Figures
Abstract
This study evaluated the association between individual fluoroquinolones and tendon disorders in Japan using two spontaneous reporting systems: the Food and Drug Administration Adverse Event Reporting System (FAERS) and the Japanese Adverse Drug Event Report (JADER). Five agents—ciprofloxacin, levofloxacin, moxifloxacin, norfloxacin, and ofloxacin—were analyzed for “Muscle, tendon and ligament injuries” and “Tendon disorders.” All fluoroquinolones showed positive signals in the FAERS, whereas only levofloxacin demonstrated consistent and robust associations in the Japanese subset of the FAERS and the JADER. In the JADER, levofloxacin exhibited strong signals for muscle, tendon, and ligament injuries (reporting odds ratio [ROR] 30.6, information component [IC] 4.5) and tendon disorders (ROR 134.2, IC 5.8). Stratified analysis revealed markedly stronger signals in males, particularly those aged ≥70 years (relative ROR 12.21 and 25.67; IC delta 3.18 and 2.85, respectively). Oral levofloxacin was associated with higher reporting odds than intravenous administration. Time-to-onset analysis indicated that most tendon-related adverse events occurred within 14 days after drug initiation. Levofloxacin exhibited the most consistent and robust association with tendon disorders, especially in males aged ≥70 years in Japan. Therefore, clinicians in Japan should exercise caution when prescribing levofloxacin to older male patients.
Citation: Nakashima T, Noguchi Y, Masuda R, Yoshimura T (2026) Levofloxacin-associated tendon disorders: Sex-, age-, and route-specific disproportionality analysis using FAERS and JADER. PLoS One 21(9): e0358427. https://doi.org/10.1371/journal.pone.0358427
Editor: Zulkarnain Jaafar, Universiti Malaya, MALAYSIA
Received: November 22, 2025; Accepted: August 31, 2026; Published: September 11, 2026
Copyright: © 2026 Nakashima et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying this study were obtained from third-party pharmacovigilance databases. JADER data are publicly available from the Pharmaceuticals and Medical Devices Agency at https://www.info.pmda.go.jp/fukusayoudb/CsvDownload.jsp The FAERS data used in this study were obtained from JAPIC AERS, a curated pharmacovigilance data service provided by the Japan Pharmaceutical Information Center (JAPIC). The authors cannot publicly redistribute the raw JAPIC AERS data because restrictions apply under the data provider’s terms of use. Researchers interested in accessing JAPIC AERS may apply to JAPIC through the information provided on the JAPIC website at https://www.japic.or.jp/service/information/jarea.html The authors did not receive any special privileges in accessing these data that other researchers would not have. The aggregate data generated in this study are included in the paper and supplementary files.
Funding: This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Number: 22K12890). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: Prof. Tomoaki Yoshimura is affiliated with an endowed course established through a donation from Tanpopo Pharmacy Co., Ltd., a commercial source. Tanpopo Pharmacy Co., Ltd. had no role in the design of the study, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare no other competing interests related to employment, consultancy, patents, products in development, marketed products, or other financial relationships. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
Fluoroquinolones are clinically important antibiotics with high oral bioavailability, extensive tissue distribution, and broad-spectrum antimicrobial activity [1]. Despite their widespread clinical use, fluoroquinolones are associated with a range of adverse effects, some of which are rare but potentially serious. These adverse events include peripheral neuropathy, aortic aneurysm or dissection, dysglycemia (hypoglycemia and hyperglycemia), QT interval prolongation with torsades de pointes, Clostridioides difficile infection, neuropsychiatric adverse effects, tendinopathy, and tendon rupture [2–8].
Fluoroquinolone-associated tendinopathy and tendon rupture have been reported across several study designs, including systematic reviews and population-based evaluations [2,9]. Disproportionate reporting of tendinopathy and tendon rupture has also been observed in large pharmacovigilance datasets such as the US Food and Drug Administration Adverse Event Reporting System (FAERS) and VigiBase, with FAERS-based analyses identifying associations with ciprofloxacin, levofloxacin, and moxifloxacin [10,11]. Contrastingly, evidence from Asia, particularly Japan, remains limited and fragmented, with most data derived from case reports and small-scale or single-center studies. In a single-center analysis in Japan, Hori et al. reported 14 tendon disorder events among 17,147 patients exposed to fluoroquinolones (levofloxacin, n = 9; moxifloxacin, n = 2; tosufloxacin, n = 2; ofloxacin, n = 1) and a higher risk of tendon disorder events with fluoroquinolones than with cephalosporins [12]. However, given the modest event counts and single-center setting, conclusions about agent-specific risks in Japanese populations should be regarded as preliminary and interpreted with caution. Accordingly, the risk of tendinopathy or tendon rupture associated with individual fluoroquinolones remains unclear in Japan.
