Background: Recombinant fibroblast growth factor-2 (rFGF-2) improves perfusion in models of myocardial and hindlimb ischaemia. We investigated whether one or two doses of intra-arterial rFGF-2 improves exercise capacity in patients with moderate-to-severe intermittent claudication. Methods: 190 patients with intermittent claudication caused by infra-inguinal atherosclerosis were randomly assigned (1:1:1) bilateral intra-arterial infusions of placebo on days 1 and 30 (n=63); rFGF-2 (30 μg/kg) on day 1 and placebo on day 30 (single-dose, n=66); or rFGF-2 (30 μg/kg) on days 1 and 30 (double-dose, n=61). Primary outcome was 90-day change in peak walking time. Secondary outcomes included ankle-brachial pressure index and safety. The main analysis was per protocol. Findings: Before 90 days, six patients had undergone peripheral revascularisation and were excluded, and ten withdrew or had missing data. 174 were therefore assessed for primary outcome. Peak walking time at 90 days was increased by 0.60 min with placebo, by 1.77 min with single-dose, and by 1.54 min with double-dose. By ANOVA, the difference between groups was p=0.075. In a secondary intention-to-treat analysis, in which all 190 patients were included, the difference was p=0.034. Pairwise comparison showed a significant difference between placebo and single-dose (p=0.026) but placebo and double-dose did not differ by much (p=0.45). Serious adverse events were similar in all groups. Interpretation: Intra-arterial rFGF-2 resulted in a significant increase in peak walking time at 90 days; repeat infusion at 30 days was no better than one infusion. The findings of TRAFFIC provide evidence of clinical therapeutic angiogenesis by intra-arterial infusion of an angiogenic protein.