Presence of active leakage on fluorescein angiography (FAG) was assessed at screening (14 to 3 days before baseline visit), month 2 and month 6. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of presence of active leakage (Yes/No). For retreated patients, the presence/absence of active leakage was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis. *NE = Not evaluable

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent fluorescein angiography at that time point.

Change from baseline in BCVA (Best Corrected Visual Acuity) was Measured by Early Treatment Diabetic Retinopathy Study (ETDRS) Letter Score. Patients with a BCVA ETDRS letter score of 78 to 24 in the study eye were included; A higher score represents better functioning of the study eye. A positive change from baseline shows improvement.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab. The study eyes of the patients belonging to the FAS were analyzed.

Mean number of ranibizumab injection is reported as number of injections per patient.

Full Analysis Set (FAS): all subjects who received at least one dose of ranibizumab

Time to re-treatment, defined as time in months from the data of first dose of ranibizumab to the date of re-treatment, was evaluated.

Full Analysis Set (FAS): all subjects who received at least one dose of ranibizumab

Number of patients with any systemic AE, with serious systemic AE, with an ocular AE, with an ocular serious AE are reported

Safety Population: all subjects who received at least one dose of ranibizumab and had at least one post-baseline safety assessment.

Patient quality of life was assessed by Impact of Vision Impairment (IVI) questionnaire. IVI is a 32-item instrument, either self- or interviewer-administered, developed to measure the impact of vision impairment on daily activities in five domains. The 32 items were divided into 5 domains as follows: Leisure and work (items 1 to 5), Social and consumer interaction (items 6 to 10 and items 23-24), Household and personal care (items 11 to 14 and items 20-21), Mobility (items 15 to 19 and item 22), Emotional reaction to vision loss (items 25 to 32). Responses to the IVI items were rated on a five-category Likert scale: not at all, 0; hardly at all, 1; a little, 2; a fair amount, 3; a lot, 4; and can't do because of eyesight, 5. Total score was an arithmetic average of the items rated between 0 (the best score) and 5 (the worst score). A negative change indicates improvement. Data was computed on items with non missing response

Full Analysis Set (FAS): all subjects who received at least one dose of ranibizumab with data at both timepoints.

Presence of macular edema from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of presence of macular edema (Yes/No). For retreated patients, the presence/absence of macular edema was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis. *NE = Not evaluable

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point.

Presence of cysts from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of presence of cysts (Yes/No). For retreated patients, the presence/absence of cysts was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis. *NE = Not evaluable

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point.

Presence of Intra-retinal fluid from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of presence of Intra-retinal fluid (Yes/No). For retreated patients, the presence/absence of Intra-retinal fluid was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis. *NE = Not evaluable

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point.

Central subfield thickness (CSFT) from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of change in CSFT versus previous visit. For retreated patients, the change in CSFT was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point and previous visit.

Central subfield volume (CSV) from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of change in CSV versus previous visit. For retreated patients, the change in CSV was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point and the previous visit.

Presence of sub-retinal fluid from optical coherence tomography (OCT) was assessed at screening (14 to 3 days before baseline visit), month 2, month 6 and month 12. The regression model for sub-retinal fluid was not valid because "Yes" was reported in almost all subjects causing a quasi-complete separation of data points. *NE = Not evaluable

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent OCT at that time point.

Presence of clinically significant abnormalities was assessed at baseline, month 1, month 2, month 3, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of presence of clinically significant abnormalities (Yes/No). For retreated patients, the presence/absence of clinically significant abnormalities was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and underwent ocular examination during that time point.

Improvement in BCVA < 5 letters (Yes/No) was assessed at month1, month 2, month 3, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of improvement in BCVA < 5 letters (Yes/No) which was reported as Gain >= 5 letters versus Gain < 5 letters. For retreated patients, Gain >= 5 letters and Gain < 5 letters were considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab with improvement in BCVA < 5 letters from baseline at that time point.

Improvement in BCVA < 10 letters (Yes/No) was assessed at month1, month 2, month 3, month 6 and month 12. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of improvement in BCVA < 10 letters (Yes/No) which was reported as Gain >= 10 letters versus Gain < 10 letters. For retreated patients, Gain >= 10 letters and Gain < 10 letters were considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab with improvement in BCVA < 10 letters from baseline to that time point.

Changes from baseline in BCVA are described for the ETDRS parameter considering the following categories at each assessment: "no change" if the change was equal to 0 letter, "worsening" if change < 0 letter , "improvement" if change > 0 letter. A univariate logistic regression model was applied expressing the presence/absence of the first retreatment in function of change from baseline in BCVA (improved/worsened/stable) which was reported as Improved versus no change and worsened versus no change. For retreated patients, this variable was considered at the closest time-point to the first re-treatment: the last scheduled assessment immediately before the first re-treatment was considered. For treated patients, the last scheduled assessment available was considered. In case of missing value on the scheduled assessment, the value was considered as missing for this analysis.

Full Analysis Set (FAS): all patients who received at least one dose of ranibizumab and evaluable BCVA at baseline and that time point.