Fluoroquinolone Exposure and Cancer Risk in Interstitial Lung Disease: A Propensity-Score-Matched Cohort Study Using Cox and Competing-Risk Models
by
,
Yi-Fan Sun
1,†, 
Yu-Ting Chiu
1,2,†,
Yung-En Ko
1,
Yu-Wei Huang
1,
Liang-Kai Hsieh
1,
Cheng-Li Lin
3,4,
Chia-Hung Kao
5,6,7,8,*,‡ and
Jun-Jun Yeh
1,*,‡
1
Department of Family Medicine, Geriatric Medicine, Chest Medicine and Medical Research, Ditmanson Medical Foundation Chia-Yi Christian Hospital, No. 539, Zhongxiao Rd., East Dist., Chiayi 600566, Taiwan
2
Department of Laboratory Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi 600566, Taiwan
3
College of Medicine, China Medical University, Taichung 40402, Taiwan
4
Management Office for Health Data, China Medical University Hospital, Taichung 40402, Taiwan
5
Artificial Intelligence and Robotics Innovation Center, China Medical University Hospital, Taichung 40402, Taiwan
6
Department of Nuclear Medicine and PET Center, China Medical University Hospital, Taichung 40402, Taiwan
7
Graduate Institute of Biomedical Sciences, School of Medicine, College of Medicine, China Medical University, No. 2, Yuh-Der Road, Taichung 40402, Taiwan
8
Department of Bioinformatics and Medical Engineering, Asia University, Taichung 41354, Taiwan
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
‡
These authors also contributed equally to this work.
Pharmaceuticals 2026, 19(7), 1067; https://doi.org/10.3390/ph19071067
Submission received: 4 May 2026
/
Revised: 6 July 2026
/
Accepted: 9 July 2026
/
Published: 10 July 2026
(This article belongs to the Section Pharmacology)
Abstract
Background: This study aimed to comprehensively investigate the complex association between the use of fluoroquinolone (FQ) antibiotics and cancer risk, with a specific focus on patients with interstitial lung disease (ILD)—a unique clinical population characterized by a high inflammatory burden and a high susceptibility to infections. Methods: We conducted a large-scale retrospective cohort study using a high-quality clinical database. A total of 7906 matched patients (3953 pairs) were included after propensity score matching (PSM). Three complementary statistical models were applied: the standard Cox proportional hazards model, the time-dependent Cox regression model, and the Fine–Gray competing-risks model, to provide a multidimensional assessment of cancer risk. Results: A total of 7906 matched patients (3953 pairs) were followed. After strictly defining the index date to eliminate immortal time bias, FQ exposure was associated with an increased risk of all-cause cancer in the standard Cox model (adjusted HR 1.45; 95% CI, 1.20–1.76) and the competing risk model (adjusted SHR 1.28; 95% CI, 1.06–1.55). Site-specific analyses revealed elevated risks for certain malignancies, notably prostate cancer. Importantly, when modeled as a continuous variable, the cumulative dose of fluoroquinolones showed no significant dose–response relationship with overall cancer risk (adjusted HR 0.99; 95% CI, 0.99–1.00). Conclusions: After correcting for immortal time bias, the previously hypothesized protective effect of fluoroquinolones on cancer risk was not observed. The increased risk observed in categorical models, coupled with a lack of a continuous dose–response, strongly suggests that these findings are driven by confounding by indication and reverse causation (i.e., frequent infections masking undiagnosed malignancies or reflecting severe underlying ILD), rather than a direct pharmacological effect.
1. Introduction
1.1. Research Background and Rationale
Since their introduction in the 1980s, fluoroquinolone (FQ) antibiotics have become a cornerstone of global clinical practice due to their broad-spectrum antimicrobial activity, excellent tissue penetration, and high oral bioavailability [1]. They are widely prescribed for respiratory, urinary tract, and intra-abdominal infections [1]. However, with their exponential increase in use, concerns regarding their long-term safety have intensified. Beyond well-documented adverse effects such as tendinopathy, aortic aneurysm, and dysglycemia, whether FQs possess carcinogenic or anticarcinogenic properties has emerged as a major point of contention in pharmacoepidemiology [2,3,4,5].
From a molecular perspective, FQs exert their antibacterial effects by inhibiting bacterial DNA gyrase and topoisomerase IV, thereby disrupting DNA replication. Although their affinity for eukaryotic topoisomerases is substantially lower, high concentrations may induce DNA double-strand breaks, oxidative stress, and mitochondrial dysfunction in mammalian cells, theoretically conferring mutagenic and carcinogenic potential [2,3,6]. Conversely, chronic inflammation is widely recognized as a central driver of carcinogenesis. Beyond their antimicrobial effects, FQs have demonstrated immunomodulatory properties, including the suppression of pro-inflammatory cytokines (e.g., IL-6, TNF-α) and matrix metalloproteinase (MMP) activity [2,3,4,7]. These properties suggest that FQs may interrupt the “infection–inflammation–cancer” axis and exert chemopreventive effects [8,9].
1.2. The Unique Risk Profile of Patients with Interstitial Lung Disease
Patients with interstitial lung disease (ILD) represent a population of exceptional clinical and pathophysiological relevance. ILD comprises a heterogeneous group of disorders characterized by pulmonary interstitial fibrosis and chronic inflammatory cell infiltration [10]. These patients inherently face a markedly elevated risk of lung cancer—estimated to be 3–5 times higher than that of the general population [11]. Moreover, structural lung damage and the widespread use of immunosuppressive agents (e.g., corticosteroids, cyclophosphamide) predispose ILD patients to recurrent respiratory infections, resulting in substantially higher rates of FQ prescriptions compared with the general population [1,12]. This dual profile of “high baseline cancer risk” and “high drug exposure” renders the ILD cohort a natural experimental setting for evaluating the long-term carcinogenic or protective effects of FQs [11,12,13]. If FQs were truly carcinogenic, their effects would be amplified in the context of ILD-related genomic instability [13,14]. Conversely, if FQs effectively suppress infection-driven inflammatory cascades, their protective effects should be most evident in this population [4,7].
1.3. Study Objectives and Core Research Questions
Recognizing that prior observational studies have been heavily influenced by confounding by indication and detection bias, we employed propensity score matching to construct a highly homogeneous comparison cohort. This study aimed to address the following key scientific questions: (1) Overall effect: After adjusting for comorbidities and concomitant medications, does FQ use increase or decrease the overall cancer risk among patients with ILD? (2) Site-specific effects: Is there heterogeneity in cancer risk across different organ systems (e.g., lung, gastrointestinal tract, genitourinary system)? (3) Dose–response relationship: Do cumulative exposure duration and cumulative dose exhibit threshold or linear associations with cancer risk? (4) Methodological validation: How do standard Cox models, time-dependent models, and competing risk models differ in the context of high mortality among ILD patients, and which approach most closely reflects biological reality?
