advertisement

Editors Selection IGR 15-3

Glaucoma and Systemic Diseases: POAG and Glucose metabolism disorders

Comment by Paul Healey on:

59521 Diabetes, glucose metabolism, and glaucoma: the 2005-2008 national health and nutrition examination survey, Zhao D; Cho J; Kim MH et al., PLoS ONE, 2014; 9: e112460

Find related abstracts

In 1997 I authored a paper which reported a positive association between Diabetes and glaucoma in the Blue Mountains Eye Study¹. This stood in contrast to a report two years earlier from the Baltimore Eye Survey which did not find any evidence for an association². We could find no clear indications of bias in either study, leaving us with the somewhat unsettling conclusion that data were inconsistent. Numerous subsequent studies were also inconsistent. Inconsistency is the worst outcome of multiple investigations as it neither offers direction for scientific research nor allows us to advise our patients. One approach to shed light on such dichotomy is the systematic review and meta-analysis which seeks to increase power by pooling like cohorts as well as showing relative results weighted for power. This is the approach taken by Zhao et al in a recent paper in JAMA Ophthalmology. The key to a good meta-analysis is to cast a wide net for studies, exclude poor quality or non-generalizable results, weight based on study quality and analyze like studies together. Zhao et have done a good job in this respect, identifying 47 cross-sectional, case-control and longitudinal studies of almost 3 million individuals from many parts of the world. Point estimates of relative risk ranged from strongly positive (4.20) to moderately negative (0.61). Most studies reported positive point estimates but many were not statistically significant. In general smaller studies reported stronger associations, a common manifestation of publication bias. But the overwhelming evidence was for a positive association when pooled by study type or dose-response. Consistent (though weaker) positive associations were also found with intraocular pressure or ocular hypertension. The authors found that risk of glaucoma increases by 5% for each year since diagnosis of diabetes with an overall risk of about 1.5 compared to individuals without diabetes. This is an almost identical finding of an earlier meta-analysis conducted on 12 studies in 2004. So did we need another meta-analysis? This one included more studies and importantly more higher quality studies published since the previous paper. Do we need more studies investigating whether an association exists between diabetes and glaucoma? Probably not. Rather, given the weak association with IOP we need research into how a vascular disease caused by a neuroprotectant can lead to neurodegeneration.

• ← Previous
• Next →