Genentech’s original marketing application was submitted in June 2000. As a result of interactions between Genentech and CBER, Genentech submitted more clinical trial data in December 2002. The original application contained 4 trials that were suitable for an evaluation of efficacy in the proposed indication, asthma; the newly submitted data contain 1 adequate and well-controlled trial as well as 2 open-label trials.

Safety data are reviewed in another document.

The proposed indication statement for omalizumab is:

XOLAIR is indicated as maintenance therapy for the prophylaxis of asthma exacerbations and control of symptoms in adults and adolescents (12 years and above) with moderate to severe allergic asthma that is inadequately controlled despite the use of inhaled corticosteroids.

Genentech defines allergic asthma as follows:

Allergic asthma is an IgE-mediated atopic disease that, in susceptible individuals, causes recurrent episodes of wheezing, breathlessness, chest tightness, and cough, particularly at night and in the early morning.

The efficacy parameter assessed in the critical efficacy trials was the occurrence of asthma exacerbations. This parameter is also assessed in the other trials reviewed for efficacy. A summary of these trials is included below. This document contains a brief description of asthma and the product, omalizumab. It reviews the major efficacy findings in 4 adequate and well-controlled trials additional information provided in 2 open-label trials. Finally, it reviews information related to the asthma related clinical outcomes and to the severity of the asthma in subjects studied in the controlled trials.

OVERVIEW OF ASTHMA

Asthma is defined in the National Institutes of Health Guidelines for the Diagnosis and Management of Asthma (“NHLBI Guidelines,” 1997)¹, as a chronic inflammatory condition of the airways. The symptoms of asthma are recurrent episodes of wheezing, breathlessness, chest tightness, and cough, particularly at night and in the early morning. Obstruction to the outflow of air from the lungs, due to bronchoconstriction, occurs variably within an individual, and may reverse spontaneously or with treatment. This obstruction, which can cause wheezing, shortness of breath, and cough, may occur acutely in response to many different kinds of stimuli, and is thought to be a consequence of inflammation-induced hyperresponsiveness of the airways. Allergens may induce bronchoconstriction in susceptible persons through an interaction with IgE. Other factors, not dependent upon an immediate reaction with IgE, may stimulate bronchoconstriction, for example, aspirin and other nonsteroidal anti-inflammatory drugs, exercise, cold air, and irritants. The pathology of asthma consists of infiltration of airways with inflammatory cells, disruption of the airway lining with deposition of collagen beneath the epithelium, and microvascular leakage. There is hypertrophy of airway smooth muscle.

Asthma broadly can be characterized into childhood-onset and adult-onset disease. Childhood onset asthma is frequently found in children who are atopic, that is, with a genetic susceptibility to produce IgE toward common environmental antigens. However, in some persons who develop asthma as adults there is no family history of asthma nor are IgE antibodies to allergens found. The presence of antigen-specific IgE is not sufficient to produce asthma. Persons with elevated serum IgE or skin reactions to allergens may not have asthma.

In general, total serum IgE values increase progressively throughout childhood, level off in adulthood, and begin to drop (concordant with other immunoglobulin levels) in old age. Antigen-specific IgE will appear in the serum following allergen sensitization. IgE may actually decrease somewhat at the time of allergen exposure during the allergen season for seasonal allergens, increase after the allergen season to 2-3 fold the pre-season baseline, then come back to baseline over the ensuing 2-3 months. There are not detailed or extensive longitudinal data on the magnitude of variability of IgE in adults in the medical literature. The variation that may be induced by other factors is also not well described (e.g., parasitic infection, other intercurrent illness, smoking, or environmental changes).

Asthma is a common disease. According to National Health Interview Survey statistics, about 27 million persons in the United States reported a physician diagnosis of asthma during their lifetime (in 1997), and about 10.5 million reported at least 1 asthmatic attack in the previous 12 months (in 1999)³. Approximately 25-30% of these cases were in persons less than 15 years old. Asthma can also be a mortal condition. In 1999, asthma was responsible for about 4600 deaths (about 4% of these in persons less than 15 years old). Due to the lack of a standard for the diagnosis of “allergic” asthma, the numbers of subjects with this condition are not established.

The subjective measures used to grade asthma generally include ability to sleep through the night, ability to participate in daily activities without breathlessness, the occurrence of acute worsenings called exacerbations, and exercise tolerance. Peak expiratory flow rate (PEFR) may be used for home monitoring of obstruction of breathing; in the clinic, expiratory volume in the first second of a forced expiration (FEV₁) is determined. The FEV₁ is measured using equipment found in a clinic or hospital that can be calibrated so that individuals may be compared rigorously to reference populations. Decrements in either measure signify a worsening in the ability to exhale rapidly and completely, due in asthma to reversible obstruction of the airways.

The NHLBI Guidelines categorize asthma as mild intermittent, mild persistent, moderate persistent, and severe persistent, based upon pretreatment symptoms and measurements. Figure 1 shows this codification scheme.

Figure 1. Severity of asthma according to NHLBI Guidelines

Clinical features before treatment

Symptoms**

Nighttime Symptoms

Lung Function

STEP 4

Severe

Persistent

· Continual symptoms

· Limited physical activity

· Frequent exacerbations

Frequent

· FEV1 or PEF £60% predicted

· PEF variability>30%

STEP 3

Moderate

Persistent

· Daily symptoms

· Daily use of inhaled short-acting beta2-agonist

· Exacerbations affect activity

· Exacerbations ³ 2 times a week; may last days

>1 time a week

· FEV₁ or PEF>60% to <80% predicted

· PEF variability>30%

STEP 2

Mild

Persistent

· Symptoms >2 times a week but <1 time a day

· Exacerbations may affect activity

>2 times a month

· FEV1 or PEF ³80% predicted

· PEF variability 20% to 30%

STEP 1

Mild

Intermittent

· Symptoms £2 times a week

· Asymptomatic and normal PEF between exacerbations

· Exacerbations brief (from a few hours to a few days); intensity may vary

<2 times a month

· FEV1 or PEF ³ 80% predicted

· PEF variability <20%

* The presence of one of the features of severity is sufficient to place a patient in that category. An individual should be assigned to the most severe grade in which any feature occurs. The characteristics noted in this figure are general and may overlap because asthma is highly variable. Furthermore, an individual’s classification may change over time.

** Patients at any level of severity can have mild, moderate, or severe exacerbations. Some patients with intermittent asthma experience severe and life-threatening exacerbations separated by long periods of normal lung function and no symptoms.

The NHLBI guidelines state that an individual should be assigned to the most severe grade in which a feature occurs. Another classification scheme, the GINA guidelines, classifies asthma as intermittent, mild persistent, moderate persistent, and severe persistent in terms of both asthma features and coincident treatment². Importantly, these classifications of asthma are necessarily broad and may contain many degrees of severity. For example, a patient with severe persistent asthma with frequent hospitalizations on treatment or requiring large doses of oral corticosteroids would be considered more refractory than another patient with severe persistent asthma who had not been hospitalized or who does not require systemic corticosteroids.

There are many therapies available to treat asthma. Long-term controller medications include b-agonists, leukotriene antagonists, 5-lipoxygenase enzyme inhibitors, cromolyn sodium and nedocromil; theophylline; short term controller medications include b-agonists and ipratropium bromide. Corticosteroids have been the mainstay of controller medications of asthma; they are considered antiinflammatory agents. Oral corticosteroids are known to have systemic side effects such as suppression of growth in children, cataract formation, osteoporosis, and disturbance of glycemic control. As such, oral corticosteroids are reserved for more severely affected asthmatics. Inhaled corticosteroids are considered much safer, due to much less systemic exposure. Troleandomycin, methotrexate, cyclosporine, and other immunomodulators have been tried in cases of glucocorticoid-resistant asthma, or in severe cases in which corticosteroids may be contraindicated. These medications are given in conjunction with modification of exposure to agents in the environment or activities that are known to trigger exacerbations.

PRODUCT INFORMATION

Omalizumab is a recombinant Chinese Hamster Ovary cell-derived humanized IgG1k monoclonal antibody. It binds to IgE and inhibits the binding of IgE to FceRI, the high affinity receptor for IgE on mast cells and basophils. In an allergic reaction, allergens bind and crosslink the IgE bound to this receptor. Aggregation of the underlying FceRI receptors triggers the cells to release histamine and other mediators of the allergic response. Omalizumab is meant to reduce the pool of IgE available to interact with FceRI and thereby reduce the allergic response.

Various formulations were tested during clinical development. Initial trials were with a liquid formulation; subsequently, a lyophilized formulation was used. The trials reviewed in this document were performed with a lyophilized formulation; however, the product used in trial Q0694g, while lyophilized, differed with respect to production method, product concentration, excipients, and vial configuration.

The adequate and well-controlled critical asthma efficacy trials (trials 008 and 009) and trial 011 were conducted using omalizumab produced from a process originally described in 1997. This process was modified in 1999 to produce the to-be-marketed product. The to-be-marketed product was used in open-label trials Q2143g and IA04. Based on submitted comparability studies, CBER has judged that product produced by the to-be-marketed process is comparable to product used in trials 008, 009, 010, and 011.

