INTRODUCTION
In spine surgery, the development of a postoperative surgical site infection (SSI) is a serious complication that can lead to increased morbidity, the need for additional surgical procedures, mortality, and cost. Overall, there has been an increasing desire to eliminate SSIs through the determinants of risk factors, perioperative risk factor control, use of perioperative antibiotics, and the use of other perioperative disinfectants. The objective of the current manuscript was to provide a systematic review regarding the epidemiology of postoperative spine infections, diagnosis of infection, major risk factors, and preventative measures. More importantly, this manuscript aimed to answer the following clinical question: Does prolonged antibiotic use (> 24 hours) after spine surgery decrease the rate of postoperative SSIs?
Epidemiology and risk factors
Overall, in orthopedic spine surgery, the risk of postoperative SSIs has been found to range between 1%-10% based on the procedure performed and has been shown to usually occur within the first 90 days postoperatively[1-4]. Numerous studies have been performed to identify risk factors for postoperative infection. In terms of age and gender, an age greater than 60 years has been shown to be a risk factor for postoperative SSI[4,5]. In addition, male gender has also been shown to be a risk factor[6]. However, a more recent retrospective study has shown that while age and gender may predispose to postoperative non-SSIs such as urinary tract infections, pneumonia, or bacteremia, these variables were not shown to be risk factors for SSI[7].
Aside from age and gender, medical comorbidities have been examined in multiple studies to determine their relationship to postoperative SSI in spinal surgery. Of all the scales used to quantify medical comorbidities, the American Society of Anesthesiologists classification system is the most commonly used. Multiple studies have shown an increasing incidence of SSI with higher American Society of Anesthesiologists scores[7,8]. Further study showed that the presence of greater than 3 medical comorbidities has been shown to increase infection rates in posterior lumbar arthrodesis[5].
In addition to measuring the overall number of comorbidities and how this relates to increased infection risk, diabetes mellitus has garnered increased attention. Some studies found diabetes mellitus alone is an independent risk factor for postoperative spine SSIs[9,10]. Furthermore, it was found that hemoglobin AIc levels greater than 7.00 mg/dL were also found to increase postoperative infection risk[9,11]. However, as shown by Guzman et al[12], even patients with good glycemic control were still found to have an elevated risk for spinal SSI [odds ratio (OR) = 1.36] when compared to non-diabetic patients in the control group. Despite this baseline elevated risk for SSI, patients who had good glycemic control were still found to have a decreased OR of SSI when compared to uncontrolled diabetic patients (OR = 1.36 vs OR = 2.61), illustrating that glycemic control may provide some portion of protection from infection[12].
Another risk factors regarding spinal SSI are related to patient lifestyle choices, including obesity, nutritional status, and tobacco use. Obesity, as defined as a body mass index (BMI) greater than 30, has been shown to be a significant risk factor for postoperative SSI, independent of other associated medical comorbidities, with an OR averaging approximately 1.6[7,13-15]. The ORs have also been shown to increase by a factor of 1.27 for every 4 kg/m2 increase in BMI[13]. It was also noted that morbidly obese patients (BMI > 35 kg/m2) had a 7.7% chance of infection compared to non-obese patients (3%-4%)[13]. Poor nutritional status, determined by low serum albumin levels, has been shown to increase the rates of infection following revision spinal surgery[16]. Another modifiable risk factor for spine SSI is smoking cessation. A recent meta-analysis by Kong et al[17] conducted a systematic review and meta-analysis, revealed ORs of 1.26-1.40 for spinal SSI in smokers vs non-smokers, and approximately 10% of all infections were associated with smoking.
