The first patient was screened in October 2007, and the last patient completed the study in March 2011. A total of 2,232 patients were enrolled into the study from 451 centres throughout Germany (Fig. 1). Of these 2,232 patients, 17 were excluded from analyses due to double documentation (n = 9) or missing baseline data (n = 8), thus, the safety population was comprised of 2,215 patients. A further 503 patients were excluded from the efficacy population, most commonly due to: missing maintenance phase follow-up data (n = 238); incorrectly performed upload phase (<3 injections or incorrect intervals between injections; n = 174); incorrectly recorded data (n = 40); and no baseline VA data gathered (n = 22). The efficacy population, therefore, included 1,729 patients.
Demographic and medical background data is displayed in Table 1. As expected from an AMD patient population, the mean age was 77.8 years, with half of the patients older than 78.8 years. There was an even distribution of disease between left and right eyes, with a diagnosis of wet AMD in the right eye of 865 patients and in the left eye of 841 patients. In 42.9 % (n = 741) of cases, both eyes showed signs of AMD. The form of CNV was occult in 41.3 % (n = 714) of patients, predominantly classic in 24.4 % (n = 422), and minimally classic in 7.8 % (n = 134). No data on CNV type was available for 26.5 % (n = 458) of patients.
Change in BCVA
Mean BCVA increased during the three-month upload phase, from a LogMAR mean of 0.201 at baseline to 0.219 at Month 4 (Fig. 2a). Improvement in BCVA from baseline was maintained up to the second follow-up visit during the maintenance phase, at Month 5. Subsequent to this visit, mean BCVA declined steadily over time during the maintenance phase, from 0.233 at Month 5 to 0.192 at Month 15 (Fig. 2a). During the study (Month 0–15), improvement in BCVA was displayed by 40.5 % of patients, deterioration by 24.5 % of patients and no change by 35.0 % of patients (Fig. 2b).
Self-assessment of VA
Changes in VA as measured by the investigator were in accordance with patient self-assessment records (data not shown). At the first follow-up visit during the maintenance phase (Month 4), patients were asked if they had experienced any vision loss. At each subsequent visit (Months 5–15), patients were asked to classify their vision as ‘better’, ‘unchanged’ or ‘worse’ as compared with the previous visit. Following the three ranibizumab injections administered during the upload phase, 29.8 % of patients evaluated their VA as ‘better’ at Month 5. Thereafter, the percentage of patients who evaluated their VA as ‘better’ declined to 14.7 % at Month 15. The percentage of patients who did not perceive any change in VA during the maintenance phase increased from 57.8 % at Month 5 to 73.4 % at Month 15, compared with the percentage of patients who evaluated their vision as ‘worse’, which remained similar throughout the study (12.3 % at Month 5 versus 13.8 % at Month 15). Patients were also asked to evaluate metamorphopsia; the majority of patients classified metamorphopsia as ‘unchanged’ throughout the study (55.8 % at Month 5; 70.1 % at Month 15).
Retreatment frequency during maintenance phase
Following the three initial monthly injections during the upload phase, 840 (48.6 %) patients received additional injections during the 12-month maintenance phase (average of 3.1 injections per patient, range 1 to 11 injections). In the efficacy population (n = 1,729), an average of 1.5 additional injections were administered per patient during the maintenance phase, resulting in an average of 4.5 ranibizumab injections received per patient in total (three injections during upload phase plus 1.5 injections during maintenance phase). The percentage of patients receiving additional injections increased steadily during the maintenance phase from Month 4 (14.2 %) to Month 8 (19.1 %), then subsequently decreased each month until the end of the study (13.6 % at Month 15).
Subgroup analysis according to the number of retreatments received during the maintenance phase demonstrated that all patients who received additional injections during the maintenance phase presented a mean BCVA at Month 15 which was below that achieved after the upload phase at Month 4 (Month 15: <4 injections, n = 1,045; ≥4 injections, n = 136; >6 injections, n = 37, Fig. 3a). However, patients who received fewer than four additional injections of ranibizumab during the maintenance phase displayed less decline from baseline in BCVA over the course of the study and levels did not fall below baseline, compared with patients receiving more than four injections (Fig. 3a). Analysis of the change in BCVA from baseline to Month 15 according to the number of additional injections received by the patient during the maintenance phase (0 injections, n = 889; 1 injection, n = 186; 2 injections, n = 195; 3 injections, n = 240; 4 injections, n = 85; 5 injections, n = 68; >5 injections, n = 66) shows that patients who received one additional injection and those who did not receive any additional injections both maintained similar levels of BCVA during this time (both displayed a change from baseline to Month 15 of -0.002 LogMAR; Fig. 3b). In comparison, patients receiving two, three and five or more additional injections experienced a loss in BCVA greater than 0.03 LogMAR during the maintenance phase (Fig. 3b). Only those patients who received additional injections at four follow-up visits during the maintenance phase showed an improvement in BCVA (+0.016 versus baseline; Fig. 3b).
Change in BCVA during the study for patients receiving additional injections was also analysed according to the following responder group definitions (Fig. 3c): ‘Gain and maintain’ (BCVA increases from baseline to Month 4 and then to Month 15), ‘Gain but not maintain’ (BCVA increases from baseline to Month 15, but with a decline from Month 4 to end of study), and ‘No initial gain’ (BCVA declines from baseline to Month 4, but increases from Month 4 to Month 15). Most of the patients classified as ‘Gain and maintain’ received fewer than four additional injections during the maintenance phase (28.3 %), whereas the majority of patients classified as ‘Gain but not maintain’ and ‘No initial gain’ received more than six additional injections (15 % and 22.5 %, respectively) (Fig. 3c). During the maintenance phase, patients showed profit from therapy as documented by the increase of VA at the visit following the injection.
Feasibility evaluation by investigators and patients
Investigators and patients were requested to rate whether monthly assessment was meaningful and if the monitoring recommendations provided in the SmPC could be easily followed from the investigator’s and the patient’s perspectives (Fig. 4). Of the 288 investigators and 1,369 patients who responded, 58.0 % of the investigators and 70.3 % of the patients found implementation of the recommended monthly assessment into routine care to be fully feasible. Amongst those investigators who found monthly assessment partly feasible (36.5 %), the most common reasons given were due to: cost, logistics and patient number (n = 89); patient age and multi-morbidity (n = 18); and compliance problems (n = 16). For investigators who found monthly assessment not feasible (5.6 %), reasons given were due to: cost, logistics and patient number (n = 9), and lack of reimbursement (n = 9). The most common reasons given by patients who found monthly assessment partly feasible (26.9 %) were: too high burden (n = 82); transportation issues (n = 76); and the high time demands (n = 58). Patients who found monthly assessment not feasible (2.8 %) most commonly gave the reasons: too high burden (n = 15); transportation issues (n = 5); and the high expenditure of time (n = 5).
The safety analysis was based on data from 2,215 patients (Table 2). Investigators reported a total of 364 AEs in 123 patients during the entire observation period (AE incidence rate 5.6 %). Of those, 255 AEs in 118 patients were serious AEs (SAEs; incidence rate 5.3 %). For the majority of SAEs (n = 171; 67.1 %), any relationship to ranibizumab treatment was not suspected. In 14.5 % of cases (n = 37), a relationship to the treatment was suspected (unknown or no information for 47 SAEs). A total of 25 (1.1 %) patients discontinued the study due to an AE.