Two studies published today in JAMA Internal Medicine suggest observed links between use of fluoroquinolone antibiotics and increased risk of a bulging or rupture of the largest artery in the body may not be as strong as previously reported.
In one the studies, US and Canadian researchers found that patients receiving fluoroquinolones for pneumonia had an increased risk of hospitalization for aortic aneurysm or aortic dissection (AA/AD) compared with those who received azithromycin, while patients who received fluoroquinolones for urinary tract infection (UTI) were not at greater risk for AA/AD compared with those who received trimethoprim and sulfamethoxazole. The overall incidence of AA/AD in all patients studied was extremely low.
In the other study, conducted in Taiwan, researchers found that an underlying infection may be the likelier risk factor for AA/AD, and that fluoroquinolones were not associated with an elevated level of risk compared with other antibiotics.
Aortic aneurysm is an abnormal, balloon-like bulge that occurs in the wall of the aorta, while aortic dissection involves a rupture of the inner layer of the aorta, which can lead to dangerous bleeding. Aortic dissection, while rare, can be fatal.
The findings of these two observational studies are noteworthy because previous studies have found that patients taking fluoroquinolones have as much as twice the risk of AA/AD compared with those not taking them, or patients on other antibiotics. Those studies led the US Food and Drug Administration to issue a warning in December 2018 that fluoroquinolones be avoided in patients at risk of aortic abnormalities, unless no other treatment options are available.
"Although the risk of aortic aneurysm or dissection is low, we've observed that patients are twice as likely to experience an aortic aneurysm or dissection when prescribed a fluoroquinolone drug," the FDA wrote. "For patients who have an aortic aneurysm or are known to be at risk of an aortic aneurysm, we do not believe the benefits outweigh this risk, and alternative treatment should be considered."
But the authors of both studies say their findings suggest fluoroquinolones may not necessarily have to be avoided in such patients.
Accounting for confounding, surveillance bias
In the first study, researchers from Brigham and Women's Hospital, Harvard Medical School, Sinai Health System, and the University of Toronto used a US commercial claims database to identify patients aged 50 or older with a diagnosis of pneumonia or UTI before initiating fluoroquinolones or other antibiotics (azithromycin for pneumonia, trimethoprim/sulfamethoxazole for UTI). They then matched them 1:1 by propensity score, controlling for 85 potential confounders.
Their aim was to assess the association of fluoroquinolones with AA/AD in these specific conditions with clinically appropriate comparators. The hypothesis of the study was that fluoroquinolones may have been associated with increased risk of AA/AD in previous studies because patients prescribed broad-spectrum fluoroquinolones tend to be sicker than patients who receive narrower-spectrum antibiotics, and are more likely to undergo advanced imaging that can detect heart issues. The primary outcome of the study was hospitalization for AA/AD.
Using data from Jan 1, 2003 through Sep 30, 2015, the researchers identified 279,554 patients with a diagnosis of pneumonia and 948,364 patients with a diagnosis of UTI. After propensity score matching, fluoroquinolone users in the pneumonia cohort had a higher rate of AA/AD compared with azithromycin users (hazard ratio [HR], 2.57; 95% confidence interval [CI], 1.36 to 4.86). The overall incidence of AA/AD was 0.03% for fluoroquinolones and 0.01% for azithromycin.
In the UTI cohort, the incidence of AA/AD was even lower (less than 0.01% in both groups), and fluoroquinolone users had similar rates of AA/AD compared with patients taking trimethoprim/sulfamethoxazole (HR, .099; 95% CI, 0.62 to 1.57).
A secondary analysis that grouped the pneumonia and UTI patients together and used amoxicillin as a comparator antibiotic did find an increased risk of AA/AD among fluoroquinolone users (HR, 1.54; 95% CI, 1.33 to 1.79). But when they limited the analysis to patients who had imaging performed to address potential surveillance bias, the association was significantly reduced (HR, 1.13; 95% CI, 0.96 to 1.33).
The authors say that while the findings suggest fluoroquinolones may increase a person's risk of AA/AD, unmeasured confounding and differential surveillance "cannot be ruled out." And they emphasize that the absolute rates of aortic aneurysm or dissection across all cohorts were low.
"Thus, given a clear indication for antibiotic use, the benefits of choosing an appropriate antibiotic in terms of coverage may outweigh a small potential increased risk for AA/AD," they wrote.
Assessing the role of infection
In the second study, researchers from National Taiwan University Hospital and National Yang-Ming University School of Pharmaceutical Science set out to answer a similar question: are fluoroquinolones linked to increased risk of AA/AD compared with other antibiotics with similar indication profiles in patients with the same type of infections. But they also looked at whether the risk of AA/AD is independently associated with the infections themselves.
In the nested case-control study, the researchers used a population-based national health insurance claims database to identify patients over 20 years old treated for AA/AD from Jan 1, 2009 through Nov 30, 2015. They then matched those patients with 10 control patients by age, sex, and follow-up duration in the database. For all cases and controls, they identified any infections and antibiotic use 60 days prior to AA/AD occurrence.
Altogether, 28,948 patients with AA/AD were identified, and 289,480 control patients were included. A total of 5,391 case patients had an infection during the 60-day risk window (18.2%), compared with 17,084 of the control patients (5.9%). Lower respiratory tract infections and genitourinary infections accounted for most of the infections.
After adjusting for confounding factors and concomitant antibiotic use among cases and controls, the odds ratio [OR] of AA/AD for any indicated infection was 1.73 (95% CI, 1.66 to 1.81), with septicemia (OR, 3.69; 95% CI, 2.63 to 3.78) and intra-abdominal infection (OR, 2.99; 95% CI, 2.45 to 3.65) having the highest increased risk.
The researchers then examined the association between antibiotic use and AA/AD among the patients with infections, comparing patients who received fluoroquinolones with those who received amoxicillin-clavulanate, ampicillin-sulbactam, or extended-spectrum cephalosporins for the same indication. After adjusting for matching factors and baseline covariates, they found no increased risk of AA/AD for fluoroquinolones when compared with amoxicillin-clavulanate or ampicillin-sulbactam (OR, 1.01; 95% CI, 0.82 to 1.24) or with extended-spectrum cephalosporins (OR, 0.88; 95% CI, 0.70 to 1.11).
The authors say their study is the first to quantify the magnitude of risk of AA/AD associated with various infections, and suggest their findings indicate that previous studies showing a link between fluoroquinolones and AA/AD risk may not have adequately adjusted for the effect of coexisting infections.
"This study's results emphasize the importance of considering coexisting infections while examining the safety of antibiotics using real-world data," the authors wrote. "Concern about AA/AD should not preclude patients with indicated infections from necessary treatment with fluoroquinolones."
An unproven link
In an editorial that accompanies the two studies, researchers with the University of California San Francisco and Kaiser Permanente Northern California argue that the findings highlight the limitations of the observational studies that have found an association between fluoroquinolone use and increased AA/AD risk, and underscore the need for more research into the link.
"Observational studies reflect real-world experience, but for rare outcomes, results are sensitive to design parameters, especially to the choice of appropriate comparison groups," they wrote. "While clinicians may choose to avoid use of fluoroquinolones in patients with aortic abnormalities, a strong causal association between fluoroquinolones and AA/AD remains unproven."