OUP user menu

★ editor's choice ★

Statins and renovascular disease in the elderly: a population-based cohort study

Daniel G. Hackam, Fangyun Wu, Ping Li, Peter C. Austin, Sheldon W. Tobe, Muhammad M. Mamdani, Amit X. Garg
DOI: http://dx.doi.org/10.1093/eurheartj/ehq452 598-610 First published online: 14 December 2010


Aims More than 90% of cases of renovascular disease (RVD) are caused by atherosclerosis; thus patients with this condition are at high risk for vascular events. We examined the association of statins with prognosis in patients with RVD.

Methods and results We performed a population-based cohort study in 4040 patients with RVD older than 65 years using province-wide health data in Ontario, Canada. The primary outcome was time to first cardiorenal event, specifically myocardial infarction, stroke, heart failure, acute renal failure, dialysis or death; the primary analysis used a time-dependent covariate for statin exposure. Despite having a greater burden of cardiovascular and renal comorbidity, the risk of the primary outcome was significantly lower in statin users than in non-users [unadjusted hazard ratio (HR) 0.51, 95% confidence interval (CI) 0.47–0.57; P < 0.0001]. This association was materially unchanged after adjusting for demographic characteristics, cardiovascular risk factors, other comorbidities, measures of health-care utilization, screening, and concomitant medications (adjusted HR 0.51, 95% CI 0.46–0.57). An analysis using the same endpoint in a propensity-matched cohort without time-dependent statin exposure revealed a lower risk of the primary outcome in statin-treated patients but with a substantially more conservative point estimate (HR 0.82, 95% CI 0.71–0.95).

Conclusion These data suggest that statins are associated with improved prognosis in elderly patients with RVD.

  • Statins
  • Renovascular disease
  • Cohort studies
  • Prognosis


Atherosclerotic renovascular disease (RVD) is a highly prevalent vascular condition, particularly among the elderly, with nearly 7% of community dwelling persons 65 years or older demonstrating RVD on duplex sonography.1 In addition, patients with RVD incur high rates of cardiovascular and renal events. In the recent Angioplasty and Stenting for Renal Artery Lesions (ASTRAL) trial, 37% of participants suffered a major cardiovascular event and 15% suffered a major renal event over a median follow-up of 34 months.2 Among elderly patients with RVD captured in US Medicare data, annual rates of stroke, acute coronary syndrome, heart failure, and death were 18, 30, 19, and 17%, respectively.3

Randomized trials in RVD have typically focused on the role of revascularization in the management of this condition; remarkably few trials have assessed the effects of medical therapy on prognosis. Because >90% of RVD is caused by atherosclerosis, most experts recommend statin therapy for affected patients, although no statin trial has been conducted in this setting. Patients with RVD are on an average sicker, older and more likely to have renal impairment than the typical participant recruited to a statin ‘mega-trial'; RVD might therefore complicate the risk-benefit ratio of statins. Alternatively, since RVD is often a marker of diffuse multisystem atherosclerosis, affected patients might have more to gain from adding a statin to their regimen.

We conducted a retrospective, population-based cohort study to evaluate the association between statins and cardiorenal outcomes in a defined sample of patients with RVD.4 Because non-adherence is common in statin users, we performed our analyses using time-dependent covariates to model statin exposure throughout follow-up. In a sensitivity analysis, we also matched statin users to controls using propensity-based matching, which accounts for the likelihood of being prescribed a statin according to measured baseline characteristics. Finally, because statins seem to exert beneficial effects across disparate vascular beds, we assessed a spectrum of major cardiac, cerebral, and renal events in the primary analysis, while also testing these outcomes separately in secondary analyses.


Setting and data sources

We conducted our study in Ontario, Canada using linked health-care databases in accordance with a fully prespecified research protocol. Ontario is Canada's most populous and ethnically diverse province with a total population of >13 million, of whom 1.8 million are older than 65. Elderly patients in Ontario have universal access to health-care services, including outpatient medical visits, hospital care, home care, and prescription drugs. The large databases that record this care have been used extensively in past research, contain little missing information and have been validated for a diverse range of cardiovascular and renal events.58

We used six health databases: the Canadian Institute for Health Information Discharge Abstract Database (CIHI-DAD), which records all hospital admissions in the province, including detailed diagnostic and procedural information; the Ontario Health Insurance Plan (OHIP) Database, which records information on outpatient medical visits and testing; the National Ambulatory Care Reporting System Database, which records emergency department visits, dialysis, oncological care, and cardiac catheterization; the Canadian Institute for Health Information Same Day Surgery (CIHI-SDS) Database, which records information on ‘same day' interventions and procedures; the Ontario Drug Benefit Database, which records all prescription medications dispensed to patients 65 years of age or older; and the Registered Persons Database, which collects vital statistics on all Ontario residents. The study was approved by the Research Ethics Board of Sunnybrook Health Sciences Centre and the Privacy Office of the Institute for Clinical Evaluative Sciences.

Sample and selection

We included consecutive patients older than 65 years with codes identifying ‘renal artery stenosis' or RVD in the CIHI-DAD, CIHI-SDS, and OHIP databases from 1 July 1994 to 1 July 2007 (a span of 13 years). Using a similar code set, Murphy et al.9 calculated a specificity of 96% and sensitivity of 80% for angiographically verified RVD in a five-state validation study. We focused on patients older than 65 years because such individuals receive universal prescription coverage, and prescriptions dispensed are recorded in the Ontario Drug Benefit database. Furthermore, these subjects are highly likely to have atherosclerotic RVD. We excluded patients with fibromuscular dysplasia (n = 21), death within 120 days of cohort entry (n = 672), invalid health card number (n = 215), missing age or sex (n = 1), non-residents of Ontario (n = 320), nursing home placement (n = 563), or end stage renal disease prior to cohort entry (n = 598). The rationale for each of these exclusions is detailed in Appendix 1.


