Risk Factors for Surgical Site Infection Complicating Laminectomy
Objective. To examine risk factors for surgical site infection (SSI) following spinal surgery and to analyze the associations between a surgeon's years of operating experience and surgical specialty and patients' SSI risk.
Design. Case‐control study.
Setting. A tertiary care facility and a community hospital in Durham, North Carolina.
Patients. Each case patient who developed an SSI complicating laminectomy was matched with 2 noninfected control patients by hospital, year of surgery, and National Nosocomial Infection Surveillance System risk index score.
Results. Forty‐one case patients with SSI complicating laminectomy and 82 matched control patients were analyzed. Nonwhite race, diabetes and an elevated body mass index (BMI) were more common among case patients than among control patients. Subjects with a BMI greater than 35 were more likely to undergo a prolonged procedure, compared with case patients who had a BMI of 35 or less. The SSI rate for patients operated on by neurosurgeons was 28%, compared with 43% for patients operated on by orthopedic surgeons (odds ratio [OR], 0.5; 
). The number of years of operating experience were not associated with SSI risk. Multivariate analysis revealed diabetes (OR, 4.2 [95% confidence interval {CI}, 1.1‐16.3]; 
), BMI greater than 35 (OR, 7.1 [95% CI, 1.8‐28.3]; 
), and laminectomy at a level other than cervical (OR, 6.7 [95% CI, 1.4‐33.3]; 
) as independent risk factors for SSI following laminectomy.
Conclusion. Diabetes, obesity, and laminectomy at a level other than cervical are independent risk factors for SSI following laminectomy. Preoperative weight loss and tight perioperative control of blood glucose levels may reduce the risk of SSI in laminectomy patients.
Received December 4, 2006; accepted March 15, 2007; electronically published June 28, 2007.
Laminectomy is the method of choice for decompression of the neural elements and maintenance of bony stability in patients with spinal stenosis, and it improves quality of life for patients with spinal stenosis.1,2 The incidence of postoperative surgical site infection (SSI) after spinal surgery is low, ranging from less than 1% for diskectomy to 6% for spinal fusion with instrumentation.3‐6 However, higher rates of SSI following laminectomy have been reported,7,8 especially in high‐risk patients with comorbid conditions, such as diabetes, and in patients who undergo surgery of long duration.9 According to a recent report from the National Nosocomial Infections Surveillance (NNIS) System, the pooled mean rate of SSI complicating laminectomy was 0.88‐2.46 infections per 100 operations.10
Postlaminectomy SSI was first described in 1947,7,11 and it routinely results in rehospitalization, multiple surgical procedures, and prolonged antibiotic therapy.12 Previous studies have delineated a wide array of risk factors for SSI following laminectomy. These studies found that the risk of SSI was increased by the following factors: advanced age, malnutrition, extended preoperative hospitalization, previous spinal surgery, obesity, diabetes, smoking, immunosuppression, prolonged duration of surgery, the degree of intraoperative tissue damage, wound seroma, and infection at remote body sites.4,12‐16 The variability of risk factors in different studies is striking and may relate to the case mix in selected practices or surgical centers. For example, the authors of a recent study of SSI following spinal fusion and instrumentation reported the following risk factors: nonambulatory status, wound hematoma, bladder and bowel incontinence, and extension of surgery into the sacral region.5 However, previous investigators have not examined the impact of a surgeon's years of operating experience on the patient's risk of infection following laminectomy. The aim of this case‐control study was to delineate risk factors for SSI following laminectomy in patients from a tertiary care hospital and an affiliated community hospital and to examine possible associations between surgeons' years of operating experience and surgical specialty and patients' risk for SSI following laminectomy.
Methods
Setting
This study included patients from 2 hospitals in Durham, North Carolina: Duke University Medical Center (DUMC), a 900‐bed hospital providing primary, secondary, and tertiary care, and Durham Regional Hospital (DRH), a 350‐bed community hospital with a teaching affiliation with DUMC. Approval for this study was obtained from the institutional review boards at both hospitals.