2. Results
After propensity score matching and applying the rigorous index date assignment to avoid immortal time bias, 3953 patients were included in each cohort, yielding a total study population of 7906 individuals. The distributions of age, sex, baseline comorbidities, and concomitant medications were well balanced between fluoroquinolone (FQ) users and non-users (Table 1; Figure 1). The mean follow-up duration was 5.43 years for FQ users and 5.95 years for non-users. During follow-up, the cumulative incidence of cancer was higher among FQ users than among non-users, as illustrated by the Kaplan–Meier curves (Figure 2), and the difference between the two groups was statistically significant based on the log-rank test.
In the matched cohort, after accounting for immortal time bias, FQ use was associated with a higher risk of overall cancer compared with non-use. In the fully adjusted Cox proportional hazards model, the adjusted hazard ratio (aHR) for overall cancer among FQ users was 1.45 (95% CI, 1.20–1.76; p < 0.001) (Table 2).
Compared with non-users, FQ users had significantly higher risks of prostate cancer (aHR 5.50, 95% CI 2.25–13.48). Conversely, a lower risk was observed for immune-related cancers (aHR 0.62, 95% CI 0.44–0.87). No statistically significant associations were observed for several other cancer types, including lung, liver, and colorectal cancers (Table 3).
When stratified by cumulative duration of FQ therapy, no significant dose–response relationship or risk reduction was observed (Table 4). For colorectal cancer, the adjusted hazard ratios across all duration strata (≤35 days, 36–55 days, and >56 days) indicated non-significant increased risks (aHRs ranging from 1.01 to 1.53, all 95% CIs crossing 1.00). Similarly, for lung cancer, liver cancer, and immune-related cancers, the risks among FQ users were not significantly different from those of non-users, regardless of the cumulative exposure duration. These findings indicate that the previously hypothesized dose-dependent protective effects were not present after accounting for immortal time bias.
Analyses stratified by cumulative dose of FQs revealed no significant dose–response relationship (Table 5). Compared with non-users, patients receiving varying cumulative doses of FQs (≤3000 mg, 3001–6000 mg, and >6000 mg) did not exhibit any statistically significant reductions in cancer risk. For colorectal cancer, liver cancer, lung cancer, and immune-related cancers, the adjusted hazard ratios across all dose strata were not statistically significant, with all 95% confidence intervals crossing 1.00. These findings further corroborate that the previously observed dose-dependent protective associations were likely artifacts of immortal time bias, and that the cumulative dose of FQs does not directly modulate cancer risk in this cohort.
In the time-dependent Cox regression model, which treated FQ exposure as a time-varying variable, the association with an increased risk of cancer was further corroborated (Table 6). FQ exposure was associated with a significantly increased risk of overall cancer (aHR 2.07, 95% CI 1.57–2.72). Markedly elevated risks were observed for prostate cancer (aHR 9.95, 95% CI 4.75–20.8) and lung cancer (aHR 3.35, 95% CI 1.87–6.01), as well as for hematologic malignancies (aHR 4.09, 95% CI 1.51–11.1) and cancers of the bladder and kidney (aHR 3.72, 95% CI 1.55–8.93).
Because of the high mortality rate in patients with interstitial lung disease, a competing risk analysis using the Fine–Gray subdistribution hazard model was performed, treating death as a competing event (Table 7). In this model, consistent with the updated primary Cox analysis, FQ use remained associated with a significantly increased risk of overall cancer (adjusted subdistribution hazard ratio [aSHR] 1.28, 95% CI 1.06–1.55). However, no significant associations were observed for specific cancer subtypes, including colorectal cancer (aSHR 1.24, 95% CI 0.69–2.22), liver cancer (aSHR 0.88, 95% CI 0.53–1.47), lung cancer (aSHR 1.04, 95% CI 0.65–1.68), and immune-related cancers (aSHR 0.91, 95% CI 0.59–1.42).
Continuous Exposure Variable Analysis
To robustly evaluate the dose–response relationship and address concerns regarding arbitrary exposure cutoffs, fluoroquinolone exposure was analyzed as a continuous variable (Table 8). In this model, the continuous cumulative dose of fluoroquinolones was not significantly associated with the risk of all-cause cancer (aHR 0.99, 95% CI 0.99–1.00). This lack of a continuous dose–response relationship further suggests that the significant associations observed in the categorical models may be driven by confounding factors rather than direct pharmacological effects.
3. Discussion
3.1. Main Findings
In this propensity-score–matched cohort study of patients with ILD, FQs exposure was associated with a higher long-term risk of overall cancer after rigorous index-date assignment to reduce immortal time bias. The association remained significant in the standard Cox model and was directionally consistent in the Fine–Gray competing-risk model and time-dependent Cox regression. These findings suggest that the previously observed protective association was not sustained after correction for immortal time bias [15,16,17,18]. Site-specific analyses showed marked heterogeneity. FQs exposure was associated with a substantially increased risk of prostate cancer, whereas immune-related cancers showed a reduced risk. In contrast, no consistent significant association was observed for several other major cancers, including lung, liver, and colorectal cancers. Importantly, cumulative duration- and dose-stratified analyses did not demonstrate a clear linear dose–response relationship. When cumulative dose was modeled as a continuous variable, no significant dose–response association with overall cancer risk was observed. These findings should therefore be interpreted cautiously. Although FQs exposure was associated with increased overall cancer risk, the lack of a clear dose–response relationship argues against a simple direct pharmacological carcinogenic effect. Instead, the observed association may reflect confounding by indication, reverse causation, infection burden, healthcare utilization, or underlying ILD severity. Patients requiring FQs may have more frequent or severe infections, greater systemic inflammation, more clinical encounters, or occult malignancies initially presenting as infectious syndromes. For example, urinary tract symptoms or prostatitis may precede the diagnosis of prostate cancer, potentially explaining the strong association observed for prostate cancer [15,16,17,18,19,20,21,22,23,24,25]. The reduced risk of immune-related cancers may reflect residual confounding, differences in immune-modulating medication use, competing risks, or heterogeneous biological pathways rather than a definitive protective effect. Because claims data do not contain detailed immune profiles, microbiome data, inflammatory biomarkers, smoking intensity, pulmonary function, imaging severity, or histopathological ILD subtype, mechanistic conclusions cannot be drawn [18,19,20,21,22,23,24,25,26]. Overall, our findings indicate that FQs exposure in ILD patients is associated with increased overall cancer risk after correction for immortal time bias, but the absence of a consistent dose–response pattern suggests that this association is more likely influenced by clinical indication and underlying disease complexity than by a direct causal drug effect.
3.2. Strengths
This study has several strengths. First, we used a large nationwide claims-based database with long-term follow-up, allowing assessment of overall and site-specific cancer outcomes in a clinically vulnerable ILD population. Second, propensity score matching was applied to balance baseline demographic characteristics, comorbidities, and concomitant medications between FQ users and non-users. Third, immortal time bias was specifically addressed by defining the index date for FQ users as the first FQ prescription after ILD diagnosis and assigning comparable randomized index dates to non-users. Fourth, multiple statistical approaches, including standard Cox regression, time-dependent Cox regression, and Fine–Gray competing-risk models, were used to test the robustness of the findings. Finally, both cumulative duration and cumulative dose analyses were performed, helping to evaluate whether the observed association followed a biologically plausible dose–response pattern.
