Does an Independent Predictor of In-Hospital Mortality Exist for Patients with Isolated Proximal Femoral Fracture? A Retrospective Two-Year Observational Study

Abstract

PURPOSE OF THE STUDYThe primary objective of the study was to find out in-hospital mortality in patients undergoing surgery for proximal femoral fracture. The secondary objective was to identify independent predictors of in-hospital mortality. MATERIAL AND METHODSA retrospective single-centre observational study PROXIMORT of patients operated on for isolated proximal femoral fracture at the University Hospital (FN) Brno in the years 2011 and 2012. The 30-day and overall one-year mortality in the study group and the impact of observed parameters on mortality were also assessed. The observed parameters were: patient age and sex, ASA score, time from injury to surgery (hr), daily (7-20 hr) or night (20-7 hr) time of surgery, type of anaesthesia (general vs spinal), initial haemoglobin and haematocrit levels, intra-operative administration of blood products and vasopressors, and erudition of the anaesthesiologist and surgeon.To evaluate the relationship of in-hospital mortality to the observed characteristics, we used univariate logistic regression modelling and odds ratio, using SPSS 22 software (IBM, USA). RESULTSData were obtained from 414 patients and 369 patients were included (male, n = 91; female, n = 278). Due to exclusion criteria, 45 patients were excluded (not an isolated injury). In-hospital mortality was 6.5% (n = 24), 30-day mortality was 8.4% (n = 31) and total mortality of the study group was 35.8% (n = 132). Statistically significant effects on in-hospital mortality included: older age of the patient (p = 0.013), ASA score of 3 or more (p = 0.002) and general anaesthesia administration (p = 0.043). For 30-day mortality, this was older age (p = 0.012), ASA score of 3 and more (p < 0.001), general anaesthesia administration (p < 0.001), lower weight (p = 0.028), lower BMI (p = 0.006) and intra-operative administration of vasopressors (p = 0.023). The influence of other observed parameters on post-operative mortality was not statistically significant. DISCUSSIONIn-hospital mortality in the PROXIMORT study was 6.5% (95% confidence interval (CI) 4.2 to 9.5%), which was significantly higher than in-hospital mortality in unselected surgically treated patients in the Czech Republic, as reported in the EuSOS study (2.3% with 95% CI 0.9 to 3.7%). Administration of general anaesthesia was determined as an independent predictor of in-hospital and 30-day mortality, which was concordant with the results of meta-analysis published by Rodgers et al. and Barbosa et al in 2013. Postponing surgery for perioperative optimisation had no effect on mortality according to the PROXIMORT study. Patorn et al. have supported this conclusion by the results of a selected group of patients with surgery delayed for more than 24 hours; the patients mortality, regardless of anaesthesia, was up to 2.5%. CONCLUSIONSThe PROXIMORT study identified the higher patient age, ASA score of 3 and more and general anaesthesia administration as independent predictors of in-hospital mortality.