Table 3.
Secondary outcome measurements of full analysis set population
| Time point | Manual acupuncture (n=58) | Sham acupuncture (n=60) | Usual care (n=29) | P value* | Pairwise comparison | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Manual v sham | P value* | Manual v usual care | P value* | Sham v usual care | P value* | |||||||
| Proportion of participants with ≥50% reduction in number of migraine days (%) | ||||||||||||
| Weeks 17-20 | 82.5 | 45.8 | 17.9 | <0.001 | 36.7 (17.0 to 56.4) | <0.001 | 64.6 (43.5 to 85.7) | <0.001 | 27.9 (4.6 to 51.2) | 0.012 | ||
| Proportion of participants with ≥50% reduction in number of migraine attacks (%) | ||||||||||||
| Weeks 17-20 | 78.9 | 44.1 | 14.3 | <0.001 | 34.9 (14.7 to 55.0) | <0.001 | 64.7 (44.2 to 85.1) | <0.001 | 29.8 (7.6 to 51.9) | 0.004 | ||
| Mean (SD) change from baseline in visual analogue scale | ||||||||||||
| Weeks 20 | −2.2 (2.5) | −0.9 (1.9) | −0.0 (1.4) | <0.001 | −1.3 (−2.2 to −0.3) | 0.005 | −2.1 (−3.3 to −1.0) | <0.001 | −0.9 (−2.0 to 0.3) | 0.224 | ||
| Mean (SD) change from baseline in MSQ, role restrictive subscale | ||||||||||||
| Weeks 20 | 24.4 (21.4) | 12.8 (16.0) | 5.8 (14.2) | <0.001 | 11.6 (3.5 to 19.8) | 0.002 | 18.6 (8.5 to 28.7) | <0.001 | 7.0 (−3.1 to 17.0) | 0.285 | ||
| Mean (SD) change from baseline in MSQ, role preventive subscale | ||||||||||||
| Weeks 20 | 22.5 (25.1) | 10.1 (14.1) | 4.5 (15.5) | <0.001 | 12.5 (3.7 to 21.2) | 0.002 | 18.1 (7.2 to 29.0) | <0.001 | 5.6 (−5.2 to 16.4) | 0.630 | ||
| Mean (SD) change from baseline in MSQ, emotional subscale | ||||||||||||
| Weeks 20 | 19.5 (26.8) | 8.4 (19.0) | 8.8 (20.9) | 0.020 | 11.2 (0.9 to 21.4) | 0.027 | 10.7 (−2.0 to 23.4) | 0.129 | −0.4 (−13.1 to 12.2) | >0.999 | ||
| Mean (SD) change from baseline in global score of PSQI | ||||||||||||
| Weeks 20 | −1.8 (2.3) | −0.7 (2.5) | −0.1 (2.6) | 0.007 | −1.1 (−2.2 to 0.0) | 0.058 | −1.7 (−3.0 to −0.3) | 0.011 | −0.6 (−2.0 to 0.8) | 0.894 | ||
| Median (IQR) change from baseline in score of MIDAS | ||||||||||||
| Weeks 20 | −15.0 (−36.0-−5.0) | −6.0 (−30.0-2.0) | −0.5 (−16.0-6.5) | 0.007 | −8.0 (−17.0 to 2.0) | 0.139 | −14.0 (−28.0 to −3.0) | 0.006 | −6.0 (−22.0 to 4.0) | 0.523 | ||
| Median (IQR) change from baseline in doses of rescue medication | ||||||||||||
| Weeks 17-20 | 0.0 (−1.0-0.0) | 0.0 (−1.0-0.0) | 0.0 (−1.5-0.0) | 0.745 | NA | NA | NA | NA | NA | NA | ||
| Median (IQR) change from baseline in BDI- II | ||||||||||||
| Weeks 20 | −2.0 (−6.0-0.0) | −1.0 (−5.0-1.0) | 0.0 (−2.5-1.0) | 0.081 | NA | NA | NA | NA | NA | NA | ||
| Median (IQR) change from baseline in BAI | ||||||||||||
| Weeks 20 | −4.0 (−8.0-0.0) | −1.0 (−5.0-1.0) | 0.0 (−3.5-1.5) | 0.005 | −2.0 (−5.0 to 0.0) | 0.051 | −4.0 (−7.0 to −1.0) | 0.007 | −1.0 (−4.0 to 1.0) | 0.980 | ||
BAI=Beck Anxiety Inventory; BDI-II=Beck Depression Inventory II; IQR=interquartile range; MIDAS=Migraine Disability Assessment Scores; MSQ=Migraine-Specific Quality-of-Life Questionnaire; NA=not applicable; PSQI=Pittsburgh Sleep Quality Index.
Pairwise comparison include differences in means, medians, and proportions and corresponding confidence intervals. For continuous variables, analysis of variance or Kruskal-Wallis test was used for comparisons among treatment groups. If Kruskal-Wallis test was significant, pairwise comparisons were made using Wilcoxon rank sum test. Difference in median is estimated on basis of Hodges-Lehmann estimator of location shift. Asymptotic (Moses) confidence intervals are provided. For categorical variables, linear probability model was used for comparisons among treatment groups. For all pairwise comparisons, P values and confidence intervals were adjusted by Bonferroni approach.