Twenty-four-month real-world visual outcomes of intravitreal aflibercept as monotherapy for the treatment of neovascular age-related macular degeneration.

BACKGROUND: Aflibercept is widely used as a treatment for neovascular age-related macular degeneration (nAMD). Nevertheless, there is no consensus in the optimal injection frequency in the 2nd year of treatment along with little real-world data on visual outcomes. On that basis, the primary aim of this study was to assess the visual acuity (VA) and the total number of injections needed on average for these patients during the 24-month follow-up.
MATERIALS AND METHODS: This is a retrospective observational study from an electronic medical record of consecutive patients treated with intravitreal aflibercept (both naïve and nonnaïve eyes) who had completed the 24-month follow-up since the commencement of treatment. Patients followed the VIEW protocol in year 1 whereas in year 2, an as required approach/Pro Re Nata (PRN) was used.
RESULTS: Eighty-seven eyes of 78 patients were analyzed. 43.7% were nonnaive eyes. Baseline VA for all eyes (logMAR) was 52.6 letters, improving to 56.2 letters at 12 months and 55 at 24 months. Almost 83.9% of the treated eyes (81.3% of the patients) did not experience any significant visual loss receiving on average of 9.9 injections in the 24 months of follow-up and attending the hospital eye service 20.3 times in total.
CONCLUSIONS: Aflibercept as monotherapy for the treatment of nAMD is associated with good 2 nd year outcomes in a real-world setting using the PRN approach in year 2 and fewer injections comparing to the clinical studies, but a higher proportion of follow-up visits compared to the treat and extend regimen.

Introduction

Age-related macular degeneration (AMD) is the most frequent cause of irreversible blindness in high-income countries of the developed world.[1] AMD is categorized into early and late AMD which includes macular neovascularization or atrophy.[2] In the UK alone, the prevalence of neovascular or “wet” AMD (nAMD) is estimated to be 263,000 cases.[3]

Antivascular endothelial growth factor (VEGF) therapy has become the main treatment option of nAMD.[456] Aflibercept got approval by the National Institute for Health and Care Excellence in July 2013, for the treatment of nAMD.[6] When compared to intravitreal (IVT) ranibizumab, aflibercept dosed every 2 months after 3 initial monthly doses produced similar efficacy and safety outcomes as monthly ranibizumab in the 1st year of treatment.[7]

Due to the expected increase in the number of patients with AMD, the associated clinical workload and costs required are substantial.[89] Hence, it is mandatory to find the optimum approach beyond the 1st year of treatment.

To the best of our knowledge, there is little real-world data on the visual outcomes of patients treated with IVT aflibercept for nAMD using a dedicated pro re nata (PRN) approach in the 2nd year of treatment. At the same time, there is no similar data from the Scottish population.

Materials and Methods

This is a retrospective observational study of patients treated with IVT aflibercept (both naïve and nonnaïve eyes) who had completed the 24-month follow-up since the commencement of aflibercept treatment.

The initial dataset included all consecutive patients currently on aflibercept as monotherapy for nAMD, who attended the return AMD clinics in our unit between February 2, 2017, and May 20, 2017. Eyes which were previously on IVT ranibizumab for the treatment of nAMD (nonnaïve eyes) before being switched to aflibercept were included as well. The decision to switch to aflibercept was decided by the lead consultant and was based on the development of tachyphylaxis to ranibizumab. This manifested clinically with the reaccumulation or no improvement in the subretinal or intraretinal fluid in spectral domain optical coherence tomography (SD-OCT) with or without a drop in the visual acuity (VA) of 5 letters or more. Patients who had completed <24 months of follow-up since the commencement of aflibercept were excluded from the study.

All patients had a fundus fluorescein angiogram (FFA) on presentation which confirmed the presence of a choroidal neovascularization; unless there was a contraindication for the FFA. On subsequent follow-ups, their progress was monitored using SD-OCT.

In our hospital, the VIEW year 1 protocol is currently used for the 1st year of treatment with 2 mg of IVT aflibercept being administered every 2 months after 3 initial monthly doses (2q8).[7] Hence, for this article, we have considered the follow-up after seven IVTs injections as being equivalent to 12 months of follow-up.

As far as the 2nd year of treatment is concerned, this subtype of patients was treated with the PRN/as required approach and when possible using a one-stop service. Patients who are deemed stable by the consultant in charge (i.e., stable or improved VA since the last injection together with no activity in the macula SD-OCT) are booked in the virtual macula clinic. The decision on whether to treat or not is based on their VA (drop in the VA of five letters or more in the logMAR chart), macular SD-OCT findings (the presence of intra- or sub-retinal fluid in the macular SD OCT), and/or the presence of new hemorrhages in the macula.

