The purpose of this study was twofold. First, we analyzed a large series of posterior cervical spine operations performed by a single surgeon to identify factors that may be associated with a higher rate of acute postoperative wound infections. Second, we evaluated the success of a two-step process initiated in an effort to reduce the rate of acute postoperative deep infections.
Materials and Methods
This is a retrospective review of a consecutive series of 1001 all-posterior cervical spine operations performed at a single institution by the senior author (K.D.R.) from 1995 to 2010. Data collection was done by retrospectively reviewing a prospectively collected database of all procedures performed at one academic institution, as well as medical records and surgical case logs. Posterior cervical procedures were selected as a personal internal review revealed an unacceptably higher rate of infections in posterior cervical procedures (2%) compared with anterior cervical procedures (<0.5%). This observation is consistent with previous studies, which, while not entirely specific to the cervical spine, have described a higher rate of postoperative infections in posterior compared with anterior approaches for spinal surgery1,2.
Data collected were grouped into several categories: patient demographics, medical comorbidities, diagnosis, and surgical data. Patient demographic data included patient age, sex, body mass index (BMI), smoking history, and alcohol abuse. Medical comorbidities included diabetes, rheumatoid arthritis, renal failure, chronic obstructive pulmonary disease, and coronary artery disease. Diagnosis was defined as the indication for surgical intervention: degenerative disease, traumatic injury, tumor, congenital anomaly, spinal cord signal change on magnetic resonance imaging (MRI), and/or spinal deformity. Lastly, surgical data included the date of surgery, operative time, procedure performed (foraminotomy, laminectomy, laminoplasty, arthrodesis, instrumentation, and/or osteotomies), primary or revision procedure, number of vertebral levels involved with the surgery, and length of follow-up (minimum of three months). Postoperative acute wound infections (involving the suprafascial and/or subfascial space) were defined as infections requiring a formal irrigation and debridement in the operating room. All infections were confirmed with intraoperative cultures taken at the washout procedure. All data were collected and analyzed by three independent reviewers (J.M.P., J.R.P., and A.H.H.), not involved in the surgical treatment.
Patients were categorized into three separate groups according to the measures used to prevent infection: Group C (the control group), Group AD (alcohol foam and drain), and Group VAD (vancomycin, alcohol foam, and drain). In Group C, which included 483 patients who had the procedure performed prior to August 2004, only standard perioperative intravenous antibiotics were administered for wound prophylaxis. In Group AD, which included 323 nonallergic patients, standard perioperative intravenous antibiotics (cephalosporins) were used, as in Group C, with the additional step of a preliminary preparation of both the surgical site and plastic drapes surrounding the site with alcohol foam. In addition, a separate superficial drain was placed at the conclusion of the procedure in obese patients whose subcutaneous fat depth was >2 cm. In Group VAD, which included 195 patients, 500 mg of vancomycin powder was added to the wound at the end of the procedure in addition to the protocol followed for Group AD.
Approval for this study was obtained from the institutional review board Human Studies Committee.
Association between postoperative surgical infection and potential risk factors were analyzed with use of the Fisher exact and chi-square tests for categorical variables. A p value of 0.05 defined significance. These tests were also utilized to determine differences in infection rates between Group C and Groups AD and VAD. The calculation of an odds ratio and 95% confidence intervals was also performed. Multivariate logistic regression analysis was used to identify independent risk factors for postoperative infection. Variables eligible for inclusion in the multivariate models included those reported to be associated with an increased risk of surgical site infection in the literature, those with clinical and/or biologic plausibility, and those with p values of <0.20 in the univariate analysis1.
Source of Funding
There was no external funding for this investigation.