DOSE SELECTION

Genentech states that the dose and dosing regimen used in the pivotal trials was selected on the basis of suggestions from previous trials that average free IgE concentrations of <25 ng/ml were correlated with efficacy and estimates from pharmacokinetic analyses of the duration of suppression based on amount of administered product. Since the product was to be targeted to children as well as to adults, a body weight correction was added.

Comment

The a priori relevance of serum free IgE suppression is unclear, since the release of histamine and other mediators from IgE-receptor-bearing cells is more dependent upon bound IgE than on free IgE, and the relation of bound to free IgE is complex. Other factors thought to be important in mediator release include the intrinsic ability of an individual’s cells to release mediators, the activation state of cells, the degree of crosslinking of the IgE receptor by antigens as well as the expression of the receptor on cell surfaces.

CLINICAL TRIALS CONDUCTED AND SCOPE OF REVIEW

Prior to the conduct of the critical efficacy trials in asthma, trials 008 and 009, Genentech conducted one 24-subject open-label trial studying asthma subjects as part of its population and two single-blind, placebo-controlled trials in asthma (n=34 and 12). Genentech also conducted 3 randomized, double-blind, placebo-controlled trials in asthma prior to the pivotal trial. Trial Q0630g studied 20 subjects, trial Q0634g studied 19, and Q0694g studied 317 subjects. Trial Q0694g was the first asthma trial to use a lyophilized formulation but was consistent with previous trials in that it involved the intravenous route. Subsequent to this trial Novartis conducted trials coded 008, 009, and 010, in which the subcutaneous route was used.

Table 1 shows a summary of the clinical trials that are suitable for a review of their effects on asthma symptomatology, exacerbations, and medication use. The open-label trials are reviewed primarily because they enrolled subjects whose concomitant medications were liberalized in comparison to the critical efficacy trials, or had worse control of their asthma.

Table 1. Summary of major trials for efficacy

Trial	n	Ages	Design
Q0694g	317	11-50	Placebo-controlled; two dose levels; 2:1 randomization; double blind
008	525	12-74	Placebo-controlled; double-blind stable steroid, steroid reduction, and extension periods
009	546	12-76	Identical to trial 008
010	334	5-12	Pediatric; placebo-controlled (2:1 randomization); double-blind stable steroid, steroid reduction, and open-label extension periods
011	341	12-75	Placebo-controlled; double-blind stable steroid, steroid reduction periods
Q2143g	1899	6-76	Open-label; 2:1 randomization to omalizumab or standard treatment
IA04	312	12-73	Open-label; 2:1 randomization to omalizumab or standard treatment

_________________________________________________________________________________

CLINICAL TRIAL DATA

EXPLORATORY TRIAL Q0694G

Trial Q0694g was a trial comparing two doses of omalizumab to placebo. Its design was similar to that of the subsequent pivotal trials 008, 009, and 010. It employed the intravenous route of administration of the product.

Title

Trial Q0694g was entitled “A Phase II, multicenter, double-blind, placebo-controlled study to evaluate the safety and efficacy of Anti-IgE recombinant humanized monoclonal antibody (rhumab-E25) in subjects with moderate to severe allergic asthma.”

Design

This was a double-blind comparison of 2 dose levels of omalizumab to placebo in 504 subjects 12-45 years old with asthma requiring corticosteroid treatment. The primary endpoint was an overall asthma score. There were to be several phases: 4-week run-in, 1-week baseline, 12-week phase with stable corticosteroid dosing, 8-week treatment phase with protocol-defined corticosteroid tapering, and a final 10-week safety period after the last dose of omalizumab was administered (Figure 2). The objectives of the trial were to examine efficacy, safety, and pharmacokinetics and pharmacodynamics.

Figure 2. Design of trial Q0694g

Comment

The critical efficacy trials had this basic design, but used asthma exacerbations as their primary endpoint.

Treatment

Product and placebo were supplied as lyophilates to be reconstituted with water for injection, 2 ml. The dose was titrated upward in the first week. Starting at day 7 and then every 2 weeks thereafter, the single dose was to be 0.006 or 0.014 mg/kg/IU (IgE)/ml.

Comment

The dosing choices included a dose similar to the 0.016 mg/kg/IU [IgE] every 4 weeks proposed for marketing (0.012 mg/kg/IU [IgE] every 4 weeks) and one approximately twice the proposed marketed dose (0.028 mg/kg/IU [IgE] every 4 weeks). The titration of the 1^st dose was abandoned in the critical efficacy trials.

Concomitant treatments

Subjects were restricted in their use of concomitant medications, as was the case in the critical efficacy trials. In trial Q0694g, subjects were only to take protocol-defined corticosteroids (inhaled triamcinolone with or without prednisone or methylprednisolone) and the b-agonist albuterol for “rescue” in case of worsened asthma. Subjects were permitted only terfenadine (Seldane) and beclomethasone nasal spray (Vancenase) for the relief of symptoms of allergic rhinitis.

Randomization and blinding

Randomization was to be 2:2:1:1 (product at 0.006 or 0.014 mg/kg/IU (IgE)/ml or corresponding placebo). Subjects were to be randomized into groups requiring either triamcinolone (inhaled) or prednisone (oral).

Trial medication was to be shipped open-label to sites, where the preparers were not to be involved in any other aspect of the trial.

Subject qualifications

Entry criteria were intended to select a population of asthmatic subjects with skin test reactivity to allergens that they would be exposed to, with minimal symptoms on moderate amounts of corticosteroids. Subjects with large or prolonged corticosteroid requirements were to be excluded. The asthma symptom score used for inclusion was also used as the primary endpoint data.

Inclusion criteria (selected)

Male or females 12–45 years old on inhaled corticosteroids or oral corticosteroids

Total serum IgE level £1785 IU/ml

Skin reactivity to two or more different allergens to which there would be exposure during trial

Documented history of reversible airway obstruction

Chronic use of oral (£20 mg of prednisone daily or £40 mg every other day or £16 mg methylprednisolone daily) or inhaled ( ³600 mg of triamcinolone) corticosteroids at enrollment

At visit 6:

FEV₁ of 50%–100% of predicted for height, age, and sex

Mean daily symptom score of ³2.5 as measured from Day –7 to Day 0

For subjects with a FEV₁ of ³70% predicted, a positive response to inhalation of methacholine (PC₂₀ FEV₁[Methacholine] equal to or less than 8 mg/ml)

Exclusion criteria (selected)

Chronic daily use of oral corticosteroids for more than 12 months

Receipt of escalating doses of immunotherapy

Greater than 150% of ideal body weight for height (adults) or weight (adolescents)

Comment

The trial intended to study asthmatics selected for skin test reactivity and exposure to an allergen or allergens. The subject population was allowed to take moderate amounts of oral corticosteroids, which would tend to allow more severe asthmatics than studied in the pivotal trials. The subject age qualification was restrictive, since asthma is a disease that occurs in the geriatric group as well.

Procedures and evaluations

The primary endpoint data, asthma symptom scores were recorded twice daily. Questions (16 for adults and 17 for adolescents) asked about the major symptoms of asthma for both groups, and values ranged from 1 (no symptom) to 7 maximal symptom or all the time) for each question).

Procedures (selected) in the trial were as follows:

· Screening, day -35

· Run-in period from weeks –4 to –2:

--prednisone/methylprednisolone and/or inhaled triamcinolone adjusted to lowest doses possible with still acceptable asthma symptoms and PEFR

--albuterol given to treatment acute symptoms

--Discontinue other asthma or rhinitis drugs (except terfenadine and nasal beclomethasone)

· Baseline, days –7 to –1:

--recording of adverse events and symptom scores, albuterol use, and PEFR

· Day 0:

--asthma evaluations and first infusion

· Treatment with stable steroid phase

--Infusions, every 2 weeks after the first week

--asthma and routine evaluations

--methacholine challenge at day 49

--total and free serum IgE and product at days 7, 21, and 49

· Treatment with corticosteroid tapering phase

--infusion every 2 weeks (last infusion on day 133)

--asthma and routine evaluations

--total and free serum IgE and product at days 91, 133, and 140

During this phase, corticosteroid tapering was done as follows:

--For subjects on inhaled triamcinolone at doses ³600 mg/d, every-2-week tapering

--For subjects on (ingested) prednisone at £20 mg/d or 40 mg QOD, or (ingested) methylprednisolone, £16 mg/d, tapering to 0 by week 8 of the phase

--For subjects on triamcinolone as well as an ingested steroid, the triamcinolone was to be continued and the ingested corticosteroid tapered

--The protocol defined reasons (PEFR, FEV₁, asthma symptoms, increase in b-agonists) for the discontinuation or reinstitution of a taper:

· follow-up, day 210:

--complete physical exam, vital signs, spirometry, methacholine challenge, clinical labs, urinalysis, total and free serum IgE and product

Comments

The design of medication standardization with a common corticosteroid and rescue b-agonist, determination of a corticosteroid dose acceptable to the subject and investigator, followed by a stable steroid phase and a subsequent steroid reduction phase, was copied in the pivotal trials.

Analytical plan

Endpoints

The primary endpoint was the change in overall symptom score from baseline to 12 weeks (stable steroid phase). The overall symptom score was the mean of the 16 scores for adults and 17 scores for adolescents.

