Posters

Cataract and visual function: a prospective observational study of surgical selection

Poster Details

First Author: S.Gholami THE NETHERLANDS

Co Author(s):    P. Aspinall   B. Dhillon   S. Klijn   N. Reus   T. van den Berg        

Abstract Details

Purpose:

Although most surgeons agree that visual acuity (VA) is a very important parameter for the indication for cataract surgery, it is also agreed there are other visual function factors to consider such as straylight (SL) and contrast sensitivity (CS) that may correspond to patients’ complaints. This presentation will focus on finding the best model for cataract surgical selection if VA (logMAR) would be combined with straylight (SL) assessment and contrast sensitivity (CS).

Setting:

Data collection was carried out in Princess Alexandra Eye Hospital, Edinburgh, Scotland, UK; Analyses were performed in Rotterdam Ophthalmic Institute, Rotterdam, the Netherlands.

Methods:

The data collection was conducted during a 12-month prospective study. Among 184 phakic eyes enrolled, 96 eyes were listed for surgery. We presently focus on a subgroup with a Snellen VA ≥ 0.5 (36 eyes out of 114 that were listed). The decision for surgery was made by the ophthalmologist based on Snellen VA and slit lamp examination. Additional tests, including visual function (logMAR VA, SL, and CS) and cataract morphology (OCGS), were performed by a research optometrist after the decision for surgery was made.

Results:

A model with SL and CS combined as a predictor for listing for cataract surgery could predict 73% of cases correctly (model improvement 4.5%); a model with only SL as predictor could predict 75% correctly. Including all three indicators in our model, it could predict 78% of cases correctly; whereas removing CS from this model increased this number to 80% (model improvement 10%). In all four models CS had no statistically significant impact on the prediction (P≥0.05). With VA as control variable, a significant correlation between SL and surgical decision was detected (r=0.19, P=0.02). No such correlation was found for other variables.

Conclusions:

In patients with a Snellen VA ≥ 0.5, a model with only SL as indicator for listing for cataract surgery can predict as well as a model with logMAR VA, SL, and CS combined. However, a model with logMAR VA and SL combined showed to be the best model in terms of correct classification. Therefore, in the future SL might be used as an indicator for cataract surgery alongside VA.

Financial Disclosure:

NONE