Associations of preoperative nutritional and inflammatory markers with length of stay after primary total knee arthroplasty

Hospital length of stay (LOS) after primary total knee arthroplasty (TKA) is influenced by both perioperative pathways and patient-level factors. Although inflammatory markers have been linked to complication-related outcomes, the association of preoperative nutritional and inflammatory markers with routine postoperative LOS remains unclear. This study examined the associations of the prognostic nutritional index (PNI), high-sensitivity C-reactive protein (hs-CRP), and the C-reactive protein-to-albumin ratio (CAR) with LOS after primary TKA.

This retrospective cohort study included consecutive patients who underwent primary TKA for osteoarthritis at a single institution in 2023. The primary outcome was index LOS, defined as days from admission to discharge. Prolonged LOS (> 8 days) was the secondary outcome. Associations were evaluated using sequential linear and logistic regression models adjusted for age, sex, body mass index, diabetes mellitus, and operative time. Time to first ambulation was assessed separately in a sensitivity analysis because it is a postoperative variable that may lie on the early recovery pathway.

Among 464 patients, higher preoperative PNI was associated with shorter LOS in the primary adjusted model (0.42-day decrease per 5-point increase; 95% CI − 0.60 to − 0.24; P < 0.001) and with lower odds of prolonged hospitalization (OR = 0.69; 95% CI 0.55 to 0.87; P = 0.002). Higher hs-CRP showed a modest association with longer LOS (β = 0.17; 95% CI 0.01 to 0.33; P = 0.042), but was not significantly associated with prolonged LOS after adjustment. CAR was associated with both outcomes, although less consistently than PNI. In sensitivity analyses additionally adjusting for time to first ambulation, the association between PNI and LOS reversed direction.

Higher preoperative PNI was associated with shorter hospitalization after primary TKA, whereas hs-CRP and CAR showed weaker or less consistent associations. The reversal observed after additional adjustment for time to first ambulation indicates that models including postoperative recovery variables should be interpreted cautiously.