Important secondary endpoints will be described below.

Analytical populations

The actual final analysis of the primary endpoint, PEFR, and b-agonist use were performed on subjects with at least 4 weeks of post-randomization data; other efficacy data were analyzed in all subjects.

Summary of statistical methods

The primary efficacy comparison was to be performed between the high-dose active group and the placebo group (pooled if no difference seen in initial statistical testing) using ANOVA stratified by randomization category (corticosteroid use/age). The analysis was to be performed on subjects with at least 4 weeks of post-randomization data. Scores of the last week completed would be used for early dropouts.

Corticosteroid use change was to be analyzed in terms of protocol-defined amounts of reduction (400 mg inhaled or 8 mg oral; or in categories of improved, unchanged, or worsened based on changes of 200 mg of inhaled or 5 mg of oral corticosteroids); the statistical test to be used was specified for those taking inhaled steroids only as the Wilcoxon rank-sum test.

Comments

The clinical meaning of changes in the asthma diary score (the primary endpoint) has not been established formally. Thus the clinical meaning of the primary endpoint results in this exploratory trial can be seen as suggestive only.

Protocol modifications

The protocol was made final on February 23, 1996. It was amended 3 times, all after trial initiation on April 5, 1996. None of the changes would have been expected to alter the judgment of the treatment effect significantly.

Results: Conduct of the trial

Dates of the trial

The trial was initiated on April 5, 1996, and completed on July 1, 1997.

Early cessation of enrollment

Planned enrollment was to be 504, equally divided among adults taking inhaled corticosteroids, adolescents taking inhaled corticosteroids, and both age categories taking oral corticosteroids. When adult enrollment was exceeded, enrollment overall was stopped, although there was underenrollment in the adolescent and oral corticosteroid groups. Final enrollment was 317.

Eligibility and other protocol violations

The nature and extent of these violations, which were equally balanced among the treatment arms, would not be expected to substantially harm the ability to detect efficacy of the product.

Enrollment by site

There were 27 sites in trial Q0694g, with no one site or small number of sites dominating enrollment. The largest site enrollment, for one site, was 27 there were 12 sites whose enrollment was between 10-20 subjects inclusive, and 14 in which site enrollment was between 6 and 9 subjects inclusive.

Demographics and baseline characteristics

Table 2 shows that the treatment arms were balanced for important demographic and baseline characteristics. The trial population was primarily Caucasian and composed of slightly more females than males; 80-85% of the trial were adults. Proportionately more subjects had a history of hospitalization than in the critical efficacy trials, and this trial enrolled subjects on oral corticosteroids, unlike the critical efficacy trials. However, mean FEV₁ was similar.

Table 2. Trial Q0694g: Demographics and baseline characteristics

Placebo

n=105**

Omalizumab

0.006*

n=106**

Omalizumab

0.014*

n=106**

Age mean (range)

<18 (n)

³18 (n)

30 (11-48)

30 (12-47)

29 (12-50)

Sex (% male)

Race (% Caucasian)

Height (m)

1.71

(1.49-1.94)

1.69

(1.52-1.95)

1.69

(1.47-1.98)

Weight (kg)

(42-136)

(44-133)

(39-140)

Baseline IgE (IU/ml)

275

(19-1390)

344

(17-1646)

374

(27-1957)

Former smoker %)

Age of asthma diagnosis

(0-38)

(1-44)

(0-41)

Hospitalized for asthma in last year, %

Asthma emergency room visits

per year

0.9

(0-15)

0.9

(0-20)

1.1

(0-30)

Overall symptom score (0-7)

4.0

(1.5-6.5)

4.0

(2.0-6.5)

4.1

(2.4-6.5)

Inhaled corticosteroid dose in adults (g)

median and range

800

(200-4000)

n=76

800

(400-3200)

n=78

800

(200-2400) n=78

Inhaled corticosteroid dose

in adolescents (mg)

median and range

800

(400-1600)

n=17

800

(600-2000)

n=14

800

(400-2600) n=19

Oral corticosteroid dose (mg),

median and range

10.0

(2.5-40)

n=12

10.0

(5.0-20.0)

n=14

10.0

(5.0-10.0)

n=19

Stratification variables

Adolescents on inhaled triamcinolone (n)

Adults on inhaled triamcinolone (n)

Adults and adolescents on (oral) prednisone (n)

Morning PEFR (l/min)

384

(150-620)

380

(151-626)

378

(143-599)

FEV₁, % predicted

(32-101)

(29-115)

(34-129)

b-agonist in subjects using MDI only (puffs)

8.2

(2.0-16.8)

n=63

8.8

(2.0-22.7)

n=66

8.8

(2.0-37.7)

n=73

* mg/kg/IU (IgE)/ml

**some cells have smaller subject numbers, as shown

Comments

The severity of some of the subjects was greater than that in the pivotal trials, since investigators felt the need for oral corticosteroid treatment, and medical care visits were greater in some. The primarily Caucasian makeup of the population mirrored that of the pivotal trials. Geriatric subjects were excluded.

Premature discontinuations

Discontinuations from the trial were slightly more common among placebo subjects (about 15% vs. 10% (omalizumab lower dose) or 7% (omalizumab higher dose)). The reasons for discontinuation were fairly well balanced among the treatment arms. The nature and extent of the discontinuations would not be expected to influence the overall judgment of the effect of omalizumab in this trial.

Interim efficacy analyses of trial

Analyses of trial results were performed by Genentech at the end of the stable steroid and steroid reduction phases prior to full completion of study. Treatment assignments were kept from investigators and subjects throughout the trial.

Results: Efficacy

Analytical populations

The results were analyzed on the pooled placebo population, as there was no statistical difference between the two placebo groups.

Primary efficacy results

Genentech conducted an ANCOVA analysis on the intent-to-treat population, with last observation carried forward, as an exploratory analysis. Because this type of analysis is subject to less bias than the planned primary (all subjects with at least 4 weeks of post-randomization data), it will be presented instead of the presented primary analysis. Table 3 shows the results.

Table 3. Baseline Overall Symptom Score and reduction**

Placebo

N=100

Omalizumab

0.006*

N=103

Omalizumab

0.014*

N=103

Baseline

Mean ± std. error

4.0 ± 0.1

4.1 ± 0.1

Week 12 reduction

Mean ± std. error

p-value

0.8 ± 0.12

1.3 ± 0.12

0.003

1.3 ± 0.12

0.004

Week 20 reduction

Mean ± std. error

p-value

1.0 ± 0.13

1.3 ± 0.13

0.09

1.4 ± 0.13

0.03

* mg/kg/IU (IgE)/ml

**ITT population

Comment

Both the pooled placebo groups and omalizumab groups improved during the trial. The difference from placebo in the change in symptom score was very small, approximately 0.5 symptom score units, consistent with the magnitude of changes seen in the critical efficacy trials. The clinical meaning of the intertreatment differences seen in total score were not clarified in the submission.

Secondary efficacy results (selected)

· Corticosteroid reductions

Genentech examined reductions in corticosteroid use among subjects with inhaled only corticosteroid use at baseline using imputation of the last observation carried forward for those who discontinued (Table 4). A somewhat larger proportion of subjects on active treatment were able to discontinue their use of corticosteroid entirely, a number that was similar in the two active dose groups.

Table 4. Trial Q0694g: Inhaled corticosteroid reductions (subjects on inhaled corticosteroid only)

Placebo

n=93

Omalizumab

0.006*

n=92

Omalizumab

0.014*

n=97

Baseline median dose (mg)

800

Median reduction (%)

p=0.022

p=0.039

Subjects with

>50% reduction

35 (38%)

45 (49%)

p=0.122

49 (51%)

p=0.074

Subjects with

100% reduction

11 (12%)

21 (23%)

p=0.048

17 (18%)

p=0.268

*mg/kg/IU (IgE)/ml

P-values compare the omalizumab groups to placebo using the Wilcoxon rank-sum test for continuous

endpoints and Pearson c² test for binary endpoints.

The numbers of subjects who were receiving oral corticosteroids at baseline was quite small (12 placebo, 14 low-dose omalizumab, and 9 high-dose omalizumab) so conclusions regarding this group are tenuous. However, there was a trend toward benefit in this group: the median percents reduction in oral corticosteroids in the placebo, low-, and high-dose groups were 0, 65% (p=0.106 compared to placebo), and 50% (p=0.045 compared to placebo).

Comment

The reduction in oral corticosteroid dosing was not seen in trial 011, which randomized 100 subjects to placebo or to the proposed regimen of omalizumab in a subcutaneous dosing.

· Morning PEFR,

Changes in PEFR were small in placebo and omalizumab groups. At baseline all groups were at about 380 l/min; at week 20 the increases, in liters/min in the placebo, lower, and higher dose treatment groups were about 10, 21, and 30 l/min, respectively. Smaller changes were seen in the evening PEFR, but the treatment effect was present, with an apparent positive relationship to dose.

· Change from baseline in FEV₁

There were no clinically important or statistically significant differences during the trial between either treatment group and placebo in change in FEV₁ from baseline to the end of the stable steroid or steroid reduction phase.

