Fitty-seven diabetic patients (37 with no apparent DR, 11 with mild to moderate NPDR and 9 with severe NPDR to PDR) and 45 controls submitted to first-eye cataract surgery were enrolled in the study. Patient demographics and vision-related preoperative characteristics were similar between the control and diabetic groups, as well as across the diabetic subgroups (Table 1); except for HbA1c levels (p<0.001, One-way ANOVA).
In the diabetic group, longer duration of DM was significantly associated with higher HbA1c levels (p<0.001, r=0.519, Spearman correlation rank). More advanced stages of DR were significantly associated with longer duration of DM (p<0.001, Kruskal-Wallis test). The HbA1c levels were significantly greater in the severe NPDR/PDR subgroup when compared to the no DR subgroup (p=0.016, One-way ANOVA bonferroni adjusted). All diabetic subjects exhibited symmetrical DR stage preoperatively.
Concerning previous DR treatements, one patient with moderate NPDR had received macular laser photocoagulation five years before surgery; one patient with severe NPDR received macular laser photocoagulation and multiple intravitreal anti-VEGF injections for diabetic macular edema over one year before surgery; all six patients with antecedents of PDR were treated with macular and panretinal laser photocoagulation, and three of them received multiple intravitreal anti-VEGF injections over 1 year before surgery.
Phacoemulsification alone was performed in all patients from the control group and in diabetic patients with no DR. In the mild/moderate NPDR subgroup, seven patients (64%) underwent phacoemulsification alone and four patients (36%) with co-adjuvant intravitreal bevacizumab. In the severe NPDR/PDR, phacoemulsification was performed simultaneously with intravitreal bevacizumab in seven patients (78%) and with intravitreal triamcinolone in two patients (22%).
No progression of retinopathy was observed in the operated and nonoperated eyes during the course of the study. At 1-month visit, only three diabetic patients had an increase in the central retinal thickness of more than 10% in the operated eye (incidence: 5%): one from the no DR subgroup (wihtout intraretinal cysts) and one from the mild/moderate NPDR subgroup (with intraretinal cysts) - both had phacoemulsification alone and edema resolved with 1mg/ml nepafenac topical eye drops three times daily for two months; another one from the PDR subgroup who had preoperative diabetic macular edema and received phacoemulsification with co-adjuvant intravitreal bevacizumab – edema did not respond to topical nepafenac but improved with triamcinolone injection at 4 months.
Preoperative diabetic macular edema in the operated eye was present in three patients with moderate NPDR, one with severe NPDR and four with PDR (prevalence: 14%). After surgery, the edema remained stable or improved in all patients up to 6 months; except for three patients with PDR who needed intravitreal injections (the one who got worse 1-month after the surgery and two others that got worse during the follow-up). The fellow nonoperated eyes of two patients with moderated NPDR and two patients with PDR showed progression of diabetic macular edema during the follow-up and were treated with intravitreal injections.
Visual outcomes
Visual Acuity
No statistically significant differences were observed in the preoperative mean corrected distance visual acuity (CDVA) of the operated and nonoperated eyes between the control group and diabetic subgroups (p=0.762 and p=0.054, respectively; One-way ANOVA) (Table 1).
Regarding the nonoperated eyes, there were no statistically significant differences across the groups for the mean CDVA change from baseline to 6 months (Fig. 1) and the mean CDVA at 6-month visit (p=0.607 and p=0.128, One-way ANOVA).
The mean CDVA of the operated eyes was observed to be significantly higher than preoperative value in all groups at 1- and 6-month follow-up (p<0.05, paired t-test) (Fig. 1). The mean CDVA change from baseline to 6-month visit of the operated eyes did not differ between the control group and diabetic subgroups (p=0.687, One-way ANOVA); however, the mean CDVA at 6-months in the severe NPDR/PRD subgroup (0.20 [95% CI, 0.03 to 0.37]) was significantly lower compared to the control group (0.01 [95% CI, 0.00 to 0.02]) and the remaining diabetic subgroups (no DR, 0.02 [95% CI, 0.01 to 0.04]; mild/moderate NPDR, 0.00 [95% CI, 0.00 to 0.00]) (p<0.001, One-way ANOVA bonferroni adjusted) (Table 1).
