Risk of Cataract in Patients Treated With Statins

Background  Studies in dogs showed that some hydroxymethylglutaryl coenzyme A reductase inhibitors (statins) are associated with cataract when administered in excessive doses. Clinical safety data of statins regarding cataract development in humans have been of limited value so far.

Objective  To determine whether long-term use of statins is associated with an increased risk of cataract.

Methods  We conducted a case-control analysis using data from the United Kingdom–based General Practice Research Database. The main outcome was a first-time diagnosis of cataract and/or cataract extraction in patients aged 40 to 79 years. Controls were matched to cases on age, sex, practice, calendar time, and duration of medical history in the database. Use of statins, fibrates, or other lipid-lowering drugs was compared with nonuse of any lipid-lowering drug, stratified by exposure duration and dose.

Results  We identified 7405 cases and 28 327 controls. Long-term use of statins (eg, ≥30 prescriptions) was not associated with an increased cataract risk (adjusted odds ratio [OR], 0.9; 95% confidence interval [CI], 0.5-1.6), nor was use of fibrates or of other lipid-lowering drugs (adjusted OR, 0.5; 95% CI, 0.3-1.1; and OR, 0.7; 95% CI, 0.1-5.6, respectively). We found evidence that concomitant use of simvastatin and erythromycin, a potent inhibitor of simvastatin metabolism, is associated with an increased cataract risk (adjusted odds ratio, 2.2; 95% confidence interval, 1.2-4.1).

Conclusions  Our study provides evidence that long-term use of therapeutic statin doses does not increase the risk of developing cataract. Concomitant use of erythromycin and simvastatin may increase the cataract risk.

INHIBITORS OF hydroxymethylglutaryl coenzyme A reductase (statins) are effective drugs to decrease morbidity and mortality in patients with hypercholesterolemia.^1-3 Although the overall safety profile of the marketed statins has been shown to be favorable in humans,⁴ concerns have arisen on the basis of long-term animal studies showing that some statins (eg, simvastatin, fluvastatin) are cataractogenic when administered at excessive doses.^5,6

Although isolated case reports have related the use of simvastatin with the development of cataract in humans,⁷ clinical examination of patients treated with statins has not demonstrated any cataractogenic risk.^8-15 However, most previous studies reporting on statin effects on the human lens have been of limited value for a variety of reasons.

We conducted a case-control analysis to explore the risk of developing cataract in patients treated with statins, compared with those treated with fibrates, other lipid-lowering drugs, or no lipid-lowering drug. In animal models, a direct relationship between systemic statin exposure (plasma drug levels) and cataract was demonstrated.⁵ Therefore, we hypothesized that increased systemic availability of statins, potentially resulting from concomitant administration of inhibitors of cytochrome P-450 with statins showing extensive cytochrome P-450 metabolism, may modify the risk of cataract.

This study is based on data derived from the General Practice Research Database (GPRD), which was previously described in detail elsewhere.^16-18 Since 1987, more than 3 million residents in the United Kingdom have been enrolled with selected general practitioners who have agreed to provide data for research purposes to the GPRD. The age and sex distribution of the patients enrolled is representative of the entire United Kingdom population. The general practitioners received 12 months of instruction on the standardized data recording on computer of anonymous information, which they agreed to supply continuously to academic researchers. The information recorded includes patient demographics and characteristics (eg, height, weight, and smoking status), symptoms, medical diagnoses, referrals to consultants, hospital admissions, and drug prescriptions, including the specific preparation, route of administration, dose, and number of tablets for each prescription. On request, hospital discharge and referral letters are available for review to validate the diagnoses recorded in the computer record. The GPRD currently encompasses some 30 million person-years of follow-up. It has been the source for numerous epidemiologic studies in recent years, including studies on statins¹⁹ or cataract²⁰; the accuracy and completeness of GPRD data have been well documented and validated.^17,21,22

