A meta-analysis yields a ranking of antihypertensive drug classes according to their association with incident diabetes.
Concerns that some antihypertensive medications, particularly thiazide diuretics and beta-blockers, can increase the risk for new-onset diabetes have been difficult to validate. Now, researchers have conducted a "network meta-analysis" to study new-onset diabetes in 22 long-term, randomized antihypertensive-drug trials (involving 143,158 subjects). This type of analysis enables comparison of different treatments across similar trials even when specific strategies have not been compared within trials.
When diuretics were used as the referent, the risk for incident diabetes was significantly lower with angiotensin-receptor blockers (ARBs; odds ratio, 0.57), ACE inhibitors (OR, 0.67), calcium-channel blockers (OR, 0.75), and placebo (OR, 0.77) — and nonsignificantly lower with beta-blockers (OR, 0.90). Compared with placebo, only ARBs were associated with significantly reduced risk for incident diabetes (OR, 0.75) and only diuretics with significantly increased risk (OR, 1.30). Excluding trials in which some subjects did not have hypertension did not change the findings substantially.
Comment: Among the limitations of this meta-analysis are that the definition of diabetes changed over time, diabetes was not a prespecified endpoint in many of the trials, and antihypertensive strategies often involved adding second or third drugs. In addition, the trials didn’t have many new diabetes cases during their roughly 3 to 5 years of follow-up, so the absolute differences in incidence between antihypertensive agents never exceeded 3.6%. However, even this risk increase might matter in high-risk populations (e.g., obese people), particularly in the longer term. Therefore, it’s reasonable to use this study’s diabetes-risk hierarchy of antihypertensive drug classes to inform medication selection in such populations.
— Harlan M. Krumholz, MD, SM
Published in Journal Watch Cardiology February 7, 2007
Elliott WJ and Meyer PM. Incident diabetes in clinical trials of antihypertensive drugs: A network meta-analysis. Lancet 2007 Jan 20; 369:201-7.