• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Vaccines, Blood & Biologics

  • Print
  • Share
  • E-mail

Appendix to Summary Basis of Approval - Aralast

Return to Aralast product page

 

I. Summary of Clinical Review

  1. Alpha Therapeutics submitted one clinical trial in support of licensure.
  2. The co-primary endpoints of the trial have been met.
  3. The product has an acceptable safety profile.
  4. The correlation between the A1PI antigenic assay and the functional assay (anti-neutrophil elastase capacity is nominally different in test and control at weeks 8 to 11. This is not thought to be of clinical significance.
  5. The vast majority of subjects maintained trough levels > 11 µmol/L for both the functional assay (anti-neutrophil elastase capacity) and antigenic assay for α1-PI .
  6. The data from the bronchoalveolar lavage are not complete. This is addressed in labeling and in a Phase 4 study the sponsor has agreed to perform.
  7. Product review issues raised by CBER during review of the initial submission resulted in a complete response letter to be issued during the initial review cycle. The latter also included requests for the sponsor to clarify certain clinical data.

Disease and Regulatory Background

Alpha 1-proteinase inhibitor (A1-PI) deficiency

Alpha 1-proteinase inhibitor (A1-PI, alpha one antitrypsin) deficiency is an autosomal co-dominant disorder in which there are below normal levels of alpha 1-proteinase inhibitor in serum and in the epithelial lining fluid (ELF) of the lung. Patients with this disorder have a high risk for the development of emphysema in the third to fifth decades. Some patients also develop liver disease, which is not thought to be ameliorated by augmentation therapy with alpha1-P.

Alpha 1-proteinase inhibitor, which is synthesized in the liver, is a glycoprotein of molecular mass 52kD. It is a serine protease inhibitor (serapin) that has the primary function of inhibiting neutrophil elastase. It is postulated that emphysema results from the imbalance between the neutrophil elastase in the lung that has the capability to destroy elastin of the alveoli and the α1-PI that is responsible for protecting the lung from the elastase.

The Alpha 1-proteinase inhibitor gene has been found to code for over 70 types of α1-PI proteins and the various subtypes of α1-PI proteins are classified on the basis of their electrophoretic motilities. Subjects with various subtypes will have differing serum levels of α1-PI .

PhenotypeMolecular/cellular defectSerum level of α1-PI
( µM/L)
Lung disease may be presentLiver disease
Pi MMNormal21-34NoNo
Pi ZZProduced in liver but cannot be transported extracellularly< 5YesYes
Pi null-nullDeletion of allele0YesNo
Pi SSIncreased intracellular protein degeneration in the liver10-20NoNo
Pi SZCombination of S and Z defects6-14YesNo
Pi MZHeterozygous normal and Z10-20NoNo

The measurement of serum levels of α1-PI is complicated by the fact that many of the commercial standards overestimate the α1-PI levels by 35-40%. In the 1980's the α1-PI levels had been expressed in the units of mg/dL. Levels of 80 mg/dL were speculated to be a threshold serum level, above which there is sufficient α1-PI to protect the lung and below which there is an increase risk of emphysema. It should be noted that this rather arbitrary choice of 11 microM as a serum level target of augmentation therapy neglects the observation by Brantly and colleagues at the University of Florida that the lungs of α1-PI -deficient individuals have an increased neutrophil burden compared to normals, and thus might require more aggressive augmentation therapy than that implied by the traditional 11 microM serum level target for antigenic α1-PI ).

Investigators now use a "true laboratory standard" that more accurately quantifies the amount of α1-PI present. It is important to point out that no controlled clinical trial data exist that substantiate the notion that maintaining a trough serum level of α1-PI of 11 µM/L with augmentation therapy necessarily has a salutary effect on the progression of emphysema/deterioration of pulmonary function.

American Thoracic Society guidelines published in the American Review of Respiratory Diseases in 1989 recommend consideration of α1-PI as augmentation therapy for patients greater than 18 years of age with evidence of ventilatory impairment upon pulmonary function testing. The therapy is based on the concept that increasing the level of α1-PI in the serum will increase the level of α1-PI in the lungs, inhibit the activity of neutrophil elastase, and prevent the damage to the lung structure.

Data from the NIH α1-PI Registry (American J Respiratory Crit. Care Med 158, 49,1998), a non-randomized, epidemiologic prospective study, suggests that augmentation therapy may be efficacious in halting the progression of disease in the subgroup of patients who have an FEV1 between 35-49% predicted. However it should be noted that the primary endpoint evaluating the slope of FEV1 change in the subgroups of the entire study population who were (ever) on augmentation therapy vs. the slope in those who were never on augmentation therapy with α1-PI did not show a difference. Although one cannot draw definitive causal inferences from any epidemiologic study, it should also be noted that the NIH Registry did suggest a possible mortality benefit associated with augmentation therapy. Researchers in the field have voiced concern that baseline imbalances in socioeconomic and other factors not fully taken into account in the analyses of this study may have led to an overestimate of possible augmentation therapy effects on mortality.

The only heretofore U.S.-licensed α1-PI concentrate, Prolastin, produced by Bayer Pharmaceuticals, was licensed in 1986. The clinical data from the Phase 3 trial supporting licensure included serum results from 19 α1-PI deficient subjects of the ZZ subtype. The treatment modality of 60mg/kg was shown to maintain plasma α1-PI levels, measured both antigenically and functionally, above 80 mg/dl in all subjects with a trough level of 120 to 130 mg/dl reached after 6 weeks. Fiberoptic bronchoscopy was performed on a total of nine subjects during the phase 2 and 3 trials. The mean antigenic level of α1-PI in ELF was 0.042+0.025 µM prior to treatment, 1.78+0.51 µM after 2 months or less of weekly infusions, 1.41+0.89 µM after 3 to 4 months of treatment. Functional anti-elastase activities were 0.77+0.036 µM prior to treatment, 1.80+0.51 µM after 2 months or less of weekly infusions, and 1.12+0.27 µM after 3 to 4 months of treatment.

Rationale for maintenance of trough level of 11 µM /L as an "accepted" surrogate measure of clinical efficacy:

An FDA Blood Products Advisory Committee (BPAC) Meeting was held in June 1998 to consider appropriate clinical trial design and endpoints for α1-PI studies. BPAC members voted 11 to 3 with 1 abstention to indicate that FDA should continue to accept maintenance of a plasma level of 11 µM /L in conjunction with demonstration of an appropriately defined increment in epithelial lining fluid as sufficient for demonstrating clinical efficacy of intravenously administered α1-PI products in pivotal phase 3 studies. However, a majority of the committee members had reservations concerning the validity and scientific basis for the conventional target trough level. Committee members were concerned that studies to validate this target level would be difficult to conduct because of the sample size required to demonstrate clinical efficacy, either by decrease in FEV1, CT changes, or death. The committee members did not specifically comment about what the increment in ELF should be and Dr. Brantly of the National Heart, Lung, and Blood Institute commented (page 151 of BPAC minutes) that bronchoalveolar lavage was a technically difficult procedure. He stated that, "As an individual's lung function begins to deteriorate, the ability to do successful and high quality BAL goes down, and there is a higher chance of inaccuracy as far as biochemical data. I think it would be very difficult to get accurate information in individuals, in my experience, that have FEV1's less than 50% of predicted."

II. Regulatory history

IND ---- was originally submitted in June 1993.
The first subject began treatment on March 18, 1997 under the clinical protocol dated December 12, 1996.

There were several protocol amendments:

  • Amendment 16 April 7, 1997:
    • Co-primary endpoints were changed to:
      • Maintenance of mean serum α1-PI trough at weeks 7 through 10 equivalent to 80% of Prolastin treated subjects.
      • Maintenance of mean serum α1-PI trough levels during weeks 11-24 when all subjects are receiving infusions of test material.
  • Amendment 20 November 7, 1997:
    • Inclusion of subjects who had previously received treatment with Prolastin provided that there was a 6 month wash-out period prior to enrollment.

     

  • Amendment 27 November 19, 1998
    • Statistical analysis plan clarifying the statistical methods to be used to evaluate the co-primary endpoints.
  • Amendment 28 December 10, 1998
    • Revision of statistical plan
  • Amendment 32 January 28,1999
    • Notice of unblinding
    • Submission of final analysis plan

May 2000
Alpha Therapeutics submitted a BLA that was designated as "refuse to file" by FDA due to product (CMC) review issues.

September 2000
Alpha Therapeutics applied for, and was granted, fast track status.

August 2001
Current BLA submission submitted.

March 2002
Complete Review letter issued to firm by CBER

June 2002
Sponsor submitted response to complete review letter.

October through December 2002
Sponsor submitted amendments to its response to CBER's complete review letter, the results of a targeted efficacy data audit, and revised draft labeling.

Study design

Clinical trial ATC 97-01 was a multicenter, randomized, double blind, active control Phase 3 non-inferiority study comparing Alpha Therapeutic Corporation's Alpha 1-Proteinase Inhibitor to commercially available Prolastin.
The overall objective was to demonstrate safety and efficacy in subjects with α1-PI deficiency.

