STATISTICAL REVIEW AND EVALUATION

NDA #: 20-838

SERIAL #: SE4-S015

DRUG NAME: ATACAND (candesartan cilexetil) Tablets

INDICATION: Hypertension

SPONSOR: AstraZeneca LP

DOCUMENT REVIEWED:

1. Vols. 1.1, 1.5, 1.37, 1.52, 1.71 (CDER REC’D Date: September 27, 2001), Supplemental new correspondence (CDER REC’D Date: 4/16/02)
2. SAS database in EDR

STATISTICAL REVIEWER: H.M. James Hung, Ph.D. (HFD-710)

MEDICAL REVIEWER: Steve Fredd, M.D. (HFD-110)

PROJECT MANAGER: Edward Fromm (HFD-110)

STATISTICAL KEY WORDS: ANCOVA, LOCF, Least square mean

Distribution: NDA 20-838, SE4-S015

HFD-110/Dr. Throckmorton

HFD-110/Dr. Stockbridge

HFD-110/Dr. Steve Fredd

HFD-110/Mr. Fromm

HFD-700/Dr. Anello

HFD-710/Dr. Chi

HFD-710/Dr. Mahjoob

HFD-710/Dr. Hung

HFD-710/chron

JHung/301-594-5436/DB1/atacand.doc/5-17-2002

0. SUMMARY *

1. INTRODUCTION *

2. OVERVIEW OF THREE CLINICAL TRIALS *

3. CONCLUSIONS *

This NDA provides three main clinical trials (two CLAIM studies and one CANDLE study) for comparison of the candesartan 16 mg to 32 qd mg regimen with the losartan 50 mg to 100 mg qd regimen. The two CLAIM studies showed that the candesartan regimen gave a statistically significantly greater reduction in blood pressures than the losartan regimen when given by forced titration. The difference was 1 to 2.2 mm Hg in trough sitting DBP and about 3.5 mm Hg in trough sitting SBP. When given by optional titration, the candesartan regimen also gave a statistically significantly greater reduction in trough sitting diastolic blood pressure. The difference was 2.2 mm Hg in trough sitting DBP. The difference in trough sitting SBP was < 2 mm Hg, not statistically significant.

Over the course of each study, the candesartan group appeared to have lower mean blood pressures than the losartan group. In all the studies, the mean blood pressures in both treatment groups were greatly reduced in the first two weeks and then seemed to stabilize afterwards.

Candesartan had a favorable trend in gender, race and age subgroups, except small subgroups.

This statistical review pertains to the three clinical studies, CLAIM (Studies 230 and 231) and CANDLE (Study 175). Most of the sponsor’s main results were confirmed by the reviewer’s analyses and reported in this review unless stated otherwise.

This was a multicenter, double-blind, randomized, parallel-group, forced-titration study to evaluate the anti-hypertensive efficacy of candesartan 16 mg forced titrated to 32 mg in comparison to losaran 50 mg forced titrated to 100 mg in the patients with mild-to-moderate essential hypertension (sitting DBP of 95 mm Hg to 114 mm Hg inclusive). The doses were selected to represent the labeled maximum doses of candesartan and losartan. All treatments were given once daily.

Following a 4 to 5 week placebo run-in period, qualified study patients were randomized in 1:1 ratio to candesartan 16 mg or losartan 50 mg. After two weeks of randomized treatment, all patients doubled their current dose of candesartan or losartan and continued treatment for an additional 6 weeks. Patients were also seen 48 hours following their last dose of study medication and 2 weeks after they discontinued therapy with the study medication for follow-up visits.

The primary efficacy measure was the mean change in trough sitting DBP from baseline to the double-blind week 8 visit. Secondary measures include sitting systolic pressures and peak blood pressures at week 1, week 2 and week 8, and proportion of responders and controlled patients based on trough sitting DBP and SBP at week 8. The primary analysis is based on ITT principle and the last value carried forward strategy for incompleters.