3.3. Limitations
Several limitations should be acknowledged. First, this was an observational study; therefore, causal inference cannot be established. Although propensity score matching and multivariable adjustment were performed, residual confounding remains possible. Second, confounding by indication is a major concern because FQs are prescribed for infections that may themselves be related to cancer risk, occult malignancy, or disease severity. Third, reverse causation cannot be excluded, particularly if early cancer manifestations were initially treated as infections before cancer diagnosis. Fourth, claims data lack important clinical information, including smoking status, alcohol use, occupational exposures, ILD severity, pulmonary function, radiological fibrosis, socioeconomic status, body mass index, environmental exposures, microbiome composition, inflammatory biomarkers, and laboratory data. Fifth, medication adherence could not be confirmed, and prescription records may not fully reflect actual drug exposure. Sixth, the lack of a clear dose–response relationship limits biological interpretation and weakens support for a direct carcinogenic effect of FQs. Finally, although competing-risk and time-dependent models were applied, unmeasured differences in mortality, healthcare-seeking behavior, cancer screening intensity, and infection burden may still have influenced the results [27,28,29,30,31,32,33].
4. Materials and Methods
4.1. Data Source
We analyzed the Longitudinal Generation Tracking Database 2000 (LGTD 2000), a representative subset of the National Health Insurance Research Database (NHIRD). The LGTD 2000 contains comprehensive demographic information, inpatient and outpatient records, and data on medications and treatments for 2 million insured individuals. All personal identification numbers were encrypted to ensure patient privacy. This study was approved by the Institutional Review Board (IRB) of the China Medical University Hospital Research Ethics Committee (CMUH112-REC1-117(CR-2)).
4.2. Study Population
Initially, we identified patients with ILD from the Longitudinal Generation Tracking Database (LGTD 2000). The ILD cohort comprised patients who had ≥3 outpatient visits or ≥1 hospitalization for ILD (newly diagnosed) between 2000 and 2020. Diagnoses were identified using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM codes, including 135, 237.7, 272.2, 277.3, 277.8, 500–505, 506.4, 508.1, 508.8, 515–516, 446.21, 446.4, 495, 517.2, 517.8, 518.3, 555, 710, 714.81, 720, and 759.5; ICD-10-CM D86, E78.2, E85, E88.89, J60-61, J62.8, J63, J64, J66, J67, J68.4, J70.1, J70.8, J82, J84, J99, M31.0, M31.3, M32, M33, M34, M35.0, M05.1, M45, K50, Q85.0, Q85.1). Full name can be found in Table S1 of Supplementary Materials.
FQs were indicated for various infections, such as complicated urinary tract infections (UTIsi)), complicated pyelonephritis, chronic bacterial prostatitis, intra-abdominal infections, and lower respiratory tract infections (including acute exacerbations of COPD and nosocomial pneumonia). Other indications included tuberculosis (TB), sinusitis, skin and soft-tissue infections (e.g., gout with bacterial infection), bone and joint infections (e.g., periodontal diseases), and febrile neutropenia (e.g., in patients with diabetes mellitus [DM]). To strictly prevent immortal time bias, we applied a rigorous index date assignment strategy. For FQ users, the index date was defined strictly as the date of their first FQ prescription following the ILD diagnosis, and follow-up commenced only from this date. Any person-time prior to this first prescription was not attributed to the exposed group. For non-users, to mimic the time-to-exposure distribution of the user cohort and ensure comparable follow-up periods, we assigned a randomized index date distributed between their ILD diagnosis and the end of the study period. Patients who developed cancer or died before their respective index dates were excluded. Those who never received FQs after the ILD diagnosis were designated as non-users. Patients aged ≤ 18 years, who had a history of cancer before participating in this study or had incomplete demographic information, were excluded. To ensure a balanced comparison, the study employed propensity score matching to select FQ users and non-users by age group, sex, and index year, resulting in a total sample of 7906 patients.
4.3. Main Outcome and Covariates
The primary endpoint of this study was the development of cancer (ICD-9-CM code: 140-208; ICD-10-CM C00-C99), as evidenced by a major illness or injury certificate for cancer. The study endpoint was defined as the occurrence of cancer, withdrawal from the insurance program, or 31 December 2021.
4.4. Selection of Comorbidities and Medications
In Taiwan, respiratory, gastrointestinal, and genitourinary syndromes are the leading causes of emergency department visits. Clinically, pneumonia, gastroenteritis, and UTI are among the most prevalent diagnoses in hospital settings [15]. It is noteworthy that certain malignancies may initially present as common infections, particularly in immunocompromised patients or those with metabolic syndrome; for instance, obstructive pneumonia may mask underlying lung cancer, UTIs may be associated with prostate cancer, and gastroenteritis may manifest in patients with colon cancer. To account for these potential confounding factors, this study included patients with cancer-associated infections—such as tuberculosis, pneumonia, peptic ulcer disease, irritable bowel syndrome, and pelvic inflammatory disease—as well as those with benign prostatic hypertrophy [15,16]. Furthermore, immunocompromised individuals and those with metabolic or chronic conditions (e.g., COPD, cirrhosis, atopic diseases, chronic kidney disease, congestive heart failure, stroke, hypertension, hyperlipidemia, and diabetes mellitus) were incorporated into the analysis. Notably, the use of anti-inflammatory medicines, or immunosuppressants, was specifically included as a key pharmacological factor in our analysis [17]. Immune-related cancers were defined as malignancies potentially associated with immune dysregulation, chronic inflammation, infection-related carcinogenesis, or impaired immune surveillance, including hematologic malignancies, infection-related cancers, inflammation-associated cancers, and cancers commonly observed under immunosuppressed conditions [14,15,16,17].
4.5. Statistical Analysis
Categorical variables were expressed as numbers and percentages, and differences between the two cohorts were examined using the chi-square test. Continuous variables were presented as a mean and standard deviation, and differences between the two cohorts were assessed using Student’s t-test. The incidence rate was expressed as per 1000 person-years. Cox proportional hazard models were used to estimate crude and adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) for cancer risk. The sub-classification of cancers includes hematologic, head and neck, esophagus, stomach, colon–rectum, liver, pancreas, lung, skin, breast, immune-related cancers, cervix/endometrium/ovary, prostate, bladder, kidney, brain, and thyroid were included in the analysis. To assess the robustness of our findings, we evaluated FQ exposure using two distinct metrics: the duration of therapy (days) and cumulative dose (mg). This dual approach allows for a comprehensive assessment of the dose-dependent impact of FQs on subsequent cancer risk. We also used a time-dependent Cox proportional hazards regression model and competing-risks models to assess the risk of cancer as a sensitivity analysis. Furthermore, to evaluate the dose–response relationship and address concerns regarding the arbitrary categorization of drug exposure, fluoroquinolone exposure was also modeled as a continuous variable. The cumulative duration (in days) and cumulative dose (in mg) of fluoroquinolone therapy were entered into the Cox proportional hazards regression to formally assess the linear dose–response relationship and to validate the robustness of the categorical findings. The Kaplan–Meier method was used to obtain the cumulative curves, and the results were then examined using the log-rank test. SAS statistical software (version 9.4, SAS Institute, Cary, NC, USA) was used to perform the analysis. A p-value was set at 0.05.