The mechanisms underlying fluoroquinolone-associated tendon disorders have not yet been fully elucidated. However, several hypotheses have been proposed. First, in extracellular matrix degradation, fluoroquinolones upregulate matrix metalloproteinases (MMPs) such as MMP-2 and MMP-3, thereby enhancing the degradation of type I collagen [13,14]. Second, in impaired repair dynamics, fluoroquinolones inhibit the speed and migration of tenocytes by downregulating focal adhesion kinase phosphorylation [14,15]. Third, oxidative stress and mitochondrial dysfunction have been implicated. Additionally, experimental models have demonstrated mechanisms such as increased reactive oxygen species, oxidation of type I collagen, and caspase-mediated apoptosis; N-acetylcysteine attenuates these effects [16].
In this study, we evaluated the association between individual fluoroquinolones and tendon disorders in Japan and assessed the differences by age, sex, and route of administration. Disproportionality analyses were conducted using spontaneous reports from the Japanese Adverse Drug Event Report (JADER) database and FAERS.
Materials and methods
All the analyses were conducted in accordance with the ethical principles of the Declaration of Helsinki for medical research. As the data were anonymized and all analyses were descriptive, informed consent from patients was not required, and ethical review by an institutional review board was deemed unnecessary. This study was conducted according to The Reporting of A Disproportionality Analysis for Drug Safety Signal Detection Using Individual Case Safety Reports in Pharmacovigilance guidelines [17,18].
Data source
We used case data from two pharmacovigilance resources: the JADER database (September 2024 release version) and JAPIC AERS, a curated distribution of the FAERS, covering Q4 1997–Q2 2024. JADER, maintained by the Pharmaceuticals and Medical Devices Agency, is publicly accessible (in Japanese only) at https://www.info.pmda.go.jp/fukusayoudb/CsvDownload.jsp (accessed on October 6, 2025). It is distributed as four comma-separated value (CSV) files: DEMO.csv (patient demographics), DRUG.csv (drug information), HIST.csv (medical history), and REAC.csv (adverse event information). FAERS quarterly files beginning with Q4 2004 are available for download on the FDA website. Earlier data (1997–2003) were archived by the US National Technical Information Service. These files are included in the JAPIC AERS database. The FAERS contains duplicate data, and the DRUG data are not standardized in terms of product name notation, ingredient name, specification, and dosage form. Some were misspelled. Therefore, when using the FAERS data for research purposes, it is necessary to clean the data and remove duplicates. The JAPIC AERS data have already been clean and deduplicated; the data can be referenced by the drug or brand name to identify the ingredient name. In this study, we excluded records in which involvement of drugs in adverse drug reactions was listed as ‘concomitant drug’ or was missing in JADER, as well as records in which the role_code was ‘concomitant’ or missing in JAPIC AERS. Patients with missing age were excluded from the analysis. Sex was restricted to male and female. In FAERS, we constructed a Japan-restricted FAERS subset (FAERS-Japan), retaining records with JP in OCCR_COUNTRY, or when OCCR_COUNTRY was missing, JPN in COUNTRY_CODE. Institutional review board permission was not required because this was an observational study using data from an open-access database containing anonymized patient information. Informed consent was not required because it was not possible to identify specific individuals among patients.
Definitions of suspected drugs and adverse events
Fluoroquinolones (ciprofloxacin, levofloxacin, moxifloxacin, norfloxacin, and ofloxacin) were selected for the investigation. The five target fluoroquinolones were identified using generic names and their salt forms, as listed in S1 Table. Because JAPIC AERS standardizes product and brand names to ingredient names during database curation, additional brand-name searching was not performed. WHO Drug Dictionary or RxNorm was not used for additional drug-name standardization. The preferred terms from the Medical Terminology for Regulatory Activities/Japanese version were used to identify adverse events. We evaluated the target adverse events using high-level terms (HLTs) [19]. The preferred terms included in the HLTs for “muscle, tendon, and ligament injuries” and “tendon disorders” were included as target adverse events in this study. The complete list of Preferred Terms included in each target HLT is shown in S2 Table.
