Dapagliflozin after TAVI was associated with benefit in frail patients but not in robust ones

Design

Post hoc analysis of a randomised trial (DapaTAVI)

Population

1,222 patients after TAVI with a baseline frailty assessment

Primary outcome

All-cause death or worsening heart failure at 1 year

Effect

Frail HR 0.53 (0.36 to 0.77); prefrail HR 1.01 (0.66 to 1.55); robust HR 1.26 (0.42 to 3.75); interaction P = 0.021

DapaTAVI randomised patients after transcatheter aortic valve implantation to dapagliflozin or standard care. This analysis split 1,222 of them by frailty using the FRAIL scale: 47% were frail, 40% prefrail and 12% robust.

Frailty predicted worse outcomes. Dapagliflozin was associated with about half the risk of death or worsening heart failure in frail patients, with no apparent benefit in the other groups. The difference between groups was statistically significant.

Subgroup results are often not reproduced, and the robust group was small. The safer reading is that frailty is not a reason to hold the drug back, rather than that only frail patients should receive it.

• Nearly half of patients undergoing TAVI in the trial were frail on a five-item scale.
• Frail and prefrail patients had about twice the risk of death or worsening heart failure at one year.
• Dapagliflozin was associated with fewer events in frail patients (HR 0.53) but not in prefrail or robust ones.
• Safety did not differ by frailty status.
• Check volume status and kidney function before starting an SGLT2 inhibitor in a frail older patient.

Why it matters

Frail older patients are the ones clinicians are most tempted to spare an extra drug, and here they seemed to gain the most.

Don't overread it

This is a post hoc subgroup analysis; it suggests where the benefit concentrates but cannot prove it.

The statistics, in plain English

An HR of 0.53 means about half the event rate. The robust group's interval runs from 0.42 to 3.75, so it is simply too small to tell. A significant interaction test makes the subgroup difference less likely to be chance, but post hoc analyses still need confirming.

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.