Surgery Before Subspecialty Referral for Periprosthetic Knee Infection Reduces the Likelihood of Infection Control.

BACKGROUND: Failure to control a periprosthetic joint infection (PJI) often leads to referral of the patient to a tertiary care institution. However, there are no data regarding the effect of prior surgical intervention for PJI on subsequent infection control. QUESTIONS/PURPOSES: (1) Is the likelihood of 2-year infection-free survival worse if an initial surgery for PJI was performed before referral to a tertiary care center when compared with after referral for definitive treatment? (2) Is the likelihood of identifying a causal organism during PJI worse if the initial surgery for PJI was performed before referral to a tertiary care center when compared with after referral for definitive treatment? (3) We calculated how many patients are harmed by the practice of surgically attempting to treat PJI before referral to a tertiary care center when compared with treatment after referral to a tertiary care center for definitive treatment.
METHODS: Among 179 patients (182 TKAs) who were referred for PJI between 2004 and 2014, we retrospectively studied 160 patients (163 TKAs) who had a minimum of 2 years of followup after surgical treatment or had failure of treatment within 2 years. Nineteen TKAs (19 patients) were excluded from the study; 13 patients (7%) had < 2-year followup, three patients had infected periprosthetic fractures, and three patients had infected extensor mechanism reconstruction. Eighty-six patients (88 TKAs, two bilateral [54%]) had no surgical treatment before referral to our institution for PJI management, and 75 patients (75 TKAs [46%]) had PJI surgery before referral. The mean followup was 2.4 ± 1.2 years for patients with PJI surgery before referral and 2.8 ± 1.3 years for patients with no surgery before referral (p = 0.065). Infection-free survival was defined as prosthesis retention without further surgical intervention or antibiotic suppression. During the period, further surgical intervention generally was performed after failure of irrigation and débridement, a one- or two-stage procedure, or between stages of a two-stage reimplantation without documentation of an eradiated infection, and antibiotic suppression generally was used when patients were not medically sound for surgical intervention or definitive implants were placed after the second of a two-stage procedure with positive cultures; these criteria were applied similarly to all patients during this time period in both study groups. Endpoints were assessed using a longitudinally maintained institutional database, and the treating surgeons were not involved in data abstraction. Relative and absolute risk reductions with 95% confidence intervals (CIs) as well as a Kaplan-Meier survival curve with a Cox proportional hazard model were used to evaluate survival adjusting for significant covariates. The number needed to harm is calculated as the number needed to treat. It is the reciprocal of the absolute risk reduction or production by an intervention.
RESULTS: The cumulative infection-free survival rate of TKAs at 2 years or longer was worse when PJI surgery was performed before referral to a tertiary center (80%; 95% CI, 69%-87%) compared with when no PJI surgery was performed before referral (94%; 95% CI, 87%-98%; log-rank test p = 0.006). Additionally, PJI surgery before referral resulted in a lower likelihood of causative microorganism identification (52 of 75 [69%]) compared with patients having surgery at the tertiary center (77 of 88 [88%]; odds ratio, 2.71; 95% CI, 1.28-4.70; p = 0.006). With regard to the infection-free survival rate of TKAs, the number needed to harm was 7.0 (95% CI, 4.1-22.5), meaning the referral of less than seven patients to a tertiary care center for definitive surgical management of PJI before intervention at the referring hospital prevents one infection-related failure. With regard to the culture negativity in PJI, the number needed to harm was 5.5 (95% CI, 3.3-16.7), meaning the referral of less than six patients to a tertiary care institution for PJI before surgery at the outside hospital prevents the diagnosis of one culture-negative infection.
CONCLUSIONS: Surgical treatment of a PJI before referral for subspecialty surgical management increases the risk of failure of subsequent surgical management. The prevalence of culture-negative PJI was much higher if surgery was attempted before referral to a tertiary care center when compared with referral before treatment. This suggests that surgical treatment of PJI before referral to a treating center with specialized expertise in PJI compromises the infection-free survival and impacts infecting organism isolation. LEVEL OF EVIDENCE: Level III, therapeutic study.