Definition and Identification of Case Patients
A case patient was defined, according to NNIS criteria, as any patient who developed a superficial, deep, or organ space SSI within 30 days following laminectomy or within 1 year of surgery if spinal implants were inserted at the time of the surgical procedure.17 Case patients at DUMC underwent their operative procedure between July 1997 and December 2000. Case patients at DRH underwent laminectomy between June 1991 and June 2001. If a patient had more than 1 episode of infection, only the first episode was considered in our analysis.
Infection control practitioners at both institutions performed prospective inpatient and laboratory‐based surveillance on all postlaminectomy patients, including those patients readmitted for SSI. Orthopedic surgeons at DRH reported additional cases of postdischarge SSI via questionnaire, as described elsewhere.18
Control Patients
For each case patient, 2 noninfected matched control patients were selected from the infection control operative database. Control and case patients were matched by calendar year of surgery,18 NNIS risk index score,19,20 and hospital where the surgery was performed. Case and control patients were not matched by the level of laminectomy performed. If numerous potential control patients existed for a case patient, 2 were selected using a random numbers chart.
Data Collection
Clinical and demographic variables. Medical records were reviewed for all study patients. Data collected included demographic data (age, sex, and race), body mass index (BMI), and information on medical comorbidities and therapies (diabetes, malignancy, immunosuppressive therapy, chemotherapy, and/or radiation therapy).
Perioperative variables. Perioperative data were collected from the infection control operative databases and the patient's medical record, including duration of initial (surgical) hospital stay, fasting blood glucose level on the day of surgery, type of surgery (including whether it involved instrumentation), duration of surgery, specialist qualifications of the surgeon (obtained from the hospital’s credentialing department), type and timing of antibiotic prophylaxis, American Society of Anesthesiologists (ASA) score, and NNIS risk index score. Operating experience was measured by the surgeons' years of operating experience and the median of the total number of laminectomies performed per year by each surgeon during the study period; this information was obtained from the infection control surgical databases. The timing of antibiotic prophylaxis administration was defined as correct if the antimicrobial was administered within 2 hours before first incision.21
Statistical Analysis
Categorical variables were expressed as numbers and percentages. Continuous variables were expressed as the median and interquartile range. Categorical variables were compared using the χ2 and Fisher exact tests. For continuous variables, median differences between case patients and their matched control patients were compared using the Wilcoxon signed‐rank test. Multivariate analysis was performed using conditional logistic regression. Variables were considered for inclusion in the multivariate model if they were considered to be important predictors of SSI by the study investigators or if the variables had a bivariate association with SSI (a P value less than .10). A 2‐sided P value of .05 or less was considered significant for all statistical tests. Data were maintained and analyzed using Excel (Microsoft), Epi Info (centers for Disease Control and Prevention), and SAS statistical software, version 8.1 (SAS Institute).
Results
Over the study period, 792 laminectomies were performed at DUMC, and 5,573 laminectomies were performed at DRH, with an overall SSI rate of 1.1% at DUMC and 0.9% at DRH. A total of 41 case patients with SSI (34 at DRH and 7 at DUMC) and 82 matched control patients were analyzed. The major pathogens causing SSI in this study were Staphylococcus aureus (in 17 cases), Staphylococcus epidermidis (9), Escherichia coli (3), and Pseudomonas aeruginosa (1). No study patients were nonambulatory or immunosuppressed.
