Angiotensin-Converting Enzyme Inhibitors or Angiotensin-Receptor Blockers for Advanced Chronic Kidney Disease : A Systematic Review and Retrospective Individual Participant-Level Meta-analysis of Clinical Trials

Abstract

Background: In patients with advanced chronic kidney disease (CKD), the effects of initiating treatment with an angiotensin-converting enzyme inhibitor (ACEi) or angiotensin-receptor blocker (ARB) on the risk for kidney failure with replacement therapy (KFRT) and death remain unclear.

Purpose: To examine the association of ACEi or ARB treatment initiation, relative to a non-ACEi or ARB comparator, with rates of KFRT and death.

Data sources: Ovid Medline and the Chronic Kidney Disease Epidemiology Collaboration Clinical Trials Consortium from 1946 through 31 December 2023.

Study selection: Completed randomized controlled trials testing either an ACEi or an ARB versus a comparator (placebo or antihypertensive drugs other than ACEi or ARB) that included patients with a baseline estimated glomerular filtration rate (eGFR) below 30 mL/min/1.73 m².

Data extraction: The primary outcome was KFRT, and the secondary outcome was death before KFRT. Analyses were done using Cox proportional hazards models according to the intention-to-treat principle. Prespecified subgroup analyses were done according to baseline age (<65 vs. ≥65 years), eGFR (<20 vs. ≥20 mL/min/1.73 m²), albuminuria (urine albumin-creatinine ratio <300 vs. ≥300 mg/g), and history of diabetes.

Data synthesis: A total of 1739 participants from 18 trials were included, with a mean age of 54.9 years and mean eGFR of 22.2 mL/min/1.73 m², of whom 624 (35.9%) developed KFRT and 133 (7.6%) died during a median follow-up of 34 months (IQR, 19 to 40 months). Overall, ACEi or ARB treatment initiation led to lower risk for KFRT (adjusted hazard ratio, 0.66 [95% CI, 0.55 to 0.79]) but not death (hazard ratio, 0.86 [CI, 0.58 to 1.28]). There was no statistically significant interaction between ACEi or ARB treatment and age, eGFR, albuminuria, or diabetes (P for interaction > 0.05 for all).

Limitation: Individual participant-level data for hyperkalemia or acute kidney injury were not available.

Conclusion: Initiation of ACEi or ARB therapy protects against KFRT, but not death, in people with advanced CKD.

Primary funding source: National Institutes of Health. (PROSPERO: CRD42022307589).

Figure 1.

Cumulative incidence of KFRT in unadjusted and adjusted analysis by treatment arm for the a) overall cohort; b) placebo-controlled trials; and c) other anti-hypertensive agent trials. The median follow-up time was 34 months in the overall analysis, 32 months in the placebo-controlled trials, and 37 months in the other anti-hypertensive agent trials.

Figure 1.

Cumulative incidence of KFRT in unadjusted and adjusted analysis by treatment arm for the a) overall cohort; b) placebo-controlled trials; and c) other anti-hypertensive agent trials. The median follow-up time was 34 months in the overall analysis, 32 months in the placebo-controlled trials, and 37 months in the other anti-hypertensive agent trials.

Figure 1.

Cumulative incidence of KFRT in unadjusted and adjusted analysis by treatment arm for the a) overall cohort; b) placebo-controlled trials; and c) other anti-hypertensive agent trials. The median follow-up time was 34 months in the overall analysis, 32 months in the placebo-controlled trials, and 37 months in the other anti-hypertensive agent trials.

Figure 2.

Forest plot of the adjusted effect of ACEi or ARB initiation on the risk of KFRT in the overall, placebo-controlled trials and other anti-hypertensive agent comparator trials, as well as in each individual trial. This plot shows the adjusted hazard ratios of ACEi or ARB initiation versus comparator on the risk of KFRT therapy in each trial and in pooled analysis. Included studies are listed on the left (See Table 1 for full study name). Adjusted hazard ratios derived from Cox proportional hazards analyses of each individual trial are shown both pictorially in the forest plot and numerically in the column on the right. A black dot indicates the hazard ratio and horizontal bars represent 95% CIs; arrows indicate that 95%CI extends beyond 3.0 or 0.3. The dashed line is the no effect value. The x-axis is presented on a log scale. The combined rows show the results of one-stage approach to IPD meta-analysis for the overall cohort, or for the placebo-controlled and other-antihypertensive agent comparator trials separately in adjusted analysis using Cox models, but including centered covariates and interaction terms between centered covariates and study. Of note, in the one-stage approach to derivation of the overall hazard ratio and subgroup hazard ratios for placebo versus other anti-hypertensive comparators, all trials were included in the adjusted models with centered covariates, though 3 trials (including Lewis, ABCD, and HKVIN trials) had very few patients and events and effectively did not contribute to estimates of the risk. No trials were excluded intentionally from analyses when deriving all point estimates. IDNT is represented twice, once in comparison to the placebo arm and another time in comparison to amlodipine (this was a 3-arm trial). ( ) after each study name refers to citations