. Author manuscript; available in PMC: 2015 Jun 1.

Published in final edited form as: J Affect Disord. 2014 Mar 13;161:30–35. doi: 10.1016/j.jad.2014.02.046

Table 1.

Effect of benzodiazepine (BZD) use at any time during follow-up^a,^b

	Baseline			Estimated change from baseline			Estimated difference in 6 month change
		BZD use during study			BZD use during study		BZD use during study
	Overall	No use	Any use	Overall	No use	Any use	Difference^c (Any use – No use)
	Mean ± SD	Mean ± SD	Mean ± SD	Mean (95% CI), p-value	Mean (95% CI), p-value	Mean (95% CI), p-value	Mean (95% CI), p-value
BISS total	56.1 ± 18.8	54.8 ± 19.0	59.2 ± 18.1	−28.39 [−30.18,− 26.59] p<.0001	−29.30 [−31.41,− 27.19] p<.0001	−25.04 [−28.35,− 21.73] p<.0001	4.26 [0.38,8.14] p = 0.03
BISS mania	9.2 ± 6.4	9.2 ± 6.4	9.1 ± 6.2	−4.98 [−5.53,− 4.44] p<.0001	−4.94 [−5.59,− 4.29] p<.0001	−4.33 [−5.35,− 3.31] p<.0001	0.62 [−0.58,1.82] p = 0.31
BISS depression	17.5 ± 7.3	17.0 ± 7.3	18.8 ± 7.3	−8.89 [−9.62,− 8.16] p<.0001	−9.25 [−10.3,− 8.38] p<.0001	−8.04 [−9.41,− 6.67] p<.0001	1.21 [−0.40,2.82] p = 0.14
BISS anxiety	15.9 ± 8.5	15.0 ± 8.5	18.2 ± 8.2	−7.10 [−7.86,− 6.34] p<.0001	−7.35 [−8.26,− 6.44] p<.0001	−6.44 [−7.88,− 5.01] p<.0001	0.91 [−0.77,2.59] p = 0.29
BISS irritability	16.8 ± 8.5	16.4 ± 8.4	17.7 ± 8.7	−7.98 [−8.80,− 7.16] p<.0001	−8.47 [−9.44,− 7.50] p<.0001	−6.60 [−8.12,− 5.08] p<.0001	1.87 [0.09,3.65] p = 0.04
CGI-EI difference	…	…	…	1.61 [1.38,1.84] p<.0001	1.68 [1.41,1.95] p<.0001	1.20 [0.77,1.63] p<.0001	−0.48 [−0.98,0.02] p = 0.06
CGI-BP overall	4.5 ± 0.9	4.4 ± 0.8	4.7 ± 0.9	−1.61 [−1.73,− 1.50] p<.0001	−1.64 [−1.78,− 1.51] p<.0001	−1.36 [−1.57,− 1.15] p<.0001	0.28 [0.03,0.53] p = 0.03

Exposure groups included patients who were prescribed benzodiazepines (BZDs) at baseline or at any time during post-randomization follow-up (any use) and those who were not prescribed BZDs at baseline or at any time during post-randomization follow-up (no use).

Within- and between-group comparisons were made using mixed effects models that included a random intercept and slope over time and fixed effects for benzodiazepine use, log-time, study site, and the interaction between benzodiazepine use and log-time.

Statistically significant results for between-group differences in estimated 6 month change for a given outcome measure appear in bold type. These results were subjected to further analysis to test robustness of findings (shown in Table 2).