Eleven RCTs (n=2,513) were included in the review, but only seven trials (n=852) were eligible for inclusion in the meta-regression. Follow-up ranged from acute phase only to 12 months. Two trials were double-blinded. One trial used intention-to-treat analysis. Assessor blinding was adequately reported in four trials. Most trials were considered to be at high risk of bias.
In patients who received subcutaneous and intravenous UFH, the maintenance measurement of activated partial thromboplastin time was significantly associated with the resolution of thrombus at the end of treatment (p<0.0001). The initial measurement was significantly associated with minor bleeding rate (p<0.01). The maintenance measurement was only significantly associated with minor bleeding rate in patients receiving subcutaneous UFH (p=0.0132). Where data were available, both treatments were significantly associated with the overall death rate at three months (p<0.0035). Activated partial thromboplastin time was not significantly associated with rates of recurrent DVT or rates of new clinical pulmonary embolism during treatment at three months, treatment-related major bleeding and death or venous thromboembolism during treatment. Subgroup analysis did not materially alter the results.
Trials that were not included in the meta-regression reported conflicting results.