"Optimal" percutaneous transluminal coronary angioplasty (PTCA) may have a late restenosis rate similar to stenting. We sought to assess short- and long-term results of a provisional stenting/optimal PTCA approach compared with elective stenting in a prospective, randomized study. A total of 97 patients with discrete, de novo lesions in native coronary arteries > or =3 mm in diameter were randomized 2:1 in PTCA with prolonged perfusion balloon inflation (n = 66) versus elective stenting (n = 31). Recoil after PTCA was assessed by routine delayed angiograms (5 and 20 minutes). Cross over to stent was allowed for an inadequate result; there was no on-line quantitative angiography. An independent core angiographic laboratory assessed all results and evaluated the adequacy of the subjective interpretation. Within the PTCA arm, there were 24 (36%) crossovers to stenting (5 of 24 [21%] due to recoil), whereas 2 stents could not be delivered to the lesion and crossed over to PTCA. As assessed by quantitative angiography, baseline reference vessel diameters were similar between the PTCA and stent groups. The immediate lumen diameter achieved with PTCA was smaller than that achieved with stenting (2.18+/-0.49 vs. 2.44+/-0.38 mm, respectively, p = 0.01). There were no differences in angiographic results between elective and crossover stenting and there were no in-hospital complications in any patient. Target lesion revascularization at 8 months was 19% (n = 6) in the elective stent arm versus 21% (n = 14) in the PTCA arm, p = NS; respective rates in PTCA alone and crossed over-to-stent subsets were 23% (n = 10) versus 17% (n = 4), p = NS. Angiographic restenosis was 47% after elective stenting versus 38% after PTCA (intention to treat), p = NS. By received treatment, it was 41% (11 of 27) in the group treated with the PTCA versus 33% (5 of 15) in the crossover-to-stent arm (p = NS). Thus, provisional stenting can be safely performed in the treatment of discrete, native de novo lesions. Early recoil after PTCA cannot be reliably assessed without quantitative angiography.