Gram-Negative Surgical Site Infections After 989 Spinal Fusion Procedures: Associated Factors and the Role of Gram-Negative Prophylactic Antibiotic Coverage

INTRODUCTION

Surgical site infection (SSI) rate is an important indicator of health care quality in spine surgery.^1,2 Quality indicators such as SSI must account for patient- and procedure-specific characteristics allowing meaningful comparison between patient- and procedure-related factors.² SSI risk factors identification in spine surgery can help to design more effective preventative measures. Some patient-related factors such as diabetes, obesity, smoking history, alcohol abuse, anemia, coronary artery disease, and coagulopathy have been correlated to SSI.^1,3,4,5 In addition, some procedure-related factors such as prolonged operative time, posterior surgical approach, osteotomy, and blood transfusions have been associated to SSI.^1,5 Furthermore, sometimes risk factors vary from institution, country, or treated population.⁵

Spine surgery infections are most commonly monomicrobial in origin, and the most common pathogen isolated is Staphylococcus aureus.^3,6,7 On the other hand, gram-negative organisms represent a considerable percentage (30.5%) of all SSI.⁷ Certain procedural risk factors in spine surgery have shown association with specific pathogens. Abdul-Jabbar et al⁷ showed that spine surgery that involves the sacrum seems to be associated with increased gram-negative organisms, and polymicrobial infections of all gram-negative organisms in same study (61.6%) were cefazolin-resistant.⁷ The use of prophylactic antibiotics has shown a significant decrease in SSI in numerous studies. A meta-analysis by Barker⁸ found that postoperative spine infection rates could be reduced from 5.9% to 2.2% with the adequate use of prophylactic antibiotics. Despite following the current clinical guidelines on antibiotic prophylaxis in spine surgery, the rate of SSI remains to be 0.7% to 10%.⁹

Cephalosporin antibiotics such as cefazolin used within 2 hours of incision has shown to improve infection rates in orthopedic surgery.¹⁰ For uncomplicated spinal procedures, prophylaxis against gram-positive organisms has been reported to decrease SSI risk.^8,11 Cefazolin is the first-line agent used for prophylactic gram-positive coverage during spine surgery at our institution.

Current clinical guidelines suggest that patients undergoing complicated spine surgery and those with comorbidities may benefit from additional measures such as gram-negative coverage and/or intrawound vancomycin or gentamycin application.⁹ Aminoglycosides including tobramycin are often used as alternative prophylactic antibiotic coverage against gram-negative organisms. Many studies have evaluated the different prophylactic regimens in the context of spine surgery, yet the superiority of 1 specific type and/or route of administration has not been clearly defined.⁹

Service reports from the orthopedic department at our institution during 2013 revealed a significant increase on the SSI rate for spinal fusion (SF) surgeries compared with previous years.¹² In 2013, SF operations accounted for 50% of the total number of SSI in the orthopedic surgery department (numbers).¹ Our infection rate on the SF group exceeded the National Healthcare Safety Network benchmark.¹ In 2014, the SF SSI rate decreased by 2.5% compared with the National Healthcare Safety Network reference rate.¹³ These service report findings^1,13 led the infection prevention and control service to consider prophylaxis coverage against gram-negative bacteria in order to reduce SSI-SF rate. The purpose of this study was to dissect and identify all of those patient- and procedure-related factors that may be associated to SSI-SF in our institution.

MATERIALS AND METHODS

We retrospectively collected data on SSI-SF cases occurring in a Level 1 trauma center from January 2013 to September 2015. The cases that have been treated for SSI either exclusively with antibiotics or with antibiotics combined with surgery were collected by the infection prevention and control surveillance service. Ethical approval from the Institutional Research Ethics Board was received for this study.

RESULTS

A total of 989 spine fusion surgical procedures were performed at our institution during 2013–2015. Twenty-four cases (2.43%) were diagnosed with SSI; the demographic details of these 24 cases were included in Table 1. Malignancy, immunocompromisation, and smoking were the most frequent associated risk factors among our population.

