppression and virological rebound at any time during the entire follow-up by fitting a logistic model that considered these outcomes as a binary response. A number of salient baseline prognostic variables were examined in all analyses including: log 1 0 transformed plasma HIV-1 R N A levels (continuous); CD4 cell count (continuous); protease-inhibitor use in the initial regimen (yes versus no); time of initiation of therapy (after versus before July 1997); a prior 188 clinical diagnosis of AIDS (yes versus no); age (continuous); gender (male versus female); and physician experience (> 6 versus < 5 patients previously enrolled in the program). July 1997 was used as a temporal cutoff in our analysis because this reflected the time when the therapeutic guidelines for antiretroviral therapy were changed to recommend universal use of triple drug regimens in this province. The definition of physician experience was a priori selected based on the findings of previous studies.20/21 In each multivariate analysis described above, we adjusted for all variables that were statistically associated with history of IDU in the preliminary analysis or the outcome of interest (discontinuation, suppression, and rebound) in a univariate analyses. A l l statistical analyses were performed using SAS software version 6.0 (SAS, Cary, N C ) . A l l tests of significance were two-sided, with a p-value of less than 0.05 indicating that an association was statistically significant. 9.5 RESULTS Between August 1, 1996 and July 1, 2000 1,583 antiretroviral naive participants aged 18 years and older began triple combination therapy including 2 nucleoside reverse transcriptase inhibitors plus either a protease inhibitor (PI) or a non-nucleoside reverse transcriptase inhibitor (NNRTI). Of these, 161 (10.2%) were excluded from this analysis for not having both baseline CD4 cell count and plasma HIV-1 R N A level measures available within 6 months prior to the start of antiretroviral therapy. Persons excluded from this analysis were more likely to be younger (p = 0.04) and taking protease inhibitors (p = 0.02). In the study population, 359 (25.2%) patients were identified as having a history of injection drug use and 1063 (74.8%) were identified as non-IDU. 189 Table 1 shows the overall study population stratified into IDU and non-IDU, and the results of univariate statistical comparisons of baseline clinical and socio-demographic characteristics. As shown here, IDU were less likely to be male (p - 0.001), to have a baseline clinical diagnosis of AIDS (p = 0.026), and to have an HIV-experienced physician (p = 0.015). Conversely, I D U were more likely to receive a protease inhibitor in the initial regimen (p = 0.026), and to have a higher CD4 cell count (p = 0.002). We detected no statistical difference with regards to date of therapy initiation, age, or baseline HIV R N A between IDU and non-IDU. After adjustment for baseline CD4 cell count, protease inhibitor use, and baseline AIDS, IDU were less likely to be male (Adjusted Odds Ratio [AOR]: 0.48 [95% CI: 0.36 - 0.66]) and less likely to be treated by experienced physicians in adjusted analyses (AOR 0.70 [95% CI: 0.53 - 0.93]). Figure 1 shows the Kaplan-Meier estimates of the cumulative therapy discontinuation rate for the overall cohort stratified by IDU and non-IDU. The curve does not drop until 1 month after the initiation of treatment since all individuals are initially dispensed 1 months worth of therapy. As shown here, significantly higher rates of therapy discontinuation were noted between IDU and non-IDU (log rank p < 0.01). Table 2(a) shows the crude and adjusted relative hazards of discontinuing therapy among IDU in comparison to non-IDU. A s shown here, I D U discontinued therapy at elevated rates even after adjustment for all relevant covariates (Adjusted Hazard Ratio: [ARH]: 1.40 [95% CI: 1.20 - 1.63]). For each multivariate analysis, the variables that were adjusted for and their statistical significance in the adjusted analyses are noted where appropriate on Table 2. 190 Figure 2 (Panel A) shows the Kaplan-Meier estimates of the cumulative virological suppression rate for the overall cohort stratified by IDU and non-IDU. A s shown here, statistically significant differences in time to virological suppression were noted between IDU and non-IDU (log rank p < 0.01). Table 2(b) shows the crude and adjusted relative hazards of achieving virological suppression for I D U in comparison to non-IDU. A s shown here, after adjustment for relevant covariates, IDU were slower to suppress plasma H I V R N A (ARH: 0.72 [95% CI: 0.62 - 0.83]). In sub-analyses, when we fit a logistic model that considered any suppression during follow-up, IDU were less likely to ever achieve suppression (AOR: 0.69 [95% CI: 0.53 - 0.91]) during the entire follow-up period. Figure 2 (Panel B) shows the Kaplan-Meier estimates of the cumulative virological rebound rate for the 970 (68.2%) patients who achieved virological suppression at least once prior to 32 weeks after the initiation of therapy stratified by I D U and non-IDU. A s shown here, statistically significant differences in time to virological rebound were noted between IDU and non-IDU (log rank p < 0.05). In Cox regression analyses shown in Table 2(c), after adjustment for relevant covariates IDU virologically failed treatment at elevated rates in comparison to non-IDU (ARH: 1.41 [95% CI: 1.13 - 1.76]; p = 0.003). In sub-analyses we found that IDU were more likely to ever virologically fail at any time during follow-up (AOR: 1.52 [95% CI: 1.11 - 1.08]; p = 0.009) when rebound was considered as a binary response in adjusted analyses. 