Analysis of Postoperative Thoracolumbar Spine Infections in a Prospective Randomized Controlled Trial Using the Centers for Disease Control Surgical Site Infection Criteria

Abstract

Introduction

Wound infections following spinal surgery place a high toll on both the patient and the healthcare system. Although several large series studies have examined the incidence and distribution of spinal wound infection, the applicability of these studies varies greatly since nearly every study is either retrospective and/or lacks standard inclusion criteria for defining surgical site infection. To address this void, we present results from prospectively gathered thoracolumbar spine surgery data for which the Centers for Disease Control (CDC) criteria were stringently applied to define a surgical site infection (SSI).

Methods

A prospective randomized trial of 314 patients who underwent multilevel thoracolumbar spinal surgery with instrumentation followed by postoperative drain placement was completed (Takemoto et al., 2015). The trial consisted of two antibiotic arms: one for 24-hours, and the other for the duration of the drain; no differences were found between the arms. All infections meeting CDC criteria for SSI were included.

Results

A total of 40 infections met CDC criteria for SSI, for an overall incidence of 12.7%. Of these, 20 (50%) were culture-positive. The most common organism was Staphylococcus aureus (4 total: methicillin-sensitive=2; methicillin-resistant=2), followed by coagulase-negative Staphylococcus (3 cases), Propionibacterium acnes and Escherichia coli (2 cases each). Six infections grew multiple organisms, most commonly involving coagulase-negative staphylococcus and enterococcus.

Conclusions

Our findings indicate that thoracolumbar SSI occurs at the higher end of the range cited in the literature (2-13%), which is largely based on retrospective data not subjected to the inclusivity of SSI as defined by the CDC. The three most common organisms in our analysis (S. aureus, P. acnes, E. coli) are consistent with previous reports. Staphylococcus aureus continues to be the most common causative organism and continued vigilance and searching for preventive measures need to be a high priority. This study provides Level I evidence.

Introduction

Wound infections following spinal surgery are fraught with morbidity, placing a high toll on both the patient and the healthcare system.^1–3 Although several large series studies have examined the incidence and distribution of spinal would infection, the applicability of these studies varies greatly since nearly every study is either retrospective and/or lacks standard inclusion criteria for defining surgical site infection.^4–7 To address this void, we present results from prospectively gathered thoracolumbar spine surgery data for which the Centers for Disease Control (CDC) criteria were stringently applied to define a surgical site infection (SSI).^8–9

Methods

The data source for this study was a prospective randomized comparative study in which patients with a deformity or degenerative condition of the thoracic and/or lumbar spine requiring spinal surgery with a postoperative drain were randomized to two antibiotic-regimen arms: one for 24 hours, and the other for the duration of drain placement.⁹ Two hundred and fifty-six received cefazolin, while 13 received vancomycin and the remaining 45 received clindamycin.⁹ Patients were enrolled from September 2008 until February 2011; a total of 314 patients met criteria for randomization and the study was approved by our institution's Institutional Review Board as previously described.⁹ All surgeries were performed by fellowship-trained orthopaedic spine surgeons, and all data was prospectively collected postoperatively. All patients had a minimum of one-year of postoperative follow-up;

Surgical site infection (SSI) was defined according to CDC criteria and was made by an attending physician specializing either in orthopaedic spinal surgery or infectious disease. As per CDC criteria, SSI was defined as an infection occurring within one year postoperatively that was related to the operative procedure and involved the skin incision, fascia, or muscle layers in a patient with at least one of the following: organisms isolated from an aseptically obtained culture of tissue or fluid, purulent drainage from a drain which had been placed into the surgical space via a stab wound, direct examination of an abscess or other evidence of infection involving the surgical space (i.e. during histological examination, radiographic examination, or during reoperation), or the diagnosis of SSI by an attending physician.¹⁰ SSI was used as the primary endpoint; there were no alterations of trial outcome measures after the trial began.⁹

Results

A total of 40 infections met CDC criteria for SSI, for an overall incidence of 12.7%. Of these, 20 (50%) grew out organisms. The most common organism was Staphylococcus aureus (4 total: methicillin-sensitive = 2; methicillin-resistant = 2), followed by coagulase-negative Staphylococcus (3 cases), Propionibacterium acnes and Escherichia coli (each with 2 total) (Table 1). Six of these 20 infections grew multiple organisms, which most commonly involved coagulase-negative staphylococcus and enterococcus.

Table 1.

