Introduction

Cataract is one of the most prevalent causes of blindness worldwide, particularly among individuals aged 50 and older, and has become the leading cause of vision impairment. According to the 2020 Global Blindness Report, cataract is responsible for nearly 45% of adult blindness cases, affecting more than 15 million people globally1. The primary feature of cataract is the gradual opacification of the lens, which leads to a progressive decline in vision and can eventually result in blindness2. With the ongoing global aging trend, the incidence of cataract continues to rise, representing a growing challenge in public health3. The pathophysiology of cataract is multifactorial, involving a range of risk factors. In addition to well-established factors such as aging, diabetes, ultraviolet radiation, and smoking, recent evidence increasingly suggests that long-term use of certain medications plays a significant role in cataract development. Particularly, corticosteroids and specific anticancer drugs, when used over prolonged periods or at high doses, may accelerate lens opacity and thereby exacerbate vision loss4,5,6,7. Given this, the effective prevention, early diagnosis, and management of cataract, especially in low- and middle-income countries with limited healthcare resources, remain pressing public health concerns that demand immediate attention8.

Although several studies have implicated certain medications in the onset of cataract, there is a paucity of large-scale, systematic research to definitively elucidate the relationship between specific drugs and cataract formation. The U.S. FDA Adverse Event Reporting System (FAERS) serves as a critical tool for monitoring drug safety, collecting reports of adverse events from across the globe, and providing valuable data for assessing the safety profile of medications9,10. While prior studies have utilized the FAERS database to investigate other drug-induced adverse effects, such as acute pancreatitis and allergic reactions, research on drug-induced cataract remains limited11,12.

The aim of this study is to explore the potential risks of cataract induction by medications through a comprehensive analysis of FAERS data spanning from the first quarter of 2004 to the third quarter of 2024. Specifically, this study seeks to evaluate the strength of the association between various medications and cataract occurrence, as well as to assess the timing of drug-induced cataract development. The findings are expected to provide more robust safety information to guide clinical drug selection and treatment strategies, ultimately aiding in the development of more precise drug-use guidelines to mitigate the risk of drug-induced cataract.

Methods

Data source

This study analyzed data from the FAERS database, which includes reports submitted between Q1 2004 and Q3 2024. In this context, Q1 refers to the first quarter (January to March), while Q3 refers to the third quarter (July to September) of each year. FAERS consolidates mandatory reports from pharmaceutical companies with voluntary submissions from healthcare professionals (including physicians and pharmacists), patients, and consumers. The dataset comprises demographic data, drug details, descriptions of adverse events, therapy start and stop dates, as well as treatment indications, all classified using the Medical Dictionary for Regulatory Activities (MedDRA). Due to its extensive coverage, FAERS is a crucial resource for pharmacovigilance and the identification of drug safety signals11,13.

Data selection and processing

The FAERS database initially contained 21,838,627 raw reports submitted between January 2004 and September 2024. After removing duplicate records, 19,541,994 unique reports were retained for further analysis. To identify cases of cataract, we applied the Preferred Term (PT) code 10007739, which corresponds to “cataract” in the MedDRA, to the PT field of the dataset. Following FDA recommendations, reports were systematically processed using a hierarchical deduplication strategy. Specifically, records were first grouped by CASEID (a unique identifier for each adverse event case), then ordered by FDA_DT (the date the FDA received the report), and finally by PRIMARYID (a unique identifier for each individual report). In instances where multiple records shared the same CASEID, the report with the most recent FDA_DT was retained. If multiple reports had both identical CASEID and FDA_DT values, the record with the highest PRIMARYID was selected, as it typically reflects the most comprehensive and updated information. This approach ensured that each adverse event was counted only once, thereby minimizing the risk of duplication and maintaining the integrity of the dataset9,14. This process yielded 54,800 reports from healthcare professionals, patients, and consumers, involving 3,551 drugs. Considering the strong association between cataract incidence and age, particularly its higher prevalence in individuals aged 45 and older, data from patients over 45 years of age were excluded to minimize age-related bias in the analysis of drug-related cataract15. Only reports from patients aged 0 to 45 were retained, resulting in 2,336 cases of drug-related cataract, involving 691 drugs. To ensure consistency, drug names were standardized using the DrugBank database16. Subsequently, four disproportionality analysis methods were employed to assess the association between drugs and cataract. This analysis identified 24 distinct drugs, which were categorized according to their mechanisms of action. The data cleaning process is illustrated in Fig. 1. A Time-to-Onset (TTO) analysis was also conducted to evaluate the temporal association between drug exposure and the development of cataract. TTO was defined as the time interval between the reported date of the adverse event (EVENT_DT) and the initiation date of the suspect drug therapy (START_DT). To ensure accuracy, records with missing, inconsistent, or implausible EVENT_DT or START_DT values (e.g., negative intervals or default dates) were excluded from the analysis.