Signal detection
In this study, disproportionality signals were assessed for drug-induced muscle, tendon, and ligament injuries and tendon disorders, using the reporting odds ratio (ROR) and relative ROR (rROR) as the frequentist statistical approach and the information component (IC) and IC delta (ICΔ) as the Bayesian confidence propagation neural network (BCPNN) method [20]. The ROR was calculated using a 2 × 2 contingency table (Table 1), according to Equations (1) and (2). In the present study, the analyses were stratified according to age and sex. Signal detection in sex or age was defined using the rROR reported by the European Medicines Agency and derived from a 4 × 2 contingency table (Table 2) and Equations (3)–(6) [20]. The ROR, rROR, and their 95% confidence intervals (CIs) were estimated. We applied a lower bound of the 95% CI (ROR025, rROR025) > 1 [21]. The ROR was not calculated when no reports were available. IC and ICΔ were calculated using Tables 1 and 2 and Equations (7)–(12). Their 95% credible intervals (95% CrIs) were estimated. The 95% CrI of the ICΔ was calculated using the same IC method. IC025 and IC975 are the lower and upper limits of 95% CrI, respectively. IC- and ICΔ-based signal was defined as the lower bound of the interval, the 2.5th percentile of CrI (IC025, ICΔ025) > 0 [22]. In this study, the criteria for signal detection were considered significant when identified using both the frequentist statistical approach and BCPNN methods.
Statistical analysis
Continuous variables were compared using Welch’s test. Statistical significance was defined as a two-tailed P-value < 0.05. All statistical analyses were performed using the R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria).
Results
Patient characteristics in FAERS.
After excluding reports that used subjective terms such as “youth” and “elderly,” as well as those lacking information on sex and age, 10,855,976 patient reports (all cases) were obtained in the FAERS dataset. Ciprofloxacin, levofloxacin, moxifloxacin, norfloxacin, and ofloxacin were used in 44,814, 39,375, 17,415, 2,115, and 4,532 cases, respectively. Tables 3 and 4 show the characteristics of the patients using these fluoroquinolones and their distribution according to sex and age.
Fluoroquinolone-associated muscle, tendon, and ligament injuries and tendon disorders in FAERS
Using the three disproportionality algorithms, we summarized the signals for muscle, tendon, and ligament injuries and tendon disorders in the overall FAERS dataset (Tables 5 and 6). Five fluoroquinolones (levofloxacin, ciprofloxacin, moxifloxacin, norfloxacin, and ofloxacin) exhibited positive disproportionality signals for both outcomes. Levofloxacin exhibited positive disproportionality signals for muscle, tendon, and ligament injuries (N11, 5,735; ROR, 38.7; 95% CI, 37.6–39.9; IC, 4.8; 95% CrI, 4.8–4.9) and tendon disorders (N11, 6,524; ROR, 82.6; 95% CI, 80.3–85.1; IC, 5.7; 95% CrI, 5.7–5.8) with the highest positive signal values. Ciprofloxacin also exhibited positive disproportionality signals for muscle, tendon, and ligament injuries (N11, 4,387; ROR, 23.9; 95% CI, 23.2–24.7; IC, 4.3; 95% CrI, 4.2–4.3) and tendon disorders (N11, 5,619; ROR, 57.6; 95% CI, 55.9–59.4; IC, 5.3; 95% CrI, 5.3–5.4). In subsequent analyses, we restricted the FAERS data to reports submitted from Japan (FAERS-Japan) and evaluated the same endpoints (Tables 5 and 6). In FAERS-Japan, levofloxacin exhibited positive disproportionality signals for muscle, tendon, and ligament injuries (N11, 24; ROR, 10.2; 95% CI, 6.7–15.4; IC, 3.0; 95% CrI, 2.3–3.5) and tendon disorders (N11, 21; ROR, 10.2; 95% CI, 6.7–15.4; IC, 3.8; 95% CrI, 3.1–4.3). There were no reports of muscle, tendon, or ligament injuries or tendon disorders involving norfloxacin or ofloxacin, and no positive disproportionality signals were detected for ciprofloxacin or moxifloxacin.
