Purchase this article with an account.
A. Hessellund, D. Larsen, B. Gjedsted, T. Bek; A Decision Model for Exudative AMD Based on Multivariate Analysis of Risk Factors. Invest. Ophthalmol. Vis. Sci. 2009;50(13):720.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The introduction of VEGF inhibitors for the treatment of exudative AMD has increased the referral rates of AMD patients for treatment considerably. However, only a subgroup of the referred patients are candidates for treatment and previous studies have shown that these patients might be identified before referral on the basis of early symptoms and clinical findings. However, these studies have identified parameters of interest from univariate analaysis only. In order determine independent parameters with high predictive value multivariate analysis is necessary.
1041 patients referred to our department with the tentative diagnosis of exudative AMD underwent a structured interview about the presence and duration of blurred vision, central dark spot, metamorphopsia, micropsia and dyschromatopsia, followed by a measurement of visual acuity, funduscopy, central retinal thickness using OCT, and fluorescein angiography. Based on the examination the patients were divided in to two groups, i.e.: 1) Exudative (wet) AMD (N=422), and 2) Atrophic (dry) AMD (N=619). Odds ratios for the association between the included parameters and exudative AMD were calculated using binary multivariate regression analysis.
The following symptoms were independently associated with exudative AMD: Blurred vision (OR=1.64, p<0.01), central dark spot (OR=1.69, p<0.01), dyschromatopsia (OR=1.71, p=0.01), sudden onset (OR=1.48, p=0.01), and worsening of symptoms since onset (OR=1.36, p=0.04). Central retinal thickness (p<0.01), the presence of subretinal hemorrhage (OR=3.53, p<0.01) and exudates (OR=3.03, p<0.01) identified by funduscopy differed among the two AMD types.
All the parameters previously identified by univariate analysis were also of significant importance using multivariate analysis indicating that the specific symptoms and clinical findings are independent. These data can be used to construct a decision model that calculates the risk of having treatment requiring wet AMD given a specific combination of early symptoms and clinical signs.
This PDF is available to Subscribers Only