Data were collected using the hospital's electronic records (Clinical Portal, Orion Health, New Zealand). Data collection included age, gender, previous ranibizumab IVTs, and number of ranibizumab injections where applicable, date of 1st aflibercept injection, total number of injections at 12, 18, and 24 months, total number of appointments at 12 and 24 months, laterality, VA on presentation (using the patient's own spectacle correction where applicable), VA following 3 monthly IVTs, 7X IVTs, along with VA at 18 and 24 months since the commencement of treatment with aflibercept. The number and proportion of eyes that lost or gained no letters or 5, 10, and 15 logMAR letters were also calculated for the above intervals. Finally, the proportion of eyes that were started on aflibercept and later switched to ranibizumab due to tachyphylaxis was calculated as well.

Data collection was anonymized. This type of analysis does not require Ethical Committee Approval as it is viewed as an audit.[10]

T-test was used to compare the means of independent values. P ≤ 0.05 was considered statistically significant.

Results

Participants

Eighty-seven eyes of 78 patients were identified. The mean age was 81.3 years (standard deviation [SD] 6.7). Nearly 53.4% were female. The minimum age was 64 years and maximum was 98 years. Totally, 19 patients received bilateral treatment.

Thiry-eight eyes previously on ranibizumab were identified (43.7% of the cohort). The mean number of ranibizumab IVTs before switching to aflibercept was 10.2 (SD 4.9). None of the naive patients were switched to ranibizumab at any point during the follow-up. There were no cases of endophthalmitis following the administration of IVT aflibercept in this cohort of patients.

Visual acuity

Mean VA at the various time intervals is presented in Table 1 and Figures 1, 2.

Table 1

Mean visual acuity at specified time intervals and mean difference comparing to baseline (letters at logarithmic minimum angle of resolution 2 m)

Time	Mean visual acuity (SD)	Mean difference comparing to baseline
Baseline	52.6 (17.6)	-
3 IVTs	56.6 (16.7)	+4
7 IVTs (12 months)	56.2 (16.9)	+3.6
18 months	55 (16.33)	+2.4
24 months	55 (17)	+2.4

IVTs: Intravitreal, SD: Standard deviation

Figure 1

Change in mean visual acuity over time for naïve and nonnaïve eyes

Figure 2

Change in mean visual acuity for all eyes

When we compared naïve and nonnaïve eyes, there was no statistically significant difference in the baseline VA or VAs of the subsequent follow-ups [Table 2]. Nevertheless, the naïve eyes tend to perform better at all-time intervals, especially in month 18 and month 24, gaining almost one more line in these time points.

Table 2

Comparing Naïve with Nonnaïve Eyes (letters at logarithmic minimum angle of resolution 2 m)

Time interval	Nonnaïve (n=39) (SD)	Difference comparing to baseline	Naïve (n=48) (SD)	Difference comparing to baseline	P (t-test)
Baseline	52.7 (18.2)	-	52.5 (17.1)	-	0.940929
3 IVTs	55.7 (16.45)	+3	57.2 (15.9)	+4.7	0.673956
7 IVTs	53.9 (17.7)	+1.2	58.3 (15.94)	+5.8	0.233825
18 months	51.8 (16.7)	-0.4	57.5 (15.5)	+5.0	0.532757
24 months	52.9 (18.8)	+0.3	56.7 (15.08)	+4.1	0.683295

IVTs: Intravitreal, SD: Standard deviation

The numbers and proportions of eyes which lost or gained any number of letters in the 24 are presented in Table 3, with a detailed comparison between naïve and nonnaïve eyes. 83.9% of the treated eyes (81.3% of the patients) did not experience significant visual loss (defined as the loss of 15 letters or more).

Table 3

Comparison of naïve and nonnaïve eyes with respect to the number of letters gained or lost (logarithmic minimum angle of resolution 2 m)

24 months	Number of eyes (%)
	Nonnaïve eyes (n=39)	Naïve eyes (n=48)
Gained >5 letters	12 (30.8)	23 (47.9)
Gained >10 letters	8 (20.5)	20 (41.7)
Gained >15 letters	4 (10.3)	17 (35.4)
Lost >5 letters	13 (33.3)	15 (31.3)
Lost >10 letters	8 (20.5)	10 (20.8)
Lost >15 letters	7 (17.9)	7 (14.6)

Number of injections and appointments

The average number of injections in the 18- and 24-month follow-up was 8.8 injections (SD: 1.7) and 9.9 injections, respectively (SD: 2.2). This translates into 2.9 injections in the 2nd year of treatment using this PRN approach. The distribution of the number of injections in the 24 months of follow-up is presented in Figure 3. In terms of the total number of appointments for patients on unilateral treatment (n = 55 eyes), patients attended the clinic on average 20.3 times in 24 months (SD: 2.2), including the injection appointments. Taking into account the total number of injections in 24 months for these eyes (10 for n = 55), patients attended 10.3 times for nurse-led or consultant review to decide whether further treatment was indicated (for the PRN approach used). The “Did Not Attend” or “DNA” rate was 1.47% (of the total number of appointments). Naïve eyes had an average of 9.4 injections in the 24 months of follow-up and nonnaïve ones 10.6 (P = 0.015). The total number of appointments in the naïve group was 19.6 and 21 in the nonnaïve eyes (P = 0.041).