Twenty-four subjects were to be randomized to receive either treatment or control product at the standard dose of 60mg/kg/week for 10 consecutive weeks.
Following the 10 weekly infusions the 12 subjects who were receiving Prolastin were to be switched to Alpha Theraputic's α1-PI test article and the 12 subjects who were receiving test article were to continue to receive the test article until week 24. During weeks 11-24 mean serum α1-PI trough levels and AE's were to be continued to be determined.

FIGURE 1: SCHEMATIC PARALLEL DESIGN FOR ATC 97-01

SCHEMATIC PARALLEL DESIGN FOR ATC 97-01

The primary objectives would be:

  • Demonstrate equivalence between treatment groups in mean serum α1-PI trough levels during weeks 7 through 10
  • Determine that serum α1-PI trough levels for all subjects are maintained during weeks 11 to 24.

The secondary objectives would be:

  • Demonstrate equivalence between treatment groups in mean serum trough anti-neutrophil elastase (NE) capacity during weeks 7 through 10
  • Demonstrate equivalence in mean change from base line to 6 weeks of α1-PI
  • Demonstrate equivalence in mean change from base line to 6 weeks of anti-NE capacity
  • Demonstrate equivalence in trough serum and ELF α1-PI and anti-NE capacity done at baseline and 7 weeks

Other assessments:

  • Pulmonary function tests
  • Chest x-rays
  • Biochemical determination of degradation products
  • Pharmacokinetic data

III. Products used

Test product:
Alpha 1 Proteinase Inhibitor [α1-PI (SD)] is manufactured by Alpha Therapeutic Corporation. The source plasma for α1-PI is tested by ---- of the plasma pools for markers of hepatitis B and C, and HIV1 and 2. The manufacturing process includes solvent/detergent treatment and nanofiltration to achieve viral reduction. The product is supplied as a sterile lyophilized powder in vials containing approximately 500 mg or 1000 mg of α1-PI. Each vial contains the labeled amount of functionally active α1-PI in mg/vial, as determined by capacity to neutralize porcine pancreatic elastase. Each vial is reconstituted with Sterile Water for Injection (SWFI) and prepared according to the lot-specific instructions provided by the manufacturer. The administration of product must begin within 3 hours after the product is reconstituted.

Control product:
The control product was Prolastin, the licensed product for the indication of augmentation of α1-PI deficiency. The product is supplied as a sterile lyophilized powder in vials containing 500 mg or 1000 mg of α1-PI. Each vial contains the labeled amount of functionally active α1-PI in mg/vial as determined by capacity to neutralize porcine pancreatic elastase. Each vial is reconstituted with Sterile Water for Injection (SWFI) and prepared according to the lot-specific instructions provided by the manufacturer. The administration of product must begin within 3 hours after the product is reconstituted.

Study procedures (refer to Tables 1 and 2 following this section)
Testing Schedule

At enrollment:

  1. Genetic screening
  2. Medical history and physical exam
  3. Blood chemistry -ALT, AST, alkaline phosphatase, total bilirubin, LDH, creatinine albumin, BUN
  4. Hematology-CBC including differential WBC, platelets
  5. Serum α1-PI level
  6. Serum Anti-Neutrophil elastase capacity
  7. Antibody determination to α1-PI
  8. Viral serology: Antibodies to hepatitis A and C, the presence of circulating HbsAg, antibodies to HIV-1, HIV-2 and Parvovirus B-19
  9. Urine Desmosine/Isodesmosine
  10. Lung CT scan
  11. Chest x-ray
  12. Full pulmonary function tests-spirometry, pre-and post-bronchodilator, lung volume measurements, DLCO, carboxyhemoglobin
  13. Arterial Blood Gas-resting, room air
  14. Bronchoalveolar lavage

Serum α1-PI level and Serum Anti-Neutrophil elastase capacity trough (prior to each infusion) and peak (one hour following each treatment infusion) were to be determined weekly for weeks 1-6.

Serum α1-PI level and Serum Anti-Neutrophil elastase capacity trough were to be determined weekly for each treatment week through week 24 and at day 7 of the last week of Months 9,12,18 and 24 of the extended treatment and follow-up.
Urine desmosine and isodesmosine were to be collected at enrollment and once per week during the first 24 weeks of enrollment and the last week of months 9, 12, 18, and 24 of extended treatment and follow-up.

Blood biochemistry and hematology were to be done every 3 weeks until week 22 and again at week 24.

Chest x-rays were to be performed at enrollment, week 7 and months 12 and 24.
Lung CT scan was to be performed at enrollment and month 12. All x-ray studies were to be read locally at each institution and then to be reviewed independently at one institution.

Pulmonary function tests were to be performed at weeks 7 and 24 and during the last weeks of months 12 and 24.

Biologic Half-life (t½) of α1-PI was to be obtained at the time of the first infusion for both test and control products by blood sampling at pre-infusion, and post infusion at 1h +0.25h, 2h+0.25h, and 6h+0.5h, and 12h+ 0.5h and then every 24 hr + 6h through day 7 prior to the next infusion

A1-PI antibody determination was to be done at enrollment, prior to infusion 7 and prior to infusion 24. Assays were to be conducted in the laboratory of Dr. Mark Brantly.

Viral serology performed at enrollment were antibodies to hepatitis A (IgG/IgM and IgM confirmatory), antibodies to hepatitis C, HbsAg, antibodies to HIV-1, HIV-2 and Parvovirus B-19 (IgM and IgG). All subjects who were negative for Parvovirus B-19 at enrollment were to be tested for Parvovirus B-19 by polymerase chain reaction (PCR) prior to Week 2. The same viral serology panel, excluding the Parvovirus B-19, was to be performed prior to treatment at Week 11 and at Week 24. An additional aliquot of serum was to be retained from all viral sampling points for retesting. Analysis of the viral serology for Weeks 24 through 96 will be submitted in a supplemental report.

Bronchoalveolar Lavage (BAL)
All evaluable subjects were to undergo a BAL at baseline preceding infusion 1 and at week 7, six to 7 days after infusion 6.
(BAL) fluid for each lobe was processed separately (there is a contradiction in the lab methods section 16) and percentage recovered was to be measured.
BALs were considered to be evaluable if they met the following criteria, which were defined by the testing laboratory prior to treatment of the first subject:

  • Return > 20%
  • Cells/mL > 5.0 x 104
  • [Urea]plasma/[Urea]BAL < 300
  • 90 nmol/L < Initial [α1-PI]BAL < 600 nmol/L
  • The epithelial lining fluid (ELF) from each BAL was to be analyzed for the presence of α1-PI level and anti-NE activity and the change from baseline in the 2 treatment groups was to be compared.

TABLE 1: SCHEDULE OF TREATMENTS AND TESTING

 

WEEK/INFUSION #M-1Month 1Month 2Month 3Month 4Month 5
E-4*1*234567*891011121314151617181920
PhenotypingX                    
Medical History and Physical ExamX§X§XXXXXX§XXXXXXXXXXXXX
Blood Biochemistry and Cell CountsXX  X  X  X  X  X  X 
Serum α1-PI LevelsXXXXXXXXXXXXXXXXXXXXX
Serum Anti-NE CapacityXXXXXXXXXXXXXXXXXXXXX
Urine Desmosine/ IsodesmosineXXXXXXXXXXXXXXXXXXXXX
Lung CT ScanX                    
Chest X-RayX      X             
Full Pulmonary Function TestsX      X             
Arterial Blood GasX                    
Bronchoalveolar Lavage(BAL)@X      X             
Biologic t½ Determination X                   
α1-PI Antibody DeterminationX      X             
Viral SerologyX X        X         
α1-PI Infusion and Vital Signs XXXXXXXXXXXXXXXXXXXX
Adverse Experiences XXXXXXXXXXXXXXXXXXXX

E (Week -4 or M-1) = Randomization and enrollment into the study to obtain baseline values for up to 4 weeks prior to infusion 1 (week 1).
@BALs will be performed at enrollment and 6 days (defined as 6 to 7 days) after the sixth consecutive infusion. Epithelial lining fluid (ELF) will be assessed for levels of α1-PI and Anti-NE capacity.
Week 2 viral serology will be for Parvovirus B-19 and only for subjects who were Parvovirus B-19 negative at enrollment.
All subjects will receive α1-PI (SD) weekly at a dose of 60 mg/kg body weight beginning at Week 11.

TABLE 2: SCHEDULE OF TREATMENTS AND TESTING

WEEK/INFUSION #Month 6 EXTENDED TREATMENT/FOLLOW-UP MONTH #
21222324*912*1824*
Medical History and Physical ExamXXXXXX§XX§
Blood Biochemistry and Cell Counts X  XXXX
Serum α1-PI LevelsXXXXXXXX
Serum Anti-NE CapacityXXXXXXXX
Urine Desmosine/ IsodesmosineXXXXXXXX
Lung CT Scan     X X
Chest X-Ray     X X
Full Pulmonary Function Tests   X X X
Arterial Blood Gas       X
α1-PI Antibody Determination   X    
Viral Serology   X X X
α1-PI Infusion and Vital SignsXXXXXXXX
Adverse ExperiencesXXXXXXXX
^ Throughout the extended treatment and follow-up period (Months 7 to 24), all subjects will receive α1-PI (SD) weekly at a dose of 60 mg/kg body weight. Vital signs and adverse experiences will be monitored during each extended treatment infusion.
§ For weekly infusions during Extended Treatment and Follow-up Months 7, 8, 10, 11, 13 to 17 and 19 to 23, modified physical exams, vital signs and adverse experiences will be be obtained.

