Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The prediction of adverse maternal outcomes in pre-eclampsia Devarakonda, Rajashree M. 2005

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0196.pdf [ 9.66MB ]
Metadata
JSON: 831-1.0092000.json
JSON-LD: 831-1.0092000-ld.json
RDF/XML (Pretty): 831-1.0092000-rdf.xml
RDF/JSON: 831-1.0092000-rdf.json
Turtle: 831-1.0092000-turtle.txt
N-Triples: 831-1.0092000-rdf-ntriples.txt
Original Record: 831-1.0092000-source.json
Full Text
831-1.0092000-fulltext.txt
Citation
831-1.0092000.ris

Full Text

The prediction of adverse maternal outcomes in pre-eclampsia By Rajashree M. Devarakonda M . B . B . S , Karnatak University, India, 1994 C C R P , Society O f Cl inical Research Associates, U S A , 2003 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Reproductive and Developmental Sciences) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A A pr i l 2005 © Rajashree M . Devarakonda Abstract Background: Pre-eclampsia (PET) continues to contribute to maternal and perinatal morbidity and mortality. Management decisions include an evaluation of maternal risk, which is assisted by expert opinion-based guidelines, while not accounting for gestational age (GA) at diagnosis. We evaluated the feasibility of developing a severity score that can predict adverse maternal outcome. Methods: Design: retrospective chart review of 2 years' PET cases in three tertiary level units. Candidate predictors of adverse maternal outcome: admission gestational age, blood pressure, proteinuria, urine output, SaC>2, seizures, uric acid, creatinine, aspartate transaminase (AST), lactate dehydrogenase, bilirubin, albumin, platelet count, MPV, MPV:platelet ratio, and fibrinogen. Combined adverse maternal outcome: maternal death or one/more of: hepatic failure/haematoma/rupture, Glasgow coma scale <13, stroke, >2 seizures, cortical blindness, positive inotrope support, myocardial infarction, infusion of any 3r d antihypertensive, dialysis, renal transplantation, >50% FIO2 for >lh, intubation, or transfusion of >10 units of blood products. Analyses: Women were classified by having achieved the combined adverse maternal outcome or not. Descriptive, parametric and non-parametric test analysis of the predictor variables was performed. The variables with sample size of >400 were selected and univariable logistic regression analysis was done. Variables with univariable p-value <0.25 were selected for multivariable analysis to determine the factors, which influence pre-eclampsia complications. 11 Results: Among 556 women, 58 women (10.4%) developed the outcome (including one death). The most common outcomes were increased oxygen requirements, 3rd infused antihypertensive, and transfusion of >10 units. G A and fibrinogen were lower and T L C , dipstick protein, bilirubin, creatinine, MPV/platelet ratio and A S T were greater in those who developed the outcome. Multivariable logistic regression revealed that higher admission G A (odds ratio 0.85), higher dipstick proteinuria (OR 1.53), and higher MPV:platelet (OR 291.0) independently predicted the outcome. Discussion: Several promising markers were identified, which need to be substantiated in a large multi centre study. Such predictors included admission G A , dipstick proteinuria and the MPV:platelet ratio. Bilirubin and fibrinogen appeared to be promising. A prospective study is required to develop a clinical prediction model for PET. Key words (MeSH): pre-eclampsia; adverse outcomes; prediction; multivariate analyses m Table of contents A B S T R A C T i i T A B L E O F C O N T E N T S iv T A B L E L I S T v i F I G U R E L I S T v i i A B B R E V I A T I O N S ix D E D I C A T I O N x A C K N O W L E D G M E N T S x i 1 B A C K G R O U N D 1 1.1 P R E - E C L A M P S I A / E C L A M P S I A : 1 1.1.1 Pre-eclampsia: 1 1.1.2 Treatment: 6 1.1.3 Pre-eclampsia - current classification: 7 1.2 A S S E S S M E N T OF BURDEN OF ILLNESS: 9 1.2.1 Severity of disease: 9 1.2.2 Diagnosis-based measures: 9 1.2.3 Therapy-based measures: 9 1.2.4 Physiology-based measurements: 9 1.3 R E C E N T DEVELOPMENTS IN OUTCOME ASSESSMENT: 10 1.3.1 Model development: U 1.3.2 Model validation: H 1.3.3 Uses of scoring systems: 13 1.4 PREDICTION MODELS IN OTHER CONDITIONS: 14 1.4.1 Scoring system in critically ill patients: 14 1.4.2 Neonatal scoring systems: 18 1.5 D R A W B A C K S OF THE CURRENT SCORING SYSTEMS FOR PRE-ECLAMPSIA POPULATION: 21 1.6 PREDICTION M O D E L FOR PRE-ECLAMPSIA: 2 2 1.6.1 The problems with the past and present classification of pre-eclampsia: 22 1.6.2 Predictors yet unexamined in the literature: 23 1.6.3 Reasons for prediction and prognostication in pre-eclampsia: 24 1.6.4 Outcome measures in prognosis and prediction of pre-eclampsia: ...25 1.7 G O A L S : 25 1.8 OBJECTIVES: 2 6 1.9 HYPOTHESIS: 2 6 1.10 R A T I O N A L E : 2 7 2 I N T R O D U C T I O N T O P R E D I C T I O N A N D O U T C O M E O F P R E -E C L A M P S I A : 28 2.1 PREDICTION OF P R E - E C L A M P S I A : 28 2.1.1 The predictors of outcome in pre-eclampsia: 31 2.2 O U T C O M E MEASURES IN PRE-ECLAMPSIA: . . 3 1 iv 2.2.1 Significant adverse maternal outcomes: 32 2.3 LITERATURE REVIEW: 3 2 2.4 D E L P H I C CONSENSUS: 3 4 2.5 PREDICTOR V A R I A B L E S : 3 4 2.5.1 Maternal variables: 37 2.6 O U T C O M E V A R I A B L E S : 59 2.6.1 Mortality: 60 2.6.2 Hepatic system: 60 2.6.3 Central nervous system: 61 2.6.4 Renal system: 63 2.6.5 Haematological system: 64 2.6.6 Respiratory system: 66 2.6.7 Cardiovascular system: 67 3 M A T E R I A L S A N D M E T H O D S : 69 3.1 S T U D Y POPULATION : 69 3.1.1 Source population and referral patterns: 69 3.1.2 Sources of data: 70 3.1.3 Inclusion Criteria: 70 3.1.4 Exclusion criteria: 71 3.1.5 Sample size calculation: 71 3.2 D A T A C O L L E C T I O N : 72 3.2.1 Database structure: 72 3.2.2 Database design: 76 3.2.3 Data entry and management: 77 3.2.4 Problems encountered during data entry: 78 3.2.5 Data cleaning: 78 3.3 D A T A A N A L Y S I S : 79 3.3.1 Multivariate analyses variable selection: 80 3.4 ETHICS A P P R O V A L : 81 4 R E S U L T S : 82 4.1 DEMOGRAPHICS: 82 4.2 R I S K FACTORS FOR PRE-ECLAMPSIA: 85 4.3 INCIDENCE OF THE COMBINED A D V E R S E M A T E R N A L OUTCOME: 88 4.4 R I S K F A C T O R / V A R I A B L E P R E V A L E N C E A N D STATISTICAL A N A L Y S I S : 9 0 4.5 B R I E F S U M M A R Y OF RESULTS: 105 5 D I S C U S S I O N : 106 5.1 BIASES: 116 5.2 LIMITATIONS OF THE STUDY: 117 5.3 BENEFITS A N D LIMITATIONS OF SCORING S Y S T E M : 119 5.4 LESSONS L E A R N T FROM THIS: 120 5.5 CONCLUSION: 121 5.6 F U T U R E DIRECTIONS: 122 5.6.1 Goals: 122 5.6.2 Primary objectives: 122 5.6.3 Secondary objectives: 123 v 5.6.4 Ongoing calibration: 123 5.6.5 A computer-based scoring system for pre-eclampsia: 124 5.6.6 Application to ongoing research studies: 124 R E F E R E N C E S : 127 A P P E N D I X 1 - GUIDELINES FOR D A T A COLLECTION 151 A P P E N D L X 2- PREDICTION OF A D V E R S E M A T E R N A L O U T C O M E IN PRE-ECLAMPSIA PAPER PUBLISHED IN JOURNAL OF OBSTETRICS A N D G Y N A E C O L O G Y C A N A D A , 2004. 163 A 2.1 PR E F A C E 163 A 2.2 T H E PREDICTION OF A D V E R S E M A T E R N A L OUTCOMES IN PRE-ECLAMPSIA. .. 163 vi Table list Table 1: Pre-eclampsia - current classification^ 1) 8 Table 2: International Hypertension definitions 5 Table 3: Candidate maternal predictor variables 36 Table 4: Combined adverse maternal outcome (primary outcome) 59 Table 5: Feasibility of study in the source population 69 Table 6: Demographics 84 Table 7: Demographics (Cont'd) 85 Table 8: Risk factors vs. Outcome 87 Table 9: Incidence of outcome by centre 88 Table 10: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for the day of admission 91 Table 11: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for the day of delivery 92 Table 12: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for postnatal day 1 100 Table 13: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for postnatal day 2 101 Table 14: Descriptive Statistics from logistic regression 104 Table 15: Multivariable logistic regression result 104 Vll Figure list Figure 1: The pathogenesis of pre-eclampsia 5 Figure 2: Access database study entry page 73 Figure 3: Access database maternal outcome page 74 Figure 4: Access daily data form or laboratory form 75 Figure 5: Incidence of outcome according to centres 89 Figure 6: Gestational age at admission compared by the occurrence or absence of the combined adverse maternal outcome 94 Figure 7: Gestational age at diagnosis compared by the occurrence or absence of the combined adverse maternal outcome 94 Figure 8: Dipstick protein compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission 95 Figure 9: Mean arterial pressure compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission 95 Figure 10: MPV/Platelets compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission 96 Figure 11: Dipstick protein compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 96 Figure 12: Mean arterial pressure compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 97 Figure 13: MPV/Platelets compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 98 Figure 14: Bilirubin compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 98 Figure 15: Creatinine compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 99 Figure 16: Fibrinogen compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery 99 Figure 17: Fibrinogen compared by the occurrence or absence of the combined adverse maternal outcome on postnatal day 1 103 Figure 18: Bilirubin compared by the occurrence or absence of the combined adverse maternal outcome on postnatal day 1 103 Vlll Abbreviations A P A C H E Development of Acute Physiology A n d chronic Health Evaluation score A P S Acute physiology score A R D S Acute respiratory distress syndrome A S S H P Australasian Society for the Study of Hypertension in Pregnancy A S T Aspartate transaminase A T N Acute tubular necrosis C H S Canadian Hypertension Society C Q I Continuous quality improvement dBP Diastolic blood pressure D I C Disseminated intravascular coagulation FI0 2- Fraction of inspired oxygen G A Gestational age G C S Glasgow coma scale G H Gestational hypertension H E L L P Haemolysis, elevated liver enzyme and low platelets Htn Hypertension I C U Intensive care unit I U G R Intrauterine growth restriction L D H Lactate dehydrogenase M I Myocardial infarction M O D S Multiple organ dysfunction syndrome M P V Mean platelet volume N H B P E P National High Blood Pressure Education Program O R Odds ratio P E T Pre-eclampsia P B L s Peripheral blood leukocytes P : C Protein: creatinine ratio P G Eicosanoids (includes prostaglandins) P I C U Paediatric I C U mortality risk assessment P R I S M Paediatric risk of mortality score PSI Physiologic stability index R O C Receiver-operator curve R O S Reactive oxygen species sBP Systolic blood pressure SIRS Systemic inflammatory response syndrome S N A P - II Score for neonatal acute physiology S N A P P E - II Score for neonatal acute physiology with perinatal extension Sa0 2 Oxygen saturation level T L C Total leukocyte count WBC White blood cell ix Dedication At the Lotus feet of my Lord Bhagawan Sri Sathya Sai Baba Acknowledgments I would like to take this opportunity to thank my supervisor, Dr. Peter von Dadelszen, for his able guidance, advice, and immense support throughout my program. He deserves much recognition for ensuring that I made it through the various stages from outline to the completion of my thesis and for believing that I could make a contribution in pre-eclampsia research. I am very grateful to Dr. Laura Magee for her excellent advises, for her time and resources made available to me, for her thoughtful, prompt and detailed comments on my ideas and thesis drafts. I would like to thank my thesis committee chair, and my supervisor Dr. Rurak for being there for me in spite of his busy schedule, for chairing a smooth thesis defence and for his valuable advice. I am grateful to Dr. Kei th Walley, thesis committee member, for his excellent advice, intelligent comments, and valuable input throughout. I would like to thank Dr. Redman and the staff at Silver Star Unit, from Nuffield department of Obstetrics and Gynaecology, Oxford, England, for providing us with the excellent data and support for our analysis. I would like to thank Dr. Andree Gruslin, University of Ottawa, for providing the data for analysis. M y appreciation goes to the department graduate secretary Mrs . Roshni Nair, for her able coordination, support, and for her encouragement throughout the program. I would like to thank the research coordinators Dr. Vesna Popovska and Shelley Soanes, for helping me with the data entry and would like to thank the statisticians Ruihua Y i n , Laurie Ainsworth, Monica Fernandez, and Boris Kuzelsvic for helping me with the analysis. I would like to thank the Health records staff at B C Women's and Children's Hospital for allowing us to borrow the chart in a timely manner. I am most grateful to my beloved husband, my wonderful parents' in-law, my revered parents, my brother and sister-in-law. I would like to thank my husband for teaching me and for helping me learn various intricacies of queries, and computer skills. Thank you for supporting me in every way, without your caring help and moral support this work would not have been possible. I would like to thank my mother-in-law, in special, for listening to me patiently during the various stages of my work and helping me in my thesis. I would like to thank my father, mother (I am ever grateful to you even though you are in India), and my father in law for all the encouragements, loving words, and the immense support in completing my thesis. xi 1 Background 1.1 Pre-eclampsia/eclampsia: The syndrome of pre-eclampsia/eclampsia remains one of the two most common causes of maternal mortality in the developed world(l;2).Pre-eclampsia complicates 3-5% of pregnancies(l), and is commonly defined as hypertension of >140/90mmHg and proteinuria >0.3g/24h that appear after 20 weeks' gestation, and regress after pregnancy(3). The severity of the disease varies widely. In the mildest form the signs and symptoms appear near term and resolve after birth, with no lasting effects on either the mother or the child. In its severest form the signs and symptoms often occur in the second or early third trimesters. If these signs and symptoms cannot be controlled then the only option is delivery, with consequent iatrogenic foetal prematurity. Because of the additional risks associated with 'early-onset' disease, pre-eclampsia and hypertensive diseases of pregnancy are leading causes of maternal mortality and morbidity and contribute significantly to premature deliveries(4). 1.1.1 Pre-eclampsia: Although pre-eclampsia has been studied extensively, considerable controversy exists over the aetiology of pre-eclampsia. Genetic, immunological, environmental, and vascular mediated factors are all hypothesised to play an important role in the development of this disease(5-16). Although the cause of pre-eclampsia is unknown, the accumulated evidence strongly implicates the placenta(14). These various available hypotheses include genetic, immunologic, and biochemical pathways causing faulty placentation and leading to a final common pathway through endothelial cell activation and dysfunction(6). Anatomical examination of the placenta shows that the affected portion is the foetal-maternal interface of the placenta. Cytotrophoblast 1 invasion of the uterus is shallow with improper invasion of the spiral arteries(9). Most cytotrophoblasts remain at some distance from these vessels. Where endovascular cytotrophoblasts are detected their invasion is limited to the portion of the vessel that spans the superficial decidua. When cytotrophoblasts gain access to the lumen they fail to destroy the myometrium of the arteries and they plug in the lumen as individual rounded cells, suggesting that they are poorly anchored to the vessel wall(10). Because of these alterations in endovascular invasion, the maternal vessels of women with pre-eclampsia do not undergo the complete spectrum of physiologic changes that normally occur (e.g. loss of endothelial lining and musculoelastic tissue); the mean external diameter of the myometrial vessels is less than half that of similar vessels from uncomplicated pregnancies(l 1;12). In addition not as many vessels show evidence o f cytotrophoblast invasion(l 1; 12). In normal pregnancy cytotrophoblast invade large bore arterioles, where they are in contact with well-oxygenated maternal blood. However, in pre-eclampsia the cytotrophoblasts are relatively hypoxic. Also , in pre-eclampsia, there is inadequate conversion of cytotrophoblasts receptors to endothelial receptor phenotypes(13). The results summarised above raise the interesting possibility that the failure of pre-eclamptic cytotrophoblasts to express vascular-type adhesion molecules, a normal cytotrophoblasts do, impairs their ability to form connections with the uterine vessels. This failure ultimately limits the supply of maternal blood to the placenta and foetus causing placental ischemia and in turn placental hypoxia. The failure of pre-eclampsia cytotrophoblasts to make a transition to a vascular cell adhesion phenotype might be part of a broader spectrum defect in which the cells 2 fail to function properly as endothelium with failure to maintain vascular integrity at the maternal-foetal interface. Another important consideration is that the local placental abnormalities eventually translate in to maternal systemic defects. Thus most investigators believe that the causative agents are widely distributed in the maternal systemic circulation. Endothelial dysfunction in pre-eclampsia may result from a variety of factors, including physical shear forces, hypoxia, or reactive oxygen products or their metabolites and other circulating constituents(17). Increased endovascular shear force perhaps caused by vasospasm or failure of spiral artery remodelling is known to affect endothelial cell morphology and function(18). Hypoxia, a plausible result of reduced placental perfusion, is a known stimulator of endothelin and vascular endothelial growth factor synthesis and secretion in other vascular beds(19). Acute atherosis was first demonstrated by Zeek and associates at placental site spiral arteries(20). The similarity of the latter to atherosclerosis led investigators to propose that endothelial injury might be a mechanism responsible for diffuse vascular disease manifest in patients with pre-eclampsia^). This endothelial injury along with other factors leads to vasospasm in turn leading to hypoxia, which in turn leads to increased secretion of vascular endothelial growth factor, endothelin-1, platelet derived growth factor, fibronectin, selectins, and other molecules that influence vascular tone and remodelling(21). In pre-eclampsia this can provoke a vicious cycle of vasospasm, micro thrombosis, disruption of vascular integrity leading to maternal syndrome of 3 pre-eclampsia and affects all the major organ systems of the body. Serious physiologic disturbances persist until the inciting factor is eliminated. So, there is a generalised systemic maternal inflammatory response in pre-eclampsia(22;23). Also it was shown in number of studies that the pro-inflammatory cytokines such as tumour necrosis factor-alfa (TNF-a)(7;24;25), circulating interleukin-6 (IL-6)(7;26) and interleukin-8 (IL-8)(27) were elevated in pre-eclampsia ; these are consistent with the inflammatory response. It is known that the normal pregnancy itself is a systemic inflammation(23), however it is milder. So, it can be said, pre-eclampsia is not a separate condition but a syndrome with extreme maternal systemic inflammatory response. There are two syndromes in pre-eclampsia: one maternal, the other foetal. It may be that the maternal syndrome of pre-eclampsia is a final common pathway with many alternative routes to its inception (Figure 1). 4 cytotrophoblast invasion immunological factors -poor placentation thrombophilia multiple pregnancy fetal macrosomia hypertension glomerular cndotheliosis/ proteinuria/ ATN liver damage/ hematoma/-rupture microangiopathic "hemolysis/ thrombocytopenia/ DIC edema Figure 1: The pathogenesis of pre-eclampsia. In this model of pre-eclampsia, the maternal syndrome develops from a number of alternative pathways leading to uteroplacental mismatch, whereby the fetoplacental demands outstrip the maternal circulatory supply. In response to the mismatch, and probably due in part to recurrent ischemia-reperfusion injury within the intervillous (maternal blood) space of the placenta and accelerated placental apoptosis, a soup of endothelium-damaging substrates is released with resulting endothelial cell activation and consequent development of the maternal syndrome of pre-eclampsia. Some elements of the soup, namely activated peripheral blood leukocytes can cause direct end-organ damage. There is cross-talk between elements of the soup (not illustrated). ARDS: acute respiratory distress syndrome; ATN: acute tubular necrosis; DIC: disseminated intravascular coagulation; PBLs: peripheral blood leukocytes; PGs: eicosanoids; ROS: reactive oxygen species. (From von Dadelszen et al, Crit Care Med 2002; 30(8): 1883-1892). The maternal syndrome of pre-eclampsia is most commonly caused by abnormal placentation(5), which becomes manifest once the foetal nutritional demands outstrip placental supply. The resultant uteroplacental mismatch(28), where the demands of the pregnancy outstrip the capacity of the maternal arterial supply, may also arise due to excessive foetal demands, as occur in multiple pregnancy(29)or foetal overgrowth(28) by the loss of functioning placental mass in the setting of thrombophilia(30). The intervillous space of the mismatched placenta releases factor(s) into the maternal circulation ('intervillous soup'(28)) 5 leading to endothelial dysfunction and microangiopathic haemolysis(6), inflammatory mediator release(7), and neutrophil activation (22;31-33). What results is the phenotypic organ dysfunction of the clinical syndrome. Although hypertension is the most common manifestation, the maternal syndrome of pre-eclampsia is considerably more than pregnancy-induced hypertension(34). In fact, in the era of adequate blood pressure control(35), pre-eclampsia-associated mortality is most commonly due to either hepatic necrosis or the acute respiratory distress syndrome, both of which are the consequences of systemic inflammation(3 6). The foetal syndrome is manifested by intrauterine growth restriction (IUGR), foetal acidaemia, and increased risk for both perinatal morbidity and mortality, particularly due to the risks of prematurity. 1.1.2 Treatment: The definitive treatment for pre-eclampsia is delivery, which is always the treatment of choice for the mother. However, this represents a double-edged sword for the foetus at gestational ages remote from term (i.e. <32-34 weeks'), at which time perinatal morbidity and mortality are substantial(37;38). It has been shown that, under these circumstances, prolongation of pregnancy by expectant therapy decreases serious perinatal morbidity without an increase in the maternal risk(35;39). However, these trials had insufficient statistical power to detect a difference in serious maternal outcomes between groups, and uncertainty about the magnitude of the maternal risk(40) has made some clinicians reluctant to use such management. 6 Therefore, to enhance the care of women with pre-eclampsia, we need to identify the women at greatest risk so that we can balance the needs of the mother in terms of burden of illness with those of the foetus remote from term with immature lungs and low birth weight, who is at increased risk of developing significant morbidity i f (s)he survives, as each week gained in utero confers tangible benefit in perinatal outcome(37;38). 1.1.3 Pre-eclampsia - current classification: Most recently, guidelines for the diagnosis and management of pre-eclampsia have been produced by the Canadian Hypertension Society(41), the US National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy(42), and the Australasian Society for the Study of Hypertension in Pregnancy (ASSHP)(43). The American and Australasian documents have been largely merged by the International Society for the Study of Hypertension in Pregnancy (ISSHP)(44). To summarize, mild pre-eclampsia is classified as a blood pressure (BP) >140/90mmHg with proteinuria of 0.3-3g/day(44). Severe pre-eclampsia is defined as mild pre-eclampsia with a single additional 'adverse feature' such as BP>160-170/100-110 mmHg, proteinuria of >3-5g/day, and/or headache(44). (See Table 1) While dichotomizing pre-eclampsia in this way presumably differentiates women with lower risk from those with higher risk, there are no shades of grey. 7 Table 1: Pre-eclampsia - current classification(41) Definition/ criteria CHS ASSHP N H B P E P Hypertension (mmHg) dBP >90 sBP >140 and/or dBP >90 BP >140/90 Severe htn (mmHg) dBP >110 sBP >170 and/or dBP >110 BP >160/110 Korotkoff IV (clinical) V (research) V V Proteinuria (g/d) >0.3 >0.3 or spot urine P:C >30mg/mmol >0.3 ASSHP: Australasian Society for the Study of Hypertension in Pregnancy; CHS: Canadian Hypertension Society; dBP: diastolic blood pressure; htn: hypertension; NHBPEP: National High Blood Pressure Education Program; P:C: protein : creatinine; sBP: systolic blood pressure This approach provides a blunt instrument that does not quantify risk over a broad range of clinical situations, and neglects other features of potential importance. A l l classifications (Table 2) are predicted on the simultaneous occurrence of hypertension and proteinuria, neither of which is present in 10% o f women within one week prior to their first eclamptic seizure(45). Table 2: International Hypertension definitions CHS ASSHP N H B P E P Pre-existing htn essential/20 Chronic htn essential/2° Chronic htn G H w/out proteinuria + adv features G H Transient htn G H with proteinuria + adv features Pre-eclampsia mild/ severe Pre-eclampsia/ eclampsia Pre-existing htn + superimposed G H with proteinuria Pre-eclampsia superimposed on chronic htn Pre-eclampsia superimposed on chronic htn ASSHP: Australasian Society for the Study of Hypertension in Pregnancy; CHS: Canadian Hypertension Society; htn: hypertension; NHBPEP: National High Blood Pressure Education Program; GH: Gestational Hypertension; Adv: Adverse 8 1.2 Assessment of burden of illness: 1.2.1 Severity of disease: Illness severity is an instinctive, self-evident concept; the challenge has been how to execute its measurement. Illness severity itself is a risk for several outcomes(46). There are few general approaches to measurement of severity; diagnosis, therapy and physiology. 1.2.2 Diagnosis-based measures: It is easy to recognise that a patient with diagnosis of pre-eclampsia is more ill than a patient without diagnosis of the same(46). However the approach overlooks the broad spectrum of severity within the international guidelines as well as in Canadian, Australasian and American guidelines for pre-eclampsia. Furthermore there is enormous variation in coding among various centres. 1.2.3 Therapy-based measures: An alternative way to measure the severity is to count the number and types of therapies and type of management whether the patient was managed expectantly or aggressively(46). The flaw in this is evident; clinicians with more interventionist style may manage pre-eclampsia in an aggressive fashion and hence those patients will be scored as more sick or severe pre-eclampsia. Paradoxically, those patients in whom such treatments were not obligatory, but rather used at clinician discretion, would have substantially better outcomes, and the interventionist clinicians would appear to have better overall results. 1.2.4 Physiology-based measurements: The rationale is that, regardless of disease or diagnosis, derangements from physiologic norm increase the likelihood of adverse outcome, and that the greater the derangements, the greater 9 the risk(46). This rationale obtains for each organ system, and the composite severity can be represented by the weighted sum of derangements across all organ systems. APACHE(47-49) did not work well in the setting of eclampsia(50). This, and other similar, scoring system are not based on the gestational reference ranges, and the outcome against which these variables are tested is mortality which is rare in pre-eclampsia(50). 1.3 Recent developments in outcome assessment: Since the olden days, scientists and clinicians have experimented with different forms of measurement. The intent is to aid interaction and transmission of ideas and, also, to ensure objectiveness, when compared with descriptive terminologies. In addition, measurement may help in the prediction of a certain outcome. Clinicians have approached the derivation of reliable measurements in a more scientific way; which led to the science of outcome analysis. From the time when the Apgar score(51) was introduced, numerous scores and scales have been developed in medicine. A score is an attempt to integrate many different pieces of information, such as blood pressure, heart rate or laboratory data, into a single one-dimensional numerical value. The more complex a clinical situation is, the more efforts are made to develop and refine score systems. A s medical costs rocket, the perspective has shifted towards other aspects, such as cost-benefit analysis, severity indices and consistency for clinical trials. There are many scoring systems that have been developed and evaluated, but only a few have been wel l validated and, therefore, are used more commonly than others. The scoring systems 10 previously developed are for health status evaluation(52;53), severity of illness(48;49;54;55), sepsis(55), assessment of injury severity(56), patients-care load(57), etc. Their main uses were/are for quality of care assessment(52;53), for prognostic evaluation(48;49;54;55;58-60), for counselling, for comparisons of patients in clinical trials(61), for development of clinical practice guidelines(62) and to plan health policies(62). Also in outcomes analysis, scores may act as the building blocks of continuous quality improvement (CQI)(52;53;61). CQI identifies faulty processes that produce poor outcomes and corrects those processes (52;53;61). 1.3.1 Model development: The development of a scoring system includes the identification of the main objective(s) and the areas of interest. The focus can be objective outcomes, subjective evaluations or generic health status measurements. Consequent steps involve the type of measurement of multi-item indexes, followed by range development for item generation and item diminution. Such indexes are supposed to be easily obtainable, such as laboratory biochemical markers. Also, some of the model development includes parameters like blood pressure, heart rate and other vital investigations in the model. Statistics helps in removing the difficult factors in the scoring system model. 1.3.2 Model validation: The model developed will then be further evaluated and validated for sensibility(46) (reasonable knowledge of pathophysiology and clinical reality of the focus), reliability(46)(same result is obtained when the same phenomenon is measured repeatedly by the same person), validity (46)(measure represents what is being sought) and responsiveness or sensitivity(46)(ability to measure clinical change). 11 The model then tests the data that has been adjusted for risk adjustment, quality and processing, and timing of assessments(46). Without risk adjustment, outcomes are uninterpretable. Contributing factors often included are demographic factors, patient's social and financial condition, health risk factors, and admitting and co-morbid diagnoses. Baseline health status is strongly associated with outcome measures. Probable recording errors may come from the area where the data was generated and quality of the data(46). This starts from the point at which the index parameter is measured, to the stage of completion, through the process of interpretation and statistical analysis and finally to conclusions and applications. In the recording of diagnosis fields, the data may be biased(46), while charting the parameters; data may be mischarted or missing(46). Non-random, systematic sampling error also occurs in settings with incomplete ascertainment. Some studies have tried to decrease this error by increasing the number of studied patients and also by doing the pilot retrospective study followed by a prospective study. The length of time of measurement is also very important(46). The follow up period and how long and how frequently the variables of interest are assessed for the outcome are also very important. Further, to assess and develop the quality of the care given, it is important to evaluate outcomes with attribution to process. Hence, in addition to longer-range follow-up, assessment across more discrete time frames is desirable for some purposes(46). 12 The focal point of a scoring system is the reduction and summary of assorted clinical data. The cutback of information can help to concentrate on the important things, but this decrease naturally causes a reduction in the amount of available information. This shows that scores w i l l not substitute the original data but provide a perspective allowing interpretation. The validation process o f a scoring system requires the use of same elements and analytical approaches as were used for the derivation (development) process. In a similar way, results have to be interpreted using the same objectives and same methodology, which allows accurate analysis. A score that was developed for a geriatric population to determine mortality risk derived from a study using a group of critically i l l patients in a multi-disciplinary unit cannot be assumed to be functional on a group of post-operative adolescent orthopaedic patients. The scoring system cannot be applied to a single patient who is from a different population (biased population), in a different time frame. Even i f we can compare the patients, the transfer of a score into a different setting may cause factual error. Calculations cannot be made with 100% confidence, especially important in morbidity or mortality prediction(48;49;54;55;58-60). Most scoring systems are more helpful in quality of care assessment and risk prediction compared to predicting mortality. 1.3.3 Uses of scoring systems: Scoring systems help in so many ways. Many study outcomes can be presented more easily in a more objective and reproducible way. Scores can be used to perform quality assessments. The main effect is on the identification of the need for diagnostic or therapeutic procedures. This may help in triaging (that is, finding the most appropriate disposition for a patient based 13 on an assessment of the patient's illness and its urgency), initiation of procedures, discharge and execution of therapy. The scientific evaluation and the realistic use of scoring systems will unquestionably continue to improve. Scores can aid comparisons by means of classification of diseases or health condition, improves decision-making, thus quality management, and helps improve clinical research. The usefulness of scores in the individual patient may be limited due to the reality that scores reduce the magnitude of a clinical situation into a single value by adding a number of variables. Therefore, the score will supplement and add a new contour to the clinical judgment and knowledge of the clinician/scientist. 1.4 Prediction models in other conditions: 1.4.1 Scoring system in critically ill patients: As high numbers of patients coming to I C U satisfy the criteria for systemic inflammatory response syndrome (SIRS) (that is systemic response to infection or trauma, with symptoms including fever, tachycardia, tachypnoea, and leukocytosis). We looked at some of these illness severity-scoring systems in our population by doing literature review. 1.4.1.1 Development of Acute Physiology And chronic Health Evaluation score (APACHE): 14 The basis for development of A P A C H E was the hypothesis, that the severity of acute disease can be measured by quantifying the degree of abnormality of multiple physiological variables. This approach was used to detect and treat life-threatening acute physiological derangements. The A P A C H E and other similar severity classification systems must be based on objective physiological measurements and be as independent of therapy as possible. A P A C H E 11(48) used a point score based upon initial values of 12 routine physiologic measurements (as compared to 34 in A P A C H E I (54)), age, and previous health status to provide a general measure of severity of disease. A n increasing score (range 0 to 71) was closely correlated with the subsequent risk of hospital death for I C U admissions. The recorded value is still based on the most deranged reading during each patient's initial 24h in an I C U . Knaus and associates(47;54)used a variation of the nominal group process to choose and weight physiologic variables. This was followed in regard to proper construction of severity scales, and took advantage o f the long-established principle of homeostasis. The A P A C H E II was developed based on 12 physiologic measures and each variable was given weight. The weighting system is based a scale of 0-4, as determined by derangement from the normal range, the sum of which yields an acute physiology score (APS) . When A P A C H E II scores were combined with an accurate description of disease, to prognostically stratify acutely i l l patients. The validity of these variables were tested against hospital mortality. Increasing in A P S was shown to be associated with increased risk of subsequent hospital death. Hence, A P A C H E was used to evaluate the use of hospital resources 15 and compare the efficacy of intensive care in different hospitals and or over time. A P A C H E was shown to be reliable and useful in classifying I C U patients. However, A P A C H E was developed for elderly patients under I C U setting and was developed to test the subsequent risk of hospital death(50). A s the too much of weight is given for age, it weighs heavily for elderly population. Also , it is well known that mortality due to pre-eclampsia has decreased considerably compared to olden days and hence it is difficult to test A P A C H E for the patients with pre-eclampsia(50). Also gestational variation in the laboratory values are not included in the classification(50). A P A C H E when tested by Bhagawanjee and associates for eclampsia population, did not prove to be predictive of mortality(50). Glasgow coma scale (part of A P A C H E scoring system variable) however, was independently predictive of mortality when used alone. In A P A C H E , age is weighed heavily. The women with eclampsia were young. Also the physiological variables are altered during gestation which is not taken in to account may be one other reason for its failure to predict mortality in the setting of eclampsia. 