Age and sex stratification of signals in FAERS compared with JADER
As in FAERS, disproportionality analyses were performed in JADER; however, there were no reports of muscle, tendon, or ligament injuries or tendon disorders with moxifloxacin or ofloxacin, and N11 was low for norfloxacin (1 and 1). For ciprofloxacin, a weak positive disproportionality signal was exhibited for tendon disorders (N11, 4; ROR, 7.6; 95% CI, 2.8–20.4; IC, 2.1; 95% CrI, 0.3–3.2), whereas no signal was detected for muscle, tendon and ligament injuries (N11, 3; ROR, 2.0; 95% CI, 0.6–6.3; IC, 0.8; 95% CrI, −1.2–2.0). For levofloxacin, disproportionality signals for muscle, tendon, and ligament injuries and tendon disorders were compared between the overall FAERS and JADER datasets (Tables 7 and 8). In JADER, levofloxacin exhibited positive disproportionality signals for muscle, tendon, and ligament injuries (N11, 247; ROR, 30.6; 95% CI, 26.6–35.2; IC, 4.5; 95% CrI, 4.3–4.6) and tendon disorders (N11, 245; ROR, 134.2; 95% CI, 112.5–160.2; IC, 5.8; 95% CrI, 5.6–6.0). For levofloxacin, sex-stratified analyses detected positive disproportionality signals for muscle, tendon, and ligament injuries and tendon disorders in both males and females in the FAERS and JADER databases. In both databases, the signals were significantly stronger in males than in females. In age-stratified analyses of JADER, there were no reports of muscle, tendon, or ligament injuries or tendon disorders among males aged <20 years, and no reports among females aged <30 years or ≥90 years. In JADER, positive disproportionality signals were identified for muscle, tendon, and ligament injuries across all strata except among females aged ≥70 years. Contrastingly, tendon disorders showed positive disproportionality signals in both males and females across both age groups (≥ 70 and < 70 years) (Table 9). In the ≥ 70 years group, a significantly stronger disproportionality signal was detected in male than in female both muscle, tendon and ligament injuries (rROR, 12.21; 95% CI, 4.85–30.72; ICΔ, 3.18; 95% CrI, 2.81–3.45 [reference: female]) and tendon disorders (rROR, 25.67; 95% CI, 8.37–78.72; ICΔ, 2.85; 95% CrI, 2.48–3.11 [reference: female]).
Comparison by route of administration in JADER
For levofloxacin, in the route-of-administration-stratified analyses, positive disproportionality signals for muscle, tendon, and ligament injuries and tendon disorders were detected irrespective of sex, and these signals were stronger with oral administration than with intravenous administration (Table 10).
Time-to-onset of levofloxacin-associated muscle, tendon, and ligament injuries and tendon disorders
For levofloxacin, the time-to-onset was available for 73 cases of muscle, tendon, and ligament injuries (54 males and 19 females) and 71 cases of tendon disorders (54 males and 17 females) in JADER. The time to onset of muscle, tendon, and ligament injuries and tendon disorders for levofloxacin is shown in Fig 1. Notably, most levofloxacin-associated muscle, tendon, and ligament injuries (median, 7 days; interquartile range [IQR], 4–11 days) and tendon disorders (median, 7 days; IQR, 4–11 days) occurred within 14 days. For levofloxacin, the time to onset did not differ by sex: muscle, tendon, and ligament injuries, 7 (4–10.75) days in males vs. 8 (3.5–11) days in females (P = 0.29); tendon disorders, 7 (4–10.75) days vs. 8 (3.5–11) days (P = 0.66). It also did not differ by age (≥ 70 vs < 70 years): muscle, tendon, and ligament injuries, 6 (4–10) days vs 7 (4–11.75) days (P = 0.66); tendon disorders, 6.5 (4–10.75) days vs 7 (4–11) days (P = 0.55).
Discussion
Tendinitis and tendon rupture may occur in patients treated with fluoroquinolones. However, the onset characteristics of these adverse events have not yet been fully elucidated. This study was conducted using the FAERS and JADER databases to clarify the association between individual fluoroquinolones and tendon disorders and to assess differences by age, sex, and route of administration in the onset of muscle, tendon, and ligament injuries and tendon disorders associated with fluoroquinolones in the Japanese population, particularly levofloxacin.
In the overall FAERS, multiple fluoroquinolones showed positive disproportionality signals for muscle, tendon, and ligament injuries and for tendon disorders: ciprofloxacin (ROR025, 23.2 and 55.9; IC025, 4.2 and 5.3), levofloxacin (ROR025, 37.6 and 80.3; IC025, 4.8 and 5.7), moxifloxacin (ROR025, 8.4 and 15.4; IC025, 2.9 and 3.8), norfloxacin (ROR025, 9.7 and 17.0; IC025, 3.1 and 3.8), and ofloxacin (ROR025, 11.0 and 23.1; IC025, 3.3 and 4.3) (Tables 5 and 6). In the FAERS restricted to reports from Japan, signals were detected only for levofloxacin (ROR025, 6.7 and 6.7; IC025, 2.3 and 3.1) (Tables 5 and 6). This pattern aligns with the small contribution of Asian reports to the FAERS, as analyses of data on ciprofloxacin, levofloxacin, and moxifloxacin showed that Asian reports comprised 1.02% of suspected tendinitis and 1.28% of tendon rupture cases [11]. Accordingly, FAERS alone may be inadequate for analyses focused on Asia or Japan; therefore, JADER was used as a complementary data source. In JADER, although a weak positive disproportionality signal for tendon disorders was observed for ciprofloxacin (ROR025, 2.8; IC025, 0.3), strong signals for muscle, tendon, and ligament injuries and tendon disorders were most consistently observed for levofloxacin (ROR025, 26.6 and 112.5; IC025, 4.3 and 5.6). Findings from the FAERS reports restricted to Japan were consistent with those from JADER, identifying levofloxacin as the agent with the most consistent signal. These differences may reflect region-specific factors, including prescription patterns, drug utilization, concomitant medications, and potential genetic background. Because the preferred fluoroquinolone agents, indications, and routes of administration may differ across countries and regions, Japan-focused pharmacovigilance analyses are important for interpreting agent-specific safety signals in Japanese clinical practice. In the present analysis, levofloxacin showed the most consistent and robust disproportionality signals among the fluoroquinolones evaluated in FAERS-Japan and JADER. Consistent with this, a nationwide population-based cohort study in Taiwan reported that, after adjusting for seven fluoroquinolone agents, levofloxacin, norfloxacin, and pefloxacin were associated with a higher risk of tendon disorders [23]. These observations indicate that levofloxacin has the most consistent association with muscle, tendon, and ligament injuries and tendon disorders. In light of the Taiwanese cohort data, this raises the possibility of region-specific risk patterns within Asia.
