Open Collections

UBC Faculty Research and Publications

Cardiovascular risk is similar in patients with glomerulonephritis compared to other types of chronic… Hutton, Holly L; Levin, Adeera; Gill, Jagbir; Djurdjev, Ognjenka; Tang, Mila; Barbour, Sean J Mar 20, 2017

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

Item Metadata

Download

Media
52383-12882_2017_Article_511.pdf [ 657.01kB ]
Metadata
JSON: 52383-1.0366837.json
JSON-LD: 52383-1.0366837-ld.json
RDF/XML (Pretty): 52383-1.0366837-rdf.xml
RDF/JSON: 52383-1.0366837-rdf.json
Turtle: 52383-1.0366837-turtle.txt
N-Triples: 52383-1.0366837-rdf-ntriples.txt
Original Record: 52383-1.0366837-source.json
Full Text
52383-1.0366837-fulltext.txt
Citation
52383-1.0366837.ris

Full Text

RESEARCH ARTICLE Open AccessCardiovascular risk is similar in patientswith glomerulonephritis compared to othertypes of chronic kidney disease: a matchedcohort studyHolly L. Hutton1,2*, Adeera Levin3,4,5, Jagbir Gill3,5, Ognjenka Djurdjev4, Mila Tang4 and Sean J. Barbour3,4,5AbstractBackground: Patients with chronic kidney disease (CKD) due to glomerulonephritis (GN) are thought to be at highrisk for cardiovascular disease (CVD). However, no study has examined whether GN directly contributes to CV riskbeyond the effects conferred by pre-existing traditional risk factors and level of renal function.Methods: Matched cohort study using the previously described prospective CanPREDDICT study cohort. 2187 patientswith CKD (eGFR 15–45 ml/min/m2) from 25 Canadian centres were divided into GN vs non-GN cause of CKD. Patientson immunotherapy for GN were not included in the study. Standardized measures of CV risk factors, biomarkers andCV outcomes were recorded over 3 years of follow-up, with the primary outcome measure being time to first all-causeCV event.Results: In the overall cohort, CV events occurred in 25 (8.7%) of the GN group and 338 (17.8%) of the non-GN group(HR 0.45, 95% CI 0.30–0.67, p < 0.01). In a Cox regression multivariable model that included age, sex, prior diabetes andCVD, baseline eGFR and onset of renal replacement therapy, the risk of CV events was similar in the GN and non-GNgroups (HR 0.71, 95% CI 0.47–1.08, p = 0.11). GN and non-GN patients were matched by age and using a propensityscore including sex, prior diabetes and CVD and baseline eGFR. In the matched group, the risk of CV events was similarin GN vs non-GN patients (N = 25/271 (9.2%) in both groups, HR 1.01, 95% CI 0.05–1.77, p = 0.9). An interaction analysisshowed that CRP, ACR and troponin conferred differing amounts of CV risk in the GN and non-GN groups.Conclusions: Patients with advanced CKD due to GN have a high 8.7% absolute 3-year risk of CVD, attributable to priorCV risk factors and level of kidney function rather than the GN disease itself.Keywords: Cardiovascular disease, Glomerulonephritis, Chronic kidney diseaseBackgroundPatients with glomerulonephritis (GN) have been trad-itionally characterised as being at high risk of cardiovas-cular disease (CVD), and this has been recentlyreiterated in the 2012 KDIGO GN guidelines [1]. GNpatients are recognised to develop traditional epidemio-logic risk factors for CVD, including hypertension andhyperlipidemia [2–4], as well as novel CV risk factorssuch inflammation, endothelial dysfunction and protein-uria [2, 5–7]. In addition, GN patients frequentlydevelop chronic kidney disease (CKD) with impairedkidney function. Irrespective of cause, CKD is associatedwith a high risk of CVD and CV death [8–10]. Priorstudies investigating CVD in GN have not accounted forthese risk factors [2, 11, 12], and as such it remainsunknown whether the high risk of CVD in GN patientsis attributable to the disease itself or to the presence ofconcurrent CVD risk factors and CKD with low renalfunction.CanPREDDICT is a prospective Canadian cohortstudy of 2544 patients with GN and non-GN CKD with* Correspondence: holly.hutton@monash.edu1Centre for Inflammatory Diseases, Monash University, Clayton, VIC, Australia2Dept of Nephrology, Monash Health, 246 Clayton Rd, Clayton, VIC 3168,AustraliaFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Hutton et al. BMC Nephrology  (2017) 18:95 DOI 10.1186/s12882-017-0511-zstandardized assessment of CV risk factors, CVD out-comes and biomarkers of inflammation, endothelial dys-function and proteinuria. We used the CanPREDDICTcohort to examine our primary hypothesis that the riskof CVD over 3 years in patients with GN is higher com-pared to those with non-GN causes of CKD, afteraccounting for traditional CVD risk factors and renalfunction. Because novel CV risk factors are proposed tobe uniquely important in glomerular diseases [7, 13, 14],we additionally explored whether the association be-tween CVD risk and proteinuria or biomarkers ofinflammation and endothelial dysfunction is different inpatients with GN compared to non-GN CKD.MethodsDetails of the CanPREDDICT study have been previouslydescribed [15]. In brief, 2544 patients with CKD (eGFR15–45 ml/min/m2) from 25 Canadian centres were re-cruited from 2008 to 2009, and were prospectivelyfollowed for 3 years for CV outcomes with standardizedmeasurements of multiple biomarkers. Patients with GNon immunosuppression were specifically excluded fromCanPREDDICT. We included in our analysis those pa-tients from the CanPREDDICT study with no missingdemographic or biomarker data at baseline and with aknown cause of primary kidney