Many studies report the incidence and prevalence of surgical site infections (SSIs) following open fractures; however, there is limited information on the treatment and subsequent outcomes of superficial SSIs in open fracture patients. There is also a lack of clinical studies describing the prognostic factors that are associated with failure of antibiotic treatment (non-operative) for superficial SSI. To address this gap, we used data from the FLOW (Fluid Lavage in Open Fracture Wounds) trial to determine how successful antibiotic treatment was for superficial SSIs and to identify prognostic factors that could be predictive of antibiotic treatment failure.

This is a secondary analysis of the FLOW trial dataset. The FLOW trial included 2,445 operatively managed open fracture patients. FLOW participants who had a non-operatively managed superficial SSI diagnosed in the 12 months post-fracture were included in this analysis. Participants were grouped into two categories: 1) participants whose superficial SSI resolved with antibiotics alone and 2) participants whose SSI did not resolve with antibiotics alone (defined as requiring surgical management or SSI being unresolved at final follow-up (12-months post-fracture for the FLOW trial)). Antibiotic treatment success and the date when this occurred was defined by the treating surgeon. A logistic binary regression analysis was conducted to identify factors associated with superficial SSI antibiotic success. Based on biologic rationale and previous literature, a priori we identified 13 (corresponding to 14 levels) potential factors to be included in the regression model.

Superficial SSIs were diagnosed in168 participants within 12 months of their fracture. Of these, 139 (82.7%) had their superficial SSI treated with antibiotics alone. The antibiotic treatment was successful in resolving the superficial SSI in 97 participants (69.8%) and unsuccessful in resolving the SSI in 42 participants (30.2%). We found that superficial SSIs that were diagnosed later in follow-up were associated with failure of treatment with antibiotic alone (Odds ratio 1.05 for every week in diagnosis delay, 95% Confidence Interval 1.004–1.099; p=0.03). Age, sex, fracture severity, fracture pattern, wound size, time from injury to initial surgical irrigation and debridement were not associated with antibiotic treatment failure.

Our secondary analysis of prospectively collected FLOW data found antibiotics alone resolved superficial SSIs in 69.8% of patients diagnosed with superficial SSIs. We also found that superficial SSIs that were diagnosed earlier in follow-up were associated with successful treatment with antibiotics alone. This suggests that if superficial SSIs are diagnosed and treated promptly, there is a higher probability that they will resolve with antibiotic treatment.

Negative pressure wound therapy (NPWT) is commonly used to manage severe open fracture wounds. The recently completed X randomized controlled trial (RCT) evaluated the effect of NPWT versus standard wound management on 12-month disability and rate of deep infection among patients with severe open fractures of the lower limb and reported no differences. Using data from the Y trial of open fracture patients, we aimed to evaluate the impact of NPWT on the odds of having deep infections and health-related quality of life (HRQL).

Our analyses included participants from the Y trial who had Gustilo II and III lower extremity fractures. To adjust for the influence of injury characteristics on type of dressing received, a propensity score was developed from the dataset. A one-to-one matching algorithm was then used to pair patients with a similar propensity for NPWT.

Mixed effects logistic regression was used to evaluate the association between type of wound dressing and development of a deep infection requiring operative management (dependent variable) in the matched cohort. Gustilo type, irrigation solution, fracture location, mechanism of injury, and degree of contamination were included as adjustment variables. To determine any differences in HRQL between the NPWT and standard wound dressing groups, we conducted two multi-level models with three levels (centre, patient, and time) and included Short Form-12 (SF-12) Physical Health Component (PCS) and SF-12 Mental Health Component (MCS) as dependent variables. Gustilo type, irrigation solution, fracture location, mechanism of injury, degree of contamination, and pre-injury SF-12 scores were included as adjustment variables. All tests were 2-tailed with alpha=0.05.

After applying propensity score-matching to adjust for the influence of injury characteristics on type of dressing used, there were 270 matched pairs of patients available for comparison. The odds of developing a deep infection requiring operative management within 12 months of initial surgery was 4.22 times higher in patients who received NPWT compared to those who did not receive NPWT (Odds Ratio (OR) 4.22, 95% Confidence Interval (CI) 2.26–7.87.

1,329 participants were included in our HRQL analysis and those treated with NPWT had significantly lower SF-12 PCS at all follow-up visits (6w, 3m, 6m, 12m) post fracture (p=0.01). Participants treated with NPWT had significantly lower SF-12 MCS at 6-weeks post-fracture (p=0.03).

Unlike the X trial, our analysis found that patients treated with NPWT had higher odds of developing a deep infection requiring operative management and that being treated with NPWT was associated with lower physical quality of life in the 12 months post-fracture. While there may have been other potential adjustment variables not controlled for in this analysis, our results suggest that the use of this treatment should be re-evaluated.

Despite long-standing dogma, a clear relationship between the timing of surgical irrigation and the development of subsequent deep infection has not been established in the literature. Traditionally, irrigation of an open fracture has been recommended within six-hours of injury based on animal studies from the 1970s, however the clinical basis for this remains unclear. Using data from a multi-centre randomized controlled trial of 2,447 open fracture patients, the primary objective of this secondary analysis is to determine if a relationship exists between timing of wound irrigation (within six hours of injury versus beyond six hours) and subsequent reoperation rate for infection or healing complications within one year for patients with open extremity fractures requiring surgical treatment.

To adjust for the influence of patient and injury characteristics on the timing of irrigation, a propensity score was developed from the data set. Propensity-adjusted regression allowed for a matched cohort analysis within the study population to determine if early irrigation put patients independently at risk for reoperation, while controlling for confounding factors. Results were reported as odds ratios (ORs), 95% Confidence Intervals (CIs), and p-values. All analyses were conducted using STATA 14 (StataCorp LP, College Station, TX, USA).

Two thousand, two hundred eighty-six of 2,447 patients randomized to the trial from 41 orthopaedic trauma centers across five countries had complete data regarding time to irrigation. Prior to matching, the patients managed with early irrigation had a higher proportion requiring reoperation for infection or healing complications (17% versus 12.8%, p=0.02), however this does not account for selection bias of more severe injuries preferentially being treated earlier. After the propensity score-matching algorithm was applied, there were 373 matched pairs of patients available for comparison. In the matched cohort, reoperation rates did not differ between early and late groups (16.1% vs 16.6%, p=0.84). When accounting for propensity matching in a logistic regression analysis, early irrigation was not associated with reoperation (OR 0.93, 95% CI 0.62 to 1.40, p=0.73).

When accounting for other variables, late irrigation does not independently increase risk of reoperation.