In JADER, signals were absent for moxifloxacin and ofloxacin, and negligible for norfloxacin; ciprofloxacin showed a weak signal restricted to tendon disorders. Contrastingly, for levofloxacin, positive disproportionality signals for muscle, tendon, and ligament injuries and tendon disorders were detectable in both sexes but were consistently stronger in males than in females (Table 9). Despite these differences in signal strength by age and sex, time-to-onset distributions did not differ by sex or age in the JADER database; most levofloxacin-associated muscle, tendon, and ligament injuries and tendon disorders occurred within 14 days. Previous studies indicated that fluoroquinolone-associated tendon injuries increase with age, particularly among older adults. A meta-analysis on fluoroquinolones reported significantly elevated risks for Achilles tendinitis and rupture, remaining significant among patients ≥ 60 years. Complementary pharmacoepidemiologic studies showed a higher absolute risk in older adults [2,9]. Similarly, a nationwide cohort study in Taiwan observed an increased risk among those aged > 60 years [23]. Sex differences in fluoroquinolone-associated tendon injuries remain controversial. A large United Kingdom case-crossover study using the Health Improvement Network database suggested slightly stronger associations in females (Achilles tendinitis odds ratio, 5.0 vs 3.6), although the difference was not statistically significant [24]. Other FAERS-based analyses have reported a marginally higher proportion of female reports [11]. Conversely, several studies have reported male predominance of fluoroquinolone-associated tendinopathy [25,26]. While VigiBase noted a slight male predominance in reports within the ≥ 65 years subgroup across fluoroquinolones [10], our Japan-focused, age- and sex-stratified disproportionality analyses demonstrated a more pronounced sex difference at older ages, with males aged ≥ 70 years exhibiting significantly stronger levofloxacin disproportionality signals than females. Baseline epidemiology indicates that Achilles tendon rupture is more common in males, which may partly account for sex-specific patterns in pharmacovigilance datasets [27]. Nevertheless, our analyses suggest that older males, especially those aged ≥ 70 years, showed stronger disproportionality signals for levofloxacin-associated tendon disorders. Accordingly, these findings support increased clinical vigilance when prescribing levofloxacin to this population.
The mechanism of fluoroquinolone-induced tendon disorders remains unclear, but several mechanistic hypotheses have been proposed, including extracellular matrix degradation, impaired repair dynamics, oxidative stress, and mitochondrial dysfunction. Although mechanisms specific to levofloxacin have not been fully established, levofloxacin-associated tendon disorders may involve mechanisms proposed for fluoroquinolone-associated tendon toxicity. These processes may weaken tendon extracellular matrix integrity and reduce tendon resilience, particularly in older tendons. Older tendons exhibit decreased cellularity, impaired extracellular matrix remodeling, increased MMP-2 and MMP-9 activities, and reduced expression of tissue inhibitors of metalloproteinases, which potentially act via pathways overlapping those implicated in fluoroquinolone-associated injury [28,29]. Therefore, aging-related tendon vulnerability may have contributed to the stronger disproportionality signals in older patients.
