Recurrence free survival (RFS) was programmatically determined based on the disease recurrence data provided by the IRC and was defined as the time between the date of randomization and the date of first recurrence or death (whatever the cause), whichever occurred first. A participant who died without reported recurrence was considered to have recurrence on the date of death. For those participants who remained alive and recurrence-free, RFS was censored on the date of last evaluable post-randomization tumor assessment. For those who remained alive and had no recorded post-randomization tumor assessment, RFS was censored on the day of randomization. Participants with disease at baseline were considered to have an event on the day of randomization. Computerized tomography (CT) and magnetic resonance imaging (MRI) were mandatory to establish recurrence. The primary analysis was event-driven and planned when at least 512 RFS events assessed per IRC were collected.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

Recurrence was defined as appearance of one or more new melanoma lesions: local, regional or distant metastasis. Computerized tomography (CT) and magnetic resonance imaging (MRI) were mandatory to establish recurrence. A participant who died without reported recurrence was considered to have recurred on the date of death. Disease was assessed at randomization and every 12 weeks (±2 weeks) for 3 years, then every 24 weeks until documented distant progression.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

Yearly recurrence-free survival rates, eg. at 1 year, defined as the probability that a participant was recurrence-free at 1 year following randomization, were estimated for each treatment group using the Kaplan-Meier product-limit method, along with their corresponding log-log transformed 95% confidence intervals. RFS was defined as the time between the date of randomization and the date of first recurrence or death (whatever the cause), whichever occurred first. A participant who died without reported recurrence was considered to have recurrence on the date of death. For those who remained alive and recurrence-free, RFS was censored on the date of last evaluable post-randomization tumor assessment. For those who remained alive and had no recorded post-randomization tumor assessment, RFS was censored on the day of randomization. Participants with disease at baseline were considered to have an event on the day of randomization. CT and MRI were mandatory to establish recurrence.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

Distant Metastasis-Free Survival (DMFS) was programmatically determined based on the first date of distant metastasis provided by the IRC and was defined as the time between the date of randomization and the date of first distant metastasis or death (whatever the cause), whichever occurred first. A participant who died without reported disease distant metastasis was considered to have distant metastasis on the date of death. For those who remained alive and metastasis-free, DMFS was censored on the date of last evaluable post-randomization tumor assessment. For those who remained alive and had no recorded post-randomization tumor assessment, DMFS was censored on the day of randomization. Participants with disease at baseline were considered to have an event on the day of randomization. Disease was assessed at baseline (randomization) and every 12 weeks (±2 weeks) for 3 years, then every 24 weeks until documented distant progression.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

DMFS was programmatically determined based on the first date of distant metastasis provided by the IRC and was defined as the time between the date of randomization and the date of first distant metastasis or death (whatever the cause), whichever occurred first. A participant who died without reported disease distant metastasis was considered to have distant metastasis on the date of death. For those who remained alive and metastasis-free, DMFS was censored on the date of last evaluable post-randomization tumor assessment. For those who remained alive and had no recorded post-randomization tumor assessment, DMFS was censored on the day of randomization. Participants with disease at baseline were considered to have an event on the day of randomization. Disease was assessed at baseline (randomization) and every 12 weeks (2 weeks) for 3 years, then every 24 weeks until documented distant progression.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

Yearly distant metastasis-free survival rates, e.g. at 1 year, defined as the probability that a participant was alive at 1 year following randomization, were estimated via the Kaplan-Meier method. Distant Metastasis-Free Survival (DMFS) was programmatically determined based on the first date of distant metastasis provided by the IRC and was defined as the time between the date of randomization and the date of first distant metastasis or death (whatever the cause), whichever occurred first. A participant who died without reported disease distant metastasis was considered to have distant metastasis on the date of death. For those who remained alive and metastasis-free, DMFS was censored on the date of last evaluable post-randomization tumor assessment. For those who remained alive and had no recorded post-randomization tumor assessment. Participants with disease at baseline were considered to have an event on the day of randomization.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

OS was defined as the time from the date of randomization to the date of death. For those participants who had not died, OS was censored at the recorded last non-missing date of contact for which the participant was known to be alive.

Intent-to-treat population: All randomized participants, analyzed in the arm to which they were allocated by randomization

OS was defined as the time from the date of randomization to the date of death. For those participants who had not died, OS was censored at the recorded last non-missing date of contact for which the participant was known to be alive.Yearly survival rates, e.g. at 3 years, defined as the probability that a participant was alive at 3 years following randomization, were estimated via the Kaplan-Meier method

Intent to Treat Population: All randomized participants,analyzed in the arm to which they were allocated by randomization

AEs: Medical Dictionary for Regulatory Activities (MedDRA) version 16.1. irAEs=unknown etiology consistent with an immune phenomenon, considered as causally related to drug. imARs=based on investigator's assessment of immune-mediated etiology [excluding novel maintenance events (ie, patients with imARs occurring for the first time during maintenance)]. AE=any new unfavorable symptom, sign, or disease or worsening of a preexisting condition that may not have a causal relationship with treatment. SAE=a medical event that at any dose results in death, persistent or significant disability/incapacity, or drug dependency/abuse; is life-threatening, an important medical event, or a congenital anomaly/birth defect; or requires or prolongs hospitalization. Drug-related (D-R)=having certain, probable, possible, or missing relationship to study drug. Grade (Gr) 1=Mild, Gr 2=Moderate, Gr 3=Severe, Gr 4= Potentially Life-threatening or disabling, Gr 5=Death.

Safety population: All randomized participants who received at least 1 dose of study therapy

AEs: Medical Dictionary for Regulatory Activities (MedDRA) version 16.1. AE=any new unfavorable symptom, sign, or disease or worsening of a preexisting condition that may not have a causal relationship with treatment. SAE=a medical event that at any dose results in death, persistent or significant disability/incapacity, or drug dependency/abuse; is life-threatening, an important medical event, or a congenital anomaly/birth defect; or requires or prolongs hospitalization.

Safety population: All randomized participants who received at least 1 dose of study therapy

P-Y = person-years of exposure. Incidence rate per 100 person-years of exposure (IR/100 P-Y) was calculated as event count * 100 /person-years of exposure. MedDRA Version: 19. Duplicate AEs have been eliminated and overlapping and contiguous occurrences of the same event have been collapsed.

All participants who received at least one dose of ipilimumab or placebo, adjusted for person-years (P-Y) of exposure; P-Y=467.4; P-Y=781.7 for ipilimumab and placebo, respectively.

Global health status was measured using European Organization for Research and Treatment of Cancer (EORTC) Quality Life Questionnaire (QLQ) C-30. This health related quality of life (HRQoL) questionnaire was comprised of 15 questions on functional scales, 13 questions on symptom scales and 2 on global health status scale. Global Health Status used a 7 point Likert-type scale of 1 (Very poor) to 7 (Excellent). All scales linearly transformed to 0-100 scales. Higher scores for Global Health Status indicate better HRQoL. An increase from baseline indicates improvement in HRQoL compared to baseline. HRQoL was administered within 1 week prior to first dose (baseline) and on Days 22, 43, 64 (+/- 3 days), Week 24 and every 12 weeks up to 2 years, independent of disease progression.

All randomized participants (ITT) analyzed in the arm to which they were allocated by randomization were analyzed. At timepoint level, all randomized participants (ITT) with a measurement at the timepoint were analyzed.