There are surgical risk factors, including surgical location, surgical approach, primary vs revision surgery, and the presence of associated trauma. Surgical location, anterior cervical surgeries have been shown to have a very low risk of infection when compared to surgeries in the lumbar spine (1%-3% vs 12%)[15,18]. Anterior approaches to the lumbar and cervical spines have been shown to decrease postoperative infection rates. One study has even stated that due to extremely low rates of infection, routine use of perioperative antibiotics during anterior procedures may not be required[19]. Spinal surgery in the setting of trauma has also been shown to increase SSI from approximately 2% in elective procedures to 10%-15% in trauma cases[20]. Furthermore, concomitant spinal cord injury has been shown to increase the risk of infection from 5% in trauma cases without spinal cord injury to up to 41% of trauma cases with complete spinal cord injury[20]. This increased rate of infection in the setting of spinal cord injury could be due to the high rates of non-SSIs (urinary tract infections, pneumonias, and decubitus ulcers) that occur at high rates within the perioperative time frame[21]. Finally, increased rates of infection in revision spine surgery and with multilevel fusions have been reported and related to increased tissue injury during surgery and prolonged operative times[22]. There are emerging data suggesting that blood transfusion is associated with increased risk of post-operative infections[23,24].
Microbiology of postoperative spine infections
In order to prevent postoperative spinal SSI, a strong understanding of the common organisms that cause these infections is a prerequisite. The first aspect of an infection that needs to be elicited is the ratio of mono vs polymicrobial infections. In postoperative spine SSI, 60%-80% of the positive cultures are monomicrobial, and 10%-30% are polymicrobial[25,26]. Staphylococcus aureus is the most commonly isolated bacterium, accounting for up to 45% of postoperative spine SSI[13,27-29]. However, high rates of Staphylococcus epidermidis, Staphylococcus propionibacterium, and Gram-negative infections have been reported as well[25,29,30]. Unfortunately, one common occurrence in postoperative SSI is culture-negative infections. This can occur up to 40% of the time, which makes prophylactic measures more important to prevent inadequate treatment or prolonged use of broad-spectrum antibiotics[26].
Diagnosis of postoperative spine SSI
A complete history, physical examination, laboratory studies, imaging modalities, and intraoperative cultures are used in the diagnosis of postoperative SSI. The most common signs and symptoms that have been reported are persistent low back pain, especially after a period of pain improvement, and wound complications, including wound drainage, abscess formation, and sinus tracts[25]. Interestingly, in one study, systemic symptoms such as pyrexia were relatively uncommon in postoperative spinal infections[31]. Laboratory studies such as complete blood count, sedimentation rate (ESR), and C-reactive protein (CRP) have become mainstays in identifying infection. Leukocytosis has been shown to be an insensitive test for the diagnosis of postoperative spine infections, and therefore, utilization of ESR and CRP levels has increased[31]. In cases of infection, both ESR and CRP can be markedly elevated. However, these biomarkers have been found to increase in response to spinal surgery alone. In a patient with baseline CRP < 10 mg/L, it has been shown that following posterolateral fusion, CRP levels can elevate to an average level of 173 mg/L[32]. These levels were then shown to normalize within 2 weeks. ESR was shown to follow a similar pattern with elevation up to postoperative day 5 with normalization within 42 days[32]. A new biomarker, presepsin, has been recently investigated as a possible indicator of infection. In one study, the author stated that this marker appears to normalize within 1 week after surgery, and elevated levels (> 300 pg/mL) could be used as a sign of SSI[33]. Imaging modalities such as radiographs, computed tomography imaging, and magnetic resonance imaging have been shown to have limited utility due to the effects of orthopedic implants on image quality. However, imaging studies are able to detect signs of infection, including early hardware loosening, loss of disk heights, and the presence of fluid collections[34]. Ultimately, diagnosis of a postoperative SSI can be achieved with intraoperative culture and tissue specimens.
ANTIBIOTIC PROPHYLAXIS
Systemic, re-dosing, and prolonged postoperative antibiotic course
Antibiotic prophylaxis for spine surgery can be divided into standard systemic as well as local intrawound use. Systemic antibiotics are administered preoperatively, intraoperatively, and postoperatively. Both the North American Spine Society and Center for Disease Control and Prevention recommended preoperative antibiotics to be dosed one hour prior to incision so that these antibiotics reach bactericidal concentration in the serum and tissues by the time skin incision is made[35-37]. Currently North American Spine Society recommends that all patients be given 1-2 g of cefazolin and 1-2 g of vancomycin preoperatively, dependent upon the patient’s weight and allergies[37]. The length of surgery has been associated with the SSI rate. For that reason, it is standard practice in prolonged cases to re-dose antibiotics at 4-hour intervals during surgery.