Fig. 1
figure 1

Raw data cleaning flowchart for drug-induced cataract. Notes: Data were processed to extract information on cataract patients and their medication usage from the FAERS database.

Full size image

Signal detection methods and statistical analysis

To evaluate the association between drugs and cataract, we employed four validated signal detection methods: Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR), Bayesian Confidence Propagating Neural Network (BCPNN), and the Multi-item Gamma Poisson Shrinker (MGPS). The criteria for signal detection were as follows: (1) ROR: a ≥ 3 and 95% CI lower > 1; (2) PRR: a ≥ 3 and 95% CI lower > 1; (3) BCPNN: IC025 > 0; (4) MGPS: EBGM05 > 2 and a > 0 (Tables 1, 2). A drug-adverse event relationship was considered significant if it met the thresholds in all four methods17,18,19,20. Each of these methods offers unique advantages: ROR helps to adjust for reporting bias, while PRR improves specificity by comparing reporting frequencies. BCPNN enhances the robustness of signals through Bayesian integration and cross-validation, making it particularly effective in managing reporting distortions. MGPS, on the other hand, excels in identifying rare signals and handling sparse data, which is especially useful in large pharmacovigilance datasets21. By applying these complementary methods, we maximized the coverage of signal detection and cross-validated the results, thereby enhancing the reliability of our assessment of the drug-cataract association. Statistical analyses were conducted using R (version 4.2.2), SPSS (version 26.0), and GraphPad Prism (version 10.1.2).

Table1 Four-grid table of disproportionality analysis method.
Full size table
Table 2 Principle of dis-proportionality measure and standard of signal detection.
Full size table

Search strategy and terminology classification

Cataract cases were classified according to MedDRA standards as follows: System Organ Class (SOC): Eye Disorders (Code: [10015919]); High-Level Group Term (HLGT): Anterior Eye Structural Change, Deposit and Degeneration (Code: [10002693]); High-Level Term (HLT): Cataract Conditions (Code: [10007772]); Preferred Term (PT): Cataract (Code: [10007739]). To ensure specificity, a narrow-scope standardized MedDRA query (SMQ) was applied, with particular emphasis on reports designated as "Primary Suspect." This approach minimized the inclusion of unrelated entries and enhanced the overall accuracy and relevance of the data, thereby improving the reliability of the analysis17,22.

Results

Demographic and clinical characteristics of cataract cases

This study included 2,336 cases of drug-related cataract adverse events, involving subjects aged 0 to 45 years. The mean age of the cohort was 30.1 ± 12.9 years, with females comprising a larger proportion (57.9%) compared to males (38.8%). The gender distribution remained generally consistent across the age groups, although slight variations were observed in some subgroups (Fig. 2A). The time trend analysis revealed a consistent increase in the number of reported drug-induced cataract cases over the years, with a notable surge beginning in 2018. This upward trend highlights the growing recognition and reporting of adverse events over time (Fig. 2B). Regarding drug administration routes, oral administration was the most commonly reported (697 cases, 29.0%), followed by subcutaneous injection (322 cases, 13.8%) and intravenous injection (161 cases, 6.9%) (Fig. 2C). In terms of patient outcomes, the most frequent result was "other serious conditions" (1,617 cases, 69.2%), followed by hospitalization (477 cases, 20.4%) and disability (155 cases, 6.6%) (Fig. 2D). Geographical distribution analysis revealed that the United States had the highest number of reported cases (776 cases, 33.2%), followed by Canada (193 cases, 8.3%), Japan (111 cases, 4.8%), and the United Kingdom (109 cases, 4.7%) (Fig. 2E). Detailed demographic characteristics of the subjects, along with subgroup analysis results, are presented in Table 3.