8.6 DISCUSSION Patients with a history of injection drug use had dramatically higher therapy discontinuation rates in comparison to non-IDU. This translated into 191 slower rates of virological suppression among IDU, and we found that IDU were less likely to ever achieve virological suppression during follow-up. Among IDU who achieved virological suppression, higher rates of virological failure were observed in comparison to non-IDU and this difference also persisted throughout follow-up. We also found that I D U are more likely to be female and treated by physicians with less experience treating HIV. Although the fact that IDU were more likely to be female is not surprising considering that the majority of non-IDU were gay men, the fact that IDU were treated by physicians with less experience treating HIV disease is cause for concern. This suggests that IDU may be accessing care in different ways than non-IDU, and is of particular relevance given that physician experience has recently been associated with survival from HIV disease in our cohort.21 We also found that IDU discontinue H A A R T at highly elevated rates in comparison to non-IDU. Although previous studies have demonstrated that IDU may have worse levels of adherence to antiretroviral therapy than non-IDU, the majority of these studies have examined adherence to daily treatment regimens among patients that continue to present for clinic visits and who participate in questionnaires regarding their daily adherence levels.22/23 TJhe present study highlights that a large number of I D U stop therapy soon after initiating H A A R T , and demonstrates the limitations of current HIV care delivery systems to IDU. Strategies that have been successful in improving access to antiretrovirals among HIV infected I D U include directly observed therapy programs, although data from these programs in Canada are not yet available.24,25 192 The clinical implications of the high rates of treatment discontinuation among IDU were evident when we evaluated virological response to H A A R T among IDU in comparison to non-IDU. We found that IDU were significantly slower to suppress plasma HIV R N A and were substantially less likely to ever suppress plasma H I V R N A during follow-up in comparison to non-IDU. Similarly, we found that IDU who achieved plasma HIV R N A suppression, also had higher rates of virological failure in comparison to non-IDU. The elevated rates of treatment discontinuation and poor virological responses indicate that IDU will likely experience more rapid disease progression in our setting.^/ 1 1 / 2 6 Access to H I V / A I D S care is most pressing in the developing w o r l d , 2 7 but also presents major challenges in developed world, even in the context of universal healthcare systems such as our o w n . 1 0 ' 2 ^ ' ^ The greatest limitation of the present study is that we have only examined patients who actually initiated triple therapy during the study period. To put the scale of this concern into perspective, while 359 IDU initiated therapy, it is estimated that approximately 40% of the province's 5,000 - 15,000 IDU are HIV i n f e c t e d . 3 0 ' 3 1 A second limitation that has likely resulted in a strong conservative bias in our analyses is that injection drug use is a stigmatized behavior that may be under-reported by IDU. Similarly, persons who use non-injection illicit drugs, such as crack cocaine users, may be at similar risk of discontinuing H A A R T and were considered as non-IDU in the present s tudy. 3 2 9.7 SUMMARY In summary, we found that IDU discontinued therapy at elevated rates, and that I D U had lower rates of virological suppression and higher rates of virological rebound. Our findings demonstrate that, despite the province's 193 universal healthcare system that delivers antiretrovirals and AIDS care free of charge, there are major flaws in the present HIV treatment delivery system for IDU. Unless substantial changes are made to the H I V / A I D S care delivery systems, in urban areas of high HIV prevalence,4'5,28 healthcare services may soon experience high levels of morbidity and mortality among patients infected with HIV through injection drug use. 194 9.8 REFERENCES 1. Hammer S M , Squires K E , Hughes M D , et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med 1997;337(ll):725-33. 2. Montaner JS, Reiss P, Cooper D , et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: the I N C A S Trial. Italy, The Netherlands, Canada and Australia Study. JAMA 1998;279(12):930-7. 3. Carpenter C C , Cooper D A , Fischl M A , et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000;283(3):381-90. 4. Karon JM, Fleming PL, Steketee RW, De Cock K M . HIV in the United States at the turn of the century: an epidemic in transition. Am J Public Health 2001;91(7):1060-8. 5. Anderson K H , Mitchell JM. Differential access in the receipt of antiretroviral drugs for the treatment of AIDS and its implications for survival. Arch Intern Med 2000;160(20):3114-20. 6. Rosenberg PS. Scope of the AIDS epidemic in the United States. Science 1995;270(5240):1372-5. 7. Montaner JS, Hogg RS, Craib KJ, Strathdee SA, O'Shaughnessy M V , Schechter M T . IDU numbers and AIDS. BC Medical Journal 1995;37(4):239. 8. Strathdee SA, Patrick D M , Currie SL, et al. Needle exchange is not enough: lessons from the Vancouver injecting drug use study. AIDS 1997;11(8):F59-65. 195 9. Mellors JW, Munoz A , Giorgi JV, et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Annals of Internal Medicine 1997;126(12):946-54. 10. Wood E, Schechter M T , Tyndall M W , Montaner JS, O'Shaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000;14(9):1229-35. 11. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 12. Carpenter C C , Fischl M A , Hammer SM, et al. Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. International AIDS Society-USA. JAMA 1996;276(2):146-54. 13. Carpenter C C , Fischl M A , Hammer S M , et al. Antiretroviral therapy for HIV infection in 1997. Updated recommendations of the International AIDS Society-USA panel. JAMA 1997;277(24):1962-9. 14. World Health Organization. Acquired immune deficiency syndrome (AIDS): interim proposal for a W H O staging system for HIV infection and disease. Wkly Epidemol Rec 1990; 65: 221-228. 15. Ledergerber B, Egger M , Opravil M , et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss H I V Cohort Study. Lancet 1999;353(9156):863-8. 16. H o g g RS, Heath K V , Bangsberg D , et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up. AIDS 2002;16(7):1051-1058. 196 17. Phillips A N , Staszewski S, Weber R, et al. HIV viral load response to antiretroviral therapy according to the baseline CD4 cell count and viral load. JAMA 2001;286(20):2560-2567. 18. Low-Beer S, Yip B, O'Shaughnessy M V , Hogg RS, Montaner JS. Adherence to triple therapy and viral load response. / Acquir Immune Defic Syndr 2000;23(4):360-l. 19. Cox DR. Regression models and life tables. / Royal Stat Soc B 1972:187-202. 20. Kitahata M M , Koepsell T D , Deyo R A , Maxwell C L , Dodge WT, Wagner E H . Physicians' experience with the acquired immunodeficiency syndrome as a factor in patients' survival. N Engl J Med 1996;334(ll):701-6. 