In the largest trial of its kind, patients receiving oral versus intravenous (IV) antibiotics for fracture-related infections (FRIs) had similar reoperation and reinfection rates, US researchers reported this week in JAMA Surgery.
But the primary and secondary analyses of trial results came to different conclusions regarding noninferiority, leaving some uncertainty about the findings.
The POvIV trial, led by investigators with the Major Extremity Trauma Research Consortium, was a prospective, multicenter randomized clinical trial conducted at 24 US trauma centers from March 2013 through September 2018. Participants were adults who had developed an FRI following fracture repair or arthrodesis with fixation with implants (fusion of two or more bones using screws or plates).
FRIs are common following fracture repair and can prolong recovery and lead to adverse events. Current guidelines recommend 6 weeks of IV antibiotics. But there are concerns about increased costs and risks associated with prolonged IV therapy, and observational studies have indicated oral antibiotics may have similar efficacy. The main aim of the trial was to see how FRI patients treated with oral antibiotics for 6 weeks fared compared with those who received IV antibiotics.
The primary outcome was the number of studied injury-related surgical interventions within 1 year after discharge from the initial infection hospitalization. Oral antibiotics were considered noninferior to IV if the mean difference in surgical interventions was less than or equal to 0.67. Secondary outcomes included recurrence of deep surgical infection. Outcomes were analyzed using a modified intention-to-treat (mITT) approach.
Analyses differ on noninferiority
A total of 233 patients (mean age, 46; 77% male) were randomized, with 115 in the oral group and 118 in the IV group. Patients in the oral group had 137 surgical interventions within 1 year of discharge, compared with 118 in the IV group. The mean number of surgical interventions within 1 year was 1.32 (95% confidence interval [CI], 1.00 to 1.66) and 1.09 (95% CI, 0.83 to 1.36) for the oral and IV groups, respectively, with an estimated difference of 0.23.
The upper bound of the 95% CI of the mean difference in the unadjusted mITT analysis was 0.59, which was lower than the prespecified noninferiority margin of 0.67, indicating noninferiority of oral to IV antibiotics. An adjusted mITT analysis also showed noninferiority.
But in a secondary per-protocol (PP) analysis, the mean number of surgical interventions within 1 year was 1.39 (95% CI, 1.05 to 1.74) for the oral group and 1.03 (95% CI, 0.78 to 1.30) for the IV group. The estimated difference in means was 0.35, and the upper bound of the 95% 1-sided CI was 0.71, indicating insufficient evidence to conclude noninferiority.
The treatment effects estimates for the key secondary outcome of reinfection showed a similar pattern as those for the primary outcome, with similar rates of reinfection in both groups in the unadjusted mITT analysis and greater differences in rates in the adjusted mITT and PP analyses.
"Despite meeting the noninferiority margin based on mITT analyses of the primary outcome, the secondary PP analysis suggested fewer overall study injury–related surgical interventions for any IV vs oral antibiotics only groups," the study authors wrote.
The authors say that an important next step would be to compare the results of the POvIV trial to those among subgroups of patients with FRIs in a previous trial involving patients with bone and joint infections. That trial (the 2019 OVIVA trial) found that oral antibiotics were noninferior to IV when used during the first 6 weeks of treatment, but results from the subgroup analysis have not been reported.
In the meantime, the authors concluded, "Clinicians and patients can use these data to inform shared decision-making regarding antibiotic care for FRIs."