Fig. 2
figure 2

Distribution of baseline data for patients reporting adverse events of drug-induced cataract in the FAERS database. Note: (A) illustrates the age and gender distribution of cataract patients, (B) depicts the temporal trend of cataract reports over the study period, (C) shows the distribution of drug administration routes among cataract cases, (D) assesses the severity of adverse drug reactions associated with drugs, (E) highlights the countries contributing to drug-related adverse reaction cases, and (F) visualizes the geographic distribution of drug-related adverse reaction reports. The map was generated using Hiplot Pro (https://hiplot.com.cn/), an online data visualization platform.

Full size image
Table 3 Baseline data of cataract patients reported in the FAERS database.
Full size table

Drug-specific reporting trends in cataract cases

To assess the association between drugs and cataract, a disproportionality analysis was performed, focusing on reports involving drugs used by subjects in the 0–45 age group (n = 691). After excluding duplicate commercial names, 24 drugs were identified as showing positive signals, which were subsequently categorized based on their mechanisms of action (Fig. 3). Among the drugs identified, glucocorticoids accounted for a substantial proportion of the reports. Fluticasone furoate had the highest number of cases (39 reports), followed by triamcinolone (31 reports) and dexamethasone (18 reports). Insulin drugs were also prominently represented, with insulin glargine showing the highest number of reports (25 cases), followed by human insulin (16 cases) and insulin lispro (5 cases). Immunosuppressants, including cyclosporine (27 cases) and pomalidomide (3 cases), were also found to be potentially associated with cataract. In the antiglaucoma preparations category, latanoprost (13 cases) and dorzolamide (3 cases) were similarly identified. While fewer reports were found in the antineoplastic agents category, notable signals were still observed. Rituximab was the most frequently reported drug in this category (5 cases), followed by erdafitinib (3 cases) and ibrutinib (3 cases). Additionally, several other drugs were identified as potentially associated with cataract. Notably, ranibizumab (23 cases), nitisinone (12 cases), and atorvastatin (10 cases) also showed positive signals for cataract formation.

Fig. 3
figure 3

Grouped bubble chart of positive drug signals for cataract. Notes: This figure presents a grouped bubble chart illustrating positive drug signals for cataract. Different colors represent distinct drug categories, while the size of each bubble corresponds to the number of adverse event reports (n) associated with each drug linked to cataract.

Full size image

Classification of cataract-associated drugs by ROR values

This study assessed drugs associated with cataract across six predefined therapeutic categories, using the ROR to evaluate potential signals. In the glucocorticoid category, difluprednate exhibited the highest ROR for cataract-related events (ROR = 890.02), followed by prednisolone (ROR = 413.34) and dexamethasone (ROR = 96.92). Among insulin drugs, insulin lispro demonstrated the highest ROR (105.39), with insulin human (ROR = 10.59) and insulin glargine (ROR = 5.95) also showing positive signals. In the antineoplastic agents category, erdafitinib displayed a significantly higher ROR (144.67) compared to ibrutinib (ROR = 7.41), suggesting a stronger association with cataract development. For antiglaucoma preparations, dorzolamide showed the strongest cataract-related signal (ROR = 117.62), followed by latanoprost (ROR = 66.79). Among immunosuppressants, pomalidomide (ROR = 9.71) and cyclosporine (ROR = 3.43) were associated with positive signals for cataract. Additionally, several other drugs were found to have significant cataract-related signals, including nitisinone (ROR = 53.91), ranibizumab (ROR = 46.98), and pentosan polysulfate (ROR = 13.67). A comprehensive summary of these findings is provided in Fig. 4 and Table 4. In addition, a subgroup analysis comparing the insulin treatment group to the non-treatment group is shown in Supplementary Fig. 1.

Fig. 4
figure 4

Forest plots and heat maps of drugs with positive signals for drug-induced cataract based on disproportionality analysis methods from the FAERS database. Notes: The darker the color of the heat map, the higher its signal value, indicating a correspondingly greater risk of cataract. Abbreviations: ROR, reporting odds ratio; PRR, proportional reporting ratio; BCPNN, Bayesian confidence propagation neural network; MGPS, multi-item gamma Poisson shrinker; CI, confidence interval.

