Exubera

1
Exubera Pfizer PharmaceuticalsNew Drug
Application 21-868 Hypoglycemia in Type 1
Patients

• Endocrine and Metabolic Drugs AC Meeting
• Silver Spring, Maryland
• September 8, 2005
• Joy Mele, MS
• FDA Statistical Reviewer
• Division of Biometrics 2

Center for Drug Evaluation and Research
2
Outline of Presentation

• Protocol 107
• Entry criteria
• Collection of the data
• Measures of hypoglycemic risk
• Focusing on severe hypoglycemia in Study 107
• The post-hoc FDA definition of hypoglycemia
• Characteristics of the events
• Conclusion regarding hypoglycemic risk for INH
  versus SC

3
The Protocol and Hypoglycemic Events

• Entry criteria
• No more than 1 severe hypoglycemic event or any
  hospitalization due to poor glycemic control in
  previous 6 months or during run-in
• Two definitions
• Hypoglycemic event
• Severe hypoglycemic event

4
The Protocol and Hypoglycemic Events

• Case Report Form Data
• Symptoms Y/N
• Glucose (several sources)
• Was subject unable to self treat?
• Did subject exhibit CNS symptoms?
• Was blood glucose 49 mg/dL or lower OR
• Did symptoms reverse with CHO?
• Severity Mild, Moderate, Severe

5
Measures of Hypoglycemic Risk

• Percentage of patients with at least one event
• Event rate as total number of events over total
  exposure (e.g. person-months)
• complementary information
• ADA Workgroup, Diabetes Care, May 2005
• BUT may not be enough information.

6
Study 107 Hypoglycemia Pts w/ at least 1
event Events per pt-month
RR 1.0 0.97 1.3
0.89 0.7
2.5
7
Study 107 Distribution of Severe Events

• of pts by total number of events
• 0 1 2 3 4 5 12
• INH 85 10 3 2 1 1 1
• SC 90 8 4 1 0 0 0
• Medians of 0 Wilcoxon rank sum test p0.30
• Recurrent events analysis w/o pt w/12 events,
  decreases estimate from 2.25 to 1.65, NS

8
Study 107 Patients with 3 events (INH 27/43,
63)
9
Study 107 INH Patients with More than 3 Severe
Hypoglycemic Episodes

• 5036-7117 35 yo M with 4 severe episodes (Wks
  3, 10, 16, 24), no clear pattern
• 5102-7639 54 yo F with 5 severe episodes all
  within 1 week no further severe episodes after
  insulin reduced and bedtime snack added
• 5007-7988 19 yo F with 12 severe episodes 9
  episodes in 2 weeks while home over college
  winter break after 7 weeks on INH

10
Conclusion for Severe Hypoglycemia

• One patient with 12 severe events overly
  influences estimates of risk based on survival
  models of recurrent events.
• No significant treatment difference between INH
  and SC in Study 107. Results are consistent with
  those seen in Study 106.

11
FDA-defined Events Glucose ?36 OR unable to
treat oneself

• What characteristics differentiate these events
  from total events?

12
Study 107 FDA-defined Hypoglycemic Events

• 91/103 INH patients with 971 events
• 94/103 SC patients with 1327 events
• INH SC
• Glucose ?36 923 1293
• Unable to treat oneself 39
  20
• Both 9 14

13
Glucose Distributions by Severity Total
Hypoglycemic Events in Study 107
FDA Events Patient Means INH 32.2 SC 31.5

14
Study 107 Severity Scores
FDA events Patient Means INH 1.5 SC 1.5
15
Study 107Symptomatic Hypoglycemic Events
FDA events Patient Means INH 0.83 SC 0.78
16
Study 107 Frequency of FDA-defined Events
17
Study 107 Comparison of FDA-events

• 91/103 INH patients with 971 events
• 94/103 SC patients with 1327 events
• Medians INH 6 events SC 8 events
• Wilcoxon rank sum test p0.09

18
Closing Comments

• Hypoglycemic events are similar in quantity and
  characteristics between the INH-treated patients
  and the SC-treated patients
• Rates of hypoglycemia should not be summarized
  and analyzed based on total events without
  carefully examining the distribution of the
  events across patients. Outliers may grossly
  skew the estimates of risk.