Risk factors vary for CNV and geographic atrophy

October 1, 2007

Data analysis from the Complications of Age-related Macular Degeneration Prevention Trial shows that choroidal neovascularization and geographic atrophy share some risk factors but have different influences as well.

Key Points

Philadelphia-Data from the Complications of Age-related Macular Degeneration Prevention Trial (CAPT) revealed differences in the risk factors for choroidal neovascularization (CNV) and geographic atrophy (GA). Older age and focal hyperpigmentation were risk factors for both. Cigarette smoking and hypertension were associated only with CNV, whereas the percent of global area covered by drusen and the extent of retinal pigment epithelium (RPE) depigmentation were risks associated only with GA, said Gui-shuang Ying, PhD, assistant professor of ophthalmology and senior biostatistician, Center for Preventive Ophthalmology and Biostatistics, Scheie Eye Institute, University of Pennsylvania, Philadelphia.

The CAPT study included 1,052 participants aged 50 or more years with at least 10 large drusen (>125 μm), visual acuity of 20/40 or better in each eye, and no other ocular diseases. One eye was randomly assigned to laser treatment and the other to observation. The primary outcome was 5-year change in visual acuity, and the secondary outcome was incidence of CNV and GA.

Follow-up visits

Blood pressure was measured at baseline, and patients provided demographic information as well as history of diabetes, cigarette smoking, and use of hypertensive medications. Trained readers evaluated color photographs for baseline drusen characteristics and pigment abnormalities.

Mean age of patients was 71 years, and 61% were women. Assessment of participant-level risk factors showed that 6% were current smokers and 49% were former smokers.

In addition, 47% had definite hypertension and 18% had suspected hypertension. A history of diabetes was noted in 9%, and 64% were aspirin users.

For ocular risk factors, drusen characteristics were evaluated within 3,000 μm of the foveal center. The largest drusen size (>250 μm) was found in 70% of eyes, and the predominant drusen size (>125 μm) was present in 50%. The percent of global area covered by drusen was <10% in 65% of eyes, 10% to 24% in 27% of eyes, and > 25% in 5% of eyes.

Drusen confluence of >10 pairs occurred in one-half of the eyes. The number of drusen (>125 μm) within 500 μm of the foveal center was zero in 8% of eyes, <10 in 85% of eyes, and >10 in 4%.

Either none or questionable hyperpigmentation was found in 29% of eyes; an area <250 μm was found in 54%, and at least 250 μm in 14%. RPE depigmentation was observed in 5% of eyes.

During the follow-up period, trained readers identified CNV through leakage on fluorescein angiography and endpoint GA (>1 dark adaptation combined area on color photographs). Estimates of relative risks and 95% confidence intervals were obtained from univariate and multivariate survival analyses of observed and treated eyes, considered separately and combined with treatment as a covariate.

Overall, 5-year results showed no benefit to laser treatment, Dr. Ying said. About 21% of both treated and observed eyes lost 3 or more lines of visual acuity, CNV developed in 13% of eyes in both groups, and GA developed in 7% and 8%, respectively.

Researchers then analyzed the data to assess the risk factors for CNV and GA, evaluating baseline participant factors and baseline drusen features and pigmentary abnormalities.

Multivariate analysis showed that in the observed eyes, current smoking was a significant risk factor (RR 1.89, 1.03 to 3.47) compared with never smoking, as was definite hypertension (RR 1.55, 1.07 to 2.25) compared with no hypertension.