An AE is defined as any untoward medical occurrence in a participant, temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product. A serious adverse event is defined as any untoward medical occurrence that, at any dose that Results in death, life-threatening; requires hospitalization or prolongation of existing hospitalization; results in disability/incapacity; a congenital anomaly/birth defect. Drug-related AE's were considered to have a reasonable possibility of being related to treatment by the investigator. AE, SAE and drug-related AEs are summarized by TP and FP.

Intention-to-Treat (ITT) Population: all participants who gave informed consent, were randomized and received at least one dose of medication.

Blood samples were collected at Day 1, Weeks 4, 8, 12, 16, 20 and 24 to estimate the following biochemistry parameters: alanine amino transferase (ALT), aspartate amino transferase (AST), albumin (Ab), total protein (ToP), creatinine (Cr), total bilirubin (TB), calcium (Ca), bicarbonate (Bi), chloride (Cl), glucose (Glu), potassium (Pot), and sodium (Sod). Laboratory abnormalities outside the reference range (high and low values) at any time post baseline were presented. Any time post Baseline = all visits (including scheduled and unscheduled). If participant had given both high and low value at least once then participant is counted under both high and low category for this visit.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Blood samples were collected pre-infusion at Day 1, Week 4 and Week 8; and 24h and 72h post-infusion at Day 1, Week 4 and Week 8 time points and at Weeks 2, 6, 10, 12, 16, 20, 24, and 34 to estimate the following hematology parameters: basophils (Bas), percentage of basophils (% Bas), lymphocytes (Lym), percentage of Lym (% Lym), monocytes (Mon), percentage of Mon (% Mon), platelet count (PC), total neutrophils (TN), percentage of TN (% TN), white blood cell count (WBC), hematocrit (He), hemoglobin (Hg), and red blood cell count (RBC). Laboratory abnormalities outside the reference range (high and low values) at any time post baseline were presented. Any time post Baseline = all visits (including scheduled and unscheduled). If participant had given both high and low value at least once then participant is counted under both high and low category for this visit.

ITT Population. Only those participants available at specified time points are analyzed.

12-lead ECG assessments were obtained at the following time points: screening, and Weeks 4, 8 and 12.. Overall ECG findings were summarized using the worst case findings without regard to visits ie. "any time post Baseline". Change from Baseline in ECG findings were categorized as clinically significant change from Baseline; no clinically significant change from Baseline and not applicable.

ITT Population

SBP and DBP measurements were obtained at the following time points: screening, pre-infusion, 10 minutes (m), 30m, 1 hour (h), 2h post-infusion on Day 1, Week 4, Week 8; and Weeks 12, 16, 20 and 24. Screening value was considered as the Baseline value. Change from Baseline was calculated as the post-Baseline value minus the Baseline value.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Heart rate measurements were obtained at the following time points: Screening, pre-infusion, 10m, 30m, 1h, 2h post-infusion on Day 1, Week 4, Week 8; and Weeks 12, 16, 20 and 24. Screening value was considered as the Baseline value. Change from Baseline was calculated as the post-Baseline value minus the Baseline value.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Temperature measurements were obtained at the following time points: Screening, Day 1, and Weeks 4, 8, 12, 16, 20 and 24. Screening value was considered as the Baseline value. Change from Baseline was calculated as the post-Baseline value minus the Baseline value.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Blood samples for testing anti-mepolizumab antibodies were collected on Day 1, Week 4 and 8 Infusion Visit (before the IV infusion) and at Week 12, 24 and 34 Week follow-up visits. The presence of anti-human mepolizumab antibodies was assessed using an immunoelectrochemiluminescent (ECL) assay. To address transient positive results, an assessment of repeated results were made. For any visit category: results were considered as positive if it was positive at any visit during the study, and results were considered as negative if it were negative at all visits during the study. For repeat visit category: results were considered as postive if the result was positive at >1 visit, and results were considered as negative if the result was negative at all visits or was positive at only one visit.

ITT Population

A responder was defined as a participant achieving a reduction in esophageal eosinophils to <5 cells per HPF as the highest count of eosinophils per HPF in all the esophageal sites biopsied at Week 12, confirmed by biopsy at Week 12 or at an early withdrawal visit prior to Week 12. A worst case (WC) approach was considered, if a particiapant withdrew prematurely : If a particiapnt dropped out of the study without having a biopsy taken, due to lack of efficacy or an adverse event, their response was imputed as not achieved. Participants who withdrew, without a biopsy, for other reasons (e.g. lost to follow-up) were considered non-evaluable for the primary analysis. For participants who withdrew early from the study and had a biopsy, the biopsy was used to determine their response.