We defined statin exposure as receipt of one or more prescriptions for a hydroxymethylglutaryl coA reductase inhibitor within 120 days following the first identifying code for RVD (hereafter labelled the ‘index date’). Since initially untreated patients might start a statin after this window, and initially treated patients might discontinue statin therapy, we modelled ongoing statin exposure as a time-dependent covariate to reduce dilution bias from these factors. On the basis of serial prescription refills, we deemed cessation to have occurred following the last drug claim (if any) in the treatment group, and deemed initiation to have occurred following the first drug claim (if any) in the control group. As described below, we replicated this analysis using the more traditional ‘intention-to-treat' approach, in which ongoing statin treatment was not reclassified following the initial exposure assessment.

Characteristics and comorbidities

For each patient in the cohort, we assessed demographic characteristics, cardiovascular risk factors, major comorbidities, measures of health-care utilization, screening, and concomitant medications by searching the health-care databases for the 3-year interval preceding the index date. Variables were selected from a literature review of prognostic factors in RVD supplemented by additional characteristics likely to affect patient outcome.3,1014 At baseline, we assessed inpatient and outpatient claims for diagnostic modalities that are typically used to test for RVD such as renal angiography and renal Doppler ultrasonography as well as diagnostic tests for related cardiovascular conditions since such manoeuvres may lead to statin prescribing. We adjusted for treatment with 15 distinct classes of medications at baseline, specifically, statins, calcium channel blockers, thiazide diuretics, alpha-blockers, beta-blockers, vasodilators, non-statin lipid drugs, anticoagulants, loop diuretics, antiplatelet agents, anti-arrhythmic agents, potassium-sparing diuretics, non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers. All variables were entered as covariates in all statistical models.


The primary outcome was time to a major cardiorenal event comprising myocardial infarction, stroke, heart failure, acute renal failure, dialysis, and death. Myocardial infarction, stroke, heart failure, and acute renal failure required admission to hospital with a most responsible diagnosis of the condition in question; death and dialysis were defined using both outpatient and inpatient databases (Appendix 2 for specific coding). We examined the six components of the primary outcome separately in secondary analyses. Follow-up for each patient began on the index date and continued until the event in question, death (for the secondary analyses) or 31 March 2008 (whichever came first). The primary analysis used time-dependent statin exposure assessment.

Sensitivity analyses

We conducted several additional analyses to test the robustness of our findings. First, we repeated our primary analysis using an intention-to-treat framework, which did not categorize statin exposure as a time-dependent covariate. The purpose of this analysis was to reduce healthy adherer bias whereby patients who are more adherent to statins may be innately healthier than patients who are less adherent to statins. Second, we replicated our primary analysis in specific settings which might influence the effect of statins, specifically dividing the cohort into subgroups by age, sex, history of diabetes mellitus, chronic kidney disease, coronary artery disease, and previous renal artery revascularization. Third, to capture the full extent of incident cardiorenal events in RVD, we prespecified three additional outcomes: hospitalizations with a most responsible diagnosis of malignant hypertension, hospitalizations with a most responsible diagnosis of renal disease (i.e. ‘any renal hospitalization'), and revascularization procedures (comprising cerebrovascular, coronary, and peripheral arterial interventions). The intent of this analysis was to test the association of statins with morbid events which might not be captured in the primary outcome and its components. Fourth, we repeated our primary analysis using a traditional propensity score matching algorithm, pairing statin-exposed patients and untreated controls by propensity score (±0.2 standard deviations), age (±1 year), and sex. Of note this analysis does not use a time-dependent statin exposure assessment covariate but rather characterizes exposure according to treatment at baseline. It is therefore inherently more conservative since it does not account for treatment uptake among controls or treatment cessation and erratic adherence among statin users. Finally, we replicated our primary analysis in patients who did not change treatment over time, specifically excluding treated patients who discontinued statin therapy even temporarily together with controls who initiated statin therapy during follow-up.

Statistical analysis

Sample size calculations suggested that 2627 patients were required to detect as statistically significant a relative risk reduction of 20% or more, based on equal numbers of treated patients and controls, and an overall primary event risk of 30% in the control group (α = 0.05, β = 0.10).3 We assumed a median follow-up of 2.5 years and attrition of 10%; this is conservative as emigration is <0.5% per year among older individuals in our region. Given that in actual fact controls were somewhat more numerous than treated patients, 2709 patients were actually required to exclude the same risk reduction. We used Cox proportional hazards regression to test the association of statins with outcomes and to compute hazard ratios (HR) with 95% confidence intervals (CI). Multivariable analyses were adjusted for demographic characteristics, cardiovascular risk factors, comorbidities, measures of health-care utilization, screening, and medications. A two-tailed P-value of <0.05 was considered statistically significant. We performed all statistical analyses using SAS version 9.1 (SAS Institute, Carey, NC, USA).


Over the 13-year accrual interval, we studied 4040 patients with RVD (Table 1). Comorbidities were highly prevalent, including hypertension (89%), coronary artery disease (54%), peripheral artery disease (46%), heart failure (46%), and cerebrovascular disease (30%). Fewer than half of all patients received a statin at baseline (n = 1682; 42%). Not surprisingly, cardiovascular risk factors and comorbidities were more common in the statin group than among controls including diabetes (33 vs. 25%), hypertension (91 vs. 87%), chronic kidney disease (63 vs. 57%), coronary artery disease (62 vs. 48%), and cerebrovascular disease (32 vs. 29%). Diagnostic methods in this cohort included renal artery ultrasound (41%), catheter renal angiography (25%), computed tomographic angiography (23%), captopril nephrography (14%), and magnetic resonance angiography (7%).

