Statistics from Altmetric.com
Knudsen and Lervang published an interesting article in the BJO recently.1 The authors concluded that a cilio-retinal artery tends to worsen diabetic maculopathy but I believe there is an alternative interpretation of the data provided. The authors mentioned that 26/29 patients had a cilio-retinal artery in the eye with worse maculopathy, which was a statistically significant result. In the results section the authors did not report whether a unilateral cilio-retinal artery was present in the 75 subjects with identical maculopathy in the right and left eyes but they stated that this was true in all 75 subjects in the first sentence of the discussion. Therefore, in 75 patients the cilio-retinal artery had not caused worsening of the maculopathy compared to the contralateral eye. Taken together with the 29 subjects with asymmetrical maculopathy, the proportion of subjects in whom the unilateral cilio-retinal artery did not worsen maculopathy would be (75 + 3)/104 or 0.75 (95% CI 0.654 to 0.827). The proportion of cases in whom maculopathy was not worsened by a cilio-retinal artery was therefore significantly more than 50%.
The authors also reported that 51 eyes had clinically significant macular oedema (CSMO), in the group of 94 subjects with maculopathy. However, the χ2 analysis should not have been performed using each eye as a case (n=962), as it is well known that the right and left eyes of each patient are not independent (that is, if the right eye has maculopathy, the left eye has a greater chance of having maculopathy than the contralateral eye of a non-affected eye in another patient). One possible approach would be to look at all the right eyes only or to randomly look at one of the two eyes of each subject and reduce n to 481. The reader of this article will not know the outcome of this analysis without more information from the authors; however, assuming all the 51 eyes with CSMO were from different subjects, the 2 × 2 table may look like Table 1. This table assumes that all the 104 eyes with unilateral cilio-retinal artery and the 15 eyes from bilateral cilio-retinal cases are included, making a total of 119 subjects with cilio-retinal arteries for analysis. In the above case clearly the result is not statistically significant (χ2 =0.017; p>0.9).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.