Levofloxacin is commonly administered orally or intravenously. By analyzing FAERS, Shu et al. detected markedly stronger disproportionality signals for tendonitis and tendon rupture with oral versus intravenous fluoroquinolones, including levofloxacin [11]. In JADER, our levofloxacin-focused analysis showed a similar pattern of differential signal strength according to the patient subgroup and route. For muscle, tendon, and ligament injuries, signals were higher with oral administration than with intravenous injections (oral: N11, 222; ROR, 34.7; 95% CI, 29.7–40.5; IC, 4.5; 95% CrI, 4.3–4.7; intravenous: N11, 7; ROR, 21.8; 95% CI, 10.2–46.5; IC, 3.1; 95% CrI, 1.8–4.0). Similar disproportionality signals were also detected for tendon disorders (oral: N11, 223; ROR, 194.7; 95% CI, 155.3–244.2; IC, 5.8; 95% CrI, 5.5–5.9; intravenous: N11, 7; ROR, 215.9; 95% CI, 93.5–478.4; IC, 3.7; 95% CrI, 2.4–4.6). Furthermore, sex-stratified analyses showed that oral levofloxacin consistently yielded stronger disproportionality signals than intravenous levofloxacin. This pattern is consistent with the FAERS-based analyses findings for levofloxacin and indicates that the effect of the route of administration is replicated across spontaneous reporting systems. However, the reasons for the stronger oral signals remain uncertain. The route-specific difference may partly reflect differences in clinical setting and patient activity in addition to the route of administration. Oral levofloxacin may be more often used in outpatient settings, where patients may remain ambulatory and have greater physical activity-related tendon loading. In contrast, intravenous levofloxacin may be more often used in hospitalized or acutely ill patients who are relatively less active. Such tendon loading could act as a co-factor in tendon injury and may contribute to differences in reporting patterns by route of administration. However, this interpretation remains speculative because FAERS and JADER do not contain information on patient activity level, mobility, or tendon loading. Accordingly, the contribution of physical activity-related tendon loading could not be directly evaluated in this study. Differences in indication, disease severity, and care setting may also have influenced the observed route-specific signals.
This study had some limitations. The analysis relied on spontaneous reporting systems, such as FAERS and JADER. The data quality depends on voluntary reporting and the reporter’s expertise. Underreporting, selective or stimulated reporting, duplicate submissions, incomplete information, and misclassification can introduce bias in the estimates [19]. Disproportionality metrics, such as ROR, and IC, reflect reporting rather than incidence or risk and do not establish causality. These metrics cannot fully account for confounding by age, indication, dose, comorbidities, or concomitant medications, and residual confounding factors may remain [30]. Among concomitant medications, systemic corticosteroids are particularly relevant because systemic corticosteroid therapy is a recognized risk factor for tendon injury and has been associated with a high risk of fluoroquinolone-associated tendon disorders [9,24]. In the present analysis, concomitant corticosteroid therapy and other concomitant medications were not used as exclusion criteria, and their timing, dose, and duration could not be reliably determined from FAERS or JADER. Therefore, residual confounding by corticosteroids and other concomitant medications may have influenced the observed disproportionality signals.
Renal function represents another important source of potential residual confounding, particularly in the interpretation of levofloxacin-associated signals, because impaired renal function may increase systemic exposure. However, FAERS and JADER do not contain structured laboratory data, such as serum creatinine or estimated glomerular filtration rate, and renal impairment-related Preferred Terms recorded as medical history, comorbid conditions, or adverse events do not necessarily indicate baseline renal function at the time of fluoroquinolone initiation. In addition, clinicians may avoid fluoroquinolones or adjust their doses in patients with renal impairment. Therefore, the reporting patterns in patients with renal impairment may not reflect the association under comparable dosing or exposure conditions. Accordingly, we could not reliably stratify reports by renal function or evaluate dose-adjusted exposure, and the present levofloxacin findings should be interpreted with caution in light of possible residual renal confounding.
Time-to-onset analyses based on spontaneous reporting systems should also be interpreted cautiously [31]. In FAERS and JADER, onset dates and treatment initiation dates may be incomplete, imprecise, or inconsistently recorded. In addition, tendon-related symptoms may develop gradually, and the reported onset date may reflect the time of clinical recognition rather than the true biological onset. Because only reports with sufficient temporal information were included in the time-to-onset analysis, selection bias may also have influenced the observed onset patterns. Therefore, the present time-to-onset findings should be regarded as descriptive and hypothesis-generating rather than precise estimates of onset timing.
Finally, exposure denominators are not available, which prevents the estimation of absolute risks and the comparison of incidence across fluoroquinolones [19]. Therefore, the findings should be viewed as hypothesis-generating and should be confirmed in population-based studies with known denominators and rigorous confounding control.