View this table:
Table 1

Selected baseline characteristics of the study cohort by initial statin exposure (n = 4040)

CharacteristicStatin group (n = 1682)Control group (n = 2358)P-value
Demographic factors (%)
 Age, y74.1 ± 5.374.9 ± 5.9<0.001
 Sex (female)769 (46)1080 (46)0.96
 Rural residence238 (14)411 (17)0.005
 Socioeconomic status (high)943 (56)1287 (55)0.53
Cardiovascular risk factors (%)
 Diabetes mellitus557 (33)583 (25)<0.001
 Hypertension1526 (91)2055 (87)<0.001
 Cerebrovascular disease540 (32)672 (29)0.014
 Peripheral artery disease750 (45)1109 (47)0.13
 Coronary artery disease1037 (62)1129 (48)<0.001
 Heart failure787 (47)1078 (46)0.50
 Aortic aneurysm606 (36)847 (36)0.94
 Chronic kidney disease1065 (63)1347 (57)<0.001
 Atrial fibrillation/flutter245 (15)387 (16)0.11
 Other dysrhythmia522 (31)535 (23)<0.001
Other comorbidities (%)
 Chronic liver disease43 (3)90 (4)0.03
 Chronic lung disease670 (40)998 (42)0.11
 Dementia83 (5)181 (8)<0.001
 Peptic ulcer disease259 (15)435 (18)0.011
 Systemic malignancy62 (4)145 (6)<0.001
 Prior venous thrombosis188 (11)273 (12)0.69
Health-care utilization and testing, n in past year
 Family medicine visits13 ± 913 ± 100.51
 Cardiology visits2 ± 32 ± 2<0.001
 Endocrinology visits0.3 ± 10.2 ± 10.07
 Nephrology visits1 ± 21 ± 2<0.001
 Internist visits1 ± 31 ± 20.30
 Vascular surgeon visits2 ± 22 ± 20.02
 Cholesterol tests1 ± 21 ± 1<0.001
 Creatinine tests4 ± 43 ± 4<0.001
 Creatinine kinase tests0.5 ± 10.1 ± 1<0.001
 Transaminase tests2 ± 21 ± 1<0.001
 Hospitalizations (past 3 years)2 ± 22 ± 2<0.001
 Number of drugs (past 6 months)12 ± 610 ± 6<0.001
 Echocardiography (%)667 (40)788 (33)<0.001
 Carotid ultrasound (%)435 (26)452 (19)<0.001
 Coronary angiography (%)294 (18)162 (7)<0.001
 Holter monitoring (%)167 (10)211 (9)0.29
 Stress testing (%)576 (34)560 (24)<0.001
 Bone density testing (%)104 (6)85 (4)<0.001
 Renal artery ultrasound (%)698 (42)969 (41)0.80
 Renal MRA (%)190 (11)79 (3)<0.001
 Renal CTA (%)411 (24)533 (23)0.18
 Plain renal angiography (%)514 (31)500 (21)<0.001
 Captopril nephrography (%)234 (14)337 (14)0.73
  • Data are mean ± SD or n (%). High socioeconomic status was defined by quintiles 3, 4, and 5. Numeric health-care data were rounded to the nearest integer.

  • CTA, computed tomographic angiography; MRA, magnetic resonance angiography.

The sample provided a total of 12 489 patient-years of follow-up with a median of 3.3 years (interquartile range 1.4–5.0 years). In the primary analysis, statins were associated with a substantially lower risk of cardiorenal events (unadjusted HR 0.51, 95% CI 0.47–0.57; P < 0.0001). The primary outcome occurred at a rate of 63 and 103 events per 100 patient years at risk in statin users and non-users, respectively. After adjusting for demographic characteristics, cardiovascular risk factors, comorbidities, measures of health-care utilization, screening, and concomitant medications, this protective association was materially unchanged (adjusted HR 0.51, 95% CI 0.46–0.57; Table 2). In fully adjusted models, statins were associated with reduced rates of stroke (HR 0.72, 95% CI 0.54–0.96), heart failure (HR 0.83, 95% CI 0.69–0.99), dialysis (HR 0.66, 95% CI 0.52–0.86), and death (HR 0.27, 95% CI 0.24–0.31), with a trend towards fewer myocardial infarctions (HR 0.82, 95% CI 0.65–1.04, P = 0.099). Statins were not associated with any protective association for acute renal failure (HR 0.89, 95% CI 0.63–1.27); however, the latter was the least frequent component of the primary endpoint and this analysis may have been underpowered. As with the primary analysis, results were largely consistent in unadjusted and adjusted models (Table 3). Propensity score analysis in 1061 matched statin user-control pairs (total n = 2122) still yielded significant (albeit more conservative) results (HR 0.82, 95% CI 0.71–0.95).