Full size image
Table 4 Statistical values and distribution of drug-induced cataract.
Full size table

Classification of cataract-associated drugs by BCPNN values

We applied the BCPNN algorithm to assess the risk of cataract development associated with various drugs, categorizing the risk based on the BCPNN value: values between 0 and 1.5 were classified as low risk, between 1.5 and 3 as moderate risk, and values above 3 as high risk9. Among the 24 drugs evaluated, 11 were classified as high-risk drugs, representing 45.8% of the total. The five drugs with the highest risk levels were difluprednate (BCPNN value = 7.83), prednisolone (BCPNN value = 6.84), erdafitinib (BCPNN value = 5.44), dorzolamide (BCPNN value = 5.15), and insulin lispro (BCPNN value = 5.01) (Fig. 5A).

Fig. 5
figure 5

Ranking of drug-induced cataract by BCPNN values and induction time. Notes: (A) Presents the ranking of drugs associated with cataract based on BCPNN values. (B) Illustrates the median induction time for cataract related to each drug. Abbreviation: BCPNN, Bayesian confidence propagation neural network.

Full size image

Time to onset analysis of cataract-associated drugs

A systematic analysis was conducted on the latency periods for cataract development associated with various drugs, utilizing drug quartiles to assess differences in onset times. The results indicated that the shortest latency periods for cataract onset were observed with the following drugs: difluprednate (74 days), prednisolone (141 days), ranibizumab (163 days), and dexamethasone (184 days). These drugs were associated with the potential to induce cataract-related adverse events within six months of use. In contrast, the drugs with the longest latency periods included pentosan polysulfate (6880 days), human insulin (2922 days), latanoprost (1156 days), and insulin lispro (1141 days). For these drugs, cataract onset typically occurred after more than three years of use (Fig. 5B). Cumulative risk curve analysis revealed statistically significant differences in cataract onset times across drug categories (P < 0.0001) (Fig. 6A). Further inter-group comparisons demonstrated that insulin drugs had the longest average time to cataract development (2270.7 days), while antineoplastic agents exhibited the shortest average onset time (533.9 days) (P < 0.05) (Fig. 6B).

Fig. 6
figure 6

Timeline of cataract adverse reactions induced by related drugs. Note: (A) Presents the Kaplan–Meier curve illustrating the timeline of cataract adverse reactions associated with related drugs. (B) Shows the median induction time for cataract.

Full size image

Discussion

Drug-induced cataract has become a growing challenge to global eye health, particularly among patients who have been using high doses of medications over extended periods, where the associated risks are more pronounced5,6. As the number of available pharmaceutical agents increases and medication regimens continue to evolve, the incidence of drug-induced cataract has steadily risen in recent years3,10. In this study, we analyzed the FAERS database and identified a total of 54,800 cases of drug-induced cataract. To minimize the confounding effect of age, we specifically focused on 2,336 cases of drug-related cataract adverse events in individuals aged 0 to 45 years. Our findings show that women make up a significant proportion of these cases, and over time, the number of reported incidents has increased significantly. By applying four different disproportionate analysis methods, we identified 24 drugs that may be associated with cataract development, which primarily fall into five therapeutic categories: glucocorticoids, insulin, antineoplastic agents, antiglaucoma preparations, and immunosuppressants. Further analysis provided insight into the risk profiles of these drugs and their role in the cataract induction process. These findings contribute valuable real-world data that can guide clinical decision-making and inform future research into drug-induced ocular complications.

Glucocorticoids are extensively used across medical disciplines for their potent anti-inflammatory and immunosuppressive effects. In this study, we identified a strong association between glucocorticoid use and drug-induced cataract formation, with these agents implicated in 42.0% of cases (119/283), consistent with prior findings. These drugs are routinely prescribed for conditions such as multiple sclerosis, rheumatoid arthritis, and uveitis, where long-term administration is often necessary. However, prolonged exposure and high-dose treatment have been associated with systemic complications, including adrenal insufficiency and endocrine dysregulation5,23. Ocular side effects remain a significant concern, with glaucoma and cataract being among the most frequently reported. While glucocorticoid-induced glaucoma results from increased resistance to aqueous humor outflow due to trabecular meshwork dysfunction, cataract formation is primarily attributed to non-genomic mechanisms, including protein-lens interactions, metabolic alterations, and cellular dysfunction, ultimately leading to progressive lens opacification and loss of transparency24. The latency to cataract onset among glucocorticoid users provides further insight into its pathogenesis. Our findings indicate a mean induction period of 808.7 days, suggesting that the risk is both delayed and cumulative5,24.

