Review: ACE inhibitors delay microalbuminuria in diabetes without nephropathy and reduce mortality in diabetic nephropathy.

and commentary also appear in ACP Journal Club . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . For correspondence: Dr G F Strippoli, University of Sydney, Sydney, New South Wales, Australia. [email protected] Sources of funding: Cochrane Renal Group and Australia/Europe Endeavour Scholarship. Angiotensin converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs) v calcium antagonists (CAs) or placebo for diabetes with and without nephropathy* Outcomes Comparisons Number of trials (n) Relative risk (95% CI) Onset of microalbuminuria in diabetes without nephropathy ACE inhibitors v CAs 4 (1210) 0.58 (0.40 to 0.84) ACE inhibitors v placebo 6 (3840) 0.60 (0.43 to 0.84) All cause mortality in diabetic nephropathy ACE inhibitors v placebo 20 (2838) 0.79 (0.63 to 0.99) ARBs v placebo 4 (3329) 0.99 (0.85 to 1.17) *CI defined in glossary. Not significant. Commentary C hoosing antihypertensive treatment can be difficult because of conflicting evidence in the literature, comorbid conditions of patients, and competing drug class benefits. Strippoli et al used meta-analysis to examine whether ACE inhibitors or ARBs have advantages over other classes for cardiorenal end points in patients with diabetes. They concluded that the use of ACE inhibitors or ARBs protected the kidneys and that ACE inhibitors reduced all cause mortality. These conclusions are similar to their previous work, 2 but stand in stark contrast to a meta-analysis by Casas et al that showed that the renal sparing effects of ACE inhibitors and ARBs are only present in placebo controlled trials and vanish with active comparators or when blood pressure control is taken into account. Conflicting conclusions can sometimes be resolved by applying the ‘‘face validity test’’—that is, does the analysis include only trials that have similar populations, interventions, and outcomes? In a meta-analysis, the test for heterogeneity can determine whether the statistical properties of the trials weigh against their combination. However, trials passing this test should not automatically be combined. In the meta-analysis by Casas et al, about 85% of the patients came from ALLHAT, a trial that excluded patients with severe renal disease and in whom ESRD was unexpected. Such a trial should not be included in an analysis of ESRD prevention. Trial selection also influenced Strippoli et al’s conclusion that ACE inhibitors but not ARBs reduce mortality in diabetes. The ACE inhibitor trials included the large cardiovascular oriented HOPE study, which excluded severe renal disease. The smaller renal oriented ARB trials only included patients with advanced renal disease—a population in which other cardioprotective therapies have failed. Both Casas et al’s conclusions regarding the lack of efficacy of ACE inhibitors and ARBs beyond control of hypertension, and Strippoli et al’s conclusions about the differential cardioprotective effects of ACE inhibitors and ARBs, are suspect for these reasons. Meta-analytic techniques are a powerful method for combining the results of clinical trials. The face validity test is more powerful than the statistical test of heterogeneity in determining whether sufficiently similar trials have been included. Philip A McFarlane, MD St Michael’s Hospital Toronto, Ontario, Canada 1 Strippoli GF, Craig M, Schena FP, et al. J Am Soc Nephrol 2005;16:3081– 91. 2 Strippoli GF, Craig M, Deeks JJ, et al. BMJ 2004;329:828. 3 Casas JP, Chua W, Loukogeorgakis S, et al. Lancet 2005;366:2026–33. 144 THERAPEUTICS www.evidence-basedmedicine.com EBM Volume 11 October 2006 group.bmj.com on October 11, 2017 Published by http://ebm.bmj.com/ Downloaded from