A Cochrane review included 16 RCTs, 10 with pharmacological and 6 with non-pharmacological intervention. A pooled analysis was precluded due to differences in reporting.
Frequency of hot flushes was significantly reduced at all dose levels of venlafaxine (37.5 mg: 30% reduction, 95% CI 0.22 to 0.53; 75 mg: 46% reduction, 95% CI 0.36 to 0.63; and 150 mg: 58% reduction, 95% CI 0.04 to 0.75; 1 trial; n=191) compared to placebo. The severity score was more reduced in the active arms at all dose levels (37.5 mg: 37% reduction, 95% CI 0.26 to 0.54; 75 mg: 61% reduction, 95% CI 0.50 to 0.68; 150 mg: 61% reduction, 95% CI 0.48 to 0.75). Two other studies on venlafaxine at 37.5 mg and 75 mg doses reported significantly superior reduction both in frequency and severity of hot flushes (P<0.001, respectively at 37.5 mg and P=0.013 and 0.001 at 75 mg) versus placebo.
One study evaluating fluoxetine at a dose of 20 mg daily in 68 patients showed a reduction in the number of flushes (19%; P=0.01) and severity score (24%; P=0.02) compared to placebo. Paroxetine reduced the frequency and severity score of hot flushes significantly compared to placebo (40.6 %; P=0.0006 and 45.6 %; P=0.0008 in the 10 mg dose arms and 51.7%; P=0.002 and 56.1%; P<0.001 in the 20 mg arms, respectively; 1 study; n =107).
Gabapentin reduced the frequency of hot flushes with the 900 mg/day dose (mean change –2.1; 95% CI –2.95 to –1.23) but not significantly with 300 mg/day (mean change –0.8; 95% CI –1.7 to 0.1; 1 study; n = 347). The same effects were observed on severity scores (low dose mean change –1.79; 95% CI –4.38 to 0.80 and high dose mean change –4.88; 95% CI –7.23 to –2.53).
Clonidine was tested in transdermal formulation in one study (n=89) and in oral formulation in one (n=163) compared to placebo. A superior reduction in both number and severity of hot flushes was achieved with both formulations (44%; P=0.04 and 56%; p=0.04 for transdermal and 38%; P=0.006 and 45%; P=0.006 for oral formulations, respectively).
Relaxation therapy (n=150) showed an effect in 1 of 2 trials (median difference 7 hot flushes per week; 95% CI 4 to 7, median difference of severity 0.54; 95% CI 0.11 to 1.01), but the significance was lost during the follow-up. Homeopathy, acupuncture and use of a magnetic device did not show any effect.
Comment: The quality of evidence is downgraded by inconsistency (heterogeneity in interventions and outcomes) and by imprecise results (limited study size for each comparison) and upgraded by dose-response relationship.
Clinical comment: It should be noted that paroxetine may interfere with the metabolism of tamoxifen and is not recommended for women taking this drug for their breast cancer. (The same may apply to fluoxetine.)