Abstract
We appreciate the careful reading of our meta-analysis on prevention of cataracts by statins by Macias Saint-Gerons and associates and understand that there are data on both sides on the relationship between statins and cataract.1,2 The anti-inflammatory and antioxidant effects of statins may mediate a decrease in the rate of cataract formation, while bidirectional effects of statins on oxidation and inhibition of appropriate epithelial cell development within the crystalline lens may increase the chance of developing cataract. We have proposed that a well-powered prospective randomized clinical trial be performed in order to confirm the positive or negative relationship of statins to cataract. A protective effect may help increase adherence to statin therapy, while an increased risk of cataract may be balanced by the marked benefits of statins in decreasing cardiovascular events.
We have acknowledged the limitations of our article including the lack of significant effect in randomized trials and the other limitations of meta-analyses. The article by Professor Blumenthal’s group published about 6 months after our meta-analysis agrees with our findings and recommendations. 3
The reason we did not use the 2010 article from the QResearch database by Hippisley-Cox and Coupland 4 is that we used a later publication from the same database (2012) by Collins and Altman.
Repeating our search strategies of MedLine, Web of Science, and the Cochrane library for the intersection of the terms statin and cataract did not retrieve the Waudby reference. 5 Waudby and associates showed an increase in age-specific cataract rate but apparently did not have an internal control group. We repeated our meta-analysis by including the data from the abstract and table 1 of the Waudby article. The overall effect as shown in figure 2 of our meta-analysis remained statistically significant with a higher P value (P = .0467 vs P = .0009). The effect on cataract extraction (figure 5) was not statistically significant (P = .7391 vs P < .0001), while the overall effect remained nearly unchanged (odds ratio [OR] 0.85, 95% confidence interval [CI] 0.81-0.88, P < .0001 vs OR 0.08, 95% CI 0.77-0.83, P < .0001). The data of Machan et al were described in the introduction of our article. They were not included in the meta-analysis since they derive from the Waterloo Eye Study database, where the prevalence of cataract was very high (above the age of 60 varying between 98.2% and 100% in the 70-79 group, 98.9%-100% in the 80+ group, and 80.4%-93.5% among those aged 60-69), rather than the population at large. The observational studies by Leuschen et al and Lai et al6,7 are discussed in our letter to the editor in JAMA Ophthalmology and the related commentary.8,9 We have addressed the issue of statins, cataracts, age, and duration of treatment in our meta-analysis and in a review article where we stated that the data on the relationship of statin therapy to the development of cataract are not consistent and that there appears to be an age effect in that statins may prevent cataract among younger individuals with long duration of exposure.1,10 In summary, the issue of the relationship of statin use and cataracts is not definitely settled, may be confounded by age and may be resolved by a well-powered prospective randomized clinical trial as we have proposed. Depending on the age at randomization and the end point selected (slit lamp examination and cataract extraction), such a study may be too long and/or too expensive to carry out. Incorporation of cataract as a secondary end point in clinical trials of statins and possibly of Proprotein convertase subtilisin/kexin type 9 (PCSK9) antibodies may yield useful information. Also, cataract may be incorporated in large epidemiologic studies such as the Framingham Heart Study, the Atherosclerosis Risk in Communities Study, and so on. A protective effect of statins may help increase adherence to statin therapy, while an increased risk of cataract may be balanced by the marked benefits of statins in decreasing cardiovascular events.