Is the Addition of Anaerobic Coverage to Perioperative Antibiotic Prophylaxis During Soft Tissue Sarcoma Resection Associated With a Reduction in the Proportion of Wound Complications?

Abstract

Background

Wound complications are common after resection of soft tissue sarcomas, with published infection rates ranging from 10% to 35%. Multiple studies have reported on the atypical flora comprising these infections, which are often polymicrobial and contain anaerobic bacteria, and recent studies have noted the high prevalence of anaerobic bacterial infections after soft tissue sarcoma resection [26, 35]. Based on this, our institution changed clinical practice to include an antibiotic with anaerobic coverage in addition to the standard first-generation cephalosporin for prophylaxis during soft tissue sarcoma resections. The current study was undertaken to evaluate whether this change was associated with a change in major wound complications, and if the change should therefore be adopted for future patients.

Questions/purposes

(1) After controlling for potentially confounding variables, was the broadening of the prophylactic antibiotic spectrum to cover anaerobic bacteria associated with a lower odds of major wound complications after soft tissue sarcoma resection? (2) Was the broadening of the prophylactic antibiotic spectrum to cover anaerobic bacteria associated with a lower odds of surgical site infections with polymicrobial or anaerobic infections after soft tissue sarcoma resection? (3) What are the factors associated with major wound complications after soft tissue sarcoma resection?

Methods

We retrospectively identified 623 patients who underwent soft tissue sarcoma resection at a single center between January 2008 and January 2021 using procedural terminology codes. Of these, four (0.6%) pediatric patients were excluded, as were five (0.8%) patients with atypical lipomatous tumors and two (0.3%) patients with primary bone tumors; 5% (33 of 623) who were lost to follow-up, leaving 579 for final analysis. The prophylactic antibiotic regimen given at the resection and whether a wound complication occurred were recorded. Patients received the augmented regimen based on whether they underwent resection after the change in practice in July 2018. A total of 497 patients received a standard antibiotic regimen (usually a first-generation cephalosporin), and 82 patients received an augmented regimen with anaerobic coverage (most often metronidazole). Of the 579 patients, 53% (307) were male (53% [264 of 497] in the standard regimen and 52% [43 of 82] in the augmented regimen), and the mean age was 59 ± 17 years (59 ± 17 and 60 ±17 years in the standard and augmented groups, respectively). Wound complications were defined as any of the following within 120 days of the initial resection: formal wound debridement in the operating room, other interventions such as percutaneous drain placement, readmission for intravenous antibiotics, or deep wound packing for more than 120 days from the resection. Patients were considered to have a surgical site infection if positive cultures resulted from deep tissue cultures taken intraoperatively at the time of debridement. The proportion of patients with major wound complications was 26% (150 of 579); it was 27% (136 of 497) and 17% (14 of 82) in the standard and augmented antibiotic cohorts, respectively (p = 0.049). With the numbers we had, we could not document that the addition of antibiotics with anaerobic coverage was associated with lower odds of anaerobic (4% versus 6%; p = 0.51) or polymicrobial infections (9% versus 14%; p = 0.25). Patient, tumor, and treatment (surgical, radiotherapy, and chemotherapy) variables were collected to evaluate factors associated with overall infection and anaerobic or polymicrobial infection. Patient follow-up was 120 days to capture early wound complications. A multivariable analysis was performed for all variables found to be significant in the univariate analysis. A p value < 0.05 was used as the threshold for statistical significance for all analyses. No patients were found to have an adverse reaction to the augmented regimen, including allergic reactions or Clostridioides (formerly Clostridium) difficile infection.

Results

After controlling for other potentially confounding factors such as neoadjuvant radiation, tumor size and anatomic location, as well as patient BMI, anaerobic coverage was associated with smaller odds of wound complications (OR 0.36 [95% confidence interval (CI) 0.18 to 0.68]; p = 0.003). Other factors associated with major wound complications were preoperative radiation (versus no preoperative radiation) (OR 2.66 [95% CI 1.72 to 4.15]; p < 0.001), increasing tumor size (OR 1.04 [95% CI 1.00 to 1.07]; p = 0.03), patient BMI (OR 1.07 [95% CI 1.04 to 1.11]; p < 0.001), and tumor in the distal upper extremity (versus proximal upper extremity, pelvis/groin/hip, and lower extremity) (OR 0.18 [95% CI 0.04 to 0.62]; p = 0.01).

