Dr. Nimesh Jain, J15133, Dr. Naresh Kumar Yadav, Dr. Ashwin Mohan, Dr. Priya Srinivasan, Dr. Santosh Gopi Krishna Gadde
Introduction
Age-related macular degeneration (AMD) is a multifactorial disease, predominantly affecting the central macula. It is a major ophthalmic concern across the globe and, with increased life expectancy coupled with growing risk factors particularly dietary habits, smoking, hypertension and obesity, incidence of AMD is expected to double by 2020. Development of molecules targeting VEGF such as ranibizumab, bevacizumab and pegaptinib was the major breakthrough for treating neovascular AMD. Landmark clinical trials, the ‘‘MARINA’’ and the ‘‘ANCHOR’’ were two multicenter, randomized, controlled, clinical trials which studied the efficacy of monthly intravitreal anti-VEGF injection in nAMD patients. However, monthly injection added to a heavy financial burden on the patients with chronic nAMD. Subsequently PrONTO and HARBOR trial demonstrated advantages of giving intravitreal anti-VEGF injection on pro re nata (PRN) or “as needed” basis. Current protocol of treatment practised across various western and European countries is “Treat and Extend” or “Treat and Observe” regimen. Inspite of various observations and protocols established one thing still remains unanswered as to what happens in a real world scenario where patients do not abide by the protocols and has an irregular follow up. We aimed to study the real world results of treating nAMD patients with anti-VEGF and evaluate if they are different from controlled trials.
The ARMOUR study (Anti-VEGF in neovascular age Related Macular degeneration-Our Understanding in a Real world Indian setting) is a retrospective, observational, study of our experience with anti-VEGF intravitreal injections in the treatment of nAMD over a period of 12 months.
Materials and Methods
Study design
ARMOUR study is a retrospective, observational, non-interventional, longitudinal study involving 119 eyes of 94 patients with nAMD. Inclusion criteria for this study consisted of patients with treatment naive choroidal neovascular membrane secondary to nAMD and subsequently treated with anti-VEGF with a minimum follow up of 12 months. Patients who had co-morbid ocular conditions or had taken previous treatments with anti-VEGF were not included in this study. Landmark studies recommended 3 loading doses at monthly interval and subsequently to be treated on PRN basis. Based on this we follow the protocol of treating all our patients in our hospital diagnosed with nAMD with 3 loading and then PRN. On retrospective analysis we however found that a sizeable proportion did not adhere to the treatment schedule described above. Hence in this retrospective real world study we divided the patients into two groups. Group I included patients who followed the protocol and were given loading dose of 3 injections and then subsequently treated on PRN basis and Group II included patients who received anti-VEGF on pro re nata basis (without the loading dose). On each visit patients underwent a detailed ophthalmic examination and were imaged by SD-OCT. Data was collected at baseline and at Month 1, post loading visit (arbitrary around 3 months), and annually (final visit at around 12 months).
Outcome measures
The primary outcome measures of this study was mean change in corrected distant visual acuity (CDVA) and central subfoveal thickness (CSFT) from baseline to months 3 and 12. Secondary outcome measures were average number of injections used in individual group, differences in response to various anti-VEGF molecules and dosing schedule and factors affecting visual outcome at end of 12 months. Results from both the groups were subsequently compared.
Results
A total of 119 eyes were studied with an average age of 72.3 (95% CI 70.8 – 73.8) years. The male female ratio was 73:46. The mean presenting visual acuity (CDVA) was 45.77 ± 26.54 ETDRS letters (95% CI 40.96- 50.59) and the average initial central foveal thickness (CSFT) was 385.9 µm (95% CI 361.4 – 410.4 µm). The average follow up was at 1, 3 and 12 months. At three months follow up post injection, the average number of letters gained were around 9 ETDRS letters (95% CI 4.91 to 12.87) and the average decrease in CSFT from baseline was 57.2 µm (95% CI 28.7 – 85.7 µm). At twelve months the average number of letters gained was 9.64 ETDRS letters (95% CI 5.39 to 13.89) and the average decrease in CSFT from baseline was 43.7 µm (15.9 – 71.4 µm).
Differences in outcomes based on dosing schedule
The average number of injections given in Group I was 5 (95% CI 4.5 – 5.5) and in Group II was 3.7 (95% CI 3.4 – 4.1) (p<0.001). The decrease in CFT at the post loading visit was more in Group 1 (90.9µm; 95% CI 51.9 – 129.8µm) than in Group 2 (30.2µm; 95% CI -10.4 – 70.7µm)(p=0.033). In our study we observed 36.97% (44 of 119 eyes) patients gained 15 or more letters at 12 months follow up. Sub-group analysis showed 43.39%(23 of 53 eyes) and 31.82% (21 of 66 eyes) in group I and II respectively who gained 15 or more than 15 ETDRS letters at 12 months follow up (p=0.193).
Differences in outcomes based on molecule
There were no differences in the pre-injection variables. The number of lines gained differed significantly between Bevacizumab and Ranibizumab. At 3 months the average number of letters gained in the Bevacizumab group was 5.58 (95% CI 1.13 to 10.04) and in the Ranibizumab group was 18.70 ETDRS letters (95% CI 10.66 to 26.74)(p =0.002).