· Asthma exacerbations

This was not a protocol-defined endpoint, and there is no definition of an exacerbation for this endpoint in Genentech’s analysis. However, due to its importance in relation to the primary endpoint of the pivotal trials, these data will be summarized here. The data suggest that the treatment reduced the number of asthma exacerbations reported as adverse events, to the same level regardless of dose. The data also suggest that asthma exacerbations that required treatment with oral corticosteroids were reduced by the same extent regardless of treatment dose among those taking only inhaled corticosteroids at baseline.

Table 5. Q0694g: Asthma exacerbations reported as adverse events

Placebo

Omalizumab

0.006*

Omalizumab

0.014*

Subjects with asthma

exacerbations (%)^a

47/105 (45%)

30/106 (28%)

p=0.01

32/106 (30%)

p=0.03

Mean asthma exacerbations/

subject ^b,c

0.77

0.41

p=0.01

0.44

p=0.02

Subjects on inhaled corticosteroids at baseline with asthma exacerbations treated with oral corticosteroids (%)^a

26/93 (28%)

15/92 (16%)

p=0.06

13/97 (13%)

p=0.01

Mean asthma exacerbations treated with oral corticosteroids in subjects on inhaled corticosteroids at baseline^{b, c}

0.38

0.21

p=0.05

0.22

p=0.02

*mg/kg/IU (IgE)/ml

^a ITT population; p-values: Pearson c² test

^b P-values: Wilcoxon rank-sum test

^c total number of exacerbations divided by the total number of subjects

Antibody Development and Efficacy

Serum was assayed for antibodies to either Fc or F(Ab')₂ portion of omalizumab. Genentech reports that no antibody development occurred.

Comments

According to the product reviewer, the data provided are inadequate to assess the correctness of Genentech’s assertion. In any event, lack of development of antibodies to the product after its intravenous administration may not be relevant to its immunogenicity after subcutaneous administration (the intended route).

Summary: Efficacy in supportive trial Q0694g

Interpretation of the results of this trial during the steroid reduction phase is compromised by unblinding of results to Genentech personnel prior to this phase, with the potential for biases in reporting and measurement of effects. In addition, production method changes between the material used in Q0694g and trials 008-010 lessen the relevance of these data to those developed in the latter trials. However, results in Q0694g were consistent with those in trials 008-010. Intertreatment differences in the changes in symptom scores were small; equivocal differences were seen in different measures of pulmonary function. Intravenous omalizumab treatment was associated with lower asthma exacerbations rates. These effects were consistent with those in the pivotal trials.

_________________________________________________________________________________

CRITICAL EFFICACY TRIALS 008 AND 009

Trials rhuMAb-E25 01 008 and 009 were the pivotal trials for adolescents and adults in asthma. They were nearly identical in design and will be reviewed in parallel.

Titles

Trials 008 and 009 were both entitled “A Phase III, 7-month, randomized, double-blind, parallel-group, placebo-controlled, multicenter study with a 5-month blinded extension period to assess the efficacy, safety, tolerability, steroid-reduction, pharmacokinetics, and pharmacodynamics of subcutaneous rhuMAb-E25 in adolescents and adults with moderate to severe allergic asthma requiring daily treatment with inhaled corticosteroids.”

Dates of the protocols

The protocols were made final on November 21, 1997 and amended formally on May 7, 1998. This review reflects the final amended version of the protocol, with other changes as noted (see section of review on protocol modifications).

Design

Trials 008 and 009 were designed as double-blind comparisons of subcutaneously administered omalizumab or placebo in asthmatic subjects with skin-test sensitivity to common allergens. Planned enrollment was 550 subjects each.

Table 6 is a schematic of the trials. Screening was performed at week –7, followed by a run-in period of 4-6 weeks during which the dose of inhaled corticosteroid (Beclomethasone dipropionate, or BDP) was to be stabilized. The evaluation “core” period was divided into a 16-week period during which corticosteroids were to be held stable followed by a corticosteroid reduction phase of an additional 16 weeks. Following this were 5 months of double-blind extension, during which subjects were to remain on their assigned treatment, but concomitant medications and the dose of prescribed corticosteroid were to be liberalized. A follow-up evaluation was performed after cessation of treatment for safety evaluations and the determination of antibodies to the product.

Table 6. Schematic of Trials 008 and 009

Screening

Run-in

Corticosteroid Stabilization

Corticosteroid reduction

Extension

Follow-up

Visit

3-7*

7-13

13-19

Week

-7

-6/-4 to 0

0-16

16-28

28-52

Treatment

none

Randomized double-blind omalizumab

or placebo

none

Inhaled

Corticosteroids

BDP ³420 mg/day

or equivalent

BDP 420-840

mg/day

BDP

stable dose

tapered BDP dose

up to 8 wks, BDP stable

dose 4 wks

BDP treatment as

appropriate

any

*Visit 2 was divided biweekly into 2.1, 2.2, and 2.3. Visits 3-6 were divided biweekly into 3, 3.1, 4, 4.1, etc.

Note: BDP is beclomethasone dipropionate

Objectives

The objectives of the trials were to examine efficacy, safety, and pharmacokinetics and pharmacodynamics of omalizumab.

Trial treatments

Omalizumab was to be supplied as a lyophilate containing 150 mg omalizumab, 108 mg sucrose, 1.3 mg L-histidine, 2.1 mg L-histidine HCl monohydrate, and 0.4 mg polysorbate 20. It was to be reconstituted with water for injection.

Omalizumab was to be administered subcutaneously at a dose normalized for body mass and serum IgE level, approximately 0.016 mg/kg/IgE [IU/ml] per month. These doses were to be selected from a chart of body mass and serum IgE categories (Table 7). Dosing in a given cell of the table was calculated to provide 0.016 mg/kg/IgE for the maximal body mass and serum IgE level that the cell referred to; that is, most subjects were to receive more than the idealized normalized dose. No subject was to receive more than 375 mg as a single administration; if a calculated monthly dose were more than 300 mg, the dose was given as 2 equal doses every 2 weeks. If a subject’s body mass and serum IgE demanded a dose of omalizumab higher than 750 mg per month, they would be excluded from the trial (Table 7).

Table 7. Dosing table for trials 008 and 009 (milligrams/dose)

Baseline IgE (IU/ml)	Body mass (kg)					Frequency of dosing
Baseline IgE (IU/ml)	30-60	>60-70	>70-80	>80-90	>90-150	Frequency of dosing
>30-100	150	150	150	150	300	Q4wk
>100-200	300	300	300	300	225	Q2wk
>200-300	300	225	225	225	300	Q2wk
>300-400	225	225	300	300		Not dosed
>400-500	300	300	375	375
>500-600	300	375
>600-700	375

Comments

This dosing goal of a minimum 4-weekly dose was replicated in every important trial of efficacy. In the pediatric trial the IgE limit was extended upwards, and in the open-label trial Q2143g, the upper body mass category was split (see review of that trial), but the overall scheme has been consistent.

Concomitant medications

Concomitant medications were to include inhaled BDP and “rescue” inhaled albuterol (trial 009 used salbutamol as rescue). If a subject were to have taken allergy vaccination therapy (desensitization immunotherapy) for ³3 months of stable doses before visit 1, he or she was to maintain this treatment unchanged throughout the trial. Short- or medium-acting antihistamines were allowed for the treatment of allergic rhinitis.

The protocol prohibited all the major medications used to treat moderate asthma (the subjects would not be expected to use medications for glucocorticoid resistant asthma): oral, parenteral, nebulized, or aerosol b-2agonists (excluding the prescribed albuterol rescue medication), theophyllines, cromolyn sodium or nedocromil sodium, oral or parenteral corticosteroids (except for treatment of asthma exacerbation as defined above), leukotriene receptor inhibitors, 5-lipoxygenase enzyme inhibitors, oral/inhaled anticholinergics, long-acting antihistamines, b-adrenergic antagonist medications, or any investigational, experimental, or nonapproved drugs. Subjects were not to start desensitization immunotherapy for allergies.

Comment

Excluding a large number of concomitant medications limited the trial population to those who could be managed reasonably on modest doses of inhaled corticosteroids alone.

Randomization and blinding

Randomization was in balanced blocks (n=4) of patient numbers for each of the two treatment groups within each center.

Treatments were to be shipped to sites open-label. Each site was responsible for reconstitution of treatments prior to administration and for ensuring that personnel responsible for reconstitution and administration were not be involved in subject evaluations. Inspection of sites by Biologics Inspection and Monitoring revealed no cause for concern due to unblinding.

Subject qualifications

Inclusion and exclusion criteria were designed to ensure that subjects were within given IgE and weight ranges so that they could be dosed according to the dosing table provided. In addition, subjects had to have a skin test reaction to a common allergen to which they would be exposed. They were to have symptomatic asthma while on a stable dose of corticosteroid and rescue bronchodilators.

