Clinical and Experimental Nephrology

, Volume 18, Issue 2, pp 278–281

Statins in chronic kidney disease: what do meta-analyses tell us?

Authors

    • Section of Nephrology, Boston Medical CenterBoston University School of Medicine
Review Article WCN 2013 Satellite Symposium ‘‘Kidney and Lipids’’

DOI: 10.1007/s10157-013-0889-2

Cite this article as:
Upadhyay, A. Clin Exp Nephrol (2014) 18: 278. doi:10.1007/s10157-013-0889-2

Abstract

Dyslipidemia is common in patients with chronic kidney disease (CKD), and previous reports indicate that a significant number of CKD patients with dyslipidemia do not receive statin therapy. This article reviews two recent meta-analyses on statin therapy in CKD which summarize results from randomized controlled trials that have reported on hard clinical outcomes and major adverse events. Despite differences in methodology, both meta-analyses show that statin therapy is safe and effective in preventing mortality and major cardiovascular events in non-dialysis-dependent CKD patients. However, there is very limited evidence to support the use of statins in patients on dialysis, and statin therapy was not found to be effective in reducing the risk of kidney failure or decline in kidney function.

Keywords

StatinDyslipidemiaChronic kidney diseaseMeta-analysis

Introduction

Chronic kidney disease (CKD) is associated with a high risk of cardiovascular disease (CVD) [1]. Dyslipidemia, an important modifiable risk factor for atherosclerotic CVD, is common in CKD patients [2], and various reports have suggested that CKD be considered the highest coronary heart disease risk category while managing lipid abnormalities [35]. Despite this, studies from the USA and Canada show that a significant proportion of patients with CKD and elevated low-density lipoprotein (LDL) cholesterol do not receive statin therapy [6, 7]. Uncertainty among health care providers on whether dyslipidemia imparts similar risk for CVD in CKD patients as in the general population and concern over higher rates of treatment-related side-effects are some of the factors potentially contributing to the lower use of statin therapy in CKD patients.

This article reviews two meta-analyses on statin therapy in CKD published in the August 2012 issue of the Annals of Internal Medicine [8, 9]. The first meta-analysis (Upadhyay et al.) was done by our group to supplement the evidence review which was being conducted for the Kidney Disease: Improving Global Outcomes (KDIGO) guideline on lipid management in CKD [8]. The second meta-analysis (Palmer et al.) was performed independently by researchers from New Zealand, Australia, USA, Canada and Italy, and builds on the systematic review published by the same group in 2008 [9, 10]. Both meta-analyses summarize results from randomized controlled trials (RCTs) that have reported on hard clinical endpoints and major adverse events associated with statin therapy in CKD patients.

Meta-analysis by Upadhyay and colleagues

The details of data sources, search, and study selection criteria are summarized in Table 1 [8]. Although this meta-analysis aimed to evaluate different lipid-lowering therapies in CKD patients, all of the studies included examined statin or ezetimibe plus statin combination therapy. Reviewers, in addition to extracting quantitative data on clinical outcomes and estimating pooled effects by using a standard random-effects model, also applied the modified grading of recommendations, assessment, development, and evaluation (GRADE) approach used in KDIGO guidelines to provide a qualitative summary of the strength of evidence for each outcome [11, 12]. A pre-specified subgroup analysis was performed for non-dialysis-dependent and dialysis-dependent CKD patients. Post-hoc meta-regression analyses were also performed to test the associations between baseline and net change in LDL cholesterol levels and cardiac mortality and CVD events.
Table 1

Summary of data sources, search and selection criteria

 

Upadhyay et al.

Palmer et al.

