Real World Data About the Treatment of Wet Age-Related Macular Degeneration in Kuopio University Hospital During 2019–2021 Using a PRN Regimen

Introduction

Age-related macular degeneration (AMD) is the leading cause of blindness in developed countries. The burden of the disease is vast affecting hundreds of millions of people globally.¹ The disease is divided into two forms: dry or non-exudative and wet or exudative AMD. The dry form progresses slowly as the years go by, leading to macular atrophy and loss of vision. In the wet form, new leaking vessels grow into the macula leading to accumulation of fluid and blood, resulting in a rapid loss of vision. The dry form accounts for approximately 80–90% of all the AMD cases while the wet form accounts for the remainder.²

Wet AMD has been successfully treated with intravitreal vascular endothelial growth factor inhibitors (anti-VEGF) for over two decades.³ As there is an exponentially increasing demand for anti-VEGF therapies and clinics all around the world are under increasing pressure to meet this growing need with limited personnel and financial resources. The pharmaceutical industry has provided us with newer, potentially more efficient drugs^4,5 and clinics have come up with different treatment regimens, mainly treat and extend (TAE) and pro re nata (PRN; literally “in the circumstances”). In TAE, the treatment interval is gradually lengthened to find the longest interval for the individual patient. In PRN, patients are treated periodically when there is a need for it, ie, when fluid begins to accumulate in the macula. Both treatment options provide unique pros and cons. In TAE, the seesaw effect, ie, fluid accumulating and drying in the macula, is eliminated, but usually more injections annually are needed. In PRN fewer injections are needed but one must accept the seesaw effect of fluid in the macula.

Some studies have accepted a certain amount of fluid in the macula and still lengthened their TAE intervals.⁴ If accepting some amount of fluid without compromising the end results is found to be tolerated in studies, this would be a tool to enhance our cost effectiveness of AMD treatment. The PRN regimen was previously used in Kuopio University Hospital. Since, we have shifted to TAE regimen. The aim of this study was to retrospectively investigate our treatment results with PRN regimen during 2019–2021. This period was considered to be exceptionally interesting, because it includes a year before and the two first years of the COVID-19 pandemic. Especially in the first year of the pandemic injection intervals were lengthened by our clinic, patients also moved their appointments because in fear of COVID thus resulting in fluid accumulation, at least to some extent. This provided us with an excellent opportunity to investigate the effectiveness of PRN regimen by studying the relation of injections given per year, the usage of bevacizumab vs aflibercept (2 mg), central retinal thickness and visual acuity.

Materials and Methods

This study was designed as a real-world retrospective cohort study. Patient data was gathered from the electronic medical records (EMR) from Kuopio University Hospital. The study cohort consisted of eyes receiving intravitreal anti-VEGF treatment for wet AMD in routine clinical practice during the predefined study period. No additional formal exclusion criteria were applied beyond this cohort definition. The included eyes were not treatment-naive at study entry but had received anti-VEGF treatment before 2019. In addition, 36 eyes without wet AMD were included as an internal control group for comparison of CRT values only and were not included in analyses of injection frequency or visual acuity outcomes. The overall cohort included eyes treated with bevacizumab only, aflibercept only, or both agents during follow-up, whereas separate subgroup analyses were performed for eyes receiving only bevacizumab or only aflibercept during 2019–2021. The number of injections per year as well as the usage of bevacizumab and aflibercept (2 mg) alone were identified from the year of the diagnosis of wet AMD as well as 2019–2021 annually. Central retinal thickness’ (CRT) in µm from the year of the diagnosis and 2019–2021 were measured from OCT (optical coherence tomography) images (Heidelberg Engineering GmbH, Germany). In addition to CRT, OCT images were reviewed for signs of wet AMD activity, including subretinal fluid, intraretinal fluid, and pigment epithelial detachment (PED). These findings were used to assess disease activity and to guide treatment decisions within the PRN regimen. From 2019 to 2021, the first and last CRT of the year were measured and pooled together to incorporate fluctuation of fluid in the macula. Visual acuity (VA) data from the time of diagnosis, early 2019, late 2021, as well as the latest available VA data from the EMRs were acquired. Mean (average) (MA), standard deviation (SD) and median (Mdn) were calculated. Statistical analysis was done using an independent samples t-test (Microsoft Corporation, The United States of America).