Conclusion

The addition of anaerobic coverage to the standard prophylactic regimen during soft tissue sarcoma resection demonstrated an association with smaller odds of major wound complications and no documented adverse reactions. Treating physicians should consider these findings but note that they are preliminary, and that further work is needed to replicate them in a more controlled study design such as a prospective trial.

Level of Evidence

Level III, therapeutic study.

Introduction

Soft tissue sarcomas comprise a large and heterogeneous group of malignancies of mesenchymal origin, with more than 100 accepted histologic subtypes [37]. Local control of most extremity soft tissue sarcomas generally consists of surgical resection, often combined with external beam radiation therapy (EBRT), and occasional use of chemotherapy [27, 28].

Postoperative major wound complications are common after surgical resection of soft tissue sarcomas, with wound complications reported to occur in 10% to 48% in the evidence [1, 5, 7, 11, 18, 19, 26]. A recent systematic review cited an incidence of major wound complications of 30.2% and an overall reoperation incidence of 13.4% [29]. Although major wound complications do not affect long-term local recurrence or overall survival, they remain a substantial source of morbidity and expense because of the resulting surgical debridement, long-term antibiotic therapy, and wound care [3]. Major wound complications that are treated with formal debridement in the operating room are estimated to increase the cost of care by roughly 20% [38].

Many factors associated with major wound complications have been reported, including patient characteristics such as age, BMI, smoking, and diabetes status, as well as tumor characteristics such as size, location (lower extremity), and proximity of the resected margin to skin [2, 5, 9, 11, 18, 19, 22, 23, 26]. Treatment variables such as EBRT (particularly preoperative) and surgical closure methods also have a major effect on major wound complications [19, 22, 23]. Compared with adjuvant EBRT, preoperatively administered EBRT is associated with an increased proportion of major wound complications by a factor of two to three [29, 31]. However, EBRT is generally preferred because of the lower dose and smaller radiation field used, which reduces the risk of long-term radiation complications [9, 22].

Surgical site infections as a subset of major wound complications after soft tissue sarcoma resection are commonly polymicrobial and have a high prevalence of atypical infectious species. Two studies have investigated the spectrum of infectious species after soft tissue sarcoma resections [26, 35], and each found that a substantial proportion of infections were polymicrobial (64% to 88%) and contained anaerobic bacteria (31%). Given the incidence and morbidity of major wound infections after soft tissue sarcoma resection, interventions aimed at decreasing the proportion and severity of these complications are needed. Notably, cefazolin and other first-generation cephalosporins commonly used as prophylaxis in patients undergoing sarcoma resections generally have poor antimicrobial activity against anaerobic bacteria. Based on our institution’s internal pilot data, a change to clinical practice was instituted in July 2018 that entailed administration of antibiotics with anaerobic coverage at the time of soft tissue sarcoma resection in addition to the standard first-generation cephalosporin. However, the efficacy of this change has not yet been evaluated, and analyzing the results of adding anaerobic coverage in this context may provide both pilot data for future randomized trials as well as preliminary evidence for use of an augmented antibiotic approach in patients undergoing resections of soft tissue sarcomas.

We therefore asked: (1) After controlling for potentially confounding variables, was the broadening of the prophylactic antibiotic spectrum to cover anaerobic bacteria associated with lower odds of major wound complications after soft tissue sarcoma resection? (2) Was the broadening of the prophylactic antibiotic spectrum to cover anaerobic bacteria associated with a lower odds of surgical site infections with polymicrobial or anaerobic infections after soft tissue sarcoma resection? (3) What are the factors associated with major wound complications after soft tissue sarcoma resection?

Patients and Methods

Study Design and Setting

We retrospectively analyzed all patients at a single sarcoma center who underwent soft tissue sarcoma resection as part of their local control strategy between January 1, 2008, and January 28, 2021. We compared the proportion of major wound complications in patients based on the use of prophylactic antibiotics with anaerobic coverage at sarcoma resection. The use of the augmented antibiotic prophylaxis began in July 2018, yielding roughly 2 years of data for patients receiving this regimen.