Planned inclusion criteria were:

  1. Greater than 18 years old
  2. Diagnosis of congenital α1-PI deficiency with any of the following combination of alleles: PiZZ, Mheerian, Mprocida, Mmalton, Plowell, Mduarte, Pduarte, Mmineral springs, Wbethesda, Mnichinan, I and all Z null, null- null and null in combination with any of the above alleles.
  3. Diagnosis of emphysema
  4. Serum α1-PI level less than 11µM
  5. Following bronchodilators:

    1. An initial FEV1 > 30% and < 80% of predicted and an initial FEV1/FVC ratio <70%
    2. Or if the initial FEV1 was >80% of predicted, a DLCO < 70% of predicted, plus an abnormal lung CT consistent with emphysema, and no other confounding disease present.
  6. Non-smoker and had not smoked for at least 6 months prior to enrollment
  7. Female subject and partner agree to use adequate contraception
  8. Agree to the terms for treatment and collection of follow-up data as scheduled in the protocol
  9. Able to sign consent form

Planned exclusion criteria:

  1. Any α1-PI augmentation therapy, either Prolastin or other investigational α1-PI product within the preceding 6 months
  2. Hospitalization >2x in the past year for lung-related problems or within 4 weeks of enrollment for pneumonia
  3. Ongoing recurrent inflammatory process such as:

    1. Diffuse infiltrative parenchymal lung disease as evidenced by lung CT scan and shown to be active by gallium scan or other clincial evidence
    2. Antibiotics within 2 weeks prior to the first infusion or received > 6 courses of antibiotic therapy during the preceding year
    3. Demonstrated inability to achieve a constant dose of steroids on a tapering regimen.
  4. Seropositivity for HbsAg or HCV IgG Antibody
  5. Seropositivity for antibody to HIV-1 and/or HIV-2
  6. ALT or AST > 3x the upper limit of normal in the preceding 6 months
  7. PaCO2 > 46mm Hg or a resting room air PaO2 < 55mmHg
  8. Serum creatinine >1.5 X ULN
  9. Selective IgA deficiency (IgA < 15mg/dl) or antibody against IgA
  10. Antibodies against α1-PI
  11. Pregnant or nursing a child
  12. Received investigational drug within 2 months of the trial or currently receiving investigational drug.

Randomization
Randomization was to be performed at the site pharmacy from a series of randomization numbers provided by Alpha Therapeutic Corporation. A separate series of randomization numbers was provided at each treatment site to allow for randomization by site.

Blinding
For the first 24 weeks of the study the subjects, investigators, radiologists, testing laboratories, on-site study nurses and home healthcare nurses treating the subjects were to be blinded regarding the subjects first 10 week treatment group and to the identity of the solutions for infusion.

The pharmacist was not to be blinded
At Alpha Therapeutic Corporation, the clinical manager, director of clinical research, statistician and anyone involved in the analysis or interpretation of the data were to be blinded regarding the subject's treatment group.

Study procedures:
Infusions of test or product were to be performed weekly.
The medication was to be reconstituted by the study site pharmacy and administered by a home health care nurse, who assessed the subject for AE's and completed the CRF.

III. Statistical Analysis

There were 2 co-primary endpoints:
 

  1. Mean serum α1-PI trough levels from weeks 8-11 for the test group were not inferior to those of the control group.

    Prior to breaking the blind, missing data for Weeks 8, 9, 10 and 11 will be imputed by taking the mean across the remaining weeks.
    A one-sided 5% level Sasabuchi t-test will be performed of the null hypothesis that the mean for test group is less than 80% of the mean for the control group.

  2. Maintenance of mean serum α1-PI trough levels during Weeks 12 through 24 while receiving weekly test article will be analyzed by computing a regression slope for each subject
    The regression slope will be fit to all available data and missing data will not be attributed. The A two-sided 90% confidence interval for the mean slope will be computed to demonstrate that the lower limit of the observed slope did not include ­0.1µmol/L/week.
    The analysis will be done separately for the subjects who were transferred at Week 11 from control to test drug in order to test for interference due to carryover effect.

Secondary endpoints:

  1. Mean serum anti-NE capacity from weeks 8 through 11 for the test group is not inferior to that of the control group, will be performed. The methods used will be the same as for the primary endpoint above.
  2. Mean change in serum α1-PI from baseline to week 7
    A test that the mean change for serum α1-PI from baseline to trough levels at week 7 for the test group is not inferior to that for the control group will be performed using the method of Sasabuchi.
  3. Mean change in serum anti-NE from baseline to week 7
    A test that the mean change for serum anti-NE from baseline to trough levels at week 7 for the test group is not inferior to that for the control group will be performed using the method of Sasabuchi.

Independent analysis done by CBER medical reviewers included::

  1. Confirmed line listings of inclusion criteria including pulmonary function tests (PFTs), reports of CT scans, and chest x-rays.
  2. Performed independent review of clinical data to ascertain that the subjects were comparable in the clinical characteristics of pulmonary disease and concomitant medications.
  3. Spot-checked data to note consistencies, reviewed areas of missing data
  4. Reviewed AE reports and classification of AEs, assessed AEs by lot numbers
  5. Reviewed BAL protocol and raw data to assess for deviations of sampling
  6. Independent analysis by CBER statistician:

    1. Confirmed analysis using imputation of data
    2. Performed analysis of first primary endpoint without imputation.
    3. Determined correlation between measurements of antigenic and functional α1-PI during weeks 1-7 and weeks 8-11 in test and control product.

IV.Clinical trial conduct

The study was initiated on February 19, 1997 and the last infusion #24 was given on August 5, 1998.

The study was opened at 5 sites. Four sites accrued a total of 28 subjects as follows:

InvestigatorSite ID NumberSiteNumber of subjects enrolled at site
Jack Clausen, MD01UC San Diego6
James Stoller, MD03Cleveland Clinic10
James Stocks, MD04U of Texas, Tyler8
Edward Campbell, MD05U of Utah, Salt Lake City4

According to the sponsor, the study was unblinded sometime between December 15, 1998 and December 24, 1998 after a discussion with FDA held December 15, 1998. The study sites were kept blinded until this date. The analysis plan was finalized prior to unblinding of the study.

Disposition of subjects:

Design of ATC 97-01

Disposition of subjects
No subjects died during the study.
Two subjects, one each from the test and control groups withdrew from the study.
Subject 301 on Prolastin discontinued the study at week one after one infusion. The subject experienced bilateral pneumonia after a bronchoscopy to remove a foreign body.
Subject 501 elected to discontinue the study after 6 infusions because she felt that the obligations of the study were too burdensome. Subject 105 experienced bloating after his weekly infusions and withdrew from the trial after 28 infusions. This is noted in the annual report submitted Sept. 2001 and is not in the BLA report.
According to the 2001 annual report 15 subjects (53.6 %) completed the planned 2-year enrollment (104 weeks). Ten subjects were terminated at weeks 78-95 due to lack of available product.

Study conduct:

Study conduct concerning the administration of medication, monitoring of subjects, collection of serum samples and clinical data appears to have been adequate. However the conduct of the study had a flaw concerning adherence to the BAL protocol. In contrast to the planned BAL protocol in which evaluations were to be performed on all samples, evaluations were done only on paired samples.

Violations and protocol deviations:

20 serum specimens were either lost or damaged by overnight courier. (See table in data analysis section)

Violations concerning both groups:

Prior to April 1998, both test and control study drug, was administered up to 24 hours following reconstitution despite the protocol instruction that it should be given within 3 hours.

Violations concerning test group:

Doses in six instances were 58 mg/kg instead of the 60mg/kg due to transcription errors

Violations concerning control group:

Due to a pharmacy error, Subject 402 received 37.8mg/kg instead of 60mg/kg for infusions 16-18 while on test product.

Protocol deviation:

Subject 104 in the test group stopped his antibiotics on 12/07/97 and the sponsor approved his participation in the study so that he received his first infusion on 12/11/97.

Inspection of study sites for integrity of data- See Bioresearch Monitoring (BIMO) report.

Good Clinical Practice Inspections were conducted at the central laboratory and clinical investigator sites at the University of Florida, as well as the clinical investigator sites at the Cleveland Clinic and University of Texas, the 2 clinical sites that had the highest enrollment,

Subject demographics:

There were 14 subjects enrolled in each treatment group. All subjects were of the ZZ phenotype with the exception of one subject in the Control group who was MmaltonZ.
Baseline serum α1-PI levels were comparable in the 2 groups.
The groups were balanced in all aspects below except for history of smoking.

Three of the 14 subjects in the test group that had not smoked but all subjects in the control group had a history of smoking.

