Periprosthetic hip fracture: early surgery showed no clear mortality advantage

Design

Target trial emulation using a national hospital administrative database

Population

4,564 patients with periprosthetic hip fracture, mean age 84 years

Primary outcome

30- and 90-day mortality, early (within 2 days) vs delayed (3 to 10 days) surgery

Effect

90-day mortality 2.35% vs 2.31%, RR 1.01 (95% CI 0.59 to 1.68); 30-day RR 2.06 (0.64 to 3.73)

Researchers used a Japanese hospital administrative database from 2014 to 2024 and a clone-censor-weight target trial emulation to compare surgery within two days of admission with surgery at three to ten days. They analysed 4,564 patients with periprosthetic hip fracture, mean age 84 years, 80% female.

The approach corrects for immortal time bias, which can make delayed surgery look safer because patients must survive to receive it. At 30 days, mortality was 1.72% with early and 0.83% with delayed strategy (risk difference 0.89%, 95% CI -0.32 to 1.72), with wide uncertainty. At 90 days it was 2.35% against 2.31% (risk ratio 1.01, 95% CI 0.59 to 1.68).

The authors do not support assuming the 48-hour standard for native hip fractures applies to these fractures for mortality alone. They add that findings do not justify avoidable delay, since complications, reoperation and function were not studied. Use it to support proper medical optimisation and planning, not to defer surgery by default.

• Surgery within two days did not show a clear 90-day survival benefit over three to ten days.
• The 30-day estimates were imprecise, so a small early effect cannot be excluded.
• Use the time needed for medical optimisation and operative planning, without drifting into avoidable delay.
• Complications and functional recovery were not studied; do not assume equal outcomes.

Why it matters

It questions whether a rule built for native hip fractures should be copied across to fractures around implants.

Don't overread it

Observational data with unverifiable assumptions; only mortality was assessed, and a guideline-level standard has not changed.

The statistics, in plain English

A confidence interval that includes 1.0, such as 0.59 to 1.68 at 90 days, means the data are compatible with a benefit, no difference or a harm. Target trial emulation reduces some biases of observational work, but cannot remove confounding from unmeasured factors such as fracture complexity.

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.