Results

Total amount of 107 patients and 162 eyes from 1.1.2019 to 31.3.2019 receiving anti-VEGF-injections for wet AMD were selected for this study. In addition, 36 eyes without wet AMD were included as an internal control group for comparison of CRT values only. These eyes were not included in analyses of injection frequency or visual acuity outcomes. Basic demographics are presented in Table 1.

Table 1 Baseline Characteristics in 2019

The average amount of total injections in the first year of diagnosis was 6.79±2.32 injections. Patients had a history of 3.01 ± 2.55 years (MA, SD) of treatment before 2019. During 2019–2021 5.09±2.33, 5.4±2.1 and 4.96±2.11 injections were given annually (MA) in the overall cohort (Figure 1). The overall cohort included eyes treated with bevacizumab only, aflibercept only, or both agents during follow-up. A subgroup analysis of patients receiving only bevacizumab or only aflibercept (2 mg) during 2019–2021 was conducted. We found out that in the bevacizumab group patients received 5.45±2.2, 5.16±2.16 and 5±2.31 injections respectively during 2019–2021. On the other hand, patients receiving only aflibercept (2 mg) had 4.1±1.92, 3.88±1.36 and 4.1±1.7 injections per year respectively in this same period. This observed difference is statistically significant (p<0.01 for 2019, p<0.05 for 2020 and p<0.05 for 2021).

Figure 1 Total number of injections given between 2019 and 2021 in the overall cohort and in the bevacizumab-only and aflibercept-only subgroups. The overall cohort included eyes treated with bevacizumab only, aflibercept only, or both agents during follow-up. The observed difference between the bevacizumab-only and aflibercept-only subgroups was statistically significant (p<0.01 for 2019, p<0.05 for 2020 and p<0.05 for 2021).

The VA on the time of diagnosis was 70±57 ETDRS letters (MA, SD), 69±57 ETDRS letters in 2019 and 62±57 ETDRS letters in the late 2021 (Figure 2). The last VA from EMRs was 60±57 ETDRS letters. There is a statistically significant decline in the VA from 2019 to 2021 (p<0.0001) but not from 2021 to the latest VA (p=0.249). Interestingly, when looking at the patients receiving aflibercept only between 2019 and 2021, a drop in the VA from 68±55 ETDRS letters to 66±56 ETDRS letters was seen but this was not statistically significant (p>0.274). Patients receiving bevacizumab only had a similar VA decline to that observed in the overall cohort (69±58 ETDRS letters to 60±57 ETDRS letters).

Figure 2 Visual acuity (VA, ETDRS letters) between 2019 and 2021 in the overall cohort and in the bevacizumab-only and aflibercept-only subgroups. In the overall cohort, VA declined from 69 ETDRS letters in 2019 to 62 ETDRS letters in late 2021. The overall cohort included eyes treated with bevacizumab only, aflibercept only, or both agents during follow-up. In the subgroup analysis, patients receiving only aflibercept had a smaller decline than those receiving only bevacizumab (68 ETDRS letters to 66 ETDRS letters and 69 ETDRS letters to 60 ETDRS letters, respectively).

CRTs (MA) from the control eyes were 282.56±140.7 µm and 388.41±139.4 µm in the treated group during time of diagnosis. The first CRT in 2019 in the treated group was 329.9±129.4 µm. The pooled CRTs in the treated group from 2019 to 2021 were 320.78±114.73 µm, 317.2±116.19 µm and 314.1±127.88 µm, respectively. The observed difference during 2019–2021, in the treated group, when compared to time of diagnosis was statistically significant (p<0.0001 in all the groups). However, there were no statistical differences between the years (2019 vs 2020: p=0.709; 2019 vs 2021: p=0.511; 2020 vs 2021: p=0.76). There were also no statistically significant differences in the CRTs between the control and treated groups during 2019–2021 (2019 p=0.124; 2020 p=0.163; 2021 p=0.209).

