Preoperative geriatric nutritional risk index predicts prognosis and postoperative complications in elderly patients undergoing non-cardiac surgery: a retrospective cohort study

Table 2 Cox proportional hazards regression analysis of the association between at-risk GNRI and overall survival

Analysis method	HR	95% CI	P value
Cox proportional hazards model regression analysis (n = 28,762)
Model 1 (univariable analysis)	2.752	2.496–3.034	< 0.001
Model 2 (preoperative patient- related covariates adjusted)	1.842	1.647–2.061	< 0.001
Model 3 (surgery-related covariates adjusted)	2.132	1.928–2.359	< 0.001
Model 4 (postoperative patient- related covariates adjusted)	2.468	2.235–2.726	< 0.001
Model 5 (fully adjusted)	1.682	1.502–1.882	< 0.001
Propensity score analysis (multivariable analysis)
PSM (n = 14,126)	1.564	1.372–1.784	< 0.001

HR, hazard ratio; CI, confidence interval; PSM, propensity score matching
Model 1 was a univariate crude model
Model 2 included age, sex, ASA classification, hypertension, diabetes mellitus, coronary heart disease, chronic heart failure, renal dysfunction, malignant tumor, last creatinine level, hemoglobin, platelet, total bilirubin
Model 3 included emergency operation, surgical procedures, duration of procedures, estimated blood loss, blood transfusion, NSAIDS, glucocorticoid, volatile anesthetic, opioid dose, dexmedetomidine
Model 4 included heart injury, AKI, stroke, pulmonary infection, surgical site infection, 30-day readmission, 30-day reoperation, ICU admission, postoperative length of stay
Model 5 included all the confounders
PSM 7063 pairs were matched by propensity score

ISSN: 1471-2318