For the roughly one million total hip arthroplasties performed globally each year, even a single extra hospital day compounds costs, infection risk, and functional decline. A new predictive tool that flags which patients are headed for prolonged recovery before they even enter the operating room could meaningfully shift how surgical teams triage perioperative support — and nutritional status appears to be a central, modifiable lever.
In a cohort of 805 total hip arthroplasty patients, researchers evaluated three established nutritional screening instruments — the Geriatric Nutritional Risk Index (GNRI), the Prognostic Nutritional Index (PNI), and the Controlling Nutritional Status score (CONUT) — as predictors of extended length of stay, defined as ten or more days post-surgery. Patients who exceeded this threshold had significantly lower GNRI and PNI scores than those discharged earlier, while CONUT showed no meaningful difference between groups. On receiver operating characteristic analysis, GNRI outperformed PNI with an area under the curve of 0.643, and combining multiple nutritional indices failed to improve predictive accuracy. Multivariate logistic regression identified GNRI, the HuaXi fall risk score, and the age-adjusted Charlson Comorbidity Index as independent predictors of prolonged hospitalization, which were then integrated into a clinical nomogram for individualized risk estimation.
This work adds useful granularity to a growing evidence base linking perioperative nutritional status to surgical outcomes, but several important caveats temper its immediate clinical application. An AUC of 0.643 for GNRI is statistically significant yet diagnostically modest — roughly one-third of the predictive variance remains unexplained, meaning nutritional indices alone are insufficient triage tools. The single-center design limits generalizability, particularly across populations with different baseline comorbidity burdens or rehabilitation infrastructures. The ten-day LOS cutoff also reflects Chinese hospital discharge norms and may not translate directly to Western settings where the median THA stay is two to three days. Still, the integration of nutritional risk into a multivariate nomogram rather than treating it as an isolated biomarker is methodologically sound. For health systems managing aging populations with high arthroplasty volumes, routine preoperative GNRI screening is low-cost, reproducible, and increasingly justifiable.