View this table:
Table 2

Multivariable predictors for the primary outcome

CharacteristicHazard ratio (95% CI)P-value
Statins0.51 (0.46–0.57)<0.0001
Age (per year)1.03 (1.02–1.04)<0.0001
Diabetes mellitus1.39 (1.25–1.54)<0.0001
Cerebrovascular disease1.17 (1.06–1.30)0.003
Peripheral arterial disease1.12 (1.01–1.24)0.03
Coronary artery disease1.16 (1.04–1.30)0.007
Heart failure1.37 (1.21–1.55)<0.0001
Aortic aneurysm (including surgery for)0.85 (0.76–0.96)0.009
Chronic kidney disease1.22 (1.10–1.36)0.0002
Chronic liver disease1.30 (1.03–1.64)0.03
Chronic lung disease1.12 (1.01–1.24)0.03
Peptic ulcer disease0.83 (0.73–0.94)0.003
Systemic malignancy1.51 (1.26–1.81)<0.0001
CADG3 (‘likely to recur’)0.80 (0.68–0.95)0.01
CADG7 (‘chronic specialty stable’)0.82 (0.72–0.94)0.004
Hospitalizations (n, past 3 years)1.07 (1.05–1.10)<0.0001
Drugs (n, past 6 months)1.02 (1.01–1.03)0.003
Calcium channel blockers1.13 (1.02–1.25)0.02
Loop diuretics1.23 (1.09–1.40)0.0008
Non-steroidal anti-inflammatory drugs0.84 (0.73–0.98)0.03
  • CADG, chronic ambulatory disease grouping.

View this table:
Table 3

Primary and secondary outcomes

OutcomeEvent rate per 1000 patient-years at riskUnadjusted analysis hazard ratio (95% CI)Adjusted analysisa hazard ratio (95% CI)
Primaryb62.6102.70.51 (0.47–0.57)0.51 (0.46–0.57)
Myocardial infarction10.512.60.82 (0.67–1.01)0.82 (0.65–1.04)
Stroke4.76.40.68 (0.52–0.89)0.72 (0.54–0.96)
Heart failure24.729.30.82 (0.69–0.97)0.83 (0.69–0.99)
Acute renal failure5.56.10.84 (0.61–1.17)0.89 (0.63–1.27)
Dialysis6.79.90.81 (0.65–1.02)0.66 (0.52–0.86)
Death31.689.00.28 (0.24–0.31)0.27 (0.24–0.31)
Renal hospitalization14.520.30.73 (0.59–0.90)0.69 (0.55–0.87)
Malignant hypertension2.010.70.37 (0.14–1.03)0.18 (0.05–0.60)
Revascularizationc68.964.91.26 (1.09–1.46)1.08 (0.91–1.27)
  • aAdjusted for demographic characteristics, cardiovascular risk factors, comorbidities, measures of health-care utilization, screening, and medications.

  • bDefined as first occurrence of myocardial infarction, stroke, heart failure, acute renal failure, dialysis, or death.

  • cCoronary, cerebrovascular, and peripheral revascularization procedures.

We conducted an intention-to-treat analysis to test the robustness of these findings, as well as to further evaluate absolute risk differences. In the unadjusted analysis, statins continued to exert a positive effect on the primary outcome although the strength of the association was approximately halved (HR 0.75, 95% CI 0.69–0.83). These results were consistent with the adjusted analysis (HR 0.77, 95% CI 0.69–0.85). Replicating our results in patients who did not cross over between treated and exposed groups over time yielded similar findings (unadjusted HR 0.85, 95% CI 0.75–0.97, P = 0.014; adjusted HR 0.85, 95% CI 0.73–0.99, P = 0.031). Specifically, of the initially statin-treated patients, 1077 patients temporarily or permanently discontinued therapy; of the initially untreated controls, 711 patients later initiated statins during follow-up.

We also found a consistently positive association between statins and prognosis across each of the predefined subgroups, in both unadjusted and adjusted analyses (Figure 1). A borderline statistical interaction was present for patients with a history of renal artery revascularization (P = 0.053), with a stronger effect in those who received revascularization (HR 0.40, 95% CI 0.34–0.48) than in those who did not (HR 0.60, 95% CI 0.53–0.69). Of the cohort, 1927 patients had a history of renal revascularization and incurred a total of 767 primary outcome events, whereas 2113 had no previous history of renal revascularization (in whom 1202 events transpired). Finally, statin treatment was associated with a lower risk of all renal hospitalizations (HR 0.69, 95% CI 0.55–0.87) and malignant hypertension hospitalizations (HR 0.18, 95% CI 0.05–0.60) but not vascular interventions (HR 1.08; 95% CI 0.91–1.27; Table 3). Time-to-event curves showed gradual and continual separation for the primary outcome (Figure 2) and mortality (Figure 3) throughout the entirety of follow-up.

Figure 1

Subgroup analyses for the primary outcome of major cardiorenal events. All analyses were adjusted for demographic characteristics, cardiovascular risk factors, comorbidities, measures of health-care utilization, screening, and medications. For each subgroup, the square represents the hazard ratio with horizontal lines representing the 95% confidence interval.

Figure 2

Time to primary outcome stratified by treatment with statins. Log-rank P < 0.0001 for the comparison of curves.

Figure 3

Overall survival stratified by treatment with statins. Log-rank P < 0.0001 for the comparison of curves.


We found that statins were associated with a significantly lower rate of cardiorenal events in older patients with RVD. This finding was consistent with reductions in secondary endpoints, was present whether renal artery revascularization was performed or not, and was observed in the intention-to-treat analysis and propensity-based algorithm. The size of the apparent benefit is in keeping with the Heart Protection Study, the largest randomized trial of statins performed to date (n = 20 536), which reported a number needed to treat of 19 for major vascular events in patients with occlusive vascular disease or diabetes. Reductions in renal events in the present study are also in keeping with several large randomized trials.1519

Relation to the previous literature

Several animal and human studies have reported on the effects of statins in RVD (Table 4). In randomized animal models of RVD, statins reduced renal fibrosis, improved kidney function and blood flow, prevented left ventricular hypertrophy, and increased myocardial perfusion.2024 Four previous cohort studies demonstrated findings similar to ours, although the largest sample size in these reports was only a fifth of the present study.11,2527 The most likely explanation is decreased progression, and possible regression, of renal artery stenosis with improved blood pressure under treatment with statins (as has been observed in one renal angiography cohort and several case reports28,29). In addition, anti-inflammatory, antithrombotic, and antioxidant effects of statins may account for these findings. Because our analytical models do not contain on-treatment cholesterol levels, we are unable to determine the precise mechanism of this relationship.

