Abstract
Purpose: :
We have previously estimated the co-prevalence of open-angle glaucoma (OAG) and age-related macular degeneration (AMD), using the Beaver Dam Eye Study (BDES) and Los Angeles Latino Eye Study (LALES) population-based data (Racette et al, ARVO 2007, #4328). The purpose of the present study is to determine whether these observed co-prevalences are similar to those predicted from the prevalences of OAG-alone and AMD-alone. Differences in observed, compared to predicted, co-prevalences would suggest some dependence between the two diseases.
Methods: :
If OAG and AMD are independent of each other, then their co-prevalence can be predicted using joint probability [(probability of OAG-alone X probability of AMD-alone) X 100]. We used the prevalences observed in the BDES and LALES for OAG-alone (definite and probable combined) and AMD-alone (both early and late stages), to predict the co-prevalences of OAG and AMD in people over 40 years of age. We then derived observed-to-predicted co-prevalence ratios. Ratios different from 1 would suggest that OAG and AMD may not be independent of each other: ratios greater than 1 would suggest that the likelihood of detecting one disease is greater if the other disease has been detected, and ratios smaller than 1 would suggest that the likelihood of detecting one disease is smaller if the other disease has been detected.
Results: :
In the BDES, the prevalences of OAG, early AMD and late AMD were 4.06, 18.06 and 1.32, respectively. The observed and predicted co-prevalences of OAG and early AMD were 0.93 and 0.73, respectively. The observed-to-predicted ratio was 1.27. The observed and predicted co-prevalences of OAG and late AMD were 0.20 and 0.05, respectively. The observed-to-predicted ratio was 3.73. In the LALES, the prevalences of OAG, early AMD and late AMD were 4.65, 9.13 and 0.41, respectively. The observed and predicted co-prevalences of OAG and early AMD were 0.70 and 0.42, respectively. The observed-to-predicted ratio was 1.65. The observed and predicted co-prevalences of OAG and late AMD were 0.07 and 0.02, respectively. The observed-to-predicted ratio was 3.67. All observed-to-predicted ratios were greater than 1.
Conclusions: :
In both the BDES and LALES studies, the co-prevalence of OAG and AMD are higher than would be predicted from the individual prevalences of OAG- and AMD-alone. This new information suggests that OAG and AMD may not be completely independent of each other.
Keywords: clinical (human) or epidemiologic studies: prevalence/incidence