However, notable differences were observed among individual glucocorticoids. Difluprednate, a potent ophthalmic corticosteroid used in the management of uveitis and scleritis, exhibited the highest cataractogenic potential, with a BCPNN value of 7.83. The median time to cataract induction with difluprednate was significantly shorter than with other glucocorticoids, at only 74 days. This rapid onset raises concerns regarding its ocular safety, particularly in patients requiring extended therapy. The necessity for close ophthalmologic monitoring should be considered when prescribing difluprednate, especially for those at risk of accelerated cataract progression25. A significant association was also observed between fluticasone furoate and cataract development (n = 39). Although prior studies have reported a favorable ocular safety profile over a two-year period, our analysis revealed a median cataract induction time of 1,096 days. This suggests that while short-term exposure may not present an immediate risk, prolonged use could contribute to gradual lens opacification. Given that fluticasone furoate is primarily administered via nasal inhalation, systemic absorption and prolonged exposure may still influence lens metabolism. The extended timeframe for cataractogenesis in these cases underscores the need for long-term ocular evaluation in patients undergoing chronic glucocorticoid therapy, even when administered via non-ocular routes. The mode of administration further influences the risk of glucocorticoid-induced cataract26. Intravitreal injection, a widely employed approach for delivering glucocorticoids in ophthalmic diseases, enhances local bioavailability while reducing systemic exposure. However, our findings suggest that this route may not mitigate the risk of cataract formation. Among the glucocorticoids associated with cataractogenesis, several—including dexamethasone and triamcinolone—are frequently used intravitreally. Prolonged intraocular drug retention may contribute to lens opacification, necessitating a careful evaluation of risk–benefit profiles when selecting treatment strategies. Refining drug delivery techniques and exploring alternative therapeutic options may be essential for minimizing glucocorticoid-induced ocular complications while preserving their clinical efficacy27.

Similarly, insulin, a cornerstone therapy for diabetes management, has been extensively studied for its metabolic effects, yet its potential ocular complications remain underexplored. While diabetes itself is a well-established risk factor for cataract formation, with affected individuals exhibiting a two- to fourfold higher incidence compared to the general population, the direct role of insulin therapy in cataractogenesis has not been well characterized. Existing evidence remains inconclusive, and few studies have systematically investigated this relationship28. Leveraging a large-scale real-world dataset, our study provides novel insights into a potential association between insulin use and an increased risk of cataract development. To account for the confounding effects of diabetes as an underlying condition, we conducted a subgroup analysis stratified by insulin indication status. Our findings revealed a consistent association between insulin therapy and cataract formation, regardless of whether insulin was prescribed within its approved indications or off-label. Notably, the risk was significantly higher in off-label use. Specifically, Insulin Glargine demonstrated a ROR of 2.75 in the therapeutic group, which escalated markedly to 7.02 in the off-label group. Similarly, insulin human exhibited an ROR of 4.11 in the therapeutic group, rising sharply to 11.11 in off-label use. These findings suggest that insulin may contribute to cataractogenesis independently of its intended therapeutic use, with off-label administration conferring a disproportionately elevated risk. To further validate these findings and minimize potential confounding by age, we extended our analysis to include cataract patients across all age groups. The observed association between insulin use and cataract risk remained robust across the broader population, reinforcing the hypothesis that insulin may act as an independent risk factor rather than merely reflecting the underlying disease state. The markedly increased risk associated with off-label use underscores the need for cautious insulin prescribing practices, particularly in populations where long-term administration is required beyond conventional diabetic indications. The underlying mechanisms by which insulin may contribute to cataract formation remain to be fully elucidated. One plausible hypothesis involves insulin-induced alterations in the protein composition of lens fiber cells, which may compromise lens transparency and accelerate opacification29. This effect appears to be cumulative over time, as evidenced by our finding that the median latency period for insulin-associated cataract was 2270.7 days. Given the prolonged nature of this latency, it is likely that chronic exposure to insulin plays a key role in the pathophysiological process. With the global prevalence of diabetes continuing to rise, the number of affected individuals is projected to reach 700 million by 2045. Considering that insulin remains the mainstay of diabetes treatment, its potential ocular risks should not be overlooked30. Clinicians must carefully weigh the long-term therapeutic benefits of insulin against its potential contribution to cataractogenesis, particularly in cases requiring prolonged or off-label use. Proactive risk assessment, routine ophthalmologic monitoring, and early intervention strategies may be essential in mitigating insulin-associated ocular complications, ultimately improving long-term outcomes for patients with diabetes.

