Four CCTs (253 patients at baseline, 104 at end point) were included. These comprised three RCTs (223 patients at baseline) and one non-randomised CCT (30 patients at baseline).
Only one RCT presented sufficient information to confirm the adequacy of the method of randomisation. None of the studies explicitly reported that the patients were blinded. All the studies treated groups equally. None of the studies reported a power calculation for sample size. Two studies included patients who died before trial completion, whereas the other two studies did not include this group. The follow-up rates were low, ranging from 31.1 to 66.6% (excluding patients who died before the study was completed).
The studies found that non-metal cannulae increased site duration compared with metal cannulae. The increase in site duration ranged from 21 to 159 hours in the individual studies.
Vialon cannulae versus butterfly needle (2 RCTs): neither study found any significant difference between the interventions in terms of site duration. One RCT found site duration was 177 hours with Vialon versus 138 hours with metal (95% CI for difference: -25, 101); the second RCT found site duration was 134 hours with Vialon versus 113 hours with metal (95% CI for difference: -8, 62).
Teflon cannulae versus butterfly needle (1 RCT, 1 CCT): both studies found that Teflon cannulae significantly increased site duration compared with butterfly needle. The RCT found that site duration was 286 hours with Teflon versus 127 hours with metal (95% CI for difference: 140, 177); the CCT found the median site duration was 94 hours with Teflon versus 42.8 hours with metal (P=0.0002).