1.4.1.2 Development of Multiple Organ Dysfunction Syndrome (MODS): Multiple Organ Dysfunction Syndrome is the leading cause of morbidity and mortality for patients admitted to an ICU(55). The hallmark of the M O D S is the development of progressive physiologic dysfunction in two or more organ systems after an acute threat to systemic homeostasis(55). The objective of the M O D S score was to develop and evaluate a scoring system to quantify the severity of the multiple organ dysfunction syndromes as an outcome in critical illness. 16 M O D S score, constructed using simple physiological measures o f dysfunction in six organ systems, mirrors organ systems as the intensivist sees them and correlates strongly with the ultimate risk of I C U mortality and hospital mortality(55). M O D S was developed to accurately reflect the clinical phenomenon, and to provide a reliable and meaningful index of the severity of the syndrome in the individual I C U patient. In M O D S , variables from a review of the multiple organ failure literature were evaluated for construct and content validity to identify the optimal descriptors of organ dysfunction. Clinical and laboratory data were collected daily to evaluate the performance of these variables individually and in aggregate as an organ dysfunction score. Descriptors meeting criteria for construct validity (is an assessment of how well you translated your ideas or theories into actual programs or measures) and content validity (is based on the extent to which a measurement reflects the specific intended domain of content) were identified for five of the seven systems and in the absence of an adequate descriptor, they developed a new variable. From the first half of the database (the development set), intervals for the most abnormal value of each variable were constructed on a scale from 0-4, so that a value of 0 essentially means normal function and was associated with an I C U mortality rate of <5%, where as a value of 4 represents marked functional derangement and I C U mortality rate of >50%. These intervals were then tested on the second half of the data set (the validation set). Maximal scores for each variable were summed to yield a multiple organ dysfunction score. This score correlated in a graded fashion with the I C U mortality rate, both when applied on the first day of the I C U admission as a prognostic indicator and when calculated as an outcome measure. The score 17 showed excellent discrimination, as reflected in area under receiver operator curve (ROC) in the validation set. The incremental increase in scores over the course of the I C U stay calculated as the difference between maximal scores and those scores obtained on the first day also demonstrated a strong correlation with I C U mortality rate. Again M O D S is not based on gestational reference ranges and is based on and developed and relevant to the I C U setting. So, what is needed is an objective scoring system based on the well-developed scoring systems as there systems are well established which includes the gestational change of pregnancy, which also takes in to consideration the gestational age. 1.4.2 Neonatal scoring systems: 1.4.2.1 Development of Paediatric RISk Mortality (PRISM) score: The Paediatric Risk Mortality (PRISM) score was developed from the Physiologic Stability Index (PSI) to reduce the number of physiological variables required for paediatric I C U (PICU) mortality risk assessment and to obtain an objective weighting of the remaining variables(59). Physiologic Stability Index (PSI) is a paediatric severity of illness measure that accurately assesses mortality risk(63). Data collection pertaining to P R I S M score analysis included I C U outcome (survival, death), admission day PSI scores, diagnoses categorised by the primary physiologic system of dysfunction, and demographic information. The PSI is a subjective score developed from 7 different organ systems based on the 34 physiologic variables. The most abnormal value of each variable during the admission day is coded into 75 pre-assigned variable ranges that reflect the clinical importance of the derangement. Derangements are assigned 1 point i f the abnormality is worthy o f concern but not necessarily 18 a change in therapy, 3 points i f the abnormality is sufficient to cause a change in therapy, and 5 points i f the abnormality is life threatening. Three physiologic variables, systolic B P , heart rate, and respiratory rate were adjusted for age. Mortality predictions calculated from the PSI score included a 4-day average score(63) and an organ system weighted score calculated from the admission day data. The PRISM(59) score was developed in a systematic manner. First the database was split into an estimation set and a validation set. The estimation set was used for model derivation and the validation set used for model verification. The variables, which were not significant in predicting death, were excluded by univariate analysis. The variables in PSI scores with less significance were removed by logistic regression. A final logistic regression analysis of outcome versus the sum of the P R I S M scores, operative status, and age determined the coefficients of each of these variants to be used in the outcome prediction. The P R I S M score is an objective simplification of the PSI. The number of physiological variables was reduced from 34 to 14 and the number of ranges was reduced from 75 to 23. The physiological variable ranges have been objectively re-weighted to directly reflect their contribution to mortality risk. Thus P R I S M score is not only easier to compute but more directly reflects severity of illness than does PSI. Again P R I S M score is used and developed for paediatric I C U population and hence we can only use them as a standard for development of a new severity of illness classification for pre-eclampsia, which includes gestational age and gestational reference range. 19 1.4.2.2 Development of Score for Neonatal Acute Physiology (SNAP)- II) and Score for Neonatal Acute Physiology with Perinatal Extension (SNAPPE)- II: S N A P - II and S N A P P E - II are simplified newborn illness severity and mortality risk scores to test the primary outcome of in-hospital mortality. S N A P , the first generation newborn illness score was cumbersome to use. Therefore, S N A P II and S N A P P E II were developed to make the score more simple, reliable and empirically weighted. In the derivation of S N A P II, cross tabulation for each o f the 34 variables from S N A P vs. mortality was performed and the variable having no association with mortality were eliminated. Within each variable adjacent risk categories were consolidated i f they were not significantly different. Additional variables were eliminated because of their infrequency and unreliability. Multivariate derivation model of illness severity was based on the strategy described by Pollack et al(58) for the P R I S M score. Mortality served as a dependent variable and all the potential physiologic predictors were tested in this model(60). Variables were said to be significant based on the significance of their coefficients. When the multiple predictors were collinear, variables were selected on their predictive power and ease of data collection finally leaving 6 variables. This was called as S N A P II and includes only physiological variables and therefore is a physiologic illness severity score. This in turn was extended to include a number of perinatal risk factors such as gestational age, birth weight, sex, white race, multiple birth, and size for gestational age, and Apgar scores. However, later on, the only significant perinatal risk factors that were selected were birth weight o f 750-999g, birth weight of <750g, and small for gestational age (<3 r d centile) and Apgar of <7. These perinatal risk factors are denominated in the same risk points as the physiologic risks. This data was validated on 3 different 20 independent cohorts and predictive performance was calculated. The discrimination of death by S N A P P E II was excellent with R O C area ranging from 0.84-0.92. However, even though S N A P E II is predictive of mortality, this is used and developed for paediatric I C U population to predict the mortality which is rare in pre-eclampsia. Hence we can only use them as a standard for development of a new severity of illness classification for pre-eclampsia, which includes gestational age and gestational reference range and which can predict the incidence and development of adverse outcome as well as mortality in pre-eclampsia. 1.5 Draw backs of the current scoring systems for pre-eclampsia population: Severity of illness scoring systems have been developed for patients critically i l l patients; the APACHE(47-49 ) , MODS(55) , and Brussels scores(64). These illness severity scores perform well when modified for defined populations(47-49), but most have been generated to reflect pathophysiology in predominately geriatric populations(50) and the A P A C H E score, specifically, does not perform well in the setting of eclampsia(50) Also these scoring systems have been developed and used in the different population, and different age groups. They do not take in to account the gestational reference range and also they don't take in to account the gestation age of the pregnant women. 21 1.6 Prediction model for pre-eclampsia: To improve clinical management, and to investigate the assertion that pre-eclampsia resembles SIRS, we require a scoring system that can predict the burden of illness that w i l l be suffered by the mother and the risk associated with the prolongation of pregnancy for both mother and the baby. Such a scoring system should be based on a number of important factors. First, the score should reflect gestational reference ranges for the biochemical and Haematological variables being studied. Second, the score should develop well-characterized scores for systemic inflammation (e.g. A P A C H E ( 5 4 ) , MODS(55) , and Brussels(64)). Third, the score should identify the burden of illness carried by the mother, that evaluates different organ dysfunctions systematically, and that can be correlated with basic science to identify the pathogenesis of pre-eclampsia. The scoring system should be prognostic, objective, syndrome-specific and reliable. Objective risk estimates are particularly important in the high-cost, emotional, and technologically demanding environments of obstetric units. In summary, a pre-eclampsia score should be syndrome-specific, taking into account both normal pregnancy changes in physiology (cardiovascular, renal, immunological, and haematological) and pre-eclampsia-induced perturbations of those physiological adaptations. This proposal is the first step of an international initiative, currently underway, to develop such a score. 1.6.1 The problems with the past and present classification of pre-eclampsia: In clinical practice, the diagnosis of pre-eclampsia requires exclusion (by renal, hepatic, and haematological investigation) when either non-proteinuric gestational hypertension present in 22 20% of women within a week of their first eclamptic seizure(45) or non-hypertensive gestational proteinuria (present in 10% of women(45)) arise. Also , in the current classification, gestational age at presentation is not a criterion for either diagnosis or severity. So, there is a need for new classification /scoring system which includes gestation age at onset of pre-eclampsia and also which includes the patients without gestation high blood pressure but, with proteinuric pre-eclampsia and also patients with gestational hypertension without proteinuria. 1.6.2 Predictors yet unexamined in the literature: That gestational age has not been accounted for in any of the current classification systems is a major problem(65). It is the most important clinical variable in predicting both maternal and perinatal outcomes. Early-onset pre-eclampsia represents considerable additional maternal risk, as maternal mortality is some 20-fold higher at <32 weeks' gestation than when pre-eclampsia occurs at term(66). That this risk is due to more severe underlying disease is supported by the pathophysiology of early-onset pre-eclampsia which differs from late-onset disease, in terms of neutrophil function(37) and cytokine levels(7;37;67). For the advancement of knowledge, this group of women may provide the most homogeneous data for differentiating the changes of pre-eclampsia from those of normal pregnancy. That perinatal morbidity and mortality are gestational age-dependent is a given, as, among diploid foetuses, gestational age is the most important determinant of perinatal outcome(35;68). A greater than 50% chance of intact survival for a foetus delivered of a 23 woman with pre-eclampsia arises only when the gestational age at delivery is >27 weeks and the birth weight >600g(37) 1.6.3 Reasons for prediction and prognostication in pre-eclampsia: The fundamental goal behind reaching a prognosis or prediction of outcome is to improve patient care, or more generally to improve the future quality o f life of the patient or child. A prognosis represents information that w i l l be used to improve decision making about how to treat a patient(69). While this is the most frequent reason for prognostication, other reasons exist. In addition, it is not known how objective markers of foetal well being, as they reflect uteroplacental pathology, may have an impact on the ability to predict adverse maternal outcomes. Such markers of foetal well-being include ultrasound estimates of foetal weight, growth velocity, amniotic fluid volume, and umbilical arterial blood flow, as wel l as assessing foetal heart rate patterns using cardiotocography ('non-stress test')(70). Conversely, the pattern of maternal disease has not been included in any predictive models of foetal/perinatal outcomes. The expectant management of pre-eclampsia at gestational ages remote from term is predicted on the intention to maximise perinatal intact survival(35). Finding maternal markers of disease activity that would assist in deciding when to deliver the foetus for foetal indications would have great clinical utility. 24 Therefore, to enhance the care of women with pre-eclampsia, we need to identify the women at greatest risk so that we can balance the needs of the mother in terms of burden of illness with those of the foetus remote from term with immature lungs and low birth weight, who is at increased risk of developing significant morbidity i f (s)he survives, as each week gained in utero confers tangible benefit in perinatal outcome(37) 1.6.4 Outcome measures in prognosis and prediction of pre-eclampsia: What needs to be considered are methods to further sub classify the various maternal presentations of pre-eclampsia. For example, pre-eclampsia could be sub classified into 'mild, moderate, severe, and extreme' disease on the basis of organ dysfunctions; into 'early- and late-onset' disease; and into disease due to 'placental, thrombophilic or other identifiable aetiologies'. A system of sub classification could help address questions surrounding the management o f pre-eclampsia with respect to expectant versus aggressive management of pre-eclampsia remote from term by balancing the needs of mother and the foetus, and predicting the maternal risks for comparison with quantifiable perinatal risks of prematurity. A burden of illness score that exemplifies the systemic nature of pre-eclampsia would also be useful both clinically and for clinical and basic science research in the field. 1.7 Goals: The goal of this proposal is to conduct a feasibility study by retrospective chart review for the development of a scoring system to identify maternal risk in pre-eclampsia that w i l l be functional, practical and influence, in a positive direction, the outcomes for women and their babies. 25 1.8 Objectives: • Identify the items/variables that have the potential for predicting adverse outcomes in women with gestational hypertension and/or gestational proteinuria, the HELLP(71 ) syndrome, or eclampsia without hypertension or proteinuria at all gestational ages. • Identify outcomes in terms of maternal and foetal mortality and morbidity from literature review, calculate the incidence of the outcome which determines the data collection needed in the participating centres, and define the outcomes that are included in the study. • Develop an Access™ database for the data entry and collection of data from the health records from the respective centres based on ICD-9 code for pre-eclampsia. • Develop exclusion and inclusion criteria. • Test maternal variables against combined maternal outcome. • We also wished to determine the feasibility of a prospective study using these variables. Areas of difficulty w i l l be identified and steps w i l l be taken to improve the data acquisition in the prospective study. 1.9 Hypothesis: The hypothesis for our study is that it is possible to develop a development set of severity score using retrospective data. 26 1.10 Rationale: To improve clinical management, and to investigate the assertion that pre-eclampsia resembles the systemic inflammatory response syndrome, or SIRS, we wish to develop a objective scoring system, based upon physiological changes of pregnancy, that describes the burden of illness suffered by the mother at presentation with the disease and the risk associated with pregnancy prolongation for both mother and baby. It is of note that established scores used for SIRS, such as A P A C H E , were not useful in pre-eclampsia. The hope is that this new score w i l l be useful in clinical practice, basic science to characterise women included in studies, and clinical research to characterise the burden of maternal illness at presentation and/or as an outcome measure. 27 2 Introduction to prediction and outcome of pre-eclampsia: 2.1 Prediction of Pre-eclampsia: Pre-eclampsia is a common, yet incompletely understood, complication of pregnancy. Its clinical features, including hypertension, proteinuria, and varying degrees of ischaemic end-organ damage, are caused by widespread endothelial dysfunction(6). The aetiology of endothelial dysfunction in pre-eclampsia is not known, but it has been postulated to be part of an exaggerated maternal inflammatory response to pregnancy(22) The onset of pre-eclampsia at or near to term is associated with low maternal and neonatal morbidity and mortality(66). In contrast, those patients (1%) who suffer early onset pre-eclampsia engender significant maternal and perinatal morbidity and mortality(66). Therefore, because of the lack of proven prophylaxis for pre-eclampsia, prediction of risks or identification of sub-clinical disease is desirable to identify patients for more intensive observations and better management. There are certain at risk groups of patients such as those with chronic hypertension (72-74), hypertension with renal disease(75), pre-gestational diabetes(74;76), multifoetal gestation(77), a previous history o f pre-eclampsia(78), and/or a family history of pre-eclampsia(15;78). These women account for the majority of cases of pre-eclampsia in multiparas(74), and only account for 14% of pre-eclampsia in nulliparous women(74). 28 Thus, the majority of cases of pre-eclampsia arise from nulliparous women without medical complications at low risks. Differences in the time of onset, severity and organ system involvement suggest there may be differing underlying aetiologies that ultimately lead to pre-eclampsia manifested as maternal hypertension and proteinuria. Distinct markers therefore may identify sub groups o f at-risk women with separate underlying causes. Based on data from patients with recognized disease, with the association of various organ systems, potential candidate markers would include cardiovascular, hepatic, renal, haematopoietic, and central nervous and coagulation systems. Although cross-sectional studies have identified some of the potential markers, they need to be evaluated in prospective longitudinal studies, with rigorous definition of outcome, to determine i f they are useful in predicting the outcome of pre-eclampsia and whether or not they can identify different subgroups with in the population of women with pre-eclampsia that w i l l develop different outcomes. Despite intensive research to elucidate the origin of pre-eclampsia, there is currently no well-validated prophylactic treatment, nor is there an effective method of identifying women at risk for adverse outcome in pre-eclampsia. Therefore, currently, the utility of predicting risk for a particular outcome lies in being able to assess that risk objectively. Prediction of risk associated with pre-eclampsia in terms of the adverse maternal and perinatal outcome w i l l identify patients for more careful monitoring but may also identify a population that is highly suited for research into the 29 aaetiology and pathophysiology of pre-eclampsia and into potential prophylaxis and /or disease-modifying treatment to prevent adverse maternal and foetal outcomes. Although previous clinical history and epidemiological factors can identify individuals or population at increased risks for pre-eclampsia, their specificity and sensitivity are not high. Therefore, the aim is to develop a scoring system which can detect burden o f illness that w i l l be suffered by the mother, and, as a secondary question, the baby, and which can predict objectively whether a particular complication w i l l occur in the woman with pre-eclampsia. The scoring system should also identify sub clinical disease before the development of severe maternal manifestations to permit the institution of, as yet unidentified, interventions that fall short of having to resort to the delivery of the patient. A major stumbling block to identifying predictors of outcome in pre-eclampsia is the relative dearth of well-performed, large-scale prospective longitudinal studies throughout gestation in individuals who develop pre-eclampsia. Most data have been gathered in cross sectional studies with small sample size and poor definition of patient groups and diagnosis of the "disease" itself. Although differences in concentration of several markers are evident in the presence of the established disease, there are very few data available to demonstrate their utility as predictors of outcome in pre-eclampsia. 30 2.1.1 The predictors of outcome in pre-eclampsia: The rationale for selection of the variables is that they must be predictive, ought to vary with changes in severity, and must be available, measurable, frequent, accurately recorded and reliable. In addition, the variables need to be reliable. The larger the numbers of score items, the more robust the score, both for detecting smaller changes and in the face of sparse data. Sampling each item across each organ system also ensures validity of the system. Measurement of specific markers for the involvement of these each of these systems expect to indicate the organ involvement at the appearance of the full maternal syndrome, but also supposed to show that there are several different causes and presentation leading to pre-eclampsia. The presence of specific markers may ultimately be used for diagnosis of disease as an alternative to the syndrome of hypertension and proteinuria. The purpose of this study is to highlight the potential predictors for the development of the complications of pre-eclampsia, based on the involvement of different organ systems within the overall disease process, which may have utility in assessing both the at risk population for adverse maternal and foetal outcome. 2.2 Outcome measures in pre-eclampsia: The two major analytical steps in developing burden of illness scoring are (1) the collection of an appropriate database and (2) analysis to establish a final system design. This requires that the 31 variables collected are tested against a predetermined endpoint. In this case, it was decided to test the variables against their ability to predict either maternal death or significant adverse maternal outcomes. 2.2.1 Significant adverse maternal outcomes: Maternal mortality has been used as a measure of the success of obstetric intervention but is now too rare for use in local practice in the developed world(79). Severe maternal morbidity has been suggested as an alternative measure(80;81). Most have been retrospective studies(80-82), with only few prospective studies(80). Most of the prospective studies have counted admissions to intensive care (80;82) and investigated only the characteristics of women receiving obstetric intensive care(80). 2.3 Literature review: First, a formal literature review was undertaken to determine how pre-eclampsia is classified, managed, predicted and how the complications were characterized in previous studies. Special focus was placed on the CHS 1997 guidelines(41). Also, we reviewed the current outcome prediction models used in categorizing patients with the systemic inflammatory response syndrome (SIRS)(47;54;55), with which pre-eclampsia shares many characteristics(22;36). This process defined the organ systems to be included in the score, the specific variables used to quantify dysfunction within each system, and the outcomes to be predicted. 32 A better understanding o f the maternal vascular and inflammatory disease in pre-eclampsia and its resemblance to other diseases of endothelial dysfunction, and the literature review helped us to identify the organ systems that should be incorporated in such a level and the exact variables that have been used to measure dysfunction within each system, and ensured the construct validity of the final score. For each organ system considered, the most descriptive variables were selected (as had been identified in the derivation of other disease severity models). We searched Medline using the following key words severe maternal morbidity, obstetric intensive care, obstetric haemorrhage, uterine rupture, obstetric sepsis, H E L L P , eclampsia and maternal mortality. We selected definitions that were clinically based and routinely measurable. When no definition specific to the condition was available we modified the standard definition to take in to account the physiological changes of pregnancy. We focused on morbidity associated specifically with pre-eclampsia and its complications. We excluded those conditions (placental abruption, DIC , pulmonary emboli) that are difficult to diagnose accurately or ascertain completely. We identified deterioration of function in a number of organs and systems presumably as a consequence of vasospasm and endothelial dysfunction. For descriptive purposes these effects were separated in to maternal and foetal outcome. However these outcomes were usually encountered simultaneously and also one or more organ complications could have occurred together. 33 2.4 Delphic consensus: In addition to reviewing more than 100 books and a number of articles, we solicited the opinions of scientists and experts in field of hypertension in pregnancy internationally. Fourteen international experts from North America, the U K , and Australasia, using the Delphic method (83) by e-mail survey, were asked to review the candidate predictor variables, the combined adverse maternal outcome, and to suggest omissions. The experts were asked to assess the clinical usefulness of 20 different patient attributes representing various organ systems in predicting the adverse maternal outcome in women with pre-eclampsia. There were 4 possible responses: l=not useful, 2=slightly useful, 3=moderately useful and 4=very useful. The average response for each attribute (range: 1.5 to 2.9) was calculated and the attributes were arranged according to the organ system. The fundamental assumption in this approach is that the clinical experts are able on average to identify correctly the predictive value of the candidate predictors. For simplicity analysis of mortality, hepatic, central nervous system, cardiovascular, renal, respiratory and Haematological systems were considered systematically. Although there are many possible maternal consequences of severe pre-eclampsia, we selected some major complications/outcomes. Those not already identified by the literature review were added to the list. 2.5 Predictor variables: These variables have been chosen as they fulfilled the criteria required by Richardson et al(46), in that they were measurable, frequently obtained, accurately recorded, and available. Because of the 34 requirement for measurability we excluded maternal symptoms from the list o f candidate variables even though they were included in some national guidelines defining disease severity(41). From the literature and expert list o f potential predictors by Delphic consensus, final variables were selected to ensure that each organ system is represented. Potentially, in addition to the patient's personal and family history, we wished to test maternal variables from seven different organ systems along with the gestational age at presentation with pre-eclampsia. Table 3 summarises the maternal variables from all the organ systems. 35 Table 3: Candidate maternal predictor variables. Organ system Variable(s) Gestational age on admission (admission during which delivery occurred) Gestational age on diagnosis of pre-eclampsia Maternal Cardiovascular Systolic BP Diastolic BP Renal Albuminuria: 24h urine Dipstick proteinuria Urine output Uric acid Creatinine Hepatic Aspartate transaminase (AST) Lactate dehydrogenase (LDH) Bilirubin Albumin Respiratory FI0 2 :Sa0 2 Haematological Platelet count Mean platelet volume (MPV) MPV:platelet count ratio Fibrinogen Central Nervous Seizures 2.5.1 Maternal variables: 2.5.1.1 Gestation: 2.5.1.1.1 Gestational age at diagnosis of pre-eclampsia: Pre-eclampsia, as stated, is an immensely variable and unpredictable condition, which obeys no rules and rarely follows a set pattern. It usually becomes apparent at any time after 20 wks of pregnancy(3), but cases of pre-eclampsia have been known to occur even earlier(34). Although the disease should be well on the decline within a day or two after delivery, women have died due to eclampsia as late as five days after delivery. Just as the time of onset varies, so does the speed with which the disease progresses. So, when a woman begins to get pre-eclampsia determines to a large extent how severe her case w i l l be. Pre-eclampsia arising around full term is normally an insignificant complication, and most cases fall in to this category. But i f the disease begins in the second trimester the disease establishes itself and often deteriorates in a feed-forward and accelerating manner, until the baby is delivered. Gestational age at onset of pre-eclampsia is the most important clinical variable in predicting both maternal and perinatal outcomes(84). Early-onset pre-eclampsia represents considerable additional maternal risk, as maternal mortality is some 20-fold higher at <32 weeks' gestation than when pre-eclampsia occurs at term(84). That this risk is due to more severe underlying disease is supported by the pathophysiology of early-onset pre-eclampsia which differs from late-onset disease, in terms of neutrophil dysfunction(22;32;33) and cytokine levels(67). For the advancement of knowledge, 37 this group of women may provide the most homogeneous data for differentiating the changes of pre-eclampsia from those of normal pregnancy. That perinatal morbidity and mortality are gestational age-dependent is an acknowledged fact, as, among diploid foetuses, gestational age is the most important determinant of perinatal outcome(35). Greater than 50% chance of intact survival for a foetus delivered o f a woman with pre-eclampsia arises only when the gestational age at delivery is >27 + 0 weeks' and the birth weight >600g(37). Hence gestational age at diagnosis of pre-eclampsia is included as an important variable in our study. 2.5.1.2 Hepatic system: The liver damage in pre-eclampsia tends to be a fairly late feature of the disease and is hard to detect, except through blood tests, which can reveal the leakage of the contents of damaged liver into the bloodstream. The causes of the liver damage are not known but the most likely explanation is that the sick placenta causes general damage to the circulation, thus provoking abnormal clotting and inflammation depriving the liver as explained earlier along with other essential organs, o f their full blood supply. Liver damage is usually linked with involvement of the clotting system, and in extreme cases, a combined breakdown of the liver and clotting system can ensue(85). Liver dysfunction is a feature of pre-eclampsia detected by elevations of circulating hepatic enzymes(85), which may progress to jaundice and hepatic impairment(86-88). Unless evidence for 38 hepatic derangement is sought in all cases of proteinuric pre-eclampsia, dangerous manifestations w i l l be missed. About two thirds of women dying from eclampsia have specific lesions in the liver (89) which are periportal ' lake' haemorrhages and various grades of ischaemic damage, including complete infarction(89). The haemorrhages arise from the arteries and arterioles of the portal tract, which show diffuse mural damage. Liver involvement in pre-eclampsia-eclampsia is grave and is frequently accompanied by evidence of other organ involvement, especially in kidney and brain, along with haemolysis and thrombocytopenia(90-92). Weinstein (71) named this the Haemolysis, Elevated Liver enzymes, and L o w Platelets ( H E L L P ) syndrome. N o reliable studies on hepatic blood flow are available, but clinical and pathological observations suggest that reduced perfusion may be an underlying cause in cases of liver dysfunction that are known to occur in pre-eclampsia (86). 2.5.1.2.1 AST (aspartate transaminase): A S T is a sensitive indicator of cellular injury. A S T is present in the heart, skeletal muscle, brain, and kidney as well as in liver(71). However A S T was selected over alanine transaminase ( A L T ) which is more specific for liver, as the levels of A S T remain stable during the course of gestation, compared with A L T , which fluctuates during the course of gestation. A S T is relatively non-specific. However high levels suggest hepatocellular injury(71). Reduced blood flow may result in hepatocellular necrosis, oedema and ischemia that stretch the Gibson capsule causing epigastric pain or right upper quadrant pain and this pain is usually associated with elevated transaminases. Periportal haemorrhagic necrosis in the periphery of the liver lobule is the most likely reason for 39 increased serum liver enzymes(93;94). Elevated levels of A S T suggest hepatic involvement and increasing levels suggest worsening severity(95;96). In a study done by Catanzarite et al it appeared that fulminant and extreme elevation of A S T was associated with high risk of maternal mortality(97). Martin et al showed that A S T elevation was significantly associated with high risk for development of maternal morbidity when tested in women with severe pre-eclampsia with and without H E L L P syndrome(98). 2.5.1.2.2 LDH (lactate dehydrogenase): L D H , commonly included in routine analysis, is insensitive as an indicator of hepatocellular injury, but is better as a marker of haemolysis, myocardial infarction, or pulmonary embolism(98). However, in Catanzarite study(97) it appeared that fulminant and extreme elevations of A S T and L D H were associated with high risk of maternal mortality. Elevated L D H levels were associated with haemolysis and hepatic involvement. This may reflect severity and may predict the potential for recovery postpartum especially in women with H E L L P syndrome(95;96). A s with A S T , L D H elevation has been significantly associated with high risk for development of maternal morbidity when tested in women with severe pre-eclampsia with and without H E L L P syndrome(96) 40 2.5.1.2.3 Bilirubin: Serum bilirubin may not be a particularly sensitive index of liver dysfunction or disease prognosis in pre-eclampsia, but it is an established test in other disease severity models(48;54;55;99).Total bilirubin should be normally less than 1.2mg/dl. Jaundice, a common sign o f liver problems, does not always occur in pre-eclampsia, but there are always exceptions and some women develop jaundice as the disease progresses. Even though bilirubin is not tested often in the screening or prediction of pre-eclampsia, we have included them in he study to test the predictability of this variable for maternal and perinatal morbidity based on A P A C H E ( 4 7 ) which was developed for the systemic inflammatory response syndrome (SIRS) patients in an I C U setting. 