Final CDVA was 20/32 or greater in 100% of the eyes in the control group and in 95% of the eyes in the diabetic group (100% in those with no DR and mild/moderate NPDR; 67% in those with severe NPDR/PRD). The operated eye became the better eye or had a visual acuity equal to that in the fellow eye in 98% of cases in the control group and severe NPDR/PRD subgroup; and 100% in the two-remaining diabetic subgroups. None of the patients in our study lost vision after cataract surgery. Overall, 98% of control subjects had at least a 2-lines of improvement in CDVA compared with 86% of patients with diabetes (89% with no DR, 100% of those with mild/moderate NPDR and 67% of those with severe NPDR/PDR).
NEI VFQ-25 Composite and Subscale results
The internal consistency reliability of the questionnaire in this sample was high. The overall Cronbach’s α statistic was 0.850 and 0.855 at baseline and 6-month visits, respectively. Driving subscale obtained a small number of responses (n=29); Cronbach’s α statistic was from 0.733 and 0.712 to baseline and 6-month visits, respectively.
The preoperative NEI VFQ-25 composite and subscale scores were comparable across the groups (Table 2). No correlation was observed between the preoperative NEI VFQ-25 composite score and the preoperative CDVA of the operated and nonoperated eyes (p>0.05, Spearman correlation). There were no differences between groups for the postoperative change of each of the questionnaire variables; however, at 6-months evaluation, there were significant differences between groups for the color vision, social functioning and ocular pain subscales (Table 2).
The mean composite score change from baseline to 6-month visit were 15.4 points (95% CI, 11.4 to 19.4), 15.5 points (95% CI, 11.4 to 19.9), 13.0 points (95% CI, 3.1 to 22.9) and 18.3 points (95% CI, 7.3 to 29.2) for the control group, no DR subgroup, mild/moderate NPDR subgroup and severe NPDR/PDR, respectively (p=0.849, one-way ANOVA) (Fig. 2). No group differences were observed regarding the categorical analysis of the NEI VFQ-25 composite score variation (p=0.280, Fisher´s exact test) (Fig. 3).
Regarding the NEI VFQ-25 subscales scores, statistically significant increases were observed in all groups for the variables general vision, near activities, distance activities and peripheral vision (Fig. 2). The general health subscale was the only item for which there was no significant improvement in any of the groups. Color vision was found to significantly improve in the control group; social functioning score increased in the control group and severe NPDR/PDR subgroup; driving, dependency and role difficulties were significantly better in the control group and no DR subgroup; ocular pain and mental health were improved in the control, no DR and severe NPDR/PDR subgroups.
Multivariate linear regression showed that age, preoperative CVDA of the operated eye, preoperative NEI VFQ-25 composite score and CDVA change from baseline to 6 months were significantly associated with NEI VFQ-25 changes at 6 months postoperatively (Table 3). In a “fixed model”, the NEI VFQ-25 composite score was found to significantly increase on average 0.31 points for each year of increase in age; to decrease on average 3.93 points for each line of increase in preoperative CDVA (logMAR); to decrease on average 0.67 points for each unit of increase in preoperative NEI VFQ-25 composite score; to increase on average 4.19 points for each line of improvement in CDVA from baseline to 6 months.
Patient satisfaction questionnaire
The internal consistency reliability of the questionnaire in this sample was high. The overall Cronbach’s α statistic was 0.891.
Overall patient´s satisfaction at 1-month visit was 93%, 89%, 100% and 100% for the control group, no DR subgroup, mild/moderate NPDR subgroup and severe NPDR/PDR, respectively (p=0.968, Fisher´s exact test).
The mean patient satisfaction score was comparable between groups with composite scores of 64.2 (95% CI, 57.7 to 70.8), 71.4 (95% CI, 63.9 to 78.8), 63.6 (95% CI, 49.6 to 77.6), and 73.1 (95% CI, 59.6 to 86.6) for the control group, no DR subgroup, mild/moderate NPDR subgroup and severe NPDR/PDR, respectively (p=0.420, one-way ANOVA). No group differences were observed with respect to patient´s self-reported improvement in vision, quality of life or surgery expectations (p=0.426, p=283 and p=0.346, respectively; one-way ANOVA).