(AK) Macrolide Treatment Failure Due to Drug-Drug Interactions: Real-World Evidence to Evaluate a Pharmacological Hypothesis

Publication Number: 327

Background: Macrolide antibiotics treat a variety of infections but have a risk of treatment failure related bacterial resistance as a result of both prolonged sub- and supratherapeutic exposure.

Objectives: We characterized the rate of macrolide antibiotic treatment failure and quantified the influence of drug-drug interactions (DDIs) on antibacterial treatment failure rates.

Methods: This was a nested case-control study using the IBM MarketScan data (2004-2015). Two study cohorts were developed: an acute infections cohort and a bacteria-specific pneumonia cohort. Patients were included if they had a macrolide (erythromycin, clarithromycin, or azithromycin) fill within five days of diagnosis, were < 65 years old, and had health coverage for 12 months prior and 60 days after diagnosis. Patients were excluded with a hospitalization within 30 days or an infection within 6 months prior to index date, a previous antibiotic claim 30 days prior to index date, or an immunocompromising disorder. Case-control pairs were matched 1:1 without replacement using incidence density sampling on the following: age, sex, infection year, number of comedications, macrolide received, macrolide dose, and comorbidity score. DDI covariates were based on CYP3A4 inhibitors and inducers where macrolide exposures either increase or decrease significantly; mild DDI (25-49% exposure change) or moderate DDI (≥50% exposure change). Analyses include calculation of the treatment failure rates, i.e. macrolide therapy didn’t eradicate the infection, and conditional logistic regression.

Results: The common infection cohort consisted of 135,683 case-control pairs, while the bacteria-specific cohort consisted of 1,115 pairs. Azithromycin was the most prescribed macrolide. The overall rate of treatment failure was 5.7% and 7.7% in the common infection cohort and bacteria-specific cohort, respectively. For the common infection cohort, the odds ratio for treatment failure was not influenced by mild AUC changes, but was for moderate macrolide AUC decrease (1.37 [1.02, 1.86]) and moderate AUC increase (1.12 [1.08, 1.17]). In comparison, the bacteria-specific cohort regression showed no statistical significance for DDIs.

Conclusions: The results of this project indicate that risk of macrolide treatment failure is not significantly impacted when macrolide exposure is mildly altered. However, moderate changes in macrolide exposure may impact treatment failure rates for common infections. When looking only at bacteria-specific CAP patients, the significance of macrolide exposure impact is decreased, suggesting that other mechanisms may influence treatment failure.