ITT Population-WC

Volume of distribution is defined as the theoretical volume in which the total amount of drug is uniformly distributed to produce the desired plasma concentration of a drug. Central volume of distribution is a hypothetical volume into which a drug initially distributes upon administration. Peripheral volume of distribution is the sum of all tissue spaces outside the central compartment. Steady state volume of distribution (Vss) is the apparent volume of distribution at steady-state. Blood samples were obtained at pre-infusion and 5m, 2h, 24h, 72-96h post-infusion at Day 1, Weeks 4, 8; and Weeks 2, 6 10, 12, 16, 20, 24 and 34 from each participant to estimate central (V1) and periperial (V2) and Steady State (Vss) volume of distribution of mepolizumab.

Pharmacokinetic Population: all participants who received study medication and for whom mepolizumab sample was obtained and analyzed.

Clearance is defined as the removal of drug from a volume of plasma in a given unit of time (drug loss from the body). Blood samples were obtained at pre-infusion and 5m, 2h, 24h, 72-96h post-infusion at Day 1, Weeks 4, 8; and Weeks 2, 6 10, 12, 16, 20, 24 and 34 from each participant to estimate plasma clearance of mepolizumab.

Pharmacokinetic Population

Par.and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A severity score of 0 was assigned for days on which pain in stomach was not experienced. If pain in stomach was reported, severity of pain was assessed as: 1=hurt a little, 2=hurt somewhat, 3=hurt quite a bit, and 4=hurt a whole lot. The average pain severity for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the pain severity scores for that interval (including days assigned as 0) divided by the number of days in the interval. Screening phase was considered as the Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric Analysis of Covariance (ANCOVA) models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction

ITT Population. The observed case (OC) datasets with incorrect questionnaires excluded were used for the analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A severity score of 0 was assigned for days on which pain in chest/throat was not experienced. If pain in chest/throat was reported, severity of pain was assessed as: 1=hurt a little, 2=hurt somewhat, 3=hurt quite a bit, and 4=hurt a whole lot. The average pain severity for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the pain severity scores for that interval (including days assigned as 0) divided by the number of days in the interval. Screening phase was considered as the Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction.

ITT Population. The OC datasets with incorrect questionnaires excluded were used for the analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of pain in stomach during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of pain in chest/throat during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A score of 0 was assigned for days on which the symptom regurgitation was not experienced. The days regurgitation experienced, the amount the symptom bothered the Par. was assessed as 1=not bothered at all, 2=bothered a little, 3=somewhat bothered, 4=bothered quite a bit, 5=bothered a whole lot. The average pain severity for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the pain severity scores for that interval (including days assigned as 0) divided by the number of days in the interval. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for Baseline score, treatment group, age group interactions

ITT Population. The OC datasets with incorrect questionnaires excluded were the primary analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of pain in regurgitation bothersome during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A participant vomiting any time was counted as one episode of vomiting, irrespective of how close they are to each other. The daily frequency of vomiting was calculated as the total number of times the participant vomited during the interval divided by the number of days in the interval. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction.

ITT Population. The OC datasets with incorrect questionnaires excluded were used for the analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of vomiting during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A score of 6 was assigned days the participant did not drink. The amount of difficulty with drinking was assessed as 1=no difficulty, 2=a little difficulty, 3=some difficulty, 4=quite a bit of difficulty, 5=a whole lot of difficulty. The average difficulty for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the drinking difficulty scores for that interval (including days assigned as 6) divided by the number of days in the interval. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A score of 6 was assigned the day participant did not drink. The severity of pain was assessed as: 1=didn't hurt at all, 2=hurt a little, 3=hurt somewhat, 4=hurt quite a bit, and 5=hurt a whole lot. The average difficulty and pain severity scores was calculated as the sum of the respective scores for that interval divided by the number of days in the interval. . Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction.