The sample size of 735 patients to be enrolled from 75 investigative sites was planned to detect a difference in mean change from baseline in trough sitting DBP of 2.0 mm Hg between the two treatment groups at 2-sided alpha level of 0.05 and power of 95%, assuming a standard deviation of 7.5 mm Hg.

A total of 613 patients out of 926 patients screened were randomized at 72 sites. Of these patients, 12.1% patients in the candesartan group and 12.4% patients in the losartan group discontinued treatment during the double-blind phase of the study. The distribution of patients discontinuing treatment by reason was similar in the two treatment groups (Table 230-1). Numerically, the losartan group appeared to have a bit larger proportion of discontinuation due to lack of response whereas the candesartan group appeared to have a bit larger proportion of discontinuation due to adverse event.

The two treatment groups appeared to be comparable with respect to baseline characteristics (Sponsor’s Table 4 of study report for Study 230) and treatment compliance (Sponsor’s Table 14.1.1.03,14.1.1.05, 14.1.1.06). The mean age was 55 years and the average duration of hypertension is 10+ years. Fifty nine percent of patients were male an 20% were blacks. The mean sitting DBP/SBP was 100/153 mm Hg.

The primary efficacy variable, change from baseline to double-blind week 8 in trough sitting DBP, showed a greater reduction with candesartan (Table 230-2). Statistical significance is borderline. The statistical analysis was based on the model containing center where centers with at least one treatment arm having less than two patients were pooled using a prespecified pooling algorithm according to the sponsor’s clarification. Since it was borderline significance, this reviewer did analysis without adjusting for centers and analysis by pooling all centers with less than two patients in a treatment arm into one center. Both analyses gave a smaller p-value and an estimated treatment difference very similar to the estimate reported in Table 230-2. Excluding baseline sitting DBP from the model made little change in the result. Candesartan also gave a statistically significantly greater reduction in trough sitting SBP and peak sitting DBP/SBP. However, the differences in all blood pressures were small.

Blood pressures seemed to trend favorably to candesartan over time (Table 230-3). The

favorable difference in blood pressures seemed to remain at 48 hours after the last dose. Two

weeks after the last dose, the mean blood pressures were similar between the two treatment groups. In both treatment groups, blood pressures were greatly reduced in the first two weeks and then seemed to stabilize afterwards.

Candesartan appeared to have a favorable trend in change from baseline in trough sitting blood pressures in most of the subgroups, except blacks or small subgroups.

The trial design of this study is identical to that of Study 230.

A total of 655 patients out of 921 patients screened were randomized. 654 patients has at least

one post-baseline site contact. These patients constituted the intent-to-treat population. Of these patients, 4.5% patients in the candesartan group and 6.2% patients in the losartan group discontinued treatment during the double-blind portion of the study. The distribution of patients discontinuing treatment by reason was similar in the two treatment groups (Table 231-1).

The two treatment groups appeared to be comparable with respect to baseline characteristics (Sponsor’s Table 4 of study report for Study 231) and treatment compliance (Sponsor’s Table 14.1.1.05,14.1.1.06, 14.1.1.03). The mean age was 54 years and the average duration of hypertension was 9 years. Fifty eight percent of patients were male and 17 percent were blacks. The mean sitting DBP/SBP was 100/152 mm Hg.

The primary efficacy variable, change from baseline to double-blind week 8 in trough sitting DBP, showed a significantly greater reduction with candesartan (Table 231-2).

Blood pressures seemed to trend favorably to candesartan over time (Table 230-3). The favorable difference in blood pressures seemed to remain at 48 hours after the last dose. Two weeks after the last dose, the mean blood pressures were similar between the two treatment groups. As in Study 230, In both treatment groups, blood pressures were greatly reduced in the first two weeks and then seemed to stabilize afterwards.

In this study, candesartan had a favorable trend in gender, race and age subgroups (Table 231-4).