Inclusion criteria

1. Male and female, aged 12-75 years, willing to sign informed consent

2. Diagnosis of allergic asthma ³ 1 year duration who also meet the following criteria:

a. Meet standards of the American Thoracic Society

b. A positive prick skin test (e.g., +3 reaction) to at least one of the following allergens to which patients are exposed to during the trial: Dermatophagoides farinae, Dermatophagoides pteronyssinus, cockroaches (whole body), dog or cat.

c. Total serum IgE level ³ 30 to £ 700 IU/ml and body weight £ 150 kg.

d. ³12% increase in FEV₁ over baseline value within 30 minutes of taking one or two puffs of albuterol (90 mg/puff)

e. Baseline FEV₁ ³ 40 to £ 80% of the predicted normal value, demonstrated 6 or more hours after short-acting b-2-agonist or 72 hours or more after long-acting b-2-agonist

f. Mean daily total symptom score of ³ 3.0 during the last 14 days prior to randomization*

g. Requiring treatment with inhaled corticosteroids in doses equivalent to beclomethasone 420 to 840 mg per day, for ³3 months prior to randomization; and as needed or regular use of bronchodilator therapy.

3. No significant change in the regular asthma medication, no acute asthma exacerbation requiring corticosteroid rescue for at least 4 weeks prior to run-in period (Visit 2.1)

4. Able to use the Mini-Wright peak flow meter for the measurement of peak flow, and a metered-dose inhaler (MDI) for administration of albuterol rescue medication.

*The asthma symptom score grades asthma symptoms by 3 periods of a day:

· morning symptoms (0=no, 1=yes)

· nocturnal symptoms

0=I did not wake up because of any breathing problems.

1=I awoke once because of my breathing problems, but did not use my rescue medication.

2=I awoke once because of my breathing problems, but my rescue medication controlled my symptoms.

3=I awoke more than once because of my breathing problems, but my rescue medication controlled my symptoms.

4=I had difficulty sleeping because of my breathing problems even though I used my rescue medication.

· Daytime symptoms

0=No symptoms at all; unrestricted activity.

1=Symptoms caused little or no discomfort; unrestricted activity.

2=Symptoms caused some discomfort, at times limiting strenuous activity.

3=Symptoms caused moderate discomfort and sometimes limited routine activity.

4=Symptoms occurred at rest, caused marked discomfort, and usually limited routine activity.

The maximum score is 9, minimum is 0

Exclusion criteria (selected items)

1. Previous treatment with rhuMAb-E25 or prior randomization into the trial

2. Hypersensitivity to any ingredients of product or to trial medication drugs related to trial medication

3. History of acute infectious sinusitis or respiratory tract infection within 1 month prior to Visit 1

4. Aspirin or other nonsteroidal anti-inflammatory drug-related asthma

5. Active lung disease other than allergic asthma (e.g. chronic bronchitis)

6. Elevated serum IgE levels for reasons other than atopia (e.g., parasitic infections, hyperimmunoglobulin E syndrome, allergic bronchopulmonary aspergillosis)

7. Currently taking allergy vaccination therapy (desensitization immunotherapy), with less than 3 months of stable maintenance doses prior to Visit 1

8. Use of antihistamines, leukotriene receptor inhibitors, 5-lipoxygenase inhibitors, cromolyn sodium or nedocromil sodium, anticholinergics, theophyllines, b2-agonists, oral or intravenous corticosteroids, or treatments for corticosteroid-resistant asthma (methotrexate, gold salts, cyclosporin or troleandomycin)

9. Use of b-adrenergic antagonist medications (e.g., propranolol)

10. Smoking within 2 years of Visit 1 or history of smoking ³ 10 pack years

11. Clinically significant abnormality on 12-lead ECG at Visit 1

12. Abnormal chest X-ray (excluding changes consistent with asthma) within the last 12 months of Visit 1

13. Significant systemic disease, or a history of such disease (e.g., cancer, infection, hematological, renal, hepatic, coronary heart disease or other cardiovascular diseases, endocrinologic, or gastrointestinal disease) within the previous 3 months

14. Clinically significant laboratory abnormalities and evaluations at Visit 1

15. Treatment with an experimental, non-approved drug or investigational drugs within the past 30 days

Comments

The diagnosis of asthma in association with skin test reactivity to common allergens restricts the population to those with allergic asthma. The standards of the American Thoracic Society as referred to in inclusion criteria generally describe definitions for diagnosis and classification. The exclusion criteria were extensive. They narrowed the population to those effectively treated with a narrow spectrum of medications, which is a concern for both the safety and efficacy evaluation. The excluded medications do not define well-recognized pathogenetic subtypes of allergic asthma.

According to the NHLBI Guidelines, based on FEV₁ alone these subjects would be expected to have moderate persistent or severe persistent asthma. However, the protocol did not limit the selection of subjects to those with a history of moderate to severe symptomalogy, those with a history of frequent medical care, or those with need for additional concomitant medications beyond medium doses of corticosteroids. In this sense, the protocol was not designed to study severely ill asthmatics.

Procedures and evaluations

Recognition of asthma exacerbations and management of exacerbations

Recognition of worsenings of asthma and their management were critical to this trial. Subjects were to notify their investigator for evaluation for any of the following

-- worsening of asthma at any time requiring an urgent (unscheduled) visit for medical

care

-- PEFR <50% of patient’s personal best -- a decrease in morning PEFR of ³20% on ³2 of 3 successive days, compared to the lowest morning PEFR in the week prior to Visit 3

-- a ³50% increase in 24-hour rescue medication use on ³2 of 3 successive days,

compared to the last week prior to randomization (must exceed 8 puffs)

-- ³2 of 3 successive nights with awakenings due to asthma symptoms requiring rescue

medication

Subjects were to monitor their PEFR at home twice a day. In addition, the protocol instructed investigators to evaluate a subject for treatment if a subject were to have a ³20% decrement in FEV₁ compared to the baseline measurement at visit 3. Subjects were to be treated as deemed appropriate by the investigator.

Comments

The criteria for subjects to notify investigators were rather complex. Case report forms, filled out at the treatment site, required that the site check whether one or more of these criteria was met. In addition, an “other” criterion was allowed. This is reasonable, since asthma exacerbations are not fully described by the criteria listed. The primary endpoint of the trial was defined not by these criteria, but by the doubling of prescribed inhaled corticosteroids or the institution of oral corticosteroids.

Treatment guidelines were provided that were minor modifications of the NHLBI Guidelines (Figure 3).

Figure 3. Treatment of asthma exacerbations during stable steroid phase

Comment

The above guidelines differ from the NIH guidelines in that the protocol guidelines recommend corticosteroid administration for more severe episodes, i.e., doubling inhaled corticosteroid dosing for the “moderate episode” category, while the NIH guidelines recommend this step in the case of a “mild episode,” and addition of oral corticosteroids for severe episodes, while the NIH guidelines recommend them for moderate and severe episodes. The protocol guidelines include the step of intravenous corticosteroids for severe episodes, while the published guidelines do not mention this step explicitly, instead recommending proceeding to an emergency department. These differences are minor, and were guidelines only, not with binding effect.

Screening (visit 1)

The screening period was for the 7 weeks prior to the evaluation period.

Run-in (visits 2.1-2.3, every 2 weeks for 4-6 weeks)

In order to standardize corticosteroid treatment, at visit 2 all subjects were switched to inhaler treatment twice a day with BDP, 84 mg/puff, at a total dose comparable to the corticosteroid the subject had been on prior to the trial. The dose was kept the same or adjusted at week 2 of the run-in period to achieve a level of symptoms and PEFR “acceptable” to the subject and investigator. The total time a subject was to be on a stable dose was 4 weeks; if an adjustment were to be required, the length of the run-in period would be extended to 6 weeks. At the end of the run-in period, patients who continued to be symptomatic (mean total symptom score ³ 3 on the last 14 days prior to Visit 3) would be randomized.

Vital signs, spirometry, adverse experiences, and concomitant medications were determined during this period.

Stable steroid phase, visits 3-7, day 0- week 12

Subjects were to remain on the dose of BDP established during the run-in period. If an exacerbation were to occur, after treatment the subject was to return to this baseline dose.

Starting at day 0 and every 4 weeks the following (selected) procedures occurred:

--physical examination

--review and collection of diary cards—data included symptom scores, twice-daily peak expiratory flow, rescue albuterol use, BDP use, and rhinitis symptoms

--spirometry

--filling out asthma exacerbation form

Subjects on the every 2-week schedule required additional clinic visits for injections. They received the following at the intervening visits:

--review and collection of diary cards as above

--spirometry

Comment

The schedule of collection of blood for serum chemistries and hematologies was such that transient laboratory effects, i.e., those of duration less than 16 weeks, would be missed.

Assessment (“core”) period, steroid reduction phase, visits 7-13, weeks 12-28

During this period, steroid reduction was to be attempted, with biweekly clinic visits for all subjects and telephone calls. Reduction rules were:

· Every 2 weeks, steroids were to be reduced by approximately 25%.

· If the subject were to meet any of the worsening criteria as defined in the discussion of recognition and management of asthma exacerbations, the dose of BDP was to be increased by 25% (or more if deemed necessary by the investigator), and albuterol rescue given. The pre-exacerbation dose of BDP could be instituted after control of asthma symptoms, at the discretion of the investigator.

· Subjects were to remain on the lowest dose of BDP tolerated without one of the above sentinel events for a minimum of 4 weeks prior to the end of the steroid reduction phase.

· Subjects unable to tolerate a steroid reduction were to remain at their baseline BDP dose.

