Risk Factors for Complications and 90-Day Mortality After Percutaneous Endoscopic Gastrostomy: The Role of Nutritional and Inflammatory Markers.

Background and Objectives: Percutaneous endoscopic gastrostomy (PEG) is a widely accepted method for long-term enteral nutrition, but procedure-related complications and early mortality remain major concerns. Nutritional and inflammatory indices such as serum albumin, C-reactive protein (CRP), Prognostic Nutritional Index (PNI), and Nutrition Risk Screening (NRS-2002) may provide prognostic value, yet comparative data in PEG cohorts are limited. This study aimed to identify predictors of complications and 90-day mortality after PEG and to compare the prognostic performance of nutritional indices. Materials and Methods: A retrospective cohort of 122 consecutive adult patients undergoing PEG between January and December 2024 was analyzed. Demographic, clinical, and laboratory parameters were collected, including albumin, CRP, PNI, and NRS-2002. Complications were categorized as early (≤30 days) or late (>30 days), and all-cause mortality was assessed at 30 and 90 days. Univariate and multivariate logistic regression models were used to evaluate predictors of complications and 90-day mortality. To address multicollinearity, albumin, PNI, and NRS-2002 were separately tested in adjusted models, with model performance assessed by AIC, BIC, Nagelkerke R2, and C-index. Results: Early complications occurred in 4.9% and late complications in 8.2% of patients, for a total complication rate of 13.1%. Thirty-day mortality was 4.1%, 90-day mortality 17.2%, and total in-hospital mortality during the study year 30.3%. Neuromuscular indication was independently associated with increased risk of complications (aOR 5.0, 95% CI 1.2-20.0, p = 0.028) but reduced 90-day mortality (aOR 0.15, 95% CI 0.03-0.80, p = 0.025). Lower baseline albumin independently predicted higher 90-day mortality (aOR 0.92, 95% CI 0.86-0.99, p = 0.034). Elevated CRP demonstrated a borderline association with mortality (p = 0.051), while NRS-2002 ≥5 and Δ-PNI showed borderline trends toward increased mortality risk. In model comparison, none of the nutritional indices achieved independent statistical significance, but all demonstrated similar performance (AIC = 114, C-index 0.72-0.74). Conclusions: PEG outcomes are strongly influenced by baseline indication and nutritional-inflammatory status. Neuromuscular patients and patients with dysphagia face higher complication risk but lower short-term mortality, while hypoalbuminemia, elevated CRP, and high NRS-2002 or declining PNI identify patients at greater risk of death. Systematic integration of albumin, CRP, PNI, and NRS-2002 may improve risk stratification and management in PEG candidates.