NT-proBNP screening found heart failure in 1 in 4 people with diabetes

Design

Prospective, multicentre, unblinded randomised controlled trial (TARTAN-HF)

Population

706 people aged 40 or over with diabetes and at least one other heart failure risk factor, no known heart failure, west of Scotland

Primary outcome

Heart failure diagnosis at 6 months

Effect

24.6% vs 1% (OR 58, 95% CI 14 to 236); number needed to screen 5

TARTAN-HF randomised 706 community-based people aged 40 or over in the west of Scotland with type 1 or type 2 diabetes, at least one further risk factor and no known heart failure. One group had NT-proBNP screening, with echocardiography if the level was 125 pg/mL or more, and the other usual care. The report was published on 19 August 2026.

At six months heart failure was diagnosed in 87 of 354 (24.6%) in the screened group against 2 of 352 (1%) with usual care (OR 58, 95% CI 14 to 236). Most cases had preserved ejection fraction. The number needed to screen was 5. SGLT2 inhibitor use rose from 24% to 39% in the screened group, and those with unrecognised heart failure had much worse health status.

The trial measured diagnoses, not outcomes. It does not show that screening reduces admissions or deaths, which is why it argues for further study and consideration, not mandatory screening.

• Heart failure was diagnosed in 24.6% of the screened group vs 1% with usual care at 6 months.
• Most diagnoses were heart failure with preserved ejection fraction.
• SGLT2 inhibitor use rose from 24% to 39% in the screened group.
• Consider NT-proBNP in people with diabetes plus additional risk factors.
• Screening for outcome benefit remains unproven, so do not make it mandatory.

Why it matters

Heart failure in diabetes is common and often unrecognised, and the question of screening is open in guidelines.

Don't overread it

The primary outcome was diagnosis at six months, and the trial was unblinded and cannot show a benefit in hospitalisation or death.

The statistics, in plain English

An odds ratio of 58 is huge only because the comparison group was almost never tested, so it mostly shows that looking finds disease. The number needed to screen of 5 is the more useful figure. Neither says anything about whether finding the disease earlier changes survival.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.