21. Wood E, Hogg RS, Yip B, Harrigan RP, O'Shaughnessy M V , Montaner JS. The Impact of Baseline CD4 T-Cell Count and Adherence on Rates of Disease Progression among HIV-infected Patients Initiating Triple D r u g Antiretroviral Therapy. 9th Conf on Retroviruses and Opportunistic Infections 2002, Feb 24-28. Seattle, W A . 22. Duran S, Solas C, Spire B, et al. 'Do HIV-infected injecting drug users over-report adherence to highly active antiretroviral therapy?' A comparison between patients' self-reports and serum protease inhibitor concentrations in the French Manif 2000 cohort study. AIDS 2001;15(8):1075-7. 23. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 24. Bangsberg DR, Mundy L M , Tulsky JP. Expanding directly observed therapy: tuberculosis to human immunodeficiency virus. Am J Med 2001;110(8):664-6. 197 25. Woodward W C . Should directly observed therapy be considered for treatment of HIV? JAMA 1996;276(24). 26. Tyndall M W , Craib KJ, Currie S, L i K, O'Shaughnessy M V , Schechter M T . Impact of HIV infection on mortality in a cohort of injection drug users. / Acquir Immune Defic Syndr 2001;28(4):351-7. 27. Wood E, Braitstein P, Montaner JS, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa [see comments]. Lancet 2000;355(9221):2095-100. 28. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 29. Wood E, Montaner JS, Chan K, et al. Socioeconomic status, access to triple therpay, and survival from HIV-disease since 1996. AIDS 2002;16(15):2065-2072. 30. Spittal P M , Craib KJ, Wood E, et al. Risk factors for elevated HIV incidence rates among female injection drug users in Vancouver. C M A / 2002;166(7):894-9. 31. British Columbia Centre for Disease Control. 2000 British Columbia Annual Summary of Reportable Diseases. 32. Edlin BR, Irwin K L , Faruque S, et al. Intersecting epidemics\u00E2\u0080\u0094crack cocaine use and HIV infection among inner-city young adults. Multicenter Crack Cocaine and HIV Infection Study Team. N Engl J Med 1994;331(21):1422-7. 198 Table 9.1 Univariate analyses of baseline characteristics among treatment naive patients initiating H A A R T stratified by history of injection drug use. History of Injection Drug Use No Yes Unadjusted Characteristic 1076, (%) 340, (%) Odds Ratio (95% C.L) p-value Date of therapy initiation Prior to July 1997 259 (24.4) 103(28.7) After July 1997 804 (75.6) 256(71.3) 0.8 (0.61-1.04) 0.104 Gender Female 139 (13.1) 85(23.7) Male 924 (86.9) 274(76.3) 0.49 (0.36-0.66) 0.001 Age Median 37.1 37.6 Inter-quartile range (31.9-44.0) (32.3-43.0) 0.99 (0.98-1.00) 0.160 Protease Inhibitor Use* No 345 (32.5) 94(26.2) Yes 718 (67.5) 265(73.8) 1.36 (1.04-1.77) 0.026 Baseline AIDS Diagnosis No 914 (86.0) 325 (90.5) Yes 149 (14.0) 34(9.5) 0.64 (0.43-0.95) 0.026 Physician Experience < 5 patients 240 (22.6) 104 (29.0) > 6 patients 821 (77.4) 255(71.0) 0.72 (0.55-0.94) 0.015 Baseline CD4 cell count (cells/mm3)** Median 260 290 Inter-quartile range (120-410) (140-460) 1.09 (1.03-1.14) 0.002 Plasma HIV-1 RNA (per log10) Median 5.1 5.1 Inter-quartile Range (4.56 - 5.48) (4.63-5.49) 1.15 (0.97-1.37) 0.118 * PI use in the initial HAART regimen, ** Odds Ratio reflects per 100 cell increase. 199 Table 9.2(a) Crude and Adjusted* Relative Hazards (RH) of treatment discontinuation. RH R H * CRUDE (95% CI) p-value ADJUSTED (95% CI) p-value Injection Drug Use (Yes versus No) 1.54 (1.32 -1.79) <0.001 1.40 (1.20-1.63) <0.001 * M o d e l was adjusted for M D experience, gender*, plasma HTV R N A , age + , Baseline AIDS*, date of therapy initiation*, baseline C D 4 cell count*, and protease inhibitor i n the init ial regimen. Note: * indicates variable remained p < 0.05 i n multivariate analyses. Table 9.2(b) Crude and Adjusted* Relative Hazards (RH) of HIV R N A Suppression. RH RH* CRUDE (95% CI) p-value ADJUSTED (95% CI) p-value Injection Drug Use (Yes versus No) 0.64 (0.56-0.74) <0.001 0.72 (0.62-0.83) <0.001 \" M o d e l was adjusted for M D experience, gender*, plasma H I V R N A * , age *, Baseline A I D S , date of therapy initiation, baseline C D 4 cell count*, and protease inhibitor i n the initial regimen*. Note: * indicates variable remained p < 0.05 i n multivariate analyses. Table 9.2(c) Crude and Adjusted* Relative Hazards (RH) of HIV R N A Rebound. RH R H * CRUDE (95% CI) p-value ADJUSTED (95% CI) p-value Injection Drug Use (Yes versus No) 1.52 (1.22-1.88) <0.001 1.41 (1.13-1.76) 0.003 * M o d e l was adjusted for M D experience, gender, plasma HTV R N A , age *, Baseline A I D S , date of therapy initiation*, baseline C D 4 cell count*, and protease inhibitor i n the init ial regimen. Note: * indicates variable remained p < 0.05 i n multivariate analyses. 2 0 0 L E G E N D FOR FIGURES Figure 9.1 Time to treatment discontinuation among patients initiating H A A R T stratified by history of injection drug use. Note, the curve does not drop until 1 month after the initiation of therapy since all individuals are initially dispensed 1 months worth of treatment. Figure 9.2 Time to HIV R N A suppression (Panel A) and HIV R N A rebound (Panel B) among patients initiating H A A R T stratified by history of injection drug use. 201 Figure 1: Time to treatment discontinuation > 3 months o \u00C2\u00A3 20 J , , , , 0 6 1 2 1 8 2 4 Time from Start of ARVs (months) Figure 2: Time to HIV RNA suppression (Panel A) and rebound (Panel B) stratified by history of injection drug use 0 6 12 18 24 0 6 12 18 24 T i m e f r o m S t a r t o f A R V s T i m e f r o m F i r s t U n d e t e c t a b l e H IV R N A 203 CHAPTER 10: SUMMARY, RECOMMENDATIONS, FURTHER RESEARCH, AND CONCLUSIONS 10.1 S U M M A R Y OF S T U D Y F I N D I N G S At the beginning of this research project, a review of available research on the topics of access to antiretroviral therapy was undertaken.I\" 3 What is clear from previous research is that there is now a great deal of evidence demonstrating that lower income populations, women, ethnic minorities, and illicit drug users have poorer access to antiretroviral therapy and higher rates of treatment discontinuation. While the majority of these studies have been conducted among living cohorts, data is emerging to suggest that limited access to antiretrovirals is contributing to the ongoing H I V / A I D S mortality rates in the developed world. In Chapter 4, I demonstrated that 33% of HIV-related deaths in the province of British Columbia during the period 1995 to 2001 occurred among untreated individuals. Furthermore, in this study Aboriginal ethnicity, female gender, and lower socio-economic status were all associated with dying without antiretroviral treatment. Among those who accessed treatment prior to death, less than half received consistent treatment, and similar socio-demographic factors were associated with elevated rates of therapy discontinuation. Although previous studies of access to antiretrovirals have tended to be conducted among living populations, these findings are consistent with the experience of other HIV-positive ethnic minorities in other settings.I\"3 In 2 0 4 addition, I found associations between lower income and female gender, and not receiving antiretrovirals, as well as receiving inconsistent treatment, prior to death. These findings are also consistent with previous studies that have found lower socio-economic status and female gender to be associated with worse access to healthcare 3\"6, and interventions aimed at improving access to antiretrovirals among females and lower income persons must also be made a priority. 