5. Conclusions
After correction for immortal time bias, FQ exposure was associated with an increased long-term risk of overall cancer among patients with ILD. A particularly strong association was observed for prostate cancer, whereas immune-related cancers showed a reduced risk. However, the absence of a consistent dose–response relationship suggests that these findings may be driven by confounding by indication, reverse causation, occult malignancy, or underlying ILD severity rather than a direct pharmacological carcinogenic effect. Further studies using active-comparator new-user designs, lag-time analyses, target trial emulation, and marginal structural models are warranted.
Supplementary Materials
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ph19071067/s1, Table S1. Full List of ICD-9-CM and ICD-10-CM Codes.
Author Contributions
Conceptualization: Y.-F.S., Y.-T.C., Y.-E.K., Y.-W.H., L.-K.H., C.-L.L., and J.-J.Y. methodology: J.-J.Y., Y.-F.S., Y.-T.C., and C.-L.L. formal analysis: C.-L.L. investigation: Y.-F.S., Y.-T.C., Y.-E.K., Y.-W.H., L.-K.H., C.-H.K., and J.-J.Y. resources: J.-J.Y., Y.-F.S., C.-L.L., and C.-H.K. data curation: Y.-F.S., Y.-T.C., Y.-E.K., Y.-W.H., L.-K.H., C.-L.L., and J.-J.Y. writing—original draft preparation: J.-J.Y., Y.-F.S., Y.-T.C., C.-L.L., and C.-H.K.; writing—review and editing: Y.-F.S., Y.-T.C., Y.-E.K., Y.-W.H., L.-K.H., C.-L.L., C.-H.K., and J.-J.Y. visualization: Y.-F.S., Y.-T.C., Y.-E.K., Y.-W.H., L.-K.H., C.-L.L., C.-H.K., and J.-J.Y. supervision: C.-H.K. and J.-J.Y. project administration: C.-H.K. and J.-J.Y. funding acquisition: C.-H.K. and J.-J.Y. All authors have read and agreed to the published version of the manuscript.
Funding
This research was supported by the Taiwan Ministry of Health and Welfare Clinical Trial Center (MOHW110-TDU-B-212-124004), China Medical University Hospital (DMR-115-075, DMR-115-076), and the Ministry of Science and Technology (MOST 110-2321-B-039-003). Ditmanson Medical Foundation Chia-Yi Christian Hospital Research Program R112-027.
Institutional Review Board Statement
This study was approved by the Institutional Review Board (IRB) of the China Medical University Hospital Research Ethics Committee (CMUH112-REC1-117(CR-2)), Approved Date: 9 July 2025.
Informed Consent Statement
Not applicable.
Data Availability Statement
The dataset is held by the Taiwan Ministry of Health and Welfare and is available upon request.
Acknowledgments
We are grateful to the Health Data Science Center, China Medical University Hospital, for providing administrative, technical, and funding support.
Conflicts of Interest
The authors declare no conflicts of interest.
Abbreviations
FQs: Fluoroquinolones; ILD: Interstitial Lung Disease; ICD-9-CM: International Classification of Diseases, Ninth Revision, Clinical Modification; NHIRD: National Health Insurance Research Database; aHRs: Adjusted Hazard Ratios; CIs: Confidence Intervals; COPD: Chronic Obstructive Pulmonary Disease; TB: Tuberculosis; DDD: Defined Daily Dose.
References
- Rodrigues, C.F.; Silva, F. The rise, fall, and rethink of (fluoro)quinolones: A quick rundown. Pathogens 2025, 14, 525. [Google Scholar] [CrossRef] [PubMed]
- Kloskowski, T.; Szeliski, K.; Fekner, Z.; Rasmus, M.; Dąbrowski, P.; Wolska, A.; Siedlecka, N.; Adamowicz, J.; Drewa, T.; Pokrywczyńska, M. Ciprofloxacin and levofloxacin as potential drugs in genitourinary cancer treatment-the effect of dose-response on 2d and 3d cell cultures. Int. J. Mol. Sci. 2021, 22, 11970. [Google Scholar] [CrossRef] [PubMed]
- Nowakowska, J.; Radomska, D.; Czarnomysy, R.; Marciniec, K. Recent development of fluoroquinolone derivatives as anticancer agents. Molecules 2024, 29, 3538. [Google Scholar] [CrossRef] [PubMed]
- Kassab, A.E. The most recent updates on the anticancer potential of fluoroquinolones: A mini review. Future Med. Chem. 2025, 17, 2067–2078. [Google Scholar] [CrossRef] [PubMed]
- Kassab, A.E.; Gomaa, R.M.; Gedawy, E.M. Drug repurposing of fluoroquinolones as anticancer agents in 2023. RSC Adv. 2024, 14, 37114–37130. [Google Scholar] [CrossRef] [PubMed]
- Chrzanowska, A.; Kurpios-Piec, D.; Żyżyńska-Granica, B.; Kiernozek-Kalińska, E.; Lay, W.X.; Ciechanowicz, A.K.; Struga, M. Anticancer activity and metabolic alteration in colon and prostate cancer cells by novel moxifloxacin conjugates with fatty acids. Eur. J. Pharmacol. 2023, 940, 175481. [Google Scholar] [CrossRef] [PubMed]
- Huang, C.Y.; Yang, J.L.; Chen, J.J.; Tai, S.B.; Yeh, Y.H.; Liu, P.F.; Lin, M.W.; Chung, C.L.; Chen, C.L. Fluoroquinolones suppress tgf-β and pma-induced mmp-9 production in cancer cells: Implications in repurposing quinolone antibiotics for cancer treatment. Int. J. Mol. Sci. 2021, 22, 11602. [Google Scholar] [CrossRef] [PubMed]
- Felicetti, T.; Di Iacovo, N.; Della Fazia, M.A.; Piobbico, D.; Pieroni, S.; Pacetti, M.; Yu, J.; Sun, Y.; Massari, S.; Barreca, M.L.; et al. New anti-ovarian cancer quinolone derivatives acting by modulating microrna processing machinery. RSC Med. Chem. 2025, 16, 98–124. [Google Scholar] [CrossRef]
- Ishaq, M.; Zafar, E.; Maqbool, M.F.; Mansha, M.; Khan, M.; Li, Y. Repurposing fluoroquinolones as cancer chemosensitizers: A way to overcome cancer therapeutic bottleneck. Naunyn Schmiedebergs Arch. Pharmacol. 2026, 399, 47–62. [Google Scholar] [PubMed]
- Yeh, J.J.; Hung, T.W.; Lin, C.L.; Chen, T.T.; Liw, P.X.; Yu, Y.L.; Kao, C.H. Colchicine is a weapon for managing the heart disease among interstitial lung disease with viral infection: Have we found the holy grail? Front. Cardiovasc. Med. 2022, 9, 925211. [Google Scholar] [CrossRef] [PubMed]