We identified all patients who had a first-time diagnosis of cataract (International Classification of Diseases, Eighth Revision [ICD-8] codes 374.x) followed by a referral to a specialist or by a hospitalization because of cataract diagnosis, surgical cataract extraction (Oxford Medical Information System procedure code 156), or both, between January 1, 1994, and September 30, 1998. We included case patients who were 40 to 79 years old at the time of a first-time diagnosis of cataract (subsequently referred to as index date), and who had a recorded medical history in the GPRD of at least 3 years before the index date. Cases were identified in the absence of any exposure information. From previous experience, we knew that a high percentage of computer-recorded cataract diagnoses (>95%) in the GPRD are correct and can be validated by surgery reports and/or documented ophthalmologic assessments.²⁰ Patients with ophthalmologic disorders defined as cataract (ICD-8 code 374.x), corneal opacities (ICD-8 code 371.x), uveitis (ICD-8 code 366.x), eye inflammation (ICD-8 code 369.x), glaucoma (ICD-8 code 375.x), retina detachment (ICD-8 code 376.x), retinal diseases (ICD-8 code 377.x), or uveal disease (ICD-8 code 378.x) before the index date were excluded. Furthermore, we excluded all patients with a history of diabetes (ICD-8 code 250.x), because diabetic patients have a high prevalence of various eye diseases and are therefore more likely than other patients to have ophthalmologic examinations on a regular basis.²³

To each case patient we randomly identified and matched 4 control patients from the base population on age (±2 years), sex, practice, calendar time (by using the same index date), and number of years of recorded medical history in the GPRD before the index date. The same exclusion criteria were applied to control as to case patients.

For each case and control patient, we assessed the exposure history for lipid-lowering drugs from the computerized patient profile. Patients were categorized as users of (1) statins (ie, atorvastatin calcium, cerivastatin sodium, fluvastatin sodium, pravastatin sodium, or simvastatin [lovastatin was not included in our analysis because it was not available in the United Kingdom]), (2) fibrates (ie, bezafibrate, ciprofibrate, clofibrate, fenofibrate, or gemfibrozil), (3) users of other lipid-lowering drugs (ie, acipimox, cholestyramine resin, or colestipol hydrochloride), or (4) "mixed users" (ie, patients who switched between different lipid-lowering drug classes or concomitant users of 2 or more lipid-lowering drug classes). Patients with a history of lipid-lowering drug therapy were further characterized according to the number of prescriptions for lipid-lowering drugs before the index date (ie, 1-9, 10-19, 20-29, or ≥30 prescriptions). We further assessed a cumulative dose based on the number of prescriptions, number of doses per prescription, and dose per tablet.

We also explored relative risk estimates of statin use when given concomitantly with erythromycin, clarithromycin, verapamil hydrochloride, cyclosporine, fluconazole, ketoconazole, or itraconazole, drugs known to significantly increase simvastatin availability to the systemic circulation.^24-26

We conducted a matched analysis (conditional logistic regression) to explore the association between type of exposure (statins, fibrates, other lipid-lowering drugs, or none), exposure duration or cumulative dose, and the risk of developing cataract. In addition to controlling for age, sex, practice attended, and calendar time (by matching), we controlled the analysis for the potential confounders smoking status, body mass index, exposure to corticosteroids, and number of general practitioner visits before the index date (as a marker for the degree of medical attention).

All analyses were performed with the statistical software SAS, version 6.12 (SAS Institute Inc, Cary, NC). Odds ratios (ORs) are presented with 95% confidence intervals (CIs); P values are 2-tailed.

We included 7405 case patients and 28 327 matched control patients in the analysis. The distributions of age, sex, body mass index, smoking status, use of corticosteroids, and number of practice visits preceding the index date of case and control patients are shown in Table 1. In the case group, 3445 patients (46.5%) had a diagnosis of cataract only; the remaining 3960 had cataract removal. The mean recorded medical history in the database before the index date was 6.2 years in both the case and control groups.

Within the case population, we identified 111 patients who were using statins (simvastatin, 72%; pravastatin, 18%; other, 10%). The relative risk estimates (ORs) of having a first-time diagnosis of cataract and/or cataract removal in relation to lipid-lowering drug use adjusted for smoking, body mass index, corticosteroid use, and number of practice visits are shown in Table 2. Overall, long-term use of statins (eg, ≥30 prescriptions) was not associated with an increased risk of cataract, yielding an adjusted risk estimate (OR) of 0.9 (95% CI, 0.5-1.6) in comparison with nonusers of lipid-lowering drugs; the adjusted ORs for fibrates and other lipid-lowering drugs were 0.5 (95% CI, 0.3-1.1) and 0.7 (95% CI, 0.1-5.6), respectively. Stratification of drug use into current (ie, ≥1 prescription for any lipid-lowering drug ≤1 year before the index date) or past (ie, last prescription for any lipid-lowering drug >1 year before the index date) did not modify the risk of developing cataract.

