April 2009
Volume 50, Issue 13
ARVO Annual Meeting Abstract  |   April 2009
Longitudinal Model of Choroidal Neovascularization (CNV) Event Risk Based Primarily on Fundus Images
Author Affiliations & Notes
  • T. R. Friberg
    Ophthalmology/UPMC Eye Cntr, Univ of Pittsburgh, Pittsburgh, Pennsylvania
  • R. A. Bilonick
    Ophthalmology/UPMC Eye Cntr, Univ of Pittsburgh, Pittsburgh, Pennsylvania
  • P. M. Brennen
    Ophthalmology/UPMC Eye Cntr, Univ of Pittsburgh, Pittsburgh, Pennsylvania
  • Footnotes
    Commercial Relationships  T.R. Friberg, None; R.A. Bilonick, None; P.M. Brennen, None.
  • Footnotes
    Support  NIH R03 EY13807; Eye and Ear Foundation (Pittsburgh, PA); Research to Prevent Blindness
Investigative Ophthalmology & Visual Science April 2009, Vol.50, 1256. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      T. R. Friberg, R. A. Bilonick, P. M. Brennen; Longitudinal Model of Choroidal Neovascularization (CNV) Event Risk Based Primarily on Fundus Images. Invest. Ophthalmol. Vis. Sci. 2009;50(13):1256.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

Purpose: : To use longitudinal data to help define risk of CNV event occurrence. At present, other than vitamin supplementation, there are few strategies to prevent visual loss from dry or non-neovascular AMD. Treatment of AMD is largely targeted toward those who have suffered from neovascular or wet AMD. Identification of the risk of conversion to wet AMD based on a patient’s longitudinal data has obvious advantages.

Methods: : We used data from 513 patients (844 eyes) who were seen longitudinally in both the ARED Study (441 patients and 772 eyes) and PTAMD Study (72 patients and 72 eyes). Some were seen over an eight-year interval. Previously, we assessed risk factors for development of wet AMD from baseline characteristics, and most studies use similar methodology. In this study, we used a longitudinal logistic regression to refine the risk profile. Using a random intercepts model fitting to event occurrences, we lagged the explanatory factors by one time period. Such a model allows for differing risks among subjects and provides estimates of co-variant risks on each subject in contrast to marginal models which only estimate population averaged effects. Furthermore, the model avoids the large influence of the neovascular event itself on the visual acuity, an effect which can obfuscate and overwhelm more subtle time-dependent changes.

Results: : Of the 513 subjects, 33 eyes (6.4%) had events. Our morphological variables included the presence of pigment, age, whether or not the fellow eye was affected by CNV, ETDRS visual acuity, drusen area, and drusen distribution. Only visual acuity and follow up time (aging) had statistically significant effects. A one-letter drop in visual acuity presages a 12% increase in the odds for a CNV event (1.012, p<0.0001). Each day in the study increased the odds for such an event by about 3 tenths of one percent (1.00308, p=0.0090). There were no detectable differences in these effects between the two study groups.

Conclusions: : A patient’s drop in visual acuity presages an increased likelihood of conversion to wet neovascular AMD. The likelihood also increases as subjects age. The lack of detectable effects due to pigment and drusen are in our model may possibly be attributable to the relatively small number of observed events.

Keywords: drusen • macular pigment • visual acuity 

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.