The same procedures occurred as were performed in the stable steroid phase.

Extension period, visits 13-19, weeks 28-52

Immediately following the steroid reduction phase, subjects could enter a double-blind extension. Subjects were to be continued on BDP, but treated at an “optimal” dose as determined by the investigator. In addition there was to be no restriction on the use of concomitant medications.

During the extension period visits continued every 2 weeks for those subjects on the every-2 week treatment schedule and every 4 weeks otherwise.

Follow-up, visit 20, week 64

During the follow-up period, no trial treatment or rescue medication was administered.

Comment

For the sake of brevity this briefing document does not mention various safety-related determinations and procedures that were done.

Analytical plan

Endpoints

Exacerbations of asthma were the cornerstone of the efficacy evaluation. An asthma exacerbation was defined as a worsening of asthma requiring treatment with oral or intravenous corticosteroids or a doubling of the inhaled beclomethasone dose from baseline.

The pivotal analysis was performed on asthma exacerbations occurring during the stable steroid and steroid reductions periods of the trial; no inferential statistics were performed during the extension phase, when exacerbations were defined slightly differently (a doubling of dose of corticosteroid was defined in the extension in relation to the dose immediately preceding the exacerbation, not in relation to the baseline dose).

Efficacy variables were defined for both the stable steroid and steroid reduction phases of the core period.

· Primary endpoints:

-- number of exacerbations during the steroid reduction phase

-- number of exacerbations during the stable steroid phase

· Secondary endpoints:

During stable steroid phase

1. number of subject experiencing at least one exacerbation

2. number of puffs of rescue medication

During steroid reduction phase

1. proportion of patients with successful reduction (³50% dose reduced) of the dose of

BDP

2. proportion of subjects with complete withdrawal of the dose of BDP

3. percent reduction in the dose of BDP

4. number of subjects with at least one asthma exacerbation

5. global evaluation of treatment effectiveness

· Exploratory variables (during stable steroid phase only):

1. asthma-free days

An asthma-free day was defined as a day on which all of the following are met:

-- morning PEFR ³90% of baseline (mean of the 14 days prior to randomization)

-- daytime asthma score £1

-- nighttime asthma score =0

-- rescue medication use £2 puffs

2. morning PEFR

3. evening PEFR

4. difference between morning and previous evening PEFR

5. FEV₁

6. forced vital capacity, FVC

7. forced expiratory flow in the middle 50% of expiration, FEF_25-75

8. total asthma symptom score (nocturnal + daytime + morning score)

9. nocturnal asthma symptom score

10. presence/absence of morning asthma score

11. daytime asthma symptom score

· Other variables

--Change from baseline in adult and pediatric asthma quality of life score (activity limitations,

symptoms, emotional function, overall)

--pharmacoeconomic variables

Comments

Endpoints were reasonable in that they were to collect clinically meaningful measures in asthma. Efficacy variables pertaining to corticosteroid reduction were repetitive, measuring slightly different aspects of the same effect. In addition, these reduction variables were to be determined only after a period of use of omalizumab. Physiological variables of greatest interest were FEV₁ and peak flow.

The determination of efficacy during the core period was appropriately separated from the evaluation of asthma exacerbations in the double-blind extension period of the trial, as concomitant medications and corticosteroid use were to be different in the two periods (liberalized in the extension). The determination of efficacy in the extension phase would be expected to be more complicated than that determined under more restrictive conditions. It would more closely resemble conditions of real use, but it would come after a period of use of the investigational agent and a steroid reduction phase immediately preceding it.

Analytical populations

The primary analysis was to be performed on subjects grouped by randomized treatment who had received at least one dose of trial treatment.

Comment

This was not a true intent-to-treat, but since all randomized subjects received at least one dose in either trial, this resulted in a true intent-to-treat analysis.

Sample size

Genentech set the sample size of 550 to achieve sample sizes of 500 after dropouts, which they calculated would give them 92% and 86% power during the steroid reduction and stabilization periods, respectively.

Summary of statistical methods

· Primary endpoint

A stepwise, conditional analysis of the two phases of the core period was to be performed. The steroid reduction phase was to be analyzed first, but only if <10% of subjects dropped out of the trial during the stable steroid phase. If the statistical criterion (p-value of 0.05 on a 2-tailed test) were met for analysis of the steroid reduction phase, the analysis would proceed for the stable steroid phase. If there were >10% dropouts during the stable steroid phase, only the stable steroid phase would be statistically analyzed.

The primary analysis was to be a between-treatment group analysis performed using the Cochran-Mantel-Haenszel (CMH) statistic stratified by treatment schedule using the standardized mid rank to assign weights to the counts. The null hypothesis was to be tested on the mean score location shift. Most of the secondary analyses (except for peak flow and spirometry) were to be performed using the CMH statistic stratified by treatment schedule. Tertiary endpoint analytical techniques were not specified

The primary endpoint analysis included imputations for subjects who discontinued prematurely. The imputation technique was as follows:

· For subjects who discontinued during a phase, the number of exacerbations attributed to the subject during that phase was the number experienced + the number of days remaining in the period divided by 14. This number was rounded to the nearest integer.

· For subjects who discontinued during the stable steroid phase, exacerbations were attributed during the steroid reduction phase. The number of exacerbations attributed during the steroid reduction phase was the maximum observed for any subject during the steroid reduction phase + 1.

· Other endpoints

Missing data from diaries (BDP use, peak flows, symptom data) and spirometry data were not to be imputed. The pretreatment average from the last 14 days prior to visit 3 (baseline) was used as the baseline for diary data, except for BDP use, where the pretreatment, visit 3 dose was considered baseline. Visit 3 pretreatment values were baseline for spirometry.

· Interim analysis

There was to be no interim analysis. The efficacy analysis was to be initiated at the completion of the steroid reduction phase, prior to completion of the extension period. The protocol states that the results would be unknown to individuals monitoring the trial.

Comments

The imputation technique for the stabilization phase was a worst-case method that was based on the average duration of asthma exacerbations during Genentech’s trial Q0694g (14 days). It represented a highly unlikely series of events, that is, an exacerbation for each 14-day period remaining in a subject’s time in the trial after discontinuation. It would have the effect in the analysis of highly weighting individuals who discontinued early in the trial. The observed maximum number of exacerbations in the stabilization period was about ½ the average of 1 every 2 weeks (it was 3, or an average of 1 every 4 weeks). For further discussion, see CBER’s analysis of the primary endpoint.

Unblinding of results during the analysis of the core period had the potential of biasing the conduct of that period. Data on any potential unblinding was not provided in the submission.

Protocol modifications

Protocol modifications were minor and would not have affected the assessment of efficacy.

Comments

The trials had an adequate duration and collected clinically meaningful data. They were sized to measure two aspects of the product’s effect: reduction in asthma exacerbation incidence, and the effect on corticosteroid dosing. The protocols made no effort to collect information on subjects selected for high degrees of severity or those on commonly used concomitant medications in asthma. Changes made to the protocols after they were implemented would be expected to have no appreciable impact on the results of the trial.

Results: Conduct of the trial

Dates of the trials

The first subject was recruited into trial 008 on February 9, 1998, and the last subject completed the trial on January 19, 2000. The first subject was recruited into trial 009 on April 26, 1998, and the last subject completed the trial on May 24, 2000.

Screening failures

Just over twice as many subjects were screened as entered into trial 008: 1117 vs. 525. The major reasons for failing to be entered into the trial were: FEV₁ over 80% predicted (152 persons, 14%) and serum IgE>700 (101 persons, 9%). Notably, 53 (4.7%) were screened out for IgE below the treatment limit, and 36 (3.2%) were screened out for having a combination of IgE and body mass outside the dosing table limits.

Out of 1356 subjects screened for trial 009, 810 failed to meet selection criteria. The major reasons for failing to be entered into trial 009 were serum IgE>700 (162 persons, 12%) and FEV₁ over 80% predicted (161 persons, 12%). Notably, 71 (4.7%) were screened out for IgE below the treatment limit, and 21 (1.5%) were screened out for having a combination of IgE and body mass outside the dosing table limits.

Comments

A substantial proportion of subjects were screened out of both of these trials due to serum IgE levels or FEV₁ higher than the limit for the trials; a smaller, and notable numbers were screened out due to IgE levels lower than the protocol limit. These IgE-related exclusions are important because of the potential variability in serum IgE. In clinical practice, patients might be tested at multiple times, and qualify on the basis of one of many determinations.

Enrollment by site

There were 26 sites in trial 008. There were 42 sites in trial 009. No single site dominated the enrollment in either trial.

Demographics and baseline characteristics

Trial 008

Table 8 shows that in trial 008 the placebo and omalizumab groups were well matched for important baseline characteristics and demographics. There were about 1½ times as many females as males in the trial (but the proportions were approximately equal in trial 009). The trial population was primarily Caucasian. The trial was primarily adult, with 7-8% adolescents. Investigators uniformly answered the question about whether subjects were going to be exposed to a relevant allergen in the affirmative.

Genentech prospectively defined two classifications for asthma: those whose FEV₁ percent predicted at visit 3 was £65% and who had an average symptom score during the 14 days prior to visit 3 that was >4, and all others. Using this stratification, the two treatment arms were well matched (22% in the more severe category in the omalizumab group vs. 21% in the more severe category in the placebo group).