There were no significant differences in the mean age and sex distribution of case and control patients. More case patients than control patients had nonwhite race (
). The presence of diabetes was more common among case patients than control patients, although this difference failed to reach statistical significance (
). However, compared with control patients, case patients had significantly higher median fasting blood glucose levels (
) and a significantly higher median BMI (
) (Table 1). In addition, compared with control patients, there were significantly more case patients with a BMI greater than 35 (12 of 41 case patients vs 5 of 82 control patients; odds ratio [OR], 6.37 [
]). Surgical factors, such as previous spinal surgery, duration of surgery, and insertion of implants, were not significantly different between case and control patients. Perioperative antibiotic prophylaxis was given to similar percentages of case and control patients (71% and 66%, respectively). Approximately half of both the case and control patients received antibiotic prophylaxis within 2 hours before the surgical procedure.
Comparable numbers of case and control patients underwent lumbar‐ or lumbosacral‐level laminectomy, (33 of 41 case patients vs 59 of 82 control patients; 
). More case patients than control patients underwent thoracic‐level laminectomy (5 of 41 vs 2 of 82; 
). In addition, significantly fewer case patients than control patients underwent cervical‐level laminectomy (3 of 41 vs 21 of 82; 
).
The surgical duration exceeded the NNIS cutoff point of 2 hours10 for 24 (59%) of 41 case patients and for 40 (49%) of 82 control patients (
). However, 15 (88%) of 17 patients with a BMI greater than 35 had a procedure duration exceeding 2 hours, compared with 47 (44%) of 106 patients with a BMI of 35 or less (
).
A total of 23 different surgeons performed operations on study patients; 14 were orthopedic surgeons, and the remaining 9 were neurosurgeons. Neurosurgeons operated on 22 (54%) of the case patients and 57 (70%) of the control patients, whereas orthopedic surgeons operated on 19 (46%) of the case patients and 25 (30%) of the control patients (P not significant) (Table 1).
Compared with surgeons who operated on control patients, surgeons who operated on case patients performed fewer total laminectomies per year but had a comparable number of years of operating experience (Table 1). A quartile analysis of procedures performed by surgeons per year found that, compared with the patients of surgeons who performed fewer than 42 laminectomies per year (the reference category), the patients of surgeons who performed 83‐122 laminectomies per year tended to have fewer SSIs (odds ratio [OR], 0.27 [95% confidence interval {CI}, 0.07‐1.08]; 
). Multivariate analysis revealed that diabetes (OR, 4.2 [95% CI, 1.1‐16.3]; 
), BMI greater than 35 (OR, 7.1 [95% CI, 1.8‐28.3]; 
), and laminectomy at a level other than cervical (OR, 6.7 [95% CI, 1.4‐33.3]; 
) were independent risk factors for SSI following laminectomy (Table 2).
Discussion
To our knowledge, there have been only 5 previous studies of SSI complicating spinal surgery; just 3 of these were case‐control studies, and only 1 included a multivariate analysis and involved as many patients as our study does (Table 3).5,12,16,22,23 Despite the differences between our study and those published previously, several important findings were present in all relevant studies, including our own.
The first of these findings was that obesity was a risk factor for the occurrence of SSI after laminectomy in our study and 3 preceding studies.4,16,23 In our study and in the study by Olsen et al.,23 morbidly obese patients had a dramatically (more than 7 times) higher risk for developing SSI, compared with patients who were not morbidly obese. One possible reason for this increased risk of SSI in obese patients is that adipose tissue is poorly vascularized. This reduced vascularity may result in local areas of incisional hypoxia, which in turn may impede bacterial phagocytosis by neutrophils.24‐26 Insulin resistance, such as that which occurs in the obesity‐related metabolic syndrome, may also increase the risk of postoperative SSI in obese patients.27 In addition, inherent technical difficulties related to performing spinal surgery for obese patients may result in prolonged operations and tissue necrosis due to the need for more forceful retraction during surgery.26
Prolonged procedure duration is a well‐established risk factor for SSI complicating laminectomy,4,16,19 and it might account for the increased risk of SSI among obese patients observed in our study and in previous studies. Although there was no significant difference in mean procedure duration between case and control patients in our study, we found that longer operations were more common among morbidly obese case patients. All but 2 subjects with a BMI greater than 35 had a procedure duration exceeding 2 hours, compared with only 44% of subjects with a BMI of 35 or less.
