Summary
Objectives:
Identifying factors influencing the growth of geographic atrophy (GA) in patients
with age-related macular degeneration (AMD).
Methods:
Data on the natural course and suspected modifying factors were collected as part
of the multicenter, longitudinal, observational FAM-study in 178 eyes of 114 patients
with atrophic AMD. The endpoint of interest – the size of GA – was measured in fundus
autofluorescence images. The influence of different putative riskfactors on progression
of GA is investigated with a forward selection procedure based on the likelihood ratio
test. In orderto interpret non-significant results of the forward selection procedure,
the power of the tests usedwas quantified bya parametric post-hoc bootstrap approach.
Results:
A mean increase in GA of 175 mm2 per year was estimated for the given population (95% CI: [1.46; 2.02]). Patient and
eye-specific random effects could be assessed. Neither patient-specific riskfactors nor ocular-specific risk factors show any significant influence on GA growth. The post-hoc bootstrap
procedure shows that only very strong effects can be detected on the basis of the
given data. For example, the hypercholesteremia which would result in an additional
increase of GA by near 4mm2 per year can be detected with a power of 80%.
Conclusions:
The use of linear mixed effects regression models offers a convenient way to explore
sources of variation in the natural course of GA. Data from further follow-up examinations
and data about other putative riskfactors than those investigated will be needed to
further investigate of the GA growth process. The procedure described in this article
is easily applicable to other putative risk factors as well as to other fields of
application.
Keywords
Macular degeneration - regression analyses - likelihood ratio test - risk factors
- natural history
