Hypertension remains one of the most undertreated chronic conditions in developed nations — not because effective drugs are unavailable, but because patients stop taking them. Understanding which drug classes and combinations are most likely to trigger adverse effects that lead to discontinuation is arguably as important as knowing which lower blood pressure most effectively. This large-scale synthesis reframes the antihypertensive conversation around real-world tolerability.
Drawing on 716 double-blind randomized controlled trials encompassing roughly 159,000 participants, this network meta-analysis systematically compared adverse-event-driven discontinuation rates across all five major antihypertensive classes — ACE inhibitors, angiotensin II receptor blockers (ARBs), beta-blockers, calcium channel blockers (CCBs), and thiazide/thiazide-like diuretics — as well as their dual combinations, over follow-up windows of 4 to 26 weeks. Odds ratios and 95% credible intervals were derived using fixed-effect network meta-analytic methods, with surface-under-the-cumulative-ranking curves used to rank each regimen. Secondary outcomes included class-specific side-effect signatures: headache, dizziness, peripheral edema, and cough — the quartet most predictably linked to patient-initiated stopping.
This analysis is methodologically significant because network meta-analysis allows indirect comparisons across trials that never directly tested certain drug pairs against each other, effectively synthesizing a richer evidence base than any individual head-to-head trial could provide. In the broader cardiovascular pharmacology literature, ARBs have long been regarded as the best-tolerated class owing to their absence of bradykinin-mediated cough, while ACE inhibitors carry a well-documented cough burden affecting 10–15% of users, particularly in East Asian populations. CCBs are frequently implicated in peripheral edema, and beta-blockers in fatigue and dizziness. What this study adds is a granular ranking across combinations — a clinical question that individual trials rarely address. The primary caveat is the short observation window: 8.6 weeks on average may not capture adverse effects that emerge with longer exposure, such as metabolic dysregulation from thiazides or sexual dysfunction from beta-blockers. Nonetheless, for a condition where first-prescription tolerability often predicts long-term adherence, this evidence base is both timely and clinically actionable at a population level.