Furthermore, other therapeutic agents, including erdafitinib, rituximab, latanoprost, cyclosporine, and pomalidomide, have also been linked to cataract formation, further expanding the spectrum of drugs associated with this ocular complication. Specifically, erdafitinib and rituximab have been previously implicated in cataract formation, and our study further corroborates these findings, demonstrating that these agents can induce cataract via multiple mechanisms6,31. Additionally, latanoprost, a widely used first-line treatment for glaucoma, is known to cause several common side effects, such as skin and iris pigmentation and conjunctival hyperemia32. However, this study is the first to highlight that latanoprost may also contribute to cataract development. Previous studies, including those by Ayaki t al. suggest that latanoprost may accelerate presbyopia by inducing lens sclerosis and altering ciliary muscle function. The onset of presbyopia is closely associated with changes in the lens and ciliary muscle, particularly when lens sclerosis occurs and the ciliary muscle’s accommodation ability declines. This raises the possibility that latanoprost could promote cataract formation by modifying the lens’s microstructure or disrupting its physiological functions. Further research is needed to validate this hypothesis and better understand the underlying mechanisms33. Similarly, cyclosporine, an immunosuppressive drug frequently used post-organ transplantation to prevent immune rejection, has also been linked to cataract progression. Long-term follow-up studies, such as the decade-long study by Dorota Raczyńska on kidney transplant recipients, demonstrate a significant association between cyclosporine use and cataract development, which aligns with the results of our study34,35. In the context of multiple myeloma treatment, the combination of Pomalidomide, dexamethasone, and belantamab mafodotin has proven to be highly effective. However, ocular adverse events, including cataract formation, are notable side effects of this therapy. Preliminary clinical data and pharmacovigilance reports suggest that prolonged pomalidomide use may elevate the risk of lens-related disorders. As a result, it is essential that patients undergoing this treatment receive regular ophthalmologic evaluations to monitor ocular health. Early detection of cataract formation is crucial to prevent potential interference with treatment efficacy. If necessary, treatment regimens should be adjusted based on the patient’s ocular status to ensure both the safety and continuity of therapeutic management36,37.

In conclusion, this study provides a comprehensive exploration of the potential risks associated with drug-induced cataract, emphasizing the risk profiles of various medications and their respective time windows for cataract development. In clinical practice, it is crucial to minimize these risks through careful medication selection, appropriate dosage adjustments, and controlling the duration of treatment. Although our disproportionality analysis offers valuable insights into identifying potential drug-event associations, it is important to note that it cannot establish causality. The observational nature of FAERS data, coupled with the absence of randomized designs, limits our ability to draw definitive causal conclusions. Additionally, the voluntary reporting system of FAERS introduces potential biases such as underreporting or overreporting, which must be considered when interpreting the findings. Moreover, while time-to-onset analysis provides useful temporal context, it may be subject to recall bias and inaccuracies in reported event timing, which are intrinsic limitations of spontaneous reporting systems38,39. To address these limitations, future research should prioritize well-designed prospective cohort studies and randomized controlled trials (RCTs) to establish stronger causal relationships between specific drugs and cataract formation. Furthermore, interdisciplinary collaboration and multi-center studies will enhance the generalizability and credibility of the findings, offering a broader and more robust understanding of the subject. Ultimately, such efforts will be critical in guiding clinical decision-making and minimizing the adverse effects of medications in cataract development.

Conclusion

In conclusion, this study, based on real-world adverse drug reaction data and employing disproportionality analysis, identified 24 drugs potentially associated with cataract induction. By systematically categorizing the mechanisms of action of these drugs and conducting an in-depth risk assessment, we uncovered key epidemiological features of drug-induced cataract and identified several high-risk drugs. The findings provide critical data to support clinical decision-making, with the ultimate goal of reducing the occurrence of such adverse reactions in clinical practice.