2.5.1.2.4 Albumin: Serum albumin is a main determinant of plasma oncotic pressure. Its serum concentration is determined by the relative rates of its synthesis and degradation or loss, by its distribution between the intra-and extra vascular beds and by the plasma volume. The normal adult liver synthesizes 10-15g o f albumin every day. In pregnancy, the albumin level is slightly lower than in non-pregnancy (100). The excessively decreased serum albumin may result from decreased blood flow to the l iver( lOl) , hepatic necrosis(lOl), chronic inflammation and also from excess loss of protein from the kidney due to renal diseases(102-105), among other causes. A s shown below, it can also result from abnormality in the glomerulus, which allows albumin to escape into the urine(102-105). The excess proteinuria is generally followed by the decrease in the level of plasma albumin. 41 2.5.1.3 Renal system in pre-eclampsia: The kidney may be a culprit or victim in the onset of hypertension during pregnancy(106;107). Successful outcome for the mother and the baby requires careful assessment of renal function in women with pre-eclampsia. Pre-eclampsia is characterized by endothelial impairment and acute atherosis with multisystem involvement(108). The renal glomeruli are frequently affected, showing a typical pathological picture of glomerular endotheliosis(109). Renal biopsy has shown a characteristic non-inflammatory lesion, primarily of swelling of the glomerular endothelial cells, which encroach on and occlude the capillary lumina, showing glomerular endotheliosis. 2.5.1.3.1 Albuminuria (urinary protein): Proteinuria during pregnancy is consistently defined in the literature as the excretion of proteins in the urine in excess of 300mg in 24 hours (0.3g/d)(l 10-112) The consequent clinical manifestation is the appearance o f proteinuria, one of three fundamental criteria for the clinical diagnosis of pre-eclampsia. Heavy proteinuria (greater or equal to 5 gm/24 hours) has traditionally served as one criterion in the definition of severe disease(44). A s with other types of proteinuria of glomerular origin, the proteinuria of pre-eclampsia involves predominantly high molecular weight proteins such as albumin. It is caused by reversible structural alterations of the glomerular filter resulting from injury of endothelial cells in the glomerular capillaries. There are different views in the development of proteinuria in pre-eclampsia. Proteinuria is one of the several signs of involvement of the renal glomerulus in pre-eclampsia and is a conventionally recognised, but late, sign of renal involvement in pre-eclampsia(102;103). Once present it indicates a poorer prognosis for both 42 mother and baby than when it is absent(102;103). On average it appears about three weeks before intrauterine death or mandatory delivery(104). Over all pre-eclampsia is one of the commonest causes of nephrotic syndrome in pregnancy(105). Kidney function is monitored during pregnancy and pre-eclampsia because departures from the norm can give an early indication of later complications^ 11). 2.5.1.3.1.1 Dipstick Protein: The most basic test is dipstick urine test: this reveals the presence o f protein, which should not be detectable in significant amounts i f the kidneys are functioning normally. Proteinuria is defined as the abnormal presence of protein in the urine. Normally a small amount of protein is present in the ultra filtrate produced by the glomerulus, but much of this protein is absorbed by the tubules (and some additional proteins are secreted into the urine). Ultimately, very little protein is present in the urine that leaves the kidney. Proteinuria is often measured using a dipstick assay. In this assay, a reagent reacts with albumin producing a colour change. The dipstick is reported on a semi-quantitative scale: negative, trace (10-20 mg/dL), 1+ (30 mg/dL), 2+ (lOOmg/dL), 3+ (300 mg/dL), 4+ (1000-2000 mg/dL). O f note, the dipstick test for proteinuria suffers from both false positive errors. False negative tests are often seen in dilute urine (specific Dipstick test for proteinuria has significant false-positive and false negative rates(110;113;114). Results should be interpreted taking into account p H , specific gravity, bacterial or leukocyte count, haemoglobin concentration and red blood cell contamination(110;113;114). I f the dipstick result is positive (> +1), a 24-hour urine collection to determine protein loss is needed to confirm 43 proteinuria. However 24 hour urinary protein is not available in all cases. Negative dipstick results do not rule out proteinuria, especially i f diastolic pressure is greater than 90 mmHg( l 10;113;114). 2.5.1.3.1.2 24 hour urinary protein: A s dipstick protein test performs poorly in terms of high false positivity and false sensitivity( 110;113;114), the 24-hour urine collection remains the most reliable method of measurement, and is the gold standard. It is stated that i f proteinuria is in excess of 2g/d, very close monitoring is warranted, and that i f it is in excess of 3g/d, delivery should be considered(113;115-117). 2.5.1.3.2 Uric acid: In pre-eclampsia renal function is impaired(l 18). The changes are biphasic involving first tubular function and later glomerular function. A n early feature of pre-eclampsia is reduced uric acid clearance, reflecting altered tubular function(118). This causes a reciprocal rise in plasma uric acid(118). Later at about the time proteinuria develops, glomerular filtration becomes impaired. Therefore, a rising urate is an early marker of pre-eclampsia which precedes late rise in plasma creatinine(104). That serum uric acid is elevated in pre-eclampsia and eclampsia was discovered by Stander et al (119), in 1925. This finding was further confirmed by other authors(120-122). The rise of serum 44 uric acid was found to be characteristic of pre-eclampsia and not of other types of toxaemias of pregnancy and was considered a more sensitive index of the severity of the clinical illness, than was the level of blood pressure, or of degree of proteinuria(121;122). Fadel(123) showed that uric acid was significantly higher in the women with eclampsia than in women with pre-eclampsia. Moreover it appears to be a sensitive index of the severity of pre-eclampsia. That is, degree of hyperuricaemia was associated with severity of pre-eclampsia and a significant negative correlation was associated with maternal plasma uric acid levels and the birth weight of the infants(68;124). Also an elevated uric acid was found to distinguish the hypertension of pre-eclampsia with poor foetal outcome, from chronic hypertension with a normal foetal outcome(68;125). Also , in a study done by Martin et al(96), it was shown that uric acid level of >7.8mg/dl were significantly associated with an increased risk for the development of maternal morbidity. 2.5.1.3.3 Creatinine: Renal function can be monitored by means of blood tests to measure the concentration of waste products such as uric acid and creatinine. Serum creatinine levels drop in normal pregnancy (70). Elevated levels, which suggest increasing severity of hypertension(70), could be within normal limits for non-pregnant women, and tend to occur late in gestational hypertension with proteinuria and are associated with reduction of renal perfusion and glomerular filtration rates. Elevated serum creatinine levels and reduced glomerular filtration rates are associated with severe cases and may represent intrinsic renal changes(106). In most women with pre-eclampsia with proteinuria > 5 gm/24 hours, renal function is reflected by serum creatinine concentration(95). A raised creatinine 45 of above 90uM is usual only in proteinuric pre-eclampsia and it helps in anticipation of the development of acute renal failure in a particular patient. Even though creatinine was the weakest link(98;126) in the prediction of maternal morbidity in pre-eclampsia it was selected in our feasibility study, as a renal marker along with proteinuria and uric acid for the assessment of its inclusion in the final severity score, because of its predictive value in APACHE(47-49;54;98). 2.5.1.3.4 Urinary output: Although most women with pre-eclampsia experience only moderate decrements in renal function, on occasion, dysfunction is severe - and leads to acute renal failure, usually tubular necrosis, and rarely cortical necrosis(106). Pre-eclampsia is often associated with reduced and / or concentrated urine volumes(127) Both the high blood pressure and the reduced systemic perfusion pose a threat to the mother and to the fetus. It is well known that sudden increases of mean arterial pressure above a critical threshold of about 140 mmHg may cause acute vascular damage with loss of auto regulation resulting in cerebral haemorrhage(115;128), impaired perfusion of maternal organs and the placenta, impaired oxygen availability and consumption leading to tissue ischemia(129;130), decreased renal perfusion and oliguria. The quantity of protein excretion and the quantity of urine production are associated with several haemodynamic factors in addition to basic renal function (68;131). 46 2.5.1.4 Haematological and coagulation system: Enhanced coagulability of the blood is a normal feature of pregnancy, facilitated by an increased concentration of several circulating clotting factors and greater reactivity of platelets, the specialized cells responsible for clotting(132;133). In pre-eclampsia platelets may become even more reactive, so that they are used up even faster than normal and their concentration in blood is consequently reduced(132-134). In some cases a persistent fall in the platelet count(90;133-136) or a dramatic reduction in the count can be detected even before proteinuria appears, and in some functioning of platelets is disturbed but their concentration remains unaffected(133). In some of the cases, the disturbances of these platelets is so much it causes destabilization of the clotting system causing disseminated intravascular coagulation (DIC) with inappropriate microvascular events in vital organs leading to damage of these organs, haemorrhage, and haemolysis o f blood cells(90; 132-136). 2.5.1.4.1 Platelet counts: A s mentioned above, platelet counts may play an important role in the prediction of complications and also play an important role in the classification of H E L L P (Haemolysis, Elevated liver enzymes and L o w Platelets) syndrome(71) of pre-eclampsia. L o w levels (<100 000 x 10 9/L) may suggest consumption in the microvasculature(137;138). Levels correspond to severity and are predictive of recovery rate in postpartum period, especially for women in H E L L P syndrome. 47 2.5.1.4.2 MPV: Platelet consumption of pre-eclampsia is a late feature but increases in platelet volume and shorter half-lives of platelets indicating in increased platelet turnover may be a more sensitive indicator. Mean platelet volume reportedly corresponds to severity of hypertensive disorder of pregnaney( 139). A mean platelet volume of >11 fl at 28 wks gestation was found to be associated with high incidence of pre-eclampsia(139;140), but substantial overlap in platelet volume with those of normal pregnant women limits the use of a cross-sectional measurement of M P V alone as a screening tool. 2.5.1.4.3 MPV:Platelet count Longitudinal studies in the normal population studies show little change in M P V throughout gestation with the 90th % ile of change > 0.8fl(140)But, 14/15 individuals in whom pre-eclampsia develop between 24 and 38 wks gestation showed a substantial increase of > 0.8 fl in M P V . Only 3% of normal pregnant individuals showed a change of this magnitude(140). Therefore, as part of this project, we have developed a summary measure of platelet consumption, mean platelet volume ( M P V ) : platelet count ratio. This ratio is designed to better reflect platelet consumption than either element of the ratio alone. A s platelets are consumed, the megakaryocytes in the marrow release immature large volume platelets, causing M P V to rise prior to ongoing consumption overwhelming the ability of the marrow to respond. Therefore, this summary measure 48 should amplify the effects of platelet consumption, although we recognize that ratios are intrinsically unstable than direct observations. A retrospective longitudinal study of 17 women with a previous history of pre-eclampsia showed the utility of measuring platelet volume with increased prediction of development of pre-eclampsia by 2-5 wks(141). We have postulated in this study that longitudinal measurement of MPV:platelet count beginning early from the day of admission for pre-eclampsia might have utility in identifying those women at increased risk of either maternal and/or foetal complications. To develop a robust normal range for this test, further studies need to be repeated prospectively in large populations that include at risk groups. 2.5.1.4.4 Fibrinogen: The coagulation system is a target for the pre-eclampsia process. Normal pregnancy is a hypercoagulable state, and this is exaggerated early in the development of pre-eclampsia^ 1;90; 133; 135). There is considerable evidence that pre-eclampsia is accompanied by a number of coagulopathic changes when compared with the normal pregnant woman. Although a declining platelet count may be an early feature of pre-eclampsia(133) it has limited diagnostic value by itself. Despite this unless pre-eclampsia is complicated by overt clinical DIC , routine coagulation tests are usually normal(142). During normal pregnancy, plasma fibrinogen concentration substantially increases and in pre-eclampsia there is a slight increase in plasma fibrinogen levels compared with normal pregnancy levels(142), and the turnover of radio labelled 49 fibrinogen is increased in women with pre-eclampsia(136). Assessment of fibrinolysis in pre-eclampsia is difficult because of the pre-eclampsia-accentuated coagulability. Due to endothelial cell dysfunction and platelet activation and free radical damage to endothelial cells the clotting cascade begins in advanced pre-eclampsia with thrombosis and fibrin deposition in various organs and blood vessels. This leads to reduction of fibrinogen from the circulation and appearance of fibrin and fibrin split products like monomers and dimers of fibrin in the circulation, which can be measured. So measurement of reduction of fibrinogen from the circulation with accentuation of fibrin split products in the circulation can be done serially to assess the coagulation system in pre-eclampsia. Hence, even though fibrinogen is not used that commonly in the prediction of complications in pre-eclampsia, we included fibrinogen in this study to test the predictability of this variable for maternal and perinatal morbidity based on A P A C H E ( 4 7 ) which was developed for SIRS patients in an I C U setting. A s the international normalized ratio (LNR) is not available on a regular basis in the retrospective database we included this predictor marker in our future prospective study. 2.5.1.5 Respiratory system: Oxygen delivery and consumption are directly related to cardiac output, and recent observations in patients with severe untreated pre-eclampsia indicate that the reduced cardiac output in these patients may result in a state of chronically reduced oxygen availability and utilization, which may ultimately lead to ischaemic tissue damage and organ dysfunction(130;137). 50 2.5.1.5.1 SaC :^ Oxygen saturation Binding sites for oxygen are the haem groups, the Fe 2 + -porphyrin portions of the haemoglobin molecule. There are four haem sites, and hence four oxygen-binding sites, per haemoglobin molecule. Haem sites occupied by oxygen molecules are said to be "saturated" with oxygen. The percentage of all the available haem binding sites saturated with oxygen is the haemoglobin oxygen saturation (in arterial blood, the Sa02). Sa02 alone doesn't reveal how much oxygen is in the blood; for that we also need to know the haemoglobin content. Whatever the Sa02, its value is simply the percentage of total binding sites on arterial haemoglobin that are bound with oxygen, and can never be more than 100%. Sa02 is not affected by either anaemia or haemolysis. It is affected by Pa02, the partial pressure of oxygen in the plasma phase of arterial blood, which in turn affected by alveolar PO2 and lung architecture. Arterial oxygen saturated was included in our study based on A P A C H E (47) even though many of the women with pre-eclampsia might have been missing this variable as it is not done routinely unless the patients are perceived to be very unwell. Even though this is not an ideal test in comparison to pulse oximetry, we included this, as it is simple and easy to perform in any population without causing alarm to the patient. In a sick patient, when Sa02 is measured it points to the problem. Arterial oxygen saturation is above 97% in a normal population. When it starts declining, it points to pulmonary oedema, and/or chronic obstructive lung disease. A s pulmonary 51 oedema is a common complication in pre-eclampsia, as well as an iatrogenic complication of aggressive fluid management, we thought that SaO"2 might act as an important variable in the feasibility study for severity score. 2.5.1.6 Cardiovascular system: The hypertension of pre-eclampsia is an early feature, not associated with a single haemodynamic pattern. Some investigators find increased cardiac output(143;144) however others do not( 145; 145; 146; 146). 2.5.1.6.1 Systolic BP, diastolic BP, MAP: 2.5.1.6.1.1 Gestational changes in blood pressure: There are changes in blood pressure related to gestational age. Most studies suggest a decrease in blood pressure during mid trimester(116;126;138;147-151), and a return to normal value by the third trimester. Diastolic pressure is clearly favoured by the majority of the definitions described by N H B E P W G , ISSHP, A S S H P , although there is evidence to suggest systolic blood pressure has a better relationship to outcome(72;110;113;152-154). Blood pressures above 140/90 mmHg have been associated with an increased perinatal mortality rate (106;114;155;156). B y definition arterial blood pressure is elevated in hypertensive pregnant women, whether they have non-proteinuric gestational hypertension, pre-eclampsia, or chronic hypertension of whatever 52 cause. Also blood pressure behaviour appears to be altered in hypertensive pregnant women, as the circadian blood pressure rhythms may be completely deranged(95) Blood pressure shows much greater variability than in normotensive pregnant women(95;157). True severe pre-eclampsia is associated with a reduced preload, low cardiac output, and elevated after load in the majority of the patients. This would be in agreement with the evidence that total blood volume is significantly reduced in patients with untreated pre-eclampsia compared to that in normotensive pregnancy and in patients with non proteinuric gestational hypertension(96;158). It was observed that there is no correlation between systemic arterial blood pressure and the cardiac index in pre-eclampsia: This may explain the clinical observation that in women with pre-eclampsia the magnitude of the arterial blood pressure is not a good indicator of the severity of the condition. Oxygen delivery and consumption are directly related to cardiac output, not to arterial blood pressure, and recent observations in women with severe untreated pre-eclampsia indicate that the reduced cardiac output in these patients may result in a state of chronically reduced oxygen availability and utilization, which may ultimately lead to ischaemic tissue damage and organ dysfunction(130;137). In a hallmark longitudinal study, Easterling et a l ( l l l ; 1 4 3 ) showed that mean arterial blood pressure ( M A P ) , heart rate and cardiac output were consistently higher throughout pregnancy in women destined to develop pre-eclampsia compared with normotensive women with 53 uncomplicated outcomes. M A P is a better indicator of the cardiac output than either systolic or diastolic blood pressure alone. A s oxygen delivery and consumption are directly related to cardiac output, not to arterial blood pressure, and as arterial blood pressure is not a good indicator of the severity of the condition we selected M A P as the variable to be included in the study. The formula used to calculate mean systemic arterial pressure is: M A P = diastolic pressure +1/3 pulse pressure (systolic - diastolic pressure) 2.5.1.7 Central nervous system: 2.5.1.7.1 Seizures: Pre-eclampsia is so named because it was originally identified as a disorder preceding eclampsia, although it is now known that eclamptic convulsions are only one o f the several potential complications ofthe disease. A s mentioned earlier, 41% of cases have been reported to have little or no proteinuria one week prior to the onset of eclampsia(45;159;160) and only a minority are thrombocytopenic^) . The seizures of eclampsia are thought to be triggered by abnormal neural activity in the brain; when these cells are damaged, as they can be by the pre-eclampsia process (probably from vasospasm, not cerebral oedema as is widely believed), their highly organized activity can become so disturbed that it explodes in to intense electrical discharges, which simulate the convulsive movements. Eclamptic seizures usually occur as a third stage complication of severe-pre-eclampsia(161). But, sometime they appear without any preceding disturbances, although in these cases other signs of pre-eclampsia tend to develop at the same time as seizure. 54 Hence we included the occurrence of a first seizure as a predictor variable and the development of recurrent seizures as an adverse maternal outcome in our study. It is possible some women develop visual disturbances, headache before the development of seizures. Since these are subjective symptoms we have excluded them from our study. 2.5.1.8 Epidemiological and other risk factors for pre-eclampsia: Epidemiological findings indicate that certain characteristics are more common in women who develop pre-eclampsia. Maternal epidemiological characteristics that have been evaluated previously in number of studies include maternal age(162-167), ethnicity(168;169), and past history in terms of smoking(170-173), cocaine use(172;174), obesity(168;175-177), thromboembolic disease(134;135), pre-pregnancy hypertension(73), and early-onset heart disease except mitral valve prolapse (178), pre-pregnancy diabetes(76) parity(168;179;180) and new paternity(181-183). Variables of interest from previous research regarding past obstetric history were history of pre-eclampsia, normotensive IUGR(184), and recurrent spontaneous abortion(184-186). In the family history, family history of pre-eclampsia(15;34;187;188), normotensive IUGR(189), thromboembolic disease(190;191), early-onset heart disease (as above)(192), and chronic hypertension(192) were included. 55 The most important risk factor for pre-eclampsia is nulliparity. Pre-eclampsia is primarily a disease of first pregnancy. A t least two thirds of cases occur in women during the first pregnancy not terminating in a first trimester loss (179). There is also a relationship between the extremes of childbearing age and the incidence of pre-eclampsia(163). Because most pregnancies, particularly the first occur in young women(164;165), most cases o f pre-eclampsia occur in this group. It is observed that in the United States this observation appears to be true. However some of the studies showed an increased incidence of pre-eclampsia in older women independent of parity(166). Other parts o f the world also support the association of higher maternal age association with pre-eclampsia(167). The relationship between pre-eclampsia with race is equally difficult to evaluate. The Jerusalem study revealed a strong correlation of race and age with pre-eclampsia(167). It was shown that the incidence of pre-eclampsia was higher in Black women by Davies et al(167) but only in nulliparous women by Eskenazi et al(168). However, Chesley et al did not find any significant difference between black race and Caucasian population as cited by Davies et al(169). Gestational age at the time of presentation^ 93) and gestational age the time o f diagnosis(193) of pre-eclampsia were included. Patient's treatment history with respect to steroids for foetal lung maturity, steroids for H E L L P syndrome, MgS04 , and antihypertensive therapy were included to describe the type of management of different centres. 56 2.5.1.8.1 Exclusion of symptoms: In pre-eclampsia, especially pre-eclampsia of increasing severity, a multitude of symptoms can arise. Because pre-eclampsia is a disease of poor perfusion to most tissues, the occurrence of symptoms related to many organ systems is not surprising. Symptoms of hepatic capsular distention(194) (e.g. epigastric pain, stomach upset, and pain penetrating to the back) and poor cerebral perfusion(194) (e.g. headache, mental confusion visual symptoms(194) [ranging from scotoma and blurred vision to blindness]), and oedema were not included. These symptoms are not quantifiable and are not objective. The sensitivity to, perception of, and tolerance of pain varies from patient to patient and there is no way of grading the symptoms retrospectively. Hence in our feasibility study we have excluded the symptoms, as the severity score w i l l be derived objectively. These candidate descriptors of organ dysfunction biochemical markers and physiological markers along with gestational age at presentation and diagnosis were evaluated for disease severity of pre-eclampsia. For this set, baseline information was collected, followed by daily data input using the most abnormal value for any 24h period as the value recorded (when multiple values are present in a given 24h period). Missing data points were filled in where possible using the 'last observation carried forward' approach used in the A P A C H E and Brussels scores and applied to clinical trials(195-197), and as is used in day-to-day clinical practice. Data were collected from the day of admission for delivery and until the day of discharge(45). 57 The validity of this study depended upon the selection of the representative group of patients from different centres with different type of management for pre-eclampsia. This would help us to determine the centres that should be included in the prospective study for the validation of the severity score. Undertaking a multi-centre study would meet the first criterion o f tool development, as the tool would be useful having been developed in different populations. 58 2.6 Outcome variables: Table 4 summarizes the combined maternal outcome variables that were included in our study. Table 4: Combined adverse maternal outcome (primary outcome) Organ system Outcome(s) Maternal (death or one/more of) Mortality Hepatic Failure Hematoma Rupture Central Nervous Glasgow coma scale (GCS) <13 Stroke Two or more seizures or status eclampticus Cortical blindness (whether transient or permanent) Cardiovascular Positive inotrope support Myocardial infarction Infusion of any third antihypertensives Renal Dialysis (whether temporary or permanent) Renal transplantation Respiratory Requirement of >50% O2 for >lh, or intubations Haematological Transfusion of >10U of blood products (in total) 59 2.6.1 Mortality: Maternal mortality is used internationally as a measure of the quality of obstetric intervention, although it is now rare in the developed world(79). Maternal mortality in association with pre-eclampsia is due predominantly to complications of abruptio placentae, hepatic rupture(198), and eclampsia. In the study done by Sibai and associates(199) there was a maternal mortality of 1.1 %(5 patients). Three patients died due to severe laryngeal oedema, difficult intubation leading to hypoxemic encephalopathy^ 99). 1 patient died due to A R D S , and 1 died due to pulmonary embolism because o f tricuspid atresia(199). In the study done by Mark Waterstone et al(200) there were 5 maternal deaths directly attributable to severe pre-eclampsia 3 from sepsis, 1 from haemorrhage, and 1 from H E L L P syndrome. 2.6.2 Hepatic system: 2.6.2.1 Hepatic failure, hepatic haematoma or rupture: Rationale: In pre-eclampsia epigastric pain or right upper quadrant pain most likely results from hepato cellular necrosis, oedema, and ischemia that stretches the Glisson capsule (hepatic capsule) and is frequently associated with rise in liver enzymes. The rise in liver enzyme is most likely due to periportal hemorrhagic necrosis(93;94). Bleeding from these necrotic lesions may cause hepatic rupture or they may extend to form a sub capsular hematoma(201). The sub capsular haemorrhage can be extensive enough to rupture the capsule(198), resulting in fatal intra-abdominal haemorrhage (201)with required extensive blood transfusions and maternal death(198). 60 2.6.3 Central nervous system: 2.6.3.1 Glasgow coma scale (GCS) <13, stroke, two or more seizures, status eclampticus, or cortical blindness (whether transient or permanent): In neglected or, less often fulminant cases of pre-eclampsia, also non-proteinuric gestational hypertension (present in 20% of women within a week o f their first eclamptic seizure(45) or non-hypertensive gestational proteinuria (present in 10% of women(45)) eclampsia may develop. The seizures are of grand mal variety and may appear before, during or after labour. 2.6.3.2 Stroke and cerebral haemorrhage: It is not known what effects pre-eclampsia has on cerebral blood flow(202). Belfort et al used trans-cranial Doppler ultrasound and found that pre-eclamptic women with headache are significantly more likely to have abnormal cerebral perfusion than women without headache. Evidence is consistent with vasospasm or with impairment of auto regulation with passive over distension of cerebral arterioles. The principal post-mortem cerebral lesions described were oedema, hyperaemia, focal anaemia, thrombosis, and haemorrhage(203). Sheehan et al examined brains of women who developed eclampsia and found 76 different type of lesions and most common were pia-arachnoid haemorrhages, cortical and subcortical petechiae, and multiple small petechiae, and so on. Massive intracerebral haemorrhage from a ruptured intracerebral vessel, an arteriovenous malformation, or a Berry aneurysm may cause coma, or death(203) was reported in 6 out of 76 women who had both eclampsia and massive white matter haemorrhage causing coma or death. They also reported a high mortality with bleeding in to basal 61 ganglia or pons. These latter lesions were common with underlying chronic hypertension and superimposed pre-eclampsia. Haemorrhage in deep areas of the brain may result from rupture of Charcoat-Bouchard micro aneurysms that occur as a result of hypertension and aging(204) showed that 39/110 patients who died due to eclampsia died due to hypertensive cerebral haemorrhage. 2.6.3.3 Two or more seizures: Brown et al (205) found that nearly half o f the eclamptic women had abnormal findings like infarction, hypo dense areas, petechial haemorrhage and so on. These findings may provide an explanation why some women with pre-eclampsia convulse and others do not. 2.6.3.4 Blindness: Compared with normotensive women, pre-eclamptic women have increased intra ocular pressure in the peripartum period(206). Although visual disturbances are common in pre-eclampsia, blindness either alone, or in accompanying convulsions, is not. Women with varying degrees of blindness have evidence of extensive occipital lobe hypo densities. Cunningham et al(207)described 15 women with pre-eclampsia or eclampsia who had blindness. This persisted for 4 hours to 8 days, but in all in it resolved completely. In some there was some incomplete healing perhaps due to retinal thrombosis. 62 2.6.3.5 Glasgow coma scale (GCS) <13: It is rare for a woman with eclampsia not to awaken after a seizure. It is also rare for a woman with severe pre-eclampsia to become comatose without an antecedent seizure. Prognosis for these women is guarded. In some women cerebral haemorrhage causes coma. Equally ominous is coma caused by cerebral oedema. In 2 eclamptic women with coma managed at Parkland Hospital(208) there was extensive general cerebral oedema. Because coma usually follows sudden and severe blood pressure elevations, it is more likely that this phenomenon represents an inability to auto-regulate cerebral blood flow with severe acute hypertension; the result being generalized cerebral oedema(209). It is said to be the cause of death in at least 20 % of those dying from pre-eclampsia and eclampsia(210). We selected G C S <13 based on(47) and also in the study done by Bhagawanjee and associates(50) it was shown that GCS<13 performed very well in eclamptic population in identifying mortality and morbidity. 2.6.4 Renal system: 2.6.4.1 Dialysis (whether temporary or permanent), or renal transplantation: During pre-eclampsia both the renal perfusion and filtration are substantially reduced. The changes are biphasic involving first tubular function and then glomerular function. The renal functions reduced in pre-eclampsia are may be due to altered haemodynamics and glomerular morphology change. Although most women experience moderate decrement in function, some women experience severe dysfunction leading to acute renal failure mostly due to acute tubular dysfunction and rarely cortical necrosis(208). However this can occur in the event of abruptio 63 placentae(211). The incidence of acute renal failure in pregnancy severe enough to require dialysis is uncommon. However, Sibai et al(212) reported observations from 31 patients with acute renal failure complicating 18 pure pre-eclamptic women, remainder with history of chronic hypertension and renal disease with superimposed pre-eclampsia. Half of the population-required dialysis and 3 women died as a direct cause of acute renal failure. About half of this population had suffered placental abruption, and almost 90% had postpartum haemorrhage. Also Sibai et al(212) observed an association of DIC with acute renal failure and dialysis(213). Renal transplantation however is rare. 2.6.5 Haematological system: 2.6.5.1 Transfusion of >10U of blood products (in total): Haematological changes consistent with intravascular coagulation, and less often erythrocyte destruction, may complicate pre-eclampsia. In early pre-eclampsia the platelet count may be moderately reduced(133) due to the reduced lifespan of the platelets because of increased consumption (214). Only a minority of women with pre-eclampsia develop thrombocytopenia defined by a platelet like count below 150,000/uX and less common significant alterations in coagulation assays such as prothrombin, activated partial thromboplastin and bleeding time(134;215). Nevertheless, findings in pre-eclampsia of fibrin and platelet thromboses in the placental and systemic microcirculation 132) suggest that platelet and coagulation abnormalities are of fundamental significance. Fibrin deposits and thromboses in vessels of various organs were described 100 years ago, in unusual cases, widespread activation of platelets(216), decreased 64 inhibitors of coagulation pathway antithrombin 111(217) and protein C(217), leading to coagulation cascade(90;217;218) and the coagulation cascade may occur, leading to clinically apparent disseminated intravascular coagulation(134). This leads to renal cortical necrosis, adrenal and pituitary haemorrhage and necrosis and periportal hepatic necrosis. Microangiopathic haemolysis(135) associated with haemoglobinaemia (135)that is reduced haemoglobin and haptoglobinaemia with distorted and reduced red blood cells(135). This can lead to increased bleeding and postpartum haemorrhage. The haematological risks of spontaneous and postpartum haemorrhage, need for blood product transfusion, and the development of superimposed DIC rank first in maternal morbidity(219). In the H E L L P syndrome (a severe form o f pre-eclampsia), found that patients with worse outcome had more frequent blood transfusion, and more serious maternal morbidity. (84) In the study done by Sibai et al(199) it was shown that 55% of the severe pre-eclampsia ( H E L L P syndrome) population required blood or blood products transfusions. In this study overall 55% required blood or blood products to correct hypovolemia , anaemia or coagulopathy. Also Sibai et al found that 21% developed D I C and 33 patients out of these 92 developed abruptio placentae and hemorrhagic complications leading to hypotensive shock needing blood or blood product transfusion. Also in the same study it was observed that sub capsular haematoma rupture resulting in shock was an indication for massive transfusion of blood, fresh frozen plasma, and platelets and for immediate laporotomy. 4 patients in this study suffered this complication and all required massive transfusions and laporotomy for control of haemorrhage. Again Sibai et al(213) it was shown that out of 55% of patients who received blood/blood product transfusion 31% received 65 blood with a mean of 10 units of blood (2-58 range) and or 18% received platelet transfusion with a mean of 12 units (10-90 range) and or fresh-frozen plasma of 28% with 5 units mean (2-27range), and or cryoprecipitate of 12%> with mean of 20 units with (6-40 range). 2.6.6 Respiratory system: 2.6.6.1 Requirement of >50% O 2 for >lh, or intubations: Severe hypertension raises left sided filling pressures(220) Moreover, aggressive fluid administration with severe pre-eclampsia causes normal left sided filling pressures to become substantially elevated, while increasing an already normal cardiac output to supra normal levels. It is usually the older, chronically hypertensive and obese women who develops pulmonary oedema with superimposed pre-eclampsia(221;222). Sibai et al(222) reported 2.9% incidence of pulmonary oedema over a period of 9 years in severe pre-eclampsia mainly due to excessive colloid and crystalloid infusions. Dexamethasone given to enhance foetal lung maturity has also been a cofactor(223). Lehman et al(224) described 89 maternal deaths due to pulmonary oedema which was potentially preventable. The management of pulmonary oedema requires intensive care monitoring with the capability to assess pulmonary and cardiac monitoring and perform intubation and mechanical ventilation as and when needed. Also in patients with convulsions, stroke and postpartum haemorrhage, hepatic haematoma, hepatic rupture requiring laporotomy, and pulmonary oedema, there is decreased oxygen saturation and hypoxia with a need for external oxygen and ventilation. 