disease.Data collectionAll patients were categorized into 2 groups (GN andnon-GN) based on their primary renal diagnosis pro-vided at the time of recruitment into the CanPREDDICTcohort. The cause of primary kidney disease was chosenfrom a list of options by the nephrologist at the time ofrecruitment into the study, with qualifying descriptionsprovided in free text format. An investigator (HH)blinded to other data and outcome status reviewed therenal diagnosis details to classify patients into non-GNor GN groups. GNs were further categorized as IgA ne-phropathy, membranous nephropathy, lupus nephritis,anti-neutrophil cytoplasmic antibody (ANCA) associatedvasculitis, focal segmental glomerulosclerosis (FSGS) orGN unspecified. Demographics, medications, bloodpressure (BP) and comorbidities were collected at base-line and every 6 months over the 3-year study period.Blood and urine samples were collected at baseline. Pro-teinuria was assessed using urine albumin to creatinineratio (uACR). Biomarkers known to be associated withCVD in the general CKD population were measured aspreviously described, including high sensitivity C-reactive protein (CRP) [16], troponin I [17], asymmetricdimethylarginine (ADMA) [18], interleukin-6 (IL-6) [19]and N-terminal pro-brain natriuretic peptide (ProBNP)[15–17].Definitions of outcomesThe primary outcome was the first occurrence of an all-cause CV event. All CV outcomes were centrally adjudi-cated based on source documentation by a blinded panelof three investigators using standardized definitions aspreviously described [15]. An all-cause CV event wasdefined as fatal or non-fatal myocardial infarction (MI),need for coronary revascularization (coronary arterybypass graft/percutaneous coronary intervention/percu-taneous transluminal coronary angioplasty), ischemicstroke or congestive heart failure.Statistical analysisAnalysis of the primary outcome was based on the timefrom entry in the CanPREDDICT study to the first oc-currence of an all-cause CV event, censored at death orthe end of follow-up.Comparing the risk of CV events in GN compared to non-GN CKD patients in the non-matched cohortSurvival without the primary CV endpoint was describedin GN vs non-GN patients using the Kaplan Meiermethod and compared using the long-rank test. To con-trol for renal function and traditional CV risk factors,we used Cox regression multivariable models that in-cluded GN vs non-GN CKD, age, sex, prior diabetes andCVD, baseline eGFR and onset of RRT as a time-dependent variable (to account for any confounding ef-fect of differential progression to end stage renal diseaseon the risk of CV events).Comparing the risk of CV events in GN to a matched cohortof non-GN CKD patientsWe matched GN to non-GN patients using a stepwiseapproach: first direct matching on age (+/− 2.5 years),and second using a propensity score that included gen-der, history of diabetes, CVD and baseline eGFR. Match-ing was 1:1 using a calliper width of 0.25 standarddeviation and a best-overall fit algorithm. Using thematched cohort, time to first all-cause CV event cen-sored at death or end of follow-up was compared in GNvs non-GN CKD patients using a shared frailty Coxmodel to account for clustering on matched pairs.Comparing the association of select biomarkers and CV riskin GN and non-GN patientsTo investigate differences in the association between se-lect biomarkers and CV risk in the GN and non-GNgroups, an interaction term between each biomarker anddisease type was included with the main effects in thematched cohort models. Using these models, weexpressed the hazard ratios for the association betweeneach biomarker and CVD separately in the GN and non-GN groups.Hutton et al. BMC Nephrology  (2017) 18:95 Page 2 of 10Because of the significant impact of diabetes on CVDrisk, we performed stratified analyses in which thematching algorithm and analysis were repeated in sub-groups with and without diabetes to explore consistencyof results. The existing literature suggests that hyperten-sion and dyslipidemia may be both directly caused byGN and result in CVD [4, 11, 14, 20–22]. As such, thesetwo variables can be considered in the causal pathwaybetween GN and CVD, and so we did not include themin our primary analyses. However, we performed a sensi-tivity analysis that included adjustment for both bloodpressure and cholesterol levels. Variables with highlyskewed distributions were transformed to the naturallogarithm scale. Categorical variables were described asfrequency (count) and compared across groups usingFischer’s exact test. Continuous variables with normaldistributions were described as mean [standard devi-ation] and compared across groups using the t-test, andvariables with non-normal distributions were describedas median [interquartile range, IQR] and comparedusing the Wilcoxon Rank Sum test. All analyses wereperformed using SAS software, version 9.3 (SAS InstituteInc., Cary, NC, USA) and R software, version 3.1.0. Alltests were two-sided with P-values <0.05 considered sta-tistically significant.ResultsThere were 2544 patients in the CanPREDDICT studywith 2187 in the analytic cohort, including 288 with GNand 1899 with non-GN CKD (see Fig. 1). Baseline char-acteristics of the cohort in the GN compared to non-GNpatients are shown in Table 1. GN patients had lesstraditional CV risk factors, with younger age (58.9 vs69.7 years, p < 0.01), less frequent diabetes (20% vs 54%,p < 