A five-item frailty index predicts 90-day mortality after pancreatoduodenectomy as well as the eleven-item one

Design

Retrospective cohort, single centre, 2015 to 2023

Population

1,926 adults undergoing pancreatoduodenectomy for periampullary lesions

Primary outcome

Major complications (Clavien-Dindo grade III or above) and 90-day mortality

Effect

Severe frailty and mortality: OR 4.1 (95% CI 1.6 to 10.3) mFI-11; OR 2.8 (95% CI 1.4 to 5.6) mFI-5; AUC 0.761 vs 0.757

This retrospective cohort covered 1,926 adults who had pancreatoduodenectomy between 2015 and 2023. Major complications (Clavien-Dindo grade III or above) occurred in 18.2% and 90-day mortality was 2.7% (52 patients).

Severe frailty was independently associated with 90-day mortality on both scores: odds ratio 4.1 (95% CI 1.6 to 10.3) with the 11-item index and 2.8 (95% CI 1.4 to 5.6) with the 5-item index. Discrimination for mortality was good and almost identical (AUC 0.761 vs 0.757). For major complications it was modest for both (about 0.63). The association was clearest for deaths adjudicated as pancreas-specific failure to rescue.

For a busy unit, the five-item score takes less time to collect and loses little. It identifies risk rather than deciding whom to operate on.

• Record the five-item modified Frailty Index before pancreatoduodenectomy.
• Expect severe frailty to carry a higher risk of death, not necessarily of major complications.
• Use the score to prepare the patient and ward for rescue, not as a stand-alone reason to decline surgery.
• Do not read a modest AUC for complications as reassurance for non-frail patients.
• Plan early recognition and escalation of pancreatic fistula and bleeding in frail patients.

Why it matters

Frailty is more tightly linked to dying after complications than to having them, which changes where to put effort.

Don't overread it

A retrospective analysis from one centre shows risk prediction, not that frailty-guided selection or prehabilitation improves outcomes.

The statistics, in plain English

An AUC of 0.76 means that, given one patient who died and one who did not, the score ranks them correctly about three times in four. That is useful for counselling but not accurate enough to decide an individual patient's operation. The wide intervals on the odds ratios reflect only 52 deaths.

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.