Vitrectomy for diabetic eye disease. The prognostic value of pre-operative electroretinography and visual evoked...

An investigation was carried out into the prognostic value of standard flash electroretinography (ERG) and visual-evoked cortical potentials (VECP) recorded for 87 diabetic eyes prior to vitrectomy. 95% of the eyes with a recordable ERG b-wave achieved a post-operative visual acuity (VA) of 0.05 or better (positive predictive value), but only 35% of the eyes with a good post-operative VA had a recordable b-wave (sensitivity). The corresponding values for the ERG a-wave were 67 and 72%, and for the VECP 77 and 69%, respectively. A non-recordable b-wave in the ERG was associated with a poor visual outcome (VA less than 0.05) in 45% of cases (negative predictive value), and 97% of the eyes with a poor visual outcome had a nonrecordable b-wave (specificity). The corresponding values for the ERG a-wave were 42 and 37%, and for the VECP 53 and 63%, respectively. Consequently, the odds ratio for an unfavourable visual outcome (VA less than 0.05) was greater for a non-recordable b-wave (15.7) than for a nonrecordable VECP (3.8) or a non-recordable a-wave (1.5). Recordability of the b-wave gave the best prediction for attached macula. Extensive photocoagulation diminished all ERG and VECP amplitudes. Vitreous opacities had no significant effect on the ERG or VECP.