Table 5: Baseline Characteristics - All Subjects

CharacteristicTest group
(n=14)
Control group
(n=14)
p
n%n%
Mean + SD
(Range)
Mean + SD
(Range)
Sex
    Male1071.41178.6 
    Female428.6321.41.000a
Race
    White14100.014100.0--
Age (years)45 + 11
(24 to 64)
49 + 7
(40 to 62)
0.319b
Weight (kg)84 + 12
(57 to 106)
91 + 15
(71 to 116)
0.199b
Height (cm)177 + 6
(166 to 187)
177 + 9
(163 to 193)
0.860b
Serum α1-PI ( µmol/L)5.7 + 1.0
(4.1 to 7.2)
5.7 + 1.3
(4.1 to 8.2)
0.929b
FEV1 (% predicted)48 + 18
(31 to 86)
47 + 15
(30 to 77)
0.835b
Phenotype
    ZZ14100.01392.91.000a
    MmaltonZ00.017.1
History of smoking
    Yes1178.614100.00.222a
    No321.400.0
Years since quitting12 + 10
(< 1 to 31)
14 + 9
(1 to 35)
0.605b
Vaccinated against Hepatitis A
    Yes17.100.01.000a
    No1392.914100.0
Vaccinated against Hepatitis B
    Yes964.3535.70.257a
    No535.7964.3
SD - standard deviation
a Groups compared with Fisher's exact test
b Groups compared with two-sided t-test

All subjects were Caucasian and over 70% of subjects were males in each group.
All subjects except 305 in the test group met the inclusion criteria for FEV1 between 30 and 80%. This subject had an FEV1 of 86% predicted and a DLCO of 68% predicted. CT scan was read as normal

Subject 105 in the test group was the only subject in the study that had been treated with Prolastin previously. This had been more than 2 years prior to enrollment in this study.

Date of Diagnosis of Illness

 Date of diagnosis
I. EmphysemaII. A1PI deficiency
Test1980-May 19971980- September 1997
ControlJuly 1980-October 1997July 1980-December 1997

Concomitant medications

All subjects enrolled had the diagnosis of "emphysema" and the majority of subjects were on inhaled or oral steroids throughout the course of the study.

Only two subjects in the test group, 104 and 407 were not receiving any therapy for their lung disease.

One subject in the control group, 101, did not receive any pulmonary therapy and 2 subjects, 304 and 503, received only bronchodilators without any steroid therapy.

During the course of the study, 14 subjects required antibiotic treatment for exacerbation of pulmonary infection. Two subjects required 3 courses of Antibiotics and 2 subjects required 2 courses of antibiotics, making for a total of 20 courses of Antibiotics. Of these, 10 courses occurred during the blinded phase of the study, 4 events in the test group and 6 events in the control group.

Efficacy Analysis

Co-primary endpoint 1

A test that the co-primary endpoint, mean serum α1-PI trough level from Weeks 8 through 11, for the test group is not inferior to that for the control group was performed.

Four subjects (two from each treatment group) were missing data for one or more of the four weeks. No subject missed an infusion, but some samples were lost or damaged. As agreed upon with the FDA, for each of these four subjects, the missing value(s) were assigned the value of the mean of weeks 8-11 for which data were available for that subject.

Group assignments:
Subject 305-test group
Subject 309-control group
Subject 310 Test group
Subject 408-Control group

Table 67: Missing data for Serum α1-PI Levels,
Weeks 8 through 11
SubjectWeek Number
891011
030515.015.015.314.6
030916.916.919.014.8
031016.115.316.017.0
040819.920.021.118.9

The null hypothesis tested was that the mean of the test group is less than 80% of the mean for the control group.

Mean serum α1-PI trough level of each group measured prior to treatment at Weeks 8 through 11

Test group             15.3 ± 2.5 µmol/L
Control group             16.9 ± 2.34 µmol/L

Mean test = 90.5%
Mean control

Lower 95% confidence limit for Mean test / Mean control = 81.7%

(p = 0.026, one-sided Sasabuchi t-test)

The null hypothesis was rejected.

The CBER statistician confirmed this value.

Analysis by CBER statistician conducted without imputation of data for endpoint 1 showed

Mean test = 90.7%
Mean control

Lower 95% confidence limit for

Mean test = 84.1%
Mean control

Trough Levels for infusions 7 through 10

V. Co-primary endpoint 2

During weeks 12-24 all subjects received test article. Maintenance of mean serum α1-PI trough levels during these weeks was analyzed by computing a regression slope for each subject employing a first-order time series autoregressive model. A two-sided 90% confidence interval for the mean slope was computed to demonstrate that the lower limit of the observed slope did not include ­0.1 µmol/L/week.

Twelve subjects (six in the test group and six in the control group) were missing α1-PI trough level data for one of the 13 infusions; one subject (Subject 0403 in the control group) was missing data for two infusions (infusions 13 and 24). These missing values were not attributed; the regression line was fit through the available data for each subject.

The null hypothesis tested was that the slope of the trough in weeks 12-24 included ­0.1 µmol/L/week for either the test or the control group.

To determine if the study drug received during the blinded portion of the study may have influenced the results obtained for all subjects, the slope of the line was calculated separately for the test and control groups.

Slope test group=­0.02490% CI ­0.088 to 0.040
Slope control group=0.01890% CI ­0.043 to 0.080
Slope combined groups=­0.00390% CI ­0.04 to 0.04

The null hypothesis was rejected.

The FDA statistician reviewed the data and concurred.

Figure 4: Trough Levels for Serum α1-PI Mean + SD

Through Levels for Serum α1-PI Mean + SD

Table 68: Missing data for Serum α1-PI Levels,
Weeks 12 through 24
Patient NumberWeek Number
12131415161718192021222324
0104            
0105            X
0304        X    
0307            
0308            
0403 X          
0404            
0405            
0406            
0407            
0502    X        
0503            
0504           

 

S

X = accidentally destroyed by Fed EX Checkmark= missing Fed Ex shipment 4/13/98
 
Table 7: Summary of Co-primary Endpoints
Co-primary EndpointsTest Group
(n = 13)
Control Group
(n = 13)
Ratio test/control
Mean + SD
( µmol/L)
Mean + SD
( µmol/L)
( µTest/
µControl
(%)
Lower limit
95% CI
pa
Serum α1-PI trough levels, Weeks 8 through 11b
    Ho: Mean test group < 80% control group
    H1: Mean test group > 80% control group
15.3 + 2.516.9 + 2.390.581.70.026
 Test Group
(n = 13)
Control Group
(n = 13)
Combined groups
(n = 26)
Slope
( µmol/L/week)
90% CI
Slope
( µmol/L/week)
90% CI
Slope
( µmol/L/week)
90% CI
Serum α1-PI trough levels, Weeks 12 through 24c
    Ho: Slope over time includes -0.1
    H1: Slope over time does not includes -0.1
    Decision: Reject Ho
-0.024
(-0.088 to 0.040)
0.018
(-0.043 to 0.080)
-0.003
(-0.04 to 0.04)
CI = confidence interval; SD = standard deviation.
a One-sided Sasabuchi t-test of Ho.
b Samples drawn prior to infusions at Weeks 8 through 11, reflecting trough levels for infusions 7 through 10.
c Samples drawn prior to infusions at Weeks 12 through 24, reflecting trough levels for infusions 11 through 23. Both groups were receiving test drug during this time period.

Secondary endpoints

  1. Non-inferiority of trough of anti-neutrophil elastase (anti-NE) capacity at weeks 8-11

    Ho: Mean test < 80%
          Mean control

    Mean serum anti-NE trough level of each group measured prior to treatment at Weeks 8 through 11:
    Test group       15.3 ± 2.4 µmol/L
    Control group       15.7± 2.6 µmol/L

    Mean test = 97.5%
    Mean control

    Lower 95% confidence limit for the ratio of test/control of 87.5%
    (p = 0.003, one-sided Sasabuchi t-test)

    The null hypothesis is rejected.
    Analysis confirmed by CBER statistician.

  2. Non-inferiority of α1-PI trough levels change from baseline to week 7

    Ho: Mean test < 80%
          Mean control

    Mean α1-PI trough levels change from baseline to week 7:
    Test group       9.7 ± 3.4 µmol/L
    Control group     11.5 ± 2.3 µmol/L

    Mean test = 84.2%
    Mean control

    Lower 95% confidence limit for the ratio of test/control of 70.2%
    (p = 0.301, one-sided Sasabuchi t-test)

    The null hypothesis is not rejected.
    Analysis confirmed by CBER statistician.

  3. Non-inferiority of anti-NE capacity trough levels change from baseline to week 7

    Ho: Mean test < 80%
          Mean control

    Mean Anti-NE capacity trough levels change from baseline to week 7:
    Test group       12.3 ± 3.0 µmol/L
    Control group       12.3 ± 2.5 µmol/L

    Mean test = 99.6%
    Mean control

    Lower 95% confidence limit for the ratio of test/control of 86.0%
    (p = 0.010, one-sided Sasabuchi t-test)

    The null hypothesis is rejected.
    Analysis confirmed by CBER statistician.

