BackgroundThis retrospective cohort study investigated perioperative C-reactive protein (CRP) value for early detection of postoperative systemic infective complications in elderly patients with intertrochanteric femoral fracture. MethodsA total of 250 patients aged ≥65 years, with intertrochanteric femoral fractures that were surgically treated between January 2011 and December 2015 were included. CRP value was measured preoperatively and on postoperative day (POD) 3, 5, and 10, and analyzed with regard to postoperative systemic infections, delirium, and death within 1 year. The patients were divided into two groups according to postoperative systemic infection, and perioperative CRP responses between the two groups were comparedusing t-test (or Man-Whitney test where appropriate). The receiver operator characteristic (ROC) curves of CRP value were used to determine the best-supported cut-off values for identifying postoperative systemic infections. Systemic infections were reclassified as pulmonary and extra-pulmonary infections. ResultsThe mean CRP values preoperatively and on POD 3, 5, and 10 were 2.82, 10.10, 3.74, and 1.89 mg/dL, respectively. Postoperative systemic infections, delirium, and death within 1 year were noted in 35 (14.0%), 30 (12.0%), and 45 (18.0%) patients, respectively. The CRP value in patients with postoperative systemic infections significantly elevated on POD 5 and 10 (p < 0.001, p < 0.001), and cut-off values were 4.71 and 1.59 mg/dL on POD 5 and 10, respectively. Postoperative delirium and death within 1 year were observed more often in the group with postoperative systemic infections (p = 0.003, p = 0.014). Although preoperatively elevated CRP values did not influence the postoperative CRP responses, they were significantly associated with delirium (p = 0.015). ConclusionsThe CRP value on POD 5 is the earliest predictor of postoperative systemic infections in elderly patients with intertrochanteric femoral fractures that are managed surgically. Moreover, when the CRP value on POD 5 is >4.71 mg/dL, the possibility of postoperative systemic infections should be considered. Level of evidenceIII, Retrospective cohort study.