Surgical Site Infection Following Neuromuscular Posterior Spinal Fusion Fell 72% After Adopting the 2013 Best Practice Guidelines

Stephen R. Stephan, MD; Kenneth D. Illingworth, MD; Kavish Gupta, MD; Lindsay M. Andras, MD; David L. Skaggs, MD, MMM


Spine. 2021;46(17):1147-1153. 

In This Article


One hundred ninety eight primary PSF for NMS were included in this study. Sixty two were performed between January 2008 and December 2012 and 136 between January 2014 and December 2018. All patients had at least 1-year of documented follow-up visits at our institution.

In regards to preoperative variables (Table 2), we found no differences in regards to age, sex, body mass index (BMI), preoperative Cobb angle, presence of bowel and bladder incontinence, presence of a tracheostomy, and use of non-invasive ventilation (P > 0.05). We found more patients in Group 2 with the presence of a gastrostomy tube (P = 0.0043). For operative variables (Table 3), we found no differences in instrumentation to pelvis, length of fusion, estimated blood loss, hematocrit, intraoperative transfusions, surgical drain use, and surgical time (P > 0.05). Implant material most commonly used in Group 1 was stainless steel (n = 43, 69.4%) and cobalt chrome (n = 100, 73.5%) in Group 2 (P < 0.001).

There were five attending surgeons in this study. In Group 1, 11 (18%) patients were operated on by junior surgeons and 51 (82%) by senior surgeons. In Group 2, 98 patients (72%) were operated on by junior surgeons. In Group 1, 17.7% of cases were done as dual attending cases, similar to the 16.9% in Group 2 (P = 0.885). Operative times were similar for Group 1 at 364.3 ± 128 and Group 2 at 360.7 ± 103 minutes (P = 0.8457).

The most common etiology of NMS in Groups 1 and 2 was CP (n = 25, 40.3%; n = 77, 56.6%). Gross Motor Function Classification System (GMFCS) levels for CP patients were similar between the groups (P = 0.608). Other etiologies are listed in Table 2.

A nutrition consult within 12 months prior to surgery was obtained for 15 (24.2%) in Group 1 and 64 (47.1%) in Group 2 (P = 0.002). Ancef or similar antibiotic was used in all patients in Groups 1 and 2 (P = 1.0). Antibiotics for gram-negative coverage were used perioperatively in 52 (83.9%) of Group 1 and 136 (100%) of Group 2 (P < 0.001). Vancomycin or tobramycin powder was used in the bone graft or applied to the surgical wound in 13 (21%) and 132 (97.1%) patients in Groups 1 and 2, respectively (P < 0.001). These data are summarized in Table 4.

Deep Surgical Site Infection

Overall, deep surgical site infection occurred in 10 (16.1%) patients in Group 1 and six (4.4%) in Group 2 (P = 0.005, Table 5). Rates of deep surgical site infection varied by diagnosis. For Group 1: four (40%) muscular dystrophy, three (30%) other diagnoses, two (20%) CP, and one (10%) syndromic. For Group 2: two (33.3%) CP, two (33.3%) syndromic, one (16.7%) muscular dystrophy, and one (16.7%) other. All CP patients who suffered an infection were GMFCS level 5. Rate of deep surgical site infection for GMFCS level 5 patients in Group 1 and Group 2 was 10% and 3.7%, respectively (P = 0.287).

Time to onset of deep surgical site infection for Groups 1 and 2 were 41.5 ± 32.9 and 115.7 ± 150.9 days (P = 0.1468, Table 5). All deep surgical site infections in both groups were managed operatively with irrigation and debridement, wound vacuum application, and with or without partial or complete removal of instrumentation.

Of the identified organisms, 13 (59.1%) were gram-negative, seven (31.8%) gram-positive, and two cases where the organism could not be identified (9.1%). Eleven (84.6%) of the gram-negative organisms were isolated from patients within Group 1 (P = 0.047). Polymicrobial infections accounted for six (37.5%) overall infections, with Group 1 having half (50%) polymicrobial infections and Group 2 only one (16.7%) (P = 0.182, Table 5). Isolated organisms are summarized in Table 6.

The groups were examined to search for any potential infection clusters. In Group 1, two patients whose primary surgeries occurred on the same day (in March 2008, by different surgeons) experienced infections. One patient developed an infection with Pseudomonas stutzeri, Methicillin Resistant Staphylococcus aureus, and Escherichia coli at 15 days postoperatively, while the other developed an Enterococcus faecalis and coagulase negative staph infection at 46 days postoperatively. In Group 2, there were two patients that developed infections with primary procedures 2 weeks apart. One of these patients developed an infection with Propionibacterium acnes at 11 days postoperatively and the other patient had no growth from cultures at 300 days postoperatively. Also in Group 2 there were two patients with their primary procedures 21 days apart that both developed infections. One of these patients had an infection with Proteus mirabilis at 32 days postoperatively and the other patient had an infection with Staphylococcus lugdunensis and Cornybacterium amycolatum that occurred at 320 days postoperatively. All other patients who developed infections had greater than 1 month between their primary procedures.