Instrumented spinal fusion is a common procedure performed to treat various spinal disorders, including spondylolisthesis, scoliosis, stenosis, instability, trauma, and tumor. These procedures may be complicated by deep wound infections in 1% to 14% of patients.1–16Postoperative wound infections typically lead to longer hospital stays, need for further surgery, increased cost, and higher morbidity and mortality.6,8,9,14,17 Fortunately, the majority of these cases can be successfully managed with irrigation and debridement and long-term regimens of parenteral and oral antibiotics.1,2,4,6,8,11,15,17,18 Instrumentation can generally be maintained while addressing early-onset postoperative wound infections, preserving stability of the spine, and providing the opportunity for successful arthrodesis.1,2,4,6,8,11,15,17,18 As a result, these patients usually achieve fusion similar to patients who had no postoperative complications. However, there is little information in the literature on how these patients do clinically in the longer term after their infection has been managed. The purpose of this study was to evaluate the 2-year health-related quality of life (HRQOL) measures of patients who have undergone instrumented spinal fusion complicated by deep wound infection, necessitating irrigation and debridement.
MATERIALS AND METHODS
This retrospective study was conducted after receiving approval from the Institutional Review Board. Surgical and clinical databases from 2001 to 2008 were reviewed for eligible subjects. Data on 1144 patients who underwent lumbar spinal fusion during this period were evaluated for eligibility. Inclusion criteria consisted of patients who underwent lumbar spinal fusion with complete preoperative and 2-year postoperative outcome measures and had acute (≤3 mo) postoperative deep wound infections, as defined by the Centers for Disease Control and Prevention,19 necessitating irrigation and debridement. The Centers for Disease Control and Prevention criterion for deep incisional surgical site infection is “the infection occurs within one year if an implant is in place and the infection appears to be related to the operative procedure and the infection involves deep soft tissues (e.g., fascial and muscle layers) of the incision. In addition, it must meet at least one of the following: purulent drainage from the deep incision but not from the organ/space component of the surgical site; a deep incision that spontaneously dehisces or is deliberately opened by a surgeon when the patient has at least one of the following signs or symptoms-fever (>38°C), localized pain, or tenderness, unless the incision is culture negative; an abscess or other evidence of infection involving the deep incision is found on direct examination, during reoperation, or by histopathologic or radiologic examination; or diagnosis of a deep incisional SSI by a surgeon or attending physician.”19 Patients who had surgery for trauma, tumor, or osteomyelitis were excluded. HRQOL measures collected and reviewed in this study included the Oswestry Disability Index (ODI),20the Medical Outcomes Study Short Form 36 (36-Item Short Form Health Survey),21 and numeric rating scales for back and leg pain.22The Physical Composite Summary (PCS) and Mental Composite Summary (MCS) of the 36-Item Short Form Health Survey were reviewed separately. Patient's demographic data recorded included age, sex, American Society of Anesthesiologists grade, body mass index, indication for surgery, workers' compensation, and smoking status. Surgical data included type of fusion and number of levels fused at index procedure, number of debridements, use of antibiotic beads, culture results, operative time, estimated blood loss, and length of hospital stay. Data regarding presenting symptoms and days between time of diagnosis and debridement were also collected.
Propensity scoring technique23,24 was then used to match the patients with surgical wound infection with a control group who did not have postoperative wound infections. This technique allows simultaneous matching for multiple characteristics to produce 2 similar comparison groups. Sex, age, body mass index, smoking status, indication for fusion, type of fusion, and number of levels fused as well as preoperative ODI, 36-Item Short Form Health Survey PCS, 36-Item Short Form Health Survey MCS, and back and leg pain scores were used to generate a control group.
Statistical analysis was carried out using SPSS software version 17.0 (SPSS Inc., Chicago, IL), with significance set at P less than 0.05 level. Student t test was used to determine any significant differences between continuous demographic variables and compare preoperative and postoperative HRQOL measure scores within and between study groups. Fisher exact test was used to compare categorical variables between the 2 groups as well as the proportion of patients achieving the minimum clinically important difference (MCID) for the different outcome measures in each group. The MCID is the smallest change that is important to patients and a concept that attempts to provide threshold values for evaluation of treatment effectiveness.25 Reported values of MCID vary according to HRQOL measures, patient population, spine pathology, treatment modality, external criteria, and calculation method. As a result, currently there is no consensus on optimal MCID values.26–29 However, this new concept improves on the analysis of HRQOL data that are clinically relevant. For purposes of this study, we chose values defined by Copay et al,30 who carefully derived MCID thresholds after statistical analysis of several calculation methods on the basis of a heterogeneous lumbar spine surgery population with a broad range of spinal disorders, which most closely represents our study group. The MCID threshold for ODI is 12.8, 4.9 for the 36-Item Short Form Health Survey PCS, and 1.2 for back pain and 1.6 for leg pain.30
Pearson correlations were also obtained to determine any correlations between 2-year HRQOL measures and the number of debridements, days between infection diagnosis and debridement, the use of antibiotic beads, and the need for flap closure.