1 ' 3 ' ' 7 These findings demonstrate that novel strategies will be required to improve access to antiretrovirals among populations with poor access to care. Given that these data were derived in a universal healthcare setting where all H I V / A I D S care and antiretrovirals are available free of charge, it is likely that similar problems exist in many other settings in the developed world . 3 In Chapter 5,1 demonstrated that large increases in AIDS mortality and decreases in life expectancy may be anticipated in urban areas where HIV prevalence is high among injection drug users if the low level of antiretroviral medication use among injection drug users persists.8 In Vancouver's Downtown Eastside, for instance, if the present low-level of antiretroviral medication use continues, the life expectancy at birth in the year 2006 will be 56.9 for males and 68.6 for females, and 503 AIDS deaths can be expected between 1999 and 2006. The life expectancies projected in this scenario, if occurring presently, would put the Downtown Eastside below all but the very least developed nations, during a time that life expectancy is rising in the developed world. Previous studies have demonstrated that all motor vehicle accidents are responsible for 0.5 years of life expectancy lost among Canadian men.9 Similarly, all lung cancer deaths are responsible for a 0.9 years of life expectancy lost among Canadian men and 2 0 5 breast cancer for 0.5 years of life expectancy lost among all Canadian women.\" Viewed in this light, the 4 year difference in potential life expectancies forecast by scenarios one and two represents an enormous difference in health status. There are a number of potential benefits attributable to the use of antiretroviral therapy beyond the documented survival benefits. For example, numerous studies have shown that antiretroviral therapy is extremely effective for decreasing plasma viral load and suppressing viral replication.lO'H A s a result, investigators have suggested that persons on antiretroviral therapy may be less infectious than individuals not on therapy. 1 2 As such, increasing access to and use of antiretrovirals in the Downtown Eastside has the potential to reduce HIV incidence rates among those that continue to engage in high-risk activities. In addition, the testing and counseling which accompany HIV treatment also has the potential to substantially reduce HIV risk behavior. These findings suggest that providing antiretroviral medications as part of a comprehensive intervention program could help to reduce future health care costs. Although efforts to increase access to antiretroviral medications should be a priority, reasonable coverage targets have not been established. Increasing the use of antiretroviral therapy in the Downtown Eastside will require the addition of a number of client tailored resources, and could most easily be achieved as part of a comprehensive preventative intervention/harm reduction strategy. The findings of this modeling exercise support the growing body of research, outlined in Chapter 2, which suggests that certain preventative interventions and harm reduction strategies are urgently needed to avert major declines in health status. 2 0 6 In Chapter 6, the methods I developed in Chapter 5 were used to model the potential impact of antiretroviral use in South Africa, as an example of a typical lesser-developed nation. 13 This study demonstrated that, despite the price reductions in the cost of antiretrovirals,!4 there are major barriers to the widespread provision of antiretroviral therapy in the developing world. However, the estimates derived suggested that the drug cost of antiretroviral prophylaxis is low and that low-level prophylaxis use could have a substantial demographic and epidemiological impact. In Chapter 7,1 demonstrated that in a population-based setting in which antiretrovirals are delivered free of charge through a universal healthcare system, there is evidence that lower socio-economic status is associated with shorter survival. 15 However, it would be incorrect to conclude that persons in lower socio-economic strata derive less benefit from H A A R T since lower income patients were more likely to be prescribed dual therapy. Previous studies have demonstrated that persons initially prescribed double therapy have a lesser virologic response and more rapid viral rebound, 16/17 a s w e n a s significantly shorter survival than persons who initiate therapy with triple therapy.!8,19 Furthermore, pre-treatment with dual therapy may preclude a worse treatment outcome even among those who later switch to a triple drug regimen because of the more rapid development of resistance.20/21 Limited access to H A A R T among persons with a history of injection drug use and persons of lower socio-economic status may reflect the concerns of physicians who have hoped that persons with less stable lifestyles may have better adherence to a less complex regimen. In addition, it is also possible that concerns regarding possible transmission of protease inhibitor or N N R T I 2 0 7 resistant virus influenced prescribing decisions. However, it was recently argued that physicians should not indefinitely withhold H A A R T from patients who are thought to be poorly adherent, 2 2 and studies have shown that providers may be poor judges of adherence. 