Outcomes of Eyes With Diabetic Macular Edema That Are Lost to Follow-up After Anti–Vascular Endothelial Growth Factor Therapy


  • In patients with nonproliferative diabetic retinopathy with diabetic macular edema treated with anti–vascular endothelial growth factor injections, vision was worse after being lost to follow-up for >6 months but recovered within 3 months after restarting therapy and remained stable until final follow-up an average of 502 days later.

  • Vision recovered regardless of pre–loss to follow-up diabetic retinopathy severity status.

  • Mean central foveal thickness increased after loss to follow-up but quickly returned to pre–loss to follow-up levels after resuming anti–vascular endothelial growth factor therapy.

Graphical abstract


To evaluate the effect of loss to follow-up (LTFU) on outcomes in eyes with diabetic macular edema (DME) treated with anti–vascular endothelial growth factor (VEGF).


Retrospective cohort study.


Single-center study of 90 eyes of 73 patients with nonproliferative diabetic retinopathy (NPDR) and DME treated with anti-VEGF injections who were LTFU for >6 months. Main outcomes were the change in mean visual acuity (VA) and central foveal thickness at the return and final visits compared with the visit before LTFU.


The mean age was 64.5 years, the mean LTFU duration was 322 days, and the mean follow-up duration after return was 502 days. Compared with the mean VA at the visit before LTFU (0.42, Snellen ∼20/52), mean VA worsened at the return visit (0.54, Snellen ∼20/69, P = .004). No significant change in the mean VA was noted at the 3-month after return visit (0.50, Snellen ∼20/63), the 6-month after return visit (0.46, Snellen ∼20/57), the 12-month after return visit (0.42, Snellen ∼20/52), or the final follow-up (0.47, Snellen ∼20/59). When analyzed by NPDR severity before LTFU, no difference in VA was found from the visit before LTFU to the final visit. Mean central foveal thickness increased when comparing the visit before LTFU (270 μm) with the return visit (305 μm, P = .012), but no difference was found by the final visit (247 μm, P = .07).


Anti-VEGF–treated patients with DME who were LTFU for a prolonged period experienced a modest decline in VA that recovered after restarting treatment.

An estimated 422 million people have diabetes mellitus worldwide, including approximately 9.4% of the United States population. With approximately 28.5% of diabetic adults in the United States having diabetic retinopathy, diabetes is a leading cause of vision loss in working-age adults. , Diabetic macular edema (DME) is one of the leading vision threatening complications affecting about 4% to 7% of diabetic adults. ,

With the introduction of anti–vascular endothelial growth factor (VEGF) agents, clinical trials have shown significant improvements in visual outcomes compared with previous treatment modalities and anti-VEGF injections have therefore become first-line therapy in the developed world. , However, these outcomes have been contingent on the strict adherence to the treatment protocol of each clinical trial. Treatment adherence in diabetic care is a well-known issue, with one meta-analysis showing only 59% of diabetics compliant with taking their medication for >80% of treatment days in a year. In a study of 136 patients with DME over a 30-month period, nearly half were found to have a treatment break of >100 days. Another study of 2595 patients with nonproliferative diabetic retinopathy (NPDR) with DME treated with anti-VEGF injections found that 25.3% of these patients were lost to follow-up (LTFU) for more than a year immediately after a treatment. Even in randomized clinical trials, noncompliance with ≥1 missed visit or early withdrawal appears to be a significant issue with 108 of 175 eyes (61.7%) meeting this definition in diabetic retinopathy studies compared with only 43 of 165 eyes (26%) in neovascular age-related macular degeneration (nAMD) studies.

Despite the high incidence of poor adherence to anti-VEGF treatment schedules in DME, little is known about the effects on visual outcomes in these patients over time. We therefore sought to evaluate visual and anatomic outcomes of patients with DME treated with anti-VEGF who had a prolonged period of LTFU before returning.


Wills Eye Hospital Institutional Review Board approval was obtained before conducting this study. The study was performed in accordance with the Health Insurance Portability and Accountability Act of 1996 and adhered to the tenets of the Declaration of Helsinki. A comprehensive code search was used with International Classification of Diseases, Ninth and Tenth Revision coding for all grades of NPDR with DME between July 2015 and February 2020. Consecutive clinical charts were then reviewed to ensure that patients had the correct diagnosis of NPDR with DME, and met the inclusion criteria of being treated with ≥1 anti-VEGF injection before LTFU and had a history of being LTFU for >6 months. Exclusion criteria included a history of nAMD, retinal vein or artery occlusion, uveitis, or previous treatment or treatment during LTFU at a different practice. Patients with <3 months of follow-up after the first return visit from LTFU were not included in final visit analyses.