RESULTS
A total of 30 subjects with postoperative deep wound infections after instrumented lumbar spinal fusion met the inclusion criteria. Seventeen patients had previously undergone posterior spinal fusion, 3 underwent anterior and posterior spinal fusion, 8 underwent transforaminal interbody fusions, and 2 underwent anterior-only lumbar interbody fusions. Using the propensity score-matching technique, an equal number of 30 patients without wound infections were used as a case-control comparison group. The demographic, surgical, and baseline HRQOL data for these groups are summarized in Table 1. Consistent with the propensity score-matching technique, there were no significant demographic, surgical, or preoperative HRQOL differences between the infection and control groups.
Image Tools
In the infection group, 90% patients (27 patients) presented with incisional wound drainage and 40% patients (12 patients) had fever. The average time from index surgery to the diagnosis of infection was 16.32 days (range, 5–35 d). The average number of days between the time of diagnosis of the infection and surgical debridement was 1.13 days (range, 0–4 d). Mean number of debridements was 1.97 (range, 1–5). Of the organisms isolated from the wound infections, gram-negative bacteria were the most common (33%) followed byStaphylococcus aureus (30%) and methicillin-resistant S. aureus(20%). Infection was polymicrobial in 3 patients. Seven of the patients had negative cultures. These patients were placed on antibiotics by a physician who saw them before presenting to the surgeon who performed the index surgery, which may account for the negative cultures. During surgical exploration, all these patients presented with drainage tracking below the fascia at the time of surgery. These patients were assumed to have a deep wound infection despite the negative cultures. In 4 patients (13%), distant foci of infection could be identified.
Forty percent of the patients had antibiotic beads placed in the wound after irrigation and debridement. All beads were removed at final debridement. Patients were placed on a course of appropriate intravenous antibiotics for 6 weeks and continued on varied regimens on the basis of culture and sensitivity results. The majority of patients were successfully managed by repeat debridements and primary closure, with only 4 patients requiring flap closure.
All patients had a minimum 24-month follow-up, with a mean follow-up period of 38 months. The 2-year postoperative HRQOL scores are summarized in Table 2. All patients had a statistically significant improvement from preoperative to 2 years postoperative for all HRQOL measures. However, the infection group had significantly higher back pain scores and smaller changes in their scores than the control group (6.45 vs. 4.70, P = 0.020) at their final postoperative evaluation. There were no significant differences in postoperative ODI, 36-Item Short Form Health Survey PCS, 36-Item Short Form Health Survey MCS, or leg pain scores between the groups. Nevertheless, a greater proportion of patients in the control group (18, 60%) achieved MCID for ODI than those in the infection group (8, 27%, P = 0.018). There were no significant differences between the proportion of patients achieving MCID for 36-Item Short Form Health Survey PCS and back or leg pain postoperatively between groups.
Image Tools
Pearson correlation analysis demonstrated no significant association between number of debridements (r = 0.18, P = 0.365), days between infection diagnosis and debridement (r = −0.10, P = 0.599), use of beads (r = 0.07, P = 0.707), or need for flap closure (r = 0.32, P = 0.095) and 2-year back pain scores.
Fusion success was not specifically addressed in this study. However, there was no evidence of hardware failure on radiographs at the final follow-up period, and no patients underwent revision surgery for nonunion.
DISCUSSION
Surgical wound infections are a known risk of surgery. The reported risk of deep wound infections after instrumented spinal fusion is generally considered higher than fusions performed without instrumentation.12,14–16 Instrumentation further complicates the management of wound infections, often requiring multiple debridements, prolonged regimens of antibiotic therapy, increased hospital stays, and concerns for latent or persistent infections when the instrumentation is not removed.
