Predicting bacterial cause in infectious conjunctivitis

Editor—The development of clinical rules that are easy to use to help discriminate between bacterial and culture negative acute conjunctivitis would be welcomed in primary care, but are the three simple questions proposed by Rietveld et al the right questions?¹ One large problem is the use of itching to count against the likelihood of bacterial infection. Itching was present in 33 of 57 patients with positive culture results and was slightly more prevalent in those with negative cultures (63% v 58%).

Figure 1.

Credit: DR P MARAZZI/SPL

Using itching to count for or against a diagnosis of bacterial conjunctivitis is wrong, as no significant difference is seen before and after logistic regression. I would not use the answer to this question to make treatment decisions on the basis of the data presented. A history of conjunctivitis seems helpful in a few patients, but this would not affect treatment choice in the majority. The best question seems to be, “Are both your eyes glued in the morning,” with the biggest odds ratio after logistic regression analysis of 14.99. I am puzzled by the odds ratio of 2.68 for one eye glued in the morning as the numbers of 53% culture positive and 62% culture negative would imply an odds ratio of below 1.0.

The authors conclude that the assessment could possibly be done over the telephone, but in their exclusion criteria, ciliary redness is mentioned, and most doctors would be hard pressed to diagnose this over the telephone.