Other authors have shown that patients with diabetes are more likely to develop SSI and have poor clinical and functional outcomes after laminectomy, compared with their nondiabetic counterparts.28‐30 In addition, diabetes and/or elevated blood glucose levels are risk factors for SSI following many types of surgery, including cardiovascular surgery.31‐33 In our study, both the presence of diabetes and the presence of elevated preoperative fasting blood glucose levels were more common among case patients than among matched control patients in univariate analysis, and diabetes was found to be an independent risk factor for SSI on multivariate analysis.
The authors of 3 prior studies reported black race as a risk factor for nosocomial infections, including SSI 34‐36 and hemodialysis‐related septicemia.37 In our study, bivariate analysis showed that nonwhite patients were more than twice as likely as white patients to develop an SSI, but this was not confirmed on multivariate analysis. These findings were not related to differences in the prevalence of diabetes or obesity among white and black patient groups. Similar proportions of whites and blacks had diabetes (12 of 95 vs 2 of 25; 
), and a similar percentage had a BMI greater than 35 (45% of blacks vs 37% of whites, 
).
The SSI rates following laminectomy at our 2 institutions were within the range reported by the NNIS System hospitals.10 Similar to studies from other centers, our study found that the pathogen most commonly causing SSI was Staphylococcus aureus.12 Previous studies have found a posterior operative approach to be a risk factor for SSI after laminectomy.16,23 Similarly, we identified laminectomy at a level other than cervical (always performed via a posterior approach) as an independent risk factor for SSI. In contrast to previous studies, we did not observe that older age, previous spinal surgery, or spinal instrumentation were risk factors for SSI in our cohort.4,13 However, some investigators in previous studies examined study populations that were substantially different from our cohort of patients (Table 3). For example, all case patients in the study by Perry et al.5 had significant spinal deformity, and gram‐negative bacteria caused the majority of infections. Thus the generalizability of the results from our study and other studies is highly dependent on the type of patients studied.
A previous large study of patients who underwent lumbar laminectomy in 1980 found that neurosurgeons performed 60% of the procedures and orthopedic surgeons performed 40% and that the specialty of the surgeon was not a factor in determining the risk of complications after surgery.6 In addition, in a recent case‐control study of risk factors for spinal SSI, the researchers found no difference in the SSI rate between procedures performed by a neurosurgeon and those performed by an orthopedic surgeon.12 The results of our study support these findings about surgical specialty. In addition, although operating experience (measured by the number of procedures performed) has been shown to reduce both the procedure duration and rates of complication significantly,38 we failed to identify surgeons' experience as a risk factor for SSI following laminectomy, although there was a trend towards significance.
Our study has several limitations. The primary limitation was a lack of statistical power. The low rate of SSI at the study hospitals made it impossible to adequately examine all variables as risk factors for SSI. Numerous variables that may have contributed to infection risk, including underlying diagnosis, history of smoking, type of anesthesia used, and complications, such as dural tears, were not included in this study. Additionally, other factors, such as referral status or patient selection factors (which are also too difficult to examine), could have biased the outcomes of procedures performed by neurosurgeons or orthopedic surgeons. We also acknowledge that at least a portion of these operative procedures were performed by medical resident staff under the supervision of neurosurgeons or orthopedic surgeons, which alters the conclusions which we can make about surgical specialty and surgical experience.
Despite the limitations listed above, our findings have both clinical and epidemiological utility. If individual surgeons have disproportionate numbers of obese or diabetic patients, the rate of SSI among those patients may reflect patient‐specific factors, rather than surgeon‐specific factors. Such patient‐specific factors are not included in the NNIS System risk index,19 but ought to be included in future studies of models to stratify risk of SSI preoperatively. In addition, the relationships between morbid obesity, diabetes, and risk of postoperative infection after this predominantly elective procedure should serve to emphasize the need for preoperative weight loss and aggressive perioperative control of blood glucose levels among laminectomy patients.