0.001) and prior history of CVD (28% vs 47%, p <0.01), and lower mean systolic BP (131 vs 134 mmHg, p< 0.01). GN patients had significantly higher uACR (59.5vs 14.4 mmol/L, p < 0.01) and total cholesterol (179 vs161 mg/dL, p = 0.03); lower ProBNP (273 vs 511 pg/mL,p < 0.01), IL-6 (3.04 vs 4.54, p < 0.01) and CRP (2.2 vs3.0, p < 0.01) but similar eGFR (27.1 vs 27.7 ml/min/1.73 m2). Progression to RRT during follow-up wasmore common in the GN compared to non-GN group(25.4% vs 15.8%, p < 0.01).The risk of CVD in GN vs non-GN CKD patients in thenon-matched cohortThere were a total of 363 CV events in the entire cohortover the 3-year period, with 25 events in the GN group(8.7%), and 338 events (17.8%) in the non-GN group.The most common event was fatal or non-fatal MI (N =166, including 9 GN and 157 non-GN), followed byCHF (N = 122, including 8 GN and 114 non-GN), stroke(N = 51, including 5 GN and 46 non-GN) and coronaryrevascularization (N = 24, including 3 GN and 21 non-GN). Figure 2a outlines the CV event free survival inGN compared to non-GN patients. GN patients had su-perior CV event free survival compared to non-GN pa-tients (unadjusted HR = 0.45, 95% CI 0.3–0.67, p < 0.01).Table 2 displays the result of the multivariate model forthe risk of CV event in GN versus non-GN patients.After adjustment for potential measured confounders,the risk of CV events was similar in the two groups (HR= 0.71, 95% CI 0.47–1.08, p = 0.11). In a sensitivityanalysis, when mean arterial blood pressure and totalFig. 1 Flow diagram of patients included in the analytic cohortHutton et al. BMC Nephrology  (2017) 18:95 Page 3 of 10cholesterol were added to the model the results were un-changed (data not shown).The risk of CVD in GN compared to a matched cohort ofnon-GN CKD patientsTo further control for renal function and CV risk factorsin the association of GN and CV risk, we matched 272patients from the GN group with 272 patients from thenon-GN group based on age, sex, eGFR and prior dia-betes and CVD. Characteristics of the matched cohortare shown in Table 3. Differences in age, BP, prior dia-betes and CVD and total cholesterol between GN andnon-GN patients that were seen in the overall cohortwere no longer present after matching. GN patients hadhigher uACR (58.7 vs 17.2 mg/mmol, p < 0.01) and lowerserum albumin (40 vs 41 g/L p < 0.01), but there wereno significant differences in troponin I, IL-6, Pro-BNP,ADMA or CRP between the groups (Table 3). Progres-sion to RRT during follow-up occurred in 24.6% of thenon-GN group and 25.4% of the GN group (p = 0.84) inthe matched cohort.The frequency of CV events over 3 years in thematched cohort was 9.2% (N = 25) in both the GN andnon-GN groups. Figure 2b shows that the CV event freesurvival was similar in the GN patients compared to thematched cohort of non-GN patients (log-rank p-value0.96). The risk of CV events was similar in the GN com-pared to non-GN patients in both univariable (HR 1.01,95% CI 0.58–1.77, p = 0.96) and multivariable models(HR 0.99, 95% CI 0.56–1.75, p = 0.96, see Table 4).Because of residual differences in proteinuria betweenthe groups after matching, we additionally added uACRto the multivariable model which did not change our re-sults (GN vs non-GN HR 0.87, 95% CI 0.49–1.58, p =0.66). In a sensitivity analysis, when mean arterial bloodpressure and total cholesterol were added to the modelthe results were unchanged (data not shown).We repeated the matching and analysis in diabetic andnon-diabetic subgroups. Of the 112 matched patientswith diabetes, 21.4% (n = 12) of both the GN and non-GN patients had a CV event. Of the 388 matchedpatients without diabetes, CV events occurred in 6.2%(N = 12) and 8.2% (N = 16) in the GN and non-GNgroups respectively. In multivariable models, there wasTable 1 Characteristics of the GN and non-GN patients in thecohort. Data are presented as mean ± standard deviation (SD),median (IQR) or count (frequency)Variable GN Non-GN P-valueNumber 288 1899Median followup (months)39.0 (33.5–39.0) 39.0 (22.5–39.0)Age (years) 58.9 ± 15.1 69.7 ± 11.7 <0.001Male (%) 189 (66%) 1184 (62%) 0.3Caucasian (%) 241 (84%) 1703 (90%) <0.001Primary cause ofkidney disease (%)Diabetes - 777 (41%)Hypertension - 823 (48%)PCKD - 108 (6%)Other non-GN - 252 (13%)GN subtypes (%)IgA Nephropathy 61 (21%) -FSGS 35 (12%) -ANCA Vasculitis 21 (7%) -Lupus Nephritis 13 (5%) -MembranousNephritis9 (3%) -GN Unspecified 149 (52%) -Diabetes (%) 58 (20%) 1025 (54%) <0.001CVD history (%) <0.001No CVD 206 (72%) 998 (53%)Ischemic HD 33 (12%) 365 (17%)CHF 29 (10%) 215 (11%)Ischemic and CHF 20 (7%) 321 (17%)Mean eGFR (ml/min/1.73 m2) ++/SD27.1 ± 9.1 27.7 ± 8.9 0.3eGFR categories (%) 0.1< 20 ml/min 81 (28%) 428 (23%)20–29 107 (37%) 757 (40%)> 30 100 (35%) 714 (40%)uACR (mg/mmol) 59.5 [13.7–172.3] 14.4 [2.9–76.8] <0.001Systolic BP(mmHg)131 ± 18 134 ± 20 <0.001Diastolic BP(mmHg)75 ± 12 70 ± 12 <0.001Albumin (g/L) 40.0 ± 4.8 40.4 ± 4.2 0.08Total cholesterol(mg/dL)4.6 ± 1.2 4.2 + 1.1 0.03Elevated TroponinI (% > LLD)71 (25%) 714 (38%) <0.001CRP (mg/mL) 2.2 [0.9–5.3] 3.0 [1.2–6.9] <0.001ADMA 0.54 ± 0.94 0.55 ± 0.12 0.1IL-6 (μg/L) 3.04 [1.00–5.80] 4.54 [1.00–7.25] <0.001Table 1 Characteristics of the GN and non-GN patients in thecohort. Data are presented as mean ± standard deviation (SD),median (IQR) or count (frequency) (Continued)NT-ProBNP (pg/mL) Pg/mL273 [119–727] 511 [213–1485] <0.001Abbreviations: PCKD polycystic kidney disease, FSGS