Table 8. Trial 008: Demographics and baseline characteristics

Omalizumab

N=268

Placebo

N=257

Sex, N (%)

Male

Female

104 (38.8)

164 (61.2)

111 (43.2)

146 (56.8)

Race, N (%)

Caucasian

Black

Other

238 (88.8)

21 (7.8)

9 (3.4)

229 (89.1)

16 (6.2)

12 (4.7)

Age group, N (%)

12 - 17 years

18 - 64 years

³65 years

20 (7.5)

241 (89.9)

7 (2.6)

21 (8.2)

229(89.1)

7 (2.7)

Mean Age, year

(range)

39.3

(12-73)

(12-74)

Mean duration of asthma, year

(range)

20.56

(1 – 61)

22.65

(2 – 60)

Smoking status (n, %)

Never smoked

Ex-smoker

204 (76.1)

64 (23.9)

181 (70.4)

76 (29.6)

BDP dose at baseline visit, mcg/day

(range)

570

(420 – 1008)

568

(336 – 840)

Mean serum total IgE, IU/ml

(range)

172

(20 – 860)

186

(21 – 702)

Mean serum total IgE, IU/ml

by treatment schedule

Q2w: 292

Q4w: 95

Q2w: 314

Q4w: 103

Mean FEV₁, % predicted

(range)

68.2

(30 – 112)

67.7

(32 – 111)

Mean qualifying FEV₁

reversibility, (%)

26.9

25.9

Subjects with hospitalization for asthma

treatment past year, N (%)

6 (2)

11 (4)

Mean number of emergency

room visits for asthma

past year

0.2

0.3

Mean number of doctor’s office visits for

urgent asthma treatment

past year

0.7

0.8

Trial 009

Table 9 shows that in trial 009 the placebo and omalizumab groups were also well matched for important baseline characteristics and demographics. In contrast to trial 008, the proportions of each gender enrolled was nearly matched. The proportion of Caucasians and adults (the large majorities) was similar to that of trial 008. Investigators uniformly answered the question about whether subjects were going to be exposed to a relevant allergen in the affirmative.

Using the stratification of severity described for trial 008 the two treatment arms were well matched and very similar to those in trial 008 (22% in the more severe category in both treatment arms).

Table 9. Trial 009: Demographic and baseline characteristics

Omalizumab

n=274

Placebo

n=272

Sex n(%)

male

female

141 (51.5)

133 (48.5)

127 (46.7)

145 (53.3)

Race n(%)

Caucasian

Black

Oriental

Other

256 (93.4)

11 (4.0)

2 (0.7)

5 (1.8)

242 (89.0)

11 (4.0)

6 (2.2)

13 (4.8)

Age n(%)

12-17years

18-64years

³65years

18 (6.6)

237 (86.5)

19 (6.9)

17 (6.3)

246 (90.4)

9 (3.3)

Age (years)

mean (range)

40.0 (12-76)

39.0 (12-72)

Duration of asthma (yrs.)

mean (range)

20.3 (2-68)

19.1 (1-63)

Smoking status [n(%)]

non-smoker

ex-smoker

213 (77.7)

61 (22.3)

207 (76.1)

65 (23.9)

BDP dose (µg/day)

mean (range)

769 (500-1600)

772 (200-2000)

Mean serum total IgE, IU/ml

by treatment schedule

Q2w: 358

Q4w: 107

Q2w: 338

Q4w: 98

% predicted FEV₁

mean (range)

69.8 (30-112)

69.9 (22-109)

Qualifying FEV₁ reversibility (%)

Mean (range)

26.4 (10-86)

25.8 (11-103)

Past year hospital or doctor visits

for asthma:

Subjects with overnight hospital admission n (%)

Mean number of emergency room visits (range)

Mean number of doctor’s office visits (range)

Mean number of missed work or school days (range)

11 (4.1)

0.23 (0-12)

1.18 (0-15)

4.34 (0-190)

20 (7.5)

0.17 (0-6)

1.21 (0-24)

2.82 (0-60)

Comments

Trials 008 and 009 enrolled a primarily adult Caucasian population of asthmatics of whom very few were in the geriatric age range.

The proportions of subjects with hospital admissions and emergency room visits was low. In addition, subjects in this trial were able to be managed solely with inhaled corticosteroid, not requiring oral (ingested) corticosteroids, without other asthma medications. Subjects whose asthma is difficult to control were not studied in these trials.

Demographic and baseline characteristics were well matched.

Premature discontinuations

Trial 008

Table 10 shows the numbers of subjects who completed trial 008, and the reasons for dropping out. Reasons for failure to complete the core period were fairly balanced, with the exception of “unsatisfactory therapeutic effect,” which was cited as a reason in noticeably more placebo subjects than omalizumab subjects. Withdrawal of consent was the most frequent reason for premature discontinuation during the core period, occurring somewhat more frequently in placebo (7 active, 11 placebo); reasons cited were similar for both groups. Administrative problems predominated reasons for failure to complete the extension period.

Although not directly pertinent to the evaluation of efficacy, it should be noted that more subjects completed their follow-up examination than completed the extension phase. This was due to the fact that early discontinuers were instructed to have a follow-up examination.

Table 10. Trial 008: Subject disposition [n (%)]

Total no. patients, n (%)	Omalizumab	Placebo
Double blind 7 months core period
Randomized	268	257
Competed stabilization	255 (95%)	234 (91%)
Completed steroid reduction	249 (93%)	223 (87%)
Discontinued	19 (7.1%)	34 (13.2%)
due to death	0	1 (0.4%)
due to AE	2 (0.7%)	2 (0.8%)
due to unsatisfactory therapy	1 (0.4%)	14 (5.4%)
due to protocol violation	1 (0.4%)	0
due to consent withdrawal	7 (2.6%)	11 (4.3%)
due to administrative problem	4 (1.5%)	2 (0.8%)
lost to follow-up	4 (1.5%)	4 (1.6%)
Double blind 5 month extension period
Completed core study but did not enter extension study	4 (1.5%)	8 (3.1%)
Enrolled in extension	245 (91.4%)	215 (83.7%)
Completed extension	233 (86.9%)	207 (80.5%)
Discontinued	12 (4.5%)	8 (3.1%)
due to administrative problem	8 (3.0%)	1 (0.4%)
due to consent withdrawal	1 (0.4%)	5 (1.9%)
due to unsatisfactory therapy	1 (0.4%)	1 (0.4%)
lost to follow-up	2 (0.7%)	1 (0.4%)
Three month no treatment follow-up period
Completed extension and completed follow-up period	231 (86.1%)	203 (79.0%)
Discontinued from extension study but completed follow-up period	4 (0.1%)	5 (0.2%)
Completed core study, did not enter extension but completed follow-up period	3 (0.1%)	6 (0.2%)
Discontinued from core study but completed follow-up period	3 (0.1%)	15 (5.9%)
Lost to follow-up	27 (10%)	28 (10.8%)

Trial 009

Table 11 shows the numbers of subjects who completed trial 009. As in trial 008, more placebo subjects discontinued. More placebo subjects discontinued for lack of therapeutic effect as in 008, but the discrepancy between placebo and omalizumab was not as great. Reasons for withdrawal of consent were primarily not illness-related.

As in trial 08, more subjects completed their follow-up examination than completed the extension phase: 22 active and 37 placebo subjects did not complete the follow-up examination.

Table 11. Trial 009: Subject disposition [n(%)]

Total no. patients, n (%)	Omalizumab	Placebo
Double blind 7 months core period
Randomized	274	272
Competed stabilization	261 (95.3%)	245 (90.1)
Completed steroid reduction	255 (93.1%)	232 (85.3%)
Discontinued	19 (6.9%)	40 (14.7%)
due to death	0	0
due to AE	0	5 (1.8%)
due to unsatisfactory therapy	3 (1.1%)	8 (2.9%)
due to protocol violation	5 (1.8%)	6 (2.2%)
due to consent withdrawal	3 (1.1%)	14 (5.1%)
due to administrative problem	1 (0.4%)	4 (1.5%)
due to abnormal lab value	1 (0.4%)	0
lost to follow-up	6 (2.2%)	3 (1.1%)
Double blind 5 month extension period
Completed core study but did not enter extension study	1 (0.4%)	3 (1.1%)
Entered into extension	254 (92.7%)	229 (84.2%)
Completed extension	244 (89.1%)	203 (74.6%)
Discontinued	10 (3.6%)	26 (9.6%)
due to adverse event	2 (0.7%)	3 (1.1%)
due to abnormal lab value	1 (0.4%)	1 (0.4%)
due to unsatisfactory therapy	0	3 (1.1%)
due to consent withdrawal	4 (1.5%)	12 (4.4%)
due to lost to follow-up	2 (0.7%)	7 (2.6%)
due to protocol violation	1 (0.4%)	0
Provided any follow-up data	252 (92.0%)	235 (86.4%)
Three month no treatment follow-up period
Completed extension and completed follow-up period	243 (88.7%)	200 (73.5%)
Discontinued from extension period but completed follow-up period	4 (1.5%)	16 (5.9%)
Completed core period, did not enter extension but completed follow-up period	1 (0.4%)	2 (0.7%)
Discontinued from core period but completed follow-up period	4 (1.5%)	17 (6.3%)
No final follow-up visit	22 (8.0%)	37 (13.6%)

Comments

The proportion of discontinuations in the placebo arm during the stable steroid phase of each trial was close to 10%, and a further 4-5% during the steroid reduction phase; corresponding reductions in the omalizumab group were around 5% and 2%. These two factors, the imbalance of the proportions of discontinuations and the relatively large proportion of these subjects to the total in each treatment arm, created a noticable difference in the magnitude of the treatment effect as analyzed by the protocol-defined technique and by other techniques (see analysis of efficacy results).

