1. 1 Lotronex Postmarketing Experience Ann Corken Mackey, R.Ph., M.P.H. Allen Brinker, M.D., M.S. Zili Li, M.D., M.P.H., formerly of ODS Office of Drug Safety (ODS) April 23, 2002

2. 2 Outline of Presentation Adverse Event Reporting System (AERS) Ischemic Colitis Small Bowel Ischemia Serious Complications of Constipation Deaths Clinical Trials vs Postmarketing Experience Conclusions

3. 3 Adverse Event Reporting System (AERS) Spontaneous, voluntary surveillance system –Voluntary reporting by health care professionals and consumers –Mandatory reporting by manufacturers Over two million reports in database

4. 4 AERS Strengths Provides for early detection of signals Identifies rarely occurring adverse events Captures off-label use, use in patient populations other than those studied, drug combinations, and use in contraindicated conditions

5. 5 AERS Limitations Cannot reliably estimate true incidence rates of events (numerator underestimated and denominator can only be projected) Subject to under-reporting, only 1 to 10% of adverse events reported to FDA 1,2 No certainty that the drug caused the event (may be due to underlying disease, concomitant meds) 1. Arch Intern Med 1988; 148: 1596-1600. 2. R I Med J 1987; 70: 311-6.

6. 6 Selected Adverse Events Associated with Lotronex Ischemic colitis (IC), small bowel ischemia, and serious complications of constipation (SCC) IC and SCC cases are mutually exclusive (if they co-exist, case linked to IC); all small bowel cases discussed separately Reports received through March 8, 2002

7. 7 AERS Report Quality Reports after market suspension have come primarily from consumers and available clinical data are not comprehensive More recently, reports have come from class action lawsuits and available clinical data also are not comprehensive Reporter follow up intensive prior to market suspension

8. 8 Ischemic Colitis (IC) Case definition used by ODS was based on any or a combination of the following: (1) the term “ischemic colitis” is explicitly used in the AERS report as a possible diagnosis, (2) any endoscopic or histologic evidence of ischemic change or necrosis, or (3) any radiologic evidence of ischemic colitis.

9. 9 Diagnostic Certainty of IC Cases* (n = 84) Biopsy confirmed: 33 (39%) Clinical diagnosis by colonoscopy without biopsy confirmation: 17 (20%) Diagnosis only: 34 (41%), the reporters stated that the patient had ischemic colitis, but did not provide documentation. *Mutually exclusive

10. 10 Description of IC Cases Gender: Female 81 (97%), Male 1 (1%), Unk 2 (2%) Age (years): 55 median/mean, 25 to 80 range Time to onset (days): 14 median, 39 mean, 1 to 200 range (n = 66)

11. 11 Description of IC Cases cont. Presenting symptoms (not mutually exclusive) –Bloody stool/diarrhea: 54 (64%) –Constipation: 16 (19%) –Abdominal pain: 63 (75%) Concomitant estrogen use: 22 (26%)

12. 12 Outcomes of IC Cases* Required hospitalization: 54 (64%) Required surgery: 11 (13%) –Segmental resection: 10 –Unknown: 1 Required transfusions: 2 (2%) Death: 2 (2%) *Not mutually exclusive

13. 13 IC and IBS-Epidemiological Data Epidemiological studies submitted by GSK suggest potential for “misdiagnosis” of selected conditions as IBS –Inflammatory bowel disease (IBD): ulcerative colitis and Crohn’s disease –Ischemic colitis

14. 14 Epidemiological Data cont. Given risk of IC due to Lotronex and potential for a “background rate” of IC in IBS population, can calculate attributable risk Based on relative risk of 5.9 for IC with Lotronex vs. placebo (from initial NDA), attributable risk is 83%

15. 15 Small Bowel Ischemia The case definition used by ODS was any ischemic change of the small bowel documented by endoscopic, surgical, or pathologic evidence.

16. 16 Description of Small Bowel Ischemia Cases (n = 6) Cases with ischemia, infarction, or necrosis of small bowel All female, ages ranged from 33 to 81 years Time to onset (days): mean = 10 (n = 4), 120 (n = 1), Unk (n = 1) 5 Surgeries 3 Deaths

17. 17 Serious Complications of Constipation (SCC) The case definition used by ODS was constipation or suspected constipation that was associated with an ER visit, hospitalization, or complications, including but not limited to, fecal impaction, bowel obstruction, necrosis, or rupture.

18. 18 Description of SCC Cases (n = 113) Gender: Female 103 (91%), Male 10 (9%) Age (years): 57 median, 54 mean, 24 to 82 range Time to onset (days): 14 median, 35 mean, 1 to 180 range (n = 79) Presenting symptoms (not mutually exclusive): –Constipation: 84 (74%) –Bloody stool: 28 (25%) –Abdominal pain: 74 (66%) Concomitant estrogen use: 19 (17%)

19. 19 Outcomes of SCC Cases* Required hospitalization: 83 (74%) Required surgery: 34 (30%) –Intestinal surgery: 25 –Anorectal surgery: 9 Transfusions: 2 (2%) Death: 2 (2%) *Not mutually exclusive

20. 20 Deaths: Total AERS Experience Total of 14 deaths in patients receiving Lotronex Association with Lotronex cannot be reasonably excluded in 7 cases –2 cases of IC –3 cases of small bowel ischemia –2 cases of SCC

21. 21 Clinical Trials (CT) vs. Postmarketing (PM) Experience

22. 22 Clinical Trials vs. Postmarketing Experience Clinical trials have strict entry criteria Use in the real world is less stringent; in this subset of Lotronex adverse events, we see the following: –Use in men: 11 –Off-label use: 22 –Use in contraindicated conditions: 18

23. 23 Conclusions In review of IBS literature and studies submitted by GSK, potential for “misdiagnosis” of selected conditions, such as IBD and IC, as IBS Most (80+%) IC cases reported in association with Lotronex can be attributed to Lotronex

24. 24 Conclusions cont. Presenting symptoms did not necessarily predict severity of outcome These data do not reveal any potential risk factors for these events Potential for unknown risks as yet identified Managing risk in the general population differs from managing risk in clinical trials