Participants

Current Procedural Terminology codes associated with soft tissue sarcoma resections identified 623 patients. Four pediatric patients were excluded, as were seven patients with inappropriate diagnoses (two primary bone tumors and five atypical lipomatous tumors). Thirty-three patients were lost to follow-up, leaving 579 patients for the final analysis. Of these, 497 received the standard antibiotic regimen and 82 received an augmented regimen with anaerobic coverage (Fig. 1).

Fig. 1.

The STROBE flow diagram used for this study is shown here.

Data Sources and Measurement

We recorded the prophylactic antibiotic regimen given at the time of resection and whether a wound complication occurred. As noted in O’Sullivan et al. [22], major wound complications were defined as any of the following within 4 months of the initial resection: formal wound debridement in the operating room, other procedural interventions such as percutaneous drainage or drain placement, treatment with intravenous antibiotics, or deep wound packing more than 120 days. One of three authors (KRG, JBH, YCD) determined whether a wound complication occurred. Infectious species were identified by a robust protocol at our institution, which uses at least five tissue samples from every wound bed, each taken with a separate, uncontaminated rongeur [10]. Swab specimens were not used. All samples were cultured separately, each in aerobic, anaerobic, and fungal media, and held for 10 days. For common skin flora, three of five samples were required to grow the same pathogen for a positive culture result to be considered true rather than a contaminate. Preoperative radiation was given at a dose of 50 Gy. Postoperative radiation was given at 66 Gy. Eighteen patients underwent a postoperative “boost” to a total of 66 Gy. Patients received the full dose except those excluded due to being lost to follow-up. There were certain surgical variables that changed during the study period. The SPY Elite Fluorescence Imaging system (Stryker), a tool to visualize tissue perfusion intraoperatively, was used on 33 patients. Manual review of operative reports showed that no patient had their incision altered in any way due to the results of its use. Similarly, 180 patients had an incisional negative pressure wound therapy device used for closure. The proportion of patients who received this differed between the standard (31% [155 of 497]) and augmented (30% [25 of 82]) regimens (p > 0.99). Further, a comparison of the annual proportions of patients who developed a major wound complication for years before the protocol change showed no difference (p = 0.47).

Other data collected included patient information (demographics, comorbidities, and smoking status), tumor characteristics (histology, size, depth, and anatomic location), and treatment specifics of radiotherapy (preoperative, postoperative, or none), chemotherapy (binary), and surgical closure method (primary closure, staged closure with negative pressure dressings, split-thickness skin graft, and rotational or free flap). Tumor size and BMI were treated as continuous variables. Anatomic location categories were defined as shoulder or shoulder girdle, upper extremity distal to the shoulder, pelvis or groin, lower extremity distal to the hip, and other, which included thoracic and chest wall tumors.

Patients’ Descriptive Data

A total of 497 patients received a standard antibiotic regimen (usually a first-generation cephalosporin) and 82 patients received an augmented regimen with anaerobic coverage (most often metronidazole). Of the 579 patients, 53% (307) were male (53% [264 of 497] in the standard regimen and 52% [43 of 82] in the augmented regimen), and the mean age was 59 ± 17 years (59 ± 17 and 60 ± 17 in the standard and augmented groups, respectively) (Table 1).

Table 1.

Patient demographics and tumor variables (n = 579 patients)

Parameter	Standard prophylaxis (n = 497)	Anaerobic coverage (n = 82)	p value
Patient demographics
Age in years	59 ± 17	60 ± 17	0.48a
Males	53 (264)	52 (43)	0.91b
BMI in kg/m2	29 ± 6.6	29 ± 7.21	0.71a
Smoker	12 (60)	6 (5)	0.11b
Diabetes mellitus	14 (69)	6 (5)	0.08b
Tumor characteristics
Tumor size in cm	9 ± 6.8	10 ± 6.9	0.049a
Number deep to fascia	73 (364)	87 (71)	0.39b
Tumor location			0.08b
Shoulder or shoulder girdle	7 (34)	9 (7)
Upper extremity	18 (88)	12 (10)
Pelvis or groin and hip	37 (184)	50 (41)
Lower extremity	26 (127)	24 (20)
Other	13 (64)	5 (4)
Treatment characteristics
Radiation, preoperativec	54 (270)	67 (55)	0.03b
Radiation, postoperativec	16 (78)	12 (10)	0.41b
Chemotherapy	29 (142)	29 (24)	0.90b
Closure method			0.86d
Primary closure	80 (399)	79 (65)
Staged	1 (3)	12 (1)
Split-thickness skin graft	3 (16)	6 (5)
Rotational flap	9 (47)	10 (8)
Free flap	5 (27)	4 (3)

Data presented as mean ± SD or % (n).