In Group C, the control group of 483 patients who received only perioperative intravenous antibiotics, the prevalence of acute postoperative deep infection was 1.86% (nine patients). The most common pathogen identified was methicillin-resistant Staphylococcus aureus (MRSA). In Group C, a significantly higher rate of acute infections was identified in patients with an active smoking history (p = 0.002), rheumatoid arthritis (p = 0.027), and a BMI of ≥30 kg/m2 (p = 0.005). Diabetes was associated with a trend toward a higher infection rate, but this did not reach significance (p = 0.2). Factors related to the surgical procedure not associated with higher infection rates included arthrodesis (p = 0.7), revision surgery (p = 0.5), instrumented procedures (p = 0.36), and procedures involving four vertebral levels or more (p = 0.21). Surgical time was significantly longer for the group with an infection (211 minutes) than for the group with no infection (145 minutes), p = 0.045 (Table I).
The results of the multivariate analysis to identify independent risk factors for acute postoperative infections after posterior cervical spine procedures are shown in Table II. A BMI of ≥30 kg/m2 and rheumatoid arthritis had the strongest association with acute postoperative infections, with an adjusted odds ratio of 4.1 (95% confidence interval [CI], 1.5 to 7.7) and 4.0 (95% CI, 1.4 to 7.9), respectively, after we controlled for other variables in the model. Smoking also remained independently associated with an increased risk of acute postoperative infection, with an adjusted odds ratio of 2.6 (95% CI, 1.0 to 7.1). Similar to the univariate analysis, the remaining parameters of diabetes mellitus, arthrodeses, revisions, instrumented procedures, and procedures involving four vertebral levels or more did not demonstrate a significantly higher infection rate (Table II).
In Group AD (alcohol foam and drain), one (0.3%) of 323 patients had an acute postoperative deep infection, representing a significant decrease compared with Group C (p = 0.047). There were no acute postoperative infections in the 195 patients in Group VAD (vancomycin, alcohol foam, and drain), also representing a significant decline compared with Group C (p = 0.048). Analysis of the predictors of infection was not performed for Groups AD and VAD because only one infection and no infection, respectively, developed in each group. There were no complications in Groups AD and VAD related to preliminary preparation with the use of alcohol foam, a second superficial drain in obese patients, and/or application of intrawound vancomycin powder prior to wound closure.
The three groups were relatively homogeneous with respect to patient demographics, comorbidities, diagnoses, and surgical procedures performed. Specifically, there were no significant differences, on the basis of the numbers available, among the three groups with regard to the percentage of patients who actively used tobacco, who had rheumatoid arthritis, and/or who had a BMI of ≥30 kg/m2, which were factors found to be associated with a higher rate of infection in the control Group C (Table III). Compared with Group C, Groups AD and VAD had significantly older patients (p = 0.002 and p = 0.003, respectively) as well as more procedures involving four or more vertebral levels (p = 0.001 for both groups). Group AD also had a higher rate of revision procedures (p = 0.02) compared with Group C. Group C had significantly more revision procedures (p = 0.02) than Group VAD. All other parameters, including surgical time, showed no significant differences among the three groups. Specific multivariate analysis of the three groups (C, AD, and VAD) did not reveal any significant differences. Further, multivariate analysis demonstrated that the infection rate differences among the three groups were not significantly confounded by the other factors evaluated (that is, rheumatoid arthritis, a BMI of ≥30 kg/m2, tobacco use, and others).
The relatively standard regimen of perioperative intravenous cephalosporins has been shown to be effective against less than half of the staphylococcal organisms found in hospitals6-8. Concomitantly, intravenous vancomycin has not been shown to decrease wound infection rates compared with intravenous cephalosporins9. Intravenous vancomycin has been associated with hypotension and renal toxicity, as well as an increased risk of resistant organisms in the oropharyngeal, respiratory, and genitourinary tracts10.
Recently, one of us (F.A.S.) reported a significant reduction in acute postoperative infections for instrumented thoracolumbar arthrodeses with the application of vancomycin powder into the wound prior to closure5. This resulted in antibiotic levels of up to 1000-fold higher than the mean inhibitory concentration for MRSA and coagulase-negative Staphylococcus aureus in the local environment5,11-13. Vancomycin was selected for its efficacy against common pathogens, but also because of its pharmacokinetics, in which it is poorly absorbed from the wound11-13. This vital characteristic enables high local wound concentrations and predominantly undetectable serum concentrations5.