View this table:
Table 4

Literature on statins and renovascular disease

StudyDesignFindings related to statin exposure
Animal studies
 Chade et al.20,21Porcine unilateral RVD modelRenoprotective effects: reduced renal fibrosis and remodelling; increased renal blood flow and glomerular filtration rate
 Zhu et al.22Porcine unilateral RVD modelCardioprotective effects: left ventricular hypertrophy prevented; myocardial perfusion increased; reduced microvascular remodelling
 Laina et al.23Rodent bilateral RVD modelRenoprotective effects: improved glomerular filtration rate, free water clearance, and fractional sodium excretion
 Lavi et al.24Porcine unilateral RVD modelRenoprotective effects: improved endothelial function, reduced renal oxidative stress, inflammation and fibrosis; attenuated endothelial progenitor cell apoptosis
Human studies
 Khong et al.29Case report (n = 1)Marked regression of renal artery stenosis over 3 years following institution of statin therapy (and despite continued heavy smoking)
 Basta et al.28Case report (n = 1)Marked regression of renal artery stenosis, including spontaneous remission of hypertension and renin levels
 Bates et al.11Cohort study (n = 748)Reduced mortality over 11 year follow-up with statins (HR 0.71; 95% CI 0.53–0.95)
 Davies et al.25Cohort study (n = 447)Reduced restenosis after renal artery stenting as well as greater freedom from recurrent symptoms (hypertension and worsening renal failure) with statins; risk ratios not provided
 Silva et al.26Cohort study (n = 104)Reduced all-cause mortality (HR 0.13; 95% CI 0.04–0.44) and renal mortality (HR 0.21; 95% CI 0.07–0.64)
 Cheung et al.27Cohort study (n = 79)Reduced angiographic progression of RVD with statins (HR 0.28; 95% CI 0.10–0.77); increased likelihood of regression with statins (HR 4.88; 95% CI 1.32–19.4)
  • Multivariable-adjusted results were provided in the table above (wherever possible).


The major limitation of our study is the non-randomized comparison of statin use with untreated controls. Indeed the overall observational design raises possibilities of selection bias and residual confounding. We adjusted for 74 variables in our analyses and, if anything, statin users appeared to have greater cardiorenal comorbidity at baseline than controls. This imbalance would be expected to bias the results against treatment, yet this was not observed in any of the analyses. Adjusting for year of diagnosis did not materially change the results, suggesting that the present findings are not due solely to secular change. However, several variables could not be adjusted for such as laboratory measures of renal function, cholesterol, and anatomic grade of stenosis. Our data also lack lifestyle measures, such as obesity, smoking, and exercise. Outcomes in the current study were not blindly adjudicated and were based on administrative data. The coding of certain outcomes—such as malignant hypertension—have been validated in other jurisdictions but not in Ontario.

For these reasons, only a large randomized trial can conclusively demonstrate that statins improve prognosis in this setting; however, such a trial is unlikely to be performed given the high prevalence of compelling indications for statins in patients with RVD. In addition, our results might not apply to excluded patients, such as individuals with end stage renal disease, patients younger than 65, or those who died within 4 months of RVD diagnosis.

We cannot know for certain why patients were screened for RVD, but we did ascertain very high frequencies of hypertension (89%), chronic kidney disease (60%), and heart failure (46%) in this cohort. The fact that RVD is being detected in health-care databases, in concert with sizeable event rates, would suggest that patients with symptomatic RVD were likely enroled.

There are several reasons why some patients with RVD in our cohort may not have received statins. Patients who did not receive statins were less likely to have coronary artery disease, diabetes mellitus, and cerebrovascular disease at baseline. Such patients had equivalent rates of ambulatory care visits to general practitioners (P = 0.51) and internists (P = .30). We believe that adherence to treatment was unlikely to be a bar to statin therapy because controls still received an average of ten different drugs in the 6 months prior to cohort accrual. In addition, time-dependent exposure adjusts for medication adherence and drug uptake in both groups. Therefore, initially untreated controls still had a chance to receive statins after ‘time zero' of follow-up, and indeed we found a general increase in statin treatment over the course of the study.

Notwithstanding the above remarks, we found some evidence of under-treatment in our study, with many patients with concomitant extra-renal atherosclerotic vascular disease not receiving statin therapy. Results in patients with or without coronary artery disease were virtually identical as shown in our predefined subgroup analyses. The beginning of our accrual period predated the results of the major statin trials in coronary disease, cerebrovascular disease, and peripheral arterial disease, during which time lipoprotein targets and thresholds became ever more rigorous. By the end of our accrual interval, statin users were outnumbering controls by a ratio of 2 to 1.


These findings suggest an association between statins and cardiorenal prognosis in patients with RVD (in keeping with data from previous studies).1127,29 Our study also underscores the high rates of morbidity and mortality in this condition; overall, 49% of patients suffered a primary event and 37% of patients died during a median follow-up of 3.3 years. Given these rates, which are consistent with the previous literature, patients with RVD require careful surveillance and diligent risk factor modification to prevent cardiorenal complications.2,3 While randomized data do not exist to guide the nature of medical therapy for this condition, our findings suggest that patients with RVD should potentially be considered for statin therapy.