Other surgical risk factors that have been described to have an association with an increase in the rate of infection after spine surgery include the type of surgery, spinal levels, blood loss, and prolonged operative time.1,3,6,7,14,16
Our study demonstrates that at 2-year follow-up, patients undergoing spinal fusion show improved postoperative HRQOL measures whether or not their surgeries are complicated by acute postoperative deep wound infections. Although the postoperative values of ODI, 36-Item Short Form Health Survey PCS, 36-Item Short Form Health Survey MCS, and leg pain were not significantly different between the groups, there was a trend toward better scores in all of the outcome measures evaluated in the control group.
This study reflects similar findings in previous studies with smaller patient populations. In a study of 7 patients with postoperative infection after surgery for adolescent idiopathic scoliosis, Rihn et al9 found no significant difference in terms of Pain, Function, Self-Image, Satisfaction, or Total Scoliosis Research Society-24 scores compared with the 229 patients who did not have an infection. Similarly, comparing a group of 16 patients with postoperative wound infections after spinal fusion with instrumentation with a matched control group with a minimum 2-year follow-up, Mok et al17 showed no difference in their postoperative 36-Item Short Form Health Survey PCS scores. However, due to the lack of preoperative HRQOL measures, it is unclear whether the groups were similar at baseline.
Despite having similar degrees of functional improvement as measured by the ODI and the 36-Item Short Form Health Survey PCS, patients who had deep wound infections had significantly worse back pain compared with the matched control group. When one considers that the additional injury to muscle is caused by infection and repeat debridements from postoperative wound infection management, it is reasonable to presume that this can cause significant disability by means of increased muscle dysfunction contributing to greater back pain. Several authors have reported on the concept that damage to muscle with surgical dissection and retraction can lead to muscle atrophy, denervation, and dysfunction contributing to the development of back pain.31–39 This explanation, however, cannot be directly concluded from our study because we did not assess muscle damage or dysfunction in our patients. Nonetheless, physical therapy, targeting exercises to recover extensor muscle strength and endurance, may improve postoperative functional recovery in these patients. Formal assisted physical therapy regimens should be considered in patients after postoperative instrumented fusions complicated by infection.
This study also revealed that the noninfected cohort had a greater proportion of patients achieving MCID for postoperative ODI and back and leg pain, but only ODI was statistically different from the infected group. This further demonstrates that the noninfected group has a trend toward better postoperative results.
Interestingly, the significant differences in back pain scores were not reflected in the 36-Item Short Form Health Survey PCS scores but seemed to correspond more with values of ODI. The ODI may be a more sensitive measure of disability in this population because it specifically measures disability from low back pain,20 whereas the 36-Item Short Form Health Survey PCS is a measure of disease burden in general.21,40 Back pain in itself may be tolerated by some patients, allowing them to perform most of their daily activities. Back pain is a symptom that can be assessed postoperatively but does not solely define the overall function of the patient.
Limitations of this study include the inherent weaknesses of a retrospective analysis as well as a small study sample size. Nevertheless, the use of a propensity score-matched control group provides the best possible comparison group in retrospective studies by matching factors typically considered confounders and limiting selection bias, thus strengthening analytic results. Given the trend toward better results in all outcome measures in the noninfected group, it is possible that a larger study sample size with greater power would reveal additional differences between patients with and without postoperative wound infections. Furthermore, the study does not include routine evaluation of fusion status with computed tomographic scan, raising the possibility that undetected nonunion might confound the postoperative outcome scores. Nonetheless, there was no evidence of hardware failure or revision surgery for pseudoarthrosis at final follow-up.
CONCLUSION
This study demonstrates that patients with acute postoperative deep wound infections after instrumented lumbar spinal fusion have improved outcome measures after surgery but have greater back pain and a decreased probability of achieving MCID for ODI than patients without infection 2 years after surgery.
Key Points
Thirty patients who underwent instrumented lumbar spinal fusion complicated by an acute (≤3 mo) postoperative deep wound infection necessitating irrigation and debridement were identified and matched to a noninfected control group, which was identified using propensity score-matching technique.
Consistent with the propensity score-matching technique, there were no significant demographic or surgical differences between the 2 groups at baseline.
ODI, PCS, and back and leg pain scores were statistically significantly better at 2 years postoperative than at baseline in both groups.
At 2 years postoperative, the infection group had statistically significantly worse back pain scores compared with the control group.
A greater proportion of patients in the control group achieved MCID for ODI than those in the infection group.