The duration of postoperative antibiotic prophylaxis is debated within the literature. Common practice involves 24 hours of postoperative antibiotics for most orthopedic procedures with hardware insertion[37]. In spine surgery, surgical drains are placed to prevent hematoma formation that may cause neurologic deficit[38]. For this reason, drains can remain in place for some time after surgery until the drain output diminishes. Carreon et al[39], in a retrospective study, recommended drain removal when their output was less than 30 mL in eight hours. Although drain use has demonstrated decreased infection rates by allowing for wound drainage, prolonged placement of drains leads to a higher rate of bacterial contamination[37-39].
Takemoto et al[38] conducted a prospective randomizes study looking at the utility of continuation of antibiotics until drain removal postoperatively vs 24-hour postoperative antibiotic prophylaxis despite a retained drain. The authors found that duration of drain use was an independent risk factor for infection; there were similar infection rates for those who had antibiotics until drain removal vs those who only had 24 hours of antibiotics in the postoperative setting. Shaffer et al[37] conducted an evidence-based clinical guideline study and found no significant decrease in infection incidence for those who had antibiotic administration for the duration of drain placement. Ohtori et al[36] conducted a comparative study of differing durations of antibiotics for spinal surgery, looking at short-term (2 days) vs long-term (9 days) of postoperative antibiotic use and found no significant differences in the occurrence of SSI within the first 2 weeks after surgery. Conducted a comparative study of differing durations of antibiotics for spinal surgery.
Pivazyan et al[40] conducted a single-center retrospective cohort study to evaluate the impact of prolonged prophylactic systemic antibiotics on SSI in degenerative spine surgery. Antibiotics were administered for the duration of the drain or for 24 hours postoperatively. Of 336 patients were included, 168 patients in the prolonged antibiotic group and 168 in the 24-hour postoperative group. The overall SSI was 5.36% (18/336). The SSI for the 24 hours postoperatively and prolonged antibiotic groups were 7.14% (12/168) and 3.57% (6/168), respectively. For the 24 hours postoperatively and prolonged antibiotic groups, the SSI for cervical [5.95% (5/84) vs 2.38% (2/84)] and lumbar [8.33% (7/84) vs 4.76% (4/84)]. They concluded that their study demonstrated a two-fold reduction of SSI with the implementation of a prolonged antibiotic regimen, and the benefit was demonstrated separately for both cervical and lumbar regions.
The reported benefits of prolonged antibiotic usage should be carefully weighed against the potential risks of adverse effects, microbial resistance, and microbiome disruption[41]. Increased duration of antibiotic prophylaxis has been associated with postoperative acute kidney injury and Clostridium difficile infection across multiple surgical specialties[41]. With regard to spine surgery specifically, the use of second-line antibiotics may contribute to the rise of resistant bacterial strains, such as methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus[42]. Long et al[43] conducted a retrospective hospital-registry study and found that Gram-positive bacteria are often responsible for cervical SSI, while Gram-negative and enteric bacteria are associated with lumbar spine infections. If standard prophylaxis does not target the appropriate bacteria, then this could play a role in the development of antibiotic resistance. It may prove beneficial to not only consider individual patient factors, but also the specific surgical site when choosing appropriate antimicrobial coverage[44].
From the aforementioned studies, while some retrospective studies support prolonged antibiotic use, prospective trials have not shown consistent benefit. It appears that there are no benefits of decreasing the SSI rate by using prolonged postoperative antibiotics. The current evidence-based guidelines from the North American Spine Society state that prolonged regimens may be considered when significant comorbidities include obesity, diabetes, neurologic deficits, incontinence, preoperative serum glucose of > 125 mg/dL or a postoperative serum glucose level of > 200 mg/dL, trauma, and prolonged multilevel instrumented surgery[37].
Local application of vancomycin powder
Vancomycin powder has a slow resorption rate, which provides an excellent local coverage against gram-positive bacteria, with no evidence of local or systemic toxicity. Vancomycin powder can be mixed with cancellous bone graft or placed directly in the surgical wound. The use of vancomycin powder has been frequently cited in the literature as a means for infection prophylaxis in surgical wounds[45]. Local vancomycin application during spine surgery has been met with mixed results.
