Antibiotic Use after Free Tissue Reconstruction of Head and Neck Defects: Short Course vs. Long Course

Patients and Methods

The medical charts of patients who underwent free flap reconstruction in the Department of Otolaryngology Head and Neck Surgery, University of Minnesota Medical Center, from May 2009 to February 2014 were reviewed after approval from the University of Minnesota Institutional Review Board. The exclusion criteria were: (1) Active infection pre-operatively or intra-operatively; (2) receipt of antibiotics within one week prior to surgery; (3) inadequate documentation of peri-operative medications; and (4) flap failure intra-operatively. Patient demographic and disease characteristics, including age, gender, body mass index (BMI), tobacco use, diabetes mellitus status, disease type, disease site, cancer stage, history of chemotherapy, and history of radiation were recorded. Intra-operative and post-operative variables, including flap types, hardware usage, tracheostomy/laryngectomy (either pre-existing or as a part of the operation), surgery length, transfusions, and post-operative antibiotic regimen, were recorded. Patient outcomes, including infection (SSI, pneumonia, urinary tract infection [UTI] and others), return to the operating room (OR), hospital length of stay (LOS), and patient death were reviewed up to six wks post-operatively.

Patient demographic and clinical data were summarized, and associations with antibiotic course length (short: ≤2 d, long: >2 d) and any SSI were assessed using the χ² and Fisher exact tests for categorical data and t-tests and Wilcoxon rank sum tests for continuous data as appropriate. In addition, the risk of any SSI was assessed using a multivariable logistic regression model, adjusting for variables with p < 0.10 in the univariate analyses, including history of chemotherapy, history of radiation, use of penicillins, use of clindamycin, and BMI. Odds ratios (ORs) and 95% confidence intervals (CIs) are presented. As this was an exploratory analysis, p values were not adjusted for multiple comparisons. Analyses were performed using SAS version 9.3 (Cary, NC), and p values <0.05 were considered statistically significant.

Results

A total of 149 free tissue transfers in 147 patients were conducted during the study period and eligible for inclusion in this study. Two patients received a second free flap reconstruction in a separate operation. Most patients were male (65.8%), and the mean age was 59.4 ± 14.7 (standard deviation) years (range 16–87 y). The indications for surgery were malignant tumor (133 cases [89%], with 123 squamous-cell carcinomas), benign tumor (seven cases [5%], with 5 ameloblastomas), and non-tumor (9 case [6%], with four cases of osteoradionecrosis). The free flap types used were radial forearm (49%), fibular (28%), anterolateral thigh (14%), latissimus dorsi (3%), scapular (2%), rectus abdominis (2%), and multiple (2%). A total of 126 cases involved the oral cavity, larynx, or pharynx; 12 cases involved craniofacial structures; seven cases involved both the oral cavity and craniofacial structures; and four cases involved the scalp. All patients received intra-operative antibiotics with either a single drug or a combination of cefazolin, ampicillin/sulbactam, clindamycin, or metronidazole. The post-operative antibiotics used were ampicillin/sulbactam (69%), clindamycin (17%), and others (14%), including multiple and no antibiotics. The overall recipient-site SSI rate was 16.8%, and the donor-site SSI rate was 5.4%, with a total SSI rate of 22.2%. One patient had both recipient and donor site SSI.

Of the 149 cases, 64 (43.0%) received short-course post-operative antibiotic prophylaxis, and 85 (57.0%) received long-course prophylaxis (Table 1). When comparing those who received short- vs. long-course prophylaxis, there were no statistically significant differences in age, gender, BMI, tobacco use, diabetes mellitus status, disease type, disease site, cancer stage, history of chemotherapy, or history of radiation therapy. There was, however, a greater proportion of patients who received short-course antibiotic prophylaxis in years 2013–2014 (68.4%) compared with years 2011–2012 (37.1%) as well as years 2009–2010 (30.6%; p = 0.001). This difference reflected the transition over time to preference for shorter antibiotic prophylaxis regimens in our institution.

Peri-operative variables, including flap type, hardware use, presence of a tracheostomy or laryngectomy, transfusion, use of steroids, length of surgery, and post-operative antibiotic choice, were not statistically significantly different in the two groups (Table 2). In addition, clinical outcomes, including SSI, tracheitis, flap dehiscence rate, flap loss rate, and hospital LOS were not statistically significantly different in the two groups (Table 3). There were, however, statistically significant differences between the groups in the rates of pneumonia and UTI. In particular, the long-course antibiotic prophylaxis group had a significantly higher rate of pneumonia than the short-course group (24.7% vs. 10.9%; p = 0.03), but a lower rate of UTI (0 vs. 9.4%; p = 0.01).

Several patient baseline characteristics were identified as risk factors for SSI of any type, including higher BMI (p = 0.05), history of chemotherapy (p = 0.01), and history of radiation (p = 0.01; Table 4). In addition, use of penicillins was associated with a lower rate of SSI (p = 0.04), whereas use of clindamycin was associated with a higher rate (p = 0.02) compared with those not given those antibiotics (Table 5). No other demographic, clinical, or peri-operative variables were associated significantly with SSI. When looking at these identified risk factors (BMI, history of chemotherapy, history of radiation, use of penicillins, and use of clindamycin) simultaneously in a multivariable model, higher BMI remained a significant risk factor (OR 1.63; 95% CI 1.16–2.29; p = 0.01). All other risk factors had p values >0.05 in the multivariable model, although a history of radiation was borderline significant (OR 2.82; 95% CI 0.84–9.14; p = 0.08). Finally, patients with any SSI were significantly more likely to return to the OR (p < 0.0001) or experience flap dehiscence (p = 0.001) and required a longer LOS (p = 0.01) than patients without an SSI (Table 6).

