Surgical site infection (SSI) rates after pediatric spine (PS) surgery: Is underlying diagnosis (UD) a predictor of infection (POI)?

Abstract

ISSUE: We have tracked SSI rates in PS patients at this university-affiliated hospital since 1995. From 1995 to 2000, we successfully decreased our overall infection rate 58.5% (from 12.23% to 5.11%). In 2001, we noted a small increase, and by 2002, our SSI rate had increased almost threefold to 13.8%. During 2003, traditional interventions only reduced the rate to 11.8%. PROJECT: Increased emphasis on patient safety led us to intensify our efforts during 2004. At the 1991 meeting of the Pediatric Orthopaedic Society of North America, Kretzler reported SSI rates of 11% in patients with cerebral palsy (CP) and spina bifida (SB) compared to 1% for those with idiopathic scoliosis (IS). We looked at our cases and infections based on UD with the intent of altering our interventions if indicated. Using discharge diagnosis (ICD-9) codes, we were able to divide our cases into the following categories: IS (including kyphosis), genetic disorders, neurological disorders (ND) (including CP, SB, and muscular dystrophy), and other. To facilitate decisions regarding interventions, we looked at depth of infection and infecting organism group as it related to UD. RESULTS: Of our cases, 62.2% had an underlying diagnosis of IS; 15.5% were ND; 3.5% were genetic, and 18.8% were other. The patients ranged in age from less than 1 year to 18 years; 66% were between ages 12 and 16. Our infection rates were 19% in our ND patients; 12.8% in our genetic patients, and 8.8% in our IS patients. We performed a multivariate analysis using logistic regression which showed that three variables were predictors of infection: being Black (versus Caucasian) (risk ratio [RR]=1.9, p=0.006) and having UD of either CP (RR 2.0, p=0.02) or SB (RR 4.2, p < 0.001). Our analysis showed that 37.5% of the deep SSIs occurred in our ND patients, 33.3% in IS, and 29.2% in other. SSI organisms in ND patients were 20.9% gut flora compared to IS patients which were 19.1% (p=0.05). LESSONS LEARNED: Our overall infection rate was higher for patients with underlying diagnosis of ND compared to those with IS. We were surprised to discover that race was a risk factor and further exploration of this variable is warranted. Based upon the findings related to infecting organisms, we will be looking at new interventions that address incontinence pre- and post-operatively as well as antibiotic prophylaxis that provides better coverage against enteric organisms such as a second-generation cephalosporin.