The study was supported by a peer-reviewed grant-in-aid from the Physician Services Incorporated Foundation. D.G.H. was supported by a new investigator award from the Canadian Institutes for Health Research. P.C.A. was supported by a career investigator award from the Heart and Stroke Foundation of Ontario. A.X.G. was supported by a clinician scientist award from the Canadian Institutes for Health Research. The Institute for Clinical Evaluative Sciences is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. The opinions, results and conclusions reported in this article are those of the authors and are independent from the funding sources. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred. D.G.H. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest: None for D.G.H., A.X.G., F.W., P.L. or P.C.A. M.M.M. has received honoraria from Pfizer and AstraZeneca and was employed at Pfizer Global Pharmaceuticals from January 2006 to April 2007. S.W.T. has received honoraria from Merck, BMS-Sanofi-aventis, and Servier, and has received research funding from AstraZeneca, Ortho Biotech, Novartis, and Amgen.


We thank Nick Daneman, Nadia Khan, Marko Mrkobrada, Donald Redelmeier, and Matthew Weir for comments on the manuscript.


View this table:
Appendix 1

Cohort exclusion criteria

ExclusionRationalePatients excluded
Fibromuscular dysplasiaFMD is a non-atherosclerotic cause of RVD typically treated with angioplasty rather than statins. It is most often present in patients under 50n = 21
Death within 120 days of cohort entryOur definition of statin exposure required an initial 120 day exposure assessment window following cohort entry. Patients who died within this window would tend to assort into the non-exposed control group, thereby introducing immortal time bias in the statin groupn = 672
Invalid health card numberCohort entry required a valid health card number to link patients across administrative databasesn = 215
Missing age or sexImportant demographic variables for cohort characterizationn = 1
Non-residents of OntarioNon-residents may have a health-care episode in Ontario and then return home for follow-up care in their own jurisdiction, potentially leading to unreliable outcome and exposure statusn = 320
Nursing home residentsThese individuals receive medication from onsite formularies which are not tracked in the administrative databasesn = 563
End stage renal diseaseBy definition, patients with end stage renal disease prior to cohort entry cannot incur new renal events, such as acute renal failure and new onset dialysisn = 598


View this table:
Appendix 2

Diagnosis coding for study outcomes

Myocardial infarctionCIHI-DAD NACRSAdmission to hospital or the emergency room with a most responsible diagnosis of acute myocardial infarction, as coded by ICD9 410; ICD10 I21, I22
StrokeCIHI-DAD NACRSAdmission to hospital or the emergency room with a most responsible diagnosis of stroke, as coded by ICD9 431, 434, 436; ICD10 I61, I63, I64, G46
Heart failureCIHI-DADAdmission to hospital with a most responsible diagnosis of heart failure, systemic volume overload, or pulmonary oedema, as coded by ICD9 276.6, 425, 428, 514, 518.4; ICD10 E87.7, I09.9, I11.0, I13, I25.5, I42, I43, I50, J81
DialysisCIHI-DAD CIHI-SDS NACRS OHIPIn-hospital or in-community performance of peritoneal dialysis, haemodialysis, or ultrafiltration (including continuous dialysis) as coded by CCP 51.95, 66.98; CCI 1PZ21; OHIP G082 G083, G085, G090, G091, G092, G093, G094, G095, G096, G294, G295, G325, G326, G330, G331, G332, G333, G860, G861, G862, G863, G864, G865, G866, H540, H740, R849
Acute renal failureCIHI-DADAdmission to hospital with a most responsible diagnosis of acute renal failure, pre-renal azotaemia, or anuria, as coded by ICD9 584; ICD10 N17, R39.2
Malignant hypertensionCIHI-DADAdmission to hospital with a most responsible diagnosis of malignant hypertension or hypertensive encephalopathy, as coded by ICD9 401.0, 402.0, 403.0, 404.0, 405.0, 437.2; ICD10 I10.1, I15.01, I15.11, I15.21, I15.81, I15.91, I67.4
Renal hospitalizationCIHI-DADAdmission to hospital with a most responsible renal diagnosis, as coded by ICD9 276.2, 276.6, 276.7, 403, 404, 580–589, 593.9, 788.5, 782.3, 791.0, 791.7, V42.0, V45.1, V56.0, V56.8; ICD10 E10.2x, E11.2x, E13.2x, E14.2x, I12, I13, I15.10, I15.11, M10.3, N00-08, N10-12, N14.1, N14.2, N16.5, N17-19, N25.0, N25.8, N25.9, N26, N27, N28.9, N29.8, R34, R39.2, R80, T86.1, T86.100, T86.101, T86102, Z94.0
RevascularizationCIHI-DAD CIHI-SDS NACRS OHIPIn-hospital or in-community performance of any arterial revascularization procedure, as coded by CCP 48.1, 48.11–48.19, 48.02, 48.03, 48.09, 48.2, 48.3, 50.1, 50.11–50.19, 51.28, 51.25, 51.29, 51.26, 51.22, 51.24, 50.28, 50.38; CCI 1ID76MU, 1IJ26, 1IJ27, 1IJ50, 1IJ54, 1IJ57, 1IJ76, 1IJ80, 1IL35, 1JE50, 1JE57, 1JE87, 1JM58, 1JW35, 1JW50, 1JW57, 1JW76, 1JX50, 1JX57, 1JX76, 1KA50, 1KA57, 1KA76, 1KE26, 1KE35, 1KE50, 1KE57, 1KE76, 1KG26, 1KG35, 1KG50, 1KG57, 1KG76MI, 1KG87, 1KR38; OHIP E645, E645C, E645B, E646, E649, E651, E652, E654, E672, E679, G262, G298, J025, J050, J058, J066, N220, N223, R741, R742, R743, R780, R783, R784, R785, R787, R791, R792, R794, R797, R804, R806, R809, R814, R815, R830, R831, R832, R855, R856, R860, R861, R867, R933, R934, R936, R937, S421, Z434
  • CCI, Canadian Classification of Health Interventions; CCP, Canadian Classification of Diagnostic, Therapeutic and Surgical Procedures; CIHI-DAD, Canadian Institute for Health Information Discharge Abstract Database; CIHI-SDS, Canadian Institute for Health Information Same Day Surgery Database; ICD9, International Classification of Diseases, Ninth Revision; ICD10, International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (Canadian enhancement); NACRS, National Ambulatory Care Reporting System; OHIP, Ontario Health Insurance Plan.