Discussion

Microvascular free-tissue reconstruction has become the mainstay of head and neck reconstructive surgery in the past four decades. However, there is still a high degree of variability in the implementation of peri-operative antibiotic prophylaxis. Given the greater risk of SSI in free flap reconstructions [9–11] and the potential highly morbid consequences of infection, many clinicians use antibiotics in an effort to prevent complications. Still, data attesting to the value of this practice are lacking. Our study showed that prolonged post-operative prophylactic antibiotic administration (>2 d) did not reduce the SSI rate significantly in free-flap reconstruction of head and neck defects.

Our study identified a history of chemotherapy, radiation therapy, and obesity as significant risk factors for SSI. However, the validity of these risk factors is controversial in the current literature [7,25–27]. Our study also demonstrated a significantly lower rate of SSI in patients treated with penicillins and a significantly higher rate of SSI in patients treated with clindamycin. This result is consistent with the data in the literature, which suggest that clindamycin monotherapy is a significant risk factor for recipient site infection [28]. In addition, a prolonged prophylactic antibiotic course was associated with a greater risk of adverse outcomes such as acquired antimicrobial resistance, Clostridium difficile infection, medication toxicity, etc. [29]. Although we did not collect data regarding antimicrobial resistance in our study, one patient who received short-course ampicillin/sulbactam developed a C. difficile infection.

Interestingly, our study demonstrated a significantly higher rate of pneumonia in the long-course group (24.7%) than the short-course group (10.9%). The most common pathogens isolated from bronchial cultures were Enterobacter and Escherichia coli. Another report found that trauma patients given prolonged (>48 h) prophylactic antibiotics were more likely to develop pneumonia with resistant or gram-negative bacteria [30]. Moreover, prolonged antibiotic regimens may alter the normal flora and phenotypes even without the selection of resistant organisms, potentially inducing normal flora to become pathogenic [31].

Current recommendations on the choice and duration of prophylactic antibiotics are based on studies excluding patients undergoing free-flap reconstruction or including only a small number of such cases. Even with the recommended prophylactic antibiotic duration of less than 24–48 h in these studies, many surgeons hesitate to adhere to short-course prophylactic antibiotics. Although our study did show a significant association of SSI with flap dehiscence and longer hospital stay, it did not associate prolonged antibiotic prophylaxis with reduced SSIs.

Carroll et al. studied the use of clindamycin after free tissue reconstruction by comparing a 1-d with a 5-d course [32]. Their study did not find a difference in outcomes in the two groups. However, as stated above, clindamycin monotherapy has been associated with a greater risk of infection. One motivation for performing our study was concern that higher rates of antibiotic resistance over the last 10 years might render prior data (i.e., Carroll et al. [32]) invalid. Yet, regarding length of therapy, our results are consistent with theirs in that no difference was found between those receiving short- or long-course antibiotics. Another, albeit smaller, study that compared antibiotic use <48 h vs. >48 h in oropharyngeal reconstructions found that clindamycin use, advanced age, and long durations of surgery were all associated with recipient site infection [28]. However, duration of therapy (<48 vs >48 h) was not a predictor of recipient site infection.

Interestingly, the conclusion that a prolonged prophylactic antibiotic course does not further reduce the SSI rate also holds true almost universally across all surgical specialties, including neurosurgery, cardiothoracic surgery, orthopedic surgery, gastroduodenal surgery, and colorectal surgery [33]. The initiation of appropriate prophylactic antibiotics prior to the surgical incision and the maintenance of serum antibiotic concentrations in the therapeutic range during the procedure appear to be more important than the duration of post-operative antibiotics [33].

The only observed benefit for long-course antibiotic prophylaxis in our study was a lower incidence of UTI. This could be explained by the positive effect of antibiotics in patients who required short-term urinary catheterization. A recent meta-analysis showed that antibiotic use at the time of urinary catheter removal carried an absolute reduction in the risk of UTI by 5.8% [34]. In our study, all patients remained on ventilation with deep sedation post-operatively to protect the delicate microvascular anastomosis. Therefore, all patients required short-term urinary catheterization. Although the timing of urinary catheter removal was not assessed in this study, patients with long-course prophylactic antibiotic were more likely to have an antibiotic at a therapeutic concentration when the catheter was removed.

The strength of our study is the inclusion of a relatively large population of patients treated at a single institution. The limitations include those inherent in a retrospective study. We were reliant on the electronic medical records and therefore limited by the quality of documentation. An example of this inadequacy relates to the nutritional status of the subjects. Although we intended to include this variable in the analysis, more than half of the subjects had missing data, and thus, we were unable to study this variable. Additional limitations relate to difficulties in addressing changes or practice patterns concerning glycemic control, resuscitation strategies, transfusion practices, and the placement of central lines or bladder catheters or both. In addition, the operations discussed in this study were performed by seven tumor ablation surgeons and four microvascular reconstruction surgeons. Therefore, surgical practices and decision making may not have been consistent.

Conclusion

Prolonged post-operative antibiotic prophylaxis in free flap reconstruction of head and neck defects does not appear to provide additional reduction of SSI risk. Moreover, long-course antibiotic prophylaxis was associated with a higher rate of pneumonia and potentially additional adverse effects such as medication toxicity and microbial resistance. Given the minimal benefit relative to the potential risk, we recommend against the use of antibiotic prophylaxis for greater than 24 h in free flap reconstruction of head and neck defects, while recognizing that a randomized trial would be required to address this issue definitively.