In this study, we analyzed FAERS data to characterize Japan-specific disproportionality for muscle, tendon, and ligament injuries and tendon disorders; however, this approach was insufficient. Therefore, we analyzed JADER data to obtain Japan-focused signal estimates. In JADER, only levofloxacin showed positive signals for both adverse effects. Signals were detectable in both sexes but were consistently higher in males, particularly in those aged ≥ 70 years. In such cases, levofloxacin should be used with caution, considering the potential for tendon-associated adverse events. Most levofloxacin-associated tendon disorders occurred within 14 days after treatment initiation. These findings support close monitoring for symptoms suggestive of tendon injury during the first 2 weeks after levofloxacin initiation, particularly in older male patients. Further population-based studies are needed to confirm these findings and to evaluate the contribution of physical activity-related tendon loading to tendon injury in this demographic.
Supporting information
S1 Table. Drug identification terms used for fluoroquinolones.
https://doi.org/10.1371/journal.pone.0358427.s001
(DOCX)
S2 Table. Preferred Terms included in the target HLTs.
https://doi.org/10.1371/journal.pone.0358427.s002
(DOCX)
References
- 1. Robson RA. Quinolone pharmacokinetics. Int J Antimicrob Agents. 1992;2(1):3–10. pmid:18611512
- 2. Alves C, Mendes D, Marques FB. Fluoroquinolones and the risk of tendon injury: a systematic review and meta-analysis. Eur J Clin Pharmacol. 2019;75(10):1431–43. pmid:31270563
- 3. Anwar AI, Lu L, Plaisance CJ, Daniel CP, Flanagan CJ, Wenger DM. Fluoroquinolones: neurological complications and side effects in clinical practice. Cureus. 2024;16(2):e54565. pmid:38516474
- 4. Briasoulis A, Agarwal V, Pierce WJ. QT prolongation and torsade de pointes induced by fluoroquinolones: infrequent side effects from commonly used medications. Cardiology. 2011;120(2):103–10. pmid:22156660
- 5. Brown KA, Khanafer N, Daneman N, Fisman DN. Meta-analysis of antibiotics and the risk of community-associated Clostridium difficile infection. Antimicrob Agents Chemother. 2013;57(5):2326–32. pmid:23478961
- 6. Chou H-W, Wang J-L, Chang C-H, Lee J-J, Shau W-Y, Lai M-S. Risk of severe dysglycemia among diabetic patients receiving levofloxacin, ciprofloxacin, or moxifloxacin in Taiwan. Clin Infect Dis. 2013;57(7):971–80. pmid:23948133
- 7. Morales D, Pacurariu A, Slattery J, Pinheiro L, McGettigan P, Kurz X. Association between peripheral neuropathy and exposure to oral fluoroquinolone or amoxicillin-clavulanate therapy. JAMA Neurol. 2019;76(7):827–33. pmid:31034074
- 8. Rawla P, El Helou ML, Vellipuram AR. Fluoroquinolones and the risk of aortic aneurysm or aortic dissection: a systematic review and meta-analysis. Cardiovasc Hematol Agents Med Chem. 2019;17(1):3–10. pmid:30947680
- 9. Morales DR, Slattery J, Pacurariu A, Pinheiro L, McGettigan P, Kurz X. Relative and absolute risk of tendon rupture with fluoroquinolone and concomitant fluoroquinolone/corticosteroid therapy: population-based nested case-control study. Clin Drug Investig. 2019;39(2):205–13. pmid:30465300
- 10. Huruba M, Farcas A, Leucuta DC, Bucsa C, Sipos M, Mogosan C. A VigiBase descriptive study of fluoroquinolone induced disabling and potentially permanent musculoskeletal and connective tissue disorders. Sci Rep. 2021;11(1):14375. pmid:34257376
- 11. Shu Y, Zhang Q, He X, Liu Y, Wu P, Chen L. Fluoroquinolone-associated suspected tendonitis and tendon rupture: a pharmacovigilance analysis from 2016 to 2021 based on the FAERS database. Front Pharmacol. 2022;13:990241. pmid:36147351
- 12. Hori K, Yamakawa K, Yoshida N, Ohnishi K, Kawakami J. Detection of fluoroquinolone-induced tendon disorders using a hospital database in Japan. Pharmacoepidemiol Drug Saf. 2012;21(8):886–9. pmid:22532516
