Oral treatment for sleep apnoea, AI-assisted echocardiography, pacing strategies and cardiovascular prevention: a clinical reading of eight selected studies presented at AHA Scientific Sessions 2026.
What does this roundup cover?
The American Heart Association Scientific Sessions brought together studies with very different implications for everyday care. This selection covers eight reports, in the same order as our accompanying carousel. It is a curated overview, rather than a complete congress review or a new treatment guideline.
Six studies have journal publications. The AI echocardiography and REDUCE findings below come from official AHA conference reports and remain preliminary. For the NEJM studies, the available publisher abstracts were checked; the complete supplied OPTIMAL-STROKE main paper was also reviewed. This distinction matters when considering details that a short report cannot resolve.
ATTAIN-OSA: Can an oral treatment improve sleep-apnoea severity?
Two phase 3, double-blind trials randomised 712 adults with moderate-to-severe obstructive sleep apnoea (OSA), excess weight and no diabetes. Separate trials studied participants not receiving positive airway pressure (PAP) and those receiving PAP. Daily oral orforglipron was compared with placebo for 52 weeks.
The primary endpoint was change in the apnoea–hypopnoea index (AHI). Placebo-adjusted differences were −13.2 events/hour without PAP (95% CI −17.7 to −8.7) and −12.0 events/hour with PAP (95% CI −17.1 to −6.9). Gastrointestinal adverse events were more frequent with orforglipron.
These are sleep-disordered-breathing findings. The trials do not establish cardiovascular event reduction or justify stopping PAP without reassessment. Original JAMA article .
AI echocardiography: What did the time saving actually represent?
A randomised comparison involving 1,200 echocardiograms assessed AI-generated versus human preliminary interpretations, followed by blinded cardiologist review. Substantial changes were required in 50.6% versus 49.7%, respectively; the official report states that the study's noninferiority criterion was met.
Cardiologists spent 4.6 versus 8.5 minutes reviewing and finalising the AI-assisted and comparison reports. Every report was finalised by a cardiologist. This is evidence about a supervised workflow, rather than autonomous reporting or improved patient outcomes.
Evidence status: conference-only and preliminary. The full peer-reviewed manuscript and detailed noninferiority analyses were not verified for this roundup. Official AHA report, abstract 28202 .
OptimPacing: Which outcome improved with left bundle-branch pacing?
OptimPacing randomised 702 patients with atrioventricular block to left bundle-branch or right-ventricular pacing. The modified intention-to-treat analysis included 685 patients. The primary outcome combined death from any cause, heart-failure hospitalisation and pacing-induced cardiomyopathy.
At median follow-up of 35.1 months, estimated four-year cumulative incidence was 8.5% versus 19.2% (HR 0.41, 95% CI 0.23–0.75). These are Kaplan–Meier estimates at four years, rather than crude event percentages during the median follow-up.
The trial stopped early following its prespecified interim analysis. The result favours left bundle-branch pacing for the composite; it should not be presented as a separately established mortality benefit. Original NEJM article .
OPTIMAL-STROKE: How should cardiovascular benefit and renal events be balanced?
Among 4,368 clinically stable participants with previous ischaemic stroke or transient ischaemic attack (TIA), systolic blood-pressure targets below 120 versus 140 mmHg were compared. Over median 3.9 years, the cardiovascular composite occurred in 10.9% versus 13.5% (HR 0.79, 95% CI 0.67–0.94).
The composite included cardiovascular death, nonfatal myocardial infarction or stroke, and hospitalisation for unstable angina or heart failure. Renal composite events occurred in 3.9% versus 2.2%, mainly sustained estimated glomerular filtration rate reductions. Formal secondary testing stopped at cardiovascular death; later renal and recurrent-stroke results require cautious interpretation.
This was secondary prevention in stable patients, not an acute-stroke BP protocol. The smaller TIA subgroup was inconclusive. Original NEJM article and our full OPTIMAL-STROKE review .
How do the remaining cardiovascular-outcome studies compare?
| Study and population | Main finding | Interpretation |
|---|---|---|
| Lp(a)HORIZON: established cardiovascular disease; lipoprotein(a) ≥70 mg/dL | Pelacarsen versus placebo: events 12.3% vs 12.1%, HR 1.02 (95% CI 0.90–1.15), median 4.7 years | Lowering the biomarker did not demonstrate a reduction in the primary clinical endpoint with this treatment in this population. |
| ISCHEMIA-EXTEND: stable coronary disease | Initial invasive versus conservative strategy: adjusted mortality HR 0.97 (95% CI 0.86–1.09), median follow-up 9 years | No statistically significant overall survival advantage was demonstrated; symptom relief and acute coronary syndromes are separate questions. |
| OPTIMAL-DIABETES: type 2 diabetes and high cardiovascular risk | SBP targets <120 vs <140 mmHg: events 6.0% vs 6.7%, HR 0.90 (95% CI 0.77–1.06), median 4 years | A nonsignificant superiority result does not prove equivalence between targets. |
Lp(a)HORIZON: Why is the biomarker–outcome distinction important?
