Background: There is currently no high grade evidence on which to base decisions about the frequency of intravenous cannula re-sites.

Objective: To assess the safety of changing peripheral venous cannulas when clinically indicated.

Design: Randomised controlled trial.

Setting: A tertiary referral hospital in Brisbane, Australia.

Participants: Two hundred and six hospitalised patients from surgical, medical and orthopaedic wards.

Interventions: Peripheral intravenous cannulas were re-sited only when complications occurred (intervention group) or every 3 days (control group).

Main outcome measures: The primary endpoint was any unplanned cannula removal, the secondary outcome was cost.

Results: Forty six patients had unplanned removals in the intervention group compared with 41 in the control group [relative risk 1.12, 95% confidence interval 0.81-1.55 (p=0.286)], a non-significant difference. Total duration of peripheral cannulation was similar in both groups (mean 123.3h in the intervention group and 125.9h in the control group: P=0.82) but significantly more re-sites occurred in the control group (167 in intervention group, 202 in the control group: p=0.022). Cost of cannula replacements in the intervention group was AUD$3,183.62 and in the control group AUD$3,837.56 (p=0.006).

Conclusion: Re-siting peripheral venous cannulas when clinically indicated compared with changing them routinely every 3 days does not lead to more complications and reduces costs.