Acknowledgments
We would like to acknowledge the contributions of Jane P. Briggs, RN, and Sharon L. Trivette, RN, in the maintenance of the infection control operative databases.
Financial support. K.S.K. was supported by grant K23 AG23621‐01A1 from the National Institute of Aging.
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article.
References
- 1. Yuan PS, Booth RE Jr, Albert TJ. Nonsurgical and surgical management of lumbar spinal stenosis. Instr Course Lect 2005; 54:303‐312.
- 2. Caliandro P, Aulisa L, Padua R, et al. Quality of life, clinical and neurophysiological picture in patients operated on for lumbar stenosis. Acta Neurochir 2005; 92 (suppl):143‐146.
- 3. Lindholm TS, Pylkkanen P. Discitis following removal of intervertebral disc. Spine 1982; 7:618‐622.
- 4. Massie JB, Heller JG, Abitbol JJ, McPherson D, Garfin SR. Postoperative posterior spinal wound infections. Clin Orthop Relat Res 1992; 284: 99‐108.
- 5. Perry JW, Montgomerie JZ, Swank S, Gilmore DS, Maeder K. Wound infections following spinal fusion with posterior segmental spinal instrumentation. Clin Infect Dis 1997; 24:558‐561.
- 6. Ramirez LF, Thisted R. Complications and demographic characteristics of patients undergoing lumbar discectomy in community hospitals. Neurosurgery 1989; 25:226‐230.
- 7. Fernand R, Lee CK. Postlaminectomy disc space infection: a review of the literature and a report of three cases. Clin Orthop Relat Res 1986; 209:215‐218.
- 8. Luer MS, Hatton J. Appropriateness of antibiotic selection and use in laminectomy and microdiskectomy. Am J Hosp Pharm 1993; 50:667‐670.
- 9. Ehrenkranz NJ. Surgical wound infection occurrence in clean operations; risk stratification for interhospital comparisons. Am J Med 1981; 70:909‐914.
- 10. National Nosocomial Infections Surveillance (NNIS) System Report. Data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004; 32:470‐485.
- 11. Thibodeau AA. Closed space infection following removal of lumbar intervertebral disc. Clin Neurosurg 1966; 14:337‐360.
- 12. Apisarnthanarak A, Jones M, Waterman BM, Carroll CM, Bernardi R, Fraser VJ. Risk factors for spinal surgical‐site infections in a community hospital: a case‐control study. Infect Control Hosp Epidemiol 2003; 24:31‐36.
- 13. Capen DA, Calderone RR, Green A. Perioperative risk factors for wound infections after lower back fusions. Orthop Clin North Am 1996; 27:83‐86.
- 14. Klein JD, Hey LA, Yu CS, et al. Perioperative nutrition and postoperative complications in patients undergoing spinal surgery. Spine 1996; 21:2676‐2682.
- 15. Polk HC Jr, Simpson CJ, Simmons BP, Alexander JW. Guidelines for prevention of surgical wound infection. Arch Surg 1983; 118:1213‐1217.
- 16. Wimmer C, Gluch H, Franzreb M, Ogon M. Predisposing factors for infection in spine surgery: a survey of 850 spinal procedures. J Spinal Disord 1998; 11:124‐128.
- 17. Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG. CDC definitions of nosocomial surgical site infection, 1992: a modification of the CDC definitions of surgical wound infections. Infect Control Hosp Epidemiol 1992; 13:606‐608.
- 18. Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton DJ. The impact of surgical‐site infections in the 1990s: attributable mortality, excess length of hospitalization, and extra costs. Infect Control Hosp Epidemiol 1999; 20:725‐730.
- 19. Culver DH, Horan TC, Gaynes RP, et al. Surgical wound infection rates by wound class, operative procedure, and patient risk index. Am J Med 1991; 91(suppl 3B):S152‐S157.
