Predictors of Referral Delay and Treatment Response to Intravitreal Bevacizumab for Wet Age-Related Macular Degeneration

Abstract

Background: Recent years have seen anti-vascular endothelial growth factor (VEGF) agents revolutionize the treatment of wet age-related macular degeneration (AMD). To ensure optimal outcomes for this time-sensitive disease, current guidelines recommend that patients should be seen in clinic within two weeks of referral and treated with a loading phase of three injections, followed by a follow-up visit for reassessment. However, meeting these guidelines is challenging due to the increasing global prevalence of wet AMD and the accumulating cohort of patients requiring treatment. In addition, services with inefficient referral pathways and limited treatment capacities have been shown to threaten patient access to prompt treatment and increase adverse patient outcomes.

Aims and Methods: This project is a two-year retrospective audit conducted at the Eye Department in Dunedin Public Hospital with the aims to (1) assess the efficiency of the current wet AMD referral pathway against contemporary guidelines and (2) identify any risk factors that are predictive of referral delay and treatment outcome. Relevant patient, referrer, referral and clinical characteristics were retrieved from retrospective analysis of clinical records and OCT scans of 113 patients. The outcome measures were referral delay (i.e. duration from point of referral to first assessment clinic); and treatment response measured by changes in visual acuity, changes in central macular thickness (CMT) and clearance of macular fluid. All potential predictors of these outcome measures were analysed via multivariable binomial logistic regression.

Results: Only 49% of patients at Dunedin Public Hospital met referral guidelines, but 85% met treatment guidelines. Overall median time from point of referral to first treatment was 10 days. A loading phase of three bevacizumab injections significantly improved mean visual acuity by 5 ± 24 letters (p=0.03) and reduced mean central macular thickness (CMT) by 55μm (p<0.001). In addition, 25.7% of patients had a marked reduction of macular fluid (≥90%) after treatment. Referral method was found to be a significant predictor of referral delay. Compared to electronic or faxed referrals, mailed referrals were associated with longer referral delays (OR 4; 95% CI 1.30, 12.5; p=0.015) whereas telephone referrals were associated with shorter referral delays (OR 8.74; 95% CI 2.03, 37.60, p=0.004). Referral delay was not associated with treatment response. Instead, worse vision at the first assessment clinic was strongly associated with 2 or more lines of visual gain after treatment compared to those with better vision at baseline (OR 4.86; 95% CI 2.16, 10.95; p<0.001). Thicker baseline CMT and higher age (>85 years) were associated with an increased likelihood of CMT reduction (OR 9.157; 95% CI 2.6, 31.747; p<0.001 and OR 4.79; 95% CI 1.11, 20.7; p=0.036 respectively), whereas patients with longer duration of symptoms (1 to 3 months) were significantly associated with a decreased likelihood of CMT reduction compared to patients with symptoms of less than one month (OR 1.65; 95% CI 0.044, 0.616, p value = 0.007). Thicker baseline CMT was also found to be significantly associated with a greater reduction of macular fluid (OR 1.475; 95% CI 0.75, 4.578; p=0.017) after treatment, and so were females compared to males (OR 3.9; 95% CI 1.2, 12.58; p=0.02).

Conclusion: This study identified that the current wet AMD referral pathway at Dunedin Public Hospital can and should be more efficient, and quality improvement work is warranted to improve compliance to contemporary guidelines.