COVID-19 Pandemic Lockdowns' Impact on Visual Acuity of Diabetic Macular Edema: A Large Cohort

Abstract

Introduction

The objective of this study was to evaluate the impact of unplanned treatment gap, secondary to COVID-19 pandemic lockdowns, on visual acuity in previously treated diabetic macular edema (DME) patients.

Methods

A multicenter, retrospective study of DME patients, previously treated with anti-VEGF injections, who were followed up during COVID-19 pandemic (2020) compared to pre-CO­VID-19 period (2019).

Results

A total of 634 DME patients with a mean age of 68.4 years met the inclusion criteria, 385 were assessed in 2019 (pre-COVID-19) and 239 patients assessed in 2020 (COVID-19). Baseline best corrected visual acuity (BCVA) among patients in 2019 and 2020 was 0.52 ± 0.44, 0.45 ± 0.43 (logarithm of the minimal angle of resolution, respectively). There was no significant difference between the years 2020 and 2019 in baseline BCVA (p = 0.07). Mean number of anti-VEGF injections was significantly lower (5 vs. 6, p < 0.01), with a major lower ratio of injections per patient in the COVID-19 first lockdown period (March–June 2020) in the COVID-19 group. Baseline BCVA (p < 0.01) was the only significant predictor of final BCVA. Number of injections, age, gender, and the year were not found as predictors of final BCVA.

Conclusions

In a large cohort of DME patients, an unplanned delay in treatment with anti-VEGF injections for 2–3 months, due to COVID-19 pandemic lockdown, had no significance impact on visual acuity. For most patients, returning to routine treatment regimen was sufficient for maintaining BCVA.

Introduction

Diabetic retinopathy (DR) is the most common ocular complication of diabetes mellitus and is the leading cause of blindness among adults aged 40 years or older. Diabetic macular edema (DME) is the most common cause of visual impairment in patients with diabetes mellitus, affecting according to the Centers for Disease Control (CDC) more than 37.3 million people in the US 2019 population aged 18 years or older with more than 1.4 million new patients diagnosed every year [1, 2, 3]. DME can pre­sent in any stage of DR, and it is the most common cause of vision loss among DR patients with a prevalence of 3–11% of DR patients [4, 5, 6, 7].

Exact mechanism and pathogenesis of DME is complex, involves several factors, and occurs mainly due to upregulation of vascular endothelial growth factor (VEGF), thus increasing vascular permeability and disrupting blood-retinal barrier, leading to accumulation of fluid within the retinal layers [1, 8]. The innovation of anti-VEGF therapy for the treatment of DME was a major success, and over the last 2 decades, it is the gold standard of treatment. All intravitreal anti-VEGF agents, bevacizumab, ranibizumab, and aflibercept, showed similar results in large pivotal trials with a major improvement in visual acuity (VA) within the first 1–3 years after treatment [9, 10, 11, 12, 13].

Anti-VEGF treatment protocols include monthly injections, pro re nata (PRN), and treat-and-extend (TAE). Monthly treatment enables rapid VA improvement, maintaining VA gain for longer time and improving neovascularization; however, time spent and financial cost are its major drawbacks. PRN regimen offers less injections but requires frequent visits. TAE may reduce the number of both injections and visits while preserving VA as it is costume made for each patient, and treatment should be given before macular edema gets worse. It cannot be concluded if more aggressive treatment can alter long-term VA outcomes in DME [14]. Recently, it was shown that transitioning from a PRN to a TAE protocol offers most patients a significant anatomical benefit with less robust functional benefits [15].

However, unlike clinical trials, real-world data have demonstrated that a significant portion of patients in clinical practice are undertreated with anti-VEGF and have subsequently lower best corrected visual acuity (BCVA) [16]. That might be attributed to the frequency of anti-VEGF injections which can be burdensome for some patients [17].

In the year 2020, the COVID-19 pandemic has undoubtedly altered patient care worldwide. In the worst phases of COVID-19 outbreak, being March, April, and May of 2020, in order to be able to reduce the quantity and severity of illness among the population, a lockdown took place and restricted the outpatient visits. Therefore, a significant number of patients avoided maintenance visits and non-life-saving treatments from the concern of being exposed [18]. First, pilot studies have revealed the negative impact of COVID-19 pandemic on treatment of patients with DME [19, 20]. The purpose of this study was to evaluate the impact of the COVID-19 pandemic lockdown in patients with DME, specifically evaluating the change in BCVA and the number of injections given to previously treated DME patients.