focal segmentalglomerulosclerosis, CVD cardiovascular disease, CHF congestive heart failure,BP blood pressure, LLD lower limit of detection, ADMA asymmetricdimethylarginine, IL-6 interleukin 6, NT-Pro-BNP N-terminal pro-brainnatriuretic peptideHutton et al. BMC Nephrology  (2017) 18:95 Page 4 of 10no difference in the risk of CV events between GN andnon-GN patients in either the diabetic or non-diabeticsubgroups (data not shown).Biomarkers as CV risk factors in GN compared to matchednon-GN CKD patientsUsing the matched cohort, we explored whether the riskof CVD associated with ACR, CRP, IL-6, ADMA, tropo-nin I and ProBNP was different in the GN compared tonon-GN groups using interaction terms. The hazard ra-tios for each biomarker by GN group are presented inTable 5 and main effects in Additional file 1: Table S1.Although the interaction terms were not statistically sig-nificant (p-values 0.06–0.94), there was a suggestion ofquantitative differences in the hazard ratios for uACR,CRP and troponin I in the GN compared to non-GNgroups (interaction p-values 0.15, 0.24 and 0.06 respect-ively). The hazard ratio for the association betweenuACR and CV risk was higher and statistically signifi-cant in the non-GN group but not in the GN group. InFig. 2 The probability of survival without an all-cause CV event in GN compared to non-GN patients in a) the overall cohort prior to matching(log-rank p-values <0.01), and b) in the matched cohort (log-rank p-values = 0.96)Table 2 The results of univariable and multivariable Cox regressionmodels for the risk of all-cause CV events in the overall non-matchedcohortHR 95%CI P-valueUnivariable ModelGN vs non-GN CKD 0.45 0.30–0.67 <0.001Multivariable ModelGN vs non-GN CKD 0.71 0.47–1.08 0.1Age 1.03 1.02–1.04 <0.001Male sex 1.05 0.85–1.31 0.6Diabetes 1.71 1.37–2.13 <0.001Prior CVD 2.28 1.82–2.87 <0.001Baseline eGFR 0.97 0.96–0.99 <0.001RRTa 2.01 1.38–2.93 <0.001a as a time-dependent variableHutton et al. BMC Nephrology  (2017) 18:95 Page 5 of 10comparison the opposite was true for CRP, which wasassociated with CV risk only in GN patients. Troponin Iwas a strong risk factor for CVD in both groups, but themagnitude of risk was three times greater in non-GNcompared to GN patients.DiscussionAlthough guidelines state that GN patients should beconsidered high risk for CVD [1], this is based on con-flicting studies that did not account for renal function,and therefore the contribution of CKD and pre-existingtraditional risk factors to CVD risk in glomerular dis-eases remained unknown. In order to address this defi-ciency, we used a large prospective CKD cohort withTable 3 Characteristics of matched GN and non-GN groups. Resultsare presented as mean ± standard deviation, median (IQR) or count(frequency)Variable GN Non-GN P valueNumber 272 272Median follow up 39.0 (33.2–39.0) 39.0 (33.8–39.0)Age (years) 60.5 ± 14 60.4 ± 14 0.9Male 177 (65%) 177 (65%) 0.9Caucasian 231 (85%) 242 (89%) 0.4Cause of KidneyDiseaseDiabetic - 49 (18%)Hypertensive - 87 (32%)PCKD - 44 (16%)Other non-GN - 92 (34%)IgA 60 (22%) -FSGS 33 (12%) -ANCA Vasculitis 21 (8%) -Lupus 13 (5%) -Membranous Nephritis 9 (3%)GN unspecified 136 (50%)Diabetes 57 (21%) 60 (22%) 0.7CVD history 0.9No CVD 193 (71%) 201 (74%)Ischemic 33 (12%) 30 (11%)CHF 27 (10%) 25 (9%)Ischemic and CHF 19 (7%) 16 (6%)eGFR (ml/min/1.73 m2) 27.0 ± 9.1 27.2 ± 9.1 0.8eGFR categories 0.7< 20 76 (28%) 68 (25%)20–29 103 (38%) 109 (40%)> 30 93 (34%) 95 (35%)ACR (mg/mmol) 58.7 [13.6, 170.2] 17.2 [3.6–90.0] <0.001Systolic BP (mmHg) 132 ± 19 131 ± 19 0.8Diastolic BP (mmHg) 75 ± 12 74 ± 12 0.4Statin 158 (58%) 147 (54%) 0.3Albumin (g/L) 39.8 ± 4.8 41.1 ± 4.2 <0.001Total cholesterol (mg/dL) 4.7 ± 1.2 4.5 ± 1.4 0.4Troponin I (> LLD) 71 (26%) 68 (25%) 0.8CRP (mg/mL) 2.4 [0.9–5.6] 2.4 [1.1–5.5] 0.6ADMA 0.537 ± 0.094 0.534 ± 0.151 0.9IL-6 (μg/L) 3.55 [1.00–6.02] 3.5 [1.00–5.27] 0.6NT-ProBNP(pg/mL) Pg/mL299 [121–799] 291 [107–757] 0.7Abbreviations: PCKD polycystic kidney disease, FSGS focal segmentalglomerulosclerosis, CVD cardiovascular disease, CHF congestive heart failure,BP blood pressure, LLD lower limit of detection, ADMA asymmetricdimethylarginine, IL-6 interleukin 6, NT-Pro-BNP N-terminal pro-brainnatriuretic peptideTable 4 The results of univariable and multivariable shared frailtyCox regression models for the risk of all-cause CV events in thematched cohortHR 95%CI P-valueUnivariable ModelGN vs non-GN CKD 1.01 0.58–1.77 0.9Multivariable ModelGN vs non-GN CKD 0.99 0.56–1.75 0.9Age 1.03 1.01–1.07 0.03Male sex 1.27 0.63–2.55 0.5Diabetes 3.11 1.59–6.07 <0.001Prior CVD 3.16 1.55–6.46 <0.001Baseline eGFR 1.01 0.98–1.05 0.4RRT a 1.51 0.51–4.50 0.4a as a time-dependent variableTable 5 In the matched cohort, the association between eachbiomarker and the risk of all-cause CV events in the GN and non-GNgroups using multivariable models that included interaction termsbetween the GN vs non-GN group and each biomarkerBiomarker HR 95% CIuACR (per log unit) GN 1.02 0.81–1.30Non-GN 1.30 1.03–1.62ADMA (per 1StD) GN 1.54 0.95–2.50Non-GN 1.06 0.82–1.37ProBNP (per 1StD) GN 2.21 1.43–3.40Non-GN 2.54 1.68–3.86CRP (per 1StD) GN 1.70 1.11–2.60Non-GN 1.22 0.81–1.84IL6 (per 1StD) GN 1.57 1.12–2.22Non-GN 1.43 0.95–2.16Troponin I(>LLD vs. <LLD)GN 3.57 1.55–8.22Non-GN 12.16 4.64–31.87uACR, ProBNP, CRP and IL6 were log-transformed for analysisAbbreviations: LLD lower limit of detection, StD standard deviationHutton et al. BMC Nephrology  (2017) 18:95 Page 6 of 10standardized measures of CV risk factors and outcomeevents to investigate the risk of CVD in GN comparedto non-GN patients matched for renal function andprior CV risk factors. We show that GN patients are at ahigh 9.2% absolute risk of CVD over 3 years, but that incontrast to our a priori hypothesis this risk was notdifferent from otherwise comparable patients with ad-vanced CKD without GN. These results suggest thatonce a sustained reduction in kidney function hasdeveloped, GN is not independently associated with add-itional CV risk beyond that explained by reduced eGFRand pre-existing traditional risk factors.The assumption that primary GN is linked to anincreased risk of vascular events derives from older caseseries with conflicting results, ranging from no increasedrisk to an 85 fold relative risk of CVD compared to thegeneral population [20, 23–27]. More recent cohortstudies described a more attenuated 2–8 fold relativerisk in CVD, but none accounted for traditional CV riskfactors or the degree of renal dysfunction [2, 4, 11, 12,28]. The association of eGFR with CVD has been wellestablished in both the general CKD and GN populations[2, 4, 5, 9, 10]. This is the first study to systematically com-pare GN to non-GN patients while considering both base-line eGFR and prior traditional CV risk factors. Inunadjusted analysis, the GN patients had a significantlylower risk of CV events compared to the non GN patients,likely due to their younger age and comparative lack oftraditional CV risk factors. However, after accounting forage, sex, diabetes, prior CVD and eGFR using two differ-ent methods (multivariable adjusted models and a two-stage propensity score matching algorithm), we showedthat CKD patients with GN had a similar risk of CVDcompared to non-GN patients. Because our results wereunchanged when we adjusted for RRT, our findings arenot likely confounded by differential rates of progressionto ESRD. In our primary analyses we did not adjust forblood pressure and dyslipidemia because these may bothmediate the risk of CVD that results from GN. However,in sensitivity analyses that included both of these variablesour results were unchanged. Nearly 50% of patients withnon-GN CKD had diabetic nephropathy, implying severediabetes that may disproportionately contribute to CVDrisk in a way not fully accounted for by the propensityscore. To address this, we performed sensitivity analysesin which we repeated the matching algorithm in sub-groups based on diabetic status. As expected, we observedsubstantially higher CVD risk in diabetics compared tonon-diabetics (21.4% vs. 6.2–8.2% respectively), but inboth subgroups CV event rates were similar in GN andnon-GN patients. Using a systematic and comprehensiveanalysis strategy, we have shown that CKD patients withGN are at a high absolute 3-year risk of CVD, but that thisis explained by the severity of renal dysfunction and theaccumulation of traditional CV risk factors rather than be-ing attributed to the GN disease itself.Because CVD is such an important cause of mortalityin patients with CKD [8, 9, 14, 21, 29], our findings havesubstantial implications to the clinical care and manage-ment of CVD in GN patients. Cardiac risk stratificationis especially important in younger patients, being neces-sary to implement appropriate prevention strategies, andin the assessment for kidney transplantation. Ourunmatched GN group had a high 8.2% absolute 3-yearrisk of CVD despite a mean age of only 58 years, and72–80% having no prior history of DM or CVD. Preven-tion strategies such as statin therapy are not likely to beemployed as routine treatment for these patients, owingto their lack of comorbidities and relatively young age.However, our study shows that GN patients with CKDare at high absolute CV risk, greater than 10% over10 years, and therefore should be considered for statintherapy according to the KDIGO lipid guidelines [30].Although biomarkers of inflammation and vascularhealth have been associated with CVD in the generalCKD population, our study offers novel insights into thepossibility that there are differences in the predictivevalue of biomarkers in those with GN compared to non-GN CKD. Our interaction analyses suggested a non-significant trend towards CRP being a stronger CVD riskfactor in CKD patients with GN, and uACR and tropo-nin being more important CVD risk factors in non-GNCKD. Reasons for this are not clear, but may include in-flammation and CRP being more pronounced in GNthereby playing a more prominent role in the develop-ment of CVD; albuminuria reflecting local renalpathology in GN patients instead of being a marker ofendothelial dysfunction and CV risk as it is in the gen-eral and all-cause CKD populations; and [5, 6, 10, 31].more common non-CVD causes of increased tropo-nin in patients with GN [32–34]. These resultsrequire confirmation in larger studies with sufficientpower to investigate and explain differential associa-tions between biomarkers and CVD risk within sub-groups of CKD etiology.Our GN group comprised of patients with advancedkidney disease with a mean eGFR of 27 ml/min/1.73 m2and who were not on immunotherapy, which must beconsidered in the generalizability of our results. Com-pared to other cohorts of patients with active GN suchas the Toronto GN registry [35], our patients are older,with lower eGFR and lower levels of proteinuria, consist-ent with study