Introduction: To detect surgical site infection (SSI) at early time point after spine surgeries, we previously reported the efficacy of postoperative lymphopenia (PL) ¹ and our original scoring system (SSI score)^2, which consists of a combination of four postoperative serological markers. Frailty has also been highlighted as a preoperative risk factor for postoperative adverse events including SSI in patients undergoing spine surgeries. The purpose of this study was to investigate if the combined use of preoperative frailty and postoperative serological assessments would be further effective to predict SSI after spine surgery. 

Methods: A total of 419 patients who underwent spine surgeries at National Defense Medical College Hospital from 2012 to 2016 were retrospectively reviewed. Excluding patients with unplanned antimicrobial agents use and those with missing serological data, 327 patients were included. Nine patients who required reoperations due to SSI were grouped into the SSI group and the other 318 patients into the non-SSI group. Using 11-item modified Frailty Index (mFI-11), preoperative frailty was determined by dichotomous definition, frail (mFI-11 ≥3) or not frail (mFI-11 <3). Patients’ characteristics at baseline (sex, age, preoperative frailty), use of metal implants, PL<1000/mL at 7 days after surgery (PL7D)¹, and SSI score were compared between the two groups. SSI score is the number of items matching the following four criteria, and SSI score ≥3 was found to be a strong indicator of SSI²: (1) neutrophil count at 7D >5095/mL; (2) increase in neutrophil-to-lymphocyte ratio from preoperative baseline to postop 7D >1.27-fold; (3) change in lymphocyte count from postop 2D to 7D <1.05-fold; (4) change in CRP level from 2D to 7D after surgery >0.39-fold. 

Results: There were no significant differences in age, sex, and use of implants between the two groups. The SSI group were likely to show higher incidence of PL7D (44.4% in the SSI group vs 17.9% in the non-SSI group, p=0.07) and significantly higher percentage of frail patients (mFI-11] ≥3; 33.3% vs 8.8%, p=0.04) and those with SSI score ≥3 (88.9% vs 7.9%, p＜0.001) compared to the non-SSI group. Univariate logistic regression analysis identified the combination of mFI ≥3 and SSI score ≥3 as the strongest predictor of SSI among the variables examined in this study: PL7D (OR 4.61, 95%CI 1.20-17.81, p=0.27); mFI-11 ≥3 (OR 5.18, 95%CI 1.23-21.84, p=0.025); SSI score ≥3 (OR 93.76, 95%CI 14.39–595.36, p<0.001), mFI-11 ≥3 and SSI score ≥3 (OR 180.67, 95%CI 12.68–2573.77, p<0.01). 

Discussion: Preoperative frailty determined by mFI-11 ≥3 showed similar predictive value for SSI with PL7D. Preoperative assessment of frailty might be useful to label patients with higher risk of SSI and to modify surgical plans. Combination of preoperative frailty and postoperative SSI score can be a feasible and reliable predictor of SSI at early postoperative time point to optimize patients care for the prevention / early detection of SSI.