German Claims Study Finds Short-Term Tendon Injury Risk With Fluoroquinolones

The findings come from an analysis of German health insurance claims data covering more than 12 million antibiotic treatment episodes, published in the journal Drugs & Therapy Perspectives.

This design is intended to reduce confounding by indication, since both groups consisted of patients being treated for bacterial infections rather than comparing antibiotic users with people who took no antibiotics at all.

The researchers drew on data from the German statutory health insurer AOK, which covers roughly 28 million people nationwide, for the years 2013 to 2019. Adults with a first antibiotic dispensing between 2014 and 2018 were followed for 365 days after exposure. After applying all inclusion and exclusion criteria, the analytic dataset contained 12,007,302 antibiotic episodes, of which about 21 percent involved fluoroquinolones such as ciprofloxacin, levofloxacin, moxifloxacin, ofloxacin, norfloxacin or enoxacin. People with previous diagnoses of the outcome of interest were excluded, and individuals had to have uninterrupted insurance coverage during the baseline year.

The outcome of interest was the first occurrence of injuries to muscle, tendon or synovium, abbreviated MTS, identified through outpatient and inpatient diagnoses coded with the German version of ICD-10. For tendon injuries, the presence of a procedure code indicating surgical treatment was also required. In total, 6.23 percent of all index episodes involved an incident MTS injury during the 365-day risk window. The most common diagnoses were synovitis and tenosynovitis, other disorders of synovium and tendon, and other disorders of muscle.

In the primary analysis, which included both inpatient and outpatient diagnoses over the full 365-day follow-up, the adjusted hazard ratio for MTS injuries with fluoroquinolones compared with the comparator antibiotics was 0.99, with a 95 percent confidence interval of 0.99 to 1.00. In other words, there was no increased relative risk over the year-long window. The standardised incidence rates were also nearly identical, at 591 events per 10,000 fluoroquinolone episodes versus 584 per 10,000 comparator episodes.

However, the picture changed in sensitivity analyses that restricted the outcome to inpatient diagnoses and shortened the risk window. Because outpatient diagnoses in the claims data are recorded only on a quarterly basis, the researchers used inpatient diagnoses, which are available at weekly intervals, to examine shorter windows of 30, 60 and 92 days. During the first 30 days after dispensing, the relative risk of inpatient MTS injuries was increased, with an adjusted hazard ratio of 1.12, and the relative risks decreased with longer follow-up. The authors interpret this pattern as suggesting a potential short-term increase in risk within the first month, and as evidence that the much longer 365-day window in the main analysis may have diluted a drug effect that is concentrated early after exposure.

When the composite outcome was disaggregated into more specific diagnostic categories, the largest relative risks were seen for spontaneous ruptures of synovium and tendon, which were 26 percent more likely in fluoroquinolone episodes, and for Achilles tendon injuries, which were 35 percent more likely, with adjusted hazard ratios of 1.26 and 1.35 respectively. The relative risk of unspecified myopathies was also modestly increased, as was myositis, while the result for the specific diagnosis code for drug-induced myopathy was inconclusive. The authors emphasise that the absolute incidence of tendon ruptures, particularly of the Achilles tendon, is low, and that the clinical relevance of these relative estimates should therefore be interpreted cautiously.

The choice of comparator antibiotic mattered. When fluoroquinolone episodes were compared separately with each individual reference drug, an increased relative risk of inpatient MTS events was apparent in the 92-day window against amoxicillin, azithromycin, clindamycin and doxycycline, whereas the relative risk was decreased in comparisons with cephalexin and amoxicillin-clavulanic acid. The authors caution that these comparator-specific estimates should be viewed as supportive evidence of differential risk, because residual confounding by indication cannot be fully ruled out, as different antibiotics may be preferentially prescribed for different infection types or severities.

The findings add nuance to a body of earlier research that generally compared fluoroquinolone users with non-users. A Canadian cohort study of people aged 65 and older reported an adjusted hazard ratio of 2.40 for tendon disorders during a 30-day window, and a UK nested case-control study found an adjusted incidence rate ratio of 1.61 for tendon ruptures, rising to 3.14 for the Achilles tendon, over 60 days. A study based on US Medicare data that, like the present study, used active comparators, found no class effect but did observe an increased risk for levofloxacin specifically. The authors suggest that the smaller estimates in their study may partly reflect the use of reference antibiotics rather than non-users, the longer follow-up period, and the older age structure of the German cohort, which had a mean age of 59 years.

As with any observational study based on claims data, several limitations apply. Drug dispensing records are only a proxy for actual exposure, adherence may vary, and inpatient antibiotic use is not captured in the AOK data, prompting several sensitivity analyses to address potential misclassification. Information on lifestyle and occupational factors such as body mass index, smoking, physical activity and occupational load was unavailable, although the authors note there is no strong evidence that these factors substantially influence the choice between fluoroquinolones and other antibiotics. Regulatory risk communications in Germany since 2012, advising against fluoroquinolone prescribing in people with a history of muscle or tendon weakness, may also have shaped prescribing patterns in the recent data. The authors conclude that no difference in overall MTS risk was found over 365 days, but that sensitivity analyses point to a potential short-term increase in the relative risk of specific tendon injuries, particularly of the Achilles tendon, within the first 30 days after fluoroquinolone exposure.

Subject of Research: Fluoroquinolone-associated injuries of tendon, muscle and synovium in a large German claims data cohort study

Article Title: Fluoroquinolone-Associated Injuries of Tendon, Muscles and Synovium: Recent Evidence From a Large German Claims Data Cohort Study

Article References: Wicherski, J., Peltner, J., Becker, C., Schüssel, K., Brückner, G., Schlotmann, A., Schröder, H., Kern, W. V., & Haenisch, B. (2026). Fluoroquinolone-Associated Injuries of Tendon, Muscles and Synovium: Recent Evidence From a Large German Claims Data Cohort Study. Drugs & Therapy Perspectives. https://doi.org/10.1007/s40267-026-01278-1

Image Credits: AI Generated

DOI: 10.1007/s40267-026-01278-1

Keywords: fluoroquinolones, antibiotics, tendon rupture, Achilles tendon, adverse drug reactions, pharmacoepidemiology, claims data, cohort study, drug safety, musculoskeletal injuries, Germany, tendon injuries