2 3\" 2^ The findings of Chapter 7 indicate that further efforts must be focussed on increasing the awareness, implementation, and monitoring of therapeutic guidelines to ensure that physicians treating HIV positive patients deliver care in accordance with the most recent evidence-based approaches. In Chapter 8, I demonstrated that mortality among patients on H A A R T was dramatically altered by very conservative adjustments for physician experience and adherence, suggesting that inappropriate care of advanced patients and patient non-adherence may be the strongest determinants of patient survival, instead of when therapy is initiated prior 200 cells /mm 3 . 2 ^ This manuscript is the first to demonstrate statistically significant differences in survival among adherent and non-adherent patients at all levels of CD4 count, and in particular the dramatic impact adherence has on the survival of highly advanced patients. These findings will likely contribute to the debate over when therapy is best ini t ia ted , 2 7 and will be useful for guidelines considering the management of the large proportion of patients that continue to present for treatment when CD4 cell counts have decreased to below 200 cells /mm3.28/29 These findings should also help to emphasize the role of patient adherence in the treatment of HIV infection. Finally, these findings also raised questions about why a large proportion of patients pick up only a limited number of their prescribed antiretrovirals after initiating therapy. 208 In Chapter 9, I demonstrated that, in comparison to other patients, injection drug users discontinued therapy at elevated rates, and that this translated into lower rates of virological suppression and higher rates of virological rebound among this population.30 Although previous studies have demonstrated that IDU may have worse levels of adherence to antiretroviral therapy than non-IDU, the majority of these studies have examined adherence to daily treatment regimens among patients that continue to present for clinic visits and who participate in questionnaires regarding their daily adherence levels.23/31 xhe present study highlights that a large number of I D U stop therapy soon after initiating H A A R T , and demonstrates the limitations of current HIV care delivery systems to IDU. These findings indicate that, despite the province's universal healthcare system that delivers antiretrovirals and AIDS care free of charge, there are major flaws in the present HIV treatment delivery system for patients with a history of injection drug use. Together, the data presented here demonstrate that unless substantial changes are made to the province's H I V / A I D S care delivery systems, in urban areas of high HIV prevalence, healthcare services may soon experience high levels of morbidity and mortality among patients infected with HIV as a result of poor access to antiretrovirals. In particular, this thesis demonstrates that Aboriginal persons, persons with history of injection drug use, women, and persons of lower socio-economic status may be at highest risk of poor access. 10.2 U N I Q U E C O N T R I B U T I O N , I M P A C T , A N D I M P L I C A T I O N S This thesis makes six unique contributions to the expansion of knowledge. First, as part of this thesis, I initiated and prepared the first draft of the review that is presented in Chapter 2 , and will be published in the international medical 2 0 9 journal AIDS' annual supplement. This work provides a comprehensive summary of the research in this area to date, and central to this review were the studies that make up the bulk of this thesis. This review paper also contains recommendations for further areas of research and intervention. To ensure that these recommendations were relevant and comprehensive, I took the initiative to gather an international team of experts to provide feedback and suggestions on these recommendations and to be co-authors on this paper. When published, the review will be available on Medline and will be a comprehensive and thought-provoking resource for all those interested in this area of HIV research. Second, I assembled the cohort of individuals who have died of HIV since 1995, and through a series of linkages, was able to determine their access to antiretrovirals prior to death. Not only is this study the first to demonstrate high levels of mortality among untreated individuals, this study also is the first to demonstrate that the problem of treatment discontinuation may be parallel to the problem of access. Specifically, that after therapy is first initiated, those populations that have been shown to be at risk of not accessing therapy are also at risk of discontinuing therapy, often soon after it is initiated. Third, the modeling exercises I prepared in Chapters 5 and 6 are the first demographic projections to demonstrate how the HIV epidemic could be dramatically altered by the use of antiretroviral therapy. This is of particular relevance in areas of high H I V prevalence among injection drug users. Furthermore, as part of this thesis, this methodology was applied to a lesser-developed country. Not only did this research quantify the potential benefits of treatment in a typical lesser-developed setting, but it also demonstrated that, despite recent price reductions, the price of antiretrovirals continues to present a 2 1 0 major obstacle to widespread use of H A A R T in developing nations. I am also proud to note that this modeling exercise resulted in a donation by a major pharmaceutical company of free antiretrovirals for the prevention of mother-to-child transmission of HIV in the developing world (see Appendix 3). Fourth, the work I undertook as part of this thesis has also augmented research at the BC Centre for Excellence in H I V / A I D S . For instance, it was central to this thesis to apply geographical information systems technology to link individual's postal codes to census based income data through the use of a postal code conversion file. This enabled an evaluation of the role of socio-economic status with respect to access to antiretroviral therapy prior to death, as well as access to triple therapy and survival among persons initiating antiretrovirals since 1996. These studies demonstrated that socio-economic status is both associated with access to antiretrovirals prior to death, and that socio-economic status is not associated with survival among patients prescribed H A A R T . However, an additional contribution this thesis makes, is the discovery that, despite adjustment for the clinical characteristics that should have driven physician's decisions, persons of lower socio-economic status were less likely to be prescribed triple therapy. This contribution to the BC Centre for Excellence in H I V / A I D S ' research activities has been, and will continue to be, applied to other studies. 