Methods

This is a retrospective, multicenter, observational study of patients with DME seen by retina specialists between 1 January 2019 and 31 December 2020, at the ophthalmology departments of Tel Aviv Sourasky Medical Center (TASMC), Tel Aviv, Israel, and Meir Medical Center, Kfar Saba. In both centers, anti-VEGF treatment protocols were quite the same and mainly based on the TAE regimen. The study and its protocol adhered to the tenets of the Declaration of Helsinki and were approved by the Institutional Review Board (IRB) TASMC Helsinki Committee, approval number TLV-1039-20.

Cases were searched and found using the electronic medical records. All data were extracted using MDClone (Israel) software which enables the extraction of precise data with specific inquiries. The inclusion criterion was the presence of DME, defined as retinal thickening within one disc diameter of the center of the macula, or definite hard exudates in this region on optical coherence tomography (OCT) [21]. All included patients received previous treatment of at least six anti-VEGF injections (ranibizumab, aflibercept, or bevacizumab) and had follow-up during the 1st and 4th quarters of either 2019 or 2020. In this study, intravitreal steroid injections and implants were not included due to their relatively small number of patients.

Follow-up visits were performed for the previously treated patients for a 12-month period in 2019 or 2020, and if patients had the appropriate data for both years, they were included for both years. For each patient, if both eyes met the inclusion criteria, one eye was randomly selected [22] as the variance between two eyes is usually smaller than that between subjects. Exclusion criteria were high myopia of above 6 D, history of retinal detachment, age-related macular degeneration, central serous chorioretinopathy, macular telangiectasia, tractional and degenerative lamellar macular holes, central or branch retinal vein occlusion, central or branch retinal artery occlusion, optic neuropathy of any kind, visually significant cataract, endophthalmitis, or retinal dystrophies.

For each patient, the following data were collected including demographics, BCVA, slit-lamp examination as well as type, dates, and number of anti-VEGF injections, and clinic visits. BCVA was documented at the first (Q1) and last (Q4) quarters of each year. The number of anti-VEGF injections per patient for each month was defined as the result of sum of injections per each month divided by total number of patients, and the injection ratio was defined as number of anti-VEGF injections per patient, per month, in the study period.

BCVA was listed in Snellen fraction and was then converted into logarithm of the minimal angle of resolution (logMAR) values for statistical analysis. Although interesting, in this study, anatomical changes and result were not included. This study was conducted to assess the real-life impact of the COVID-19 lockdown on the final visual outcome of patients and whether they had any visual inferiorities; hence, it was concerned only with VA and number of injections and not with anatomical changes seen in the OCT.

Statistical Analysis

All data collected in the study were inserted into an electronic database via Microsoft Excel 2013 (Microsoft Corporation). Statistical analyses were performed using Minitab Software, version 17 (Minitab Inc., State College, PA, USA). Results are expressed as mean ± SD, median (range), or N (%). For the comparison of continuous and categorical data at final visit versus baseline, the paired T test and McNemar's test were used, respectively. For the comparison of continuous and categorical data between non-paired groups, the Student t test and the χ² test were used, respectively. A p value of less than 0.05 was considered statistically significant.

Results

A total of 624 eyes of 624 patients were treated with anti-VEGF injections throughout the study period and met the inclusion criteria. 45% of them were female and mean age was 68.4 ± 10.7 years. 385 eyes of 385 patients were assessed in 2019 and 239 eyes of 239 patients were assessed in 2020. No significant difference was found in mean age (p = 0.33) and gender (p = 0.78) between 2019 and 2020.

Visual Acuity

Table 1 describes the comparison of baseline demographic and VA characteristics of 2019 and 2020 patients. In 2019, Q1 mean BCVA was 0.52 ± 0.44 logMAR (20/66) and by Q4, the mean BCVA was 0.55 ± 0.41 logMAR (20/71) (p = 0.082). In 2020, COVID-19 period, Q1 mean BCVA was 0.45 ± 0.43 logMAR (20/56) and by Q4, the mean BCVA was 0.47 ± 0.44 logMAR (20/60) (p = 0.323). Table 2 describes the VA changes during 1 year in 2019 and 2020; no significant difference was found between the COVID-19 and pre-COVID-19 in terms of baseline BCVA (p = 0.07); however, a significant small change was found in terms of final BCVA (p = 0.02).

Table 1.

Baseline comparison of characteristics of 2019 and 2020 patients

	2019 (n = 385)	2020 (n = 239)	p value
Female gender, n (%)	173 (45.4)	106 (44.3)	0.78
Age	68.1±10.3	68.9±11.5	0.33
Average of yearly anti-VEGF injections	6±2.1	5±2.4	0.002
Baseline BCVA	0.52±0.44	0.45±0.43	0.07
Final BCVA	0.55±0.41	0.47±0.44	0.02

BCVA, best corrected visual acuity.