Table 8: Summary of Secondary Endpoints

Secondary EndpointsTest Group
(n = 13)
Control Group
(n = 13)
Ratio test/control
Mean + SD
( µmol/L)
Mean + SD
( µmol/L)
( µTest/
µControl
(%)
Lower limit
95% CI
pa
Serum anti-NE trough levels, Weeks 8 through 11b
    Ho: Mean test group < 80% control group
    H1: Mean test group > 80% control group
Decision: Reject Ho
15.3 + 2.415.7 + 2.697.587.50.003
Serum α1-PI trough levels, change from baseline
(Week 1) to Week 7c
    Ho: Mean test group < 80% control group
    H1: Mean test group > 80% control group
Decision: Do not reject Ho
9.7 + 3.411.5 + 2.384.270.20.301
Serum anti-NE trough levels, change from baseline (Week 1) to week 7c
    Ho: Mean test group < 80% control group
    H1: Mean test group > 80% control group
Decision: Reject Ho
12.3 + 3.012.3 + 2.599.686.00.010

CI = confidence interval; SD = standard deviation.
a One-sided Sasabuchi t-test of Ho.
b Samples drawn prior to infusions at Weeks 8 through 11, reflecting trough levels for infusions 7 through 10.
c Samples drawn prior to infusion at Week 7, reflecting change from baseline through Infusion 6.
 

Individual data concerning α1-PI trough less than 11 µmol/L
Subject 0104 was the only subject in the test group who had α1-PI trough level <11µM after week 3.

Subject 104 α1-PI Trough levels <11µM

Week45810111314
α1-PI trough (µmol/L)
10.710.310.09.99.610.610.6

While the first 6 infusions of this subject were with the same lot, and each infusion was 7 days apart, the variability in trough levels during this period, which included values < 11 microM, was greater than during the remainder of the 24 week observation period. Therefore the variability cannot be explained by lot-to-lot variability and is consistent with assay variability plus biological intra-subject variability.

VI. Subject 104 had Gilbert's syndrome, received amoxicillin and was on Prilosec for the duration of the study. The only other subject in the study who took Prilosec was subject 404 in the control group and this subject's levels were not abnormal.

Omeprazole (Prilosec) may interact with other drugs metabolized by the cytochrome P-450 system and can cause mild and, rarely marked elevations of liver function tests

VII. Serum trough levels fell below the target levels in 2 subjects in the control group while on Prolastin and in subject 306 while on test drug.

Individual data concerning anti-NE capacity trough less than 11 µmol/L

Subject 104 anti-NE capacity trough levels <11µM

Week4571114
anti-NE capacity trough (µmol/L)8.910.610.710.810.6

A few Control Subjects had sporadic anti-NE capacity trough level <11µM after week 3 .

In addition to the above, subject 401 had anti-NE capacity trough levels < 11µM. Due to a pharmacy error, this subject received only 37.8mg/kg/wk of product instead of 60 mg/kg/wk of product during weeks 16,17, and 18.

The trough levels of α1-PI measured both antigenically and functionally were above the level of 11µM in the vast majority of measurements.

Additional CBER analysis of correlation of the assays for total α1-PI and anti NE capacity

The CBER statistician performed an analysis of the correlation of the functional assay (anti-neutrophil elastase capacity) and the antigenic assay for α1-PI. Comparisons were made of the control and test articles only for weeks 1-11 because all subjects received test article after week 11. In weeks 1-7 this correlation is 0.9 in the subjects receiving Prolastin and 0.87 for the subjects receiving test article. For weeks 8-11 the correlation is 0.71 for the subjects receiving Prolastin and 0.47 for the subjects receiving test article.

The lower correlation for test product in weeks 8-11 and the differences in the correlations between the test and control article were the subject of further analysis by the sponsor at the request of FDA. When other time frames were examined, it could not be concluded that any nominal between-product differences in correlation coefficients for antigenic vs. functional α1-PI were of significance.

The sponsor was asked to "analyze, compare, and discuss the relationship between antigenic and functional levels of A1­PI in the final container test product and Prolastin and in serum samples from patients receiving the test product and the control Prolastin. The sponsor was asked to compare the results obtained for the final container test product to results obtained for subject serum samples for each lot of product administered. For subjects enrolled in the control arm of the study and administered test product for the open label phase of the study, the sponsor was asked to analyze and compare the relationship between antigenic and functional levels of A1­PI for both treatment periods."

The sponsor clarified that the A1PI levels for the Prolastin control lots were not assayed by the sponsor.

The sponsor stated that the final container potency values, in g active A1PI per vial, met all final container specs. In the trial, the infusion volume was calculated based on the labeled potency of the specific A1PI lot and the subject's body weight. Because of the adjustment of infusion volume for potency, there is no apparent correlation between potency of functional A1PI in mg/vial to mean functional A1PI trough level in subjects receiving product from a particular vial.

The sponsor compared the subgroup mean (a) antigenic and (b) A1PI levels between weeks 3-11 (reflecting trough levels from infusions 3 to 10) and weeks 13-24 (reflecting trough levels from infusions 13 - 23) for those subjects randomized to receive Prolastin control during the initial 10 weeks and ATC A1PI during weeks 11-24. The sponsor calculated the ratio of antigenic OR functional levels (weeks 4-11)/(weeks 14-24). This ratio was 1.0257 (95% CI 0.9874-1.0652) for antigenic A1PI and was 0.9816 (95% CI 0.945-1.1094) for functional A1PI (anti-neutrophil elastase capacity [ANEC]). Similar ratios and confidence intervals were obtained using log transformed serum values. I calculate a ratio of mean functional to mean antigenic A1PI levels for the Prolastin steady-state period (weeks 4-10) of 15.48/16.61 = 0.93 and for the ATC A1PI period in this Prolastin randomization subgroup 15.78/16.19 = 0.975. Thus, the ratio of functional to antigenic is slightly less for Prolastin than for ATC A1PI by this analysis. It should be noted that the comparison requested by CBER of the ratio of antigenic to functional A1PI in the final container versus subject's plasma should take into account the contribution of the subject's endogenous A1PI. In the case of severely-deficient A1PI patients, this ratio is not expected to be 1.

Analysis of Results of Bronchoalveolar Lavage

Bronchoalveolar lavage (BALs) was performed twice on 26 subjects.

The fluid from each lobe was processed separately according to the BAL worksheet and in contradiction to laboratory methods 16.5.4, which states that the fluid from the lobes was to be combined.

BAL fluid was evaluated for adequacy according to the following criteria:

  • Return > 20%
  • Cells/mL > 5.0 x 104
  • [Urea]plasma/[Urea]BAL < 300
  • 90 nmol/L < Initial [α1-PI]BAL < 600 nmol/L

Cells were separated from the ELF fluid, a cell count was determined, and the fluid was frozen for further analysis. A sample was considered "unpaired" and was not analyzed if the subject's other BAL sample was deemed inadequate.

Only 5 of 13 subjects in the test group (38.5%) and 3 of 13 subjects in the control group (23.1%) had BALs both at baseline and at week 7 that met the criteria for evaluation established in the protocol.

For those subjects with unpaired BALs, the unpaired sample was not analyzed. This occurred in 13 pairs.

Two samples from the Tyler site were not analyzed because no second sample was sent due to inadequacy of the BAL return. Nine of the 10 subjects at the Cleveland Clinic site had unevaluable BAL because of low cell counts.

The degree of emphysema may have affected the quality of a BAL. Seven of 14 (50.0%) subjects with FEV1 > 40% had evaluable BALs while only one of 12 (8.3%) subjects with FEV1< 40% had an evaluable BAL.

The data are difficult to interpret because of the small sample size and the wide variability of the values. Subject 105 in the test group had levels of anti-NE capacity that decreased from 1,532 nmol/L at baseline to undetectable levels at week 7. The sponsor attributed this unexpected drop to the fact that this subject was taking antibiotics for bronchitis at week 1 and also had an infection at week 7.

The sponsor conducted an analysis on both the untransformed data and on loge (natural logarithm) transformed data. In spite of these manipulations of the data, no firm conclusions were drawn from the BAL data from the original submission. The sponsor submitted in June 2002 further analysis using the confidence interval approach on the differences in the six parameters listed below.

  1. ELF α 1-PI change from baseline to week 7

    Mean α 1-PI in ELF at baseline:
    Test group     190 ± 108 nmol/L
    Control group 452 ± 92 nmol/L

    Mean α 1-PI in ELF at week 7:
    Test group 1,294 ± 885 nmol/L
    Control group 1,640 ± 511 nmol/L

    Mean change in α 1-PI in ELF at week 7:
    Test group 1,104 ± 905 nmol/L
    Control group 1,188 ± 432 nmol/L

    There is a small sample size and a very large standard deviation. The difference in the changes between the 2 groups is not statistically significant. (p = 0.888)

  2. ELF Anti-NE levels

    Mean Anti-NE in ELF at baseline:
    Test group 1,086 ± 320 nmol/L
    Control group 737 ± 280 nmol/L

    Mean Anti-NE in ELF at week 7
    Test group 1,635 ± 1,168 nmol/L
    Control group 1,516 ± 839 nmol/L

    Mean change in Anti-NE in ELF at week 7
    Test group 549 ± 1,419 nmol/L
    Control group 779 ± 575 nmol/L

    The large standard deviation in the test group makes the data uninterpretable.