3 2 Fifth, the study I presented in Chapter 8, is one of the first studies to demonstrate the role of adherence to H A A R T in patient survival after the initiation of therapy. This paper demonstrated relative hazards on the order of a 35 fold greater risk among non-adherent patients with low CD4 cell counts. This paper is also the first to demonstrate the role of physician experience in survival 211 with HIV disease in the H A A R T era. These findings make a unique contribution in several ways. First, they provide critical data to inform the debate over when H A A R T should optimally be initiated. In particular, the study demonstrated that a CD4 cell count of 200 - 350 does not represent an irreversible biological threshold beyond which response to therapy is compromised as has been suggested in therapeutic guidelines. Conversely, this work demonstrated that patient non-adherence may be the strongest determinant of patient survival, rather than when antiretroviral therapy is initiated prior to 200 cells/mm 3 . Finally, this study also provided a unique contribution by raising the question of why patients are non-adherent. This question was partially evaluated in the literature review when it was demonstrated that many patients who were previously defined as non-adherent, actually discontinued therapy within a month of starting treatment. In addition, this concern was partially evaluated in the final study when it was demonstrated the role that history of injection drug use plays in treatment discontinuation. Sixth, the work I conducted as part of this thesis makes a unique contribution with regards to the close examination of virological response and the role of treatment discontinuation for patients who have a history of injection drug use. Although at present, only a fraction of British Columbia's HIV-positive injection drug users have initiated antiretroviral therapy, this study demonstrated the major flaws in our current model of care for HIV-infected patients with a history of injection drug use. 212 10.3 R E C O M M E N D A T I O N S The strengths and limitations of the studies presented here have been outlined in Chapter 3, and in detail as part of each study. For instance, the fact that antiretrovirals are centrally distributed enables population level analyses to be conducted, whereas using an ecological measure of socio-economic status may result in a bias from misclassification of individuals who do not live in an area that is representative of their socio-economic status. Nevertheless, the studies provide a great deal of insight into the challenges of delivering antiretrovirals to marginalized populations and three recommendations can be confidently made. First, throughout the developed world, I believe interventions to improve access to and retention in antiretroviral therapy programs are urgently required to prevent high levels of ongoing AIDS mortality. In particular, the barriers and strategies outlined in Chapter 2 deserve immediate attention. For individuals who have not accessed therapy, strategies to improve contact with HIV care providers are needed. In addition, when contact is made, guidelines for physicians must be based on available evidence. 15,23-25 This evidence suggests that physicians should seek to address modifiable barriers to adherence and retention in HIV treatment prior to the start of therapy among patients not requiring immediate treatment 33,34 j n addition, it suggests that no patient should be denied the opportunity initiate H A A R T regardless of perceived or real barriers to optimal adherence including continued illicit drug use.35 Locally, given the challenges required to be optimally adherent and the high prevalence of HIV among injection drug users, addressing these barriers 213 will require a major increase in programs that aim to stabilize the lives of those addicted to illicit drugs, including an expansion of drug treatment services. This will most easily be accomplished through a provincial and national drug strategy that acknowledges the high prevalence of HIV among illicit drug users, and as the epidemic evolves, the growing need for H I V treatment among this population. Given the challenges in delivering HIV treatment to IDU, it is clear that this strategy must also incorporate greater HIV prevention efforts, and during my PhD training, I have considered these issues extensively.36-40 Second, I believe that vigilance is required to monitor the delivery of antiretrovirals to ensure that all patients access care in accordance with the most recent guidelines. Specifically, efforts must be focussed on increasing the awareness, implementation, and monitoring of therapeutic guidelines when treatment strategies are updated to ensure that physicians treating HIV positive patients deliver care in accordance with the most recent evidence-based approaches. Third, if the findings presented in Chapter 8 can be independently confirmed, I believe that international therapeutic guidelines for the treatment of HIV disease should be revised. Specifically, the data presented here suggest that therapy can be safely delayed until a CD4 cell count of 200 cells/mm 3 , and that patient non-adherence may be the strongest determinant of patient survival, instead of when antiretroviral therapy is initiated prior to this time. This research provides an important warning about the increased healthcare resources that will be required to address increased morbidity among HIV-infected IDU, and should help to demonstrate that strategies are required to improve antiretroviral delivery to IDU. 214 10.4 F U T U R E R E S E A R C H This thesis points towards several areas for future research. Given the findings of Chapter 4, it will be critical that surveillance efforts continue to evaluate the proportion of persons who have died without accessing therapy. As part of this surveillance, there should be research to evaluate those strategies that are put in place to improve access and adherence for IDU. For instance, it will be interesting to see if prescribed heroin is associated with improved access to and retention in HIV treatment.