Surgical Site Infections in Posterior Lumbar Surgery: A Controlled-Cohort Study of Epidural Steroid Paste

A retrospective, single-center cohort study of consecutive patients undergoing posterior lumbar decompression between 2007 and 2013 was conducted.

To compare rates of surgical site infection between matched cohorts of patients undergoing lumbar surgery with and without intraoperative application of epidural steroid pastes.

Epidural steroid agents reduce postoperative pain and inflammation following lumbar surgery, reducing the use of postoperative narcotics and improving McGill pain scores. Preliminary studies have, however, suggested an increase in surgical site infections following the use of these steroid-containing pastes.

We reviewed 758 patients undergoing decompression performed at a single center by surgeons who either routinely used or never used an analgesic steroid paste. Patients undergoing instrumentation or revision surgery were excluded, and surgical and postoperative protocols were uniform. Two hundred eighty-three patients met specific inclusion and exclusion criteria. Demographic, clinical, and surgical data were assessed and correlated to the incidence of postoperative infections. Multivariate logistic regression controlled for confounding characteristics and identified independent predictors of postoperative surgical-site infections.

Patient demographics, comorbidities, and perioperative protocols were similar between groups. There were six acute infections among 103 patients receiving steroid paste (5.83%), and two infections among 180 patients not receiving paste (1.11%), a statistically significant difference which remained after controlling for confounding characteristics (odds ratio 6.74, P = 0.01). All but one infection occurred among patients with identifiable preoperative risk factors for infection.

The observed increase in infection among patients receiving pain paste is clinically significant, but infection was primarily observed among at-risk patients. The present study confirms suspicions raised in preliminary studies, and we recommend caution when treating patients with identifiable, comorbid risk factors.

3.