  3. ELF α 1-PI:NE complex levels

    Mean α 1-PI:NE in ELF at baseline:
    Test group 16 ± 16 nmol/L
    Control group 79 ± 33 nmol/L

    Mean α 1-PI:NE in ELF at week 7
    Test group 129 ± 219 nmol/L
    Control group 215 ± 160 nmol/L

    Mean change in α 1-PI :NE in ELF at week 7
    Test group 114 ± 206 nmol/L
    Control group 136 ± 177 nmol/L

    The sponsor states that, "These increases in α 1-PI: NE complexes suggest that the appha1-PI provided by the augmentation therapy was functional and that is it was able to inactivate elastase in the lung tissue."

    However, the differences are not significant, probably because of the large standard deviation, and, as seen below, no significant changes in NE were demonstrated as a result of treatment..

  4. ELF NE levels

    Mean NE in ELF at baseline:
    Test group 405 ± 640 nmol/L
    Control group 149 ± 69 nmol/L

    Mean NE in ELF at week 7:
    Test group 427 ± 849 nmol/L
    Control group 109 ± 94 nmol/L

    Mean change in NE in ELF at week 7:
    Test group 21 ± 215 nmol/L
    Control group -40 ± 60 nmol/L

    Mean ELF NE levels did not change significantly in either the test or control groups.

  5. ELF IL-8 levels

    Mean IL-8 in ELF at baseline:
    Test group 14,316 ± 4,996 ng/mL
    Control group 4,111 ± 1,107 ng/mL

    Mean IL-8 in ELF at week 7:
    Test group 4,012 ± 1,547 ng/mL
    Control group 5,160 ± 3.676 ng/mL

    Mean change in IL-8 in ELF at week 7:
    Test group -10,304 ± 4,558
    Control group 1,048 ± 2,619 ng/mL

    The clinical significance, if any, of the large decrease in the test group mean IL-8 level is uncertain. It was noted that the magnitude of the SD was quite variable.

  6. ELF neutrophil counts

    Mean neutrophil counts in ELF at baseline:
    Test group 12.9 ± 17.4 x 107/mL
    Control group 6.6 ± 4.0 x 107/mL

    Mean neutrophil counts in ELF at week 7:
    Test group 6.7 ± 8.2 x 107/mL
    Control group 5.3 ± 3.7 x 107/mL

    Mean change in neutrophil counts in ELF at week 7:
    Test group -6.2 ± 10.9 x 107/mL
    Control group -1.4 ± 3.5 x 107/mL

    The change from baseline was not statistically significant in either group nor was the comparison of change between groups.

Table 11: Summary ELF Data

Test
UnitsTest group
(n=5)
Control group
(n=3)
Comparisons between groups
Change Week 1 to Week 7
Week 1aWeek 7bpcWeek 1aWeek 7bpbWeek 1
pd
Test groupcontrol grouppe
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
α1-PI
nmol/L190 + 108
(94 to 368)
1,294+885
(361 to 2,495)
0.053452+92
(366 to 550)
1,640+511
(1,056 to 2,006)
0.0410.0131,104+905
(267 to 2,350)
1,188+432
(690 to 1,456)
0.888
Anti-NE
nmol/L1,086+320
(774 to 1,532)
1,635+1,168
(BDLf to 3,101)
0.436737+280
(450 to 1,010)
1,516+839
(566 to 2,156)
0.1440.172549+1,419
(-1,532 to 2,238)
779+575
(116 to 1,146)
0.803
α1-PI:NE
nmol/L16+15
(1 to 41)
129+219
(5 to 519)
0.312g79+33
(48 to 114)
215+160
(72 to 388)
0.3150.009114+206
(-13 to 478)
136+177
(-42 to 313)
0.880
NE
nmol/L405+640
(69 to 1,548)
427+849
(4 to 1,942)
0.836149+69
(81 to 219)
109+94
(37 to 215)
0.250g0.42421+215
(-139 to 393)
-40+61
(-111 to -4)
0.656
IL-8
ng/mL14,316+4,996
(6,963 to 20,448)
4,012+1,547
(2,212 to 6,268)
0.0074,111+1,107
(3,125 to 5,309)
5,160+3,676
(2,663 to 9,381)
1.000g0.01510304 +4,558
(-14,180 to -2,503)
1,048+2,619
(-465 to 4,072)
0.008
Neutrophils
x 107/mL12.9+17.4
(2.7 to 43.6)
6.7+8.2
(0.1 to 20.0)
0.2756.6+4.0
(2.8 to 10.8)
5.3+3.7
(1.3 to 8.5)
0.5710.574-6.2+10.9
(-23.6 to 6.0)
-1.4+3.5
(-4.8 to 2.2)
0.499
BDL = below detectable limits; SD = standard deviation.
a Sample drawn prior to first infusion. Week 1 is considered baseline.
b Sample drawn prior to infusion at Week 7, reflecting trough level from infusion 6.
c Week 1 (preinfusion) compared to Week 7 (preinfusion), two-sided t-test.
d Week 1 (preinfusion), test group compared to control group, two-sided t-test.
e Change from Week 1 (preinfusion) to Week 7 (preinfusion), test group compared to control group, two-sided t-test.
f Treated as zero for calculations.
g Data were not normally distributed, so signed rank test was performed.
Since no adjustments have been made for multiple comparisons, the tests of significance should be considered suggestive only.

Urinary levels of elastase breakdown products

  1. Desmosine
  2. Isodesmosine
  3. Collagen degradation products

The desmosine:lysylpyridinoline ratios for both groups were nearly constant and similar for both groups.

Other categories evaluated

  1. Pulmonary function
    PFT's were performed at screening, week 7 and week 24. There was a great deal of variability in the PFT's and one would not expect a significant change in the PFT's over the 23-week period studied.

Table 20: Pulmonary Function Tests

TestScreeningWeek 7Week 24Analysis of variance
Test
(n = 13a)
Control
(n = 13a)
Test
(n = 13)
Control
(n = 13)
Test
(n = 13)
Control
(n = 13)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Mean + SD
(Range)
Factorpb
FEV149.2 + 17.945.0 + 14.047.2 + 18.645.1 + 14.048.2 + 19.945.5 + 12.4Group0.64
(% predicted)(31 to 86)(30 to 77)(29 to 91)(24 to 77)(28 to 97)(26 to 73)Week0.59
 
FVC93.5 + 24.389.3 + 12.191.5 + 23.989.8 + 13.993.5 + 24.192.0 + 14.0Group0.74
(% predicted)(65 to 144)(68 to 109)(59 to 141)(73 to 124)(63 to 145)(70 to 120)Week0.45
 
FEV1FVCc53.1 + 15.151.2 + 16.950.7 + 17.051.2 + 17.251.9 + 14.950.5 + 16.2Group0.88
(% predicted)(25 to 87)(31 to 85)(26 to 90)(33 to 85)(27 to 87)(35 to 92)Week0.30
 
DLCO61.9 + 15.963.2 + 15.062.2 + 15.562.8 + 19.263.9 + 16.762.1 + 17.0Group1.00
(% predicted)(28 to 80)(43 to 104)(29 to 85)(31 to 107)(28 to 93)(41 to 107)Week0.95

SD = standard deviation.
a Subjects 0501 (test group) and 0301 (control group), who discontinued treatment after six and one infusion(s), respectifely, are not included in this table.
b Probabilities are based on analysis of variance (ƒ statistic); interaction between groups and weeks was not significant.
c Subject 0401 (test group) was missing a value at Week 7; therefore, this subject was excluded from the analysis of variance.

Pharmacokinetics
Summary of FDA pharmacology review

  1. t1/2 of RespitinTM was longer compared to that of Prolastin (the currently licensed α1-PI preparation); t1/2 of RespitinTM = 5.9 +/- 1.2 days; t1/2 of Prolastin = 5.1 +/- 0.5 days.
  2. Bioequivalence for RespitinTM and Prolastin was not demonstrated as C.I. for Tmaxtest/Tmaxcontrol was 0.40 to 1.23 (<0.80)
  3. Differences in other parameters (Tmax, AUC, Cmax, CL, MRT, Vd) between RespitinTM and Prolastin were not statistically significant

Antibiotics:

14 subjects, 7 from each group required antibiotic treatment throughout the course of the study.
One subject in the test group was on continuous tetracycline throughout the study because of a skin condition.
2 subjects required 3 courses of antibiotics, 2 subjects required 2 courses of antibiotics for a total of 20 courses of antibiotics

Other medications for asthma or COPD:

In the test group 2 subjects received no therapy for asthma-104 and 407.
In the control group, 1 subject received no therapy for asthma.
In the test group 3 subjects did not receive steroid therapy, either local or oral for their asthma.
In the control group, 3 subjects did not receive steroid therapy, either local or oral for their asthma.
In the test group 2 subjects did not receive bronchodilators.
In the control group 3 subjects did not receive bronchodilators

Safety analysis (initial 24 week trial period)

All subjects were treated with the same dosage of test and control drugs, 60mg/kg administered intravenously per week. Twenty-six subjects completed all 23 scheduled weekly infusions. Subject 0301 (control group) withdrew after experiencing a serious adverse event (bilateral lower lobe pulmonary infiltrates) following his BAL at Week 1 (after one infusion of control drug). Subject 0501 (test group) withdrew at Week 7 (after receiving six infusions of test drug) due to the extensive follow-up required of study participants.

