RESUMO
PURPOSE: The goal of this systematic review and meta-analysis was to identify the main risk factors for periprosthetic joint infection (PJI) in patients undergoing total hip or knee arthroplasties. METHODS: A systematic review was conducted of the potential risk factors for PJI in total hip or total knee arthroplasty. Risk factors were compared and grouped according to demographics, comorbidities, behavior, infections, native joint diseases and other patient-related and procedure-related factors. Meta-analysis (random-effects models) was conducted using odds ratio (OR) and mean difference (MD). Risk of bias (ROBBINS-I) and strength of the evidence (GRADE) were assessed. RESULTS: The study included 37 studies (2,470,827 patients). Older age was a protective factor (MD = - 1.18). Male gender (OR 1.34), coagulopathy (3.05), congestive heart failure (2.36), diabetes mellitus (1.80), obesity (1.61), systemic neoplasia (1.57), chronic lung disease (1.52), and hypertension (1.21) increased the risk for PJI. Behavioral risk factors comprised alcohol abuse (2.95), immunosuppressive therapy (2.81), steroid therapies (1.88), and tobacco (1.82). Infectious risk factors included surgical site infections (6.14), postoperative urinary tract infections (2.85), and prior joint infections (2.15). Rheumatoid arthritis, posttraumatic native joint disease, high National Nosocomial Infections Surveillance (NNIS) system surgical patient index score, prior joint operation, American Society of Anesthesiologists score ≥ 3 and obesity were also significantly associated with higher risk of PJI. Osteoarthritis and blood transfusion were protective factors. CONCLUSIONS: The main risk factors for PJI in each category were male gender, coagulopathy, alcohol abuse, surgical site infection (highest score) and high NNIS system surgical patient index score. Protective factors were age, female gender in TKA, osteoarthritis and blood transfusion. Optimization of modifiable risk factors for PJI should be attempted in clinical practice. LEVEL OF EVIDENCE: II.