Table 2.

VA changes during 1 year in 2019 and 2020

	Baseline BCVA	Final BCVA	p value
2019	0.52±0.44 (20/66)	0.55±0.41 (20/71)	0.082
2020	0.45±0.43 (20/56)	0.47±0.44 (20/60)	0.323

BCVA, best corrected visual acuity.

Trends in Intravitreal Injections

Significantly lower mean amount of anti-VEGF injections was found in the COVID-19 period in comparison with year 2019, the pre-COVID-19 period (6 ± 2.1 vs. 5 ± 2.4, p < 0.01). When exploring the absence of patients out of clinic for a period of 3 months or more, a significant larger population did not arrive at the clinic nor receive anti-VEGF injections during the COVID-19 era (25.37% vs. 42.99%, p < 0.01). For all 3 anti-VEGF agents, there was no significant change between 2019 and 2020.

Figure 1 depicts the trends in the ratio of anti-VEGF injections per patient, per month, in the study period. During the months of March–June 2020, when the major lockdowns took place, there was a significant decrease in injecting per patient, comparing this period to 2019, and no other significant difference between the rest of the months was found.

Fig. 1.

Trends in the ratio of anti-VEGF injections per patient, per month.

In 2020, the ratio of injections decreased to 0.4, and in 2019, in the corresponding months, the ratio was 0.65 (p < 0.05). Figure 2 illustrates the gap with its corresponding confidence intervals between injections index of each year per month, with a corresponding significant negative higher gap in March and April.

Fig. 2.

Illustration of the gap with its corresponding confidence intervals (CIs) between injections index of each year per month.

Figure 3a, b presents OCT imaging of a 58-year-old patient, treated with ranibizumab, with worsening following an unplanned 5-month gap in injections in 2019 due to COVID-19 lockdown. Figure 3c presents significant improvement in the OCT image after the next 2 consecutive injections.

Fig. 3.

a OCT imaging of a 58-year-old patient, treated with ranibizumab, before COVID-19 lockdown. b OCT imaging of a 58-year-old patient, treated with ranibizumab, after an unplanned 5 months gap in injections due to COVID-19 lockdown. c OCT imaging of a 58-year-old patient, treated with ranibizumab, 2 months after lockdown (after completing 2 monthly injections).

Loss of Vision

Table 3 depicts the distribution of vision loss before and during the COVID-19 period. A similar distribution of vision loss was found in both groups. Although not significant, the results showed that a higher rate of patients lost 10 letters or more in 2019 rather than in the COVID-19 period (19.1% vs. 13.4%, respectively, p = 0.051).

Table 3.

VA loss in 2019 and 2020

Snellen letters lost	2019	2020	p value
0–5	279 (72.5%)	183 (76.6%)	0.251
5–10	32 (8.3%)	24 (10%)	0.472
>10	74 (19.1%)	32 (13.4%)	0.051

Multiple Regression Analysis

Table 4 depicts the outcome of the multiple regression analysis with age, Q1 BCVA, gender, period of injection (pre-COVID-19 or during COVID-19), and number of injections as potential predictors of Q4 BCVA. Q1 BCVA (0.701, p < 0.01) was the only significant predictor of Q4 BCVA, whereas number of injections, age, gender, and year were not (p = 0.233, 0.105, 0.102, 0.086, respectively). Betas represent the “impact proportion” of each; for example, the Q1 BCVA has a beta of 0.701, meaning every delta of 0.1 logMAR will add 0.07 to Q4 BCVA.

Table 4.

Final VA predictors

Regression independent variables	Beta (std)	p value
Baseline VA	0.701 (0.656, 0.746)	<0.01
Number of injections	–0.006 (–0.016, 0.004)	0.233
Age	0.002 (–0.001, 0.004)	0.105
Gender	0.04	0.102
Year (2019 vs. 2020)	–0.04	0.086

Beta, the proportion of each predictor that influences final VA; Std, standard deviation; VA, visual acuity.

Discussion

This multicenter real-world study investigated the effect of COVID-19 pandemic lockdown on prior treated DME patients with at least 1 year of previous anti-VEGF treatment and was the first large study proving no VA damage due to unplanned delay in treatment in these patients. We found that the number of annual injections was significantly lower in COVID-19 period compared to 2019, the pre-COVID-19 period with a major decline during the March–April lockdown. Although no significant difference was found in baseline BCVA between the 2 years, we found a small significant positive difference in final BCVA outcomes in COVID-19 period compared to 2019. A mean reduction of two letters (0.04 logMAR) in BCVA was presented in both 2019 and 2020. In addition, when dividing the patients into three categories based on the number of letters lost (0–5, 5–10, >10), no significant difference was found between the 2 years. These results consist of long-term studies such as the MEAD and VISION [23, 24] and can be attributed to the undertreatment in real life or to the progression of the disease despite treatment.