Factors associated with a significantly higher rate of acute postoperative deep infections in our patients receiving only perioperative intravenous antibiotics (Group C) were an active smoking history (p = 0.002), rheumatoid arthritis (p = 0.027), and a BMI of ≥30 kg/m2 (p = 0.005). Diabetes and procedures involving four vertebral levels or more were associated with a trend toward a higher infection rate (p = 0.2 and 0.21, respectively). Multivariate analysis demonstrated that a BMI of ≥30 kg/m2, rheumatoid arthritis, and smoking were independently associated with higher acute postoperative infections (adjusted odds ratio of 4.1, 4.0, and 2.6, respectively). Surgical time was significantly longer (211 versus 145 minutes) in the group with infection than in the group without infection (p = 0.045). MRSA was the most common pathogen for Group C.
After the transition to the Group-AD protocol, postoperative deep infection occurred after surgery in only one (0.3%) of 323 consecutive patients. Subsequently, there have been no postoperative infections after 195 consecutive operations, in which vancomycin had been added to the wound (Group VAD). Both Group AD and Group VAD demonstrated a significant reduction in postoperative infections (p = 0.047 and p = 0.048, respectively) compared with Group C. To eliminate selection bias, all posterior cervical spine surgical procedures that were performed by the senior author (K.D.R.) at one academic institution were analyzed from 1995 to 2010. All three groups maintained a relatively homogeneous patient population with regard to patient demographics, surgical procedures, surgical data, and operative times.
These prophylactic measures may also represent a potentially significant cost-savings, as the cost of treating deep surgical site infections can be as high as four times the cost of the initial surgery6,8,14,15. Conversely, the cost of 500 mg of vancomycin powder is approximately $12.0016. Alcohol foam is commonly used in the clinical setting and is safe for use in the operating room.
While both Groups AD and VAD demonstrated a significant decline in acute postoperative infection rates compared with Group C, no significant decline in infection rates was detected between Groups AD and VAD. Thus, one can argue that the Group-AD protocol alone may be sufficient to lower postoperative infection rates. Given the exceedingly low infection rates with both Groups AD and VAD, our study was underpowered to identify a significant difference between these two treatments. Our institution continues to prospectively collect infection data on all cervical surgical procedures to evaluate the efficacy of our current protocols.
We believe that the potential downsides to the application of intrawound vancomycin are minimal, if any, as no side effects were demonstrated nor were vancomycin-resistant organisms created from the use of vancomycin in the combined group of >1100 patients from the present investigation and the study by Sweet et al.5. Also, no specific complications were noted when vancomycin came in direct contact with the dura in either study. We are currently evaluating the effect of high-dose vancomycin on dural tissue.
A potential flaw in our investigation is that it is a retrospective study with the control patients having had the operation earlier in the career of the senior author (K.D.R.). As such, it is possible that improved surgical technique might be one of the reasons for the decrease in infections. The surgical times did not differ among the three groups. Further, nearly half of the infections in Group C (four of nine) occurred nine years into the senior author’s career, and a marked reduction in postoperative infections occurred in the year following the protocol change of Group AD. The surgical preparation, surgical technique, and operative environment did not change throughout the series, with the exception of the AD and VAD protocols. Lastly, we examined the records of all total knee arthroplasties done at our institution to determine if secular changes may have resulted in significant declines of surgical site infections. The acute postoperative infection rates for total knee replacements have remained stable, varying between 0.5% and 2.1% per year, in the time frame evaluated for this study.
In this study, we present a protocol that significantly reduced the postoperative infection rate in posterior cervical procedures to zero in the previous two years. We recommend the use of a preliminary preparation with alcohol foam, multiple drains in obese patients, and vancomycin powder placed locally in the wound for nonallergic patients undergoing posterior cervical surgery.
NOTE: The authors thank John Gaughan, PhD, Department of Biostatistics, Temple University School of Medicine, and Jahangir Asghar, MD, Department of Orthopaedics, Miami Children’s Hospital, for aid in preparation of this manuscript.
