Abstract
We are happy to clarify the issue raised by professor Hippisley-Cox in identifying a mistake in our letter to the editor regarding statin use and risk of cataracts: firm conclusions are still off. 1 We performed a new meta-analysis including both the Collins and Altman and the Hippisley-Cox and Coupland articles 2,3 using data from table 4 of the Hippisley-Cox and Coupland article. 3 In this analysis, the odds ratio (OR) is closer to 1, the confidence interval (CI) is wider, and the P value is higher than .05 (OR 0.86 95% CI 0.68-1.07, P = .1744) compared to our previous report (OR 0.82 95% CI 0.72-0.92, P = .0011). This was expected from the directionally opposite statistically significant effects of the Collins and Altman article (OR 0.83, 95% CI 0.80-0.87, P < .0001) compared to the Hippisley-Cox and Coupland article (OR 1.31, 95% CI 1.27-1.36, P < .0001). 3
We had questions about the article of Collins and Altman and tried to obtain clarification via e-mail. 2 In addition, a reviewer of our original submission commented on the skewed distribution of the numbers of participants included in the meta-analysis (Collins et al, over 2 million persons; data from all other studies ranged from 58 to 105 454 participants). For these reasons, we performed an additional meta-analysis excluding the data of Collins. 4 This analysis showed a significant protective effect of statins (OR 0.81, 95% CI 0.69-0.95, P = .0105).
As we stated in our letter, 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 proposed previously. 5