 Number of InfusionsMaximum infusion rate
(mL/kg/min)
Range of infusion rate
(mL/kg/min)
"Blinded Phase"Test1330.070 ± 0.0130.038 to 0.085
Control1310.072 ± 0.0110.042 to 0.088
Combined PhaseCombined4710.070 ± 0.0130.038 to 0.088

There were 605 infusions during the 24-week portion of the study. 474 of the infusions were of test material and 131 were of the licensed product Prolastin.

Serious adverse events:

Two serious events occurred during the course of the 24-week study phase, both in the control group.
1. Subject 103 was hospitalized after developing pneumonia after an off protocol bronchoscopy which was performed to remove a foreign body.
2. Subject 404 was hospitalized after infusion week 1 due to a severe headache deemed to be secondary to DJD and not product infusions.

Weeks 1 through 10 incidences of adverse events, regardless of causality

 TestControl
Mild1119
Moderate17
Severe02

During Weeks 1 through 10,these AEs were reported in the test group

The following symptom was associated with 3 infusions

Headache

The following symptoms were each associated with one infusion:
Chills and fever (temp to 100.2)
Malaise
Back pain
Dizziness
Pruritus
Rash
Abnormal vision (c/w migrainous visual symptom)
 

In the combined group of all Respitin infusions (all infusions in the test group during Weeks 1 through 23 and all infusions in the control group during Weeks 11 through 23), the frequency of infusions associated with an adverse event, regardless of causality, was 20 of 474 infusions.

The most common symptoms were:

SymptomNumber of infusions
Headache
5(1% of infusions)
Dizziness (includes lightheadedness)
3(0.6% of infusions)
Somnolence
3(0.6% of infusions)
Rash
2(0.4% of infusions)

The following occurred with one infusion (0.2%) each:
Abdominal pain, back pain, chest pain, chills and fever, malaise, vasodilatation (facial flushing), vomiting, leg cramps, pharyngitis, rhinitis, pruritus, sweating, and abnormal vision.

Thirty lots of test product were used for the 474 infusions. The number of infusions/lot varied from 1 to 57. Only 2 lots were associated with more than 1 AE. No lots were associated with more than 2 AEs.
Lots 7011A had 2 AE's: sleepiness and headache.
Lot 7002A had 2 AEs: stomach pain and headache

Laboratory Abnormalities

Hematology values were similar between the test and control groups. Mild hematological abnormalities were frequent but showed no consistent pattern of either elevation or depression of these values.
Thrombocytopenia was the most frequent hematological abnormality seen and was present in some subjects at the time of screening. The lowest platelet count of 78,000 was seen in subject 304 in the control group. This subject also had low hemoglobin.
Chemistry abnormalities in subjects while receiving test article were mild and all were less than 3.7 x ULN. Subject 104 with the diagnosis of Gilbert's Syndrome had persistently elevated transaminases to 2-3 x ULN. Subject 402 had an isolated serum creatinine of 1.7 at week 13 when all other values were within the normal range.
Subject 408 had a creatinine value of 1.3 at week 13 and 1.2 at week 22 with all other values being normal.

Viral safety:

All subjects were both HCV and HIV negative at entry into the study and none had converted at week 23. None of the subjects became HBsAg positive during the study although five of 13 evaluable subjects in the test group and eight of 13 subjects in the control group, were not vaccinated against hepatitis B. None of the 21 subjects who were seronegative to HAV (12 in the test group and 9 in the control group) seroconverted during the study.

One subject in the control group converted to Parvovirus positive at week 11 prior to receiving test product.

Following the CBER medical review of the initial 24-week portion of the study, the sponsor was sent a complete review letter on 8 March 2002. Selected clinical issues from that letter, highlights of the sponsor's responses, and a key aspects of the CBER analysis of the sponsor's responses are summarized below:

  1. Subject 104 had persistently low serum levels of alpha-1-proteinase inhibitor (A1­PI) as determined antigenically and functionally (with the anti-neutrophil elastase capacity assay) during the 24 week study period. The levels of both antigenic and functional A1­PI were below 11µM on 12 occasions. There were additional subjects who had serum levels of either antigenic or functional A1­PI below 11µM on more than one occasion. These . may have been due to subject variability and/or test product lot-to-lot variability. The sponsor was asked to comment.
    This subject's A1PI levels hovered around the 11 micromolar cutoff. While the first 6 infusions of this subject were with the same lot, and each infusion was 7 days apart, the variability in trough levels during this period, which included values < 11 microM, was greater than during the remainder of the 24 week observation period. Therefore the variability cannot be explained by lot-to-lot variability and is consistent with assay variability plus biological intra-subject variability.
  2. Protocol ATC 97-01 requires that endothelial lining fluid (ELF) from all BAL samples meeting certain pre-specified criteria be analyzed. FDA requested the sponsor to calculate confidence intervals for the differences between ELF data for week 1 and week 7 for each of the six measured parameters in the paired samples. These calculations were performed on both the original data and the log transformed data. In addition, this should be done for the population with and without the "outlier" subject, subject 105.
    The sponsor presented the requested confidence intervals on original and log-transformed ELF data in Table 3.1 of the submission (attached to original review of complete review of complete response to complete review letter).
  3. The sponsor was asked to submit all data collected between week 24 and two years as described in Protocol Table 2.
    The sponsor has submitted a supplemental clinical report of study ATC 97-01 covering the extension phase from weeks 24-96 dated 10 Feb 2002.

    Summary: of Safety Update:

    Dates: The first subject entered 19 Feb 1997 and the last terminated 10 Dec 1999. Note that the study was terminated early due to a shortage of Respitin due --------------------------------------------------------------, and Bayer's Prolastin was then substituted for the ATC test article.

    Therapy: Subjects received ATC AIPI IV (Respitin) 60 mg/kg weekly through week 96. Smoking in the 6 months prior to entry was prohibited. Lots used during the extended treatment phase numbered 49. Note: because the supply of Respitin was not adequate, there were times when subjects received Prolastin in the same dose schedule in lieu of ATC Respitin. Sixteen lots of Prolastin were used for this purpose.

    Safety assessments: AEs, hepatic, renal, hematologic function at weeks 36, 48, 72, and 96, viral serology, and vital signs.

    Surrogate efficacy assessments during weeks 24-96: Serum AAT levels, anti-NE capacity, PFTs, urine elastin degradation products

    Sponsor's conclusions from extension phase of study ATC 97-01:

    • Serum levels of A1PI and anti-NE capacity (ANEC) were essentially unchanged from the initial 24 week period of the study.
    • Rates of antibiotic usage and respiratory infections were little changed from baseline.
    • Pulmonary Function Tests (PFTs), urine elastin breakdown products, and radiographic analyses were insensitive measures of efficacy, showing little change over time.
    • Of 2226 Respitin ITT infusions over the whole trial period, 188 (8.4%) were associated with AEs, of which 19 (0.9%) signs or symptoms were considered at least possibly related to the product.
    • Serious Adverse Events (SAEs) were reported, including one case of pericarditis, none considered by the investigator to be product-related.
    • Mild elevations in ALT and AST were common and occasional mild bilirubin elevations were seen. The sponsor found no cause for these, other than one case of Gilbert's syndrome. The sponsor speculated that AAT-deficiency liver disease may have been involved in some affected subjects.
    • Vital sign changes were not considered product-related.
    • One control subject seroconverted for parvovirus B19 (IaG and IgM at week 11. Subjects were screened for parvovirus B19 at baseline and at week 11. A total of 20 subjects were positive for parvovirus B19 at baseline (77%). No other seroconversions to hepatitis A, B, or C or to HIV were detected. A total of 13 subjects were enrolled without vaccination against HBV.
    • No subject developed antibodies to A1PI.
    • The safety profile of ATC Respitin is similar to that of Prolastin. AEs are of low incidence and generally only mild or moderate in intensity.

    Twenty-six subjects completed the 26 week initial study period and are included in the extension study report. Ten subjects (5 from each original randomization group) withdrew during the extension phase, leaving 16 subjects who completed 96 weeks of treatment. Nine of the 10 were withdrawn by the sponsor due to lack of availability of Clinical Trial Material (CTM). These withdrawals occurred after 19.9 to 23.5 months into the study. Subject 0105 withdrew voluntarily 7.8 months after the initial infusion. Prior to withdrawing from the study, 3 study drug infusions were withheld due to bloating. The investigator later determined that the bloating was due to end-stage pulmonary disease "and an erratic pattern of use of prescribed medications." The subject then withdrew for "personal reasons" after this assessment. Five subjects were treated (with Respitin and/or Prolastin) longer than 96 weeks, as provided by the protocol (subjects 0302, 0303, 0401, 0402, 0403, and 0404).

    Protocol violations: 17 infusions were missed. "No subject missed more than three consecutive infusions. A total of 26 infusions in 5 subjects were slightly under or overdosed due to using the wrong body weight, or because of the unavailability of sufficient CTM. Subject 0104 received 50% of the protocol dose for 4 weekly infusions due to product shortage. This under dosage was approved by the IRB.

