Procalcitonin and C-reactive protein for bacterial infection diagnosis in elderly patients after traumatic orthopedic surgery.

Background: Biomarkers prove valuable for diagnosing postoperative bacterial infection, but data in elderly patients are scarce. Here we analyze how procalcitonin and C-reactive protein (CRP) perform for bacterial infection diagnosis after traumatic orthopedic surgery in elderly patients. Methods: We included all patients admitted to our perioperative geriatrics unit after traumatic orthopedic surgery. Patients on antibiotics, presenting preoperative bacterial infection, or without procalcitonin measurement were excluded. Clinical and biological data were collected prospectively. Medical charts were reviewed by three experts blinded to biomarker results to assess bacterial infection diagnosis. Areas under the curve and 90%-specificity thresholds were analyzed for baseline procalcitonin and CRP levels and relative variations. Results: Analysis included 229 patients (median age 86 years, hip fracture 83%), of which 40 had bacterial infection (pneumonia [n = 23], urinary tract infection [n = 8]; median delay to onset: 2 days post-admission). For bacterial infection diagnosis, the computed areas under the curve were not significantly different (procalcitonin-baseline 0.64 [95% confidence interval: 0.57-0.70]; procalcitonin-relative variation 0.65 [0.59-0.71]; CRP-baseline 0.68 [0.61-0.74]; CRP-relative variation 0.70 [0.64-0.76]). The 90%-specificity thresholds were 0.75 mu g/L for procalcitonin-baseline, +62% for procalcitonin-variation, 222 mg/L for CRP-baseline, +111% for CRP-variation. Conclusions: Diagnostic performances of procalcitonin and CRP were not significantly different. Baseline levels and relative variations of these biomarkers showed little diagnostic value after traumatic orthopedic surgery in elderly patients.