- 20. Delgado‐Rodriguez M, Sillero‐Arenas M, Medina‐Cuadros M, Martinez‐Gallego G. Nosocomial infections in surgical patients: comparison of two measures of intrinsic patient risk. Infect Control Hosp Epidemiol 1997; 18:19‐23.
- 21. Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical‐wound infection. N Engl J Med 1992; 326:281‐286.
- 22. Kuo CH, Wang ST, Yu WK, Chang MC, Liu CL, Chen TH. Postoperative spinal deep wound infection: a six‐year review of 3230 selective procedures. J Chin Med Assoc 2004; 67:398‐402.
- 23. Olsen MA, Mayfield J, Lauryssen C, et al. Risk factors for surgical site infection in spinal surgery. J Neurosurg 2003; 98(suppl 2):149‐155.
- 24. Brashers VL, Davey SS. Alterations of pulmonary function. In: McCance KL, Huether SE, eds. Pathophysiology: The Biologic Basis for Disease in Adults and Children. 3rd ed. St. Louis: Mosby, 1998:1158‐1200.
- 25. Groszek DM. Promoting wound healing in the obese patient. AORN J 1982; 35:1132‐1138.
- 26. Wilson JA, Clark JJ. Obesity: impediment to postsurgical wound healing. Adv Skin Wound Care 2004; 17:426‐435.
- 27. Wisse BE. The inflammatory syndrome: the role of adipose tissue cytokines in metabolic disorders linked to obesity. J Am Soc Nephrol 2004; 15:2792‐2800.
- 28. Airaksinen O, Herno A, Turunen V, Saari T, Suomlainen O. Surgical outcome of 438 patients treated surgically for lumbar spinal stenosis. Spine 1997; 22:2278‐2283.
- 29. Kawaguchi Y, Matsui H, Ishihara H, Gejo R, Yasuda T. Surgical outcome of cervical expansive laminoplasty in patients with diabetes mellitus. Spine 2000; 25:551‐555.
- 30. Simpson JM, Silveri CP, Balderston RA, Simeone FA, An HS. The results of operations on the lumbar spine in patients who have diabetes mellitus. J Bone Joint Surg 1993; 75:1823‐1829.
- 31. Harrington G, Russo P, Spelman D, et al. Surgical‐site infection rates and risk factor analysis in coronary artery bypass graft surgery. Infect Control Hosp Epidemiol 2004; 25:472‐476.
- 32. Peersman G, Laskin R, Davis J, Peterson M. Infection in total knee replacement: a retrospective review of 6489 total knee replacements. Clin Orthop Relat Res 2001; 392:15‐23.
- 33. Wilson SJ, Sexton DJ. Elevated preoperative fasting serum glucose levels increase the risk of postoperative mediastinitis in patients undergoing open heart surgery. Infect Control Hosp Epidemiol 2003; 24:776‐778.
- 34. Harger JH, English DH. Selection of patients for antibiotic prophylaxis in cesarean sections. Am J Obstet Gynecol 1981; 141:752‐758.
- 35. Soltau JB, Rothman RF, Budenz DL, et al. Risk factors for glaucoma filtering bleb infections. Arch Ophthal 2000; 118:338‐342.
- 36. Watts DH, Lambert JS, Stiehm ER, et al. Complications according to mode of delivery among human immunodeficiency virus‐infected women with CD4 lymphocyte counts of ⩽500/μL. Am J Obstet Gynecol 2000; 183:100‐107.
- 37. Abbott KC, Agodoa LY. Etiology of bacterial septicemia in chronic dialysis patients in the United States. Clin Nephrol 2001; 56:124‐131.
- 38. Knight BP, Oral H, Chugh A, et al. Effects of operator experience on the outcome and duration of pulmonary vein isolation procedures for atrial fibrillation. Am J Cardiol 2003; 91:673‐677.