Culture-positive surgical site infections following thoracolumbar spine surgery with drain placement

Organism	Depth of infection	Timing of Infection	Antibiotic Treatment Route	Revision Performed?	Incision and Drainage Performed?
MSSA	Superficial	Delayed	IV	No	No
MSSA	Deep	Acute & Delayed	PO & IV	No	Yes
MRSA	Deep	Acute	PO & IV	No	Yes
MRSA	Deep	Delayed	PO & IV	No	Yes
CNS	Superficial	Acute	PO & IV	No	No
CNS	Superficial	Acute	PO	No	No
CNS	Superficial	Acute	PO & IV	No	Yes
Propionibacterium acnes	Superficial	Acute	PO	No	No
Propionibacterium acnes	Superficial	Acute & Delayed	PO	No	Yes
Escherichia coli	Superficial	Acute	PO	No	Yes
Escherichia coli	Deep	Acute	IV	No	Yes
Enterococcus	Superficial	Acute	PO & IV	No	Yes
Enterobacter	Superficial	Acute	PO & IV	No	Yes
Klebsiella	Superficial and Deep	Acute	IV	No	Yes
Polymicrobial + CNS	Deep	Acute	PO & IV	No	Yes
Polymicrobial + Enterococcus	Superficial	Acute	PO & IV	No	Yes
Enterococcus + Candida albicans	Deep	Acute	IV	No	Yes
Gram-negative rods + Klebsiella + Enterococcus	Superficial	Delayed	PO & IV	No	No
Klebsiella + CNS	Superficial	Delayed	PO	Yes	No
Pseudomonas + CNS	Superficial	Delayed	IV	No	No

MSSA = Methicilin-sensitive Staphylococcus aureus; MRSA = Methicilin-resistant Staphylococcus aureus; CNS = Coagulase-negative Staphylococcus; Acute = within 30 days of surgery; Delayed = after 30 days following surgery; Acute & Delayed = one acute infection + one delayed infection in the same patient; IV = intravenous; PO = by mouth.

Of the 20 SSIs that grew out organisms on culture, 13 were superficial, six were deep, and one was both superficial and deep (Table 1). Five of these infections were delayed (occurring later than 30 days after surgery), 13 occurred acutely (defined as within 30 days following surgery), and two were classified as both acute and delayed (Table 1).

Five of the 40 infections required operative revision; only one of these had an organism that grew out on culture. 17 infections required subsequent incision and drainage; only two infections required both revision and I&D (neither of these grew out organisms on culture). The antibiotic route and I&D distribution among infections with culture-positive organisms can be found in Table 1. 37 of the 40 infections (93%) involved operative fusion. Exactly half of the 40 infections occurred in revision cases.

Discussion

Despite the preponderance with which operative management of the spine occurs, there have been relatively few reports in the literature examining the incidence of SSI in thoracolumbar spine surgery. Given the completion of our recent trial examining this issue, we believed that an investigation into the type of infections encountered in our study would provide a significant contribution to the literature given how few reports of SSI are collected prospectively and with rigorous adherence to CDC SSI criteria.

The established range of SSI in thoracolumbar spine surgery is 2-13% based on the existing literature.^{1–2, 11–13} Some centers have used prophylactic vancomycin powder intraoperatively in an attempt to lower infection rates, with mixed results.^14–16 None of the 314 patients in our study were exposed to prophylactic vancomycin powder intraoperatively.

The importance of this study’s reliance on the CDC definition of SSI cannot be overstated, as reported SSI rates have been shown to vary greatly when a standard SSI definition is not used.^17–18 Despite our infection rates being subjected to the inclusiveness of the CDC SSI definition, our overall infection rate of 12.7% falls within that of the published literature, which is based largely on nonstandard SSI criteria which would be expected to artificially lower their overall infection rates. It is our hope that our findings will encourage others to publish their own infection data without fear of infection rates being (falsely) presumed to be abnormally high in comparison with literature not relying on strict SSI definitions. Encouraging is the recent publication of a prospectively gathered spinal SSI population meeting CDC criteria, although those patients did not have the same increased risk as our patient population of universally having an intraoperative drain placed.¹⁹

An interesting finding was the prevalence of P. acnes, which occurred as frequently as MSSA and MRSA in our series and comprised 10% of the culturepositive organisms found in our series (Table 1); this incidence is consistent with recent literature.^20–21 Also consistent with the majority of the postoperative spinal infection literature was that our most common causative organism was S. aureus (comprising 20% of the culture-positive organisms in our study) with coagulase-negative staphylococcus and E. coli comprising a significant representation as well.^{19, 22–25}

In conclusion our series, which represents one of the first analyses of prospectively collected thoracolumbar postoperative spine infections meeting CDC criteria, revealed an overall infection incidence of 12.7%, of which only 50% grew organisms on culture. The three most common organisms in our analysis (S. aureus, P. acnes, E. coli) are consistent with previous reports. Our findings indicate that thoracolumbar SSI occurs at the higher end of the range cited in the literature (2-13%), which is largely based on retrospective data not subjected to the inclusivity of SSI as defined by the CDC. Staphylococcus aureus continues to be the most common causative organism and continued vigilance and searching for preventive measures need to be a high priority. Further studies of prospectively collected data using a larger sample size will be needed to better gauge the true incidence and most common causes of thoracolumbar SSI.

Disclosures

No author has any conflicts of interest; the trial from which the data was derived received no funding by any external funding source.