The preoperative diagnoses of the 24 SSI cases were mainly degenerative disc disease (DDD; 29% [7 patients]), followed by tumors 25% (6 patients) (Figure 1). Interestingly, there were 2 infection-related diagnoses: spondylodiscitis and osteomyelitis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Preoperative diagnoses.

All of the SSI were cultured and analyzed, and the classes of organism cultured for each specimen were monomicrobial in 62.5% (predominantly gram-positive organisms, 73%), and polymicrobial being 37.5% (Figure 2). Bacteremia was found to be present in 3 of the 24 SSI patients (12.5%).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Class of organisms reported in surgical site infection cases.

Larger surgeries involving multiple spinal regions were most likely to become infected with thoraco-lumbo-pelvic having the highest incidence of infection of 9.1% (Table 2; Figure 3). Of the 24 SSI cases, 21 (87.5%) were done via a posterior approach, 3 were dual approach procedures, none were performed exclusively via an anterior approach, and all of the 24 cases had instrumentation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Surgical site infection incidence according to spinal region.

In terms of the time of the day during which the surgeries resulting in SSI were performed, 22 (92%) were performed between 6 am and 6 pm, and 2 were performed after hours between 6 pm and 6 am (both patients were polytrauma, required intensive care unit admission, and had unstable spine fractures).

A total of 21 of 24 SSI patients received some kind of antibiotic prophylaxis (Figure 4), 17 (71%) received cephazolin (Ancef), and the rest had combination of Ancef with either vancomycin or gentamycin. Of the 21 patients who received antibiotic prophylaxis, 19 (90%) received the initial dose prior to the first incision (up to 100 minutes before), and 2 received it postincision at the 2- and 3.5-hour marks. One of the 2 patients who received cephazolin after surgical incision, the antibiotic was delayed until specimen collection for the reason of spondylodiscitis. Of the SSI patients, 10 received blood transfusions, which represented 40% (Figure 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Antibiotic prophylaxis use.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Transfusions.

Data on the treatment methods for the SSI patients were gathered as well. It was determined that 19 patients (79%) required surgical debridement (incision and drainage), and the remaining 5 patients (21%) simply received a course of intravenous antibiotics (Figure 6). All patients were noted to have improved post SSI treatment. It is important to note that 4 of the 24 patients died during the time span of the study. However, their causes of death were not related directly to the SSI in itself.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Surgical site infection treatment methods.

DISCUSSION

The incidence of SSI at our center between January 2013 and September 2015 was 2.43%. This result is congruent and with the reported 0.7% to 12.0% incidence of SSI after adult spinal surgery reported in the literature.¹⁴

The preoperative diagnosis that was shown to have the highest number of SSI was DDD with 7 cases, and interestingly the diagnosis with the lowest number of SSI was “spinal infection” (osteomyelitis and spondylodiscitis) with 2 cases. With our total amount of patients with SSI being 24, the difference between these 2 groups (DDD and spinal infection) is insignificant (29% vs 8% respectively). In a previous study with a sample of patients greater than 6000, the diagnosis of neoplasm was identified as being a risk factor for SSI.¹⁵ This was associated to the overall immune and nutritional state of oncological patients.²⁰ In our series, we observed the same trend. Although, the fact that our sample was smaller may have affected the result leaving “oncological” diagnosis as the second most common group presenting with SSI in our study.

Other main risk factors that have been associated with higher incidences of SSI are surgeries involving 8 or more levels and surgeries involving the sacrum or the pelvis.¹⁵ This was also found in our results as the majority of SSI (9.1%) involved thoraco-lumbo-pelvic instrumentations and therefore, longer surgical time and greater surgical area exposed and estimated blood loss.

