Is Penicillin Allergy a Risk Factor for Surgical Site Infection After Oral and Maxillofacial Surgery?
Purpose: The selection of perioperative antibiotics for prevention of surgical site infection (SSI) is often limited by the presence of a reported penicillin allergy. The purpose of this study was to determine if oral and maxillofacial surgery patients who report allergy to penicillin are at an increased risk of developing SSI.
Methods: A retrospective cohort study was performed of patients who underwent oral and maxillofacial surgical procedures in the operating room setting at a single institution between 2011 and 2018. The following categories of procedures were investigated: dentoalveolar, orthognathic, orthognathic with third molar extraction, pathology and reconstruction, and temporomandibular joint. The primary predictor and outcome variables were reported penicillin allergy and surgical site infection, respectively. Bivariate and multiple logistic regression analysis were performed. P < .05 was considered to be significant.
Results: The cohort was composed of 2,058 patients of which 318 (15.5%) reported allergy to penicillin. Beta-lactam antibiotics were administered less frequently to penicillin allergic patients perioperatively compared with those without penicillin allergy (7.9 vs 97.1%, P < .001), while clindamycin was more commonly administered (76.4 vs 2.5%, P < .001). Clindamycin was associated with a higher SSI rate compared with beta-lactam antibiotics (5.6 vs 1.4%, P < .001). Penicillin allergy was significantly associated
with SSI at an adjusted odds ratio of 2.61 (95% CI 1.51 to 4.49, P=.001). After holding perioperative antibiotic usage equal between the 2 groups, penicillin allergy per se was no longer associated with SSI (P=.901), suggesting that the outcome was mediated by antibiotic selection.
Conclusions: Penicillin allergy was associated with development of SSI due to receipt of non−beta-lactam antibiotics as perioperative prophylaxis. Formal allergy evaluation should be considered for patients
with putative penicillin allergy.
Penicillin allergy is among the most common self-reported allergies. Approximately 10% of patients report penicillin hypersensitivity; however, 90% of these cases are found not to be true allergies. Potential cross-reactivity with cephalosporins has been a concern due to the presence of similar sidechains to penicillin in some early generation agents, but the risk is estimated to be minimal.
A patient’s penicillin allergy status is an important consideration when choosing perioperative antibiotics. Diagnosis of a penicillin allergy often precludes use of penicillins and other beta-lactams, including cephalosporins, and instead results in the use of broad-spectrum antibiotics that are more associated with toxicities and the development of antimicrobial resistance.6 The proper choice of antibiotics is also a contributing factor in the prevention of surgical site infection (SSI). Report of a penicillin allergy has been found to be associated with a significantly increased odds of SSI in patients undergoing orthopedic, gynecologic, colorectal, and cardiac surgery; this has been found to be brought about by the use of alternative perioperative antibiotics.
Surgical site infection is a risk for patients undergoing oral and maxillofacial surgical procedures. The reported risk of SSI following orthognathic surgery ranges from 0.5 to 18% and has been linked to selection and duration of antibiotic prophylaxis. For total replacement of the temporomandibular joint, the SSI rate is reported to be 1.5%. Data for SSI following third molar removal is also reported with wide variability.
The purpose of this study was to answer the following clinical question: among patients undergoing oral and maxillofacial surgical procedures, do those who report penicillin allergy when compared with those who do not report being penicillin allergic, have an increased risk of developing SSI? The authors hypothesized that patients with a reported penicillin allergy would have an increased odds of SSI after oral and maxillofacial surgical procedures when compared with similar patients without a penicillin allergy designation. The authors also hypothesized that there would be a predilection for SSI on the basis of the class of prophylactic antibiotic administered with an increased incidence associated with the use of non −beta-lactam antibiotics. The specific aims of this study were to: 1) estimate and compare the rate of SSI in patients with and without a reported penicillin allergy; and 2) assess the association between choice of antibiotic prophylaxis and development of SSI.
Methods
STUDY DESIGN
This was a retrospective cohort study of patients who underwent oral and maxillofacial surgical procedures in the operating room setting between 2011 and 2018 at Massachusetts General Hospital (Boston, MA). Institutional review board approval was sought through Partners Human Research and an exemption was granted in writing (Protocol: 2019P000357). Patients who underwent dentoalveolar, orthognathic, pathology and reconstruction, and temporomandibular joint procedures were identified through the Massachusetts General Hospital patient data registry. Inclusion criteria were complete medical record data identifying allergy history; complete medical history; operative records that included antibiotic management; and sufficient clinical follow-up documenting development or absence of SSI. Exclusion criteria were incomplete medical records; inadequate clinical follow-up; absence of receipt of perioperative antibiotics; and the presence of a preexisting infection at time of surgery. Only the first operation for patients who underwent subsequent procedures during the study period was included.
