Levine 1996.

Methods	Study design: randomised controlled trial. Method of randomisation: stratified according to centre, mode of diagnosis (venography or ultrasonography), and category of participants (outpatients, admitted at weekend or at night, hospitalised). Concealment of allocation: not blinded for treatment allocation, blinded for outcome assessment. Exclusions post‐randomisation: none. Lost to follow‐up: none.
Participants	Country: Canada. Setting: hospital. No.: 500 outpatients and inpatients. Age: mean 58 years. Sex: males and females (breakdown not supplied). Inclusion criteria: acute proximal DVT. Exclusion criteria: 2 or more previous episodes of DVT or PE; active bleeding; active peptic ulcer disease; familial bleeding disorder; concurrent symptomatic PE; > 48 hours heparin treatment; inability to be treated with LMWH as outpatient because of coexisting condition (e.g. cancer, infection, stroke) or likelihood of non‐compliance; inability to make follow‐up visits because of geographical inaccessibility; presence of known deficiency of anti‐thrombin III, protein C or protein S; pregnancy.
Interventions	Treatment: LMWH: enoxaparin (Rhone‐Poulenc Rorer) body weight‐adjusted fixed dose (1 mg/kg body weight), s.c., twice daily, at home. 1 vial: 1 mL/100 mg = 100 anti‐factor Xa IU/mg). Control: UFH: APTT‐adjusted, continuous i.v. infusion (started with 20,000 Units in 500 mL of 5% dextrose solution) in hospital after an initial i.v. bolus of 5000 Units. Treatment duration: at least 5 days; discontinuation if INR was 2 or above and maintained for 2 consecutive days. Oral anticoagulation warfarin sodium started on day 2 and continued for 3 months.
Outcomes	Primary: symptomatic recurrent VTE within 90 days of follow‐up; major haemorrhage during the initial treatment or 48 hours after treatment cessation. Secondary: minor haemorrhage; mortality.
Notes	Some participants received 1 or 2 days UFH before randomisation; this was considered part of the overall duration of heparin treatment. Follow‐up: 3 months.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information.
Allocation concealment (selection bias)	Low risk	Allocation over the telephone from a central site.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Not blinded for treatment allocation. Comment: given the clinical outcomes of the study, review authors judged that the non‐blinding of the participants and staff was unlikely to have affected the outcomes.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All reported outcome events were reviewed by a central adjudication committee whose members were unaware of the treatment assignments.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data.
Selective reporting (reporting bias)	Low risk	Published report includes all expected outcomes.
Other bias	Low risk	The study appears to be free of other sources of bias.