- Naccache, J.M.; Gibiot, Q.; Monnet, I.; Antoine, M.; Wislez, M.; Chouaid, C.; Cadranel, J. Lung cancer and interstitial lung disease: A literature review. J. Thorac. Dis. 2018, 10, 3829–3844. [Google Scholar] [CrossRef] [PubMed]
- Manoharan, S.; Iyappan, O.R.; Prabahar, A.; Bhasuran, B.; Raja, K. Gut-lung axis microbiome: Towards precision medicine in respiratory disorders (review). World Acad. Sci. J. 2025, 7, 88. [Google Scholar] [CrossRef]
- Goto, T. Airway microbiota as a modulator of lung cancer. Int. J. Mol. Sci. 2020, 21, 3044. [Google Scholar] [CrossRef] [PubMed]
- Xu, H.; Yin, L.; Bian, W.; Kang, M.; Adami, H.-O.; Ye, W. Interstitial lung disease and risk of lung cancer. JAMA Netw. Open 2025, 8, e2519630. [Google Scholar] [CrossRef] [PubMed]
- Wu, T.-S.J.; Shih, F.-Y.F.; Yen, M.-Y.; Wu, J.-S.J.; Lu, S.-W.; Chang, K.C.-M.; Hsiung, C.; Chou, J.-H.; Chu, Y.-T.; Chang, H.; et al. Establishing a nationwide emergency department-based syndromic surveillance system for better public health responses in taiwan. BMC Public Health 2008, 8, 18. [Google Scholar] [CrossRef] [PubMed]
- Chen, J.; Ye, F.; Shang, K.; Li, N.; Li, C.; He, H. The mendelian randomized study revealed the association of prostatitis with prostate cancer risk. Sci. Rep. 2024, 14, 24643. [Google Scholar] [CrossRef] [PubMed]
- Yeh, J.J.; Liw, P.X.; Wong, Y.S.; Kao, H.M.; Lee, C.H.; Lin, C.L.; Kao, C.H. The effect of colchicine on cancer risk in patients with immune-mediated inflammatory diseases: A time-dependent study based on the taiwan’s national health insurance research database. Eur. J. Med. Res. 2024, 29, 245. [Google Scholar] [CrossRef] [PubMed]
- Airola, C.; Severino, A.; Spinelli, I.; Gasbarrini, A.; Cammarota, G.; Ianiro, G.; Ponziani, F.R. “Pleiotropic” effects of antibiotics: New modulators in human diseases. Antibiotics 2024, 13, 1176. [Google Scholar] [CrossRef] [PubMed]
- Chen, N.; Li, L.; Han, Y.; Chen, Z. The role of gut microbiota in the modulation of pulmonary immune response to viral infection through the gut-lung axis. J. Inflamm. Res. 2025, 18, 11755–11781. [Google Scholar] [CrossRef] [PubMed]
- Noble, A.; Pring, E.T.; Durant, L.; Man, R.; Dilke, S.M.; Hoyles, L.; James, S.A.; Carding, S.R.; Jenkins, J.T.; Knight, S.C. Altered immunity to microbiota, b cell activation and depleted γδ/resident memory t cells in colorectal cancer. Cancer Immunol. Immunother. 2022, 71, 2619–2629. [Google Scholar] [CrossRef] [PubMed]
- Iaccarino, J.; Mignini, I.; Maresca, R.; Giansanti, G.; Esposto, G.; Borriello, R.; Galasso, L.; Ainora, M.E.; Gasbarrini, A.; Zocco, M.A. The impact of air pollution on the lung–gut–liver axis: Oxidative stress and its role in liver disease. Antioxidants 2025, 14, 1148. [Google Scholar] [CrossRef] [PubMed]
- Deepika; Alsharari, Z.D.; Ahmad, M.F.; Singh, J.; Yadav, M.; Kumari, A.; Kumar, A.; Matthews, K.R.; Aluko, R.E.; Raposo, A.; et al. Gut-lung axis, probiotics, and prebiotics: Insights on dysbiosis, mechanism, and prevention of lung cancer. Front. Nutr. 2025, 12, 1624803. [Google Scholar] [CrossRef] [PubMed]
- Ohtani, N.; Hara, E. Gut-liver axis-mediated mechanism of liver cancer: A special focus on the role of gut microbiota. Cancer Sci. 2021, 112, 4433–4443. [Google Scholar] [CrossRef] [PubMed]
- Shao-Yu, Y.; Niu, D.; Chen, J.; Li, W.-Y.; Wang, X.; Meng, Q.-W.; Song, W.-J.; Yang, Y.-G.; Wang, H.; Li, R.-R.; et al. Antibiotic cocktail-induced changes in gut microbiota drive alteration of bile acid metabolism to restrain th17 differentiation through the fxr–nlrp3 axis. Gut Microbes 2025, 17, 2582944. [Google Scholar] [CrossRef] [PubMed]
- Abdel-Aal, M.A.A.; Abdel-Aziz, S.A.; Shaykoon, M.S.A.; Abuo-Rahma, G.E.-D.A. Towards anticancer fluoroquinolones: A review article. Arch. Pharm. 2019, 352, 1800376. [Google Scholar] [CrossRef]
- Ferrario, N.; Marras, E.; Vivona, V.; Randisi, F.; Fallica, A.N.; Marrazzo, A.; Perletti, G.; Gariboldi, M.B. Mechanisms of the antineoplastic effects of new fluoroquinolones in 2d and 3d human breast and bladder cancer cell lines. Cancers 2024, 16, 2227. [Google Scholar] [CrossRef] [PubMed]
- Pazzagli, L.; Linder, M.; Zhang, M.; Vago, E.; Stang, P.; Myers, D.; Andersen, M.; Bahmanyar, S. Methods for time-varying exposure related problems in pharmacoepidemiology: An overview. Pharmacoepidemiol. Drug Saf. 2018, 27, 148–160. [Google Scholar] [CrossRef] [PubMed]
- Pottegård, A.; Friis, S.; Stürmer, T.; Hallas, J.; Bahmanyar, S. Considerations for pharmacoepidemiological studies of drug-cancer associations. Basic Clin. Pharmacol. Toxicol. 2018, 122, 451–459. [Google Scholar] [CrossRef] [PubMed]
- Anwar, A.; Assaad, M.; Jos Walsh, J. Post-obstructive pneumonia as a presenting feature of endobronchial squamous cell carcinoma in a 42-year-old patient. CHEST 2023, 164, A4447–A4448. [Google Scholar] [CrossRef]
- Pottegård, A.; Hallas, J. New use of prescription drugs prior to a cancer diagnosis. Pharmacoepidemiol. Drug Saf. 2017, 26, 223–227. [Google Scholar] [CrossRef] [PubMed]
- Tamim, H.; Monfared, A.A.; LeLorier, J. Application of lag-time into exposure definitions to control for protopathic bias. Pharmacoepidemiol. Drug Saf. 2007, 16, 250–258. [Google Scholar] [CrossRef] [PubMed]
- Hicks, B.; Kaye, J.A.; Azoulay, L.; Kristensen, K.B.; Habel, L.A.; Pottegård, A. The application of lag times in cancer pharmacoepidemiology: A narrative review. Ann. Epidemiol. 2023, 84, 25–32. [Google Scholar] [CrossRef] [PubMed]
- Thommes, E.W.; Mahmud, S.M.; Young-Xu, Y.; Snider, J.T.; van Aalst, R.; Lee, J.K.H.; Halchenko, Y.; Russo, E.; Chit, A. Assessing the prior event rate ratio method via probabilistic bias analysis on a bayesian network. Stat. Med. 2020, 39, 639–659. [Google Scholar] [PubMed]
Figure 1.