66 2.6.7 Cardiovascular system: 2.6.7.1 Positive inotrope support, myocardial infarction, or infusion of any third anti-hypertensive: 2.6.7.1.1 Positive inotrope support: In women with severe pre-eclampsia especially with H E L L P syndrome, renal insufficiency from tubular necrosis is more common and is invariably associated with hypovolemic shock (212;213), commonly resulting from haemorrhage(212;213) at delivery, for which timely and adequate blood replacement is not given. Similarly severe pre-eclampsia complicated by placental abruption further increases the chances of renal failure and chances of hypovolemic shock(225) and hence usage of inotropes support to increase the blood pressure and cardiac output(226;227). 2.6.7.1.2 Myocardial infarction: Myocardial infarction may occur due to the severe hypertension, or due to the disease process of pre-eclampsia it self, or due to generalized vasoconstriction with decreased perfusion to the heart, or due to the result of DIC with coronary artery blockage with infarction or due to the result of iatrogenic hypotension as a result of excessive use of antihypertensives or tubular necrosis and acute renal failure or due to cerebral haemorrhage with inefficient auto regulation. 67 2.6.7.1.3 Infusion of third antihypertensives: Therapy is reserved for women in whom blood pressure is elevated to a degree that might be associated with intracranial bleeding or uncontrollable hypertension in spite of therapy with other antihypertensives. The agent most commonly used to lower blood pressure in severe pre-eclampsia is hydralazine. This is also used when there is a need for vasodilatation as well as increase in cardiac output to increase the uterine blood flow. 68 3 Materials and Methods: 3.1 Study population: This was a retrospective cohort study of women admitted to three tertiary level perinatal centres, two in Canada (Children's and Women's Health Centre of British Columbia and the Ottawa Hospital - General Campus) and one in the U K (the Maternal Medicine Unit, John Radcliff Hospital, Oxford). 3.1.1 Source population and referral patterns: The Children's and Women's Health Centre of B C , Vancouver, Canada is a university hospital specializing in the care of patients with hypertension in pregnancy as well as pre-eclampsia research. It consists of obstetricians, gynaecologists, researchers, research coordinators, nurses etc. Patients seen at this hospital are mainly residents of British Columbia. It has been estimated that the B C Women's Hospital sees about 600 women with pre-eclampsia per annum. Referrals to the hospital come mainly from within British Columbia. Such referrals take place for a variety of reasons including issues of diagnosis, decision over therapy or due to complications due to pre-eclampsia. Table 5: Feasibility of study in the source population Country Centre No of deliveries per No of G H / G P annum deliveries per annum Canada Ottawa 4000 400 Vancouver 7500 600 U K Oxford/Silver 700 200* Star Unit Total 12,200 1200 *EO PET only 69 The Maternal Medicine Unit in Oxford preferentially cares for women with pre-eclampsia diagnosed prior to 34 completed weeks of pregnancy and does not care for all cases of pre-eclampsia admitted to the hospital. A l l consecutive cases of women admitted to the respective units with a diagnosis of a hypertensive disorder of pregnancy from January 2000 - December 2001 were reviewed. The Ottawa hospital sees about 400 women with pre-eclampsia per annum. It is a tertiary unit and cares for women with pre-eclampsia especially with early onset. 3.1.2 Sources of data: Data on all women with pre-eclampsia from the above mentioned centres were recorded by retrospective chart review, into a standardized access database. The medical chart contains medical reports, laboratory reports, investigation/laboratory reports, and letters to the referring physician and or general practitioner; follow up notes including discharge summaries, results, and communication with the patient. 3.1.3 Inclusion Criteria: Women were eligible i f they had been admitted to hospital and had either: (1) hypertension (sBP >140mmHg and/or dBP >90mmHg, twice, >4h apart) after 20wk, and/or (2) proteinuria defined as >0.3g/d or >2+ dipstick proteinuria after 20wk(41). If either hypertension or proteinuria occurred in isolation, then either of the criteria, 3 or 4 (below), had to be achieved. (3) Non-hypertensive and non-proteinuric H E L L P syndrome, using Sibai's criteria (84). (4) A n isolated 70 eclamptic seizure without preceding hypertension or proteinuria, using the British Eclampsia Survey Team (BEST) criteria to define eclampsia (45). Either (3) or (4) in isolation were sufficient to meet our criteria. 3.1.4 Exclusion criteria: Women were excluded i f they were admitted in spontaneous labour or had achieved the maternal outcome prior to fulfilling the eligibility criteria. 3.1.5 Sample size calculation: For a predictive model, the sample size calculation is based on the number of variables (n) x 10, divided by the frequency of the outcome of interest in the population of interest (y). Therefore, the sample size = n x 10/y. Therefore, as an example, for 20 variables (n = 20), with a frequency of the outcome of 10% (0.1 = y), the sample size is 20 x 10/0.1 = 2000. The number of cases required for the predictive score was calculated based on the incidence of the combined maternal outcome among women with gestational hypertension and/or proteinuria as 10% and the number o f variables being tested 17. For robust variable selection, we require, on average, 10 adverse outcomes per candidate variable (228). Therefore, Using the above equation, we estimated that we require approximately 1700 charts to develop a predictive severity score. Numbers of studies have used the first half of the data for the development of development set of the severity score, which are 850 samples. However we realised at our preliminary analyses, that it is not possible to develop a development set of severity score from retrospective data, hence we aborted our data collection at 556 samples. However to conduct a 71 feasibility study to identify the frequency of the variables available in different centres, the frequency o f maternal outcome occurrence, and the participating centres that needed to be included in the study (considering the different type of management approaches). We calculated that we needed approximately 500 charts from different centres to conduct a feasibility study for preliminary testing of the predictability of these variables in the participating centres. 3.2 Data Collection: We developed a Windows Access™ database for data entry in collaboration with the clinical data managers Dave Jung from Vancouver General Hospital and Boris Kuzjelvic, B C Research Institute from Children's and Women's Health Centre of B C , Vancouver, for systematic data entry and analysis. The database was designed to have a user-friendly front end, and to provide robust information for analysis. For the day of delivery, the 'last observation carried forward' method was used, by which any preceding observation was considered current unless replaced by a more current value. For example, 24h urine proteinuria of 0.6g/d-measured 4d prior to delivery would be considered the degree o f proteinuria on the day of delivery for the purpose of the analyses. 3.2.1 Database structure: The Access database was designed to have four different tables. Each table had a user end where the data were entered. Table 1 corresponded to the datasheet form (Figure 2); Table 2 corresponded to maternal outcome form (Figure 3), Table 3 to neonatal outcome form (not shown) and laboratory results table to the Labs form (Figure 4), respectively. 72 Table l ( f ig 2) incorporated the data with respect to demography, family history, personal history, past obstetric history, treatment history, gestational age (GA) at the time o f admission, G A at the time of diagnosis of pre-eclampsia, G A at the time of termination of pregnancy, and anthropometry. H Microsoft Access - [Tabli-/1 JE! File Edit View Insert Format Records Tools Window Help IM,• y I m a y $, © & «;[%I * i si . IS x ISlxJ Patient ID [MRN)| Location , (~~ Lastname First Name PHN' • I Study ID f ,D0A . .)> Mother's DOB I EO PET 8 • I LO PET tt I nlUGRt [ Control Normal pregnancy EO/LO tt P nlUGR tt I j Non-pregnancy 8, >[ G T P SA TA ,L Past obstetric history: pervious Pie-eclampp nlUGR p Personal History: .Smoking p DVT/PE • p Cocaine .... P" sRenal:di;sa:e - | ' Pre-piegnancy diabetes 3 J R S ' r ~ "~3J _iI5§p IHD/other heart disease except MVP |fjbe|ity^^Sl New Partner Hypertension Other "31 Family history of Pre-eclampsia nlUGR heart disease ~7vEl DVT/PE \EJ Hypertension JjJ Other I ' Pre-pregnancy weight f" height P -1 Present pregnancy Weight 33j ' height (feet) | 'I 31 height (inches) Weeks Days Gest age @present'n (US): Gest age PET/nlUGR ' . R o « . l a r a fn* l a Record: l< I < H r Treatment with • Weeks Days Postpartum Days jteuids tcr fc.3' irgma-j | | | a Antihypertensives si | 3] | | | 521 > I H h*j of 521 JLL - i r • | o — • - I , patient id NUM Figure 2: Access database study entry page Fields incorporated into the study. DOB: date of birth; DVT/PE: deep venous thrombosis/pulmonary embolism; HELLP: haemolysis, elevated liver enzymes and low platelets; IHD: ischaemic heart disease; nlUGR: normotensive intrauterine growth restriction; MVP: mitral valve prolapse; PET: pre-eclampsia; RSA: recurrent spontaneous abortion. 73 H Microsoft Access - [Table3] I S I File Edit View Insert Format Records lools Window Help;" - \ a-?"-* ^m¥ - %, nivy I RESP | Record: H I < Form View IJBstartHj 0 © »| ! ®l| y a] gfej fgj gjj g]rj fflg EH) g|fj UB j ^ f l J O c l ^ - ^ ^ S l 4:12PM Figure 3: Access database maternal outcome page Fields incorporated into the study. CNS: central nervous system; CVS: cardiovascular system; HAEM: haematological; RESP: respiratory. The maternal outcome form (Table 2/Figure 3) integrated the variables about adverse maternal outcome as defined in the second chapter selection of variables and maternal outcome of my thesis. It also included were date of admission and discharge reflecting the number of days of stay in the hospital. For the above two tables information were collected and entered only once for the given patient for that admission. 74 The laboratory table (fig 4) was designed to include the data collected daily or when not available daily, as many data points as available in the chart. For the mother the variables collected were gestational age, number of days postpartum, clinical variables, biochemical and haematological variables. Under clinical variables the variables collected were blood pressure, number of seizures, saturated arterial oxygen, urine output and dipstick urinary protein. Under biochemical variables the variables collected were uric acid, creatinine, 24-hour urinary albumin, aspartate aminotransferase, lactate dehydrogenase, bilirubin, and plasma albumin. E3 Microsoft Access - [Labs: Form] File Edit View Inserts Fgrmat Records Tools Window Help ' B ' I m a y- V C P , V - '• % I a n MRN ' I -Test dale |. "''•U i WBC (cels/mL) | BP Systolic (mmHg) | BP Diastolic (mmHg)| Sa02 (*ile) RA f"^" Eclampsia tt of Seizures | Unc acid (uM) | Cieatmme (uM) j Unnary albumin g/d | Dip stick protein j AFISSe for GA (0-1 ttBti Diastolic flow [~ EXCEPTIONS T~ Study ID ..Gestational Urine oulput^M • Toial Bilirubin luM) —Mbummlg'U • —•• Platelets le*P.9A-V MPVUU- .... ^—— Fibrinogen tQ^?-Postpartum days "3 EFW %ile for GA (0-100*) Recordi i< I < sForm View 3730 • |>ll» + l of 3730 H l ^ ^ ^ e *'iBi^SaMaMIF© j jjOrjJij |NUM | Figure 4: Access daily data form or laboratory form Access database study daily follow-up page. Fields incorporated into the study. API: amniotic fluid index; AST: aspartate transaminase; BP: blood pressure; EFW: estimate of foetal weight; LDH: lactate dehydrogenase; MPV: mean platelet volume; WBC: white blood cells. 75 The haematological markers collected were white blood cells, platelets, mean platelet volume, and fibrinogen. Some of the foetal variables that were collected were umbilical artery diastolic flow, amniotic fluid index, estimated foetal weight etc. The above data were entered from the day of admission to the day o f discharge as many data points as possible. 3.2.2 Database design: The database was designed using Study ID as the primary key. It was designed so that the data needed to be entered in the data sheet before going to the maternal and neonatal section of the form. Mandatory variables were highlighted in blue colour. If the information was incomplete in the mandatory field, then the validation text message came up asking to complete the missing information. It was designed so that until the missing information was completed the program would not allow the data entry into other forms. This was designed to prevent accidental omission of the data. In case the form was closed prior to entering all the required data, a user alert message was displayed indicating that the data could not be stored. The boxes in white were designed so that they were not mandatory. Some of the fields were equipped with drop-down menus to prevent data entry errors, to hasten the data entry, and also to help systematise the data entry and analysis. 'Other' field was created so that any additional relevant medical information could be entered in the 'other' field. Each box was designed to have some validation rules and reference ranges set for that particular field. This was done to prevent or reduce the number of errors that was expected during data entry. Some of the fields were restricted to enter numbers, some for text, 76 and some for only drop-down menus depending on the variable. Laboratory tables with boxes were restricted with minimum and maximum values to prevent keying errors with extreme values with huge biases and skewed results. Some of the fields were designed to have both Metric and SI units so that the database could be used internationally. The patient was allocated a unique study ID, which acted as a primary key for other tables. This was allocated in order to remove the personal identifiers in order to maintain patient confidentiality during data analysis. The identifier/initial next to the number allotted was used to distinguish the patient records from one centre with the other centre. We used the identifier ' U ' for Oxford, ' V for Vancouver, and ' A ' for Ottawa. Guidelines and information sheet for data entry were developed and were given to all the centres, and the research co-ordinators were guided through the same at the time of preliminary stages of data entry. 3.2.3 Data entry and management: Data entry was done by research coordinators and by me according to the protocol and guidelines (see Appendix 1). 556 patient files meeting the inclusion criteria as mentioned above in the inclusion section of the thesis were entered in the database. The first 40 files were entered in the paper format of the database that was developed to crosscheck the access database and then the same was entered in the access database and was analysed for any problems and errors. The data was stored at the end of the day in the hard drive as well as backed up to the network drive. Data was monitored regularly for the missing data, data entry errors, and repetitions. Some 77 random crosschecking was done in between to verify the data base data from the patient chart data. The data was maintained and monitored by the data base manager for any problems. At the end of the data entry, data from access database was extracted to excel and then to SPSS, statistics software program for data analysis. 3.2.4 Problems encountered during data entry: Some of the patients were missing maternal outcomes as multiple people entered the data initially. There were some keying errors, and repetitions. For some of the patients the wrong study ID was allocated. In the initial phase of data entry letters were typed in the drop in section/box. For a few patients the decimal points were missing. A l l the errors were rectified after verifying it with the original charts. Some of these initial problems were fixed by improving the database structure and by tightening the validation rules and texts. Also one coordinator was asked to do the data entry for one patient. 3.2.5 Data cleaning: Random cross checking of the data that was entered in to the database was done regularly. This was done to identify any keying errors, wrong entry, spelling mistakes. Those files were identified and the cleaning was done on a regular basis. Data cleaning was done again just before the final analysis to identify the missing data points, data keying errors at the time of data entry, for repetitions, in the access database. Query is a tool to make a request the information from a database. Queries are developed to help clean the data and also for help with the analysis. The Development of the queries to identify all the abnormal data points helped in achieving this goal systematically. The appropriate corrections were made for the keying errors after verifying the 78 original patient files. Some of the errors were checked and corrected manually by going through every column and row. The extreme data points with very high and low values that were entered in the exception box were copied and entered manually in the appropriate field by removing the validation rule for that box at the time of data cleaning. 3.3 Data analysis: Data analysis was done in collaboration with Ruihua Y i n (data manager, Centre for Healthcare Innovation and Improvement), Laurie Ainsworth (Statistician, Cl inical Research Support Unit (CRSU) , B C R I ) , Monica Fernandez, (graduate student, SFU) , Boris Kuzjelvic (Data manager, C R S U ) Research Institute) and Dave Jung (Data manager, V G H ) Vancouver, B C , Canada. The Access database was designed in collaboration with Dave Jung, and Laurie Ainsworth and Monica Fernandez wrote the data analysis initial coding, and syntax. The final analysis was done in collaboration with Ruihua Y i n . The major steps in the data analysis were as follows. Access databases from different centres were merged and then the data was extracted into Excel. The data from Excel were extracted to SPSS using Study ID as primary key. Using the primary key, day of admission, day of delivery, postnatal day 1& 2 for all the patients were identified. A code was developed to do the last observation carry forward method (195; 197) for the missing data points in the lab variables. The code was written to run the analysis for all the maternal variables against the maternal outcome. Syntax for maternal age, gestational age was developed and called age.sav. Syntax for all the laboratory variables except M A P and MPV/platelets were developed and called 'get result.sps.' 79 Syntax for M A P and MPV:platelets was developed and called get result ratio.sav. The appropriate variable from last observation carried forward method was opened in SPSS in the data view section and then the corresponding syntax was opened and 'run a l l ' button was clicked. The box plot was developed using box plot graphs. Women were classified by having achieved the combined adverse maternal outcome or not, and descriptive analysis of the predictor variables was performed. Univariable analysis was done to see the frequency of data available and their relationship with the outcome. Parametric (t-test)(when the distribution was normal) and non-parametric (Mann-Whitney U test) analyses (when the distribution was skewed) were performed using SPSS for Windows (version 10). This was done by whether patient had an adverse outcome (yes) or not (no). Centre-by-centre analyses were done by selecting the study ID's for that appropriate centre and by excluding other centre study ID's at the time of running the syntax. Both parametric and non-parametric test results were presented. 3.3.1 Multivariate analyses variable selection: To develop the multivariable modelling, of all the variables listed, only the gestational age at the time of the delivery admission, mean arterial pressure ( M A P ) , dipstick protein, total leukocyte count, uric acid, creatinine, aspartate transaminase (AST) , lactate dehydrogenase ( L D H ) , and MPV:platelet count were eligible to be included in the multivariable analysis as they were reported in at least 400 cases. Univariable logistic regression analysis was used to select all variables with a univariable p-value <0.25 for subsequent multivariable analysis. Multivariable 80 logistic regression was done to estimate a model to determine the factors, which influence pre-eclampsia complications. First, we excluded the variables, which had sample size less than 400, to have adequate power for multivariable regression analyses. B y doing univariable logistic regression analysis of each variable we selected variables with univariable test of p-value less than 0.25. The variable set (white blood cells, M A P , creatinine, dipstick protein, aspartate aminotransferase, lactate dehydrogenase, mean platelet volume and platelet ratio and gestational age at the time of admission) was used for multivariable stepwise logistic regression analysis. The dependant variable was the combined adverse maternal outcome (Figure 3). A s the dependant variable was dichotomous, logistic regression modelling was applied. Both the forward and backward elimination of the variables was done using stepwise logistic regression to estimate a model to determine the factors, which influence pre-eclampsia complications. The dependent variable, i.e. the combined maternal outcome, is equal to 1 i f the patient developed the complications, 0 otherwise. In this one variable at a time was included in the regression equation; then additional variables are added one at a time until all statistically significant variables are included in the equation. After each addition of a new variable in to the equation, however, all previously entered new X variables are checked to see whether they maintain their level of significance. Previously entered X variables were retained only i f their removal w i l l cause significant cause of reduction in R . 3.4 Ethics approval : The study was approved by the University of British Columbia Clinical Research Ethics Board and the Children's and Women's Health Centre of British Columbia Ethics Board. 81 4 Results: 4.1 Demographics: There were a total of 556 (294 Vancouver, 151 Ottawa, 111 Oxford) women with pre-eclampsia for the feasibility study. The demography of the study population is shown in Tables 6 and 7. 291 of the 556 women (52%) had early onset pre-eclampsia (<34 weeks' gestation). O f these 291 women, 51 had pre-existing hypertension (18%>) and 29 had renal disease (10%). 265 of the 556 women (48%) had late-onset pre-eclampsia (>34 weeks' gestation). O f these 265 women, 29 of them had pre existing hypertension (11%) and 7 had renal disease (3%). A s shown in the demographics table (Table 6), the average maternal age at the time of admission and delivery was similar to other pre-eclampsia studies and was keeping with the observation that women with pre-eclampsia are predominantly older. The average maternal age for admission for delivery for pre-eclampsia was 30.4 for early onset of pre-eclampsia and 31.3 yrs for late onset of pre-eclampsia. The average gestational age at the time of admission for delivery for pre-eclampsia was 30.8 weeks' gestation. Extremes of age (p=0.857) were not related to the adverse maternal outcome although advanced maternal age was associated with the development of pre-eclampsia. Closer examination between early onset and late onset of pre-eclampsia cohorts revealed other similarities and differences in risk factor prevalence when compared with published cohorts. The individuals with early onset pre-eclampsia suffered from more risk factors such as obesity, family and personal history of pre-eclampsia, multiple pregnancy, personal history of hypertension compared with those women who had late onset pre-eclampsia. There were no significant differences between the early- and late onset groups with respect to ethnicity, parity, 82 and multiple pregnancies. However, a history of hypertension (17.5% vs. 11%), current smoking (13.4% vs. 8.0%), obesity (23.7% vs. 9.1%) and family history of pre-eclampsia (10.6%o vs. 4.1%)) were higher in patients who developed early onset of pre-eclampsia compared with women who developed late onset of pre-eclampsia. Patients with early onset pre-eclampsia were treated more frequently with steroids for prematurity (81.4% vs.7.6%), antihypertensives (85.2% vs. 39.69%), and magnesium sulphate (65.29% vs. 33.2%) compared with women with late onset pre-eclampsia. Not unexpectedly, the individuals with early onset of pre-eclampsia were delivered 7.6 weeks earlier compared with individuals with late onset of pre-eclampsia. 83 Table 6: Demographics Total N=556 Pre-eclampsia <34wks Pre-eclampsia >34wks PHTN+ Pre-eclampsia N=80 Renal Disease +pre-eclampsia N= 36 <34wks >34wks <34wks >34wks N=556 N=291 (52.4%) N=265 (47.6%) N=51 (63.7%) N=29 (36.2%) N=29 (80.5%) N=7 (19.4%) Maternal age (yr) 528 30.31(6.1) 31.44 (6.0) Gestational age (wks) at the time of presentation 546 30.2(3.3) 38.3 (2.5) 29.4 (2.2) 37.2 (2.4) 30.8 (3.1) 35.6 (3.2) Gestational age at the time of PET(wks) 531 28.8(4.1) 37.25 (4.2) 26.0 (3.8) 31.9 (4.0) 27.1 (3.8) 34.7 (4.1) Gestational age (wk) at the time of termination of pregnancy 556 30.4 (3.1) 38.0 (2.7) 30.7 (3.5) 38.2 (3.7) 31.9 (4.0) 38.5 (4.1) Ethnicity Caucasian Black Asian South Asian 1 s t nations Latin Pacific islanders Others Unknown 556 105 28 (9.6%) 77 (29%) 2 (3.9%) 9 (31.3%) 3 (8.3%) 3 (42.8%) 7 4 (1.37%) 3 (1.13%) 1 (1.9%) 0 (0%) 1 (2.7%) 0 (0%) 51 13 (4.46%) 38 (14.3%) 2 (3.9%) 3 (10.3%) 1 (2.7%) 0 (0%) 51 23 (0.68%) 28 (10.5%) 6 (11.7%) 2 (6.8%) 1 (2.7%) 0 (0%) 14 1 (0.34%) 13 (4.9%) 1 (1.9%) 1 (3.4%) 1 (2.7%) 0 (0%) 6 1 (0.34%) 5 (1.8%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 1 (0.34%) 1 (0.37% 0 (0%) 0 (0%) 0 (0%) 0 (0%) 10 6 (2.6%) 4 (1.5%) 2 (3.9%) 2 (6.8%) 1 (2.7%) 1 (14.2%) 310 213 (73.0%) 92 (34%) 36 (70.5%) 12 (41.3%) 22 (75.8%) 2 (28.5%) Parity Nulliparous Multiparous 551 270 146 (51%) 124 (47%) 281 144 (50%) 137 (52%) Multiple pregnancy Twin Triplet 556 15 (5.2%) 2 (0.6%) 3 (1.13%) 0 (0%) Past obstetric h/o Pre-eclampsia 555 48 (16.5%) 33 (12.4%) 17 (33.3%) 9 (31.3%) 4 (13.8%) 2 (28.5%) NIUGR 555 5 (1.71%) 11 (4.1%) 4 (7.8%) 3 (10.3%) 2 (6.8%) 1 (14.2%) Pre-pregnancy diabetes 555 5 (1.71%) 8 (3.0%) 1 (1.9%) 0 (0%) 2 (6.8%) P (0%) Recurrent spontaneous abortion 555 11 (3.8%) 9 (3.4%) 2 (3.9%) 1 (3.4%) 0 (0%) 0 (0%) 84 Table 7: Demographics (Cont'd) Total Pre-eclampsia <34wks Pre-eclampsia >34wks P H T N + Pre-clampsia N=80 <34wks >34wks Renal Disease +pre-eclampsia N = 36 <34wks >34wks N=556 N=291 52.4% N=265 47.6% N=51 (63.7%) N=29 (36.2%) N=29 (80.5%) N=7 (19.4%) Past medica l h / o Hypertension D V T / P E Early onset heart disease except M V P Obesity N e w partner Cocaine use Smoker 555 39 (13.4%) 21 (8%) 6 (12%) 4 (14%) 6 (21%) 2 (29%) 555 0 (0%) 6 (2.3%) 0 (0%) 2 (6.9%) 0 (0%) 0 (0%) 553 10 (3.4%) 8 (3.01%) 3 (5.9%) 0 (0%) 1 (3.4%) 0 (0%) 543 69 (23.7%) 24 (9.05%) 22 (43%) 5 (17%) 8 (28%) 2 (29%) 509 11 (3.07%) 7 (2.6%) 3 (5.8%) 0 (0%) 2 (6.8%) 0 (0%) 553 1 (0.34%) 2 (0.75%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 554 39 (13.4%) 21 (7.9%) 6 (12%) 4 (14%) 6 (21%) 2 (29%) Fami ly h / o Pre-ec lamps ia 511 31 (10.65%) 11 (4.1%) 8 (16%) 2 (6.9%) 5 (17%) 0 (0%) n l U G R 512 5 (1.71%) 7 (2.6%) 0 (0%) 0 (0%) 1 (3.4%) 0 (0%) E a r l y onset heart disease except M V P 513 67 (23%) 63 (23.7%) 13 (26%) 9 (31%) 1 (3.4%) 0 (0%) H y p e r t e n s i o n 514 104 (35.7%) 101 (38.1%) 24 (47%) 15 (52%) 12 (41.3%) 2 (29%) D V T / P E 513 2 (0.68%) 2 (0.75%) 0 (0%) 1 (3.4%) 0 (0%) 0 (0%) Treatment with Steroids for prematur i ty 548 237 (81%) 20 (7.5%) 46 (90%) 0 (0%) 14 (82.7%) 0 (0%) A n t i hyper tensive 553 248 (85%) 104 (39%) 46 (90%) 22 (76%) 12 (75.8%) 1 (14%) MgSo4 545 190 (65%) 87 (32.8%) 26 (51%) 8 (28%) .5 (51.7%) 1 (14%) 4.2 Risk factors for pre-eclampsia: We compared traditional risk factors in pre-eclampsia in mothers who developed the adverse maternal outcome vs. mothers who did not develop the adverse outcome. There was a positive relation between family history of pre-eclampsia and the combined adverse maternal outcome (p=0.0007) in our study. Interestingly, having a personal history of pre-eclampsia was not. When tested for the gestational age at onset o f pre-eclampsia by v 2 test, early onset pre-eclampsia was clearly associated with higher maternal adverse outcome compared with late onset pre-eclampsia (p=0.003). Also there was a positive relationship to the number and 85 frequency of usage of antihypertensives and MgSCu for hypertension and seizures for development of adverse maternal outcome as expected (p<0.00001 and p=0.0001 respectively). It is noteworthy that although nulliparity is a well-accepted risk factor for pre-eclampsia, in our study nulliparity was not found to be significantly associated with adverse maternal outcome (p=0.2). Also well known risk factors like pre-pregnancy hypertension (p=0.6), pre-pregnancy diabetes (p=0.4), having past obstetric history of pre-eclampsia (p=1.0), having a new partner (p=0.8) were not significantly associated with adverse maternal outcome. Smoking was also not shown to be protective (p=0.14) for adverse development of maternal outcome. 86 Table 8: Risk factors vs. Outcome Comparison or Outcome No. Of Partici pants Effect Size 'P' value 01 EO PET vs. LO PET 01 E O PET vs. L O PET to predict adverse maternal outcomes 556 2.29[1.32,3.98] 0.003*** 02 Primiparous vs. Mutliparous 01 Parity to predict adverse maternal outcomes 556 0.73[0.45,1.20] 0.2 03 Past obstetric history 01 H / O PET 555 1.00[0.53,1.88] 1.0 02 H / O n lUGR 555 0.69[0.17,2.85] 0.6 03 PHx RSA 555 0.46[0.06,3.34] 0.4 04 Past medical history 01 H / O D V T / P E 555 0.57[0.03,9.81] 0.7 02 H / O renal disease 549 0.72[0.23,2.26] 0.6 03 H / O heart disease 553 0.49[0.07,3.59] 0.5 04 H / O pre-pregnancy hypertension 555 1.19 [0.65,2.17] 0.6 05 H / O pre-pregnancy diabetes 551 0.32[0.02,5.25] 0.4 05 Family history 01 FHx PET 504 3.04[1.60,5.78] 0.0007*** 02 FHx n lUGR 513 0.46[0.03,7.93] 0.6 03 FHx hypertension 511 0.99[0.72,1.34] 0.9 04 FHx D V T / P E 510 0.67 [0.04,11.66] 0.8 05 FHx heart disease 534 0.62[0.34,1.13] 0.12 06 Current pregnancy 01 H / O smoking in this pregnancy 554 0.49[0.18,1.28] 0.14 02 H / O cocaine in this pregnancy 553 0.65[0.04,11.44] 0.8 03 H / O new partner in this pregnancy 509 0.87[0.21,3.61] 0.8 04 H / O steroids for F L M +/- H E L L P in this pregnancy 548 1.20[0.95,1.53] 0.13 05 H / O antihypertensive use in this pregnancy 553 1.44[1.29,1.60] 0.00001*** 06 H / O MgS04 use in this pregnancy 545 1.43 [1.19,1.72] 0.0001*** Assessment was done using Chi-square analysis 'P' value >0.05 ns, 0.01-0.05 =*(significant), 0.001 to 0.01=** (very significant), <0.001= *** (extremely significant) DVT/PE: deep venous thrombosis/pulmonary embolism; EO: early-onset; FLM: fetal lung maturity; HELLP: haemolysis, elevated liver enzymes, low platelet syndrome; H/O: history of; LO: late-onset; nlUGR: normotensive intrauterine growth restriction; PET: pre-eclampsia 87 4.3 Incidence of the combined adverse maternal outcome: Among these 556 women with pre-eclampsia, 58 (10.4 %) developed the combined adverse maternal outcome, including one maternal death (Table 9). Figure 5 is the graph describing the incidence of outcomes. The most common outcomes (>10 cases) reached were: >50% O2 for > lh (n=21), 3 r d infused antihypertensive (n=20), and transfusion >10 units (total) of blood products (n=l 1). N o women required either dialysis or renal transplantation. There were more number of women with requirement of >50% O2 for > lh from Ottawa, more cases of infusion of 3 r d antihypertensives from Oxford centre, and number of placental abruption (not included in the adverse combined maternal outcome) from Vancouver. Also there were more number of blood and blood product transfusions in Vancouver compared to other centres. There were equal number of hepatic failure and seizures cases in Oxford and Vancouver. Table 9 describes the incidence of outcome based on the centre. Table 9: Incidence of outcome by centre. Variable name Oxford (111) Vancouver (294) Ottawa (151) Total Mortality 1 0 0 1 Hepatic failure 2 2 0 4 Hepatic haematoma/rupture 0 0 1 1 Glasgow Coma Score <13 1 1 0 2 Stroke 0 2 0 2 2 or more seizures 3 3 0 6 Cortical blindness 0 0 1 1 Myocardial infarction with chest pain 0 0 3 3 Dialysis 0 0 0 0 Transplantation 0 0 0 0 Positive inotropic support 1 1 0 2 Infusion of 3rd antihypertensive 14 4 2 20 Requires >50% 0 2 > 1 hr 2 0 19 21 Intubation (Vent, EIT, CPAP) 2 0 1 3 > 10 U of blood or blood products 5 5 1 11 Platelet transfusion 2 7 1 10 Fresh frozen plasma infusion 4 0 2 6 Cryoprecipitate infusion 0 1 2 3 88 Incidence of outcome accordina to centres 25 20 o 15 0) 3 cr £ 10 u. i 1 1 1 1 n 1 «"B v<* Maternal outcome Figure 5: Incidence of outcome according to centres. The woman who died was a 31 year-old primigravida, was admitted at 39 weeks (delivered at 4 0 + 0 weeks) with pre-eclampsia (day of delivery: maximum B P 180/100; 18g proteinuria/d; urate: 488u.M; creatinine: 118p.M; bilirubin: 10u,M). Postnatally, she developed severe H E L L P (haemolysis, elevated live enzyme, low platelet) syndrome(71) and died on her 37 t h postnatal day following a cerebral haemorrhage, having developed a paradoxical 12 thrombocytosis [highest postnatal platelet count (postnatal day 8): 1114 x 10 / L ] . 89 4.4 Risk factor/variable prevalence and statistical analysis: The relationships between laboratory values on the day o f admission, day o f delivery, postnatal day 1, postnatal day 2 and, the subsequent development of significant maternal morbidity in the categories of haematology, respiratory, central nervous system, renal, hepatic, and coagulation systems were investigated by parametric and non parametric analysis. A n example of a parametric test is the Student T test. These were done when the distribution was normal (that is when the mean was closer to median). The non parametric Mann-Whitney U test was done when the distribution was not normal/skewed. When the test was significant, the appropriate p value was selected and * sign was put next to that corresponding ' P ' value. Table 10 (page no.99) summarizes the distribution of laboratory values from 556 women with admission day and Table 1 l(page no. 100) for the day of delivery laboratory values who either did (n=58) or did not (n=498) develop the combined adverse maternal outcome. Table 12 (page no.107) and 13 (page no.108) summarizes the distribution of laboratory values from the 556 women on postnatal days 1 and 2 according to their development, or not, o f the combined adverse maternal outcome respectively. 90 Table 10: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for the day of admission. Variable Day of admission Parametric and non parametric test results N Mean (SD) Median [range] T test (P value) MWUTest Maternal age (Yrs) no adverse outcome adverse outcome 528 472 56 30.9 (6) 30.9 (6) 31.0 [17, 46] 30.6 |20, 43] 0.857 0.834 Gestational age at diagnosis (Wks) no adverse outcome adverse outcome 531 474 57 33.0 (6) 30.2 (6) 33.3 (18, 42] 31.0 [20,40] <0.001*** <0.001 Gestational age on admission (Wks) no adverse outcome adverse outcome 546 488 58 34.3 (5) 31.8(5) 34.6 |20, 42] 31.8 [20, 41] <0.001*** <0.001 Total leukocyte count (xlO'/L) no adverse outcome adverse outcome 464 411 53 11.7(4) 12.5 (4) 11.0 [4, 25] 11.9 [6,24] 0.201 0.133 Mean arterial pressure (mmHg) no adverse outcome adverse outcome 504 451 53 121 (14) 128 (14) 120 [83, 217] 127 [100, 170] 0.001*** 0.001 Sa0 2 (%ile) no adverse outcome adverse outcome 63 47 16 96 (3) 96 (2) 96 |87, 100] 97 [92, 100] 0.435 0.576 Urate (• M) no adverse outcome adverse outcome 375 328 47 372 (93) 371 (101) 360 [136, 648] 364 [172, 666| 0.809 0.919 Creatinine (p M) no adverse outcome adverse outcome 366 314 52 70 (20) 73 (23) 69 [27, 236] 73 [28, 139] 0.348 0.213 Dip stick protein no adverse outcome adverse outcome 458 411 47 2.9(1.9) 3.8(1.2) 3[0.0, 5.0] • 4[1.0, 5.0] <0.001*** <0.001 Urinary albumin (g/D) no adverse outcome adverse outcome 46 41 5 2.0(2.7) 2.4(2.7) 0.86[0.0, 11.6] 2.8|0.5, 4.26] 0.646 0.223 Urinary output (ml/H) no adverse outcome adverse outcome 219 197 22 109 (66) 87 (37) 90 [18, 375] 88 [15, 161] 0.024* 0.291 Bilirubin (0 M) no adverse outcome adverse outcome 92 72 20 8.8 (12) 17.6 (31) 6 [1, 93] 9 [1, 139] 0.052 0.066 Aspartate transaminase (u/L) no adverse outcome adverse outcome 456 406 50 69 (195) 149 (408) 26 [9, 3340] 31 [12, 2704] 0.174 0.271 Lactate dehydrogenase (u/L) no adverse outcome adverse outcome 351 315 36 588 (477) 794 (1091) 526 [103, 4592] 329 [164,3759] 0.271 0.146 Plasma albumin (g/L) no adverse outcome adverse outcome 114 93 21 31.6 (3) 31.0(4) 31 [21, 40] 31 [24, 36] 0.254 0.354 Fibrinogen (• M) no adverse outcome adverse outcome 85 75 10 4.9 (1.5) 3.9 (1.4) 4.8 |1,9| 4.1 [2, 6] 0.050* 0.077 Mean platelet volume: platelet count ratio (fl X 10') no adverse outcome adverse outcome 388 348 40 0.07 (0.05) 0.09 (0.11) 0.06 [0.02, 0.44| 0.06 [0.01, 0.48] 0.015* 0.629 91 Table 11: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for the day of delivery. Variable Day of delivery Parametric and non parametric test results N Mean (SD) Median [range] T test (P value) MWU Test Total leukocyte count (xlO'/L) no adverse outcome adverse outcome 515 459 56 13.2 (4) 15(6) 12.53 [0.4, 40] 14.5 [7,35] 0.014* 0.059 Mean arterial pressure (mmHg) no adverse outcome adverse outcome 532 478 54 112 (14) 123 (16) 112 [67,172] 122 [95,177] 0.001** 0.001 Sa0 2 (%ile) no adverse outcome adverse outcome 145 120 25 96(2) 95(2) 96 [83,100] 95 [92, 98] 0.064 0.066 Urate (|iM) no adverse outcome adverse outcome 410 360 50 383 (97) 401 (109) 380 [136,705] 384 [105, 661] 0.394 .352 Creatinine (p,M) no adverse outcome adverse outcome 424 372 52 73 (26) 82 (50) 71 [18,359] 75 [32, 332] 0.014* 0.039 Urinary output (ml/H) no adverse outcome adverse outcome 325 294 31 144 (81) 145 (86) 128 [23, 575] 129 [15, 323] 0.679 0.177 Dip stick protein no adverse outcome adverse outcome 502 451 51 2.6(1.5) 3.5 (1.3) 3[0.0, 5.0] 4[1.0, 5.0] <0.001** * <0.001 Urinary albumin (g/D) no adverse outcome-adverse outcome 68 60 8 2.3(2.7) 3.6(5.9) 1.1[0.0,11.6] 1.8[0.4,17.99] 0.530 0.452 Bilirubin (uM) no adverse outcome adverse outcome 147 114 33 10 (11) 16 (17) 6 [1,79| 9 [3, 64] 0.002 0.021* Aspartate transaminase (u/L) no adverse outcome adverse outcome 515 461 54 60 (128) 381 (1671) 29 [10,1305 46 [10,12132] 0.019 0.006* Lactate dehydrogenase (u/L) no adverse outcome adverse outcome 426 378 48 643 (584) 934 (1484) 536 [115, 5847| 439 |138,8497| 0.002 0.396 Plasma albumin (g/L) no adverse outcome adverse outcome 146 119 27 29 (5) 27(6) 29 [15,45] 28 [19,42] 0.665 0.568 Fibrinogen (p.M) no adverse outcome adverse outcome 129 111 18 5.0 (1.5) 3.7 (1.5) 4.8 [1,9] 3.5 [1,7] 0.050* 0.077 Mean platelet volume: platelet count ratio (fl X 109) no adverse outcome adverse outcome 445 399 46 0.07 (0.06) 0.10 (0.12) 0.06 [0.01, 0.44] 0.07 [0.02, 0.35] 0.015* 0.629 Italicised comparisons: (T-test or Mann-Whitney U test as appropriate). MPV: mean platelet volume; Sa02: oxygen saturation (pulse oximetry, room air). Mean arterial pressure: diastolic blood pressure + (pulse pressure/3). 