In the present study, we identified the risk factors for wound complications, wound infection, and reoperation for wound complications after spine metastasis surgery and deployed the resultant model as a web-based calculator.
Patients treated at a single comprehensive cancer center during a 7-year period were included. The demographics, pathology, comorbidities, laboratory values, and operative details were collected. Factors with P < 0.15 on univariable regression were entered into multivariable logistic regression to generate predictive models internally validated using 1000 bootstrapped samples.
Of the 330 patients included, 29 (7.6%) had experienced a surgical site infection. The independent predictive factors for wound-related complications were a higher Charlson comorbidity index (CCI; odds ratio [OR], 1.41 per point; P < 0.01), Karnofsky performance scale score ≤70 (OR, 2.14; P = 0.04), lower platelet count (OR, 0.49 per 105/μL; P < 0.01), revision versus index surgery (OR, 3.10; P = 0.02), and increased incision length (OR, 1.21 per level; P = 0.02). Wound infection was associated with a higher CCI (OR, 1.60 per point; P < 0.01), a lower platelet count (OR, 0.35 per 105/μL; P < 0.01), revision surgery (OR, 4.63; P = 0.01), and a longer incision length (OR, 1.25 per level; P = 0.03). Unplanned reoperation for wound complications was predicted by a higher CCI (OR, 1.39 per point; P = 0.003), prior irradiation (OR, 2.52; P = 0.04), a lower platelet count (OR, 0.57 per 105/μL; P = 0.02), and revision surgery (OR, 3.34; P = 0.03), The optimism-corrected areas under the curve were 0.75, 0.81, and 0.72 for the wound complication, infection, and reoperation models, respectively.
Low platelet counts, poorer health status, more invasive surgery, and revision surgery all independently predicted the risk of wound complications, including infection and unplanned reoperation for infection. Validation of the calculators in a prospective study is merited.