enrolment at a later stage in the diseaseprocess. Low levels of CRP (2.4 μg/L) and IL-6 (3.55 μg/L) were seen in the GN patients and were comparable tolevels in the matched non-GN group. Our study clearlyshows that when glomerular disease becomes quiescent,GN patients are at comparable CV risk to other patientsHutton et al. BMC Nephrology  (2017) 18:95 Page 7 of 10with CKD and similar CV risk factors. Future research isrequired to determine if our results apply at an earlystage of glomerular disease, in which renal function ispreserved but more severe inflammation or proteinuriahave been hypothesized to disproportionately contributeto CVD in the absence of other CV risk factors.Although no study has specifically addressed this issue,the results of Mahmoodi et al. suggest this may not bethe case. In this study of 298 nephrotic patients withmean eGFR of 59 ml/min/1.73 m2, the absolute risk ofCVD was high at 1.48% per year, however this was dis-proportionately dominated by those with diabetic ne-phropathy. In the subgroup of GN patients withoutdiabetes or prior CVD, the risk was only 0.82% per year,suggesting that even amongst nephrotic patients withpreserved renal function, the development of CVD issubstantially related to pre-existing CV risk factors [2].The considerably higher incidence of CV events seen inour GN cohort compared to that of Mahmoodi, at 3.12%per year, is probably a result of older age, lower eGFRand greater accumulation of traditional CV risk factors.If future research confirms our finding that CVD inpatients with early GN is predominantly determined byprior traditional risk factors and level of renal function,then GN patients without other comorbidities andnormal renal function may not have an increasedabsolute risk of CVD. This finding would substantiallyimpact therapeutic decisions regarding primaryprevention strategies with statins at earlier stages ofglomerular disease.Our study has several other limitations that should beconsidered in the interpretation of the results. The spe-cific type of glomerular disease was unknown in ap-proximately 50% of the GN patients. This limits ourability to draw conclusions about specific GN types andCV risk. Certain types of GN such as minimal changedisease may not be associated with CVD due to infre-quent nephrotic flares with long periods of intermittentdisease quiescence [2, 27]. However, it is unlikelythat such patients would have been included in ourGN cohort since progression to advanced CKD isunlikely in the absence of persistent ongoing pro-teinuria and disease activity. There was no availableinformation about cigarette smoking as a CV riskfactor, and so this could not be included in our mul-tivariable models. Centrally adjudicated peripheralvascular disease outcome events were not availablefor analysis; however, our composite CV outcomeevent definition is nonetheless consistent with thatused in major CV clinical trials [36–38]. Finally, it ispossible that the GN and non-GN groups will differin CV events occurring late after study enrolment,and this difference would not be detected by the3 year follow up of our study.ConclusionsWe have shown that GN patients with CKD and de-creased eGFR who are not on active immunosuppressivetherapy have a high 9.2% 3 year risk of CV events. How-ever, this risk does not appear to be different from other-wise similar CKD patients without GN, suggesting thatthe elevated risk of CVD in GN patients may be attribut-able to prior CV risk factors and level of kidney functionrather than the GN disease itself.Additional fileAdditional file 1: Table S1. The biomarker hazard ratios using Coxproportional hazards models that included GN vs non-GN CKD and eachbiomarker individually. ACR, CRP, IL-6, ProBNP and FGF-23 were log-trans-formed for analysis. (DOCX 15 kb)AbbreviationsADMA: Asymmetric dimethylarginine; BP: Blood pressure; CKD: Chronic kidneydisease; CRP: C-reactive protein; CVD: Cardiovascular disease; FSGS: Focalsegmental glomerulosclerosis; GN: Glomerulonephritis; IL-6: Interleukin-6;MI: Myocardial infarction; Pro-BNP: N-terminal pro-brain natriuretic peptide;RRT: Renal replacement therapy; uACR: Urine albumin to creatinine ratioAcknowledgementsWe thank all the CanPreddict investigators and coordinators for data andsample collection: Mohsen Agharazii, CHUQ: L’Hôtel-Dieu de Québec; AyubAkbarii, MD, Judy Cheesman, Jennilea Courtney, Sabrina Hamer, Edita Delic,Valerie Cronin, University of Ottawa; Paul Barré, MD, Jeffrey Golden, RoyalVictoria Hospital; Brendan Barrett, MD, Elizabeth Langille, Sandra Adams,Janet Morgan, Eastern Regional Health Authority, Health Sciences Centre;Catherine Clase, MD, Cathy Moreau, St Joseph’s Hospital; Susan Cooper, MD,Brian Forzley, MD, Susan Caron, Shauna Granger, Susan Valley, Helen Sather,Penticton Regional Hospital; Serge Cournoyer, MD, Lorraine Menard, MichèleRoy, Hélène Skidmore, Dolores Beaudry, Charles Le Moyne Hospital; JanisDionne, MD, Josephine Chow, Valla Sahraei, BC Children’s Hospital; SandraDonnelly, MD, Niki Dacouris, Rosa Marticorena, St. Michael’s Hospital; BrendaHemmelgarn, MD, Sharon Gulewich, Troy Hamilton, Foothills Hospital; PaulKeown, MD, Nadia Zalunardo, MD, Daniel Rogers, Reena Tut, MatthewPaquette, Rossitta Yung, Vancouver General Hospital; Adeera Levin, MD,Nancy Ferguson, Helen Chiu, Kathleen Carlson, Lina Sioson, Taylor Perry,Zainab Sheriff, Naama Rozen, St. Paul’s Hospital; Charmaine Lok, MD, MichelleCross, Cathy