'P'value >0.05 ns, 0.01-0.05 =*(significant), 0.001 to 0.01=** (very significant), <0.001= *** (extremely significant) 92 On the day of admission, gestational age at admission (p=<0.001), gestational age at diagnosis (p=<0.0001), fibrinogen (p=0.05) urinary output (p=0.02) were lower, and mean arterial pressure (p=0.001), dipstick protein (p=<0.001), and MPV/Platelet (p=0.01) were greater, in women destined to develop the combined adverse maternal outcome. For the day of delivery, Total leukocyte count (p=0.01), mean arterial pressure (p=0.001), dipstick protein (p=<0.001), bilirubin (0.02), creatinine (p=0.01), aspartate transaminase (AST) (p=0.006), and MPV/Platelet (p=0.01), were greater, and fibrinogen (p=0.05) were lower in women achieving the combined adverse outcome. Even though fibrinogen (n=85 for day of admission and n=T29 for day of delivery) and bilirubin (n=92 for day of admission and n=147 for day o f delivery) were significant, there were small numbers of samples available for drawing a conclusion. Similarly 24-hour urinary protein (n=46 for day of admission and n=68 for day of delivery) and plasma albumin (n=l 14 for day of admission and n=146 for day of delivery) were not done regularly. F ig 6 to 16 shows the box plot for patients who developed the outcome vs who did not for the day of admission and day of delivery. 93 o W3 T3 CD < Gestational age at admission vs final outcome Normal maternal outcome Adverse maternal outcome 50 40 30 20 10 PO.001 N= 488 Normal maternal outcome 58 Adverse maternal outcome Figure 6: Gestational age at admission compared by the occurrence or absence of the combined adverse maternal outcome. Gestational age at diagnosis of pre-eclampsia vs. outcome Normal maternal outcome Adverse maternal outcome Figure 7: Gestational age at diagnosis compared by the occurrence or absence of the combined adverse maternal outcome. 94 Dipstick protein on the day of admission Normal maternal outcome Adverse maternal outcome Figure 8: Dipstick protein compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission. MAP on the day of admission 24 0 Normal maternal outcome Adverse maternal outcome Figure 9: Mean arterial pressure compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission 95 MPV/platelet ratio on the day of admission or Q < Normal maternal outcome Adverse maternal outcome Figure 10: MPV/Platelets compared by the occurrence or absence of the combined adverse maternal outcome on the day of admission Dipstick protein on the day of delivery Normal maternal outcome Adverse maternal outcome Figure 11: Dipstick protein compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. 96 MAP on the day of delivery P=0.001 C 6 1 5 1 s o J 120 J rr: LU O102 Normal maternal outcome Adverse maternal outcome Figure 12: Mean arterial pressure compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. 97 MPV/Platelets on the day of delivery Normal maternal outcome Adverse maternal outcome Figure 13: MPV/Platelets compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. Bilirubin for the day of delivery 100 CD > "55 T3 M— o >. ro Q 52 18 Normal maternal outcome Adverse maternal outcome Figure 14: Bilirubin compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. 98 Creatinine on the day of delivery P=0.01 •was •mtt an Normal maternal outcome Adverse maternal outcome Figure 15: Creatinine compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. Fibrinogen on the day of delivery Normal maternal outcome Adverse maternal outcome Figure 16: Fibrinogen compared by the occurrence or absence of the combined adverse maternal outcome on the day of delivery. 99 Table 12: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for postnatal day 1. Variable Postnatal dayl Parametric and non parametric test results N Mean (SD) Median [range] T test (P value) MWU Test Total leukocyte count (xlO'/L) no adverse outcome adverse outcome 520 464 56 14.1 (5) 16(6) 13.5 [0.4, 33] 15 [5.5,24] 0.009** 0.014 Mean arterial pressure (mmHg) no adverse outcome adverse outcome 504 478 54 11314) 122(16) 113[67,153] 123 [85, 163] <0.001** * <0.001 Sa0 2 (%ile) no adverse outcome adverse outcome 155 128 27 96(3) 94(2) 96 [83,100| 94 [88, 98] 0.299 0.006 Urate (uM) no adverse outcome adverse outcome 494 440 54 384 (96) 409 (102) 381 [136, 682] 417 [105, 661] 0.156 0.108 Creatinine (uM) no adverse outcome adverse outcome 432 378 54 74 (26) 99 (125) 71 [30, 328] 75 [35, 937] 0.001*** 0.240 Urinary output (ml/H) no adverse outcome adverse outcome 331 300 31 140 (79) 137 (80) 124 [23, 579] 129 [15, 323] 0.859 0.817 Dip stick protein no adverse outcome adverse outcome 506 455 51 2.2(1.4) 3.1(1.6) 2[0.0, 5.0] 3|0.0, 5.0] 0.001*** <0.001 Urinary albumin (g/D) no adverse outcome adverse outcome 73 65 8 2.2(2.9) 3.6(5.9) 0.9[0.0,15.8] 1.8(0.37,17.99] 0.508 0.372 Bilirubin (uM) no adverse outcome adverse outcome 156 121 35 9.6 (11) 20 (25) 6 [2, 79] 9 [3, 92] 0.001 0.055* Aspartate transaminase (u/L) no adverse outcome adverse outcome 518 463 55 86 (196) 583 (1756) 29 [10,1715] 56 [10,11802] <0.001 0.001*** Lactate dehydrogenase (u/L) no adverse outcome adverse outcome 435 386 49 711 (692) 1328(2211 559 [115, 6000] 412 [133,11110] 0.058 0.759 Plasma albumin (g/L) no adverse outcome adverse outcome 156 126 30 28 (5) 26 (6) 28 [15, 39| 26 [17, 42] 0.026* 0.021 Fibrinogen (uM) no adverse outcome adverse outcome 131 113 18 4.9 (1.5) 3.6 (1.3) 4.7 [1,9] 3-5 [1,6] 0.001*** 0.001 Mean platelet volume: platelet count ratio (fl X 109) no adverse outcome adverse outcome 454 408 46 0.09 (0.09) 0.13 (0.14) 0.06 [0.01,0.82] 0.07 [0.02, 0.73[ 0.001 0.027* Italicised comparisons: (T-test or Mann-Whitney U test as appropriate). MPV: mean platelet volume; Sa02: oxygen saturation (pulse oximetry, room air). Mean arterial pressure: diastolic blood pressure + (pulse pressure/3). 'P'value >0.05 ns, 0.01-0.05 =*(significant), 0.001 to 0.01=** (very significant), <0.001= *** (extremely significant) 100 Table 13: Data from women with pre-eclampsia compared by the occurrence or absence of the combined adverse maternal outcome for postnatal day 2. Variable Postnatal day 2 Parametric and non parametric test results N Mean (SD) Median [range] T test (P value) MWU Test Total leukocyte count (xl09/L) no adverse outcome adverse outcome 521 465 56 13.2 (4) 15(6) 12.53 [0.4, 40] 14.5 [7, 35] 0.004** 0.006 Mean arterial pressure (mmHg) no adverse outcome adverse outcome 532 478 54 112(14) 123 (16) 112 [67,172] 122 [95,177] <0.001*** <0.001 Sa0 2 (%ile) no adverse outcome adverse outcome 157 129 28 96(2) 95(2) 96 [83,100] 95 [92, 98] 0.014* 0.002 Urate (p.M) no adverse outcome adverse outcome 495 441 54 383 (97) 401 (109) 380 [136, 705] 384 [105, 661] 0.333 0.390 Creatinine (uM) no adverse outcome adverse outcome 432 378 54 73 (26) 82 (50) 71 [18,359] 75 [32, 332] 0.028* 0.426 Urinary output (ml/H) no adverse outcome adverse outcome 333 302 31 144 (81) 145 (86) 128 [23, 575] 129[15, 323] 0.966 0.955 Dip stick protein no adverse outcome adverse outcome 511 459 52 2.2(1.4) 2.6(1.6) 2[0.0, 5.0] 3[0.0, 5.0] 0.030* 0.050 Urinary albumin (g/D) no adverse outcome adverse outcome 92 81 11 2.2(2.7) 3.2(5.1) l.OOfO.0,15.8] 1.1 [0.3,17.99] 0.310 0.448 Bilirubin (uM) no adverse outcome adverse outcome 169 132 37 10 (11) 16 (17) 6 [1,79] 9 [3, 64] 0.011 0.034* Aspartate transaminase (u/L) no adverse outcome adverse outcome 518 463 55 60 (128) 381 (1671) 29 [10,1305] 46 [10,12132] 0.161 0.005** Lactate dehydrogenase (u/L) no adverse outcome adverse outcome 445 395 50 643 (584) 934 (1484) 536 [115, 5847] 439 [138, 8497] 0.009 0.880 Plasma albumin (g/L) no adverse outcome adverse outcome 159 127 32 29 (5) 27 (6) 29 [15,45] 28 [19,42] 0.031* 0.032 Fibrinogen (uM) no adverse outcome adverse outcome 133 115 18 5.0 (1.5) 3.7 (1.5) 48 [1,9] 3.5 [1,7] 0.001** 0.001 Mean platelet volume: platelet count ratio (fl X 109) no adverse outcome adverse outcome 472 424 48 0.07 (0.06) 0.10 (0.12) 0.06 [0.01,0.44] 0.07 [0.02, 0.35] <0.001*** 0.018 Italicised comparisons: (T-test or Mann-Whitney U test as appropriate). MPV: mean platelet volume; Sa02: oxygen saturation (pulse oximetry, room air). Mean arterial pressure: diastolic blood pressure + (pulse pressure/3). 'P'value >0.05 ns, 0.01-0.05 =*(significant), 0.001 to 0.01=** (very significant), <0.001= *** (extremely significant) 101 On postnatal day 1, Total leukocyte count (p=0.009), M A P (p=<0.001), dipstick protein (p=0.001), bilirubin (0.05) (fig 17), aspartate transaminase ( A S T ) (p=0.001), and MPV/Platelet (p=0.02), were greater, and plasma albumin (p=0.02), fibrinogen (p=0.001) were lower in women achieving the combined adverse outcome. On postnatal day 2, Total leukocyte count (p=0.004), M A P (p=<0.001), creatinine (p=0.02), bilirubin (p=0.03), dipstick protein (p=0.03), A S T (p=0.005) and MPV/Platelet ratio (p=<0.001) were elevated and fibrinogen (p=0.001), plasma albumin (p=0.03) and Sao2 (p=0.014) were decreased in women achieving the combined adverse outcome. O f all the variables listed, only the gestational age at the time of the delivery admission, mean arterial pressure ( M A P ) , dipstick protein, total leukocyte count, uric acid, creatinine, aspartate transaminase (AST) , lactate dehydrogenase ( L D H ) , and MPV:platelet count were subjected to univariable analysis (>400 cases). This revealed that, on the day of admission, gestational age on admission, M A P , dipstick proteinuria, and MPV:platelets were associated with the combined adverse outcome. In addition to these variables, on the day of delivery, total leukocyte count, creatinine, A S T , and L D H were associated with the outcome (p<0.25). 102 Fibrinogen on postnatal day 1 N - 113 Normal maternal outcome Adverse maternal outcome Figure 17: Fibrinogen compared by the occurrence or absence of the combined adverse maternal outcome on postnatal day 1. Bilirubin on postnatal day 1 F I N A L O U T Figure 18: Bilirubin compared by the occurrence or absence of the combined adverse maternal outcome on postnatal day 1. 103 Table 14: Descriptive Statistics from logistic regression N Minimum Maximum Mean Std. Deviation Complication 556 0 1 0.1 0.3 WBC 510 4.489 34.779 14.154 5.207 M A P 536 76.666 164.666 120.105 13.414 Uric acid** 495 136 682 382.28 94.34 Creatinine* 415 27 239 74.426 25.69 Dipstick protein 505 0 5 2.705 1.49 Bilirubin* 145 1 139 11.58 17.24 AST* 508 10 5000 124.99 387.836 L D H * 427 103 5847 698.4 698.56 Albumin 136 15 43 28.41 4.86 MPV/Platelet 449 0.01 .48 .083 .069 Fibrinogen 130 1 9.43 4.63 1.53 G A at admission 546 19.57 42.43 34.0162 5.02105 G A at diagnosis 531 6.00 42.43 32.6981 5.90939 Table 15: Multivariable logistic regression result Variables in the Equation B S.E. Wald df Sig. Exp(B) 95.0% C.I.for EXP(B) Lower 95.0% C.I.for EXP(B) Upper G A at admission -.159 .040 15.470 1 <.0001 .853 .789 .924 MPV/Platelet ratio* 7.975 2.154 13.703 1 <.0001 2907.885 42.629 198358.427 Dipstick protein .429 .152 7.989 1 .005 1.535 1.140 2.066 G A at admission/MPV/Plat 1.081 .396 7.442 1 .006 2.947 1.356 6.405 Dipstick protcin/GA at admission .041 .024 2.913 1 .088 1.042 .994 1.092 *For a 0.01 rise in ratio, odds for adverse outcome increased by 29-fold Multivariable logistic regression was done to estimate a model to determine the factors, which influence pre-eclampsia complications. The multivariable logistic regression revealed that the gestational age at admission (odds ratio 0.853 [95% confidence ratio 0.789, 0.924], dipstick protein (OR 1.535 [1.140, 2.066]), and MPV:platelet ratio (OR 291.4 [42.629, 198358.4]) were independent predictors of the combined adverse maternal outcome. This indicated that for each week of advanced gestation prior to admission, the odds for 104 developing the combined adverse maternal outcome was lowered by 0.853. Similarly, for every '+' increase in proteinuria, the odds for developing the outcome increased by 1.54-fold. For the M P V : platelet ratio, a one-unit rise in the ratio would not be biologically plausible, while a 0.01 unit rise would be (Table IV). Such a rise increased the chance of adverse outcome by 29 -fold. 4.5 Brief summary of results: In summary, gestational age on the day of admission for delivery, gestational age at the time of diagnosis of pre-eclampsia, dipstick protein, M A P , MPV/Platelet ratio appeared to be significant in predicting the adverse combined maternal outcome. Bi l i rubin and fibrinogen appeared to be significant on the day of delivery and postnatally, however, their importance needs to be verified by collecting the variables regularly prospectively as their numbers were not enough to draw any conclusion. Also , by parametric analyses, A S T and creatinine were significant for the day o f delivery and postnatally. However, the distribution of each was skewed and we did not do any correction for the correct interpretation of the data as it is only a descriptive analysis. We were unable to derive a model for prospective validation. Development set must be developed prospectively. 105 5 Discussion: A s suggested by Sibai and associates(229) analysis of the data from hundreds of women with 'm i ld ' and 'severe' pre-eclampsia with or without outcome provides some direction for clinicians as they perform a series of laboratory tests early in the assessment process of women with pre-eclampsia. These tests may have the ability to estimate the risk, in an individual case, of the subsequent development of significant maternal morbidity in the cardiopulmonary, hepatic, haematological and renal systems. Regardless of the original classification of pre-eclampsia, especially in women deemed to be at low risk because of signs and symptoms, performance of supplemental broad battery of tests may reveal that woman is at high risk of developing significant morbidity. This information could be helpful i f and when entered into the scoring system to the clinician for management purposes, in terms of prognostication, the type of management whether to deliver the baby or manage expectantly and possibly as grounds for transfer of the patient to a tertiary unit. This approach to patient assessment and care could have considerable utility for a number of reasons. Although clinical and laboratory criteria have been developed to differentiate severe from mild pre-eclampsia and H E L L P from severe pre-eclampsia and to determine severity o f pre-eclampsia(230-232), the disease process is reversed only by delivery(230). The ultimate goal of any protocol for management of pre-eclampsia must be maternal safety first, followed by delivery of a live, mature newborn infant in optimal condition(233). The optimal time for delivery of pre-eclampsia remains elusive and, therefore, some amount of expectant management exists in all protocols. Ideally, it would be advantageous to predict and thus prevent development of severe complications of this disease process. 106 A s stated, delivery is the definitive treatment for pre-eclampsia, the maternal treatment of choice, but a double-edged sword for the foetus remote from term(37). Randomised controlled trial (RCT) evidence shows that, remote from term, prolongation of pregnancy by expectant therapy (delaying delivery until necessitated by either maternal or foetal condition) decreases serious perinatal morbidity without, increased maternal risk(35;36). However, these R C T s had insufficient power to detect a difference in serious maternal outcomes between groups, and uncertainty about the magnitude of the maternal risk(40) has made some clinicians reluctant to use such management. Closer to, or at term, expectant management has been advocated to reduce obstetric consequences (Caesarean section) of a policy of routine induction for women with mild gestational hypertension or pre-eclampsia >34wk(234). Barton et al(234) advise admission and delivery for women at 34-37wk should they develop persistent hypertension or proteinuria, abnormal laboratory results, poor foetal growth, or i f affected women are unreliable. Between 37-40wk, they do not advocate induction unless the woman has a favourable cervix, there is evidence of foetal compromise, or there are either visual disturbances or persistent headaches(234). In order to enhance the care of women with pre-eclampsia, we need a classification rule which w i l l identify which women can be safely managed by expectant therapy(34). The decision to terminate the pregnancy with severe pre-eclampsia with low platelets for maternal indications could be based not only on descent of platelets to a certain level of range but also on the achievement of the high risk thresholds of other variables like bilirubin, 107 MPV/Platelet ratio, higher score and so on. On the other hand, persistence of laboratory values with a low score could be used as grounds to continue attempts at vaginal delivery. A s suggested, earlier decision to transfer a patient with severe pre-eclampsia could be based on the same information. Comparison of patient populations among different institutions, and countries might be enhanced i f studies were based not only on one laboratory variable but on a new classification system or scoring system which includes all the different variables of importance/sub-groupings defined by the achievement of important predictors like gestational age, 24 hour urinary protein/dipstick, M P V : platelet ratio, fibrinogen, bilirubin level thresholds for high risk morbidity status. This also can be used in various prospective clinical trials to investigate the utility of medical interventions. This study has shown that it is possible to predict adverse maternal outcomes in women admitted to tertiary centres with the diagnosis of pre-eclampsia. Especially important is the recognition that gestational age at diagnosis (as reflected by admission) is an important and independent predictor of adverse maternal outcomes. Early-onset pre-eclampsia is more dangerous for both the woman and her foetus, and we have proposed the sub classification of pre-eclampsia to reflect this reality(65). A number of studies have suggested a different pathophysiology in women who develop pre-eclampsia early in term(67;228) and also studies have been done to show that gestational age is an important predictor of adverse maternal and perinatal outcome(84). These data confirm that gestational age at diagnosis (as reflected by admission) is an important and independent predictor of adverse maternal outcomes; Also , as shown in the results section, women with early onset of pre-eclampsia more frequently had risk factors and adverse pregnancy complications. The gestational age is not currently acknowledged in the 108 classification systems in general use(48;203;220). Early-onset pre-eclampsia is more dangerous for both the woman and her foetus, and the Vancouver research group has proposed the sub classification of pre-eclampsia to reflect this reality(65). The novel M P V : platelet ratio, a summary measure of platelet consumption, was designed to better reflect platelet consumption than either element of the ratio alone. A s platelets are consumed, the megakaryocytes in the marrow release immature large volume platelets, causing M P V to rise prior to the ongoing consumptive coagulopathy overwhelming the ability of the marrow to respond (at which time platelet counts fall). Therefore, this summary measure appeared to amplify the effects of platelet consumption as predicted, although we recognise that ratios are intrinsically more unstable mathematically compares to direct observations. This ratio would be easily introduced, as both elements are included in complete blood count analyses, and at no extra cost. Normal ranges for pregnancy need to be developed. It is possible that this ratio could have utility in other settings where platelet consumption is of concern (e.g. sepsis, immune thrombocytopoenic purpura). This ratio was predictive of the adverse maternal outcome for all the days o f admission, delivery to postnatal days as expected. Bi l i rubin was revealed as a similarly promising variable in this regard, although insufficient data were collected to permit its inclusion in the multivariable analysis. Proteinuria, which is included in the C H S definition of 'adverse features'(41), was associated with worse maternal outcomes. This was despite the fact that the proteinuria was measured by dipstick alone in almost 80% of cases. Although dipstick measurement of proteinuria is associated with false positive and false negative rates(41), it is cheap, and remains the standard 109 of care even in centres where protein: creatinine ratios are available. While the C H S does not recommend dipstick screening for proteinuria(41), these data show that dipstick assessment may indeed be useful in triaging women with either suspected or confirmed pre-eclampsia. In our study, urinary dipstick protein appears to parallel the rise in maternal risk in this patient population with severe pre-eclampsia. However, interestingly, 24hour urinary protein, one of the variable required for diagnosis of pre-eclampsia according to current classification guidelines was not available for 90% of our study population. So, the importance o f this variable can be assessed only by doing a prospective study, with systematic collection of the data in all the centres. The oldest and the most studied laboratory test, other than urinary protein determination, in the investigation of pre-eclampsia is the serum concentration of uric acid(104;l 18;119). Whether the cause of increased serum uric acid concentration in pre-eclampsia is secondary to tubular damage because of renal vasoconstriction and ischemia, or to a pure functional adaptation due to the well known hypovolaemia existing in this disease, is not known. Despite the large number of reports dealing with uric acid in women with diagnosed pre-eclampsia, few data are available with regard to the predictive value for this disease. Redman et al(104) have suggested that patients who subsequently develop pre-eclampsia have significantly higher levels of uric acid starting from 28 weeks o f gestation. Other findings suggest that uric acid rises significantly only in the week before delivery in-patients who develop pre-eclampsia. Finally, it has been reported that clinical proteinuria exceeding 300mg/24h, a late sign in the evolution of pre-eclampsia associated with poor perinatal prognosis, is usually preceded by reduced uric acid clearance and rising uric acid concentration(68;125). 110 Uric acid (n=375) however was not a good predictor of outcome in our study when, the analysis was done on the day of admission, delivery day or postnatal days. The data were skewed and hence it was difficult to make an interpretation. A number of studies have examined use of blood pressure as a predictor of adverse outcome in hypertensive disorders of pregnancy. The studies suggest that blood pressure is altered early in pregnancy in women who w i l l develop any hypertensive disorders(148;235). In our study, M A P was elevated at least by lOmmHg (p<0.001) in women who developed adverse maternal outcome compared to who did not for the day of admission, day of delivery, and postnatal days 1 and 2. The difference in M A P between those women who developed the combined adverse maternal outcome and those who did not probably reflects the peaks of blood pressure that occur despite the fact that blood pressure is the only component of the maternal syndrome of pre-eclampsia for which there is effective therapy (35;236). A s mentioned earlier, bilirubin is not tested often in either the screening for or prediction of pre-eclampsia. We included bilirubin in this study to test the predictability of this variable for maternal and perinatal morbidity based on its utility within the A P A C H E model(47). A P A C H E was developed for patients admitted to I C U , often with the systemic inflammatory response syndrome (SIRS). Bi l i rubin data were missing for 90% of the women from Vancouver and Ottawa. However it was collected systematically in the women from Oxford. Bil i rubin appeared 111 significant for the day of admission, delivery, and postnatal days. Also , there was a pattern of increased significance for the day of delivery and postnatal days. But only collecting the data systematically in all the centres by doing a prospective study we can draw conclusion. A s discussed previously, the albumin level is falls slightly during pregnancy to levels slightly lower than in non-pregnancy (100). The excessively decreased serum albumin of pre-eclampsia may result from decreased blood flow to the l iver( lOl) , hepatic necrosis(lOl), chronic inflammation and also from excess loss of protein from the kidney due to renal diseases(102-105), among other causes. A s shown below, it can also result from decreased renal function, which allows albumin to escape into the urine(102-105). The excess proteinuria is generally followed by the decrease in the level of plasma albumin. In our study, plasma albumin, was significant for the postnatal days. However this variable again was not collected frequently. Plasma albumin was done regularly at Oxford, however that information was missing for 90% of the Vancouver and Ottawa centre study population. During normal pregnancy, plasma fibrinogen concentration substantially increases and in pre-eclampsia there is a slight increase in plasma fibrinogen levels compared with normal pregnancy levels(142). Measurement of reduction of fibrinogen from the circulation with accentuation of fibrin split products in the circulation can be done serially to assess the coagulation system in pre-eclampsia. Hence, even though fibrinogen is not used that commonly in the prediction of complications in pre-eclampsia, we included fibrinogen in this study to test the predictability of 112 this variable, and to see how many centres are collecting this information. We included this for maternal and perinatal morbidity based on A P A C H E ( 4 7 ) . A s LNR was not available on a regular basis in the retrospective database we included this predictor marker in our future prospective study. Fibrinogen was lower for all the delivery and postnatal days. But, again, fibrinogen was done regularly for the women from Oxford and it was not done regularly for the other two centres. So, we were unable to draw a conclusion, as there is insufficient data. We need to do this by systematic prospective study. A number of studies shows the association of family history o f pre-eclampsia to development of pre-eclampsia (15;34;187;188). In our study we looked for any association of family history of pre-eclampsia to the development of adverse maternal outcome and the association was significant. Interestingly, having a personal history of pre-eclampsia was not predictive of adverse maternal outcome. This may be due to the error in collecting the information, or due to the effect o f family history, women remember it well , and have given that information, we do not know. However it is impossible to draw a conclusion. However, prospective study needs to be done to draw any firm conclusion. A s shown in the results section, treatment with anti-hypertensives, steroid and MgS04 was associated with adverse maternal outcome, which is to be expected. A s the mother is sick, she w i l l be treated with more number of medications. Previous studies have observed an increased risk of maternal morbidity in women aged 35 years and older when they deliver their first child 113 (162-167). However we did not see any association with adverse maternal outcome in our study and also parity did not show any relationship with morbidity. We deliberately excluded from the combined maternal outcome, placental abruption, which is recognized as a common cause of maternal morbidity in pre-eclampsia, as it is difficult to diagnose with certainty. The method of recognition of placental abruption differs from centre to centre and unit to unit and some unit in our region may diagnose entirely on clinical suspicion or abdominal pain. In view of these factors it would be impossible to ascertain i f cases gathered represented an accurate reflection of the incidence. But there were a significant number of women in Vancouver who appeared to have had placental abruption compared with other centres; this increased incidence was associated with an increased use of transfusion in Vancouver. Also there were a number of women (n=19) from Ottawa who received, > 50% O2 for more than 1 hour. This may be due to the management difference with regards to pre-eclampsia rather than due to the complication o f the disease by itself. In pre-eclampsia there is a decrease in the cardiac output. So, centres with aggressive fluid management might have ended up with fluid overload (in an attempt to reverse the reduced intravascular volume known to complicate pre-eclampsia), in turn leading to pulmonary oedema/respiratory distress, with resultant excessive requirement of 02. Also , there were more number of women (n=14) with third infused antihypertensives from Oxford compared with other centres. This again may be due to the difference in the management 114 practice with regards to pre-eclampsia rather than pre-eclampsia as a disease alone. Some of the clinicians who are comfortable with prolonging the pregnancy by expectant management might try to reduce the maternal blood pressure by adding a third infused antihypertensives compared with the clinicians with a trend for aggressive management who may want to deliver the patient at the first sign of elevation of blood pressure. Doing a prospective management can assess this; however, the surveillance itself might affect the results. Also , the contributions of symptoms like nausea, epigastric pain, headache and so on to the risk status of a pregnant patient is difficult to quantify, unlike the battery of tests presented here. Hence, it was decided to exclude maternal symptoms from this study, despite their inclusion in the C H S definition of 'adverse features' and recent evidence to support their utility in predicting adverse maternal outcomes(237;238). We recognise that they remain clinically influential and have potential importance in the surveillance of women with pre-eclampsia as the disease evolves, but did not feel that symptoms could be reliably cleaned through a retrospective chart review. However, they, and other factors such as family history, should be assessed for their utility in predicting adverse maternal outcomes in a prospective manner. In summary, the monitoring of pre-eclampsia pregnancies in this study was idiosyncratic within and between centres. Most variables were available in <80% of women, even 24h urinary protein estimation, upon which the firm diagnosis of pre-eclampsia is predicated, was measured in only 21% of cases. Therefore, we were underpowered to test whether or not variables such as fibrinogen and bilirubin, which are powerful predictors of death in adult intensive care unit (ICU) settings, when incorporated into predictive models(47-49;54;55), might have true utility in 115 the setting of pre-eclampsia. Predictive models (e.g. APACHE(47-49;54) and M O D S (55)) have been developed for patients with SIRS, which, as stated, pre-eclampsia resembles to a remarkable degree(22;36). Although these scores perform well when modified for defined populations(99), most have been generated to reflect pathophysiology in predominately geriatric populations(239). The A P A C H E score, specifically, did not perform well with eclampsia in the I C U setting, with the Glasgow coma scale outperforming A P A C H E II in predicting maternal mortality(50). This study has implications for the monitoring of women with pre-eclampsia, and argues for the development of a disease specific outcome prediction model, as has been suggested(239). There is international support for this approach. Such a prediction model could be developed using the combined adverse maternal outcome used in this study and should be tested against the current dichotomous definitions o f mild vs. severe' or 'with vs. without adverse features' using area under the receiver operator curve analyses(240). 5.1 Biases: The estimates of incidence of maternal outcome probably underestimates the true incidence as case ascertainment is unlikely to be complete, especially i f events occur outside the delivery suite and just before admission and also just after discharge or transfer to a different hospital and because of an inability to find notes of information. However, we used several measures to minimise this loss of ascertainment by collecting data as much as possible from all the sources available when information is incomplete or missing. Also , the inclusion of > 50% O2 for more 116 than 1 hour and third infusion of antihypertensives in the adverse maternal outcome may affect the true incidence of maternal complications, it is difficult to tease them out as they are also the complications seen in women with pre-eclampsia population, even though their incidence w i l l be influenced by management variations. 5.2 Limitations of the study: The data presented do reveal that some variables changed between admission and delivery, reflecting the need to monitor these women closely across a range of organ systems. At present, however, it is unclear which tests in which organ systems might provide the most utility in predicting the clinical course of any particular woman with pre-eclampsia. In the survey done by Magee and associates(241) in 1999-2000, of Canadian obstetricians, G P obstetrician members of the S O G C , midwife members of the S O G C , nephrologists, and a random sample of general internists (10%)(241), using a series of clinical scenarios, practitioners were asked to describe how they diagnose, evaluate, and manage women suffering from the hypertensive disorders of pregnancy. A finding of this survey was that the CHS(41) recommendations and the stated practice of Canadian practitioners are quite disparate. Some of the tests in the candidate predictor list (e.g. albumin) from our study are not graded by the C H S , but it is noted that in the presence of hypoalbuminaemia, the risk of pulmonary oedema is high. The one grade A recommendation made by the CHS(41) was the use of urinalysis by the laboratory, and it was not commonly done. In contrast, the C H S specifies that urinary dipstick testing, which is performed routinely, not be 117 done. The CHS(41) also recommends that tests of coagulation not be performed routinely. However, this is in the evaluation of women with gestational hypertension, not pre-eclampsia specifically, in whom the likelihood of surgery and/or disseminated intravascular coagulation is higher; the majority of Canadian practitioners perform coagulation studies, with 80% performing them at least once/week. The A S S H P does recommend coagulation studies in the setting of pre-e c l a m p s i a ^ ) . In summary, there is insufficient evidence available to guide the clinicians in recommending which tests to do, and how frequently, in the evaluation of women with pre-eclampsia. Individual practitioners may find that guidelines differ from their practice, but it is difficult to justify most of what we do. Hence, it is understandable and expected that, lab variables and investigations were not performed uniformly within the centre and between the centres. In this study, there were lot of variations with respect to the samples and management and outcome. We were unable to assess the influence o f medical intervention on the outcomes being tested, especially as there was within centre and between centre variability in the expectant management of pre-eclampsia remote from term. There was considerable heterogeneity in the care of women enrolled in this study. The study is based on groups that are ordinarily variable and, therefore, are representative samples of the disease as it presents in most centres; variations in case-mix and management reflect the real world and have made this study generalisable. 