Investigation performed at the Cervical Spine Service of Washington University Orthopaedics, St. Louis, Missouri
1. Olsen MA Nepple JJ Riew KD Lenke LG Bridwell KH Mayfield J Fraser VJ. Risk factors for surgical site infection following orthopaedic spinal operations. J Bone Joint Surg Am. 2008 Jan;90(1):62–9.
2. Olsen MA Mayfield J Lauryssen C Polish LB Jones M Vest J Fraser VJ. Risk factors for surgical site infection in spinal surgery. J Neurosurg. 2003 Mar;98 (2)(Suppl):149–55.
3. Zeidman SM Ducker TB Raycroft J. Trends and complications in cervical spine surgery: 1989-1993. J Spinal Disord. 1997 Dec;10(6):523–6.
4. Watanabe M Sakai D Matsuyama D Yamamoto Y Sato M Mochida J. Risk factors for surgical site infection following spine surgery: efficacy of intraoperative saline irrigation. J Neurosurg Spine. 2010 May;12(5):540–6.
5. Sweet FA Roh M Sliva C. Intrawound application of vancomycin for prophylaxis in instrumented thoracolumbar fusions: efficacy, drug levels, and patient outcomes. Spine (Phila Pa 1976). 2011 Nov 15;36(24):2084–8.
6. Klevens RM Edwards JR Richards CL Jr Horan TC Gaynes RP Pollock DA Cardo DM. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep. 2007 Mar-Apr;122(2):160–6.
7. Klevens RM Morrison MA Nadle J Petit S Gershman K Ray S Harrison LH Lynfield R Dumyati G Townes JM Craig AS Zell ER Fosheim GE McDougal LK Carey RB Fridkin SK; Active Bacterial Core surveillance (ABCs) MRSA Investigators. Invasive methicillin-resistant Staphylococcus aureus infections in the United States. JAMA. 2007 Oct 17;298(15):1763–71.
8. Noskin GA Rubin RJ Schentag JJ Kluytmans J Hedblom EC Jacobson C Smulders M Gemmen E Bharmal M. National trends in Staphylococcus aureus infection rates: impact on economic burden and mortality over a 6-year period (1998-2003). Clin Infect Dis. 2007 Nov 1;45(9):1132–40.
9. Finkelstein R Rabino G Mashiah T Bar-El Y Adler Z Kertzman V Cohen O Milo S. Vancomycin versus cefazolin prophylaxis for cardiac surgery in the setting of a high prevalence of methicillin-resistant staphylococcal infections. J Thorac Cardiovasc Surg. 2002 Feb;123(2):326–32.
10. Moise PA Smyth DS El-Fawal N Robinson DA Holden PN Forrest A Sakoulas G. Microbiological effects of prior vancomycin use in patients with methicillin-resistant Staphylococcus aureus bacteraemia. J Antimicrob Chemother. 2008 Jan;61(1):85–90.
11. Wu P Grainger DW. Drug/device combinations for local drug therapies and infection prophylaxis. Biomaterials. 2006 Apr;27(11):2450–67.
12. Chilukuri DM Shah JC. Local delivery of vancomycin for the prophylaxis of prosthetic device-related infections. Pharm Res. 2005 Apr;22(4):563–72.
13. Murray BE. Vancomycin-resistant enterococcal infections. N Engl J Med. 2000 Mar 9;342(10):710–21.
14. Calderone RR Garland DE Capen DA Oster H. Cost of medical care for postoperative spinal infections. Orthop Clin North Am. 1996 Jan;27(1):171–82.
15. Sasso RC Garrido BJ. Postoperative spinal wound infections. J Am Acad Orthop Surg. 2008 Jun;16(6):330–7.
16. McNamara DR Steckelberg JM. Vancomycin. J Am Acad Orthop Surg. 2005 Mar-Apr;13(2):89–92.
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.Copyright 2013 by The Journal of Bone and Joint Surgery, Incorporated