ITT Population. The OC datasets with incorrect questionnaires excluded were used for the analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of difficulty and pain when participant drank during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP and FP. A score of 6 was assigned for that symptom when Par. did not eat solid foods. When Par.eat solid foods, the amount of difficulty was assessed as 1=no difficulty, 2=a little difficulty, 3=some difficulty, 4=quite a bit of difficulty, 5=a whole lot of difficulty. The average pain severity for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the pain severity scores for that interval (including days assigned as 6) divided by the number of days in the interval. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interactions.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of eosinophilic esophagitis on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A score of 6 was assigned for that symptom when Par. did not eat. The severity of pain was assessed when Par. eats food as: 1=didn't hurt at all, 2=hurt a little, 3=hurt somewhat, 4=hurt quite a bit, and 5=hurt a whole lot. The average pain severity for the interval (Baseline, Weeks 9-12, Weeks 21-24) was calculated as the sum of the pain severity scores for that interval (including days assigned as 6) divided by the number of days in the interval. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age group interaction.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with the symptom of difficulty and pain when eating solid foods during each analysis interval (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Par. and/or parent/guardian recorded daily symptoms of the feeling like something is stuck in throat on a hand held personal digital assistant (electronic diary) during the Screening Phase, TP, and FP. A score of 0 was assigned for days on which the symptom of feeling of something stuck was not experienced. On days that feeling of something stuck in the throat was experienced, the amount the symptom bothered the Par. was assessed as 1=not bothered at all, 2=bothered a little, 3=somewhat bothered, 4=bothered quite a bit, 5=bothered a whole lot. Average bothersome score was calculated as the sum of the respective scores for that interval divided by the number of days. Screening phase was considered as Baseline interval. Change from Baseline was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA models with terms for the relevant Baseline score, treatment group, age group and treatment by age interaction.

ITT Population. The OC datasets with incorrect questionnaires excluded were the primary analysis. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

The percentage of days with feeling of something stuck in throat during each analysis period (Baseline, Weeks 9-12 and Weeks 21-24) was calculated as the number of days the symptom was experienced divided by the number of days in the analysis interval, and presented as a percentage (ie, the proportion X 100%). Screening phase was considered as Baseline interval. Change from Baseline for each analysis interval was calculated as the value for that interval minus the value for the Baseline interval. Analysis was performed using parametric ANCOVA model with terms for Baseline score, treatment group, age group and treatment by age group interaction. The OC datasets with incorrect questionnaires excluded were used for the analysis.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Participants who achieved a response of <5 esophageal eosinophils/HPF at Week 12 by worst case analysis, were evaluated for maintenance of response of <20 cells/HPF at Week 24. Response categories were defined as: non-responder (did not respond at Week 12 or Week 24); delayed responder (did not respond at Week 12 but responded at Week 24); relapsed (responded at Week 12 but not at Week 24); maintained (responded at Week 12 and Week 24). The following assumptions were made for worst case: if a Participant dropped out of the study due to lack of efficacy or an adverse event and had a missing response, their response was imputed as not achieved (i.e. failure). However for Participants withdrawn for other reasons (e.g. lost to follow-up) with a missing response (i.e. did not have the biopsy) the response was made as missing and not imputed.

ITT Population. Only those participants were responders at Week 12 were analyzed.

Participants underwent an esophagogastroduodenoscopy (EGD) with biopsies at Screening and at Weeks 12 and 24. Peak esophageal eosinophils were calculated as the maximum count across all esophageal biopsies at each time point. Screening value was considered as the Baseline value. Change from baseline was calculated as the post-Baseline value minus the Baseline value.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Participants underwent an EGD with biopsies at Screening and at Weeks 12 and 24. Mean esophageal eosinophils were calculated as the mean number across all esophageal biopsies at each time point. Screening value was considered as the Baseline value. Change from Baseline was calculated as the post-Baseline value minus the Baseline value.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Blood samples were obtained at Screening, pre-infusion and 24h and 72-96h post-infusion at Day 1, Weeks 4 and 8; and at Week 2, 6, 10, 12, 16, 20, 24 and 34 visits or Early Withdrawal Visit to estimate blood eosinophil count.

ITT Population. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Blood samples were obtained at pre-infusion and 5m, 2h, 24h, 72-96h post-infusion at Day 1, Weeks 4, 8; and Weeks 2, 6 10, 12, 16, 20, 24 and 34 to estimate the plasma concentration of mepolizumab. Only those participants available at specified time points are analyzed (represented as n=X,X,X in the category titles).

Pharmacokinetic Population: all participants who received study medication and for whom mepolizumab sample was obtained and analyzed.