^aMann-Whitney U test.

^bChi-square test.

^cPatients who received both neoadjuvant and adjuvant therapy were counted in each group.

^dFisher exact test.

With regard to the most common histologic types, 26% (153 of 579) of patients had undifferentiated pleomorphic sarcoma, 21% (124 of 579) of patients had myxofibrosarcoma, and 8% (48 of 579) of patients had leiomyosarcoma. There was no difference in age, gender, BMI, smoking status, or diabetes status between those who received the augmented prophylaxis and those who received standard prophylaxis. Tumors in the anaerobic coverage group were larger than those in the standard prophylaxis group (10.2 cm versus 8.8 cm; p = 0.049), and patients in this group were more likely to receive neoadjuvant radiation (67% [55 of 82] versus 54% [270 of 497]; p = 0.03). These were controlled for in the multivariable analysis. All other treatment variables were similar between the two prophylaxis groups.

Twenty-six percent (150 of 579) of patients had major wound complications: 27% (136 of 497) in the standard prophylaxis group and 17% (14 of 82) in the group with anaerobic coverage (p = 0.049). Six patients received IV antibiotics as their only intervention.

Four hundred ninety-seven patients received 24 hours of a standard agent that did not adequately cover anaerobic bacterial infection, either cefazolin (472 patients), vancomycin (22 patients), nafcillin (one), ciprofloxacin (one), or levofloxacin (one). Eighty-two patients received an augmented regimen of a second antibiotic, with anaerobic coverage given once before surgical incision. This second antibiotic was either metronidazole (62 patients), clindamycin (10 patients), piperacillin-tazobactam (seven patients), ertapenem (one patient), cefoxitin (one patient), or amoxicillin and clavulanic acid (one patient). Of the seven patients who received piperacillin-tazobactam, six received it as a sole agent. Twenty patients had previously received an augmented regimen with anaerobic coverage (ten clindamycin, three piperacillin-tazobactam, four metronidazole, and one each with amoxicillin/clavulanate acid, ertapenem, and cefoxitin) before widespread adoption due to allergies to cephalosporins or surgeon preference. Thirty-five (6% of total) patients received the standard regimen after the augmented adoption date due to miscommunication with anesthesia given the protocol change. No patients had a clinical infection with Clostridioides (formerly Clostridium) difficile documented, and no adverse reactions to the augmented regimen, such as allergies, were noted.

Ethical Approval

We obtained institutional review board approval for this study.

Statistical Analysis

We analyzed variables using chi-square, Fisher exact, or Mann-Whitney U tests where appropriate. Logistic regression was used for a multivariable analysis. In the univariable analysis, BMI, tumor size, tumor location, and receipt of neoadjuvant radiation and chemotherapy were also associated with major wound complications (Supplementary Table 1; http://links.lww.com/CORR/A862). These were included in the multivariable analysis. For continuous (independent) variables, odds ratios represent the increase in odds of the dependent variable based on a one-unit increase in the continuous variable. The study was planned and implemented in accordance to Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies [34]. All statistical analyses were performed with the R statistical software [33].

Results

Major Wound Complications

After controlling for other potentially confounding factors such as preoperative radiation, tumor size and anatomic location, as well as patient BMI, anaerobic coverage was associated with smaller odds of wound complications (OR 0.36 [95% confidence interval (CI) 0.18 to 0.68]; p = 0.003).

Surgical Site Infections With Polymicrobial or Anaerobic Infections After Soft Tissue Sarcoma Resection

With the numbers we had, we could not document that the addition of antibiotics with anaerobic coverage resulted in a decrease in anaerobic infections (4% [3 of 82] and 6% [31 of 497], respectively; p = 0.51) and polymicrobial infections (14% [70 of 503] and 9% [7 of 82]; p = 0.25).