^l In addition, further study of programs that seek to improve access to healthcare by reducing socio-cultural and geographic barriers, such as safer injecting facilities, also deserve additional study with regards to HIV testing, and referral to HIV treatment.36,38 To ensure guidelines reflect the available evidence, it will be critical that studies evaluating the optimal time to initiate H A A R T are ongoing and updated as additional follow-up becomes available. For instance, while baseline plasma viral load was not statistically associated with mortality in Chapter 8, the role of baseline HIV R N A in survival after the initiation of H A A R T may become more relevant with additional follow-up. It will also be necessary to evaluate if differences in survival emerge between patients initiating therapy at different baseline CD4 cell counts with additional events and follow-up. Although not presented as part of this thesis, during my PhD training I investigated the relationship between CD4 cell count and plasma HIV R N A ^ O / l l and in addition to the ongoing evaluations described above, my post-doctoral training proposes to conduct additional evaluations of the role of baseline CD4 cell count and HIV R N A with regards to virological response to therapy. 2 1 5 In addition, in Chapter 8 physician experience was identified as predictor of survival after the initiation of H A A R T . For the purposes of this thesis, this variable was primarily included as an adjustment factor and to further argue that a baseline CD4 cell count of 350-200 cells/mm 3 does not represent an irreversible biological threshold. In this study, we argued that the association was likely due to the fact that physicians who have treated < 6 HIV infected patients are significantly less knowledgeable about prescription of prophylaxis against opportunistic infections (O.I.'s), provision of preventative vaccinations, as well as the treatment of acute O.I.'s.^ 2 However, the independent association demonstrated in this study deserves further study and is also presently continuing as part of my post-doctoral training. The above are areas for future research that have emerged from this specific research project. There are, however, many other critical areas for future research, and I would be remiss to not point out some general statistics and areas of research that are in urgent need of intervention. Globally, H I V / A I D S is now the leading cause of disease burden and death. Since the beginning of the epidemic it is estimated that 60 million people have been infected and 20 million have died. Among the 40 million people living with H I V / A I D S , 28 million live in sub-Saharan Africa, 6'million in Asia, 1.5 million in Latin America, 0.5 million in the Caribbean, 1.0 million in eastern Europe, 0.5 million in western Europe, and 1.0 million in North America. Among the 40 million people living with HIV, U N A I D S estimates that fewer than one million people are on antiretroviral therapy. These figures demonstrate that strategies to improve access to antiretrovirals should be among the major public health priorities globally. 2 1 6 10.5 C O N C L U S I O N S What is clear from previous research, is that there is now a great deal of evidence demonstrating that lower income populations, women, ethnic minorities, and illicit drug users are at risk of poor access to antiretroviral therapy and higher rates of treatment discontinuation among living cohorts. This thesis demonstrates that in British Columbia, despite a universal healthcare system, access disparities persist until death, and that women, Aboriginal persons, and persons residing in low-income neighborhoods may be at greatest risk. Although injection drug use could not be evaluated directly among the cohort of persons with HIV who have died, it is noteworthy that these characteristics are associated with having a history of injection drug use among persons with HIV in the province. Furthermore, the research presented here indicates that, if the low level of antiretroviral medication use persists, large increases in AIDS mortality and decreases in life expectancy may be expected in urban areas where HTV prevalence is high among IDU. With regard to efforts to improve access, strategies must also involve greater vigilance and monitoring of compliance with therapeutic guidelines to ensure that all patients access care that is in line with the most recent evidence. For the population that has yet to initiate therapy, data presented here suggest that all patients who have a CD4 cell count below 200 should immediately be offered therapy, but that there is no survival benefit of initiating therapy prior to this time. Finally, to derive the full benefit of therapy when treatment is initiated, data presented here also demonstrate that strategies to optimize adherence and retention in therapy are urgently required. 217 10.6 R E F E R E N C E S 1. Andersen R, Bozzette S, Shapiro M , et al. Access of vulnerable groups to antiretroviral therapy among persons in care for HIV disease in the United States. H C S U S Consortium. HIV Cost and Services Utilization Study. Health Serv Res 2000;35(2):389-416. 2. Moore RD, Stanton D, Gopalan R, Chaisson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994;330(ll):763-8. 3. Anderson K H , Mitchell JM. Differential access in the receipt of antiretroviral drugs for the treatment of AIDS and its implications for survival. Arch Intern Med 2000;160(20):3114-20. 4. Wood E, Sallar A M , Schechter M T , Hogg RS. Social inequalities in male mortality amenable to medical intervention in British Columbia. Soc Sci Med 1999;48(12):1751-8. 5. Shapiro M F , Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States: results from the H I V Cost and Services Utilization Study. /AMA\"1999;281(24):2305-15. 6. Strathdee SA, Palepu A , Cornelisse P G , et al. Barriers to use of free antiretroviral therapy in injection drug users. JAMA 1998;280(6):547-9. 7. Bangsberg DR, Mundy L M , Tulsky JP. Expanding directly observed therapy: tuberculosis to human immunodeficiency virus. Am J Med 2001;110(8):664-6. 218 8. Wood E, Schechter M T , Tyndall M W , Montaner JS, O'Shaughnessy M V , Hogg RS. Antiretroviral medication use among injection drug users: two potential futures. AIDS 2000;14(9):1229-35. 9. H o g g RS, Heath K V , Strathdee SA, Montaner JS, O'Shaughnessy M V , Schechter M T . H I V / A I D S mortality in Canada: evidence of gender, regional and local area differentials. AIDS 1996;10(8):889-94. 10. Wood E, Yip B, Hogg RS, et al. Full suppression of viral load is needed to achieve an optimal CD4 cell count response among patients on triple drug antiretroviral therapy. AIDS 2000;14(13):1955-60. 