Data source

MEDLINE, Cochrane central register of controlled trials, Cochrane database of systematic reviews, and systematic review for 2003 KDOQI lipid guideline [4]

EMBASE, Cochrane central register of controlled trials, Cochrane Renal Group’s Specialized Register, 2008 systematic review by the same group on lipid-lowering in CKD [10]

Last search date

November 2011

February 2012

Population studied

CKD, including CKD subgroups of general population trials

CKD, including CKD subgroups of general population trials

Study design

RCT

RCT

Interventions studied

Lipid-lowering therapy

Statins

Minimal sample size

100/arm

No restriction

Minimum follow-up

6 months

8 weeks

Inclusion of trials published only in abstract form

No

Yes

KDOQI kidney disease outcomes quality initiative, CKD chronic kidney disease, RCT randomized controlled trials

After reviewing more than 13,000 abstracts, a total of 20 articles reporting results from 18 RCTs were included in the meta-analysis. Five of 18 RCTs were conducted exclusively in the CKD population while 13 RCTs were conducted in the general population but provided results for the CKD subgroup. Of 5 RCTs conducted in the CKD population, 2 were conducted in patients receiving hemodialysis, 2 were conducted in patients with mild to advanced CKD (including those receiving hemodialysis or peritoneal dialysis), and 1 was conducted in kidney transplant recipients. In general, patients with CKD in 13 general-population RCTs providing results for the CKD subgroup had an early Stage 3 CKD.

The meta-analysis results for clinical outcomes are summarized in Table 2. Statin therapy was found to be effective in preventing cardiac mortality and atherosclerosis-mediated CVD events like myocardial infarctions and revascularization procedures. There was also a suggestion of benefit for all-cause mortality but this was limited to studies in non-dialysis-dependent CKD patients and the results were highly heterogeneous. Meta-regression analyses for cardiac mortality and CVD events did not find significant associations between the positive results and baseline LDL cholesterol level or net change of LDL cholesterol level in trials. Statin therapy, however, was not effective in preventing kidney failure [risk ratio (RR) 0.97, 95 % confidence interval (CI) 0.90–1.05] or decline in kidney function (RR 0.91, 95 % CI 0.78–1.06). The results also showed that statins have a favorable safety profile in CKD. A qualitative summary of the rates of pre-specified adverse events (total adverse events, drug discontinuation from adverse events, rhadomyolysis, clinically significant liver function abnormality, and new onset cancer) showed that they were mostly similar between the statin and placebo arms of the trials analyzed.
Table 2

Summary of results for major clinical outcomes

Outcome

Upadhyay et al.

Palmer et al.

Quality

Risk ratio (95 % CI)

Quality

Risk ratio (95 % CI)

All-cause mortality

 All trials

Moderate

0.91 (0.83–0.99)

Moderate to high

0.89 (0.82–0.97)

 ND-CKD

 

0.83 (0.70–0.98)

 

0.81 (0.74–0.88)

 D-CKD

 

0.96 (0.90–1.02)

 

0.96 (0.88–1.04)

CVD mortality

 All trials

High

0.96 (0.87–1.06)

Moderate to high

0.86 (0.78–0.95)

 ND-CKD

 

1.00 (0.25–3.95)

 

0.78 (0.68–0.89)

 D-CKD

 

1.00 (0.87–1.14)

 

0.94 (0.82–1.07)

Cardiac mortality

 All trials

High

0.82 (0.74–0.91)

NA

NA

 ND-CKD

 

0.67 (0.48–0.94)

 

NA

 D-CKD

 

0.78 (0.68–0.89)

 

NA

CVD events

 All trials

Moderate

0.78 (0.71–0.86)

Moderate to high

0.78 (0.72–0.85)

 ND-CKD

 

0.77 (0.71–0.83)

 

0.76 (0.73–0.80)

 D-CKD

 

0.96 (0.80–1.15)

 

0.95 (0.87–1.03)

CVD events, excluding revascularization procedures

 All trials

High

0.94 (0.86–1.03)

NA

NA

 ND-CKD

 

0.60 (0.36–0.99)

 

NA

 D-CKD

 

0.96 (0.88–1.05)

 

NA

Myocardial infarction

 All trials

Moderate

0.74 (0.67–0.81)

NA

0.76 (0.68–0.86)

 ND-CKD

 

0.74 (0.65–0.83)

 

0.76 (0.42–0.72)

 D-CKD

 

0.72 (0.56–0.92)

 

0.87 (0.71–1.07)

Stroke

 All trials

Very low

0.90 (0.63–1.27)

NA

0.86 (0.62–1.20)

 ND-CKD

 

0.72 (0.48–1.07)

 

0.61 (0.38–0.98)

