Background: Patients with a patient-reported penicillin allergy may be at greater risk for postoperative prosthetic joint infection (PJI) after total joint arthroplasty of the hip, knee, or shoulder. The increased risk of PJI in these patients has been attributed to these patients receiving a less-effective perioperative antibiotic. However, prior reports did not fully address the clinical characteristics of these unique patients, who may inherently be at greater risk of having a PJI, which may confound prior findings.
Questions/purposes: After controlling for risk factors for PJI such as BMI, anxiety, depression, and other comorbidities, we asked: Are patients with a patient-reported penicillin allergy more likely to have a PJI after THA, TKA, or total shoulder arthroplasty than patients without such a reported allergy?
Methods: We queried patient records from 2010 to 2017 from a nationwide administrative claims database of 122 million patients to adequately power an investigation comparing the 1-year incidence of PJI after TKA, total shoulder arthroplasty, and THA in patients with patient-reported penicillin allergy versus patients without a patient-reported penicillin allergy. Operative treatments for deep joint infection, identified by Current Procedural Terminology and ICD-9 and ICD-10 codes were used as a surrogate for PJI. Clinical characteristics such as age, sex, BMI, length of stay, and Charlson comorbidity index and specific comorbidities including alcohol abuse, anemia, anxiety, cardiac disease, diabetes, immunocompromised status, rheumatoid arthritis, depression, liver disease, chronic kidney disease, tobacco use, and peripheral vascular disease were queried for each study group. The odds of PJI within 1 year of THA, TKA, or total shoulder arthroplasty were compared using multiple logistic regression after adjusting for potential confounders.
Results: After adjusting for potential confounding factors such as BMI, anxiety, depression and other comorbidities, we found that patient-reported penicillin allergy was independently associated with an increased odds of PJI after TKA (odds ratio 1.3 [95% confidence interval 1.1 to 1.4]; p < 0.01) and total shoulder arthroplasty (OR 3.9 [95% CI 2.7 to 5.4]; p < 0.01). However, patient-reported penicillin allergy was not independently associated with an increased odds of PJI after THA (OR 1.1 [95% CI 0.9 to 1.3]; p = 0.36) after controlling for the same risk factors.
Conclusions: In this study, we found that patients with patient-reported penicillin allergy were at an increased risk for PJI after TKA and total shoulder arthroplasty, which we suspect-but cannot prove-is likely a function of those patients receiving a second-line antibiotic for presurgical prophylaxis. Since prior research has found that many patients listed in medical records as having a penicillin allergy are in fact not allergic to penicillin, we suggest that surgeons consider preoperative allergy testing, such as using an intraoperative test dose, to aid in choosing the most appropriate antibiotic choice before knee or shoulder arthroplasty and to amend patient medical records based on testing results. Future studies should determine whether this additional diagnostic maneuver is cost-effective.
Level of evidence: Level III, therapeutic study.