Six publications describing seven RCTs were included in the updated review (861 patients). Five RCTs were located through the original searches (349 patients) and two RCTs were located though the update searches (512 patients). The five trials assessed for quality in the original review were all assessed as having low risks of bias for randomisation and incomplete outcome data/selective reporting. Results for other quality domains were variable. Length of follow-up across all seven RCTs ranged from six months to five years.
Drug-eluting balloons versus drug-eluting stents (three comparisons): No statistically significant differences were observed between groups in relation to target lesion revascularisation, major adverse cardiac events, mortality, myocardial infarction and stent thrombosis. Similar results were shown when analyses were restricted to populations with in-stent restenosis. Only one RCT reported in-stent angiographic observations (results not reported). No statistically significant differences between groups were observed in relation to late lumen loss, minimal lumen diameter and binary restenosis.
Drug-eluting balloons versus plain old balloon angioplasty (five comparisons): Statistically significant reductions in target lesion revascularisation (OR 0.27, 95% CI 0.15 to 0.48), major adverse cardiac events (OR 0.30, 95% CI 0.20 to 0.44) and mortality (OR 0.39, 95% CI 0.16 to 0.92) were observed with the use of drug-eluting balloons. No statistically significant difference between groups was found in relation to myocardial infarction. Stent thrombosis reportedly occurred only in one RCT; rates for this outcome were comparable across the trial arms. Use of drug-eluting balloons was also associated with statistically significant reductions in late lumen loss (in-stent MD -0.61, 95% CI -0.77 to -0.46; in-segment MD -0.55, 95% CI -0.74 to -0.35, Ι²=71%), minimal lumen diameter (in-stent MD 0.68, 95% CI 0.50 to 0.85; in-segment MD 0.60, 95% CI 0.48 to 0.72) and binary restenosis (in-stent OR 0.12, 95% CI 0.06 to 0.34; in-segment OR 0.14, 95% CI 0.06 to 0.34, Ι²=70%).
No substantial changes to the primary outcome results were observed when the meta-analyses were performed using fixed-effect models instead (to explore the influence of small studies). Similarly, no substantial changes were apparent when primary outcome results were calculated separately according to length of follow-up (6-12 months, up to 2 years, up to 5 years).