This meta-analysis used all original articles from 1966 to June 1996 that fit the preset inclusion criteria to examine the clinical effectiveness of intermittent pneumatic compression (IPC) devices in preventing deep vein thrombosis (DVT) and pulmonary embolism and to compare the results of knee-high sleeves to thigh-high sleeves. IPC devices decreased the relative risk of DVT by 62 per cent when compared with placebo, 47 per cent compared with graduated compression stockings, and 48 per cent compared with mini-dose heparin. IPC devices significantly decreased the relative risk of DVT compared with placebo in high-risk patients such as neurosurgery and major orthopedic surgery patients and in modest risk patients such as general surgery patients. In major orthopedic surgery patients, the incidence of DVT was similar for IPC- and warfarin-treated patients; however, IPC was significantly better than warfarin at decreasing the incidence of calf only DVT, whereas warfarin seemed to be better at decreasing proximal DVT. IPC devices are effective in decreasing the incidence of DVT in patients at moderate to high risk and are probably more efficacious than graduated compression stockings or mini-dose heparin; however, IPC devices are not protective against pulmonary embolism. The data directly comparing the various methods of compression (knee-high versus thigh-high sleeves and graded-sequential versus uniform compression) are sparse and conflicting.