Study flow diagram for cohort construction and propensity score matching. Patients with interstitial lung disease (ILD) were identified from the database. After applying eligibility criteria (including exclusion of patients with prior cancer before the index date), patients were classified as fluoroquinolone (FQ) users or non-users. To avoid immortal time bias, the index date for FQ users was defined as the date of their first prescription, while non-users were assigned a randomized index date. Patients were then matched 1:1 using propensity scores based on demographics, comorbidities, and concomitant medications. The final propensity-score-matched cohort included 3953 FQ users and 3953 non-users (total N = 7906). Patients were followed from their respective index dates until incident cancer, death, or the end of the study period; the mean follow-up duration was 5.43 years for FQ users and 5.95 years for non-users.
Figure 1.
Study flow diagram for cohort construction and propensity score matching. Patients with interstitial lung disease (ILD) were identified from the database. After applying eligibility criteria (including exclusion of patients with prior cancer before the index date), patients were classified as fluoroquinolone (FQ) users or non-users. To avoid immortal time bias, the index date for FQ users was defined as the date of their first prescription, while non-users were assigned a randomized index date. Patients were then matched 1:1 using propensity scores based on demographics, comorbidities, and concomitant medications. The final propensity-score-matched cohort included 3953 FQ users and 3953 non-users (total N = 7906). Patients were followed from their respective index dates until incident cancer, death, or the end of the study period; the mean follow-up duration was 5.43 years for FQ users and 5.95 years for non-users.
Figure 2.
Kaplan–Meier cumulative incidence curves for overall cancer in fluoroquinolone (FQ) users and non-users after propensity score matching. The cumulative incidence of overall cancer was higher in FQ users than in non-users throughout follow-up. After rigorous index-date assignment to minimize immortal time bias, the Kaplan–Meier analysis demonstrated a statistically significant difference between the two groups (log-rank p < 0.001). These findings indicate that FQ exposure was associated with an increased long-term risk of cancer in patients with interstitial lung disease (ILD).
Figure 2.
Kaplan–Meier cumulative incidence curves for overall cancer in fluoroquinolone (FQ) users and non-users after propensity score matching. The cumulative incidence of overall cancer was higher in FQ users than in non-users throughout follow-up. After rigorous index-date assignment to minimize immortal time bias, the Kaplan–Meier analysis demonstrated a statistically significant difference between the two groups (log-rank p < 0.001). These findings indicate that FQ exposure was associated with an increased long-term risk of cancer in patients with interstitial lung disease (ILD).
Table 1.
Baseline characteristics of fluoroquinolone (FQ) users and non-users (N = 3953).
| Fluoroquinolone | |||
|---|---|---|---|
| Variable | No (N = 3953) N (%) | Yes (N = 3953) N (%) | p-Value |
| Age | 0.58 | ||
| ≤49 | 607 (15.4%) | 641 (16.2%) | |
| 50–64 | 1205 (30.5%) | 1193 (30.2%) | |
| 65+ | 2141 (54.2%) | 2119 (53.6%) | |
| Mean Age (SD) # | 66.1 (15.0) | 65.8 (15.1) | 0.43 |
| Sex | 0.43 | ||
| Female | 1934 (48.9%) | 1899 (48.0%) | |
| Male | 2019 (51.1%) | 2054 (52.0%) | |
| Comorbidities | |||
| Tuberculosis | 499 (12.6%) | 524 (13.3%) | 0.40 |
| Pneumonia | 2503 (63.3%) | 2510 (63.5%) | 0.87 |
| Peptic ulcer disease | 2825 (71.5%) | 2157(54.6) | 0.77 |
| COPD | 2170 (54.9%) | 3077 (56.7%) | 0.56 |
| Cirrhosis | 515 (13.0%) | 534 (13.5%) | 0.53 |
| Atopic disease | 2678 (67.8%) | 2661 (67.3%) | 0.68 |
| Diabetes | 2244 (56.8%) | 2265 (57.3%) | 0.63 |
| Hypertension | 3197 (80.9%) | 3187 (80.6%) | 0.78 |
| Hyperlipidemia | 2997 (75.8%) | 4000 (73.7%) | 0.35 |
| Chronic kidney disease | 1911 (48.3%) | 1853 (48.7%) | 0.19 |
| Stroke | 1673 (42.3%) | 1694 (42.9%) | 0.63 |
| Pelvic inflammatory disease | 5 (0.13%) | 8 (0.20%) | 0.41 |
| Heart failure | 1247 (31.6%) | 1298 (32.8%) | 0.22 |
| Periodontal diseases | 3093 (78.2%) | 3066 (77.8%) | 0.46 |
| IBS | 893 (22.6%) | 871 (22.0%) | 0.55 |
| Prostate inflammation (incl. BPH) | 1414 (35.8%) | 1394 (35.3%) | 0.64 |
| Anti-inflammatory medicines a | 3858 (97.6%) | 3857 (97.6%) | 0.94 |
| Immunosuppressants b | 2547 (64.4%) | 2493 (63.1%) | 0.21 |
Chi-square test; # t-test; SD: Standard deviation. BPH: Benign Prostatic Hyperplasia. a Includes aspirin, statins, NSAIDs, ACE inhibitors, ARBs, CCBs, beta-blockers, diuretics. b Includes ICS, prednisolone, sulfasalazine, cyclophosphamide, methotrexate, hydroxychloroquine, and TNF-α antagonists.
Table 2.
Risk of all-cause cancer by fluoroquinolone use.
| Fluoroquinolone | N | Person-Years | Follow-Up (Mean, SD) a | Events | Rate/1000 PY b | Crude HR (95% CI) | Adjusted HR (95% CI) c |
|---|---|---|---|---|---|---|---|
| No use | 3953 | 23,504 | 5.95 (4.29) | 189 | 8.04 | 1.00 (Reference) | 1.00 (Reference) |
| Use | 3953 | 21,462 | 5.43 (4.22) | 247 | 11.5 | 1.45 (1.20, 1.75) *** | 1.45 (1.20, 1.76) *** |
a SD: Standard deviation. b PY: Person-years. c Adjusted for age, sex, comorbidities, and medication use. *** p < 0.001.