Factors Associated With Major Wound Complications After Soft Tissue Sarcoma Resection

Tumor location in the distal upper extremity (versus proximal upper extremity, pelvis/groin/hip, and lower extremity) was associated with fewer wound complications (OR 0.18 [95% CI 0.04 to 0.62]; p = 0.009). In addition, major wound complications were associated with the use of neoadjuvant radiation (versus no neoadjuvant radiation) (OR 2.66 [95% CI 1.72 to 4.15]; p < 0.001), tumor size (OR 1.04 [95% CI 1.00 to 1.07]; p = 0.03), and patient BMI (OR 1.07 [95% CI 1.04 to 1.11]; p < 0.001) (Fig. 2).

Fig. 2.

This forest plot shows ORs from the multivariable analysis, along with 95% CIs. ORs to the right of 1 are associated with an increased odds of major wound complications, whereas ORs to the left are associated with decreased odds of major wound complications; UE = upper extremity; EBRT = external beam radiation therapy.

Discussion

Surgical site infections and major wound complications are common after resection of soft tissue sarcomas of the extremities, contributing substantially to morbidity and cost of care. New interventions are needed to help prevent or temper the severity of these complications. The high rate of atypical anaerobic and polymicrobial infections further augments this morbidity because anaerobic infections in orthopaedic surgery are known to be particularly difficult to eradicate, even after appropriate surgical and medical therapy [36]. This study evaluated a change in clinical practice that added antibiotics with anaerobic activity to the standard first-generation cephalosporin for surgical prophylaxis in soft tissue sarcoma resections. It was our belief that this would lead to a decrease in overall major wound complications in general and a decrease in anaerobic and polymicrobial infections in particular. The addition of anaerobic coverage correlated with a decrease in the proportion of major wound complications but not in anaerobic or polymicrobial surgical site infections. Given the augmented regimen’s association with smaller odds of wound complications and the lack of adverse reactions to it, we posit that consideration of such a regimen is reasonable while verification in a prospective trial is done.

Limitations

Several limitations of the present study warrant mention, principal among which is its retrospective design used to evaluate a change in clinical practice. Selection bias affected the study in several ways. The first concerns baseline characteristics: The augmented group overall had greater mean tumor size and a greater proportion of patients who received preoperative radiation. These are associated with increased odds of wound complications both in our study and in the reports of others [2, 5, 22], suggesting that selection bias could have underestimated the magnitude of the association between the augmented regimen and the (lower) odds of wound complications. Further, 20 patients received an augmented regimen before the date of protocol change, and 35 patients did not receive it after the change. These groups comprised 3.5% and 6.0% of the total study size, respectively. The reasons for these regimen changes were because of allergies or surgeon preference (in the group of 20 patients) and a lack of communication to the anesthesia team (in the group of 35 patients); they could have conferred bias as well. Similarly, although metronidazole was by far the most common antibiotic with anaerobic coverage, there was variation which should be standardized in any future trial. Although the definition of “major wound complications” used in this study is commonly used in orthopaedic oncology, it comprises several endpoints that encompass a wide spectrum. Several endpoints, such as when and for how long to pack wounds or when to give antibiotics, are susceptible to assessment bias on the part of the treating physician as well. This is somewhat mitigated by the fact that only a small number of patients (six) received IV antibiotics only without supporting data from deep cultures or percutaneous drainage.

Some methods of treatment, resection techniques, or even individual surgeon techniques may have changed longitudinally during the study period. This includes the use of surgical devices such as the SPY device or negative pressure incisional wound therapy. We confirmed that no intervention took place due to any instance of use of the former, and although negative pressure incisional wound therapy may have benefits in wound healing, it was not distributed unevenly between the two groups as noted above [32]. Other variables that could conceivably affect infection rates over time that could not be studied due to poor data acquisition include laminar flow rooms, general operating room protocol changes, and antibiotic skin preparation technique. A regimented protocol in future prospective work would be useful in minimizing these effects.