Of the 24 SSI cases, all of them required instrumentation thereby suggesting that the introduction of a foreign body such as instrumentation is a risk factor for SSI. This is consistent with previous studies that have demonstrated how the use of instrumentation is high risk factor associated with SSI.¹⁶

We hypothesized that nonpatient related factors such as surgeon or operating team variation during the night shift (6 pm to 6 am) could have led to an increased risk of SSI. However, our results did not show that the time of surgery is a remarkable risk for SSI as 22 of the 24 SSI cases were performed during day time hours. As expected, almost all the patients in the SSI group received antibiotic prophylaxis.

We know from the literature that using a cephalosporin antibiotic such as cefazolin within 2 hours of incision has shown to improve infection rates in orthopedic surgery.¹¹ Our data show that 19 of the 21 SSI patients who received prophylaxis (gram-positive coverage only) received it according to this recommendation. We then asked ourselves whether adequate therapeutic antibiotic levels were maintained throughout the entire duration of the procedures, especially for the longer cases. However, we assumed the subsequent doses were given. While collecting data, we realized that the only prophylaxis protocol for anesthesia is to document the time at which the initial antibiotic was given. Despite this difficulty, the results we would have gathered may not have affected our results. Previously reported evidence showed that patients with prolonged procedures that are re-dosed have a similar infection rate to those were there was not a “subsequent dose” given.¹⁷

In addition, as seen in Figure 4, more than half of the SSI cases (13 out of 24) were found to be infected with either exclusively gram-negative bacteria or a combination of gram-negative and gram-positive. It is worth noting that out of the 10 SSI patients who had spine surgery involving the sacrum, 7 of them had a polymicrobial infection, 1 had a gram-negative infection and the remaining 2 had a gram-positive infection. This is consistent with the study by Abdul-Jabbar et al,⁷ which showed that spine surgery involving the sacrum seems to be associated with increased gram-negative organisms and polymicrobial infections.⁸ This supports our idea that there may be a benefit of using gram-negative coverage in spinal procedures, especially those involving the sacrum.

In our study, we did not find any association to blood transfusions during the procedure and SSI. However, it has been previously reported by Koutsoumbelis et al¹⁸ that the use of packed red-blood cells for volume replacement could increase the risk for SSI following spinal procedures, but more specifically for posterior lumbar instrumented arthrodesis.

In terms of the specific SSI treatment, 19 out of 24 patients (80%) required surgical debridement (incision and drainage), and the remaining 5 (20%) patients were treated only with intravenous antibiotics. The fact that 20% of the patients responded to medical management alone is interesting. However, there are no studies that randomized the modalities of treatment. This may be a factor that is still dependent on the staff surgeon in charge of the patient most probably implemented on a case by case scenario.

It is important to note that the 4 patients who died during the time span of the study did not die due to the SSI in itself. It is well documented that the SSI-related mortality rate among spine surgical patients ranged from 1.1% to 2.3%.¹⁹ Given these low percentages and our small sample of 24 SSI cases, it is not surprising that the number of SSI-related mortalities is 0.

LIMITATIONS

The results presented here were obtained from a retrospective analysis of prospectively collected data. Data extraction from charts should be interpreted carefully because human systematic errors can occur; for example, failure to state if a second antibiotic dose was given or just omitted. Also, retrospective cohort studies require large sample sizes especially if outcomes are rare. Our study sample of 24 cases was too low to run statistical analysis and conclude significant findings. This is a descriptive analysis and larger randomized studies are necessary to confirm our results.

CONCLUSION

The incidence of SSI at our center after reviewing 989 patients between January 2013 and September 2015 was 2.43%. SSI were associated to SFs that required multilevel constructs of instrumentation via a posterior approach. Polymicrobial infections were identified to be present when sacral levels were involved. Blood transfusions and operating at night (from 6 pm–6 am) were not shown to be associated with increased SSI. The role of gram-negative prophylaxis in multilevel surgeries involving the sacrum is recommended but its effectivity remains to be elucidated.