Figure 3

Distribution of the total number of injections in 24 months of follow-up for all eyes (n = 87)

Discussion

The above results represent real-world experience from a single-center in Scotland. Contrary to other studies, we have included nonnaïve eyes that are eyes which had been previously on ranibizumab but were switched to aflibercept because of inadequate response to ranibizumab. Our results indicated that this was done after an average of 10.2 IVT ranibizumab injections. In our opinion, this element of our study reflects more accurately the routine clinical practice comparing to clinical studies which included only treatment naïve eyes since it is known that repeated IVT injections of anti-VEGF in exudative AMD seemed to be associated with tachyphylaxis, and clinicians often switch treatment to another anti-VEGF agent.[11]

With respect to the baseline characteristics of our cohort, these are comparable to the VIEW studies with respect to the VA, but the mean age and the proportion of female patients are higher in our cohort.[7]

As far as patients' visual outcomes are concerned, our results demonstrate that IVT aflibercept for the use of nAMD appears to continue to have beneficial effects in terms of preservation of vision in the 2nd year of treatment. The mean gain in the VA from baseline is bigger after the initial loading dose of 3 monthly IVTs (+4 letters), but even though this gain drops in the subsequent follow-ups, there is a sustained improvement throughout follow-up. In addition, our results do demonstrate that there was a trend for the naïve eyes to perform better compared to the nonnaïve ones in terms of gaining clinically significant improvements in their VA at the various time intervals, although this is not statistically significant. Overall, in our study, 83.9% of the patients (or 81.3% of the cohort) of the eyes did not experience significant visual loss as defined by the loss of 15 letters in the logMAR chart.

Comparing our 12 months results to the VIEW study, the change in mean VA in our cohort is not as impressive as in that study. Almost one line less was gained in our cohort, but this is in keeping with the fact that real-world data are consistently less favorable compared to clinical trial results.[1213] One of the reasons for these differences is related to the study populations, as clinical studies have strict entry criteria and less heterogeneity.[1314] At the same time, it is not always possible to see or treat patients on a monthly basis due to capacity issues or patients' inability to attend a follow-up at a specified date. Furthermore, retrospective analyses such as this one rely on clinical notes and full data are not always available. This is a limitation of this study.

When we compared our results to a recently published large retrospective study in England on the subject, there was a gain of 2.4 letters in our study (vs. 2.8 letters in that study) in the 2nd year of treatment of nAMD patients on aflibercept monotherapy.[14] In that study, there was a variable approach in the 2nd year of treatment. There is also a study by Barthelmes et al. which reports 24-month outcomes of a treat and extend (T and E) regimen using aflibercept in eyes with nAMD.[15] In that study, the mean gain for the 123 eyes which were analyzed was 6 letters. Nevertheless, the dropout rate was 28% of the studied eyes though and an additional 8% of the studied eyes were switched to ranibizumab, so there could be an element of bias in favor of the treated group as suggested by the authors.

As far as the number of injections was concerned, on average, our patients received 8.8 injections in the 18 months of follow-up and 9.9 at the end of year 2. About 58.3% of the patients received between 7 and 10 injections and in total more than two-thirds of the cohort received <12 injections in the 24 months of follow-up. Comparing our results to the UK aflibercept users group results, their average number of injections in year 2 was 3.7, meaning that our cohort received 0.8 injections less in year 2.[14] Barthelmes et al., in the aforementioned study of IVT aflibercept for nAMD using a T and E regimen, reported an average of 13.6 injections in 24 months and 15.1 appointments in total.[15] This compares with more appointments in total in our study using the PRN approach (20.3 vs. 15.1) but less IVTs (13.6 vs. 9.9). In terms of safety, there were no cases of endophthalmitis reported in our cohort during the study period.

The main limitations of this study are that it is a retrospective, single-center audit. In addition, there are a relatively small number of patients. As far as the strengths are concerned, we present real-world data without strict inclusion criteria and baseline characteristics of patients similar to other large UK-based audits. In addition, there are no recent data from the Scottish population on the subject.

Conclusions

IVT aflibercept as monotherapy for the treatment of nAMD is associated with good 2nd year outcomes both in naïve and nonnaïve eyes. 9.9 injections were needed on average in total in the 24 months of follow-up using a PRN approach in the 2nd year of treatment, with 2.9 injections in year 2 after 7 injections in year 1. Over 80% of the patients did not experience significant visual loss. Naïve eyes received 1.2 injections less in the 2 years of follow-up and attended the clinic 1.4 times less, and these results are statistically significant.

More data are needed on the subject with larger number of eyes to establish the ideal approach in the 2nd year of treatment, taking into consideration the effects of more proactive approaches such as treat and extent and less proactive approaches such as the PRN regime in the patients' quality of life as some patients are less keen on more frequent injections.

Financial support and sponsorship

Nil.

Conflicts of interest

RP and ME have none. TS has received a travel bursary from Bayer. MG has received travel bursaries and lecture fees from Bayer, Novartis, and Allergan. Bayer was not involved in data collection, analysis, or writing of this paper and has not provided support of any kind.