Table 3.
Adjusted risk of site-specific cancer by fluoroquinolone use among patients with interstitial lung disease.
Table 3.
Adjusted risk of site-specific cancer by fluoroquinolone use among patients with interstitial lung disease.
| Fluoroquinolone | Fluoroquinolone Compared to Non-Fluoroquinolone | |||||
|---|---|---|---|---|---|---|
| No | Yes | Crude HR | Adjusted HR | |||
| Cancer Site | Events | Rate a | Events | Rate a | (95% CI) | (95% CI) b |
| Hematologic malignancy | 10 | 0.43 | 17 | 0.79 | 1.85 (0.85, 4.04) | 1.93 (0.88, 4.23) |
| Head and neck cancer | 8 | 0.34 | 14 | 0.65 | 1.94 (0.81, 4.63) | 1.66 (0.68, 4.01) |
| Esophagus | 10 | 0.43 | 6 | 0.28 | 0.65 (0.24, 1.80) | 0.61 (0.22, 1.74) |
| Stomach | 7 | 0.3 | 12 | 0.56 | 1.94 (0.76, 4.92) | 2.14 (0.82, 5.62) |
| Colon, rectum | 20 | 0.85 | 26 | 1.21 | 1.44 (0.80, 2.58) | 1.37 (0.76, 2.46) |
| Liver | 30 | 1.28 | 26 | 1.21 | 0.97 (0.57, 1.64) | 1.03 (0.61, 1.76) |
| Pancreas | 6 | 0.26 | 4 | 0.19 | 0.75 (0.21, 2.65) | 0.77 (0.21, 2.84) |
| Lung | 34 | 1.45 | 34 | 1.58 | 1.12 (0.70, 1.80) | 1.22 (0.75, 1.97) |
| Skin | 5 | 0.21 | 6 | 0.28 | 1.32 (0.40, 4.32) | 1.16 (0.34, 4.01) |
| Breast cancer | 16 | 0.68 | 21 | 0.98 | 1.46 (0.76, 2.80) | 1.38 (0.71, 2.66) |
| Immune-related cancers | 92 | 1.38 | 57 | 0.85 | 0.62 (0.45, 0.87) ** | 0.62 (0.44, 0.87) ** |
| Cervix/Endometrium/Ovary | 6 | 0.26 | 6 | 0.28 | 1.12 (0.36, 3.46) | 1.23 (0.39, 3.88) |
| Prostate | 6 | 0.26 | 30 | 1.40 | 5.52 (2.30, 13.26) *** | 5.50 (2.25, 13.48) *** |
| Bladder, kidney | 12 | 0.51 | 22 | 1.03 | 2.04 (1.01, 4.13) * | 2.03 (0.99, 4.14) |
| Brain | 0.09 | 0.09 | 1.11 (0.16, 7.85) | 1.36 (0.16, 11.70) | ||
| Thyroid | 4 | 0.17 | 5 | 0.23 | 1.41 (0.38, 5.25) | 2.12 (0.48, 9.23) |
| Others | 40 | 1.7 | 36 | 1.68 | 1.00 (0.64, 1.57) | 1.05 (0.67, 1.66) |
a Incidence rate per 1000 person-years b Adjusted for age, sex, comorbidities, and medications. * p < 0.05; ** p < 0.01; *** p < 0.001.
Table 4.
Incidence and adjusted hazard ratio of sub-cancer stratified by cumulative use days per year of fluoroquinolone therapy in patients with interstitial lung disease-virus infection among the propensity score-matched cohort.
Table 4.
Incidence and adjusted hazard ratio of sub-cancer stratified by cumulative use days per year of fluoroquinolone therapy in patients with interstitial lung disease-virus infection among the propensity score-matched cohort.
| Medication Exposed | Event | Person-Year | Rate a | Adjusted HR (95% CI) b |
|---|---|---|---|---|
| Colon, rectum | ||||
| Non-fluoroquinolone | 20 | 23,504 | 0.85 | 1.00 |
| Fluoroquinolone # | ||||
| ≤35 days | 14 | 10,141 | 1.38 | 1.53 (0.76, 3.06) |
| 36–55 days | 7 | 5447 | 1.29 | 1.43 (0.60, 3.41) |
| >56 days | 5 | 5874 | 0.85 | 1.01 (0.37, 2.72) |
| Liver | ||||
| Non-fluoroquinolone | 30 | 23,504 | 1.28 | 1.00 |
| Fluoroquinolone # | ||||
| ≤35 days | 13 | 10,141 | 1.28 | 1.12 (0.58, 2.17) |
| 36–55 days | 5 | 5447 | 0.92 | 0.76 (0.30, 1.98) |
| >56 days | 8 | 5874 | 1.36 | 1.15 (0.52, 2.54) |
| Lung | ||||
| Non-fluoroquinolone | 34 | 23,504 | 1.45 | 1.00 |
| Fluoroquinolone # | ||||
| ≤35 days | 13 | 10,141 | 1.28 | 1.05 (0.55, 2.02) |
| 36–55 days | 9 | 5447 | 1.65 | 1.26 (0.60, 2.64) |
| >56 days | 12 | 5874 | 2.04 | 1.42 (0.72, 2.77) |
| Immune-related cancers | ||||
| Non-fluoroquinolone | 40 | 23,504 | 1.7 | 1.00 |
| Fluoroquinolone # | ||||
| ≤35 days | 19 | 10,141 | 1.87 | 1.18 (0.68, 2.06) |
| 36–55 days | 8 | 5447 | 1.47 | 0.91 (0.42, 1.95) |
| >56 days | 9 | 5874 | 1.53 | 0.98 (0.47, 2.03) |
a Incidence rate per 1000 person-years b Adjusted HR: multivariable analysis including age, sex, comorbidities, and medications. # The cumulative use days are partitioned into 3 segments by the median and the third quartile.
Table 5.
Incidence and adjusted hazard ratio of sub-cancer stratified by cumulative use dose per year of fluoroquinolone therapy in patients with interstitial lung disease-virus infection in the propensity score-matched cohort.
Table 5.
