Background: Fluoroquinolones (FQs) are used commonly for patients (pts) with community acquired bacterial pneumonia (CABP). A recent FDA Drug Safety Communication strengthened labeling regarding tendinopathy/tendon rupture for FQs. The data prompting this change lacked a comparator group of pts using other antibiotics, like ceftriaxone/azithromycin (CTX-AZ) for similar indications. The objectives of this study were to compare the incidence of adverse tendon events (TE) between FQ and CTX-AZ among pts with CABP and determine if FQ treatment is independently associated with TE.
Methods: A retrospective cohort study was performed among pts in the Upstate New York Veterans Healthcare Administration. Inclusion criteria: 1) age ≥ 18 years, 2) diagnosis of CABP (ICD9 code with manual confirmation) from 1/2014 to 12/2015, 3) receipt of IV/oral FQ or CTX-AZ ≥ 1 day, and 4) treatment initiated as inpatient. Data were collected from pts medical records. Occurrence of TE was defined using a natural word search algorithm of pts clinical progress notes within 90 days of starting FQ or CTX-AZ therapy. Search terms were: tendinopathy, tendon pain, tendon rupture, tendinitis, and Achilles heel pain/tear/torn/rupture. Classification and regression tree (CART) was used to identify breakpoints in continuous variables associated with TE.
Results: There were 379 FQ and 274 CTX-AZ recipients. Mean ± standard deviation (SD) ages for FQ and CTX-AZ recipients were, 73.0 ± 12.7 vs 72.8 ± 12.7 years, respectively. Mean (SD) APACHE-II was significantly higher for FQ than CTX-AZ recipients, 10.2 ± 5.1 versus 8.5 ± 3.6, respectively (p<0.001). Residence in the intensive care unit at start of therapy did not differ (FQ: 11.6% vs CTX-AZ: 10.2%, p=0.58). The incidence of TE did not differ between groups (FQ: 9/379 [2.4%] vs CTX-AZ: 4/274 [1.5%], p=0.57). In multivariate analyses (Figure), treatment was not independently associated with TE (aOR: 1.78, 95% confidence interval: 0.51-6.21, p=0.37) after adjustment for treatment duration, APACHE-II, age ≥52 years and BMI ≥27.5.
Conclusion: Incidence of TE did not significantly differ between FQ and CTX-AZ recipients. After adjustment, FQ treatment was not independently associated with an increased risk of TE.
N. Stornelli, None
G. Belfiore, None
T. P. Lodise, paratek: Consultant and Scientific Advisor , Consulting fee .