- 13. Corps AN, Harrall RL, Curry VA, Fenwick SA, Hazleman BL, Riley GP. Ciprofloxacin enhances the stimulation of matrix metalloproteinase 3 expression by interleukin-1beta in human tendon-derived cells. A potential mechanism of fluoroquinolone-induced tendinopathy. Arthritis Rheum. 2002;46(11):3034–40. pmid:12428247
- 14. Tsai W-C, Yang Y-M. Fluoroquinolone-associated tendinopathy. Chang Gung Med J. 2011;34(5):461–7. pmid:22035890
- 15. Tsai W-C, Hsu C-C, Chen H-C, Hsu Y-H, Lin M-S, Wu C-W, et al. Ciprofloxacin-mediated inhibition of tenocyte migration and down-regulation of focal adhesion kinase phosphorylation. Eur J Pharmacol. 2009;607(1–3):23–6. pmid:19232343
- 16. Simonin MA, Gegout-Pottie P, Minn A, Gillet P, Netter P, Terlain B. Pefloxacin-induced achilles tendon toxicity in rodents: biochemical changes in proteoglycan synthesis and oxidative damage to collagen. Antimicrob Agents Chemother. 2000;44(4):867–72. pmid:10722483
- 17. Fusaroli M, Salvo F, Begaud B, AlShammari TM, Bate A, Battini V, et al. The REporting of A disproportionality analysis for DrUg safety signal detection using individual case safety reports in pharmacovigilance (READUS-PV): explanation and elaboration. Drug Saf. 2024;47(6):585–99. pmid:38713347
- 18. Fusaroli M, Salvo F, Begaud B, AlShammari TM, Bate A, Battini V, et al. The reporting of a disproportionality analysis for drug safety signal detection using individual case safety Reports in PharmacoVigilance (READUS-PV): development and statement. Drug Saf. 2024;47(6):575–84. pmid:38713346
- 19. Noguchi Y, Tachi T, Teramachi H. Detection algorithms and attentive points of safety signal using spontaneous reporting systems as a clinical data source. Brief Bioinform. 2021;22(6):bbab347. pmid:34453158
- 20. Noguchi Y, Yoshimura T. Detection algorithms for simple two-group comparisons using spontaneous reporting systems. Drug Saf. 2024;47(6):535–43. pmid:38388828
- 21. Rothman KJ, Lanes S, Sacks ST. The reporting odds ratio and its advantages over the proportional reporting ratio. Pharmacoepidemiol Drug Saf. 2004;13(8):519–23. pmid:15317031
- 22. Bate A, Lindquist M, Edwards IR, Olsson S, Orre R, Lansner A, et al. A Bayesian neural network method for adverse drug reaction signal generation. Eur J Clin Pharmacol. 1998;54(4):315–21. pmid:9696956
- 23. Chang C-K, Chien W-C, Hsu W-F, Chiao H-Y, Chung C-H, Tzeng Y-S, et al. Positive association between fluoroquinolone exposure and tendon disorders: a nationwide population-based cohort study in Taiwan. Front Pharmacol. 2022;13:814333. pmid:35387340
- 24. Wise BL, Peloquin C, Choi H, Lane NE, Zhang Y. Impact of age, sex, obesity, and steroid use on quinolone-associated tendon disorders. Am J Med. 2012;125(12). pmid:23026288
- 25. Khaliq Y, Zhanel GG. Fluoroquinolone-associated tendinopathy: a critical review of the literature. Clin Infect Dis. 2003;36(11):1404–10. pmid:12766835
- 26. Seeger JD, West WA, Fife D, Noel GJ, Johnson LN, Walker AM. Achilles tendon rupture and its association with fluoroquinolone antibiotics and other potential risk factors in a managed care population. Pharmacoepidemiol Drug Saf. 2006;15(11):784–92. pmid:16456878
- 27. Lemme NJ, Li NY, DeFroda SF, Kleiner J, Owens BD. Epidemiology of achilles tendon ruptures in the united states: athletic and nonathletic injuries from 2012 to 2016. Orthop J Sports Med. 2018;6(11):2325967118808238. pmid:30505872
- 28. Kwan KYC, Ng KWK, Rao Y, Zhu C, Qi S, Tuan RS, et al. Effect of aging on tendon biology, biomechanics and implications for treatment approaches. Int J Mol Sci. 2023;24(20):15183. pmid:37894875
- 29. Yu T-Y, Pang J-HS, Wu KP-H, Chen MJ-L, Chen C-H, Tsai W-C. Aging is associated with increased activities of matrix metalloproteinase-2 and -9 in tenocytes. BMC Musculoskelet Disord. 2013;14:2. pmid:23281803
- 30. Noguchi Y, Tachi T, Yoshimura T. Caveats of covariate adjustment in disproportionality analysis for best practices. Drug Saf. 2025;48(1):1–5. pmid:39154117
- 31. Noguchi Y, Ino Y, Yokoyama S, Tachi T, Yoshimura T. Theoretical framework and key considerations for time-to-onset analysis in spontaneous reporting systems. Drug Saf. 2026. pmid:42118500