Skip to main content
. 2019 Oct 18;43(4):zsz252. doi: 10.1093/sleep/zsz252

Table 2.

Analysis results for Drug Liking (“at this moment”) VAS Emax (primary endpoint) in the completers population (N = 38)

Comparison Mean (SE) or median (Q1-Q3) of the paired difference P
Phentermine HCl 60 mg vs placeboa 22.7 (2.86) < 0.0001
Phentermine HCl 60 mg vs pitolisant 35.6 mga 21.4 (3.16) < 0.0001
Phentermine HCl 60 mg vs pitolisant 213.6 mga 19.7 (3.52) < 0.0001
Pitolisant 35.6 mg vs placebob 0.0 (0.0 to 6.0) < 0.0001
Pitolisant 213.6 mg vs placebob 0.0 (0.0 to 11.0) 0.0013

aPaired t-test was used to assess the difference between the two treatments. For these comparisons, the null hypothesis was that the mean difference for phentermine vs placebo was ≤5 and that the mean difference for phentermine vs pitolisant was ≤0.

bSign test (which evaluated statistical significance based on the proportion of patients for whom the pitolisant–placebo difference in Drug Liking Emax exceeded the prespecified threshold of 11) was used to assess the difference between the two treatments, because the paired differences were not normally distributed or quite symmetric. For these comparisons, the null hypothesis was that the difference between medians for pitolisant vs placebo was ≥11. Therefore, the significant results indicate that the null hypothesis was rejected and the Drug Liking Emax scores were similar for pitolisant and placebo.

Emax = peak maximum effect; Q1 = 25th percentile; Q3 = 75th percentile; SE = standard error; VAS = visual analog scale.