118 In light of shortcomings and also missing data points it is not surprising that this some of the tests were not as strong a risk reflector for major morbidity as others. A s this was a retrospective study, there were lot of missing information to draw a solid conclusion for the study. This study has shown that it is feasible to develop a model to predict adverse maternal outcomes in women admitted to tertiary centres with the diagnosis of pre-eclampsia. This study did not achieve that end, but provides evidence that the prospective development of such a model is possible. Given the limitations of the present study, we believe that this should be a prospective study that would ensure full documentation of the candidate variables at each epoch of interest. A s clinical care varies both within and between centres, prospective data collection in a number of centres would identify those variables that are both most useful in predicting the outcome and generalisable. Because this study included retrospectively collected data from a large cohort of women from different centres, this study population represents many varied phases o f the disease process. Also there appears to be some "good news" in the study about this disease. There is only one case of maternal death among the population studied and also there is a decrease in the incidence of eclampsia. This may be due to decline in the severity of pre-eclampsia seen at these tertiary centres as a result o f aggressive early identification and delivery before the florid deterioration in the disease process. 5.3 Benefits and limitations of scoring system: Indirect benefits are many. First o f all, research results can be presented much easier and results of clinical trials become more objective and reproducible. Secondly, scores can be used to 119 perform quality audits. Thirdly, younger physicians w i l l benefit from the educational effect of scoring systems. A direct effect is present i f a therapeutic or diagnostic procedure is induced or withdrawn or denied due to a score value. This may occur in triaging, initiation of procedures, discharges and termination of therapy. The scientific assessment and the practical use of score systems w i l l undoubtedly continue to improve. Scores can facilitate comparisons via classification of diseases or health status, enhances decision-making and thus quality management, and improves clinical research. The usefulness of scores in the individual patient is limited due to the fact that scores reduce many dimensions of a clinical situation into a single value. Therefore, a score may supplement but w i l l never replace clinical judgment, due to these limitations. 5.4 Lessons learnt from this: A predictive tool can be developed only after analysis of robust data. In prediction of pre-eclampsia, this requires collecting data prospectively using robust data entry and monitoring techniques to ensure complete data collection. We have refined the database for the prospective study to overcome data entry and analysis problems identified during the chart review. This database is now being piloted prospectively by vonDadelszen and associates in Vancouver. The study is feasible as planned, given the frequency of gestational hypertension and/or proteinuria and the combined adverse maternal outcome in the participating centres. Certain key predictive variables have been identified that are not routinely collected in women with pre-eclampsia (e.g. bilirubin); the findings of the retrospective review 120 justify their inclusion in the prospective study as part of routine clinical care. The retrospective data have helped us to determine the minimum frequency of data collection for the variables of interest for the prospective study. 5.5 Conclusion: In conclusion, we have shown that the women at greatest risk for adverse outcomes in pre-eclampsia are those at earlier gestational ages and those with greater dipstick proteinuria. Platelet consumption (reflected by M P V : platelet count ratio) also appears to be predictive. Also, bilirubin and fibrinogen appeared to be predictive though the sample size was inadequate to draw a conclusion. Although, only the clinical signs and laboratory profiles of women at the time of hospital admission, day of delivery, postnatal day 1 and 2 was done for this study, it is probable that similar assessment of the data on a recurring basis longitudinally during the care of the patient prospectively could be used to assess the pathogenesis of pre-eclampsia and as well as to enhance the patient care by anticipating complications. However, what is required is the comprehensive prospective evaluation of the association between all candidate predictor variables including maternal symptoms and adverse maternal outcomes. Only then w i l l we be able to identify those women with pre-eclampsia for whom delivery is clearly mandated and those most appropriately managed expectantly. 121 5.6 Future directions: A valid, sensitive, specific score, predictive of adverse outcomes within l w k of admission, w i l l provide, for the first time, an evidence base upon which to plan the management of pregnancies complicated by pre-eclampsia. The C I H R has recently funded the development phase of this project (Pre-eclampsia Integrated Estimate of R iSk [PIERS] model; PI: P von Dadelszen). 5.6.1 Goals: The goal of this proposal is the prospective development of a functional and practical predictive score to identify maternal risk in pre-eclampsia. We w i l l use the model that gives the best sensitivity and specificity. 5.6.2 Primary objectives: • Develop and validate a model, using maternal and foetal variables, predictive of a combined adverse maternal outcome occurring within one week o f hospital admission, in pregnancies complicated by suspected or confirmed pre-eclampsia (gestational hypertension and/or proteinuria, idiopathic gestational proteinuria, the H E L L P (haemolysis, elevated liver enzymes, and low platelets) syndrome, or eclampsia). We consider this to be a clinically relevant time frame as every week gained has an appreciably beneficial impact on foetal outcomes. 122 5.6.3 Secondary objectives: • Investigate models, using maternal and foetal variables, predictive of a combined adverse maternal outcome occurring within 48h (the timeline for antenatal steroid administration for foetal lung maturation), 2wk (the anticipated gain from expectant management remote from term), and throughout admission following hospitalisation admission, in pregnancies complicated by suspected or confirmed pre-eclampsia. • To explore factors affecting time to an adverse outcome (survival analysis). • To investigate the predictive score as a day-to-day burden of illness score (sensitivity to change). • To investigate predictors of foetal outcome • To investigate whether or not a model can be developed to predict a combined adverse fetal outcome. 5.6.4 Ongoing calibration: • The utility of using the score at different times (say, weekly) during a woman's course will be explored using this database. Every few years the severity score will require re-calibration as new indices of disease severity are introduced into clinical practice and are eligible for assessment for their inclusion in the score (e.g. plasminogen activator inhibitor (PAI) 1 :PAI2 ratio, tumour necrosis factor-a, and interleukin 6). 123 5.6.5 A computer-based scoring system for pre-eclampsia: A clinically relevant database that wi l l permit practitioners to enter data in real time and be given graphic outputs for individual variables and for the total score w i l l be developed in a subsequent study for which funding w i l l be sought. This w i l l be developed using a program, which can be loaded onto a Palm Pilot. This w i l l allow updated risk estimates of adverse maternal outcomes that take into account that day's score and also the rate of change in the score over the past 24h, 48h and since entry into the database. The commercial implications of this have led us to apply for a patent to protect the computer programme and its use. 5.6.6 Application to ongoing research studies: • This severity score w i l l be immediately useful for a number of studies currently underway or planned at B C Women's Hospital. A l l candidate variables are being collected prospectively for the women in the studies outlined below, so that the score, once derived, w i l l be incorporated into the study analyses. • First, given the efficacy and safety of recombinant human activated protein C (rhAPC) in SIRS (196), von Dadelszen et al reviewed the evidence for a role for A P C in the pathogenesis of pre-eclampsia (37). Pre-eclampsia is a proinflammatory and procoagulant state, and is a pregnancy-specific condition that mimics SIRS. r h A P C reduces mortality in patients with SIRS, and could potentially have a role as disease-modifying therapy in pre-eclampsia. To determine which patients would be offered rhAPC, the literature pertaining to foetal/neonatal outcomes for pre-eclampsia remote from term, transplacental transport of protein C, and pregnancy experience with the compound were reviewed. Sufficient data exist to support the use of r h A P C in phase II 124 clinical studies for women with either early-onset pre-eclampsia or severe and/or deteriorating postpartum disease. The severity score w i l l be an integral part of this study, and the courses of women in the severity score database w i l l provide historical controls for the women enrolled in the phase II trial. • The second study is a submission planned for March 2003. This application w i l l investigate, mechanistically, the maternal inflammatory response in pre-eclampsia. This study involves the investigation of potential infectious triggers for pre-eclampsia (242). Women in the study (those with early onset pre-eclampsia, late onset pre-eclampsia, normotensive IUGR, and normal control pregnancies) w i l l be compared for the presence of active cytomegalovirus and/or Chlamydia pneumoniae infection, and their worst daily severity score value w i l l be taken as an index of disease severity. We have shown that women with early-onset pre-eclampsia have higher titres of antibodies to these two pathogens than do women with late-onset disease, normotensive I U G R , and matched normal pregnancy (243). It is postulated that reactivation of chronic maternal infection during pregnancy could lead to pre-eclampsia, as pre-eclampsia is linked with the later development of cardiovascular disease(244;245), which in turn is linked to the same infectious agents(246). • The third study is a blinded assessment of placental histopathology from women delivered of pregnancies complicated by pre-eclampsia and/or I U G R and from women with normal pregnancy outcomes. Dr Fergall Magee, perinatal pathologist, U B C , w i l l use the severity score to gain a global assessment of maternal and foetal disease severity, and correlate that with 125 the development of placental grading system. This w i l l aid in understanding the clinico-pathological correlations of the index pregnancy and assist in planning for future pregnancies. 126 References: Reference List: (1) Department of Health. Report on confidential enquiries into maternal deaths in the United Kingdom. Report H M S O . 1989. London. (2) Department o f Health. Report on confidential enquiries into maternal deaths in the United Kingdom. Report H M S O . 1996. London. (3) Mark A Brown, Marshall D Lindenheimer, Michael de Sweit, Andre Van Assche, Jean Marie Moutquin. The classification and diagnosis of the hypertensive disorders of pregnancy: Statement from the international society for the study of hypertension in pregnancy (ISSHP). Hypertens Pregn 2001; 20(1):9-14. (4) Department o f Health. Report on confidential enquiries into maternal deaths in the United Kingdom. H M S O . 1999. London. (5) Khong T Y , de W o l f F , Robertson W B BI . Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-gestational age infants. B r J Obstet Gynaecol 1986; 93:1049-1059. (6) Roberts J M , Taylor R N , Musc i TJ , Rodgers G M , Hubel C A , McLaughl in M K . Pre-eclampsia: an endothelial cell disorder. A m J Obstet Gynaecol 1989; 161:1200-1204. (7) Vince G S , Starkey P M , Austgulen R, Kwiatkowski D , Redman C W G . Interleukin-6, tumour necrosis factor and soluble tumour necrosis factor receptors in women with pre-eclampsia. B r J Obstet Gynaecol 1995; 102:20-25. (8) Redman C W . Current topic:Pre-eclampsia and the placenta. Placenta 1991; 12:301-308. (9) Robertson W B . The pathological response o f the vessels o f the placental bed to hypertensive pregnancy. J Pathol Bacteriol 1967; 93:581-592. 127 (10) Susan J Fisher J M R . Defects in placentation and placental perfusion. Aetiology of pre-eclampsia 1999;377-394. (11) Brosens IA, Robertson W B , Dixon H G . The role of spiral arteries in the pathogenesis of pre-eclampsia. Obstet Gynaecol arm 1972; 1:177-191. (12) Gerretson G . Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsia and growth retardation. B r J Obstet Gynaecol 1981; 88:876-881. (13) Zhou Y . Pre-eclampsia is associated with failure of human cytotrophoblasts to mimic a vascular adhesion phenotype.One cause of defective endovascular invasion in this syndrome? J C l i n Invest 1997;2152-2164. (14) de Groot C J TR. New insights into the etilogy of pre-eclampsia. A n n 1993; 25:243-249. (15) Chesley L , Cooper D . Genetics of hypertension in pregnancy:Possible single gene control of pre-eclampsia and eclampsia in the descendents of eclampstic women. Br J Obstet Gynaecol 1986; 93:898-908. (16) Scott J. Immunologic aspects of pre-eclampsia. A m J Obstet Gynaecol 1976; 125:418-425. (17) de Groot C J , Taylor R N . New insights into the aetiology o f pre-eclampsia. A n n M e d 1993; 25(3):243-249. (18) Resnick N . Hemodynamic forces are complex regulators of endothelial expression. F A S E B J 1995; 9:874-882. (19) Namik i A . Hypoxia induces vascular endothelial growth factor in cultured human endothelial cells. J B i o l C H E M 1995; 270:31189-31195. (20) Zeek P M . Vascular changes in the decidua associated with eclamptogenic. A m J C l i n Pathol 1950; 1099-1109. (21) Berk B C . Vaso constriction: a new activity for platelet derived growth factor. Science 1986; 232:87-90. 128 (22) Redman C W G , Sacks GP , Sargent IL. Pre-eclampsia: an excessive maternal inflammatory response to pregnancy. A m J Obstet Gynaecol 1999; 180:499-506. (23) Sacks G P , Studena K , Sargent IL, Redman C W G . Normal pregnancy and pre-eclampsia both produce inflammatory changes in peripheral blood leukocytes akin to those of sepsis. A m J Obstet Gynaecol 1998; 179:80-86. (24) Conrad K P , Miles T M , Benyo D F . Circulating levels of immunoreactive cytokines in women with pre-eclampsia. A m J Reprod Immunol 1998; 40:102-111. (25) Benyo D F , Smarason A , Redman C W G , Sims C, Conrad K P . Expression of imfiammatory cytokines in placentas from women with pre-eclampsia. J C l i n Endocrinol Metab2001; 86:2505-2512. (26) Greer IA, Lya l l F, Perera T, Boswell F, Macara L M . Increased concentration of cytokines interleukin-6 and Interleukin -1 receptor antagonist in plasma of women with pre-eclampsia, a mechanism for endothelial dysfunction? Obstet Gynaecol 1994; 84(937):940. (27) El l i s J, Wennerholm U B , Bengtsson A , Li l ja H , Petterson A , Sultan B et al. Levels of dimethyl arginines and cytokines in mild and severe pre-eclampsia. Acta Obstet Gynaecol Scand 2001; 80:602-608. (28) Magee L A vonDadelszen P. Hypertension and pregnancy. Part I. Can J Diagn 1999; 16(June)(1999):75-80. (29) Bulf in M J L P . Problems associated with toxemia in twin pregnancies. A m J Obstet Gynaecol 1957;37-73. (30) Greer IA. Thrombosis in pregnancy: maternal and foetal issues. Lancet 1999; 353:1258-1265. (31) Fenakel K , Fenakel G , Appelman Z , Lurie S, Katz Z , Shoham. Nifedipine in the treatment of severe pre-eclampsia. Obstet Gynaecol 1991; 77(3):331-337. 129 (32) von Dadelszen P, Watson R W G , Noorwali F , Marshall JC , Parodo J, Farine D et al. Maternal neutrophil apoptosis in normal pregnancy, pre-eclampsia, and normotensive intrauterine growth restriction. A m J Obstet Gynaecol 1999; 181:408-414. (33) von Dadelszen P. Activation of maternal peripheral blood leukocytes in pre-eclampsia. Oriel College, Oxford, 2000. (34) Roberts J M R C . Pre-eclampsia: more than pregnancy-induced hypertension. Lancet 1993; 341(8858):1447-1451. (35) Magee L A , Ornstein MP,von Dadelszen P. Fortnightly review: management of hypertension in pregnancy. B r M e d J 1999; 318:1332-1336. (36) von Dadelszen P, Magee L A , Marshall JR, Rotstein O D . The maternal syndrome of pre-eclampsia: a forme fruste of the systemic inflammatory response syndrome. Sepsis 2000; 4:43-47. (37) von Dadelszen P, Magee L A , Lee S K , Stewart S D , Simone C, Koren G et al. Activated protein C in normal human pregnancy and pregnancies complicated by severe pre-eclampsia: a therapeutic opportunity? Crit Care M e d 2002; 30:1883-1892. (38) Stevenson D K , Wright L L , Lemons J A , Korones S B , Papile L A et al. Very low birth weight outcomes of the National Institute of Chi ld Health and Human Development Neonatal Research Network, January 1993 through December 1994. A m J Obstet Gynaecol 1998; 179:1632-1639. (39) Odendaal H J , Pattinson R C , Barn R, Grove D , Kotze TJvW. Aggressive versus expectant management for patients with severe pre-eclampsia between 28-34 weeks' gestation: a randomised controlled trial. Obstet Gynaecol 1991; 76:1070-1075. (40) Olah K S , Redman C W , and Gee H . Management of severe, early pre-eclampsia: is conservative management justified? Eur J Obstet Gynaecol Reprod B i o l 1993; 51(3): 175-180. 130 (41) Helewa M E , Burrows R F , Smith J, Will iams K , Brain P, Rabkin SW. Report of the Canadian Hypertension Society Consensus Conference: 1: definitions, evaluation and classifications of hypertensive disorders in pregnancy. Can M e d Assoc J 1997; 157:715-725. (42) Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. A m J Obstet Gynaecol 183, S1-S22. 2000. (43) Brown M A , Hague W M , Higgins J, Lowe S, M c C o w a n L , Oats J et al. The detection, investigation and management of hypertension in pregnancy: full consensus statement. Aust N Z J Obstet Gynaecol 2000; 40(2)( ; :):139-155. (44) Brown M A , Lindheimer M D , de Swiet M , V a n Assche A , Moutquin J - M . The classification and diagnosis of the hypertensive disorders of pregnancy: statement from the International Society for the Study of Hypertension in Pregnancy (ISSHP). Hypertens Pregn 2001; 20(1)(1):9-14. (45) Douglas K A , Redman C W G . Eclampsia in the United Kingdom. B r M e d J 1994; 309(6966)(.):1395-1400. (46) Douglas Richardson W O T - M S K L . Section 2: Measurement Risk adjustment for quality improvement. Paediatrics 1999; 103(l):255-265. (47) Knaus W , Wagner D , Draper E . A P A C H E III study design: analytic plan for evaluation of severity and outcome in intensive care unit patients. Implications. Crit Care M e d 1989; 17(12 Pt 2):S219-S221. (48) Wi l l i am A Knaus, Elizabeth A Draper, Douglas P, Wagner, Jack E Zimmerman. A P A C H E II: A severity of disease classification system. Crit Care M e d 1985; 13:818-829. (49) Wagner D D E K W . A P A C H E III study design: analytic plan for evaluation of severity and outcome in intensive care unit patients. Development of A P A C H E III. Crit Care Med 1989; 17(12 Pt 2):S199-S203. 131 (50) Bhagwanjee S, Paruk F , Moodley J, Muckart D J . Intensive care unit morbidity and mortality from eclampsia: an evaluation of the Acute Physiology and Chronic Health Evaluation II score and the Glasgow Coma Scale score. Crit Care M e d 2000; 28(1):120-124. (51) Crawford JS. The Apgar Scoring system. Dev M e d Chi ld Neurol 1962; 4(441):444. (52) Murray D J , Boulet JR, Kras JF, Woodhouse J A , Cox T, McAll i s te r JD. Acute care skills in anaesthesia practice: a simulation-based resident performance assessment. Anaesthesiology 2004; 101 (5): 1084-1095. (53) Smulders Y M , Spijkerman A M , Kostense PJ, Bouter L M , Stehouwer C D . [Old and new scoring systems for assessing cardiovascular risks: problems with the validity, the precision and the homogenicity of the risk categories]. Ned Tijdschr Geneeskd 2004; 148(50):2480-2484. (54) Knaus W A , Zimmerman JE, Wagner D P , Draper E A , Lawrence D E . APACHE-acu te physiology and chronic health evaluation: a physiologically based classification system. Crit Care M e d 1981; 8:591-597. (55) Marshall JC , Cook D J , Christou N V , Bernard G R , Sprung C L , Sibbald W J . Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care M e d 1995; 23(10): 1638-1652. (56) Reiter A , Mauritz W , Jordan B , Lang T, Polz l A , Pelinka L et al. Improving risk adjustment in critically i l l trauma patients: the T R I S S - S A P S Score. J Trauma 2004; 57(2):375-380. (57) Hjortso E , Buch T, Ryding J, Lundstrom K , Bartram P, Dragsted L et al. The nursing care recording system. A preliminary study o f a system for assessment o f nursing care demands in the I C U . Acta Anaesthesiol Scand 1992; 36(7):610-614. (58) Pollack M M , Ruttimann U E , Getson PR. Pediatric risk of mortality(PRISM)score. Crit Care M e d 1988; 16:1110-1116. 132 (59) Murray M.Pollack, Urs E.Ruttimann, Pamela R Getson. Pediatric risk of mortality (PRISM) score. Crit Care M e d 1988; 16:1110-1116. (60) Douglas K Richardson, John D Corcoran, Gabriel J Escobar, Shoo K Lee. S N A P II and S N A P P E II: Simplified newborn illness sverity and mortality risk scores. J Paediatr 2001; 138:92-100. (61) Robson W P . A n evaluation of the evidence base related to critical care outreach teams—2 years on from Comprehensive Critical Care. Intensive Crit Care Nurs 2002; 18(4):211-218. (62) Oortwijn W J , Vondeling H , van Barneveld T, van Vugt C, Bouter L M . Priority setting for health technology assessment in The Netherlands: principles and practice. Health Pol icy 2002; 62(3):227-242. (63) Y e h TS, Pollack M M , Ruttimann U E et al. Validation of a physiologic stability index for use in critically i l l infants and children. Paeditr res 1984; 18:445. (64) Vincent, Jean-Louis, de Mendonca Arnaldo, Cantraine, Francis, Moreno et al. Use of S O F A score to assess the incidence of organ dysfunction/failure in intensive care units: Results of a multicenter, prospective study. Crit Care M e d 1998; 26(11):1793-1800. (65) von Dadelszen P, Magee L A , Roberts J M . Subclassification o f pre-eclampsia. Hypertens Pregn 2003;22:143-148. (66) M a c K a y A P , Berg C J , Atrash H K . Pregnancy-related mortality from pre-eclampsia and eclampsia. Obstet Gynaecol 2001; 97:533-538. (67) Opsjln S L , Wathen N C , Tingulstad S, Wiedswang G , Sundan A , Waage A et al. T N F , IL-1 and IL-6 in normal human pregnancy. A m J Obstet Gynaecol 1993; 169:397-404. (68) Sibai B M , Mercer B M , Schiff E , Friedman S A . Aggressive versus expectant management of severe pre-eclampsia at 28 to 32 weeks' gestation: a randomised controlled trial. A m J Obstet Gynaecol 1994; 171:818-822. 133 (69) Fries JF, Ehrlich G E . Some conceptual considerations. Prognosis: Contemporary outcomes of disease. Bowie (Maryland) U S : Charles press, 1981: 5-8. (70) Hallak M . Hypertension in pregnancy in High risk pregnancy: management options. London: W.B.Saunders, 1998: 639-663. (71) Weinstein L . Pre-eclampsia-eclampsia with haemolysis, elevated liver enzymes, and thrombocytopenia. Obstet Gynaecol 1985; 66:657-660. (72) Chesley L C , Annitto JE, Jarvis D G . Pregnancy in the patient with hypertensive disease. A m J Obstet Gynaecol 1947; 53:372-381. (73) Roberts J M , Perloff D L . Hypertension and the obstetrician-gynaecologist. A m J Obstet Gynaecol 1977; 127(316). (74) Sibai B M , Rodriguez J. Pre-eclampsia: Diagnosis and Management. 1999: 871-879. (75) Fielding C F . Obstetric aspects in women with histories of renal disease. Acta Obstet Gynaecol Scand 1969; 48(suppl 2): 1-43. (76) Chesley L C . Hypertensive disorders in pregnancy. New York: Appleton Century Crofts, 1978. (77) Long P, Oats J. Pre-eclampsia in twin pregnancy-severity and pathogenesis . Aust N Z J Obstet Gynaecol 1987; 27(1). (78) Sibai M , E l Nazer A , Gonzolez Ruiz A . Severe pre-eclampsia and eclampsia in young primigravidas:subsequent pregnancy outcome and remote prognosis. A m J Obstet Gynaecol 1986; 1555(1011). (79) Department of Health. Why mother's die. Report on confidential enquieries into maternal death in the united kingdom 1994-1996. 1st. 1998. London. (80) Fitzpatrick C, Halligan A , McKenna P, Coughlan B M , Darling M R , Phelan D . Near miss maternal mortality ( N M M ) . Ir M e d J 1992; 85(1):37. 134 (81) Stones W , L i m W , A l Azzawi F, K e l l y M . A n investigation of maternal morbidity with identification of life-threatening 'near miss' episodes. Health Trends 1991; 23(1):13-15. (82) Creighton SB. 'Near-miss' obstetric enquiry. J Obstet Gynaecol 1997; 17(l):26-29. (83) Milhol land A V , Wheeler SG , Heieck JJ. Medical assessment by a Delphi group opinion technic. N Engl J M e d 1973; 14;288(24):1272-1275. (84) Audibert A , Friedman S, Frangieh A , Sibai B . Clinical utility of strict diagnositic criteria for the H E L L P (haemolysis, elevated liver enzymes, and low platelets) syndrome. A m J Obstet Gynaecol 1996; 175(2):460-464. (85) Shukla P K , Sharma D , Mandal R K . Serum lactate dehydrogenase in detecting liver damage associated with pre-eclampsia. B r J Obstet Gynaecol 1978; 85:40-42. (86) Davies M H . Acute liver disease with encephalopathy and renal failure in late pregnancy and the early puerperium: a study of fourteen patients. B r J Obstet Gynaecol 1980; 87:1005. (87) K i l l i a m A , Dillard S, Patton R, Pederson P. Pregnancy induced hypertension complicated by acute liver disease and disseminated intravascular coagulation. A m J Obstet Gynaecol 1975; 123:823-828. (88) Long R G , Sheuer PJ , Sherlock S. Pre-eclampsia presenting with deep jaundice. J O f C l i n Pathol 1977; 30:212-215. (89) Sheehan H L , Lynch JB. Pathology of toxemia in pregnancy. London: Churchill-Livingstone, 1973: 328-330. (90) De Boer K , Buller H R , Ten Cate JW, Treffers P E . Coagulation studies in the syndrome o f haemolysis, elevated liver enzymes and low platelets. B r J Obstet Gynaecol 1991; 98:42-47. (91) Sibai B M . The H E L L P Syndrome(haemolysis, elevated liver enzymes and low platelets):much ado about nothing? A m J Obstet Gynaecol 1990; 162:311-316. 135 (92) Pritchard J A , Cunningham F G , Mason R A . Coagulation changes in eclampsia:their frequency and pathogenesis. A m J Obstet Gynaecol 1976; 124:855-864. (93) Cunningham F G . Liver disease complicating pregnancy. In: Norwalk C T , editor. Wil l iams Obstetrics. Appleton and Lange, 1993. (94) Knox T A , Olans L B . Liver disease in pregnancy. N Engl J M e d 1996; 335:569-576. (95) Sibai B . Maternal and perinatal outcome associated with syndrome of haemolysis, elevated liver enzymes and low platelets in severe pre-eclampsia and eclampsia. A m J Obstet Gynaecol 1986; 155:501-509. (96) Martin J. Pregnancy complicated by pre-elampsia/eclampsia with syndrome of haemolysis, elevated liver enzymes, and low platelet countHow rapid is postpartum recovery? Obstet Gynaecol 1990; 76:737-741. (97) V a l A Catenzarite, Steve M Steinberg, Coleman A Mosley, Charles F Landers, Larry M Cousins, Jack M Scneider. Severe pre-eclampsia with fulminant and extreme elevation of A S T and L D H levels high risk for maternal death. A m J Perinatol 1995; 12:310-313. (98) James N.Mart in , Warren L , M a y E F , Magann, D o m A , Terrone. Early risk assessment of severe pre-eclampsia: Admission battery of symptoms and laboratory tests to predict likelihood of subsequent significant maternal morbidity. A m J Obstet Gynaecol 1999; 180. (99) Jacobs S, Zuleika M , Mphansa T. The Multiple Organ Dysfunction Score as a descriptor of patient outcome in septic shock compared with two other scoring systems. Crit Care M e d 1999;27(4):741-744. (100) Makuyana D , Mahomed K , Shukusho F D , Majoko F. Liver and kidney function tests in normal and pre-eclamptic gestation—a comparison with non-gestational reference values. Cent Afr J M e d 2002; 48(5-6):55-59. 136 (101) Gojnic M , Petkovic S, Papic M , Mostic T, Jeremic K , Vilendecic Z et al. Plasma albumin level as an indicator of severity of pre-eclampsia. C l i n Exp Obstet Gynaecol 2004; 31(3):209-210. (102) Butler N R , Bonham D G . Perinatal mortality. Edinburgh: E & S Livingstone, 1963: 87-100. (103) Naeye R L , Friedman E A . Causes of perinatal death associated with gestational hypertension and proteinuria. A m J Obstet Gynaecol 1979; 133:8-10. (104) Redman C W G , Bei l in L J , Bonnar J. Renal function in pre-eclampsia. J C l i n Pathol 1976; 10:91-104. (105) Fisher K A , Ahuja S, Luger A , Spargo B , Lindheimer M . Nephrotic proteinuria with pre-eclampsia. A m J Obstet Gynaecol 1977; 129:643-646. (106) Gaber L W . Renal pathology in pre-eclampsia. C l i n Obstet Gynaecol 1987; 4:971-995. (107) Imbasciati E , Ponticelli C . Pregnancy and renal disease:predictors for foetal and maternal outcome. A m J of Nephrol 1991; 11:353-362. (108) Friedman S A , Taylor R N , Roberts J M . Pathophysiology of pre-eclampsia. C l i n Perinatal 1991; 18:661-682. (109) Spargo B H , Lichtig C, Luger A M , Ka tz l , Lindheimer M D . The renal lesion in pre-eclampsia. In: Ka tz l , Lindheimer M D , editors. Hypertension in pregnancy. New york: John Weiley & Sons, 1976: 129-137. (110) Davey D A , MacGi l l ivray I. The classification and definition of the hypertensive disorders of pregnancy. A m J Obstet Gynaecol 1988; 158:892-898. (111) National highblood pressure education program working group report on high blood pressure in pregnancy. A m J Obstet Gynaecol 1990; 163:1689-1712. (112) Hypertension in pregnancy. A C O G 1996; Tech bulletin(no219). 137 (113) Sibai B . Pitfalls in diagnosis and management of pre-eclampsia. A m J Obstet Gynaecol 1988; 159:1-5. (114) Kuo V S . Proteinuria and its assessment in normal and hypertensive disorders of pregnancy. A m J Obstet Gynaecol 1992; 167:723-728. (115) National High blood pressure Education Working Group. High blood pressure during pregnancy. A m J Obstet Gynaecol 1991; 163:1689-1712. (116) Nisel l H . Blood pressure and renal function seven years after pregnancy complicated by hypertension. B r J Obstet Gynaecol 1995; 102:876-881. (117) Management of pre-eclampsia. A C O G 1986; Tech bulletin(no 91). (118) Chesley L C , Wil l iams L O . Renal glomerular and tubular functions in relation to the hyperuricemia of pre-eclampsia and eclampsia. A m J Obstet Gynaecol 1945; 50:367-375. (119) Stander H J , Duncan E E , Sisson. W E B u l l John Hopkins Hosp 1925; 36:411. (120) Chesley L C . Ibid 1939; 38:430. (121) Lancet M , Fisher I.L. J Obstet Gynaecol B r Emp 1956; 63:116. (122) McFarlane, C . N . Ibid 1963; 70:63. (123) H.E.Fadel, Sabour, M.Mahran, D.Sei f E l D in , M.N.El .Maha l l av i . Serum Uric acid in pre-eclampsia and eclampsia. 12-23. 1964. (124) A.Anasuya and Leela Raman. Significance of hypeuricemia in pre-eclampsia. Indian Journal O f Medical Research 1979; 70:767-770. (125) Redman C W G , Bei l in L J , Bonnar J, Wilkinson R H . Plasma urate measurements in predicting foetal death in hypertensive pregnancy. Lancet 1976; i : 1370-1373. 138 (126) Moutquin J M . Report of the Canadian Hypertension Society Consensus Conference: Non pharmacological management and prevention of hypertensive disorders in pregnancy. Can M e d Assoc J 1997; 157:907-949. (127) John C.Hauth and F.Gary Cunningham. Pre-eclampsia-eclampsia. 1998: 169-198. (128) Department of Health, Welsch Office. Report on confidential enquiries in to maternal death in the United Kingdom. H M S O 1994;22-33. (129) Anderson G D , Sibai M . Hypertension in pregnancy: Obstetrics normal and problem pregnancies. New York: Church h i l l and Livingstone, 1998. (130) Belfort M A . The oxygen consumption/Oxygen delivery curve in severe pre-eclampsia:evidence for a fixed oxygen extraction state. A m J Obstet Gynaecol 1993; 169:1448. (131) Schiff E M , Friedman S A M , Kao L R , Sibai B M M . The importance of urinary protein excretion during conservative management of severe pre-eclampsia. American Journal of Obstetrics and Gynaecology 1996; 175(5): 1313-1316. (132) M c K a y D G . Chronic intravascular coagulation in normal pregnancy and pre-eclampsia. Contr Nephrol 1983;25:108-113. (133) Redman C W G , Bei l in L J , Bonnar J. Early platelet consumption in pre-eclampsia. B r M e d J 1978; i:467-469. (134) Giles C. Intravascular coagulation in gestational hypertension and pre-eclampsia: The value of haematological screeining tests. C l i n Lab Haemat 1982; 4:351-358. (135) Pritchard J A , Weisman R Jr, Rtanoff O D , Vosburgh G . Intravascular haemolysis, thrombocytopenia and other hematological abnormalities associated with severe toxemia of pregnancy. N Engl J M e d 1954; 250:89-98. (136) Wallenburg H C S . Changes in the coagulation system and platelets in pregnancy-induced hypertension and pre-eclmpsia. In: Sharp F, Symonds E M , editors. Hypertension in 139 pregnancy proceedings of 16th study group on R C O G . Ithaka. New York: Perinatology press, 1986: 227-248. (137) Pavel Hamet M P . Hypertension in pregnancy: new reccomendation for management. Can M e d Assoc J, Sep 15 1997; 157(6):709-1254. (138) National high blood pressure education program working group report on high blood pressure in pregnancy. A m J Obstet Gynaecol 1990; 163:1689-1712. (139) Walker JJ, Cameron A D , Bjornsson S, Singer C R , Fraser C. Can platelet volume predict progressive hypertensive disease in pregnancy? A m J Obstet Gynaecol 1989; 161:676-679. (140) Ahmed Y , Van Iddekinge B , Paul C et al. Retrospective analysis of platelet numbers and volumes in normal pregnancy and in pre-eclampsia. B r J Obstet Gynaecol 1993; 100:216-220. (141) Hutt R, Ogunniyi SO, Sullivan M H F , etal. Increased platelet volume and aggregation precede the onset of pre-eclampsia. Obstet Gynaecol 1994; 83:146-149. (142) Davies J A , Prentice C R M . Coagulation changes in pregnancy induced hypertension and growth retardation. In: Greer IA, Turpie A G G , Forbes C D , editors. Hemostasis and thrombosis in Obstetrics and Gynaecology. London: Chapman an Ha l l , 1992: 143-162. (143) Easterling TR. Maternal hemodynamics in normal and pre-eclampstic pregnancies; a longitudinal study. Obstet Gynaecol 1990; 76:1061. (144) Mabie W C , Ratts T E , Sibai B M . The central hemodynamics o f severe pre-eclampsia. A m J Obstet Gynaecol 1989; 161:1443-1448. (145) Lang R M , Pridijian G , Feldman T, Neumann A , Lindenheimer M D , Borow K M . Left ventricular mechanics in pre-eclampsia. A m Heart journal 1991; 121:1768-1775. (146) Kuzniar J, Piela A , Skret A , Szmigiel Z B , Zaczec T. Echocardigraphic estimation of hemodynamics in hypertensive pregnancy. A m J Obstet Gynaecol 1982; 144:430-437. 140 (147) Management of pre-eclampsia. A G O G , Tech Bulletin feb 1986; 91. (148) Moutquin J M . A prospective study of blood pressure in pregnancy: prediction of pre-eclampsia. A m J Obstet Gynaecol 1985; 151(2): 191-196. (149) Gallery E D . Predicting the development of pregnancy associated hypertension. The place of standardised blood pressure measurement. Lancet 1977; I(8025):1273-1275. (150) Mac gillivray I. Blood pressure survey in pregnancy. C l i n Sci 1969; 37(2):395-407. (151) Page E W . Influence of blood pressure changes with and without proteinuria with outcome of pregnancy. A m J Obstet Gynaecol 1976; 126(7):821-833. (152) Pitfalls in diagnosis and management of pre-eclampsia. A m J Obstet Gynaecol 1988; 159:1-5. (153) Tervila L . Estimation of gestosis of pregnancy(EPH-gestosis). Acta Obstet Gynaecol Scand 1973; 52(3):235-243. (154) Fisher C M . The ascendency of diastolic blood pressure over systolic. Lancet 1985; 2(8468):1349-1350. (155) Rippman ET . Gestosis of late pregnancy, Nomenclature and scoring. Gynaecologia 1968; 165(l):12-20. (156) Browne FJ . Pregnancy in the patient with chronic hypertension. J Obstet Gynaecol B r Emp 1942;49:1-17. (157) Redman C W G . Reversed diurnal blood pressure rhythm in hypertensive pregnancies. C l i n Scie M o l M e d 1976; 5L687S. (158) Silver H M . Comparison of total blood volume in normal, pre-eclamptic, and non proteinuric gestational hypertensive pregnancy by simultaneous measurement of red blood cell and plasma volumes. A m J Obstet Gynaecol 1998; 179. 141 (159) Eden T W . Eclampsia: a commentary on the reports presented to the British Congress of Obstetrics and Gynaecology. J Obstet Gynaecol B r Emp 1922; 29:386-401. (160) Sibai B M , McCubb in J H , Anderson G D , Lipschitz J, Dilts P V . Eclampsia I. Observations from 67 recent cases. Obstet Gynaecol 1981; 58:609-613. (161) Redman C W . Treatment of hypertension in pregnancy. Kidney Int 1980; 18(2):267-278. (162) Hansen JP. Older maternal age and pregnancy outcome: a review of the literature. Obstet Gynaecol Surv 1986; 41(11):726-742. (163) Abdu l -Az iz A l Mulh im, Ade l A b u Heija, Fatia A l Jamma, E l Harith A , E l Harith. Pre-eclampsia: Maternal risk factors and perinatal outcome. Foetal Diagnosis and Therapy 2003; 18:275-280. (164) Vol lman R F . Rates of toxemia by age and parity. In Die Spat Gestose 1970. (165) Saftlas A F , Olson D R , Franks A L , Atrash H K , Pokras R. Epidemiology of pre-eclampsia and eclampsia in the United States, 1979-1986. A m J Obstet Gynaecol 1990; 163(2):460-465. (166) Hay D M B J J . A study of the obstetric performance of the adolescent Jamaican primigravida. A m J Obstet Gynaecol 1973; 116(l):34-38. (167) Davies A M , Czaczkes JW, Sadovsky E , et al. Toxemia of pregnancy in Jerusalem: I. Epidemiological studies of a total community. Isr J M e d Sci 1970; 253(6). (168) Eskenazie B , Fenster L , Sydney S. A multivariate analysis of risk factors for pre-eclampsia. J A M A 1991; 266:237. (169) Davies A G . Geographical epidemiology o f the toxemias o f pregnancy. Springfield: Charles C Thomas, 1971. (170) Klonoff-Cohen H , Edelstein S, Savitz D . Cigarette smoking and pre-eclampsia. Obstet Gynaecol 1993; 81(4):541-544. 142 (171) Underwood P B , Kesler K F , O'Lane J M , Callagan D A . Parental smoking empirically related to pregnancy outcome. Obstet Gynaecol 1967; 29(1): 1-8. (172) Samuels P, Steinfeld JD, Braitman L E , Rhoa M F , Cines D B , McCrae K R . Plasma concentration of endothelin-1 in women with cocaine-associated pregnancy complications. A m J Obstet Gynaecol 1993; 168(2):528-533. (173) Russell C S , Taylor R, Law C E . Smoking in pregnancy, maternal blood pressure, pregnancy outcome, baby weight and growth, and other related factors. A prospective study. B r J Prev Soc M e d 1968; 22(3): 119-126. (174) Towers C V , Pircon R A , Nageotte M P , Porto M , Garite TJ . Cocaine intoxication presenting as pre-eclampsia and eclampsia. Obstet Gynaecol 1993; 81(4):545-547. (175) Stone J, Lockwood C, Berkowitz G , Alvarez M , Lapinski R, Berkowitz R. Risk factors for severe pre-eclampsia. Obstet Gynaecol 1994; 83:357. (176) Wolfe H M , Zador IE, Gross T L , Martier SS, Sokol R J . The clinical utility of maternal body mass index in pregnancy. A m J Obstet Gynaecol 1991; 164(5 Pt 1):1306-1310. (177) Sibai B M , Gordon T, Thorn E , Caritis S N , Klebanoff M , McNel l i s D et al. Risk factors for pre-eclampsia in healthy nulliparous women: a prospective multicenter study. The National Institute of Chi ld Health and Human Development Network o f Maternal-Foetal Medicine Units. A m J Obstet Gynaecol 1995; 172(2 Pt l):642-648. (178) Ness R B , Markovic N , Bass D , Harger G , Roberts J M . Family history o f hypertension, heart disease, and stroke among women who develop hypertension in pregnancy. Obstet Gynaecol 2003; 102(6): 1366-1371. (179) Roberts J M . Pregnancy Related Hypertension. 