Since there was no material evidence that use of fibrates or of other lipid-lowering drugs was associated with an altered risk of developing cataract, we combined users of these drugs with nonusers of any lipid-lowering drug into one reference group for further analyses. Patients in the mixed group who had exposure to statins were categorized as statin users. As compared with nonuse of statins, the longest duration of statin exposure (ie, use of statins with ≥30 prescriptions) yielded an adjusted OR of 1.0 (95% CI, 0.6-1.5). No evidence of effect modification was found when we stratified by age (40-59 vs 60-79 years), sex, and diagnosis (cataract only vs cataract removal). We further assessed the total cumulative doses of simvastatin and pravastatin, the 2 statins most often prescribed in our study population, and found no evidence of an elevated risk in relation to increasing cumulative doses.

In the analysis of potential risk modification by coadministration of cytochrome P-450–inhibiting drugs, we found some evidence that concomitant use of erythromycin with statins may be associated with an increased risk of cataract (OR, 2.2; 95% CI, 1.2-4.1; adjusted for body mass index and smoking); this association was slightly diminished by further adjusting for corticosteroid use and number of general practitioner visits (Table 3). For subjects with 2 or more concomitant courses of erythromycin, the OR was 3.3 (95% CI, 1.0-10.9); in this stratum, all but 1 patient were simvastatin users (1 used fluvastatin). To exclude the possibility of observing an erythromycin effect, we also analyzed the risk of developing cataract in association with use of erythromycin in the absence of statin exposure. As compared with nonusers of erythromycin, the relative risk estimate for subjects who received 4 or more erythromycin courses before the index date was 1.1 (95% CI, 0.9-1.3).

The use of other drugs known to inhibit the metabolism of certain statins did not modify the risk of cataract, or the numbers of subjects in substrata were too small to yield informative results.

We also found an association between use of inhaled corticosteroids and cataract development. Users of 30 or more prescriptions of inhaled corticosteroids only (ie, without exposure to oral corticosteroids) had a relative risk estimate of 1.2 (95% CI, 1.0-1.6; adjusted for body mass index, smoking, number of general practitioner visits, and use of statins). Users of 30 or more prescriptions of oral corticosteroids only (ie, without exposure to inhaled corticosteroids) had a relative risk estimate of 1.8 (95% CI, 1.4-2.3; adjusted for body mass index, smoking, number of general practitioner visits, and use of statins).

Long-term administration of simvastatin and other statins (eg, lovastatin and fluvastatin) has been associated with cataract development in dogs,^5,6 while atorvastatin and pravastatin lacked a cataractogenic effect in dogs.^27,28 Data from observational and clinical studies did not show an increased cataract risk associated with use of statins in humans so far.^8,13-15,29 However, most previous studies reporting on statin effects on the human lens have been of limited value for a variety of reasons, such as absence of a control group, small numbers of patients involved, short duration of treatment, and/or the deficiencies of the ophthalmic assessments conducted.

This large case-control analysis provides evidence that therapy with statins at therapeutic doses is not associated with an increased risk of cataract development in humans. However, since the majority of patients analyzed in this study were exposed to statin doses at the lower therapeutic range (eg, 20 mg of simvastatin per day), we do not know whether our finding is also true for patients exposed to high therapeutic statin doses. The exact mechanism of statin-induced cataract in animals is unclear. In animals, no relationship could be established between decrease in circulating cholesterol levels at pharmacologically equipotent statin doses and the incidence of lenticular opacities, but a direct relationship between plasma statin levels and cataract incidence was observed. Statins producing high circulating plasma levels were associated with a higher incidence of cataract formation.⁵ The currently approved maximal simvastatin dose for use in patients with hyperlipidemia (40 mg per day) produces plasma levels approximately 14 times lower than the minimally cataractogenic dose in dogs (50 mg/kg per day) and is approximately 5 times lower than the noncataractogenic dose in dogs (10 mg/kg per day).³⁰ Since a dose-dependent effect has been established for statins with proved cataractogenic potential,³⁰ we hypothesized that while long-term administration of therapeutic statin doses may not increase the risk of cataract, concomitant administration of drugs that significantly increase the systemic availability of statins might modify the cataract risk in association with statin therapy. Human in vivo and in vitro data have shown that statins, with the exception of pravastatin,³¹ are substrates of the cytochrome P-450 system (CYP450),⁴ especially CYP3A4 (ie, atorvastatin, cerivastatin, lovastatin, and simvastatin)^31-33 or CYP2C9 (ie, fluvastatin).³³ Drug interaction studies have shown that concomitant administration of CYP3A4 inhibitors with simvastatin is associated with elevated systemic availability with the potential to increase the risk of serious adverse effects of statins, such as myopathy and rhabdomyolysis.^4,34 Among the substances that have been shown to lead to a 5-fold to 10-fold increase in systemic availability or peak concentrations of simvastatin are erythromycin,²⁴ itraconazole,²⁵ ketoconazole,³⁵ verapamil,²⁴ and cyclosporine.³⁴ The enzymes responsible for the biotransformation of pravastatin are less well known,³⁶ but CYP3A4 contributes only negligibly to its metabolic fate, if at all.³⁷