With respect to data acquisition, some patients may have presented with a wound complication after 120 days and were not captured in our time horizon. This effect is likely minimal, as other work has reported median number of days between surgery and presentation with wound complication being 28, 34, and 88 days [19, 26, 35]. Further, anaerobic and other atypical bacteria in these infections tend to be fastidious and notoriously difficult to culture [12, 20]. Although this would have a minimal effect on the identification of total major wound complications, it could have hampered the identification of anaerobic infections. Further, partitioning groups according to the anatomic location variable may be different from that in other reports; this should be noted before any direct comparison is done. Finally, we were not able to identify specific subgroups that would benefit most from this intervention due to the size of the study. Robust subgroup analysis should be a major goal of future trials, if undertaken.

Antibiotic stewardship requires attention to possible effects of widening any antibiotic use such as allergies, C. difficile infections, or the emergence of antibiotic resistance. Although complications related to the augmented antibiotic regimen were not found in the present study, this is an inherent risk to consider if such a regimen is adopted. However, an augmented regimen consisting of the addition of metronidazole has the benefit of being the first-line treatment for C. difficile infections.

Major Wound Complications

We found that after controlling for potentially confounding variables like BMI, preoperative radiation, tumor size, tumor location, and chemotherapy, the odds of developing a major wound complication were lower in patients who received additional anaerobic antibiotic coverage. Most wound infections in nononcologic orthopaedic surgery are because of wound colonization by gram-positive skin flora at the time of surgery [24]. It is not known whether the atypical infections after soft tissue sarcoma resection result similarly from seeding at the time of resection or from hematogenous seeding in the subacute period after surgery. Anaerobic-containing infections in this population present later than purely aerobic infections, although whether this is because of late hematogenous seeding of the former or is simply owing to the slower-growing, less virulent nature of those organisms is not known [26]. The fact that there was a decreased proportion of major wound complications in the current study was associated with changes to the perioperative antibiotic regimen lends evidence to intraoperative rather than late hematogenous seeding as the mechanism. It is also unknown whether this decrease would further benefit from a longer duration of coverage postoperatively, a question that can be addressed in future work.

Surgical Site Infection With Polymicrobial or Anaerobic Infections After Soft Tissue Sarcoma Resection

Augmented antibiotic coverage was not associated with a reduction in the odds of surgical site infection with polymicrobial or anaerobic infections in this series. Given the rarity of these infections compared with major wound complications, it is possible that this lack of association was because of the small number of such infections and that a larger study may detect an association. Regardless, the present study highlights the nontrivial number of polymicrobial and anaerobic infections after these surgeries that has been previously reported [26, 35]. The difficulty in culturing anaerobic bacteria [20] and the suboptimal rate of treatment success versus anaerobic bacteria despite appropriate medical and surgical intervention [36] reinforces the potential benefit of deep tissue cultures at wound debridement or drainage to better tailor therapeutic antibiotics in the case of surgical site infection.

Factors Associated With Major Wound Complications After Soft Tissue Sarcoma Resection

Several other variables were associated with lower or higher odds of major wound complications after controlling for relevant confounding variables. The strongest variable was preoperative radiation, as has been widely reported. Other reported factors associated with major wound complications are tumor size, BMI, and upper extremity location distal to the shoulder girdle, the latter of which has a negative association [5, 8, 13, 16, 17, 19, 22, 29, 31]. These other factors are not novel in the published evidence on the subject. However, their repetition elsewhere lends strength to the external validity of our results. Although these are not modifiable variables, many factors that have been associated with wound complications in the evidence are. These include patient factors such as smoking cessation and tight glycemic control, the use of vascularized tissue coverage using free or rotational flaps, alterations to radiation dosage and fractionation, or the use of negative pressure dressings, all of which may affect major wound complication rates as well [4,5,6,14,15,21,23,25,28,30]. Future prospective studies examining these in a controlled setting may be of benefit to patients.

Conclusion

We observed a lower odds of major wound complications in patients treated with additional anaerobic antibiotic therapy than in those treated with standard prophylactic antibiotic protocols after soft tissue sarcoma resection. We cannot fully account for the potential adverse effects of these additional antibiotics, particularly on a population level, and our results are preliminary. Treating physicians should consider these findings but use caution in adopting our protocol as these findings should be confirmed in a larger, controlled study.