View this table:
Appendix 3

Adjusted covariates in multivariable models

Demographic factorsOther investigations
 Age Echocardiography
 Income Carotid ultrasound
 Locale Coronary angiography
 Sex Holter monitoring
 Stress testing
Cardiovascular risk factors and comorbidities Bone density testing
 Diabetes mellitus Renal artery ultrasound
 Hypertension Renal magnetic resonance angiography
 Cerebrovascular disease Renal computed tomographic angiography
 Peripheral arterial disease Renal angiography (catheter-based)
 Coronary artery disease Captopril nephrography
 Heart failure Renin immunoassay
 Aortic aneurysm Aldosterone immunoassay
 Chronic kidney disease
 Atrial fibrillation/flutterMedications
 Other dysrhythmias Statins
 Calcium channel blockers
Other comorbidities Thiazide diuretics
 Chronic liver disease Alpha-blockers
 Chronic lung disease Beta-blockers
 Dementia Vasodilators
 Peptic ulcer disease Non-statin lipid drugs
 Systemic malignancy Anticoagulants
 Venous thromboembolism Loop diuretics
Chronic ambulatory disease groupings Anti-arrhythmics
 Acute major Potassium-sparing diuretics
 Acute minor Non-steroidal anti-inflammatories
 Likely to recur Angiotensin-converting enzyme inhibitors
 Asthma Angiotensin receptor blockers
 Chronic medical unstable
 Chronic medical stable
 Chronic specialty stable
 Chronic specialty unstable
 Prevention, administration
Health-care access and screening (n, past year)
 Family doctor visits
 Cardiologist visits
 Nephrologist visits
 General internist visits
 Vascular surgeon visits
 Cholesterol tests
 Creatinine tests
 Hospitalizations (past 3 years)
 Distinct drugs (past 6 months)
 Creatinine kinase tests
 Liver transaminase tests


