TRIDENT
(2026)Objective
To determine whether a single pill combining three low-dose antihypertensive agents, added to standard care, lowers blood pressure and reduces recurrent stroke risk in patients with prior intracerebral hemorrhage.
Study Summary
• Mean systolic BP during follow-up: 127 vs 138 mm Hg (between-group difference 9 mm Hg)
• Major cardiovascular events: 6.6% vs 9.8% (P=0.04)
• Serious adverse events: 23.2% vs 26.0%; early discontinuation due to adverse events: 13.6% vs 6.0% (most commonly ≥20% rise in serum creatinine)
Intervention
Once-daily single pill containing telmisartan 20 mg, amlodipine 2.5 mg, and indapamide 1.25 mg, in addition to standard antihypertensive care.
Inclusion Criteria
Adults ≥18 years with a history of spontaneous intracerebral hemorrhage, clinically stable, with systolic BP 130–160 mm Hg at baseline (on any background antihypertensive therapy), and no contraindication to any component of the triple pill.
Study Design
Arms: Triple pill — telmisartan 20mg + amlodipine 2.5mg + indapamide 1.25mg (n=833) vs Matching placebo (n=837)
Patients per Arm: Triple pill: 833; Placebo: 837
Outcome
• Mean follow-up SBP: 127 vs 138 mm Hg (difference 9 mm Hg, 95% CI 7–10)
• Major cardiovascular events: 6.6% vs 9.8%, P=0.04
• Discontinuation for adverse events: 13.6% vs 6.0% (most often ≥20% rise in serum creatinine)
Bottom Line
In patients with a history of intracerebral hemorrhage and SBP 130–160 mm Hg, adding a once-daily low-dose triple pill (telmisartan 20 mg + amlodipine 2.5 mg + indapamide 1.25 mg) to standard care lowered SBP by ~9 mm Hg and reduced recurrent stroke by ~39% (HR 0.61) and major cardiovascular events over 2.5 years, with an increase in early treatment discontinuation, most commonly due to a ≥20% rise in serum creatinine.
Major Points
- Triple pill reduced recurrent stroke from 7.4% to 4.6% over a median 2.5-year follow-up (HR 0.61; 95% CI 0.41–0.92; P=0.02).
- Mean follow-up SBP was 127 mm Hg with the triple pill vs 138 mm Hg with placebo (between-group difference 9 mm Hg, 95% CI 7–10).
- Major cardiovascular events (nonfatal MI, nonfatal stroke, or CV death) were lower with the triple pill: 6.6% vs 9.8% (P=0.04).
- Serious adverse events were similar (23.2% vs 26.0%), but early discontinuation due to adverse events was more frequent with the triple pill (13.6% vs 6.0%), driven mainly by a ≥20% rise in serum creatinine.
- BP differences attenuated over time as concomitant antihypertensive use increased in the placebo group.
- Trial population was predominantly Asian (72.6%), with two-thirds residing in Sri Lanka.
Study Design
- Study Type
- Multinational, double-blind, randomized, placebo-controlled trial
- Randomization
- Yes
- Blinding
- Double-blind (with single-blind active run-in phase)
- Sample Size
- 1670
- Follow-up
- Median 2.5 years (IQR 1.6–4.4); planned up to 72 months
- Centers
- 61
- Countries
- Australia, Brazil, Georgia, Malaysia, Netherlands, Nigeria, Singapore, Sri Lanka, Switzerland, Taiwan, United Kingdom, Vietnam
Primary Outcome
Definition: First recurrent stroke (time-to-event analysis)
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 62/837 (7.4%) | 38/833 (4.6%) | 0.61 (0.41 to 0.92) | 0.02 |
Limitations & Criticisms
- Sample size was reduced from 3782 to 1500 in 2020 due to recruitment challenges, limiting power for some analyses.
- Population was predominantly Asian (72.6%), with two-thirds residing in Sri Lanka, which may limit generalizability to other populations.
- Active single-blind run-in phase excluded patients who were intolerant or non-adherent (24.3% did not proceed to randomization), potentially enriching for tolerant patients and underestimating real-world discontinuation rates.
- Higher early discontinuation in the triple-pill group (13.6% vs 6.0%) driven by serum creatinine rises raises concerns about renal safety in broader use.
- Between-group BP differences attenuated over time as concomitant antihypertensives were added in the placebo group, which may have diluted the effect estimate.
- Modest violation of the proportional-hazards assumption noted, though sensitivity analyses were consistent.
- Subgroup analyses by race/ethnicity not feasible due to small numbers.
Citation
N Engl J Med 2026;394:1571-82