Incidence, risk factors, prevention and prediction of postoperative pneumonia after hip fracture surgery in older adults: a systematic review and meta-analysis

Postoperative pneumonia (POP) is a frequent and serious complication of hip fracture surgery in older adults. We synthesised evidence on its incidence, risk factors, prevention and prediction. PubMed, Web of Science, SinoMed and CNKI were searched for studies published from 1 January 2021 to 1 August 2026 (date of the final search), supplemented by citation searching (PROSPERO CRD420261513572). Incidence was pooled with random-effects models on the logit scale. Adjusted risk factors were pooled only when at least two independent cohorts used the same exposure definition, and overlapping cohorts contributed one estimate each. Prediction models were synthesised without pooling areas under the curve. Risk of bias was assessed with QUIPS, PROBAST and ROBINS-I, and certainty of evidence with GRADE. We included 23 independent cohorts reported in 35 articles. The pooled incidence was 8.7% (95% CI 7.5–10.2%) in 20 cohorts (24,574 patients), with substantial heterogeneity (I² = 93%; 95% prediction interval 4.1–17.8%). Incidence was higher in cohorts at high than at moderate risk of bias (11.6% vs. 7.8%; P = 0.02); no other moderator explained the heterogeneity. COPD (OR 7.70, 95% CI 2.61–22.71), hypoalbuminaemia below 35 g/L (OR 7.01, 95% CI 4.37–11.25), ASA physical status ≥ 3 (OR 2.50, 95% CI 1.44–4.34), history of stroke (OR 1.92, 95% CI 1.34–2.76), postoperative ICU admission (OR 2.79, 95% CI 2.13–3.66), older age (OR 1.041 per year) and fracture-to-surgery delay (OR 1.091 per day) were associated with POP; certainty was moderate for age and delay and low for the other factors. Perioperative prevention protocols were associated with fewer pneumonias in two before–after studies (risk ratio 0.60, 95% CI 0.44–0.82; very low certainty). All 12 prediction-model studies were at high risk of bias, and only 3 models had been externally validated. About one in eleven older adults develops pneumonia after hip fracture surgery, with wide variation between settings. COPD, hypoalbuminaemia, ASA physical status ≥ 3, prior stroke and surgical delay can help target intensified prevention, but the evidence is observational and mostly of low certainty. Existing prediction models are not ready for routine use, and prevention bundles require controlled evaluation. We have completed the registration in the International Prospective Register of Systematic Reviews (PROSPERO), and the assigned registration number is CRD420261513572.