Forrester, Alexandra Cotoi, University Health Network; FrançoisMadore, MD, Manon Maltais, Hôpital du Sacré-Cœur; Louise Moist, MD, KerriGallo, Sarah Langford, Leah Slamen, Danielle Cram, London Health ScienceCentre- Victoria Campus; Norman Muirhead, MD, Mary Jeanne Edgar, TaylorGray, Cameron Edgar, Karen Groeneweg, Eileen McKinnon, Erin McRae, KylaBlackie, London Health Science Centre- University Campus; Bharat Nathoo,MD, Kimmy Lau, York Central; Malvinder Parmar, MD, Sylvie Gelinas, Timmins& District Hospital; Martine Leblanc, MD, Lucie Lépine, Maisonneuve-Rosemont Hôpital; Claudio Rigatto, MD, Dolores Friesen, Marla Penner, St.Boniface Hospital; Steven Soroka, MD, Susan Fleet, Jeanette Squires, QEIIHealth Sciences Centre; Siva Thanamayooran, MD, Michael Binder, ChristineHines, Brenda McNeil, Patrice McDougall, Joy Howard, Deborah Gillis,Kathleen Hines, Cape Breton District Health Authority; Sheldon Tobe, MD,Mary Chessman, Nancy Perkins, Martha Agelopoulos, Stacey Knox, TiffanyRichards, Sunnybrook Hospital; Marcello Tonelli, Susan Szigety, DawnOpgenorth, University of Alberta; Karen Yeates, MD, Karen Mahoney,Kingston General Hospital.FundingCanPREDDICT was funded by an educational grant from Janssen-Ortho Inc.No additional funding was received for this retrospective analysis.Availability of data and materialsAt the current time, CanPREDDICT data is not publicly available.Hutton et al. BMC Nephrology  (2017) 18:95 Page 8 of 10Authors’ contributionsAll authors contributed significantly to the study conception and design,and the intellectual content of the manuscript. OD conducted all statisticalanalyses and generated the figures used in the manuscript. SB and HHdeveloped and wrote the manuscript, and AL, JG, OD and MT providedconstructive criticism and edits of draft manuscripts. HH categorised subjectsinto GN and non-GN subgroups. All authors read and approved the finalmanuscript.Competing interestsThe authors declare that they have no competing interests associated withthis work. All of the authors agree with the content of the manuscript andstate that results contained herein have not been published previously, inwhole or part. Dr. Barbour is funded by the Carraresi Foundation. Drs.Barbour and Gill are funded by the Michael Smith Foundation for HealthResearch.Consent for publicationNot applicable.Ethics approval and consent to participateWritten consent was obtained from all study participants for CanPREDDICT; furtherconsent was not necessary for this retrospective analysis. All research was carriedout in accordance with the Declaration of Helsinki and the study was approvedby all Providence Health Care Research Ethics Board (ID: H07-02457).Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Centre for Inflammatory Diseases, Monash University, Clayton, VIC, Australia.2Dept of Nephrology, Monash Health, 246 Clayton Rd, Clayton, VIC 3168,Australia. 3Division of Nephrology, University of British Columbia, Vancouver,BC, Canada. 4BC Provincial Renal Agency, Vancouver, BC, Canada. 5Centre forHealth Evaluation and Outcome Sciences, St Paul’s Hospital, Vancouver, BC,Canada.Received: 22 September 2015 Accepted: 14 March 2017References1. Kidney Disease: Improving Global Outcomes (KDIGO) GlomerulonephritisWork Group. KDIGO clinical practice guideline for glomerulonephritis.Kidney Int. 2012;S2:159.2. Mahmoodi BK, ten Kate MK, Waanders F, Veeger NJ, Brouwer JL, Vogt L,Navis G, van der Meer J. High absolute risks and predictors of venous andarterial thromboembolic events in patients with nephrotic syndrome: resultsfrom a large retrospective cohort study. Circulation. 2008;117(2):224–30.3. Ihm CG. Hypertension in chronic glomerulonephritis. Electrolytes BloodPress. 2015;13(2):41–5.4. Myllymaki J, Syrjanen J, Helin H, Pasternack A, Kattainen A, Mustonen J.Vascular diseases and their risk factors in IgA nephropathy. Nephrol DialTransplant. 2006;21(7):1876–82.5. Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de JongPE, Coresh J, Gansevoort RT. Association of estimated glomerular filtrationrate and albuminuria with all-cause and cardiovascular mortality in generalpopulation cohorts: a collaborative meta-analysis. Lancet.2010;375(9731):2073–81.6. Astor BC, Matsushita K, Gansevoort RT, van der Velde M, Woodward M,Levey AS, Jong PE, Coresh J, Astor BC, Matsushita K, et al. Lower estimatedglomerular filtration rate and higher albuminuria are associated withmortality and end-stage renal disease. A collaborative meta-analysis ofkidney disease population cohorts. Kidney Int. 2011;79(12):1331–40.7. Matsuzawa Y, Guddeti RR, Kwon TG, Lerman LO, Lerman A. Treatingcoronary disease and the impact of endothelial dysfunction. ProgCardiovasc Dis. 2014;57:431.8. Bello AK, Hemmelgarn B, Lloyd A, James MT, Manns BJ, Klarenbach S,Tonelli M. Associations among estimated glomerular filtration rate,proteinuria, and adverse cardiovascular outcomes. Clin J Am Soc Nephrol.2011;6(6):1418–26.9. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney diseaseand the risks of death, cardiovascular events, and hospitalization. N Engl JMed. 2004;351(13):1296–305.10. van der Velde M, Matsushita K, Coresh J, Astor BC, Woodward M, Levey A,de Jong P, Gansevoort RT, van der Velde M, Matsushita K, et al. Lowerestimated glomerular filtration rate and higher albuminuria are associatedwith all-cause and cardiovascular mortality. A collaborative meta-analysis ofhigh-risk population cohorts. Kidney Int. 2011;79(12):1341–52.11. Heaf J, Lokkegaard H, Larsen S. The epidemiology and prognosis ofglomerulonephritis in Denmark 1985–1997. Nephrol Dial Transplant.1999;14(8):1889–97.12. Orodonez JHR, Kilebrew E, Fireman B. The increased risk of coronary heartdisease associated with nephrotic syndrome. Kidney Int. 1993;4:638–42.13. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. NEngl J Med. 2005;352(16):1685–95.14. Liu M, Li XC, Lu L, Cao Y, Sun RR, Chen S, Zhang PY. Cardiovascular diseaseand its relationship with chronic kidney disease. Eur Rev Med Pharmacol Sci.2014;18(19):2918–26.15. Levin A, Rigatto C, Brendan B, Madore F, Muirhead N, Holmes D, Clase CM,Tang M, Djurdjev O, Can P. Cohort profile: Canadian study of prediction ofdeath, dialysis and interim cardiovascular events (CanPREDDICT). BMCNephrol. 2013;14:121.16. Vickery S, Webb MC, Price CP, John RI, Abbas NA, Lamb EJ. Prognostic valueof cardiac biomarkers for death in a non-dialysis chronic kidney diseasepopulation. Nephrol Dial Transplant. 2008;23(11):3546–53.17. Hickman PE, McGill DA, Talaulikar G, Hiremagalur B, Bromley J, Rahman A,Koerbin G, Southcott E, Potter JM. Prognostic efficacy of cardiac biomarkersfor mortality in dialysis patients. Intern Med J. 2009;39(12):812–8.18. Young JM, Terrin N, Wang X, Greene T, Beck GJ, Kusek JW, Collins AJ, SarnakMJ, Menon V. Asymmetric dimethylarginine and mortality in stages 3 to 4chronic kidney disease. Clin J Am Soc Nephrol. 2009;4(6):1115–20.19. Barreto DV, Barreto FC, Liabeuf S, Temmar M, Lemke HD, Tribouilloy C,Choukroun G, Vanholder R, Massy ZA. European Uremic Toxin Work G:Plasma interleukin-6 is independently associated with mortality in bothhemodialysis and pre-dialysis patients with chronic kidney disease. KidneyInt. 2010;77(6):550–6.20. Vosnides G, Cameron JS. Hyperlipidemia in renal disease. Med J Aust. 1974;ii:855.21. Alani H, Tamimi A, Tamimi N. Cardiovascular co-morbidity in chronic kidneydisease: Current knowledge and future research needs. World J Nephrol.2014;3(4):156–68.22. Wanner C. Importance of hyperlipidaemia and therapy in renal patients.Nephrol Dial Transplant. 2000;15 Suppl 5:92–6.23. Alexander JH, Schapel GJ, Edwards KD. Increased incidence of coronary heartdisease associated with combined elevation of serum triglyceride and cholesterolconcentrations in the nephrotic syndrome in man. Med J Aust. 1974;ii:119–22.24. Gilboa N. Incidence of coronary heart disease associated with nephroticsyndrome. Med J Aust. 1976;i:207–8.25. Berlyne GM, Mallick NP. Ischaemic heart-disease as a complication ofnephrotic syndrome. Lancet. 1969;2(7617):399–400.26. Wass VJ, Jarrett RJ, Chilvers C, Cameron JS. Does the nephrotic syndromeincrease the risk of cardiovascular disease? Lancet. 1979;2(8144):664–7.27. Hopper Jr J, Ryan P, Lee JC, Rosenau W. Lipoid nephrosis in 31 adultpatients: renal biopsy study by light, electron, and fluorescence microscopywith experience in treatment. Medicine. 1970;49(4):321–41.28. Faurschou M, Mellemkjaer L, Starklint H, Kamper AL, Tarp U, Voss A,Jacobsen S. High risk of ischemic heart disease in patients with lupusnephritis. J Rheumatol. 2011;38(11):2400–5.29. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group.KDIGO clinical practice guideline for the evaluation and management ofchronic kidney disease. Kidney Int. 2013;S3:1–150.30. Kidney Disease: Improving Global Outcomes (KDIGO) Lipid Work Group.KDIGO clinical practice guideline for lipid management in chronic kidneydisease. Kidney Int. 2013;3:259–305.31. Matsuzawa Y, Lerman A. Endothelial dysfunction and coronary artery disease:assessment, prognosis, and treatment. Coron Artery Dis. 2014;25(8):713–24.32. Hamwi SM, Sharma AK, Weissman NJ, Goldstein SA, Apple S, Canos DA,Pinnow EE, Lindsay J. Troponin-I elevation in patients with increased leftventricular mass. Am J Cardiol. 2003;92(1):88–90.33. Paradis P, Dali-Youcef N, Paradis FW, Thibault G, Nemer M. Overexpressionof angiotensin II type I receptor in cardiomyocytes induces cardiachypertrophy and remodeling. Proc Natl Acad Sci U S A. 2000;97(2):931–6.Hutton et al. BMC Nephrology  (2017) 18:95 Page 9 of 1034. Urushihara M, Kinoshita Y, Kondo S, Kagami S. Involvement of the intrarenalrenin-angiotensin system in experimental models of glomerulonephritis.J Biomed Biotechnol. 2012;2012:601786.35. Barbour SJ, Greenwald A, Djurdjev O, Levin A, Hladunewich MA, NachmanPH, Hogan SL, Cattran DC, Reich HN. Disease-specific risk of venousthromboembolic events is increased in idiopathic glomerulonephritis.Kidney Int. 2012;81(2):190–5.36. Group SR, Wright Jr JT, Williamson JD, Whelton PK, Snyder JK, Sink KM, Rocco MV,Reboussin DM, Rahman M, Oparil S, et al. A randomized trial of intensive versusstandard blood-pressure control. N Engl J Med. 2015;373(22):2103–16.37. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, Gomez-Gracia E,Ruiz-Gutierrez V, Fiol M, Lapetra J, et al. Primary prevention of cardiovasculardisease with a Mediterranean diet. N Engl J Med. 2013;368(14):1279–90.38. Investigators O, Yusuf S, Teo KK, Pogue J, Dyal L, Copland I, Schumacher H,Dagenais G, Sleight P, Anderson C. Telmisartan, ramipril, or both in patientsat high risk for vascular events. N Engl J Med. 2008;358(15):1547–59.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Hutton et al. BMC Nephrology  (2017) 18:95 Page 10 of 10

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.52383.1-0366837/manifest

Comment

Related Items