The placebo-controlled trial randomised 8,350 patients, with 8,322 included in efficacy analysis, to monthly subcutaneous pelacarsen 80 mg or placebo. The primary endpoint combined cardiovascular death, nonfatal myocardial infarction, nonfatal stroke and urgent coronary revascularisation.
Despite substantial lipoprotein(a) lowering, clinical benefit was not demonstrated. This finding should remain specific to pelacarsen, its tested regimen and this trial population; it is not proof that every approach to lowering lipoprotein(a) is ineffective. Original NEJM article .
ISCHEMIA-EXTEND: Does this remove the role of revascularisation?
The original trial included 5,179 participants. With extended follow-up, estimated 11-year all-cause mortality was 25.8% with an initial invasive strategy versus 26.9% with an initial conservative strategy. The adjusted confidence interval crossed the null.
The analysis concerns the initial management strategy in stable coronary disease. It does not address acute coronary syndromes or negate revascularisation for symptom relief. Neither arm should be described as an absence of medical treatment. Original JAMA article .
OPTIMAL-DIABETES: Why should it be read separately from OPTIMAL-STROKE?
This Brazilian open-label trial enrolled 9,476 adults. After excluding the first three months, achieved mean systolic BP was 121.8 versus 135.4 mmHg. The primary composite was cardiovascular death, nonfatal myocardial infarction or stroke, or hospitalisation for unstable angina or heart failure; P=0.20 for the primary comparison.
The confidence interval includes both potential benefit and little effect. Different trial results do not, by themselves, establish that diabetes changes the treatment effect. Original NEJM article .
REDUCE: Does better BP control establish recurrent-stroke prevention?
REDUCE randomised 160 people with uncontrolled hypertension after ischaemic stroke or intracerebral haemorrhage to added spironolactone or usual care. At three months, the between-group home systolic BP difference was −7.7 mmHg (95% CI −11.6 to −3.8). BP below 130/80 mmHg was reached in 52% versus 30%.
The open-label study enrolled fewer than planned, and about 15% lacked three-month BP measurements. It was not designed to establish fewer recurrent strokes or heart attacks. Potassium and renal function require attention when interpreting a spironolactone strategy.
Evidence status: conference-only and preliminary. Official AHA research report, abstract 21380 and AHA Scientific Sessions numerical report .
Which distinctions matter most in practice?
| Headline temptation | More accurate reading |
|---|---|
| “No significant difference means the treatments are the same.” | A superiority trial with a confidence interval crossing the null does not establish equivalence. |
| “A composite improved, so mortality improved.” | Check the components; a composite result does not establish a separate benefit for each one. |
| “A faster report is a better clinical outcome.” | Workflow efficiency, diagnostic performance and patient outcomes are different endpoints. |
| “A congress result is a new guideline.” | Consider evidence status, population, safety and the wider evidence before changing a treatment strategy. |
Clinical takeaway
These eight reports refine different clinical questions. Their value lies in matching the result to the patient, comparator, endpoint and timeframe actually studied. Keep preliminary reports distinct from journal publications, retain the uncertainty around nonsignificant results, and weigh benefits alongside safety signals rather than treating a congress headline as a universal instruction.
Original sources
- Malhotra A, Gottlieb DJ, Woehrle H, et al. Orforglipron for Obstructive Sleep Apnea: Results From 2 Randomized Clinical Trials (ATTAIN-OSA). JAMA. Published online 9 November 2026. doi:10.1001/jama.2026.24080 .
- American Heart Association. AI-interpreted echocardiograms comparable to human reports, cardiologists' time cut by about 50%. Scientific Sessions 2026, abstract 28202. 9 November 2026. Official report .
- Qian Z, Jiang L, Hou X, et al. Left Bundle-Branch Pacing versus Right Ventricular Pacing for Atrioventricular Block. N Engl J Med. Published online 8 November 2026. doi:10.1056/NEJMoa2612972 .
- Sampaio Silva G, Machline-Carrion MJ, Santo K, et al. Intensive Blood-Pressure Control in Patients with Previous Ischemic Stroke. N Engl J Med. Published online 7 November 2026. doi:10.1056/NEJMoa2612924 .
- Cho L, Nicholls SJ, Frndak S, et al. Pelacarsen and Cardiovascular Outcomes in Patients with High Lipoprotein(a). N Engl J Med. Published online 7 November 2026. doi:10.1056/NEJMoa2610159 .
- Hochman JS, Anthopolos R, Reynolds HR, et al. Long-Term Survival After Initial Invasive or Conservative Management of Stable Coronary Disease in ISCHEMIA-EXTEND. JAMA. Published online 7 November 2026. doi:10.1001/jama.2026.20375 .
- Santo K, Martins E, Silva GS, et al. Intensive Blood-Pressure Lowering in Patients with Type 2 Diabetes. N Engl J Med. Published online 7 November 2026. doi:10.1056/NEJMoa2612922 .
- American Heart Association. Can a low-cost diuretic medication help to lower blood pressure after a stroke? Scientific Sessions 2026, abstract 21380. 7 November 2026. Official report .