Discussion

In this cohort, the mean number of injections during the first year of treatment was consistent with previously published studies. Wecker et al showed in their cohort that the mean number of injections in the first year (after 2013) was 75.⁶ On the other hand, In the CATT study it is shown that the mean number of injections annually using a PRN regimen is around 7, which is more than in our cohort after the first year (around 5 in the bevacizumab group and 4 in the aflibercept group).⁷ In TAE regimen, for example in the LUCAS study the mean number of injections in the first year was 8.9 for bevacizumab, and 18.2 injections in total at 24 months.⁸

The VA in our cohort was average stable from the time of diagnosis to 2019 (70 ETDRS letters vs 69 ETDRS letters, respectively). Interestingly, the VA dropped during our main investigation period (2019 onwards) from 69 ETDRS letters to 62 ETDRS letters in the overall cohort, which included eyes treated with bevacizumab only, aflibercept only, or both agents during follow-up, but stayed stable after that to the latest VA (60 ETDRS letters). Interestingly, in a subgroup analysis of patients receiving only aflibercept, a smaller decline in VA from 68 ETDRS letters in 2019 to 66 ETDRS letters in 2021 was seen but this was statistically insignificant.⁹ This is either due to normal decline of VA in neovascular AMD using a PRN regimen or more likely could be explained by the drop in annual injections during the first two years of the COVID pandemic.^9–11 This interpretation is supported by previous reports from the COVID-19 period. Yeter et al reported that treatment delay during the pandemic was associated with worsening visual acuity and increased central macular thickness in patients with nAMD and that despite subsequent improvement in OCT findings, short-term visual recovery remained limited.¹⁰ Similarly, a systematic review and meta-analysis by Im et al found that delayed anti-VEGF injections during the pandemic were associated with worsening visual acuity in nAMD, particularly when injection intervals were prolonged to 4 months or longer.¹¹ In contrast, Wickman et al reported similar 1-year visual acuity gains in pre-pandemic and pandemic aflibercept-treated treatment-naïve eyes in the Swedish Macula Register, suggesting that prioritization of injection visits may mitigate some of the adverse effects of the pandemic.¹² One can only postulate whether this decline could have been avoided using a TAE regimen, which has shown superiority in keeping the VA more stable in the later years of treatment than PRN.⁹ More recent evidence also supports this interpretation. Debourdeau et al reported in a large real-world comparative study that treat-and-extend was associated with greater 24-month visual acuity improvement than PRN, although more injections were required.¹³ Similarly, Rosenberg et al found in a systematic review and meta-analysis that T&E achieved better visual acuity outcomes than PRN at both 1 and 2 years, again at the cost of a higher injection burden.¹⁴

In PRN, as previously mentioned, one must accept a certain amount of fluid in the macula at some time points of the treatment. In our study, OCT was used primarily to assess disease activity, including subretinal fluid, intraretinal fluid, and PED, whereas CRT was used as the main quantitative anatomical outcome. In the newer models of AMD treatment, for example with faricimab,⁴ some amount of fluid has also been accepted without changing or shortening the injection interval. Looking at our data, it is worth noting that we need more studies with longer evaluation periods to see if this (acceptance of fluid with TAE protocols) is safe to the eye overall.

The economic advantage of using a PRN regimen would be that it allows fewer injections annually. However, PRN treatment still requires monitoring visits between injections and therefore does not necessarily reduce the overall use of clinical resources. Based on data gathered in this study, it is highly likely that the surplus made with fewer injections is eventually obliterated by the loss of vision and cost it brings to society. Thus, we need more studies on the economical aspect of AMD treatment.

This study highlighted that indeed using a PRN regimen yields in fewer injections even after years of use. On the other hand, we see a substantial drop in VA most likely due to fewer injections. In the future, we are planning to study our patients under TAE protocol and evaluate the differences in cost-effectiveness between these protocols.

Conclusions

The PRN regimen maintained anatomical outcomes but resulted in gradual visual decline, likely due to fewer injections, particularly during the COVID-19 period. Aflibercept required fewer injections without compromising outcomes. Although PRN may reduce the number of injections, it still requires monitoring visits and may not reduce the overall use of clinical resources. Transitioning toward a treat-and-extend protocol may enhance visual stability and cost-effectiveness in nAMD care.