Vestibular rehabilitation (VR) is increasingly popular, but few data exist to support enthusiasts' claims of efficacy in improving functional abilities of patients with bilateral vestibular hypofunction (BVH). A double-blind, controlled study of eight subjects (mean, 64 +/- 12 years; seven females, one male) with bilateral vestibular hypofunction was conducted. Subjects in group A received 8 weeks of VR followed by 8 weeks of home VR exercises, whereas those in group B received 8 weeks of control treatment (isometric strengthening exercises) followed by 8 weeks of VR. At the end of 8 weeks, group A walked 8% faster and, during paced gait and stair-climbing, with greater stability, evidenced by a 10% larger maximum moment arm and a 17% decreased double-support duration during gait and stair stance. Group B improved less than 1% during the control treatment. Self-reported Dizziness Handicap Inventory scores did not differ significantly between control and active VR. All subjects improved compared with baseline tests at the 16-week post-test on both functional testing and on the Self-reported Dizziness Handicap Inventory scale. We conclude that in this small sample, VR effectively improved functional, dynamic stability during locomotion, but even strengthening exercises result in self-reported symptomatic improvement.