1999: 833-871. (180) MacGi l l ivray I. Some observations on the incidence of pre-eclampsia. J Obstet Gynaecol B r Emp 1958; 65(4):536-539. 143 (181) Robillard P Y , Hulsey T C , Alexander G R , Keenan A , de Caunes F , Papiernik E . Paternity patterns and risk of pre-eclampsia in the last pregnancy in multiparae. J Reprod Immunol 1993; 24(1):1-12. (182) Ikedife D . Eclampsia in multipara. B r M e d J 1980; 280(6219):985-986. (183) Chng P K . Occurrence of pre-eclampsia in pregnancies to three husbands. Case report. B r J Obstet Gynaecol 1982; 89(10):862-863. (184) Grzywacz M , Ronin-Walknowska E . [The significance of H L A - G for the miscarriages]. M e d Wieku Rozwoj 2003; 7(3 Suppl 1):105-114. (185) Niittyvuopio R, Juvonen E , Kaaja R, Oksanen K , Hallman H , Timonen T et al. Pregnancy in essential thrombocythaemia: experience with 40 pregnancies. Eur J Haematol 2004; 73(6):431-436. (186) Burton G J , Jauniaux E . Placental oxidative stress: from miscarriage to pre-eclampsia. J Soc Gynaecol Investig 2004; 11(6):342-352. (187) Adams E . M . , Finlayson A . Familial aspects of pre-eclampsia and hypertension in pregnancy. Lancet 1961; 2:1375-1378. (188) Kilpatrick D C , Gibson F, Livingston J, Liston W A . Pre-eclampsia is associated with H L A - D R 4 sharing between mother and foetus. Tissue Antigens 1990; 35(4):178-181. (189) Torry D S , Hinrichs M , Torry R J . Determinants of placental vascularity. A m J Reprod Immunol 2004; 51(4):257-268. (190) Ament L . Factor V Leiden: a review of the literature. J Perinat Neonatal Nurs 2003; 17(3):190-195. (191) Walker ID. Thrombophilia in pregnancy. J C l i n Pathol 2000; 53(8):573-580. (192) Ness R B , Markovic N , Bass D , Harger G , Roberts J M . Family history o f hypertension, heart disease, and stroke among women who develop hypertension in pregnancy. Obstet Gynaecol 2003; 102(6):1366-1371. 144 (193) V o n Dadelszen P, Magee L A , Devarakonda R M , Hamilton T, Ainsworth L M , Y i n R et al. The prediction of adverse maternal outcomes in pre-eclampsia. J Obstet Gynaecol Can 2004; 26(10):871-879. (194) Peralta-Pedrero M L , Guzman-Ibarra M L , Cruz-Avelar A , Basavilvazo-Rodriguez M A , Sanchez-Ambriz S, Martinez-Garcia M C . [Usefulness of establishing diagnosis and severity of the most frequent signs and symptoms in preeclamptic patients]. Gac M e d M e x 2 0 0 4 ; 140(5):513-517. (195) Bernard G R , Wheeler A P , Russell J A , Schein R, Summer W R , Steinberg K P et al. The effects of ibuprofen on the physiology and survival of patients with sepsis. The lbuprofen in Sepsis Study Group. N Engl J M e d 1997; 336(13):912-918. (196) Bernard G R , Vincent JL , Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J M e d 2001; 344(10):699-709. (197) Arons M M , Wheeler A P , Bernard G R , Christman B W , Russell J A , Schein R et al. Effects of ibuprofen on the physiology and survival of hypothermic sepsis. Ibuprofen in Sepsis Study Group. Crit Care M e d 1999; 27(4):699-707. (198) Bis K A , Waxman B . Rupture of the liver associated with pregnancy: a review of the literature and report of 2 cases. Obstet Gynaecol Surv 1976; 31(11):763-773. (199) Sibai Baha M , Ramadan, Mohammed K , Usta Ihab, Salama, Mostafa et al. Maternal morbidity and mortality in 442 pregnancies with Haemolysis, Elevated Liver Enzymes and L o w Platelets(HELLP Syndrome). A m J Obstet Gynaecol 1993; 169(4): 1000-1006. (200) Waterstone M , Bewley S, Wolfe C. Incidence and predictors of severe obstetric morbidity: case-control study. B M J 2001; 322(7294): 1089-1093. (201) Isler C M , Rinehart B K , Terrone R W , et al. Maternal mortality associated with H E L L P Syndrome. A m J Obstet Gynaecol 1999. 145 (202) Belfort M A , Sadde G R , Moise K J Jr. The effect of magnesium sulfate on maternal retinal blood flow in pre-eclampsia: A randomised placebo controlled study. A m J Obstet Gynaecol 1992; 167:1548-1553. (203) Sheehan H L , Lynch JB. Cerebral lesions. Pathology of toxemias of pregnancy. Baltimore: Wil l iams &Wilk ins , 1973. (204) Govan A D T . The pathogenesis of eclamptic lesions. Pathol Microbio l 1961; 24:561-575. (205) Brown C E L , Purdy P D , Cunningham F G . Head comuter tomographic scans in women with eclampsia. A m J Obstet Gynaecol 1988; 159:915-920. (206) Gianninna G , Belfort M A , Abedejos P, Dorman K . Comparison of intra ocular pressure between normotensive and pre-eclamptic women in the peripartum period. A m J Obstet Gynaecol 1997; 176:1052-1055. (207) Cunningham F G , Fernandez C O , Hernandez C. Blindness associated with pre-eclampsia and eclampsia. A m J Obstet Gynaecol 1995; 172:1291-1298. (208) Pritchard J A C F P S . The parkland memorial hospital protocol for treatment of eclampsia: Evaluation of 245 cases. A m J Obstet Gynaecol 1984; 148:951-963. (209) Hinchey J, Chaves C, Appignani B ea. A reversible posterior leukoencephalopathy syndrome. N Engl J M e d 1996; 334:494. (210) Hibbard L T . Maternal mortality due to acute toxemia. Obstet Gynaecol 1973; 42(2):263-270. (211) Smith K , Browne J C M c , Shackman R, Wrong O M . Renal failure of obstetric origin. B r M e d B u l l 1968;24:49-58. (212) Sibai B M , Vi l l a r M A , Mabie B C . Acute renal failure in hypertensive disorders of pregnancy. Pregnancy outcome and remote prognosis in 31 consecutive cases. A m J Obstet Gynaecol 1990; 162:777-783. 146 (213) Sibai Baha M , Ramadan, Mohammed K . Acute renal failure in pregnancies complicated by Haemolysis, Elevated Liver Enzymes, and L o w Platelets. A m J Obstet Gynaecol 1993; 168(6):1682-1690. (214) Rakoczi I, Tallian F, Bagdany S, Gati I. Platelet life-span in normal pregnancy and pre-eclampsia as determined by a non-radioisotope technique. Thromb Res 1979; 15(3-4):553-556. (215) Cunningham F G , MacDonald P C , Gant N F . Will iams Obstetrics. Norwalk: Appleton & Lange, 1993: 763-817. (216) Redman C W , Allington M J , Bolton F G , Stirrat G M . Plasma-beta-thromboglobulin in pre-eclampsia. Lancet 1977; 2(8031):248. (217) De Boer K , Ten Cate JW, Sturk A , Bormm JJ, Treffers P E . Enhanced thrombin generation in normal and hypertensive pregnancy. A m J Obstet Gynaecol 1989; 160:95-100. (218) Redman C W , Denson K W , Bei l in L J , Bolton F G , Stirrat G M . Factor-VIII consumption in pre-eclampsia. Lancet 1977; 2(8051):1249-1252. (219) Carl J Saphire, John T Repke. Haemolysis, Elevated Liver Enzymes, and Low Platelets(HELLP Syndrome): A review of Diagnosis and Management. Seminars in perinatology 1998; 22(2):118-133. (220) Cotton D B , Gonik B , Dormann K F . Cardiovascular alterations in severe pregnancy-induced hypertension: Acute effects of intravenous magnesium sulfate. A m J Obstet Gynaecol 1984; 148:162-165. (221) Cunningham F G , Pritchard J A , Hankins G C V E A . Idiopathic cardiomyopathy or compounding cardiovascular events? Obstet Gynaecol 1986; 67:157-168. (222) Sibai B M , Mabie B C , Harvey C J , Gonzalez A R . Pulmonary edema in severe pre-eclampsia and eclampsia:Analysis of 37 consecutive cases. A m J Obstet Gynaecol 1987; 156:1174-1179. 147 (223) Desai D K , Moodley J, Naidoo DP , Bhorat I. Cardiac abnormalities in pulmonory edema associated with hypertensive crises in pregnancy. B r J Obstet Gynaecol 1996; 103:523-528. (224) Lehmann D K , Mabie W C , Mi l l e r J M , Pernoll M L . The epidemiology and pathology of maternal mortality: Charity hospital of Louisiana in New Orleans, 1965-1984. Obstet Gynaecol 1987; 69:833-840. (225) Bouveir-Colle M - H S B A P - Y , Varnoux N , Fernandez H , Papiernik E ea. Obstetric patients treated in intensive care units and maternal mortality. Eur J Obstet Gynaecol Reprod B i o l 1996; 65:121-125. (226) Alexopoulos E , Tambakoudis P, B i l i H , et al. Acute renal failure in pregnancy. Renal Fai l 1993; 15:609-613. (227) Grunfeld JP, Pertuiset N . Acute renal failure in pregnancy. A m J K i d Dis 1987; 9:359-362. (228) Wasson J H , Sox H C , Neff R K , Goldman L . Clinical prediction rules. Applications and methodologic standards. N Engl J M e d 1985; 313:793-799. (229) Wit l in A G , Saade G R , Mattar F, Sibai B M . Risk factors for abruptio placentae and eclampsia: analysis of 445 consecutively managed women with severe pre-eclampsia and eclampsia. A m J Obstet Gynaecol 1999; 180(6 Pt 1): 1322-1329. (230) American college of Obstetricians and Gynaecologists. Hypertension in pregnancy. Technical Bulletin 1996; 219. (231) Sibai B M , Anderson G D , McCubbin JH . Eclampsia II. Cl inical significance of laboratory findings. Obstet Gynaecol 1982; 59(2):153-157. (232) Magann E F , Martin J N , Jr. The laboratory evaluation of hypertensive gravidas. Obstet Gynaecol Surv 1995; 50(2):138-145. 148 (233) Wi t l in A G , Sibai B M . Hypertension in pregnancy: current concepts of pre-eclampsia. A n n u R e v M e d 1997;48:115-127. (234) Barton JR, Wi t l in A G , Sibai B M . Management of mi ld pre-eclampsia. C l i n Obstet Gynaecol 1999; 42(3):455-469. (235) Sibai B M , Ewel l M , Levine R J , Klebanoff M A , Esterlitz J, Catalano P M et al. Risk factors associated with pre-eclampsia in healthy nulliparous women. The Calcium for Pre-eclampsia Prevention (CPEP) Study Group. A m J Obstet Gynaecol 1997; 177(5):1003-1010. (236) Magee L A , Cham C, Waterman E J , Ohlsson A , V o n Dadelszen P. Hydralazine for treatment of severe hypertension in pregnancy: meta-analysis. B M J 2003; 327(7421):955-960. (237) Piguel D , Pierre F , Pourrat O, D'Halluin G , Magnin G . Are the symtpoms relevant enough to predict the occurrence of pre-eclampsia. Hypertens Pregn 2002; 21:51. (238) Stoger S W B . Diagnostic value of the clinical history in pre-eclampsia. Hypertens Pregn 2002; 21:48. (239) Belfort M A . Scoring systems for eclampsia. Crit Care M e d 2000; 28:272-273. (240) Richardson D K , Gray JE, McCormick M C , Workman K , Goldmann D A . Score for Neonatal Acute Physiology: a physiologic severity index for neonatal intensive care. Paediatrics 1993; 91(3):617-623. (241) Caetano M , Ornstein M P , V o n Dadelszen P, Hannah M E , Logan A G , Gruslin A et al. A survey of Canadian practitioners regarding the management of the hypertensive disorders of pregnancy. Hypertens Pregnancy 2004; 23(l):61-74. (242) V o n Dadelszen P, Magee L A . Could an infectious trigger explain the differential maternal response to the shared placental pathology of pre-eclampsia and normotensive intrauterine growth restriction? Acta Obstet Gynaecol Scand 2002; 81(7):642-648. 149 (243) V o n Dadelszen P, Magee L A , Krajden M , Alasaly K , Popovska V , Devarakonda R M et al. Levels of antibodies against cytomegalovirus and Chlamydophila pneumoniae are increased in early onset pre-eclampsia. B J O G 2003; 110(8):725-730. (244) Chesley L C , Annitto JE, Cosgrove R A . The remote prognosis of eclamptic women. A m J Obstet Gynaecol 1976; 6:446-459. (245) Chesley L C . Recognition of the long-term sequelae of eclampsia. A m J Obstet Gynaecol 2000; 182:249-250. (246) Ross R. Atherosclerosis - an inflammatory disease. New Engl J M e d 1999; 340:115-126. 150 Appendix 1 - Guidelines for data collection Please enter the data into the computer on a timely basis. This is especially important when you first begin so that any misunderstanding can be clarified. The information entered w i l l be reviewed on a regular basis by the research coordinator. If there are problems or data is missing you w i l l be sent a data clarification form via e-mail or you may be phoned. Data sheet: (Table 1) Table 1 or data sheet includes: - patients demography, patient's family history, personal history, past obstetric history, current and past treatment history, other relevant medical history, maternal age, gestational age (GA) at the time of admission, G A at the time o f diagnosis of pre-eclampsia, G A at the time of termination of pregnancy, weight, height etc. Data should be entered in the following order: Table 1, Table 2, and lab data respectively. Blue box I N F O R M A T I O N IS M A N D A T O R Y . If the blue boxes are not entered none of the data in that record w i l l be saved. Always use the drop in menus wherever it is available. Each box is set with a range for that variable. If the value exceeds the range the field w i l l not accept that value. This is done to prevent accidental errors while entering the data. If the box cannot take the value, the data to be entered in the exceptions box at the bottom of the page. For example, L D H value was 12,000 and the L D H box w i l l not allow you to enter that data, then go to exceptions box and type in L D H - 1 2 , 000. Y o u can enter multiple variables by using comma in between variables in this box. 151 PHN: In British Columbia it is the ten-digit personal health number. In Ontario it is the OHIP#. In the United Kingdom it is the patients social security number. Study ID: Study identification number. Please use the letters assigned for the centre, and then, assign a number to the patients beginning with 1. For example in Ottawa the first patient or the chart would be given the first identification number C l ; the next patient file would be C2 for patient 2. If the mother was admitted twice for two different deliveries please allot her a different study ID from the previous year. The letters assigned for the centres are: • C - Ottawa, Canada. • U - Oxford, U K . • V - Vancouver, Canada Dates: • DOA: Date of admission for the delivery. The entry will be dd/mm/yy • Mother's DOB: mother date of birth .The entry will be dd/mm/yy • EDD: expected date of confinement. The entry will be dd/mm/yy Onset of pre-eclampsia: • EO PET #: Early onset pre-eclampsia. Onset of PET at < 34 weeks of gestation. Please allot 1 i f it is early onset in the early onset box and enter 0 in other boxes. • LO PET: Late onset pre-eclampsia. Onset of PET >34 weeks of gestation. Please allot 1 i f it is late onset in the late onset box and enter 0 in other boxes. Parity: (Number of pregnancies) • G: Gravida-total number of pregnancies. • T: Term delivery -delivery at >37 weeks. • P: Pre term delivery-delivery at 0-36 weeks. 152 • SA: Spontaneous abortion-greater than or equal to three consecutive spontaneous abortions in less than 20 weeks gestational age. • T A : Therapeutic abortion; ending of a pregnancy through therapeutic means. • L : Living Gestational age: • Gestational age for admission for delivery: Gestational age on the day of admission for delivery. To be entered as dd/mm/yy. • Gestational age at the diagnosis of pre-eclampsia gestational hypertension and or/ proteinuria: Gestational age at which all the criteria for the definition of pre-eclampsia are satisfied or when the Dr. first notices that these syndromes are present. To be entered as dd/mm/yy. • Gestational age at termination of pregnancy: Gestational age on the day baby is delivered or the pregnancy is terminated. To be entered as dd/mm/yy. Anthropometry: • Pre-pregnancy weight: Weight of the patient prior to pregnancy. May be recorded in kilograms or pounds. Enter what is present in the chart. Do not convert as the computer is programmed to convert the entry into kilograms • Present pregnancy weight: The Weight of the patient on admission or within 24 hours. It may be recorded in Kilograms or pounds. Enter what is present in the chart. Do not convert as the computer is programmed to convert the entry into kilograms • Height: May be recorded in centimetres or feet and inches. Enter what is present in the chart. Do not convert as the computer is programmed to convert the entry into centimetres. Ethnicity: • Caucasian: White, individuals from the Middle East. • B lack : African, or Afro-American • East Asian: Individuals from Southeast Asia, China, Japan, and the Philippines. • South Asian: Individuals from the Indian subcontinent. E.g. India, Bangladesh, Pakistan, Sri Lanka and Nepal. • First Nations: Canadian indigenous people. 153 • Latino: Individuals from South America, Mexico, Spain and Portugal. • Pacific Islanders: Individuals from Hawaii, Polynesia, Micronesia, Melanesia (New Zealand, and Australia) • Aborigines: Australian indigenous people • Others: Martial Status: • Married/stable relationship: Individuals who are married or living together in a stable relationship. • Single: an individual who currently does not have a partner. Past Obstetric history: Refers to previous pregnancies only, not the current pregnancies. When entering the data in to the computer this area will contain drop down boxes indicating whether the value is: 'yes', 'no', or 'unknown'. If the condition was present then the answer is 'Yes', i f the condition was not present then the answer is 'no'. If it is not known if the condition existed then the answer is unknown. • Gestational hypertension +/_ proteinuria/ Pre-eclampsia: BP > 140/90 on 2 separate occurrences or 4 hours apart. Proteinuria of >0.3 grams per day or 2 + protein by dipstick. • HELLP: Hemolysis (abnormal peripheral blood smear), increased bilirubin (> greater than 1.2 mg/dl), elevated L D H (> 600 IU/litre), Elevated Liver enzymes (AST >100 IU or A L T > 70 IU /litre). Low Platelets (< 100,000 per cumm). • Previous nlUGR: Previous normotensive intrauterine growth restriction- with birth weight of less than 3 r d percentile for the baby and normal blood pressure for the mother. • Known Aneuploidy - Numerical chromosomal abnormalities. • Major structural anomaly • Pre-pregnancy diabetes: • Recurrent Spontaneous Abortion (SA): Recurrent Spontaneous Abortion. >3 consecutive spontaneous abortions in less than 12 weeks. 154 Personal History: Mother's history pertaining to current pregnancy. When entering the data in to the computer this area w i l l contain drop down boxes indicating whether the value is 'yes', 'no' , or 'unknown'. If the condition was present then the answer is 'yes', i f the condition was not present then the answer is 'no' . If it is not known i f the condition is or is not present then the answer is 'unknown'. • Smoking: Smoking during pregnancy. A n y amount of smoking w i l l be considered yes. • DVT/PE: Deep vein thrombosis or pulmonary embolism confirmed by history, physical exam or scan. I f present the answer w i l l be yes. • Cocaine: Use of cocaine during pregnancy. A n y amount used w i l l be considered a yes answer. • Renal disease: • IHD or other heart disease except MVP: Ischaemic, structural or functional heart disease other than Mitra l valve prolapses. • Obesity: Body mass index (BMI)>25 at the time of conception. Calculate the B M I using metric measurements: K g / m 2 . e.g. Pre-pregnancy weight of 70kgs and height of 150 cms that w i l l be 1.50m. 70/1.50 2 =31.1.1 B M I • New partner: Refers to this pregnancy • Hypertension: History of persistent elevated blood pressure of > 140/90 mmHg or a >90 diastolic blood pressure before 20 weeks of pregnancy. • Other: Please specify what other medical problems the patient may have had during this pregnancy or previous to the pregnancy. This information w i l l need to be entered directly into the space supplied. For example: • Epilepsy: Controlled by medications. • Thyroid disease: Physician diagnosis of hyperthyroid, hypothyroid or euthyroid disease. • Migraine: Diagnosed by a physician. • S L E : Physician diagnosis of systemic lupus erythematous • A P S : or A P L A antiphospholipid 155 Family history: Significant maternal or perinatal medical history. When entering the data in to the computer this area w i l l contain drop down boxes indicating whether the value is yes, no, or unknown. If the condition was present then the answer is yes, i f the condition was not present then the answer is no. If it is not known i f the condition existed then the answer is unknown. • FHx Pre-eclampsia: hypertension of > 140/90 and proteinuria > 0.3g/day in 24 hours that appears after 20 weeks i f gestation. • FHx Pre-eclampsia: BP>140/90 m m H G on 2 separate occurrences or 4 hour apart with proteinuria of > 0.3g/day or > 1 by dipstick. • FHx n l U G R : normotensive intrauterine growth factor restriction with ultrasound foetal birth weight of less than the 3 r d centile and normal blood pressure. • Previous nlUGR: Previous normotensive intrauterine growth restriction- with birth weight of less than 3 r d centile for the baby and normal blood pressure for the mother. • Heart disease: Myocardial infection at <60 years, angina at <60 years, or congenital heart disease other than mitral valve prolapse. • DVT/PE: Deep vein thrombosis or pulmonary embolism. • Hypertension: Persistent elevated blood pressure of 140/90mmHg • Other: Complete this field by entering the condition and relevant information. E.g. Mother had renal disease before 40. Treatment in this pregnancy • Steroids for foetal lung maturity: • Steroids for H E L L P syndrome: • Antihypertensives: • M g S 0 4 for Gestational hypertension: • Methyldopa: 156 Labetolol: Beta Blockers: Calcium Channel Blockers: ASA: Heparin: Other: LAB Form: This form must be completed for each day that the patient is in the hospital from admission to discharge. Enter data for all the days of hospital stay and for as many variables as possible. Whenever the data is not available, leave the space empty. On the data worksheet indicate missing data by writing unknown or missing data. • Patient ID (MRN): same as previous page. • Test Date: The day the lab tests were done. • WBC: White Blood Cells count, measurement in cells/ml. Record the highest value of the day. • Measuring BP: The highest blood pressure of the day will be recorded. This may need to be calculated using the mean arterial pressure (MAP) formula of 2 (diastolic BP)/3. Example:BP: 180/100, 2(100) + 180 divided by 3= 126.6 map. The M A P is one that should be recorded for the day.**if the subject is in shocks then the lowest BP must be recorded. • BP systolic (mmHg): Systolic blood pressure in mmHg. • BP diastolic (mmHg): Diastolic blood pressure in mmHg. • Sa02 (%ile): Haemoglobin oxygen saturation in arterial circulation expressed in a percentile. Record the lowest value for the day. • # of Seizures: The number of seizures the subject has had in a day. • Uric Acid (uM): Record highest value of the day. • Creatinine (LIM): Record highest value of the day. • Urinary Albumin (g/D): Record the highest value of the day. • Dipstick protein: The measurement of protein in the urine using the dipstick method. Drop down box. Record the highest value of the day. • AFI %ile for GA (0-100%): percentile of amniotic fluid for gestational age. • Urine output (ml/H): the measure of urine output per hour. Total output in millilitres 24 hours. Example: total output is 1200 millilitres. 1200/24=50ml per hour. 158 • Total Bilirubin (p-M): Record the highest value of the day. • AST (u/L): Aspartate aminotransferase. Record the highest value of the day. • LDH (u/L): Lactate dehydrogenase or lactic acid dehydrogenase. Record the highest value of the day. • Albumin (g/L): Plasma albumin • Platelets (exp9/L): Record the lowest value of the day. • MPV (fL): Mean platelet volume. Record the highest value of the day. • Fibrinogen (g/L): Record the highest value of the day. • EFW for GA ( 0 - 1 0 0 % ) : Estimated foetal weight for gestational age. This is a drop box with ranges of percentiles. Choose the correct range and click. The range will be entered into the computer. Record the lowest value of the day. • Diastolic Flow: This is a drop down box. Indicate status of diastolic blood flow: present -absent-reverse Treatment history: Gestational age at which treatment with anti-hypertensive/ steroids for H E L L P syndrome and for the foetus for prematurity/ MgS04 were first started. Please enter G A in weeks and days i f known. At the end of each page there are forward and backward arrows, which wil l take you to the new form or the previous form as, needed. The button next to forward with a line next to the forward arrow will take you one record in front and the one button left side to the backward button takes you one record previous to your current record. By clicking on the button with forward arrow and star will take you to a new form for new patient file data entry. Once you complete the data entry for that form you can click on the X on the top right hand corner of each page and you can go to the next form. If the blue box is incomplete you will get a prompting from the system saying the required field is incomplete and you can go back to the form and you can complete the required 159 data. If you close this window without completing the required the data will not be saved. Then you can go to Table 2 by double clicking. You can work on multiple forms simultaneously. Lab data form abbreviations: Variables Expansions Record Highest value Record Lowest value Abbreviations used for the day. for the day. W B C White blood cells X (cells/ml) Systolic Systolic blood pressure X BP*(mmHg) Diastolic B P * Diastolic blood pressure X (mmHg) Sa0 2 Saturated arterial Oxygen X (% RA) for room air # of seizures X Uric acid X ( • M ) Creatinine X ( D M ) Urinary albumin X (g/D) Dipstick protein X AST Aspartate amino X ( • M ) transferase L D H Lactate dehydrogenase X ( • M ) Bilirubin X ( • M ) Platelets X (X 109/L) M P V Mean platelet volume X Fibrinogen X (g/L) EFW % ile Estimated foetal weight X in percentile AFI %ile Amniotic fluid index in X percentile Diastolic flow Umbilical artery diastolic X flow *Highest BP yielding the highest M A P for that day will be recorded unless the patient is in shock or failure where in lowest BP as well as M A P is recorded. 160 Maternal outcome form: (Table 2) These forms are self-explanatory. Please enter all the sections using drop boxes and select the appropriate answer. The person's ID and other M R N numbers are defined in the data sheet page. • Number of times: Refers to the number of times the incidence occurred. Example the patient had 4 seizures in one day. The correct answer would be 4. E.g. If the person requires >50% of oxygen 5 times during the period of hospital stay, then answer w i l l be 5. • How long (days): H o w long the incidence was there. E.g. H o w many days the patient was blind from cortical blindness. • Mortality: This refers to the death of the mother only. Please indicate the date dd/mm/yy. If death did occur click 'yes' in the drop down box and mention the date in the date box as dd/mm/yy. If 'no,' select 'no' in the drop in menu. • Placental abruption: Clinical ly diagnosed by abdominal pain or uterine contractions with one or more of vaginal bleeding, intrauterine foetal death and/or disseminated intravascular coagulation. This is a drop down box indicating 'yes'; there was placental disruption or 'no' , i f there was not. Please indicate yes or no as appropriate. • Hepatic failure: Includes elevated liver enzymes, conjugated hyperbilirubinemia and either hypoglycemia or LNR greater than 1.2. Indicate i f this occurred by using the drop down box. • Hepatic hematoma and /hepatic rupture: Defined by the presence of blood collection under the hepatic capsule and/ or rupture of the hepatic capsule as confirmed by ultrasound. Indicate i f this occurred by using the drop down box. • GCS: Glasgow coma scale of <13. Indicate i f this occurred by using the drop off box. • Stroke: Acute neurological event with deficits lasting greater than 24 hours, not due to a post- ictal stage. Indicate i f this occurred by using the drop down box. • Two or more seizures: Record only i f the patient had 2 or more seizures. Indicate i f this occurred by using the drop down box. Record the number of times the patient had 2 or more seizures from the time of admission to discharge. Example* during hospital stay. • Cortical blindness: A lack of visual function in spite of anatomically and structurally intact eyes because of decreased blood supply to cortical region of brain. Indicate i f this occurred by using the drop down box. Record the number of times the incidence occurred from the time of admission to discharge. 161 Dialysis: M a y include renal or peritoneal dialysis. Indicate i f this occurred by using the drop down box. Record the number of times the incidence occurred from the time of admission to discharge. Record the number of times in the box given. Transplantation: Indicate i f the patient had a renal transplant by using drop down box. Positive ionotropic support: The use o f vasopressors to maintain a systolic blood pressure of >90 mmHg or a mean arterial pressure >70mmHg. Indicate by using the drop down box. Give gestational age or postpartum day at time of incidence. Infusion of 3rd antihypertensive: Indicate i f the patient has received infusion of 3 r d antihypertensive because of uncontrollable hypertension with other oral or injectable antihypertensives. This does not include Magnesium Sulphate. Give the number of times the incidence occurred from admission to discharge. MI: Myocardial Infraction: Indicate by using the drop down box. Give the number of times the incidence occurred form admission to discharge. Requires >50% O2 >1 hour: The patient requires Oxygen at > 50% for longer than 1 hour. Indicate as appropriate. Give the number of times the incidence occurred from admission to discharge. Intubation(Vent, EIT, CPAP): Ventilation, monitoring of lung function with electronic image tomography or continuous positive air pressure. Indicate by using the drop down box. Give the number of times the incidence occurred from admission to discharge. >10 units of blood or blood product: Indicate as appropriate using drop down box. Give the number of times the incidence occurred from admission to discharge. • Platelet infusion: Indicate number of times during hospital stay. Indicate by using the drop down box. Give the number o f times the incidence occurred from admission to discharge. • FFP infusion: Fresh frozen plasma infusion. Indicate by using the drop down box. Give the number of times the incidence occurred from admission to discharge. Indicate using drop down box. • Cryo infusion: Cryoprecipitate infusion. Indicate by using the drop down box. Give the number of times the incidence occurred from admission to discharge. Date of mother's discharge: Date mother was discharged from the hospital. 162 Appendix 2- Prediction of adverse maternal outcome in pre-eclampsia paper published in Journal of Obstetrics and Gynaecology Canada, 2004. A 2.1 Preface n addition to the 556 women, 38 more women were added for this part of the study. The adverse maternal outcome was developed in 60 women (10.1%). The study statistician did analysis for this portion of the study. This article was published in Journal of Obstetrics and Gynaecology, 2004. A 2.2 The Prediction of adverse maternal outcomes in pre-eclampsia. 163 THE PREDICTION OF ADVERSE MATERNAL OUTCOMES IN PREECLAMPSIA Peter von Dadelszen, M B C h B , DPHil, F R C S C , 1 " Laura A . Magee, MD, F R C P C , J - 4 Rajashree M. Devarakonda, M B C h B , ' Trevor Hamilton, BEng, 1 Laurie M. Ainsworth, M S c , 5 Ruihua Yin, M S c , 4 Monica Norena, BSc , s Keith R.Walley, MD, F R C P C , 1 Andree Gruslin, MD, F R C S C , 6 jean-Marie Moutquin, MD, F R C S C , 7 Shoo K. Lee. MBBS, F R C P C , 5 4 James A . Russell, MD, F R C P C 2 'Departments of Obstetrics and Gynaecology, ^Medicine, and 'Paediatrics, and the 'Centre for Healthcare Innovation and Improvement, University of British Columbia,Vancouver BC 'Clinical Research Support Unit, British Columbia Research Institute for Children's and Women's Health.Vancouver BC 'Department of Obstetrics and Gynecology. University of Ottawa, Ottawa ON 'Department of Obstetrics and Gynecology, University of Sherbrooke, Sherbrooke QC Abstract Objectives: (I) To evaluate whether clinical variables reflecting the mukiorgan dysfunctions of preeclampsia can predict adverse maternal outcomes of preeclampsia: (2) to determine the usefulness of the mean platelet volume (MPV):platelet ratio as a novel measure of platelet consumption in predicting the severity of preeclampsia. Method: A retrospective chart review was conducted of. cases of preeclampsia seen in 3 tertiary level units from January 2001 to December 2001. Candidate predictors of adverse maternal outcome were gestational age (GA) on admission to hospital, b lood pressure, proteinuria, urine output, uric acid, creatinine, aspartate transaminase (AST), lactate dehydro-genase, bilirubin, albumin, fraction of inspired oxygemoxygen saturation (F lO^SaOj ) ratio, platelet count. MPV, MPVplatelet ratio, fibrinogen, and seizures.The combined adverse maternal outcomes included maternal death; I or more of hepatic fail-ure, hematoma, or rupture; Glasgow coma scale <I3; stroke; 2 or more seizures; cortical blindness; positive inotrope support; myocardial infarction; infusion of any third antihypertensive: dialysis: renal transplantation; >S0% FIOj for >l hour; intuba-tion; or transfusion of 210 units of blood products. Descrip-tive, univariable, and multivariable analyses were performed, with signficance set at P < ,05, Results: Of a total of S94 women with preeclampsia, 60 (10.1 %) developed at least I element of the combined adverse.outcome; I of these 60 women died. The most common outcomes were increased oxygen requirements, the. use of a third infused ami-hypertensive, and transfusion > 10 units. In women who devel-oped an adverse outcome, G A and fibrinogen were lower, and total leukocyte count, creatinine, and A S T we<*e greater, Multivariable logistic regression revealed that admission G A Key Words Pre-eclampsia. risk factors, HELLP syndrome, multivariate analysis Competing interests: None declared. Received on February 16, 2004 Revised and accepted on May 28, 2004 (odds ratio [OR] , 0.91). dipstick protein (OR, 1.31), and MPV:platelet ratio (OR, 391.0) independently predicted the outcome. Conclusions: Several promising markers were identified: admis-sion G A , dipstick proteinuria, and the MPV:piatelet ratio, MPV:platelet ratio also showed .promise as a marker of platelet consumption. A prospective study is required to develop a clin-ical prediction model for preeclampsia. Resume Objectifs : (I) Evaluer si les variables ciiniques qui refletent le dysfonctionnement affectant de multiples organes attribuable a la preeclampsie peuvent predire les issues indesirabtes maternelles de celle-ci; (2).determiner I'utilite du rapport vo-lume plaquettaire moyen (VPM)-.plaquettes a titre de mesure novatrice de ia consommation de plaquettes en matiere de prediction de la grayite de la preeclampsie. Methode : Une analyse retrospective de dossiers s'est penchee sur des cas de preeclampsie constates au sein de trois unites tertiaires ent're Janvier at decembre 2001. Parmi les pre-dicteurs candidats d'une issue indesirable maternelle, on trouvait I'Sge gestationnei (AG) au moment de I'hospitalisa-t ion, la tension arterfella, la proteinuric, la diurese. I'acide urique. la creatinine, I'aspartate transaminase (AST), la.Jactate-deshydrogenase. la bilirubine, I'alburnine, le rapport fraction d'oxygene du gaz inspire:saturation arterielle en oxygene (FIOjiSaQj), la numeration plaquettaire,. le V P H , le rapport VPH:plaquettes, le fibrmogene et les crises. Parmi ies issues indesirables maternelles combinees, on trouvait le dices maternel; une insuffisance, un hematome ou une rupture hepa-tiques ou plus; un resultat <I3 i lechelle de Glasgow; I'ac-cident vascufaire cerebral; deux crises ou plus; la cecite corticale; le soutien inotrope positif; i'infarctus du rnyocarde: la perfusion de quelque antihypertenseur de troisieme ligne que ce soit; la dialyse; la. transplantation renale; une valeur 2: SO % FIOj pendant >l heure; 1'incubation; ou la transfusion de >I0 unites de produits sanguins. Des analyses descriptives, univariees et multivariees ont ete effectuees, selon un seuil de signification etabli a P < 0,05. [ Resultats : Sur un total de 594 femmes presentant une pre-I ectampsie. 60 (10.1 %) d'entre elles ont presents au mains •1 OCTOBER 2004 urt element des issues indesirables cornbinees; une de ces feO-fatrsmes est decedec. les issues les plus courarites one er.e 1'accroissement des besoms en oxygene, ia perfusion d'un anti-hypertenseur de troisieme ligne et ia transfusion de ->I0 unites, Chez les fernmes qui ont present* une issue indesirable, TAG et le fibrinogette etaiem moindres, tandis que la teucocytemie \ totals,- la creatinine .et 1'AST etaienc accrues. La regression logis-tique mulrivartec a revelee que TAG a I'adrnission (rapport de cotes [RC], 0,91), les proteines rnesurees par bandelette reac-tive (RC, 1.31) et le rapport VPM:plaquecces (RC, 391.0) per-mcttaienc. de facon indepcndante. de predtre Tissue. Conclusions : Plusieurs marqueurs prometteurs ont ete identi-fies : AG a ('admission, proteinuric mesuree par bandelette reactive et rapport VPM.plaquettcs. Ce dernier s'est egalemenc revele prometteur a litre de marqueur. de la consommation de plaquettes. Une etude prospective s'avere necessaire pour la mise au point d'un modele de prediction cliriique pour la preeclampsie. j Obstet Gynaecol Can 2004;26(i0):87l~9. The syndrome of preeclampsia/cclampsia remains 1 of the 2 common causes of maternal death in the developed world.1"3 Preeclampsia complicates 3% to 5% of pregnancies and is com-monly defined as hypertension of > 1 40/90 mm Hg and pro-teinuria >0.3 g/d that appear at >20 weeks" gestation and regress after pregnancy. There arc 2 syndromes in preeclampsia: maternal and fetal. The maternal syndrome is characterized by multiple organ dys-functions. Although hypertension is the.mosc common clinical manifestation,4 with good blood pressure control 5 , 6 associated mortality is most commonly due to either hepatic necrosis or the acute respiratory-distress syndrome, both consequences of systemic inflammation/ It may be chat: the maternal syndrome of preeclampsia reflects a global,systemic response with many alternative routes to its inception.8 The definitive treatment for preeclampsia is delivery, which, although always the treatment of choice for the mother, may be a double-edged sword for the fetus less than 34 weeks" gestation.8,9 Each- weekgained in utero. confers benefit in .peri-natal outcome.