Dropouts during the extension phase, whose efficacy results will be reviewed in a separate section from the stable steroid and steroid reduction primary analyses, occurred at about a further 3-4% rate for each treatment arm during trial 008, but were noticeably imbalanced during trial 009 (rates of about 3-4% in the omalizumab arm and 10-11% in the placebo arm). The efficacy evaluations during this phase of the trial were not considered primary.

Eligibility, dosing, and other protocol violations

Table 12 and Table 13 show frequent and important protocol violations for the core and extension periods of trials 008 and 009. Violations were summarized separately for the core and extension periods of trial 009; some of the subjects may have been the same as in the core period, potentially leading to redundancy in the table. However, the numbers of violations in the extension period was small, so the error would not be great.

Table 12. Trial 008: Most common protocol violations during core and extension periods (proportion of subjects affected per group*)

Violation	Omalizumab N=268	Placebo N=257
Run-in period <4 weeks	45 (17)	49 (19)
Beta agonist <6 hours before spirometry	32**	64**
Excluded concomitant med**	22	30
Run in period stable BDP dose 3 to <4 weeks	15 (5.6)	15 (5.8)
Reduced BDP dose in last 4 weeks of reduction phase	14 (5.2)	13 (5.1)
Baseline serum IgE <30	18 (6.7)	7 (2.7)
Dosing error (missed or extra dose)	7**	16**
Baseline mean symptom score <3	10 (3.7)	15 (5.8)
IgE/weight out of dosing table range	7 (2.6)	3 (1.2)
FEV₁ % >80%	6 (2.2)	2 (0.8)

* Protocol violations not listed by subject in submission; calculated by CBER when likely to be single occurrence; proportions expressed as proportions of enrolled population

**possible multiple occurrences; incidence by subject not calculated by CBER

Table 13. Trial 009: Most common protocol violations during core and extension periods (proportion of subjects affected per group)

Period	Violation	Omalizumab	Placebo
Core		N=274	N=272
	Run-in period stable BDP dose 3 to <4 weeks	75 (27)	75 (28)
	Beta agonist <4 hours before spirometry	58 (17)	104 (26)
	Run-in period <4 weeks	39 (14)	33 (12)
	Baseline mean symptom score <3	46 (17)	39 (14)
	IgE/body weight outside dosing table range and dosing >=0.007 mg/kg/IU/ml twice weekly***	23 (8.4)	23 (8.5)
	Excluded concomitant medication	14 (4.7)	22 (6.3)
	FEV₁% >80%	12 (4.4)	5 (1.8)
	Reduced BDP dose in last 4 weeks of reduction phase	5 (1.8)	9 (3.3)
	FEV₁% <40%	11 (4.0)	1 (0.4)
Extension		N=254	N=229
	Beta agonist <4 hours before spirometry	12 (5)	6 (3)
	Missing or additional dose	3 (1)	1 (0.4)

· For both trials, run-in period duration violations were relatively frequent but were balanced between arms.

· Rescue medication violations prior to spirometry were not well balanced and might have affected FEV₁, a secondary endpoint, biasing against the product.

· Unstable doses of corticosteroid at the beginning of the trial would be expected to create more variability in the corticosteroid endpoint of the trial; the number of these violations was balanced.

· Reductions in corticosteroid dosing too close to the end of the steroid reduction phase were infrequent and approximately equal in occurrence in the two treatment arms.

· Excluded concomitant medications violations were infrequent and reasonably balanced.

· FEV₁criteria violations in trial 009 were unbalanced, but relatively infrequent, and occurred in opposite directions in a balanced way.

· Genentech expressed dosing violations for mean dose as numbers of subjects with mean dose less than 0.008 mg/kg/IU/ml every 2 weeks. For trial 008, 5 active-treated subjects and 0 placebo subjects fell into this category; for trial 009, 1 subject in each treatment arm fell into this category.

Comments

Violations of run-in periods were the most common protocol violation in both trials. A too-brief run-in period could potentially have introduced uncertainty in the corticosteroid dosing at the start of the steroid stabilization period. The most problematic outcome of this would be to introduce variability in efficacy measures, but the effect would likely be equal since the violation was equally distributed between treatment arms and was in the same direction (too short a period for both arms). Thus this frequent violation would not have been expected to have an impact in the assessment of efficacy. Other violations with potential impact on the efficacy results of the trials were uncommon, so the impact on the overall results would have been expected to be small. Overall, protocol violations were unlikely to have affected the assessment of efficacy in either trial.

Exposure to product

Compliance to trial agent administration according to protocol requirements was very good, with a reasonable balance between the treatment arms.

Completeness of data collection

Data collection was very good in trials 008 and 009.

Comments on conduct of the trial

In summary, the conduct of the trial was good, and will allow justifiable conclusions from a review of efficacy data as presented.

Results: Efficacy

Primary endpoint

The primary analytical population for these trials was all randomized subjects who received trial medication. Since all subjects received at least one injection, this was equivalent to an intent-to-treat population.

Primary endpoint: asthma exacerbations during stable steroid phase

Trial 008

Table 14 shows exacerbations as determined in trial 008 using the protocol-defined technique of imputation. The CMH test stratified by dosing frequency favored omalizumab with a p-value of 0.006.

Table 14. Trial 008: Asthma exacerbations in stable steroid phase (subjects, %)*

	Q2w dosing		Q4w dosing		Overall
Number of exacerbations	Omlzmb n=106	Placebo n=101	Omlzmb n=162	Placebo n=156	Omlzmb n=268	Placebo n=257
0	87	75	142	122	229	197
	82%	74%	88%	78%	85%	77%
1	11	9	16	23	27	32
	10%	9%	10%	15%	10%	13%
>1	8	17	4	11	12	28
	8%	17%	2%	7%	4%	11%
total ³1	19	26	20	34	39	60
	18%	26%	12%	22%	15%	23%

*ITT population; imputation according to protocol

Trial 009

Table 15 shows exacerbations as determined in trial 009 using the protocol-defined technique of imputation. The CMH test stratified by dosing frequency favored omalizumab with a p-value of <0.001.

Table 15. Trial 009: Asthma exacerbations in stable steroid phase (subjects, %)*

Number of exacerbations	Q2w dosing		Q4w dosing		Overall
Number of exacerbations	Omlzmb n=127	Placebo n=122	Omlzmb n=147	Placebo n=150	Omlzmb n=274	Placebo n=272
0	112	87	127	102	239	189
	88%	71%	86%	68%	87%	69%
1	9	19	16	30	25	49
	7%	16%	11%	20%	9%	18%
>1	6	16	4	18	10	34
	5%	13%	3%	12%	4%	13%
total ³1	15	35	20	48	35	83
	12%	29%	14%	32%	13%	31%

*ITT population; imputation according to protocol

Comments

Most subjects in both groups in both trials experienced no exacerbations (see also comment under the review of different imputation analyses), so the treatment effect was seen in a relatively small number of subjects. There was a greater placebo rate of exacerbations in trial 009, leading to a greater treatment effect. Omalizumab was effective in reducing the numbers of subjects with exacerbations in both treatment schedules.

Primary endpoint: asthma exacerbations during steroid reduction phase

Trial 008

Table 16 shows exacerbations as counted using the protocol-defined technique of imputation. The CMH test stratified by dosing frequency favored omalizumab with a p-value of 0.003.

Table 16. Trial 008: Asthma exacerbations in steroid reduction phase (subjects, %)*

Number of exacerbations	Q2w dosing		Q4w dosing		overall
Number of exacerbations	Omlzmb n=106	Placebo n=101	Omlzmb n=162	Placebo n=156	Omlzmb n=268	Placebo n=257
0	85	58	126	116	211	174
	80%	57%	78%	74%	79%	68%
1	9	23	25	17	34	40
	9%	23%	15%	11%	13%	16%
>1	12	20	11	23	23	43
	11%	20%	7%	15%	9%	17%
total ³1	21	43	36	40	57	83
	20%	43%	22%	26%	21%	32%

*ITT population; imputation according to protocol

Trial 009

Table 17 shows exacerbations as counted using the protocol-defined technique of imputation. The CMH test stratified by dosing frequency favored omalizumab with a p-value of <0.001.

Table 17. Trial 009: Asthma exacerbations in steroid reduction phase (subjects, %)*

Number

of exacerbations

Q2w dosing

Q4w dosing

overall

Omlzmb

n=127

Placebo n=122

Omlzmb

n=147

Placebo n=150

Omlzmb

n=274

Placebo n=272