VARIABLES
A data intake form was used to record candidate variables which included demographic factors, medical history (including history of drug allergy), operative records (type of surgery, wound classification, duration of procedure) and antibiotic management (selection and duration). The primary predictor variable was reported penicillin allergy, defined as documentation of an allergic reaction to any antibiotic within the penicillin’s group at the time of surgery. Both patient self-report and clinically verified allergy were considered to have met this criterion. The outcome variable was the development of SSI. The Centers for Disease Control and Prevention defines an SSI as either superficial or deep. A superficial SSI is one that occurs no deeper than the subcutaneous tissue, occurs within 30 days after the procedure, and meets 1 of the following criteria: purulent drainage, an organism is identified aseptically, dehiscence or purposeful opening by the surgeon of the superficial incision due to infection, or a diagnosis of SSI is made by the surgeon. A deep SSI is one that occurs deeper than the subcutaneous tissue, occurs within 30 and up to 90 days postoperatively and meets 1 of the following criteria: purulent drainage, dehiscence or purposeful opening by surgeon of deep incision with aseptic identification of organisms, or development of an abscess. We adapted and simplified these criteria to include submucosal tissue in addition to subcutaneous tissue and to monitor for a uniform period of 30 days postoperatively. Development of either a superficial or deep SSI per these criteria was considered to have met our definition for SSI.
DATA ANALYSIS
Descriptive statistics were calculated for each variable. Categorical values were analyzed by x2 or Fisher exact test as appropriate. Mann−Whitney U test was used to compare continuous variables. A multiple logistic regression model was used to obtain adjusted odds ratios. Independent variables associated with the outcome at a P value of <.15 in a bivariate analysis were included in the regression model. Variables were excluded if they were collinear with other variables included in the model. P < .05 was considered to be statistically significant for all analyses. Data analysis was performed using SPSS Version 25 (IBM Corp., Armonk, NY).
Results
A total of 2,058 patients were included in this study of which 318 (15.5%) reported an allergy to penicillin (Table 1). The penicillin allergy group was older (median age 38 vs 30, P < .001) and had a higher proportion of females (66.7 vs 55.5%, P < .001). Patients who reported a penicillin allergy were also more likely to report an allergy to cephalosporins (6.6 vs 2.6%, P < .001), have history of malignancy (10.7 vs 7.4%, P=.042) or a status of immune compromise or dysfunction (8.5 vs 3.9%, P < .001). The stated type
of reaction to penicillin was categorized as: 1) hypersensitivity (n = 248); 2) side effect or intolerance (n = 31); or 3) unknown (n = 39).
The following surgical procedures were represented: dentoalveolar (219 patients), orthognathic (478 patients), combined orthognathic and third molar extraction (69 patients), pathology and reconstruction (836 patients), and temporomandibular
joint (456 patients). The penicillin allergy group had a higher proportion of wound classification of clean (30.5 vs 22.9%, P=.003) and less of clean-contaminated (68.9 vs 76.7%, P=.003). Duration of surgery was shorter for the penicillin allergy group (median
of 98 vs 113 minutes, P < .001). Additionally, chlorhexidine was used less commonly in the penicillin allergy group (65.4 vs 74.5%, P=.001). The prophylactic perioperative use of beta-lactam antibiotics was less common in penicillin allergic patients (7.9 vs
97.1%, P < .001) and clindamycin was more commonly used (76.4 vs 2.5%, P < .001). Other alternatives to beta-lactam antibiotics were also more common in the penicillin allergic patients (15.7 vs 0.4%, P < .001). The duration of postoperative antibiotics was not statistically different between the 2 groups (P=.079).
Table 2 shows the bivariate correlations between study variables and the development of SSI. Cephalosporin allergy (P=.020), wound classification (P=.097), chlorhexidine use (P=.012), and duration of procedure (P=.026) were associated with SSI at a P value of less than .15. The selection of perioperative antibiotic was also associated with the development of SSI. SSI was more likely to develop after administration of clindamycin than after receipt of beta-lactam antibiotics (5.6 vs 1.4%, P < .001). Ampicillin-sulbactam was associated with 13 SSIs (rate of 1.3% of recipients), penicillin G was associated with 8 (2.4%) and the combination of penicillin G and nafcillin was associated with 3 (4.0%). Cefazolin, a first-generation cephalosporin, was associated with zero infections in this cohort. No SSIs were noted for patients receiving prophylaxis with other alternative antibiotics.
Table 3 shows the bivariate relation between the primary predictor variable (reported penicillin allergy) and the development of SSI. Patients who reported a penicillin allergy were more likely to develop an SSI (4.1 vs 1.6%, P=.004) with a relative risk of 2.63 (95% CI 1.37 to 5.05).
A multiple logistic regression analysis was used, and variables found to be associated with development of SSI at P < .15 were included. Chlorhexidine use was excluded from this model due to collinearity with wound classification. Table 4 shows the results of the multiple logistic regression model. After adjusting for cephalosporin allergy, wound classification, and duration of procedure, penicillin allergy was found to be associated with SSI development with an adjusted odds ratio of 2.61 (95% CI 1.51 to 4.49, P=.001). Cephalosporin allergy (P < .001), wound classification (P=.010), and duration of procedure (P=.001) were also significantly associated with development of SSI. When adjusting for perioperative antibiotic selection, penicillin allergy per se was no longer associated with development of SSI (P=.901), suggesting that it was the alternative antibiotic choice that was associated with the increase in SSI.