*,9,Randlomizcd-co»ti'6]led trial (RCT) evidence shows that when the fetus-is remote from term, delaying deliv-ery until necessitated by either the maternal or fetal condition decreases serious perinatal morbidity without increased mater-nal risk.5*'0 However, these RCTs 5 , 1 0 bad insufficient power to detect a difference in serious-maternal outcomes between groups. Further, uncertainty about the magnitude of the mater-nal risk11 has made some clinicians reluctant to use such man-agement. In order to enhance the care of women with preeclampsia, it is necessary to predict those women at risk of adverse events who can be safely managed by expectant therapy.5 Guidelines for the diagnosis, classification, and manage-ment of preeclampsia have been produced by the Canadian Hypertension Society (CMS), 1 2 the US National High Blood Pressure Education Program (NHBPEP) Working Group on High Blood Pressure in Pregnancy,13 and the Australasian Soci-ety for the Study of Hypertension in Pregnancy (ASSHP) ' 4 (Table 1). The latter 2 form the basis,of the guidelines by the. International Society for the Study of Hypertension -in- .Preg-nancy (1SS.HP).'5.A1I these guidelines are based largely on expert opinion. Mild preeclampsia is classified as a blood pressure. (BP) >140/90 mm Mg with proteinuria of 0.3 g/d to 3 g/d. Severe preeclampsia is defined as mild preeclampsia with a single addi-tional "adverse feature," such as BP >160-170/100-110 nun Hg, proteinuria of >3 g/d, and/or headache. There are limitations to these guidelines. First, although most: classifications require, both hypertension and proteinuria, for 40% of women this combination of. conditions fails to occur within the week prior to an eclamptic seizure.-6Therefore, in practice, the diagnosis of preeclampsia .needs co be considered when either non-proteinuric gestational hypertension (present in 20% of women within.a week of their first eclamptic seizure16) or non-hypertensive gestational proteinuria (present in 1.0% of women1 6) exists. Furthermore, in a .secondary analysis of the Table 1. Classification of the Hypertensive Disorders of Pregnancy* CHS" NHBPEP1 4 .ASSHP ,S » Pre-existing htn, essential/secondary « Chronic htn • Chronic hin, essential/secondary » Gestational htn-without proteinuria. ± adverse features! • Transient hin • Pregnancy-induced hypertension * Gestational hin with proteinuria ± adverse features * » Preecl ampsia/edampsia * Preeclampsia, mild/severe • Preexisting htn superimposed gestational htn with proteinuria * Preeclampsia superimposed on chronic htn • Preeclampsia superimposed on chronic htn « Unclassifiable antenatally • C H S ; Canadian .' iypsr tcmimi Sax-jay; NHBPEP: Natfooj} High Blood Pressure Education Pwgrat-n: ASSHP; Australasian Socfetv tor iho Study of HypeVrewkjn in tVt>gj}am:y: bm: h>*,k?nf?mkan. t A d w w e r « ; « u r e s : ! ' co .WJs iusu : d B P > ! I t ) mm Hg, ihronibocyiuu«nia [<iot> x 10 , ! , - ' l o l igur i a{<S00 mUdS; pulmonary edema: e levawMiver efayiiieS; severe r a i s e s a m i v w n i i i a g ; S o w s * ! « s w f a c h « ; visual d i s iu rbaf ios ;pes i i t en i riirttt upperquadf i tmpa in :che t pain; dyspnea; suspected abruption: homtsiysis, elevatedIcvrsr enemies, l o w p b s r t e i a m u rt-in.ti ' isyndrom*:int/auierincgrovwhrestriction ClUCKS; o%shydrammo5;a teen!of reversed umbi l ica laner ia le ra !d ias to l ic i to iv iDnppier). X Adverse kMuiet". s,<env ;t> above plus proieiiHrtia >l ^ 'd and h y ^ i b t s m i n v r n i a ?«1rj | O G C » F W a O C T O B E R 5004 National Institute of C h i l d Health and Human Development (NICH.D) Aspir in for the Prevention o f Preeclampsia t r ia l , 1 7 women who developed severe oom-proteijuiriegestatbnal hyper-tension had higher rates of both preterm delivery (<37 weeks) and small-fbr-gestatiooal-age infants, compared.wich women who developed -mild preeclampsia.' 8 Second, dichotomizing preeclampsia into mild or severe disease, presumably differenti-ates women with lower risk from chose with higher risk, but there arc no shades of grey over a broad, range of clinical situations. Third, gestational age, although the most important predictor of both maternal and perinatal outcornes, , , 8" l f i has not been accounted for in any of the current classification systems.'1* Early onset preeclampsia (<32 weeks) is associated with a 20-fold high-er risk of maternal mortality compared with preeclampsia that occurs at term,'"'' and is associated with greater perturbations in biological markers of disease activity. 8 , 2 1 Also, gestational age-is the most .important determinant of perinatal outcome among dipioid fetuses.5- !<- , ; i ; ! A greater than 50% chance of intact fetal survival in preeclampsia arises only when the gestational age at delivery is i>27'1' weeks and the birth weight is 2600 g . s Therefore, we investigated die feasibility of identifying objec-tive clinical and laboratory variables that reflect: the mukiorgan dysfunctions of preeclampsia.'iuid couldpredict a combined adverse maternal outcome. The. secondary outcome recognizes that the platelet counts fall prior to ami wi th the onset ol preeclampsia,•''i,2'> the M l ' V can rise prior to such a fall in platelet count,'**2'' as megakaryocytes release immature (and large) platelet forms in response to platelet consumption before that consump-tion overwhelms this increased, supply.. As a result:, the ratio of MPV:platelet count would be better able to.detect, platelet con-sumption than the-measuringof'either platelet parameter alone, METHODS recorded. However, some elements* from Richardson's criteria were not met, namely that the variables were necessarily fre-quently obtained (e.g., -fibrinogen) and reliable (e.g:, blood pres-sure measurement). The requirement for rneasurability led us Tabic 2. Candidate Maternal and Fetal Predictor Variables* Organ System Variablet Gestational age on admission (admission during which delivery occurred) Maternal Cardiovascular Systolic BP Diastolic BP Renal Albuminuria: • 24-l»our urine • spot proiemxrenlinine ratio • dipstick proteinuria Urine output Uric acid Creatinine Hepatic Aspartate-transaminase (AST} Lactate dehydrogenase (LDH) Bilirubin Albumin Respiratory -FlOjtSaOj. Hematological Platelet count Mean platelet volume (MPV) MPV:platelet count ratio Fibrinogen Central nervous Seizures Fetal Growth Estimated fetal weight (percentile) Renal perfusion Amniotic fluid index (percentile! Vascular maladaptation Umbilical artery Doppler (present vs. ^ absenl'/reversedj *D3S<t watt* cdte-c'ed on ihe day of d^mis^ ityt ami the day at ddivoty. IBi1: blood fj!fc$surt>; RO,:SaO*: fratilon of mspirea r^yge?iiO\'\<ger;.vrfiiiWtiuri. A literature review was conducted to determine how preeclamp-sia is classified, managed, and predicted, and how the compli-cations werecharacterized in previous studies. Special focus was placed on the C H S 1997 guidelines. 1 2 The current outcome prediction models.used in categorizing women with the sys-temic inflammatory response syndrome (SiRS),2 8*3'"1 with which preeclampsia shares many characteristics/*31 were, also reviewed. This process defined the organ systems to be included in the. score, the specific variables used to quantify dysfunction with-in each system, and the outcomes to be predicted. Using the. Delphic method**''*'*'''' bye-mail survey, 14 international experts were asked to review the candidate predictor variables (Table 2) and the combined adverse maternal outcomes (Table 3) and: to suggest omissions. CANDIDATE PREDICTOR VARIABLES The variables-chosen fulfilled 3 of the 5 criteria of Richardson et ai.?4 in that they were available, measurable, and accurately Table 3. Combined Adverse Maternal Outcome (Primary Outcome) Organ System Hepatic Central nervous Cardiovascular Renal Respiratory Hematological Maternal Outcome Mortality Failufe Hematoma Rupture Glasgow coma scale (CCS) <13 Stroke Two or more seizures or status eclampticus Conical blindness 'transientw permanent! Positive inotrope support. Myocardial infarction Infusion of any third antihypertensive Dialysis (temporary or permanent) Renal 'transplantation Requirement of >S0% Oj for >! hour, or intubation Transfusion of £10 units of blood products (in total) jOGC j fJJ^J ! OCTOBER JtXM 1 • I to exclude, from among die C H S guidelines' "adverse fea-tures,''" 1 ; maternal symptoms front the list of candidate variables. 1 herefore, we investigated objective maternal variables from 7 different organ systems along with the gestational- age at the time of the delivety admission and the day of delivery, as well as ultra-sound fetal, variables for women -with preeclampsia. These 2 time points are clinically important, as admission assessments guide decision-making around delivery or expectant manage-ment,.and the day-of delivery, recognizes that women may devel-op worse disease in die'tnimediatepuerperitiiti. Feral parameters were chosen as they might reflect the severity'of the underlying placental disease, and thereby indirectly predict worse ..maternal outcomes. ln'addition to-standard clinical and laboratory markers of disease activity, we tested the ratio of M P V (fL):platelet. count (x ] 0 9 / L ) , as'a novel marker of platelet consumption. COMBINED ADVERSE MATERNAL OUTCOME 'The combined adverse maternal outcome-was developed to reflect the multisystem nature of preeclampsia. Maternal mor-tality is a rare event, even among women with preeclampsia.1"3 What is more relevant to predict is serious maternal morbidity that would preclude safe.pregnancy prolongation. We proposed a list .of .outcomes based on known serious end-organ compli-cations o f preeclampsia that we felt would change clinical man-agement, by being considered worthy of avoidance, even in the face of extreme prematurity. STUDY DESIGN We. conducted a retrospective cohort study o f women admit- -ted to 3 tertiary-level perinatal centres: the Children's and Women's Heal th Ou t re o f Bri t ish C o l u m b i a , Vancouver, British Columbia ; the Ottawa, Hospital - GeneraiCampus, Ottawa, Ontario; and the Maternal Medicine Unit , John-Rad-cliffe Hospital, Oxford, United Kingdom, The Maternal Med-icine U n i t in Oxfo rd preferentially cares for women with preeclampsia diagnosed prior to '34 completed weeks of preg-nancy and does not care for all cases of preeclampsia admitted ro the -hospi tal. AH consecu ti ye cases of women "admi f ted- to the respective units with a diagnosis of a hypertensive disorder o f pregnancy from January 2000 to December 2001 were reviewed. Women were eligible if they had been admitted to hospital and had at least: 2 of the following: (1) hypertension (sBP >l40 mnvHg aiJd/or. dBi''S90 mm H g , twice, >4 hours apart)-after 20 weeks, (2) proteinuria-defined- as:>0.3 g/d or>2* dipstick proteinuria .after 20 weeks, 1 2 (3) non-hypertensive and-non-proteinuric hemolysis, elevated liver enzymes, low platelet count (11ELLP) syndrome, using the .criteria o f -.Audibert etal.P or (4) an isolated eclamptic seizure without preceding hyperten-sion or proteinuria* using the British Eclampsia Survey Team (BEST) criteria to define eclampsia. 1 6 Women were excluded i f they were admitted i n spontaneous labour or had achieved the maternal outcome prior to fulfilling the:eljgibility crtletia. A Windows Access database was developed for the data entry. For the day of delivery, the "last observation carried for-warf ' 'method was-used, by which any-preceding observation-was considered current unless:replaced by-a more current-value. For example, a 24-hour urine proteinuria of 0.6 g/d measured 4 days prior to delivery would be considered to be the existing degree, o f proteinuria on .the .'day of delivery for the. purpose of the analyses. Women were classified by having achieved the. combined adverse maternal outcome .or-not, and-descriptive analysis o f the predictor variables was performed. The frequency o f data avail-ability was determined. Parametric, (rtesc) and non-parametric (Mann-WltiriieyTJ jtest) analysts were-performed using SPSS (version 10) for Windows (Table 4). Al l P values < .05:are high-lighted in.Table 4, not withstanding that multiple comparisons are made, as Bonferro'ni correction was felt to be too strict in a • feasibility study of this nature. Univariable and multivariable analyses were conducted for all the candidate predictor variables available.in at least 4i)i) women on the dayvof admission. T h e dependent variable was the combined adverse maternal outcome (Table 3). As the dependent variable was dichotomous, logistic regression mod-elling was applied. Univariable logistic regression analysis was used to select all variables with a univariable, / 'value < .25 lor subsequent- multivariableanalysis. 3 6 T h e stttdy'was approved by.'the Univers i ty of Bri t ish Columbia Clinical Research Ethics Board and the Children's and Women's Health Centre of British Columbia Ethics Board. RESULTS O f the 594.(295 Vancouver, 149 Ottawa. 150 Oxford) women with preeclampsia recorded in-our database, 60 women (10.1 "<.) developed the combined adverse maternal outcome, includ-ing 1 maternal death. The most common outcomes t>10ra.ses) were >50% Q2 for mote than 1 hour In » 22,- o f whom 14 had at-least 1 other component* of the combined adverse outcome),, •wird,infused antihypertensive (n =20),'and transfusion o f mote • than 10 units : (total) of blood ..products- (n - • 10). -0.ne.oi t he 2 women who developed eclampsia suffered from recti n e n t eclampsia. None of the 594 women required either dialysis or renal-transplantation-. There was no;significam difference between centre variabil i ty in che' in 'cidenceof the adverse outcomes. The woman who died was 31, years old and'in her first preg-nancy, She had been admitted at 39* ! weeks wtthpreeclamp'sia. O n the day of delivery, at 40*" weeks, her 'maximum-blood, pres-sure was 180/100 mm Hg, and her blood tests showed -18 g pro-teinurta/d, 488 p M urate, 1. l 8 ; p M creatinine, and 10 u M bilirubin, Postnatally, she developed severe H E L L P syndrome.3'' H0 She developed a paradoxical thrombocytosis with her postnatal platelet count reaching 1.1.1.4 x 1 Q u / L on postnatal day 8, and died on her 37th postnatal day, fallowing a cerebral hemorrhage The data in.'Table 4 represent the descriptive statistics for selected variables on the day of admission and day of delivery, by occurrence {or not) .of the combined adverse maternal out-come. Included are the results of investigative parametric and non-parametric analyses. In women destined to develop the combined adverse maternal .outcome, gestattonalage'and fib-rinogen wereiower, and mean arterial pressure, total leukocyte count, and bilirubin were greater, on the day o f admission. In addition, to 'those 5 variables, die 'MPV-.platelet count ratio was lower, ancl levels of creatinine, aspartate/transaminase (AST), and lactate dehydrogenase .(I'.DH) were greater in women achieving the combined adverse outcome, on their day of deliv-ery. For some variables, such as 24-hour urinary protein, fib-rinogen, and bilirubin, only a minority had data collected. O f all the variables listed, only the gestational age at the time •of the.delivery admission, mean arterial pressure ( M A P ) , dip-stick protein, total leukocyte count, uric acid, creatinine, AST, L D H , and MPV:platelet count were subjected to univariable analysis (>-1C)0 cases), which revealed that, on the day of admis-sion, gestational age on admission,.MAP, dipstick proteinuria, and MPV:plat:elet count were associated,with the combined adverse outcome, In addition to these variables, on the day of delivery,, total leukocyte count, creatinine, AST, ancl L D H were associated w i t h the outcome (/:'< .25). A lower platelet count alone on admission was not associated with the outcome ( / '« -37), whereas an increased M P V was(.P = .02). At delivery, neither platelet count (P = ,001) nor M P V {/•* ~ .001) discrimi-nated as'well between those women developing the adverse out-' come and those who.did not, compared to the :MPV:platelet count ratio {/' < .001). As M P V and platelet count are die com-ponents o f the MPVtplateiet count ratio, their inclusion in the regression analysis-removed, the ratio as an independent predictor. The multivariable logistic regression, inc lud ing the MPV:plateler count ratio, but not including its components, revealed that the gestational age at admission (OR, 0.853; 95% confidence, interval [CI), 0.789-0.924), dipstick protein (OR, J .54; 9 5 % C l , 1.34-2.07), and MPVtplateiet count ratio (OR, 2908; 95% C l , 42.6-198358) were independent predictors of the combined-adverse maternal outcome. These factors indi-cate that for each week .of gestation gained prior to admission, the odds for developing the combined adverse maternal out-come was 0.853. meaning that increasing gestational age on admission protects against developing the adverse outcome. Similarly, lor every " + " increase in proteinuria, the odds for developing the outcome' increased by .t. 54-fold. F o r the MPVtplateiet-ratio, a t unit rise-in the ratio would not. be bio-logically plausible, whereas a 0.01 unit rise would be (Table 4). Such a rise increased the odds for the adverse outcome by 29-fold. PtSCOSSION This study provided evidence that the prospective development o f a model to predict adverse maternal outcomes in women admitted to ternary centres wirh the diagnosis of preeclampsia is feasible. It confirmed that gestational age atdiagnosisi'as reflected by admission) isan•important and independent predic-tor of adverse maternal outcomes, a fact: not currently acknowl-edged in the classification systems in general use,12"-'' further, o u t study indicated, that early'onset preeclampsia is more danger-ous for both the, woman •'arid beffetus.and we have proposed the subclass! fixation of preeclampsia to reflect this risk.1'-' The novel MPVpla tde t ratio, a summary measure of platelet consumption, wasxlesigoeet to better reflect platelet consump-tion than eithct element of the ratkratone. As platelets are coti-• su.med,,:the-megakaryocytes' in theoiarrow release immature large-volume platelets,'causing M P V to rise-prior rothe ongoing consumptive coagulopathy, overwhelming the ability of the mar-row to respond, at which ti me platelet counts fall. "I"herefore, th is summary measure appeared to amplify the effects of platelet o >n-sumption as predicted, although-ic is recognized that ratio's are intrinsically less stable mathematically than, are direct observa-tions. This ratio can be-easily introduced, as both elements-are included in complete blood count analyses, and at no extra-cost. Normal ranges for pregnane;,', need to be developed. It is possible that this ratio could have utility in other settings where platelet consumption-is of concern, such as for sepsis or immune thrombo-cytopocnic-purpura. Bilirubin was revealed as a'simiiarly prornis-itig variable in. this regard,.although.insufficient data were collected to.permit its inclusion in the mai.hivariable analysis. Proteinuria, which is included in.the C H S definition of "adverse features,"'™ was associated with worse maternal out-comes,despite the feet that-the proteinuria was measured by dipstick alone in almost 8 0 % o f cases. A l though dipstick measurement of proteinuria is associated with false-positive and false-negative rates,12 it.is inexpensive, and remains the standard o f care even in centres where protein-.creatinine ratio.-, are avail-able. While the C H S does not recommend dipstick screening for proteinuria,'1 our data show that -dipstick assessment may indeed be useful i n . triaging women with either, suspected or confirmed-prccclanipsia. The difference in M A P between tho-se'women who devel-oped the combined adverse maternal outcome and. those who did not probably- reflected the peaks.'of blood pressure'that occur, despite the'fact that blood pressure is die only compo-nent.of the.marernd:syndrcmitM>fpreeclanipsia-for-which there is effective therapy. Unfortunately, the monitor!ng.'of preeclampsia pregnancies in this study was idiosyncratic wi thin and between centres. Most variables were available in fewer than 80% o f women. Indeed, even 24-hour urinaiy protein estimation, upon which the firm diagnosis o f preeclampsia is:predicated, 'was-measured''in on ly P & C K W J | O C T O M R 5 0 0 4 Table 4. Data from Women With Preeclampsia, Compared by the Occurrence* or Absence of the Combined Adverse Maternal Outcome* Day of Admission Day of Delivery Variable N Mean (SD) Median (range] N Mean (SD) Median [range) Maternal Age (Year) 591 NO „. No adverse outcome 531 30.8 (6) 30.8 116, 461 ' NO ' ND ND Adverse outcome 60 31.1 (6) 30.7 (20, 431 NO ND' ND Gestational Age on Admission (Week) 545 - - NO - -No adverse outcome 487 34.4 (5)+ 34.7 (20, 421 ND ND ND Adverse outcome SB ND Mean Arterial Pressure (ram Hg)* 544 - - 585 - -No adverse outcome 491 121 (I3)§ 120 [83, 217] 527 120(13)§ 120 (76, 168! Adverse outcome 53 127 (14>§ 2? [100, 170] 58 128(14}§ 126 [100, 160| Sao.. (%)« 61 ... 157 - _ No adverse outcome 47 96 (3) 96 (87, 1001 130 9612) 96 [87, 1001 Adverse outcome 95 (2) 96 [92, 99) 27 95 (3) 9.4 [90, 100| Total Leukocyte Count (X10VU 496 - - 55.9 - ... No adverse outcome 444 11.6 (4)§ 10,8 14, 25| 501 ?3.8{5)§ 13.0 [4, 33! Adverse outcome 52 12.7 (4)§ 12.0 |'6, 30| 58 16.0 (6)§ Urate (uM) 494 561 No adverse outcome 442 370 (90) 3.39 (136, 648) 504 382 (94) 376 [136, 683) Adverse outcome 52 375 (98) 365 [172, 666) 57 400 19.4) 3861190, 666) Creatinine (jiM) 392 - - 47'0 - _ No adverse outcome 341 71 (20) 70 [27, 204] 415 *74Q4)§ 73 |27, 236] Adverse outcome „ . , ^ „ 73 (23) 71 (28, 1391 55 S2'36)§ ' 80 [34, 2161 Urinary Oulpul 221 - ... 35 3 _ _ No adverse outcome 200 109 (65) 90 (18, 3751 320 12-1 (67) 108 {15, 575] Adverse outcome 21 85 (36) 87 (1S, 1611 33 117(96) 87 [5, 5161 Bilirubin (tiM)** 121 ... _ 203 - _ No adverse outcome 98 .8.9 (13)1 6 11,93) 165 9,9 (14)+ 6 |2, 93| Adverse outcome '23 21,6 '38)1 .9 |1, 139| 38 21.8 (34S+ " ' 3 J 3 , l 3 9 ) Aspartate Transaminase (U/L) 494 - - 555 - .... No adverse outcome 445 67(191) 26 |9, 3340J 500 94 (24! it 28 (10, 3340)1 Adverse outcome 49 120 (388) 29 112, 2704) 55 395 <975)t 40 (10, 5000)1 lactate Dehydrogenase (U/L) 371 - - 495 -No adverse outcome 335 573 (470) 304 [103, 4.392! 4 43 619 (549)t 513 169,58471 Adverse outcome 36 612 (853) 296 (164, 3759] 52 1321 (2926)4 .386 1137, 200201 Plasma Albumin (g/L)** 146 - - 187 -No adverse outcome 124 31.5(3) 31 121,40! 155 28.9 (5) 29.115, 43! Adverse •oulcome 22 31.0.4) 30 )24, 36) 32 27.4 C4) 29 (19, 35) Fibrinogen. (jiM)** 86 - - 139 _ No adverse outcome 76 4.8 (Mi§ 4.8 11,9] 121 4,7 (1,5)+ • 4,7 11, 91 Adverse outcome 10 3.9(!.3)§ 4.0 12, 6| '18 .3,5 (1.3)+ 3.5 [1,7| MPV:Platelet Ratio (fL/10'i 418 ... 520 _ No adverse outcome 379 0.O6!) i0.03) 0.057 (0.02, 0,44) 471 0.077 (0.06)+ _ 0.059 10.02, 0.48) Adverse; outcome 39 it 073 i0.06,i 0,060 10.02, 0.33] 49 0.135 (0.12)+ i j Table 4. (continued from previous page) Day of Admission Day of Delivery Variable N Mean (SD) Median [rangel N Mean (SD) Median (range) Estimated Fetal Weight <%ile CA)** 119 . _ _ 1.58 . . . No adverse outcome 107 38.7(26.4) 35.5 11,97] 137 30.4 C2&.6. 33 (1,971 Adverse outcome " i f " ' 32,4(29.0) 13 11, 92] 21 31.1 (27.85 1.3 st, 92i Amniotic Fluid index CA)** 119 - - 127 -K'o adverse outcome 107 23.4 (23.6) 16 (1, 88] 113 24.0(24.1) 16 1.1,88] Adverse outcome 12 23.8 (24.8) '11.5 [1, 631 14 23.9(22.4) .15.5 [1, 63] Present Absent/Reversed Present Absent/Reversed Umbilical Artery End Diastolic Flow** 119 127 No adverse outcome 102 86 16 107 91 16 Adverse outcome 1*2" 5 20 14 6 'SO: standard deviation: NO: no ijiffereiCB. SrtQj: oxygen saturation (pulse oximotrv, imm »!<!; .MPV: mean platelet volume- OA: mu&ana age •*F«.00s. itea or Mann-Whtaey U lest as appropriate, iAte-iin after!,,! preso.'e; diastolic bland preswe • Cputse pressure/a} |P «.t», t test or M*nr»-Whiit>cy U test as appropriate. -'Mot Included in umvaniible analysis. t /"<i)i,test or Mann-Whitney u (est, as appropriate. 21% of cases. Therefore, we were underpoweted to test whether or nor variables such as fibrinogen and bilirubin, which are pow-erful predictors o f death in adult intensive care unit ( ICU) set-tings, when incorporated into predictive models , 2 8 " 3 0 , 3 8 might have true utility in the setting o f preeclampsia. Predictive mod-els, such as acute physiology and chronic health.evaluation { A P A C H E ) 2 8 , 2 9 , 3 8 and multiple organ dysfunctional syndrome ( M O D S ) , * have been developed for patients with SIRS, which preeclampsia resembles to a remarkable degree. 7 3 1 Although these scores perform well when modified for defined popula-tions,"*'' most have been generated-to reflect pathophysiology in predominately geriatric populations.'" 0 The APACHE.sco re , specifically, did not. perform well with eclampsia in the I C U set-ting, with the Glasgow coma scale outperformttig A P A C H E II in predicting maternal mortality. 4 ' We were also unable to assess the influence of medical inter-vention on the outcomes-being tested, especially as- there was variability within and between centres in the expectant man-agement of preeclampsia remote from term. It is anticipated that there was considerable heterogeneity in the care of women enrolled in this study. The study was based on groups that were ordinarily variable and, therefore,'were representative samples of the disease as it presents in most centres. Variations in case-mix and management reflect the real world and have made this study gcneralizable. The data presented did. reveal that some variables changed between admission and delivery, reflecting the need to monitor these women closely across a range of organ systems. At present, however, it is unclear which tests in which organ systems might provide the most utility in predicting the clinical course of any particular woman with preeclampsia. This study has implications for the monitoring o f women with preeclampsia and argues for the development o f a disease-* specific outcome prediction m o d e l * There is international sup-port for this approach. Such a predic t ion model couic lbe developed using die combined adverse maternal outcome used in this study and should be tested against die current dichoto-mous definitions of "mild vs. ;severeM'or "with'vs: without adverse features" using the area under the receiver operator curve.analy-ses.4" Funding has recently- been received for the .development: -phase of such a project (Preeclampsia Integrated Estimate .of RiSk [PIERS] model; principal investigator: P. von Dadelszen)'. Given the limitations of the. present-study, i t is .believed-that this new project-should be a prospective study that would ensure full documentation.of-the candidate variables at each epoch of interest. As clinical care varies both within and between centres, prospective data collection in a number o f centres would iden-tify those variables thatare both most useful in predicting the outcome and gerieralizable. Maternal symptoms were excluded from thisstudy, despite their inclusion in the C H S definition o f "adverse'features" and recent evidence to support their utility .-in predicting adverse maternal outcomes.' 1 3''' 4 Although we recognized that maternal symptoms remain clinically influential and are important in the surveillance of women with preeclampsia as the .disease evolves, it was felt tiiat symptoms could not be reliably gleaned through a retrospective chart review. Nevertheless, maternal symptoms and other factors, such as family history, should be assessed for their uti l i ty in predict ing adverse maternal outcomes in a prospective manner. j O G C [ITS? I OCTOBER 2004 CONCLUSIONS 3. Pregnancy-related mortality surveillance: United States, 1987-1990. MMWR Morb Mortal WkSy Repl997;46(SS-4):l7-36. l i t e women at greatest risk for adverse outcomes in preeclamp-sia are those presenting at early gestational ages and those with greater dipstick/proteinuria. Platelet consumption, as reflected by MPV;platekt count ratio., also appears to be predictive. How-ever, a comprehensive prospective evaluation of the association between all candidate .predictor variables is required, includ-ing maternal symptoms and adverse maternal outcomes, such: as is currently being evaluated usingthe PIERS model. Only then will those women with preeclampsia lor whom delivery is clearly mandated, as well as those most appropriately managed expectantly, be identified. ACJrWOWLED^MiB^S We gratefully acknowledge the. efforts of Shelley Soanes, Vcsna Popovska, Terry Viczko, Margaret Gluxynski, and the Medical Records Departments in Vancouver and Ottawa in data retrieval and entry. The support of the Maternal .Medicine Unit staff', Oxford, is also acknowledged, especially that of Professor Chris Redman, for giving us access to the Units charts and for his use-ful advice during preparation of the manuscript; David Jung and Boris Kuzeljvic helped co develop the database, and Ruth Mi l -ner gave advice and guidance throughout the project-This pro-ject was funded by grants to Peter von Dadelszen, Ltura Magee, the Centre for Healthcare Innovation-and Improvement, and the Clinical Research Support Unit from the BC Research Insti-tute for Children's and Women's Health and the BC Women's Hospital and Health Centre Foundation. Shoo Lee is a Medical Research C o u n c i l (Canada) scholar, and Laura Magee is a Michael Smith Foundation for Health Research scholar. The members of the Delphic consensus were, from Canada: I! von Dadelszen, Vancouver (maternal-feral medicine [MFM]); L . A . Magee. Vancouver (obstetric internal medicine [OlM]) ; M . .)• Douglas.. Vancouver (obstetric anaesthesia); K. R. Walley, Vancouver (critical care medicine { C C M ) ) ; J . A . Russell, Van-couver (CCM); A. Gruslin, Ottawa ( M F M ) ; J.-M. Moutquin, Sherbrooke ( M F M ) ; from the United States: J. M . Roberts, Pittsburgh ( M F M ) ; from the United Kingdom:.S. Robson, Newcasdc-on-Tyne ( M F M ) ; M . de Swiet. London ( O l M ) ; ] , J , Walker, Leeds ( M F M and O I M ) ; from Austral ia: M . A . Brown, Sydney (OlM); G . Davis, Sydney ( M F M ) ; and from Hew Zealand: L. A . McCowan, Auckland ( M F M ) . REFERENCES 1. Department of Health.Why women die. Report on confidential enquiries into maternal deaths in the United Kingdom 1994-1996. London: HI-ISO; 1999. 2. National Institute for Clinical Excellence.Why women die. Report on confidential enquiries-into maternal deaths in the United Kingom, 1997-1999. London: RCOG Press; "2001. 4. Roberts JM, Redman CWG. Pre-eclampsia: more than pregnancy-induced hypertension. Lancet 1993:341:1447-51. •5; Mages LA,.Oi*i-($t€irtMP,'von:DadoIszen ft Fortnightly review:manage-ment of hypertension in pregnancy. BMJ 1999:318:1332-6. 6. Magee LA, Cham C.Waterman EJ. Ohlsscn A. von Dadelszen R Hydralazine for treatment of severe hypertension in pregnancy, meta-analysis, BMj 20Q3;327:9S5~60. 7. von Dadelszen P, Magee LA, Marshall jC, Rotstein OD.The maternal syndrome of preeclampsia: a forme fmste of the systemic inflammatory response syndrome. Sepsis 2000;4:43-7. 8. von Dadelszen P, Magee LA, Lee SK. Stewart SD. Sirnone C, Koren G, et al. Activated protein C in normal human pregnancy and pregnancies complicated by severe preeclampsia: a therapeutic opportunity? Crit Care Med 2002;30:1883-92. 9. Stevenson DK.Wright LL Lemons JA. Oh W, Koror.es SB, Papilc LA. et al.Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network, January 1993 through December 1994. Am] Obstet Gynecol 1998:179:1637.-9. 10. Odendaa! HJ.Pattinson RC.Bam R,Grove D.KotzeTj. Aggressive or expectant management for patients with severe.preeclampsia between 28-34 weeks" gestation: a randomized controlled trial. Obstet Gynecol 1990:76:1070-5. ! 1. Olah KS. Redman CW, Gee H. Management of severe, early pre-eclampsia: is conservative management Justified? Eur J Obstet Gynecol Reprod Biol 1993:51:175-80; 12. : Helewa ME, Burrows RF. Smith J,Williams K. Brain R Rabkih-SW, Report of the Canadian Hypertension Society Consensus Conference: I. Definitions, evaluation and classification of hypertensive disorders in pregnancy. CMAj i 997:157:71S-25. 13. Report of the National High Blood Pressure Education Program Work-ing Group on High Blood Pressure in Pregnancy. Am j Obstet Gynecol 2Q0Q;I83:SI~522. 14. Brown MA. HagueWM.Higgins). Lowe S. McCowan L.Oats J.et ai. The detection, investigation and management of hypertension in preg-nancy: full consensus statement Aust-N Z J Obstet Gynaecol 2000:40: (39-55. 15. Brown MA, Lindheimer MD, de Swiet M.Van Assche A. Moutquin JM. The classification and diagnosis of the hypertensive disorders of preg? nancy: statement from the Internationa! Society for the Study of Hypertension in Pregnancy (ISSHP). Hypertens Prcgn 2001 ;20:ix-xiv. 16. Douglas KA. Redman CWG. Eclampsia in the United Kingdom. BMJ 1994;309:1395-400. 17. Caritis S. Sibai B. Hauth J. Lindheimer MD, Klebanoff M.Thom E, et a!. Low-dose aspirin to prevent preeclampsia in women at high risk. National Institute of Child Health and Human Development Network of MaterralTetal Medicine Units.N Engl'j Med 1998;338:701-5. 18. Buchbinder A, Sibai BM, Caritis S. Macpherson C. Hauth j , UndhefmerMD.et al, Adverse perinatal outcomes are significantly, higher in severe.gestational hypertension than in mild preeclampsia. Am J Obstet Gynecol 2002; 186:66-71. 19. von Dadelszen P. Magee LA, Roberts JM. Subclassification of preeclamp-sia. Hypertens Pregn 2003:22:143-8. 20. MacKay AP, fierg CJ. A trash HK. Pregnancy-related mortality from preeclampsia and eclampsia. Obstet Gynecol 2001 ;97:533-8, jee-c gTQijj octosa JOW 21. Opsjon SL. Austgulen R,Waage A. Interleukin-1, interleukin-6 and tumor necrosis factor at delivery in preeclamptic disorders; Acta Obitet Gynecol Scand 1995:74:19-26. 22. von Oadelsren R Magee LA Krajden H. Afasaly K, Popovska V, Devarakonda RH, et al. Levels of. antibodiesagairKt cytomegalovirus and Qikmy'dophifa pneumoniae are increased in-early onset pre-eclampsia. Br J Obstet Gynaecol 2003; 110:725-30. 23. Sibai BM, Mercer BM, Schiff E, Friedman SA. Aggressive versus expec-tant management of severe preeclampsia at 28 to -32 weeks' gestation: a randomized controlled trial. Am J Obstet Gynecol 1994; 171:818-22. 24,. Redman CWG.Platelets and the beginnings of preeclampsia. N Engl jMed 1990;323:478-80. ' 25. Redman CW, Bonnar J, Beilin I, Early platelet consumption in pre-eclampsia. BMj 1978;1:467-9. 26. Rosevear SK. LigginsGC. Platelet dimensions in pregnancy-induced hypertension. N Z Med j 1986;99:356-7. 27. Walker jj, Cameron AD, Bjornsxon S, Singer CR. Fraser C. Can platelet volume predict progressive hypertensive disease in pregnancy? Am j Obstet Gynecol 1989;161:676-9. 28. Knaus W.Wagner D, Draper £. APACHE 111 study design:analytic plan for evaluation of severity and outcome in intensive care unit patients. Implications. Crit Care Med 1989; 17:S219-S221, 29. Knaus W.Wagner D, Draper E APACHE III study design: analytic plan for evaluation of severity and outcome in intensive care unit patients. Development of APACHE. Crit Care Med 1989:17:S 181-S185. 30. Marshall J C Cook Dj.Christou NV, Bernard GR, Sprung CL. Sibbald Wj. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med 1995;23:1638-52. 31. Redman CWG, Sacks GP, Sargent IL Preeclampsia: an excessive maternal inflammatory response to pregnancy. Am J Obstet Gynecol 1999:180:499-506, 32. Miiholland AV.Wheeler SG, Heieck Jj. Medical assessment by a Delphi group opinion technic, N Engi J Med 1973;288:1272-5. 33. Pill J.Tlie Delphi method: substance, context, a critique and an annotat-ed bibliography. Socso-Economic Planning Science.1971;5:57-71, 34. Richardson D.Tarnow-MordiWO, Lee SK. Risk adjustment for quality, improvement Pediatrics 1999;I03;255~6S.-35. Audibert fi FriedmanSA, Frangieb AY, Sibai BM; Clinical utility, of strict diagnostic criteria for the HELLP (hemolysis, elevated liver enzymes, and low platelets) :syndrome. Am j Obstet Gynecol 1996:175:460-4. 36. Hosmer D, Lemeshow S. Assessing the fit of the model, Iri: Hosmer D. Lemeshow S, editors: Applied logistic regression. New York: John Wiley and Sons; 1989. p. 135^75, 37. Weinscein L Syndrome of hemolysis, elevated liver enzymes, and few platelet count a severe consequence,of hypertension in pregnancy. Am j Obstet Gynecol 1982:142:159-67. 38. KrausW.Wagner D. APACHE 111 study design: analytic plan for evalua-tion of severity and outcome in. intensive care unit patients. Individual patient decisions, Crit Care Med 1989; 17:S204^S209;; 39. Jacobs 5, Zuleika M, Mphansa T.The Multiple Organ Dysfunction Score as a descriptor of patient outcome in septic shock compared with two other scoring systems, Crit Care Med I999;27:74l~4, 40. Belfort MA. Scoring systems for eclampsia. Crit Care Med 2000:28: 272-3. 4 i , Bhagwanjee S, paruk F, Moodley J, Muckart DJ. Intensive care unit mor-bidity and mortality from eclampsia: an evaluation of.theAcute Physiol-ogy and Chronic Health Evaluation II score and the Glasgow Coma Scale score. Crit Care Med 2000;28:120-4. 42. Richardson DK,GrayJE,McCormick MC.Workman K. Goldmann DA. Score for Neonatal Acute Physiology; a physiologic severity index for neonatal intensive care,Pediatrics;;l?93;9|:6|7-23, 43. Piguel D, Pierre F, Pourrat O, D'Halluin G. Magrtin G; Are the symp-toms relevant enough to predict the occurrence of pre-eclampsia? Hypertens Pregn 2002:21:51. 44. Stoger S.Walters 8N, Diagnostic value of the clinical history in pre-eclampsia; Hypertens Pregn 2002:21:48. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092000/manifest

Comment

Related Items