View this table:
Appendix 4

Baseline characteristics for the propensity-matched cohort

VariableValueControl (n = 1061)Statin (n = 1061)Overall (n = 2122)
Income quintile 1 (%)231 (21.8)242 (22.8)473 (22.3)
Income quintile 2 (%)227 (21.4)225 (21.2)452 (21.3)
Income quintile 3 (%)218 (20.5)222 (20.9)440 (20.7)
Income quintile 4 (%)197 (18.6)186 (17.5)383 (18.0)
Income quintile 5 (%)188 (17.7)186 (17.5)374 (17.6)
Female (%)494 (46.6)494 (46.6)988 (46.6)
Rural location (%)158 (14.9)173 (16.3)331 (15.6)
AgeMean ± SD74.07 ± 5.3674.09 ± 5.3674.08 ± 5.36
Median (IQR)73 (70–78)73 (70–78)73 (70–78)
Family doctor ambulatory visits in the past year, nMean ± SD12.61 ± 8.7412.88 ± 8.9412.74 ± 8.84
Median (IQR)11 (7–17)11 (7–17)11 (7–17)
Cardiologist ambulatory visits in the past year, nMean ± SD2.03 ± 2.702.03 ± 2.312.03 ± 2.51
Median (IQR)1 (0–3)1 (0–3)1 (0–3)
Endocrinologist ambulatory visits in the past year, nMean ± SD0.27 ± 1.030.25 ± 0.960.26 ± 1.00
Median (IQR)0 (0–0)0 (0–0)0 (0–0)
Nephrologist ambulatory visits in the past year, nMean ± SD1.03 ± 2.200.94 ± 1.940.99 ± 2.08
Median (IQR)0 (0–1)0 (0–1)0 (0–1)
General internist ambulatory visits in the past year, nMean ± SD1.55 ± 2.411.44 ± 3.091.49 ± 2.77
Median (IQR)1 (0–2)0 (0–2)1 (0–2)
Vascular surgeon ambulatory visits in the past year, nMean ± SD1.70 ± 2.261.69 ± 2.061.70 ± 2.16
Median (IQR)1 (0–3)1 (0–3)1 (0–3)
Number of cholesterol laboratory tests in the past yearMean ± SD0.95 ± 1.070.96 ± 1.000.96 ± 1.04
Median (IQR)1 (0–1)1 (0–1)1 (0–1)
Number of creatinine laboratory tests in the past yearMean ± SD3.08 ± 3.243.04 ± 3.543.06 ± 3.39
Median (IQR)2 (1–4)2 (1–4)2 (1–4)
Number of hospitalizations in past 3 yearsMean ± SD2.34 ± 1.982.24 ± 2.022.29 ± 2.00
Median (IQR)2 (1–3)2 (1–3)2 (1–3)
Number of distinct drugs in the past 6 monthsMean ± SD11.66 ± 6.2011.70 ± 5.4511.68 ± 5.83
Median (IQR)11 (7–15)11 (8–15)11 (8–15)
Number of CK tests in the past yearMean ± SD0.21 ± 0.630.25 ± 0.560.23 ± 0.60
Median (IQR)0 (0–0)0 (0–0)0 (0–0)
Number of ALT/AST tests in the past yearMean ± SD1.06 ± 1.451.10 ± 1.411.08 ± 1.43
Median (IQR)1 (0–1)1 (0–2)1 (0–2)
Diabetes mellitus (%)288 (27.1)306 (28.8)594 (28.0)
Hypertension (%)955 (90.0)951 (89.6)1906 (89.8)
Cerebrovascular disease (%)340 (32.0)331 (31.2)671 (31.6)
Peripheral arterial disease (%)487 (45.9)487 (45.9)974 (45.9)
Coronary artery disease (%)613 (57.8)618 (58.2)1231 (58.0)
Heart failure (%)488 (46.0)495 (46.7)983 (46.3)
Aortic aneurysm (%)407 (38.4)392 (36.9)799 (37.7)
Chronic kidney disease (%)635 (59.8)635 (59.8)1270 (59.8)
Atrial fibrillation/flutter (%)160 (15.1)169 (15.9)329 (15.5)
Other arrhythmias (%)292 (27.5)280 (26.4)572 (27.0)
Chronic liver disease (%)22 (2.1)30 (2.8)52 (2.5)
Chronic lung disease (%)441 (41.6)433 (40.8)874 (41.2)
Dementia (%)47 (4.4)57 (5.4)104 (4.9)
Peptic ulcer disease (%)177 (16.7)181 (17.1)358 (16.9)
Systemic malignancy (%)40 (3.8)44 (4.1)84 (4.0)
Prior venous thrombosis/thromboembolism (%)111 (10.5)121 (11.4)232 (10.9)
CADG—‘acute minor’ (%)999 (94.2)1007 (94.9)2006 (94.5)
CADG—‘acute major’ (%)1045 (98.5)1044 (98.4)2089 (98.4)
CADG—‘likely to recur’ (%)968 (91.2)966 (91.0)1934 (91.1)
CADG—‘asthma’ (%)115 (10.8)118 (11.1)233 (11.0)
CADG—‘chronic medical unstable’ (%)1060 (99.9)1060 (99.9)2120 (99.9)
CADG—‘chronic medical stable’ (%)1042 (98.2)1046 (98.6)2088 (98.4)
CADG—‘chronic specialty stable’ (%)144 (13.6)129 (12.2)273 (12.9)
CADG—‘eye dental’ (%)395 (37.2)401 (37.8)796 (37.5)
CADG—‘chronic specialty unstable’ (%)296 (27.9)305 (28.7)601 (28.3)
CADG—‘psychosocial’ (%)547 (51.6)557 (52.5)1104 (52.0)
CADG—‘prevention, administration’ (%)718 (67.7)726 (68.4)1444 (68.0)
Echocardiography (%)384 (36.2)409 (38.5)793 (37.4)
Carotid ultrasound (%)269 (25.4)251 (23.7)520 (24.5)
Coronary angiography (%)130 (12.3)135 (12.7)265 (12.5)
Holter monitoring (%)107 (10.1)104 (9.8)211 (9.9)
Stress testing (%)320 (30.2)312 (29.4)632 (29.8)
Bone density testing (%)45 (4.2)52 (4.9)97 (4.6)
Renal artery ultrasound (%)440 (41.5)441 (41.6)881 (41.5)
Renal MRA (%)69 (6.5)73 (6.9)142 (6.7)
Renal CTA (%)270 (25.4)260 (24.5)530 (25.0)
Plain renal angiogram (%)276 (26.0)278 (26.2)554 (26.1)
Captopril nephrography (%)178 (16.8)161 (15.2)339 (16.0)
Renin immunoassay (%)7 (0.7)6 (0.6)13 (0.6)
Aldosterone immunoassay (%)10 (0.9)8 (0.8)18 (0.8)
Calcium channel blocker (%)616 (58.1)618 (58.2)1234 (58.2)
Thiazide diuretic (%)333 (31.4)336 (31.7)669 (31.5)
Alpha-blocker (%)107 (10.1)116 (10.9)223 (10.5)
Beta-blocker (%)576 (54.3)558 (52.6)1134 (53.4)
Vasodilator (%)94 (8.9)82 (7.7)176 (8.3)
Oral anticoagulant (%)155 (14.6)148 (13.9)303 (14.3)
Loop diuretic (%)520 (49.0)537 (50.6)1057 (49.8)
Antiplatelet (%)383 (36.1)389 (36.7)772 (36.4)
Anti-arrhythmic (%)169 (15.9)174 (16.4)343 (16.2)
Potassium-sparing diuretic (%)105 (9.9)104 (9.8)209 (9.8)
NSAID (%)119 (11.2)120 (11.3)239 (11.3)
ACE-inhibitor (%)499 (47.0)515 (48.5)1014 (47.8)
ARB (%)123 (11.6)114 (10.7)237 (11.2)
  • ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker; CADG, chronic ambulatory disease group; CK, creatine kinase; CTA, computed tomographic angiography; IQR, interquartile range; MRA, magnetic resonance angiography; NSAID, non-steroidal anti-inflammatory drug.


View Abstract