Incidence and adjusted hazard ratio of sub-cancer stratified by cumulative use dose per year of fluoroquinolone therapy in patients with interstitial lung disease-virus infection in the propensity score-matched cohort.
| Medication Exposed | Event | Person-Year | Rate a | Adjusted HR (95% CI) b |
|---|---|---|---|---|
| Colon, rectum | ||||
| Non-fluoroquinolone | 20 | 23,504 | 0.85 | 1.00 |
| Fluoroquinolone # | ||||
| ≤3000 mg | 20 | 14,635 | 1.37 | 1.57 (0.84, 2.94) |
| 3001–6000 mg | 6 | 0.57 | 0.60 (0.14, 2.61) | |
| >6000 mg | 1.2 | 1.35 (0.45, 4.05) | ||
| Liver | ||||
| Non-fluoroquinolone | 30 | 23,504 | 1.28 | 1.00 |
| Fluoroquinolone # | ||||
| ≤3000 mg | 18 | 14,635 | 1.23 | 1.04 (0.58, 1.90) |
| 3001–6000 mg | 3 | 3491 | 0.86 | 0.73 (0.22, 2.43) |
| >6000 mg | 5 | 3337 | 1.5 | 1.31 (0.49, 3.46) |
| Lung | ||||
| Non-fluoroquinolone | 34 | 23,504 | 1.45 | 1.00 |
| Fluoroquinolone # | ||||
| ≤3000 mg | 22 | 14,635 | 1.5 | 1.13 (0.65, 1.95) |
| 3001–6000 mg | 7 | 3491 | 2.01 | 1.61 (0.70, 3.67) |
| >6000 mg | 5 | 3337 | 1.5 | 1.22 (0.46, 3.18) |
| Immune-related cancers | ||||
| Non-fluoroquinolone | 40 | 23,504 | 1.7 | 1.00 |
| Fluoroquinolone # | ||||
| ≤3000 mg | 24 | 14,635 | 1.64 | 1.01 (0.61, 1.70) |
| 3001–6000 mg | 4 | 3491 | 1.15 | 0.73 (0.26, 2.05) |
| >6000 mg | 8 | 3337 | 2.4 | 1.60 (0.73, 3.49) |
a Incidence rate per 1000 person-years b Adjusted HR: multivariable analysis including age, sex, comorbidities, and medications; # The cumulative use days are partitioned into 3 segments by median, and third quartile.
Table 6.
Risk of all-cause cancer by fluoroquinolone use (time-dependent Cox regression model).
| Variable | Fluoroquinolone Compared to Non-Fluoroquinolone | |
|---|---|---|
| Crude HR | Adjusted HR | |
| Cancer Site | (95% CI) | (95% CI) a |
| Cancer | 1.99 (1.52, 2.61) *** | 2.07 (1.57, 2.72) *** |
| Hematologic malignancy | 3.68 (1.40, 9.68) ** | 4.09 (1.51, 11.1) ** |
| Head and neck cancer | - | - |
| Esophagus | 0.52 (0.07, 3.99) | 0.46 (0.06, 3.61) |
| Stomach | 1.90 (0.42, 8.61) | 2.36 (0.50, 11.1) |
| Colon, rectum | 0.74 (0.23, 2.42) | 0.77 (0.23, 2.55) |
| Liver | 1.72 (0.76, 3.87) | 1.71 (0.75, 3.88) |
| Pancreas | - | - |
| Lung | 3.56 (2.00, 6.33) *** | 3.35 (1.87, 6.01) *** |
| Skin | - | - |
| Breast cancer | 0.64 (0.15, 2.68) | 0.70 (0.16, 2.97) |
| Immune-related cancers | 1.22 (0.55, 2.68) | 1.21 (0.55, 2.70) |
| Cervix/Endometrium/Ovary | 1.23 (0.15, 9.96) | 1.42 (0.17, 11.7) |
| Prostate | 8.89 (4.36, 18.2) *** | 9.95 (4.75, 20.8) *** |
| Bladder, kidney | 3.39 (1.43, 8.04) ** | 3.72 (1.55, 8.93) ** |
| Brain | - | - |
| Thyroid | - | - |
| Others | 1.31 (0.51, 3.35) | 1.37 (0.53, 3.53) |
a Adjusted for age, sex, comorbidities, and medication use. ** p < 0.01. *** p < 0.001.
Table 7.
Risk of all-cause cancer by fluoroquinolone use (competing risk model).
| Variable | Fluoroquinolone Compared to Non-Fluoroquinolone | |
|---|---|---|
| Crude SHR | Adjusted SHR | |
| Cancer Site | (95% CI) | (95% CI) a |
| Cancer | 1.31 (1.08, 1.58) ** | 1.28 (1.06, 1.55) ** |
| Colon, rectum | 1.31 (0.73, 2.34) | 1.24 (0.69, 2.22) |
| Liver | 0.87 (0.51, 1.47) | 0.88 (0.53, 1.47) |
| Lung | 1.00 (0.62, 1.62) | 1.04 (0.65, 1.68) |
| Immune-related cancers | 0.90 (0.58, 1.42) | 0.91 (0.59, 1.42) |
a Adjusted for age, sex, comorbidities, and medication use. ** p < 0.01.
Table 8.
Hazard ratios and 95% confidence intervals for all-cause cancer risk by continuous dose of fluoroquinolones.
Table 8.
Hazard ratios and 95% confidence intervals for all-cause cancer risk by continuous dose of fluoroquinolones.
| Variable | Fluoroquinolone Compared to Non-Fluoroquinolone | |
|---|---|---|
| Crude HR | Adjusted HR | |
| Cancer Site | (95% CI) | (95% CI) a |
| Cancer | 0.99 (0.99, 1.00) | 0.99 (0.99, 1.00) |
a Adjusted for age, sex, comorbidities, and medication use.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.
Share and Cite
MDPI and ACS Style
Sun, Y.-F.; Chiu, Y.-T.; Ko, Y.-E.; Huang, Y.-W.; Hsieh, L.-K.; Lin, C.-L.; Kao, C.-H.; Yeh, J.-J. Fluoroquinolone Exposure and Cancer Risk in Interstitial Lung Disease: A Propensity-Score-Matched Cohort Study Using Cox and Competing-Risk Models. Pharmaceuticals 2026, 19, 1067. https://doi.org/10.3390/ph19071067
AMA Style
Sun Y-F, Chiu Y-T, Ko Y-E, Huang Y-W, Hsieh L-K, Lin C-L, Kao C-H, Yeh J-J. Fluoroquinolone Exposure and Cancer Risk in Interstitial Lung Disease: A Propensity-Score-Matched Cohort Study Using Cox and Competing-Risk Models. Pharmaceuticals. 2026; 19(7):1067. https://doi.org/10.3390/ph19071067
Chicago/Turabian StyleSun, Yi-Fan, Yu-Ting Chiu, Yung-En Ko, Yu-Wei Huang, Liang-Kai Hsieh, Cheng-Li Lin, Chia-Hung Kao, and Jun-Jun Yeh. 2026. "Fluoroquinolone Exposure and Cancer Risk in Interstitial Lung Disease: A Propensity-Score-Matched Cohort Study Using Cox and Competing-Risk Models" Pharmaceuticals 19, no. 7: 1067. https://doi.org/10.3390/ph19071067
APA StyleSun, Y.-F., Chiu, Y.-T., Ko, Y.-E., Huang, Y.-W., Hsieh, L.-K., Lin, C.-L., Kao, C.-H., & Yeh, J.-J. (2026). Fluoroquinolone Exposure and Cancer Risk in Interstitial Lung Disease: A Propensity-Score-Matched Cohort Study Using Cox and Competing-Risk Models. Pharmaceuticals, 19(7), 1067. https://doi.org/10.3390/ph19071067
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