11. Wood E, Hogg RS, Yip B, et al. \"Discordant\" increases in CD4 cell count relative to plasma viral load in a closely followed cohort of patients initiating antiretroviral therapy. / Acquir Immune Defic Syndr 2002;30(2):159-66. 12. Quinn T C , Wawer MJ, Sewankambo N , et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group [see comments]. New England Journal of Medicine 2000;342(13):921-9. 13. Wood E, Braitstein P, Montaner JS, et al. Extent to which low-level use of antiretroviral treatment could curb the AIDS epidemic in sub-Saharan Africa [see comments]. Lancet 2000;355(9221):2095-100. 14. U N A I D S Press Release. New public/private sector effort to initiated to accesserate acces to H I V / A I D S care adn treatment in developing countries. May, 2000.. 2 1 9 15. Wood E, Montaner JS, Chan K, et al. Socioeconomic status, access to triple therpay, and survival from HIV-disease since 1996. AIDS 2002;16(15):2065-2072. 16. Opravil M , Cone RW, Fischer M , et al. Effects of early antiretroviral treatment on HIV-1 R N A in blood and lymphoid tissue: a randomized trial of double versus triple therapy. Swiss HIV Cohort Study. / Acquir Immune Defic Syndr 2000;23(l):17-25. 17. Rhone SA, Hogg RS, Yip B, et al. Do dual nucleoside regimens have a role in an era of plasma viral load- driven antiretroviral therapy? / Infect Dis 1998;178(3):662-8. 18. Hogg RS, Heath K V , Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA 1998;279(6):450-4. 19. Hogg RS, Yip B, Kully C , et al. Improved survival among HIV-infected patients after initiation of triple-drug antiretroviral regimens. CMAJ 1999;160(5):659-65. 20. Miller V , Phillips A , Rottmann C , et al. Dual resistance to zidovudine and lamivudine in patients treated with zidovudine-lamivudine combination therapy: association with therapy failure. / Infect Dis 1998;177(6):1521-32. 21. Deeks SG. Determinants of virological response to antiretroviral therapy: implications for long-term strategies. Clin Infect Dis 2000;30 Suppl 2(7):S177-84. 2 2 0 22. Sollitto S, Mehlman M , Youngner S, Lederman M M . Should physicians withhold highly active antiretroviral therapies from HIV-AIDS patients who are thought to be poorly adherent to treatment? Aids 2001;15(2):153-9. 23. Bangsberg DR, Hecht F M , Clague H , et al. Provider assessment of adherence to HIV antiretroviral therapy. / Acquir Immune Defic Syndr 2001;26(5):435-42. 24. Miller L G , L iu H , Hays RD, et al. How well do clinicians estimate patients' adherence to combination antiretroviral therapy? / Gen Intern Med 2002;17(1):1-11. 25. Gross R, Bilker WB, Friedman H M , Coyne JC, Strom BL. Provider inaccuracy in assessing adherence and outcomes with newly initiated antiretroviral therapy. AIDS 2002;16(13):1835-7. 26. Wood E, Hogg RS, Yip B, Harrigan RP, O'Shaughnessy M V , Montaner JS. The impact of baseline CD4 T-cell count and adherence on rates of disease progression among HIV-infected patients initiating triple drug antiretroviral therapy. AIDS 2003;in press. 27. Pomerantz RJ. Initiating antiretroviral therapy during H I V infection: confusion and clarity. JAMA 2001;286(20):2597-2599. 28. Gupta SB, Gilbert RL, Brady A R , Livingstone SJ, Evans BG. CD4 cell counts in adults with newly diagnosed HIV infection: results of surveillance in England and Wales, 1990-1998. C D 4 Surveillance Scheme Advisory Group. AIDS 2000;14(7):853-61. 221 29. Hogg RS, Yip B, Chan KJ, et al. Rates of Disease Progression by Baseline CD4 Cell Count and Viral Load After Initiating Triple-Drug Therapy. JAMA 2001;286(20):2568-2577. 30. Wood E, Montaner JS, Chan K, et al. Highly Active Antiretroviral Therapy and the HIV-infected Injection Drug User: The State of Care in a Universal Healthcare System. Under review 2003. 31. Duran S, Solas C , Spire B, et al. 'Do HIV-infected injecting drug users over-report adherence to highly active antiretroviral therapy?' A comparison between patients' self-reports and serum protease inhibitor concentrations in the French Manif 2000 cohort study. AIDS 2001;15(8):1075-7. 32. Wood E, Chan K, Montaner JS, et al. The end of the line: has rapid transit contributed to the spatial diffusion of HIV in one of Canada's largest metropolitan areas? [In Process Citation]. Soc Sci Med 2000;51(5):741-8. 33. Bangsberg DR, Moss A . When should we delay highly active antiretroviral therapy? / Gen Intern Med 1999;14(7):446-8. 34. Bangsberg D , Tulsky JP, Hecht F M , Moss A R . Protease inhibitors in the homeless. JAMA 1997;278(l):63-5. 35. Wood E, Montaner JS, Tyndall M W , Schechter M T , O'Shaughnessy M V , Hogg RS. Race, gender, socio-economic status, and access to H I V treatment among persons who have died in the era of antiretroviral therapy. Under review 2003. 222 36. Wood E, Tyndall M W , Spittal P M , et al. Unsafe injection practices in a cohort of injection drug users in Vancouver: could safer injecting rooms help? CMA] 2001;165(4):405-10. 37. Wood E, Tyndall M W , Spittal P M , et al. Factors associated with persistent high-risk syringe sharing in the presence of an established needle exchange programme. AIDS 2002;16(6):941-3. 38. Wood E, Kerr T, Spittal P M , et al. The potential public health and community impacts of safer injecting facilities: Evidence from a cohort of injection drug users. JAIDS (In Press) 2003. 39. Wood E, Tyndall M W , Spittal P, et al. Needle exchange and difficulty with needle access during an ongoing HIV epidemic. International J Drug Policy 2002;13(2):95-102. 40. Wood E, Tyndall M W , Spittal P M , et al. Supply-side drug control policy in the face of the AIDS and overdose epidemics: investigation of a massive heroin seizure. Can Med Assoc J 2002;In Press. 41. Brissette S. Medical prescription of heroin\u00E2\u0080\u0094a review. Can HIV AIDS Policy Law Rev 2001;6(1-2):1,92-8. 42. Montaner JS, Hogg RS, Heath K V , et al. Heterogeneity of physician agreement with recommended therapeutic guidelines for the management of HIV-associated disease. Antivir Ther 1996;l(3):157-66. 223 APPENDIX 1: STATEMENT OF AUTHORSHIP APPENDIX 2: CERTIFICATE OF ETHICAL APPROVAL "@en . "Thesis/Dissertation"@en . "2003-05"@en . "10.14288/1.0091210"@en . "eng"@en . "Health Care and Epidemiology"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "The impact of access and adherance on mortality from HIV disease in the era of modern antiretroviral therapy"@en . "Text"@en . "http://hdl.handle.net/2429/14923"@en .