 D-CKD

 

1.47 (1.09–2.00)

 

1.30 (0.79–2.11)

ND-CKD non-dialysis-dependent chronic kidney disease, D-CKD dialysis-dependent CKD, NA not available

Meta-analysis by Palmer and colleagues

The details of data sources, search, and study selection criteria are summarized in Table 1 [9]. In addition to quantitatively pooling effect estimates using the DerSimonian and Laird random-effects model, reviewers of this meta-analysis also summarized risk of bias in individual trials using a pre-specified criteria, and reported on the quality of evidence for some outcomes using the GRADE approach [11]. A priori subgroup analyses were performed by dialysis and kidney transplant status.

The study selection criteria for Palmer et al.’s meta-analysis was broader than that of Upadhyay et al.’s meta-analysis, and 86 comparisons from 80 RCTs involving more than 50,000 patients met the inclusion criteria for statin versus placebo or no treatment comparison, with 48 comparisons in patients not receiving dialysis, 21 in patients on dialysis, and 17 in kidney transplant recipients. Assessment of study bias revealed that only 15 trials had adequate allocation concealment, 36 had blinded patients or investigators, and 18 had used intention-to-treat analysis.

The meta-analysis results for clinical outcomes are summarized in Table 2. In essence, the results showed that the benefits of statin therapy for mortality and CVD outcome differed depending on the stage of CKD. Moderate-to-high-quality evidence indicated that statin therapy reduced all-cause mortality and CVD events in patients with non-dialysis-dependent CKD whereas similar quality evidence indicated that statin treatment, despite markedly lowering serum cholesterol level, had little or no benefit for individuals receiving dialysis. Evidence for statin therapy in kidney transplant recipients was noted to be sparse. In contrast to Upadhyay et al.’s meta-analysis, Palmer et al.’s meta-analysis did not find a benefit of statin therapy in dialysis patients, even for preventing atherosclerotic-mediated events like myocardial infarction. The results for kidney function and proteinuria were noted to be highly heterogeneous and meaningful interpretation could not be made. Similar to Upadhyay et al.’s meta-analysis, statins were found to have a favorable safety profile in CKD patients. Statin treatment was not associated with an increased risk of cancer, myalgia, elevated creatinine kinase level, abnormal liver function, or treatment withdrawal.

Strengths and limitations of two meta-analyses

The main strength of the two meta-analyses is their methodological rigor. The results and conclusions from the meta-analyses are also very similar, and this consistency, even when the study selection criteria and some of the analytic methods used are quite different, provides confidence in their results. Some of the minor differences between the two meta-analyses are likely due to differences in the study inclusion criteria and differences in the adjudication method for CVD outcomes [13, 14]. In spite of these strengths, there are also important limitations. The studies reviewed were heterogeneous in population, interventions, and reporting of outcomes. These differences could have hindered comparability across studies and may have limited the reliable interpretation of pooled results. In particular, the results of studies comparing statin therapy to placebo in patients with non-dialysis-dependent CKD were mostly from post-hoc CKD subgroup analyses of large lipid trials, and inclusion of these unplanned subgroup analyses could have introduced bias. There was some statistical heterogeneity as well, but analyses by Palmer and colleagues showed that the heterogeneity in treatment effect for major outcomes was not seen when trials were analyzed by CKD stage. Finally, the reporting of adverse events was not uniform across studies and it was not always clear whether the lack of an adverse event was because of its true absence or inadequate assessment.

Conclusions

The results of these two meta-analyses provide reasonably strong evidence supporting the use of statins for the prevention of mortality and CVD events in non-dialysis-dependent CKD patients. There is limited evidence to support statin use in patients on dialysis, and treatment decisions in this group need to be taken after the careful consideration of patients’ underlying cardiovascular co-morbidities as well as their values and preferences. These meta-analyses also show that statin therapy is not effective in reducing the risk of kidney failure or decline in kidney function. The upcoming KDIGO guideline is expected to evaluate the results of these meta-analyses and other available evidence to provide clinical practice recommendations for the management of lipid abnormalities in CKD patients.

Conflict of interest

None.

Copyright information

© Japanese Society of Nephrology 2013