Sweet et al[46] conducted a retrospective cohort study from a single institution of a consecutive series of spine surgery patients. There was a statistically significant reduction in infection rate in those treated with vancomycin powder and intravenous prophylaxis as compared to intravenous antibiotic prophylaxis alone. Lameire et al[47] conducted a systematic review of 26 comparative studies that assessed the efficacy of local vancomycin in spine surgery, and there was a statistically significant decrease in overall infection rate in fifteen of the publications. However, one study found an increased infection rate when vancomycin was used locally, and the majority of infections that occurred were due to Gram-negative bacteria[48]. The inconsistent outcomes seen with nonsystemic vancomycin prophylaxis may be due to its limited activity against Gram-negative and enteric bacteria. While there is a potential risk of the development of antibiotic resistance, previous work has found that intraoperative vancomycin powder does not increase the risk of methicillin-resistant Staphylococcus aureus following spine surgery[48-51].
Sun et al[50] conducted a database research study to investigate the association between intrawound vancomycin use and infection following open spine surgery. Their study included seven randomized controlled trials with 2235 patients; 1095 (49%) of them received intrawound vancomycin. They reported the overall rate of superficial and deep SSIs in the treatment group was 3.38%, compared with 4.08% in the control group. The overall rates of deep infection were 2.5% and 1.8% in the treatment and control groups, respectively, and the overall superficial infection rates were 0.9% and 1.8% in the treatment and control groups, respectively. They occluded that intraoperative vancomycin may not be associated with significantly decreased rates of superficial or deep SSI in patients undergoing open spine surgery.
Wang et al[51] conducted a prospective randomized case-controlled study at a single medical center to examine the effects of vancomycin powder mixed with local autogenous bone graft and bone substitute on deep SSIs and fusion rate. There were no deep SSIs in the vancomycin group and five in the control group. All five patients with deep SSIs had diabetes (100%). None of the patients with diabetes in the vancomycin group developed deep SSIs. At the final follow-up, functional outcomes and bone fusion rates were similar between the two groups. They conclude that vancomycin mixed with local autogenous bone graft and bone substitute maintains high vancomycin levels at the surgical site and appears safe and effective for preventing deep SSI in lumbar degenerative instrumented fusion surgery without affecting bony fusion, especially in diabetic patients. In the author’s practice, 1 gm of vancomycin powder is mixed with the bone graft before insertion in the posterolateral gutter and the disc space when performing interbody fusion. A prolonged postoperative antibiotic course is prescribed in high-risk patients that consists of 48 hours of IV cefazolin followed by 10 days of oral doxycycline taken twice daily.
CONCLUSION
SSI following spine surgery is associated with significant morbidity, length of stay, and increased cost. Intraoperative application of vancomycin and prolonged use of postoperative prophylactic antibiotics have been used by some surgeons to decrease the rate of postoperative SSI, especially in high-risk patients. Having said that, there is still no current standard of care, especially in patients with high risks of postoperative wound infection. When deciding on antibiotic surgical prophylaxis, one has to consider multiple factors, including patient and surgical factors. Patient factors include age, BMI, medical comorbidities such as diabetes, renal disease, history of previous infection, medications such as steroids, immunosuppressive drugs after organ transplantation, and inflammatory arthritis disease-modifying drugs. Surgical factors such as primary vs revision, approach posterior vs anterior or lateral, length of surgery, whether hardware was inserted vs decompression without instrumentation. Using local as well as prolonged postoperative systemic antibiotics may be considered in these patients to decrease the rate of SSI. Recommendation to extend postoperative prophylaxis antibiotic beyond 24 hours has to be weighed against the current global emphasis on rational antibiotic use, to decrease potential risks of antibiotic resistance. This highlights the need for risk-stratified antibiotic protocols and prospective randomized trials to evaluate the best recommendation for combined prolonged postoperative systemic and application of intrawound antibiotic to achieve optimum surgical wound prophylaxis without risking the development of antibiotic-resistant strains.
