Table 1.

Randomized Controlled Trials comparing the effectiveness of a taper program alone versus a taper program associated with CBT.

Study	Condition	Population	Benzodiazepine Use	Design	Withdrawal Schedule	Main Outcomes
Otto et al., 1993 [118]	Panic disorder	n = 33, 22 women	Alprazolam or clonazepam use >6 months	Slow taper alone (5 to 7 weeks) (n = 16) vs. slow taper + 10 weekly sessions (60–90 min) of group CBT (n = 17)	Alprazolam: Reduction of 0.25 mg or 0.125 mg every 2 days (depending on the initial dose) Clonazepam: reduction of 0.25 mg every 4 or 8 days (depending on the initial dose)	Successful discontinuation (=completion of the taper and no use of BZD beyond “minimal p.r.n use” during the 2 post-discontinuation weeks) in 76% with CBT vs. 25% without CBT (p < 0.01) At 3 months: persistent effect of CBT
Spiegel et al., 1994 [120]	Panic disorder	n = 21, 17 women	Alprazolam use, 1 to 10 mg/j	Supportive drug maintenance and slow flexible taper (n = 10) vs. same taper + 12 weekly sessions of individual CBT (n = 11)	Reduction of 0.125 mg to 0.5 mg/1–2 week, mean duration 6.5 weeks	Successful discontinuation (=completion of the taper and no use of BZD through the follow-up) 2 weeks after treatment: 90% with CBT vs. 80% without CBT (ns) At 6 months: 90% with CBT vs. 40% without CBT (p < 0.05)
Baillargeon et al., 2003 [122]	Chronic insomnia in older adults	n = 65, >50 years, 38 women	Daily use of BZD >3 months (Molecules not specified)	Slow taper alone (n = 30) vs. slow taper + 8 weekly sessions (90 min) of group CBT (n = 35)	25% reduction of dosage every 1–2 weeks	Successful discontinuation (= BZD cessation confirmed by blood screening) in 77% with CBT vs. 38% without CBT after treatment completion (p = 0.002), and results maintained at 3 and 12 months
Voshaar et al., 2003 [121]	Not mentioned	n = 180, 128 women	BZD use >3 months BZDs were switched for an equivalent dose of diazepam	Usual care (letter with advice to stop) (n = 34) vs. taper (n = 73) vs. taper + 5 weekly sessions (120 min) of group CBT (n = 73)	25% reduction of dosage every week	Successful discontinuation (=no self-reported BZD use at 3 months follow-up) in 58% with CBT vs. 62% tapering off without CBT (no additional benefice of CBT) vs. 21% with usual care
Morin et al., 2004 [33]	Chronic insomnia in older adults	n = 76, >55 years, 38 women	BZD use >50% of nights >3 months (different molecules: lorazepam, alprazolam, bromazepam, oxazepam, temazepam, clonazepam, flurazepam, triazolam)	Supervised withdrawal program (n = 25) vs. CBT for insomnia (weekly 90 min sessions) (n = 24) vs. supervised withdrawal program + CBT (n = 27) For all groups: program duration 10 weeks	25% reduction of dosage every 2 weeks and introduction of an increasing number of drug-free nights	Drug-free patients (confirmed by blood and urine samples): 85% taper + CBT vs. 48% taper alone vs. 54% CBT alone at post-treatment (p < 0.002) and results maintained at 3 and 12 months Reduction of weekly quantity of BZD use (dosage_overall 90% reduction_and number of nights_overall 80% reduction) in the 3 groups with lower frequency of medicated night in the CBT + taper vs. taper alone group
Gosselin et al., 2006 [112]	Generalized Anxiety Disorders (GAD)	n = 61, 36 women	BZD use >4 days/week for >12 months (different molecules: clonazepam, lorazepam, alprazolam, bromazepam, oxazepam, temazepam, diazepam, clorazepate)	Non-specific psychological treatment (NST) + taper (n = 30) vs. CBT + taper (n = 31), 12 weekly 90 min sessions	25% reduction of dosage every 2–3 weeks	Drug-free patients at post-treatment: 74% CBT + taper vs. 37% NST + taper group, p < 0.001. Results maintained at 3, 6 and 12 months. Greater proportion of patients no longer with GAD criteria in the CBT group
Otto et al., 2010 [119]	Panic Disorders	n = 47, 31 women	Alprazolam or clonazepam use >6 months	Taper alone (5 to 9 weeks) (n = 15) vs. taper + CBT (8 weekly 60–90 min sessions followed by 3 booster sessions separated by 2 weeks) (n = 16) vs. taper + relaxation (same number/duration of session as CBT sessions) (n = 16)	Alprazolam: Reduction of 0.25 mg or 0.125 mg every 2 days (depending on the initial dose) Clonazepam: reduction of 0.25 mg every 4 or 8 days (depending on the initial dose)	Successful discontinuation (=completion of the taper and no use of BZD beyond “minimal p.r.n use” during the month post-discontinuation) in 56% CBT vs. 31% relaxation vs. 40% taper alone (ns), and maintained at 3 (44% vs. 13% vs. 27%, ns) and 6 months (63% vs. 13% vs. 27%, p < 0.01)
Lichstein et al., 2013 [123]	Chronic insomnia in older adults	n = 70, >50 years, women	Hypnotic dependance (BZD, non-BZD receptor agonists, sedating antidepressants)	Withdrawal only (4–8 biweekly 30 min sessions) vs. withdrawal (n = 23) + placebo biofeedback vs. withdrawal (8 weekly 45 min sessions) (n = 23) + CBT (8 weekly 45 min sessions) (n = 24)	Conversion of the dose of hypnotics in a number of “lowest recommended dosages” (LRD): gradual reduction to nightly dose at 1 LRD then gradual elimination of nightly dose	Drug-free patients at post-treatment: 67% CBT vs. 61% placebo feedback vs. 52% withdrawal only (ns) At follow-up (1 year): 50% vs. 35% vs. 43% (ns)

BZD: benzodiazepine; CBT: cognitive behavioral therapy; p.r.n: pro re nata; GAD: generalized anxiety disorders; NST: non-specific psychological treatment; LRD: lowest recommended dosage.