Our analysis indicates that concomitant exposure to erythromycin with statins (which was almost exclusively exposure to simvastatin) may be associated with an approximately 2-fold to 3-fold increased risk of developing cataract in comparison with nonusers of statins. An additional analysis provided evidence that the relative risk of developing cataract was not increased for subjects who repeatedly used erythromycin, but not statins, suggesting that the elevated relative risk estimate was seen only in subjects using both simvastatin and erythromycin concomitantly. Because of relatively small numbers in these cells, however, these results need to be interpreted cautiously. Additional results such as increased cataract risk associated with smoking or exposure to oral or inhaled corticosteroids are in agreement with previous findings.^23,38,39

As with observational studies in general, we cannot rule out unknown biases or confounders as possible alternative causes for our findings. However, our results were adjusted for age, sex, geography, and calendar time (by matching cases to controls) as well as for smoking status, body mass index, corticosteroid use, and degree of medical attention (by adjusting the multivariate analysis), and substantial distortions by these potential confounders seem unlikely. A potential difficulty of this observational study is that cataract is a disease of slow onset; thus, the index date is relatively poorly defined. This can potentially lead to a certain exposure misclassification, ie, exposure before the index date is taken into account in the analysis even though use of study drugs occurred after the real onset of the disease. To reduce the risk of getting distorted results, we did an additional analysis in which we did not take any exposure to lipid-lowering drugs into account that occurred in the year immediately preceding the index date. This assessment did not yield different results.

In conclusion, the present large case-control analysis provides further evidence that therapy with statins at therapeutic doses is not associated with an increased risk of cataract development. However, we cannot extrapolate our findings to those beyond the observed period of exposure. It has been suggested that the ocular safety of statins can be established only in light of 10 to 20 years of clinical experience.⁴⁰ We found evidence that concomitant administration of simvastatin and erythromycin, a drug known to significantly increase simvastatin bioavailability, may be associated with an increased risk of developing cataract, similar to the well-documented interaction of certain statins with cytochrome P-450 inhibitors leading to myopathy.⁴ This potentially important finding needs to be explored further.

Accepted for publication December 5, 2000.

Dr Meier is the recipient of grant 32-056 751 from the Swiss National Science Foundation, Berne, Switzerland. Research projects on drug-drug interactions of the Department of Internal Medicine VI, Clinical Pharmacology and Pharmacoepidemiology, University Hospital of Heidelberg, Heidelberg, Germany, are supported by BMBF grant 01EC9902 from the Federal Ministry for Education and Research in Bonn, Germany.

The Boston Collaborative Drug Surveillance Program is partly supported by grants from AstraZeneca, Sodertalje, Sweden; Bayer AG, Leverkusen, Germany; Berlex Laboratories, Wayne, NJ; Boots Healthcare International, Nottingham, England; Bristol-Myers Squibb, Princeton, NJ; GlaxoWellcome Inc, Research Triangle Park, NC; RW Johnson Pharmaceutical Research Institute, Raritan, NJ; McNeil Consumer Products, Fort Washington, Pa; and Hoffman-La Roche and Novartis Pharmaceuticals, Basel, Switzerland. This study was not directly funded by any of these companies.

Presented in part at the 16th International Conference on Pharmacoepidemiology of the International Society for Pharmacoepidemiology, Barcelona, Spain, August 21, 2000.

We thank the participating general practitioners for their excellent cooperation.

Corresponding author and reprints: Christoph R. Meier, PhD, MSc, Basel Pharmacoepidemiology Unit, Division of Clinical Pharmacology, University Hospital of Basel, Petersgraben 4, CH-4031 Basel, Switzerland (e-mail: Christoph.Meier@unibas.ch).