Baseline characteristics at the visit before LTFU including age, sex, hemoglobin A1c, anti-VEGF agent used, number of previous injections, and history of focal laser were collected. Ocular data, including VA, diabetic retinopathy severity, the presence of subretinal fluid and intraretinal fluid on optical coherence tomography (OCT), and central foveal thickness (CFT) on OCT, were collected at the visit before LTFU, the first return, 3 months after return, 6 months after return, 12 months after return, and final visits on record. In a limited number of visits, OCT scans were unable to be analyzed because of poor quality or lack of availability. CFT was measured from the inner border of the retina to the inner border of the hyper-reflective layer of the retinal pigment epithelium using the caliper tool in the Heidelberg Eye Explorer software (Heidelberg Eye Explorer; Heidelberg Engineering, Inc., Heidelberg, Germany) with all OCT scans aligned with the baseline scan using the automated tracker. All OCT measurements were performed by 2 masked graders (M.S., R.M.) and an average value of 2 measurements was considered for analysis. VA was recorded using the best available Snellen VA based on the better of habitual correction or pinhole testing. Snellen VA was converted to logarithm of the minimum angle of resolution for analysis.

All statistical tests were performed using SPSS software (version 24; IBM Corp., Armonk, New York, USA). Comparisons between continuous and categorical variables were made using a generalized estimating equation to account for intercorrelation between 2 eyes from the same patient. A Pearson correlation coefficient was used to determine the strength and direction of association between 2 continuous variables. P < .05 was considered statistically significant.


A total of 90 eyes from 73 patients were included in the study. Baseline characteristics are summarized in Table 1 . The mean (standard deviation [SD]) age was 64.5 (10.6) years, with 43.8% being female. The average LTFU time was 322 days (range 181-1126), and the average follow-up time from return to the final visit was 502 days (range 126-1141). The mean number of anti-VEGF injections before LTFU was 6.8 ± 5.3. Focal laser had been performed in 17 of the 90 eyes (18.8%) before LTFU.


Baseline Characteristics

Age (y), mean ± SD 64.5 ± 10.6
Race, n (%)
Caucasian 42 (57.6)
African-American 28 (38.3)
Asian 1 (1.4)
Other 2 (2.7)
Sex, n (%)
Male 41 (56)
Female 32 (44)
HbA1c (%), mean ± SD 8.2 ± 2.0
Previous anti-VEGF therapy, n (%)
Bevacizumab 11 (12.2)
Ranibizumab 21 (23.3)
Aflibercept 58 (64.5)
No. of injections, mean ± SD 6.8 ± 5.3
History of focal laser, n (%)
Yes 17 (18.8)
No 73 (81.2)
Lens status, n (%)
Phakic 51 (56.7)
Pseudophakic 39 (43.3)

HbA1C = hemoglobin A1C; SD = standard deviation; VEGF = vascular endothelial growth factor.

The grade of diabetic retinopathy at baseline, return, and final visit is described in Table 2 . The grade of NPDR before LTFU worsened in 13 eyes by the return visit—11 eyes by 1 stage and 2 eyes by 2 stages. By the final visit, 18 eyes had worsening severity of NPDR compared with the visit before LTFU—14 eyes by 1 stage and 4 eyes by 2 stages. Of the 10 eyes that progressed to proliferative diabetic retinopathy by the final visit, 3 were graded as moderate NPDR and 7 were graded as severe NPDR at the visit before LTFU. Eyes that progressed to PDR were complicated by vitreous hemorrhage in 3 eyes at the return visit, 1 eye at 2 months after return, 2 eyes between 6 to 12 months after return, and 2 eyes at >12 months after return. One eye was complicated by vitreous hemorrhage and neovascular glaucoma at the return visit. No eyes developed traction retinal detachments.

Jan 3, 2022 | Posted by in OPHTHALMOLOGY | Comments Off on Outcomes of Eyes With Diabetic Macular Edema That Are Lost to Follow-up After Anti–Vascular Endothelial Growth Factor Therapy

Full access? Get Clinical Tree

Get Clinical Tree app for offline access