Background and purpose Current evidence suggests that the time lag from the publication of randomised clinical trial results to changes in prescribing behaviour for drugs is gradually reducing. However, the effect of results of clinical trials of devices and non-pharmacological interventions on clinical practice is less clear.

Methods Prospective data from the ongoing international ‘Efficacy of Nitric Oxide Stroke’ (ENOS) trial were analysed to assess the use of graduated compression stockings (GCS) for deep vein thrombus (DVT) prophylaxis in acute stroke patients before and after publication of the large ‘Clots in Legs Or sTockings after Stroke’ (CLOTS-1) trial.

Results Data on GCS use were available for 1971 patients with acute stroke enrolled into ENOS from February 2003 to April 2011; of these, 498 (25.3%) wore GCS. Prior to publication of CLOTS-1, GCS use was common (>50%) in the UK, Australasia and Canada but infrequent in Asia and the rest of Europe. After publication of CLOTS-1, use of GCS in the UK declined from 398/656 (61%) to 20/567 (4%) (p<0.001) but not elsewhere (eg, in Australasia (57% before publication vs 70% after publication, p=0.24, but based on small numbers). Practice change was apparent within 3 months of the study publication and was sustained thereafter. There was no change in DVT rates before and after CLOTS-1 (0.8% vs 1.0%).

Conclusions GCS use declined dramatically following the reporting of the CLOTS-1 trial. The results support the notion that a neutral trial of a device can influence clinical practice rapidly, which is important with a widely used and moderately expensive (time and finance) intervention.