Zillich · Hypertension (Dallas, Tex. : 1979) 2006 · systematic review and meta-regression · n=59 studies (83 study arms)

Thiazide diuretics, potassium, and the development of diabetes: a quantitative review.

Cited 454 times in the scientific literature.

Level 1 - systematic review of randomized trials

Systematic review and quantitative synthesis of clinical trials

PubMed 16801488 · doi:10.1161/01.HYP.0000231552.10054.aa · record verified 2026-08-29

What was done

A systematic literature search from 1966 to June 2004 identified clinical trials evaluating thiazide diuretics that reported metabolic outcomes for both potassium and glucose. Data from 59 clinical trials comprising 83 thiazide treatment arms were synthesized using a sample-size-weighted statistical model to evaluate the joint association between mean changes in potassium and glucose. Sensitivity analyses examined subset effects, covariates, and inverse-variance weighting.

What was found

The Pearson correlation coefficient, weighted by study sample size, between changes in potassium and changes in glucose across the 83 study arms was -0.54 (95% CI, -0.67 to -0.36; P < 0.01). Greater drops in serum potassium were significantly associated with greater increases in blood glucose. Sensitivity and inverse-variance analyses supported this association.

Why it matters

The findings identify potassium depletion as a key correlate of thiazide-induced hyperglycemia, suggesting that preventing or correcting hypokalemia may attenuate glucose intolerance and lower the risk of developing diabetes during thiazide therapy.

Limits

The analysis relied on aggregate study-arm data rather than individual participant data, precluding detailed control for individual baseline risk factors. Included trials varied substantially in sample size, follow-up duration, thiazide type, and dosage. The abstract reports correlation coefficients but lacks absolute glucose values or clinical diabetes event counts.

Cited by