Tolerability patterns diverge across blood pressure drug classes in 716 trials
A large network meta-analysis of short-term randomized trials has drawn a sharper map of a familiar problem in hypertension care: blood pressure-lowering drugs do not all behave the same when patients start taking them. Across 716 double-blind trials involving 159,362 participants, the central outcome was treatment discontinuation due to adverse events, a practical endpoint that captures whether a regimen can survive beyond the prescription pad and into everyday use.
The study focused on the five major antihypertensive classes—angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, β-blockers, calcium channel blockers, thiazide and thiazide-like diuretics—and selected combinations built from those agents. Mean follow-up was 8.6 weeks, which places the analysis squarely in the early treatment window when dizziness, edema, cough, and other symptoms are most likely to determine whether a patient stays on therapy. Baseline blood pressure was high, with a mean of 158/100 mm Hg, and the cohort reflected a typical hypertension trial population: mean age 54.6 years and 44% female.
One result stood out immediately. Calcium channel blockers were associated with more discontinuation than placebo, as were angiotensin-converting enzyme inhibitors combined with calcium channel blockers and β-blockers combined with thiazide diuretics. In contrast, ARB-containing regimens were consistently better tolerated than placebo on the discontinuation endpoint, with statistically significant reductions seen for ARB monotherapy and ARBs plus calcium channel blockers.
Withdrawal risk was not uniform, even within combination therapy
The discontinuation signal varied enough to challenge the common assumption that more intensive therapy automatically brings more intolerance. Compared with placebo, discontinuation due to adverse events increased with calcium channel blockers alone, with an odds ratio of 1.43 and a risk difference of 1.2%. The ACE inhibitor plus calcium channel blocker pairing carried a similar direction of effect, with an odds ratio of 1.46 and a risk difference of 1.1%. The highest estimate among the combinations reported here came from β-blockers plus thiazide diuretics, with an odds ratio of 1.58 and a risk difference of 1.7%.
One sentence captures the bedside consequence: some antihypertensive combinations are more likely than placebo to push a patient off therapy before blood pressure control has a chance to improve.
The pattern was not simply a matter of one drug class being harsh and another being gentle. Several combination regimens performed better than monotherapy in the ranking analysis, suggesting that tolerability may improve when lower doses or complementary mechanisms are paired. In network meta-analysis terms, five combination regimens and two monotherapy regimens had higher surface under the cumulative ranking curve values than placebo for treatment discontinuation due to adverse events. In plain clinical language, some dual-drug strategies appear to reduce the symptom burden enough to keep patients engaged with treatment.
ARB-based regimens carried the lowest withdrawal burden
Among the clearest findings, ARB-containing regimens showed fewer discontinuations than placebo. ARB monotherapy produced an odds ratio of 0.73 and a risk difference of −0.8%, while ARBs plus calcium channel blockers produced an even stronger signal, with an odds ratio of 0.61 and a risk difference of −1.2%. Those numbers matter because they suggest that tolerability may not merely be acceptable with ARB-based treatment; it may be better than placebo in the short-term trial setting.
That result is biologically plausible. ARBs avoid the cough and angioedema pathway linked to bradykinin accumulation with ACE inhibition, and they often deliver effective blood pressure reduction without the vasodilatory and autonomic symptoms that can limit adherence in other classes. When a therapy lowers pressure without provoking the physical discomfort that patients notice at home—lightheadedness, fatigue, unpleasant swelling—it is more likely to remain on board long enough to matter for cardiovascular prevention.
The finding does not mean ARBs are symptom-free. It means that, in aggregate across short-term trials, they generated fewer withdrawals than placebo and lower withdrawal rates than several other regimens. For clinicians choosing an initial agent for a patient worried about side effects, that distinction can influence whether treatment starts at all.
Dizziness rose across all regimens while headache generally fell
The adverse-event profile was not limited to the primary withdrawal endpoint. All regimens significantly increased dizziness compared with placebo, a result that fits the hemodynamic effect of lowering blood pressure during the early phase of therapy. Patients who move from untreated hypertension to rapid pressure reduction may feel the drop before they see any long-term benefit, especially if volume status, age, or concurrent medications amplify orthostatic symptoms.
Headache moved in the opposite direction for most regimens. All but calcium channel blockers significantly decreased headache compared with placebo. The mechanism is not difficult to imagine: elevated pressure itself can contribute to head pain, so effective treatment may relieve a symptom that patients otherwise attribute to medication. This duality is clinically important. A patient may experience dizziness from the therapy while simultaneously losing the headache that brought them to care in the first place.
Edema and cough were part of the prespecified adverse-event landscape, but the summary emphasized the broader pattern rather than isolated symptom counts. That approach reflects the central problem in hypertension pharmacotherapy: adherence is usually decided by the overall symptom experience, not by a single adverse event in isolation.
What short-term trials can and cannot answer
The analysis was built on fixed-effect network meta-analysis, with results expressed as odds ratios, 95% credible intervals, and ranking probabilities. Trial data were extracted independently by two reviewers, and the final statistical analysis was completed in April 2026. Those methods provide a useful comparative framework, but they also bring the standard assumptions of network meta-analysis: indirect comparisons, trial-level aggregation, and dependence on how well the underlying studies fit together.
These are short-term trials, with follow-up ranging from 4 to 26 weeks. The analysis therefore speaks most directly to the first weeks of treatment, when adverse effects most often drive early discontinuation. It does not settle long-term tolerability, real-world adherence over years, or the extent to which patient-specific factors such as age, frailty, prior medication exposure, or comorbid illness should modify drug choice.
The most actionable message is narrow. Blood pressure-lowering drugs are not interchangeable when tolerability is the outcome. Some regimens, especially ARB-based combinations, were associated with fewer withdrawals than placebo, while others increased the likelihood of stopping treatment because of adverse events. For clinicians trying to get a patient through the first month of therapy, that difference can determine whether blood pressure control begins at all.
The next question is not whether antihypertensive drugs work—they do—but which regimen best balances efficacy and tolerability for the patient sitting in front of the clinician, with the understanding that trial-level averages cannot predict every individual response.