    Some subjects received product that had been reconstituted in a local pharmacy rather than the subject's home and thus exceeded the 3 hour time limit. PFTs were delayed if a respiratory infection were present.

    Respitin and Prolastin infusions post week 24 were given at very similar infusion rates and durations (Table 19, attached to Medical review). The rate of infusion was a mean of 54 +/- SD 15 min for ATC A1PI (Respitin).

    Comments on efficacy measures during extension phase:

    Serum AAT levels and ANEC were measured at months 9, 12, 18, and 24. See table 8 (attached to review). Levels were maintained with some variation, but drop somewhat from week 72 to week 96 to a level of 14.76 micromol/L for AAT and to 8.64 mi cromol/L for ANEC.

    The sponsor asserted that the observed mild drop in ANEC at end of study was likely due to delays in blood sampling/visit date due in part to unavailability of product. A somewhat larger drop was seen at week 96 for AAT levels

    Urine desmosine and isodesmosine were anticipated to possibly fall with AAT augmentation treatment. No consistent trend over time was observed for urine desmosine:creatinine ratio (Table 9, attached to review) at the U. of Texas lab. An unplanned analysis was also conducted at BU which suggested a slight rise over time in urine desmosine and desmosine:creatinine ratio.

    Antibiotic Courses by subject are presented for weeks 2-24 and weeks 25-96 in table 17 (attached to Medical review). No information is provided as to the proportion of these that were prescribed for lower respiratory infections.

    Comments on safety measures during study extension phase:

    CXR was repeated at weeks 7, 48, and 96. CT was repeated at weeks 48 and 96. Baseline abnormalities in addition to emphysema were comparatively frequent and included bronchiectasis, fibrosis, and nodules. Two subjects had infiltrates (0306 and 0309).

    PFT values and changes over time are given in Tables 13 and 15. Mean DLCO (pre-bronchodilator) actually rose from 61.9 at week 0 to 64.6 at week 96 over the course of the trial, but the variation was considerable (SDs of 15.1 and 17.9, respectively).

    Mean FEV1 fell from 42.54 +/- SD 14.84 to 40.40 +/- 17.55 % of predicted at week 96. A similar change over time was seen in the post-bronchodilator values.

    Vital Capacity dropped from 93.1 to 88.54 % of predicted, but with large SDs.

    ABGs showed a mean drop in PaO2 from 68.85 +/- SD 10.11 at week 0 to 63.44 +/- 14.18 at week 29 (n = 23), suggesting some deterioration over time in gas exchange despite therapy with ATC Respitin and Prolastin.

    PaCO2 rose from a mean of 38.04 (max 81) to 40.61 (max 80).

    The sponsor re-coded all COSTART terms that appeared compatible with URIs as pharyngitis.

    The overall summary of AE totals throughout the study is given in Table 23 (attached). Twenty-six subjects reported AEs, of which 8 reported AE(s) at least possibly related to study drug. Of the 189 AEs reported, 141 (75%) were rated mild, 27 (14%) moderate, and 5 (2.65%) were rated severe. Sixteen were rated "unknown" in intensity.

    Nineteen AEs were regarded at least possibly related to study drug, including 16 mild and 3 moderate. Those rated mild were:

    Headache (3 AEs in 2 subjects)
    Chest pain (1)
    Chills & fever (1)
    Vasodilatation (1)
    Paresthesias (2 AEs in 2 subjects)
    Somnolence (3 episodes in 1 subject)
    Dizziness (1)
    Increased cough (2 AEs in 1 subject)
    Dyspnea (2 AEs in 1 subject)
    Rash (1)
    Pruritis (1)
    Abnormal vision (1)
    Those were regarded at least possibly related to study drug and rated moderate in intensity were:
    Increased cough
    Dyspnea
    Chest pain

    The moderate intensity AEs were all reported in the same subject (#0106).

    No unusual or worrisome pattern was evident among the non-serious AEs reported.

    Five subjects reported 8 AEs during the extension phase, of which none was regarded as related to the study drug. These are listed in Table 27 (attached to Medical review). Two subjects had reported SAEs during the original 26 week portion of the trial. Subject 0302 was hospitalized for 2 days for chest pains attributed to COPD. Subject 0306 was hospitalized 3 times at weeks 59, 85, and 93 for respiratory infections. Subject 0401 was hospitalized at week 68 for pericarditis that developed 5 days after his most recent infusion. His signs and symptoms lasted 49 hours and were considered due to his previously existing medical condition. He had also had pericarditis in 1990.

    AEs by Intent-to-treat (ITT) are displayed by body system in Table 24 of the CR to FDA's CR letter. The sponsor also undertook an ad hoc "per protocol" (misnomer) analysis of AEs by censoring AEs reported after the subject first received during the extension phase the first infusion of Prolastin (due to shortage of ATC A1PI). The number or infusions in the "per protocol" AE analysis divided by the number in the ITT AE analysis is 1127/1799. From this one can gain an impression of the extent of Prolastin use during the extension phase.

    Respiratory infections numbered 28 among 19 subjects during weeks 2-24 and 66 among 24 subjects during weeks 25-96.

    The use of specific categories of concomitant medications tended to decrease slightly over the trial course.

    Laboratory Data:

    Mean hematology values were stable over time. Mean AST rose from 36.8 to 40.4 with wide AD. Mean AST rose from 30.9 at week 26 "baseline" to 32.1 at week 96. Mean alk phos rose from 171.6 to 182.2 at week 96. Serum mean creatinine was stable over time, as was BUN. Mean total bilirubin was stable over time. One subject had elevated eosinophil counts at various time points. No terribly remarkable treatment-emergent hematology abnormalities were noted among those with abnormalities flagged by the sponsor.

    Subject 0101 had ALT AND AST elevations at baseline (1034 and 66, respectively) and on study at all timepoints for unknown reasons. Subject 0102 had a single rise in alk phos to 277 U at week 72. Subject 0104 had ALT elevations of 105, 70, 64, and 48 at weeks 36, 48, 72, and 96. AST was elevated at 82, 56, 52, and 41 units at the same respective time points, but also had elevations at screening (ALT 61 and AST 89). Other subjects had milder AST and/or ALT elevations that were not noted at baseline, and were not of clinical significance, but the vast majority of subjects with abnormal aminotransferases had abnormalities also at baseline. As stated by the sponsor, some of these AST and ALT elevations could be due to previously unrecognized AAT-deficiency-related liver disease.

    Vital Signs

    Some vital changes during infusion were noted in Table 40 but were not judged to be of particular clinical significance.

  4. At FDA's request, the sponsor submitted chest x-ray reports and CT scan reports for all subjects in the study.

    No mention of emphysema was present in 5 of the CT reports, 4 of which were from a single study site. However, the protocol only required the CT exam be compatible with emphysema should the subject not meet the inclusion criteria for pulmonary function

REVIEW OF SPONSOR'S 18 Oct 2002 amendment to the BLA submitted in response to CBER's information request relayed to the sponsor during the teleconference of 20 Sept 2002:

During the September telecon, CBER requested the sponsor undertake a 100% audit of the primary endpoint data relating to antigenic A1PI trough blood levels between weeks 8 and 11 to account for and provide 2 supplementary analyses for subjects who had repeated lab analysis of A1PI levels at the University of Florida's central contract lab. The requested analyses were: to (1) recalculate the primary endpoint using only the first value obtained in any samples analyzed multiple times and (2) recalculate the endpoint using the average of all test values obtained for each sample. ATC undertook a 3rd analysis, which used the results of the repeat value with the last test date.

A total of 130 reported antigenic A1PI values were audited (1 test/week x 5 weeks x 26 subjects). Of these, multiple test results were available for 40 samples (31%). All of those had duplicate values, save for 3 that had 3 test results available per sample. Of the multiple test result values, 23/40 reported the first value and 17/40 reported the 2nd value. Two transcription errors lead to 2 of the latter reports. Reasons recollected by the University of Florida central laboratory staff for the repeat testing were provided by the sponsor.

The results of the re-analyses are shown in Table One (attached to Medical review of CR to CR letter). The supplementary analyses, like the original primary endpoint analysis, reject the null hypothesis and support the biochemical surrogate efficacy of the product.

The sponsor also revised its response to question #2 from the original CR letter, providing correlation coefficients between antigenic and functional A1PI for blood samples from weeks 4-11 separately for both treatment groups. As can be seen in Table 4 (attached to Medical review), no meaningful or statistically significant difference was seen between the correlation coefficients for the ATC product vs. Prolastin. The Pearson correlation coefficients were 0.66 and 0.56 for Prolastin and ATC product, respectively, during weeks 4-11, with overlapping confidence intervals. All subjects' data from weeks 14 -24 ) (ATC product only) yielded a CC of 0.66 (95% CI 0.59 to 0.97).

The data in the sponsor's response to the CBER complete review letter were deemed to support the sponsor's conclusions regarding the efficacy and safety of the product and support the conclusion reached in the CBER review of the original study report from the original filed BLA submission regarding the product having a satisfactory benefit:risk profile. Although somewhat supportive, the sponsor's bronchpulmonary lavage data were not considered conclusive. The sponsor has committed to perform a phase IV study to further verify ELF-related endpoints.