Background: Surgical site infection is a serious complication of posterior spinal fusion for neuromuscular scoliosis, with a reported prevalence of 6% to 24%. A single-institution experience over a thirty-year period was reviewed to determine the prevalence of surgical site infection after posterior spinal fusion for neuromuscular scoliosis, and to identify patient and treatment-related risk factors.
Methods: Our retrospective review included all patients treated with posterior spinal fusion (alone or in combination with an anterior procedure) for neuromuscular scoliosis from 1980 to 2009 and followed for a minimum of two years. Univariate and multivariate statistical analysis was performed to identify significant risk factors for occurrence of deep surgical site infection (p < 0.05).
Results: The study included 428 patients with an average duration of follow-up of 4.9 years. The mean Cobb angle was 74.3°. Most (74%) were treated with posterior spinal fusion alone. Deep infection developed in forty-four patients (10.3%); 57% of the infections occurred within three months after the surgery and 73%, within twelve months. Nearly half (45%) of the infections were polymicrobial; 59% of the organisms were gram-positive and 41% were gram-negative. Implant removal was required in 58% of the patients. Surgical site infection was more frequent from 1980 to 1989 (20.3%) than it was from 1990 to 2009 (8.4%) (odds ratio [OR] = 2.8, p = 0.01 in univariate analysis). Surgical site infection was more common in patients with spina bifida (21.5%) than in those with other diagnoses (8.3%) (OR = 3.0, p = 0.001). Other patient factors associated with surgical site infection were a body mass index (BMI) of >25 kg/m2 (OR = 2.4, p = 0.04) and incontinence (OR = 2.4, p = 0.009). Treatment factors associated with surgical site infection were inadequate prophylactic antibiotic dosing (cefazolin ≤20 mg/kg) (OR = 3.3, p = 0.0002), length of fusion (p = 0.002), pelvic fixation (OR = 2.4, p = 0.04), length of hospital stay (p = 0.005), and other complications (OR = 3.2, p = 0.0003). Drain output (p = 0.04) and lower hemoglobin levels (p = 0.008) were significantly associated with surgical site infection in patients with spina bifida, and drain use (superficial to the fascia) was protective in those without spina bifida (OR = 0.5, p = 0.046).
Conclusions: This study identified modifiable factors, especially antibiotic dosing and drain use, associated with surgical site infection in patients with neuromuscular scoliosis.
Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
1Texas Scottish Rite Hospital for Children, 2222 Welborn Street, Dallas, TX 75209. E-mail address for B.A. Ramo: firstname.lastname@example.org