This was a multicenter, double-blind, randomized, parallel-group, optional-titration (titration-to-effect) study to evaluate the anti-hypertensive efficacy of candesartan 16 mg titrated to 32 mg in comparison to losaran 50 mg titrated to 100 mg in the patients with mild-to-moderate essential hypertension (sitting DBP of 95 mm Hg to 114 mm Hg inclusive). Patients with a mean of sitting DBP ³ 90 mm Hg after 4 weeks of treatment were up-titrated to candesartan 32 mg or losartan 100 mg for the remaining 4 weeks of the study. Patients with a mean of sitting DBP < 90 mm Hg after 4 weeks of treatment continued on candesartan 16 mg or losartan 50 mg. All treatments were given once daily.

Following a 4 to 5 week placebo run-in period, qualified study patients were randomized in 1:1 ratio to candesartan 16 mg or losartan 50 mg. After four weeks of randomized treatment, all patients either continued on the same dose or up-titrated according to the criterion described above for an additional 4 weeks. Patients were also seen 2 weeks after they discontinued therapy with the study medication for follow-up visits.

The primary efficacy measure was the mean change in trough sitting DBP from baseline to the double-blind week 8 visit. Secondary measures include sitting systolic pressures and peak blood pressures, standing blood pressures, and proportion of responders and controlled patients based on trough sitting DBP and SBP at week 8. The primary analysis is based on ITT principle and the last value carried forward strategy for incompleters.

The sample size of 330 patients to be enrolled from 45 investigative sites was planned to detect a difference in mean change from baseline in trough sitting DBP of 3.0 mm Hg between the two treatment groups at 2-sided alpha level of 0.05 and power of 95%, assuming a standard deviation of 7.5 mm Hg.

A total of 332 patients out of 460 patients screened were randomized at 40 sites. 329 patients has at least one post-baseline site contact. These patients constituted the intent-to-treat population. Of these patients, 4.3% patients in the candesartan group and 9.4% patients in the losartan group discontinued treatment during the double-blind portion of the study. The distribution of patients discontinuing treatment by reason was summarized (Table 175-1). The losartan group appeared to have more discontinuations due to lack of response or adverse event.

The two treatment groups appeared to be comparable with respect to baseline characteristics (Sponsor’s Table 4 of study report for Study 175) and treatment compliance (Sponsor’s Table 10.1.1.05,10.1.1.06, 10.1.1.03). The mean age was 55 years and the average duration of hypertension was 9+ years. Fifty eight percent of the patients were male and 12% blacks. The mean sitting DBP/SBP was 100/154 mm Hg.

The primary efficacy variable, change from baseline to double-blind week 8 in trough sitting DBP, showed a significantly greater reduction with candesartan (Table 175-2). Other blood pressures showed a numerically trend in favor of candesartan but only the reduction in peak sitting DBP showed a statistically significant treatment difference.

Blood pressures seemed to trend favorably to candesartan over time (Table 175-3). This pattern was observed in the patients who did not have their doses up titrated and who had (Table 175-4).

Two weeks after the last dose, the mean blood pressures were similar between the two treatment

groups. Standing blood pressures showed a similar pattern (Sponsor’s Tables 10.2.4.02, 10.2.4.05). Also, the mean blood pressures in both treatment groups were greatly reduced in the first two weeks and then seemed to stabilize afterwards.

In this study, candesartan had a favorable trend in gender, race and age subgroups (Table 175-4), except in small subgroups.

Two CLAIM studies showed that the candesartan 16 mg to 32 mg regimen gave a statistically significantly greater reduction in blood pressures than the losartan 50 mg to 100 mg regimen when given via force-titration. The difference was 1 to 2.2 mm Hg in trough sitting DBP and about 3.5 mm Hg in trough sitting SBP. The CANDLE study showed that when given by optional titration, the candesartan regimen also gave a statistically significantly greater reduction in trough sitting diastolic blood pressure. The difference was 2.2 mm Hg in trough sitting DBP. The difference in trough sitting SBP was < 2 mm Hg, not statistically significant. Over the course of each study, the candesartan group appeared to have lower mean blood pressures than the losartan group. In all the studies, the mean blood pressures in both treatment groups were greatly reduced in the first two weeks and then seemed to stabilize afterwards.