Factors influencing number of lines gained (multivariate analysis)
At 3 months follow up the baseline visual acuity (beta = 1.57; p<0.001), change in CSFT (beta = 0.003; p=0.002) and the choice of drug (beta 1.04; p=0.008) significantly influenced the number of lines gained. The dosing schedule did not have an influence on the number of letters gained. At 12 months follow up the number of letters gained at 3 months (beta = 0.63; p<0.001) and the CSFT at 12 months (beta = 0.002; p=0.037) significantly influenced the number of letters gained. The choice of drug did not independently influence the number of letters gained at 12 months in the multivariate analysis. It must be remembered though that a significant positive correlation was seen (r = 0.188; p = 0.04) and that the drug plays a significant role in the number of letters gained at 3 months. The dosing schedule again did not have any influence.
Other significant correlations
Choice of drug molecule ranibizumab significantly improved visual outcome at 3 months (r = 0.252; p = 0.006) and at 12 months (r = 0.366; p<0.001).
Discussion
In our study only treatment naive patients were included with no other ocular co-morbid conditions, while in other real world observations inclusion criteria was not clearly defined as in whether or not only treatment naive patients were included and were patients with other co-morbid ocular conditions excluded or included..
In this retrospective study we observed an overall 9.9 ETDRS letter gain (at 12 months) which is comparable with vision gain reported by various pivotal clinical trials like ANCHOR, MARINA, PrONTO and HARBOR (11.3, 7.2, 9.3 and 8.2 letters respectively).However, other real world observations like LUMIERE study and TWIN study have observed gain of only +3.2 and +4.3 ETDRS letters at 1 year respectively, which is less as compared to our observation.
As we noticed in our study population group many patients (66 eyes, 55.46%) didn’t follow the protocol, so it becomes necessary to evaluate vision gain in both the arms (Loading and PRN) to understand the outcomes in them. We observed 13.3 ± 24.77and 6.6970 ± 22.04ETDRS letters gained at 12 months follow up in group I (loading) and II (PRN) respectively which are in line with vision gain in landmark clinical trials and better than other real world observation mentioned earlier. Inspite of better mean letters gained in loading group, inter group variability at 12 months showed no statistical significance (p=0.138).
In our study we observed 36.97% (44 of 119 eyes) patients gained 15 or more letters at 12 months follow up. Sub-group analysis showed 43.39%(23 of 53 eyes) and 31.82% (21 of 66 eyes) in group I and II respectively who gained 15 or more than 15 ETDRS letters at 12 months follow up. This gain was comparable in magnitude between both the groups with no statistical significance (p=0.193). Overall we observed 77.31 % (92 of 119 eyes) patients who either maintained (no change from baseline visual acuity at 12 months) or improved visual acuity at 12 months follow up.
Average number of injection required in group I were 5 while fewer injections (3.7 injections) were required in group II over 12 months (P<0.001). Even with lesser number of injections in PRN group, vision gain was similar in both the groups. Eventhough the study group is small, with such encouraging results this observational study can be considered as a reference for future randomized controlled trials.
Factors affecting visual gain have never been dealt with in various randomized clinical trials. Rationale for using loading dose for 3 months has not been understood well. In our clinical observation CDVA at 12 months is influenced by CSFT (beta= 0.195, p=0.003) at 12 months and CDVA (beta = 0.674; p=0.001) at 3 months.
MARINA and ANCHOR trials observed significant gain in visual acuity with monthly anti-VEGF injections for nAMD. Results from these trials set the standard for treatment outcome and led to monthly ranibizumab treatment being approved for patients with nAMD in the U.S. in 2006.Overburden with the number of injections and visits, few trials like PrONTO and HARBOR were designed to study the efficacy of PRN dosing schedule, which showed equal success with “as needed” dosing schedule. As patients were treated after signs of clinical activity on OCT in PrONTO trial “treat and extend” regimen was designed in European countries. In “Treat and Extend” regimen anti-VEGF injections are given till the signs of exudation are resolved and then the follow up interval is individualized. If no sign of activity was seen then the duration of follow up was increased, but at every follow up anti-VEGF was given. Rationale of treating in absence of signs of clinical activity was unacceptable in current Indian clinical practice. From time to time there has always been a discrepancy and approach to treat nAMD patients. Inspite of extensive research and trials in this field there is always a dilemma about patients who do not abide by the protocol in the real world situation. Also, ethnicity has a definite role to play in disease progression but does it really affect the treatment response has never been studied and answered. In this retrospective study, however we observed no difference in vision gain between the loading and PRN groups complementing the current evidence of treating nAMD which considers no superiority of loading schedule to PRN schedule
This study shows that the visual acuity gained is comparable with previous studies and is not affected by the type of treatment schedule. Pro re nata treatment with significantly lesser injections achieved similar anatomical and functional outcomes when compared to the loading dose group. Most important parameter affecting vision gain at 12 months is vision gain at 3 months. This real world observation shows that an individualized PRN approach produces favourable patient outcomes in this real-life setting.