Development of a Prediction Model for Postoperative Hypoproteinemia in Patients Undergoing Microsurgical Reconstructive Surgery: A Prospective Observational Study.

Convenience sample from one orthopaedic department in Shanghai, so findings may not apply to other hospitals or patient groups. Eight of 200 recruited patients were excluded for incomplete records; the analysed sample was 192. The prediction model was developed on the whole cohort and only internally validated, with no separate derivation and validation sets or external validation. Risk was assessed only for the first three postoperative days, so longer-term hypoproteinemia was not measured. Possible confounders such as nutritional status, inflammatory markers and perioperative medications were not measured. The outcome was a laboratory definition of hypoproteinemia, not clinical complications such as infection, wound healing, length of stay or mortality. The study did not test whether using the model or any intervention prevents hypoproteinemia or improves outcomes.

The study was funded by the Innovative Research Team of High-level Local Universities in Shanghai and the Nursing Discipline Construction Project, Shanghai Jiao Tong University School of Medicine. The authors declared no conflicts of interest.

Do not conclude that albumin infusion or another treatment will prevent postoperative hypoproteinemia. The study identified predictors of a low albumin level in the first three postoperative days and validated the model only internally in 192 patients from one hospital; it did not test treatment effects or patient outcomes.