We found a significant decrease in the number of annual injections during the COVID-19 pandemic in 2020 compared to 2019, with a delta of yearly 1 injection between them (6 vs. 5, p < 0.01). A significant decrease in injections per patient was found in March till June period in 2020 which is aligned with COVID-19 spreading time in Israel and the first national lockdown (March–April). This reduction consists with other studies examining COVID-19 impact on treatment for DME and for other retinal diseases which require intravitreal frequent injections such as neovascular age macular degeneration [19, 20]. Surprisingly, a similar number of injections were found in the second half of 2020 compared to the same months in pre-COVID-19 pandemic, which might be attributed to national health campaigns calling citizens to continue medical routine treatments.

However, while we found no significant difference in vision loss due to COVID-19 impact in a large cohort, other studies have shown significant changes in BCVA or OCT outcomes [20, 25], possibly due to our major use of TAE treatment regimen which is more tolerable for long pauses rather than monthly regimen. For most patients, the unplanned gap was larger than 2 months, with a stable recovery since then. Similar results were reported in the pivotal trials VIVID and VISTA [9] in which patients received intravitreal aflibercept every 8 weeks, following several monthly injections, and BCVA improvement was achieved.

Treatment intervals can vary among DME patients, as previously reported in DRCR network protocols I and T [26, 27] which demonstrated good VA outcome in PRN regimen following 3–4 monthly injections. As opposed to regular PRN or TAE regimen, when routine visits are scheduled and a retina specialist decides whenever treatment is necessary, in our study, patients missed routine treatments and follow-up visits. To our knowledge, this is the first real-life study demonstrating that an unplanned delay in treatment for 2 or more months, in patients with central involved DME, would not harm BCVA outcomes in a large group of patients. For most patients, returning to routine treatment regimen is sufficient for maintaining their BCVA without suffering from a significant visual loss.

Baseline BCVA was the only significant predictor for BCVA changes among previously treated subjects in a multivariable regression analysis. Several recent studies have stated that most of the missing follow-up visits in the COVID-19 period belonged to the elderly population [28, 29]; however, we showed no significant age difference between 2020 and 2019. Better baseline BCVA is a known predictor for final BCVA, and in a lower baseline BCVA, a bigger improvement can be achieved [11, 30], and we showed among study cohorts that any previous BCVA, within the second year of treatment forward, is a significant BCVA predictor for the following year. Although the correlation between number of injections and BCVA change was studied before and a linear correlation was found between mean letters gained and mean number of annual anti-VEGF injections [31, 32], we found that the change in number of annual injections (between 6 and 5) did not have a significant impact on VA changes.

Limitations

This study has several limitations. First, the retrospective nature of the study. Second, following its retrospective nature, the scheduled follow-up visits and treatment intervals were unknown to the investigators, thus making it difficult to establish when a scheduled interval was exceeded and the exact deviation time from scheduled visits. Due to routine regimen of monthly injections or TAE used in our institutes, usually an interval gap exceeding 3 months could be considered an unplanned gap, causing underestimation of the number of patients missing their scheduled visits. Third, the relatively short unplanned gap in this study makes it difficult to establish the impact of prolonged unplanned gap on VA in our study cohort. In addition, this study included anti-VEGF treatment and not steroid injections or steroid implants.

Conclusions

In a large cohort of centrally involved DME patients, an unplanned delay in treatment with anti-VEGF injections for 2–3 months might be tolerable, with no significant impact on VA. For most patients, returning to routine treatment regimen is sufficient for maintaining BCVA.

Statement of Ethics

This study protocol was reviewed and approved by the Institutional Review Board (IRB) TASMC Helsinki Committee, approval number TLV-1039-20. According to the IRB decision, no informed consent was needed due to the retrospective nature of this study.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

No funding was received for this article.

Author Contributions

Nir Gomel and Gilad Rabina: design, data collection, analysis, interpretation of data, drafting, revision, and final approval. Reut Shor: data collection, analysis, drafting, revision, and final approval. Naama Lippin: data collection, interpretation of data, drafting, and final approval. Ori Segal: design, data collection, analysis, interpretation of data, drafting, and final approval. Eran Greenbaum: data collection, Analysis, interpretation of data, and final approval. Shulamit Schwartz: design, data collection, analysis, drafting, revision, and final approval. Omer Trivizki: design, data collection, analysis, revision, and final approval. Anat Loewenstein: design, data collection, revision, and final approval.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

Funding Statement

No funding was received for this article.