Patients with diabetes have a higher risk for in-stent restenosis after coronary stent implantation. Drug-eluting stents (DES) are highly effective in reducing in-stent restenosis. Once neointimal hyperplasia is suppressed with DES, the impact of stent underexpansion becomes magnified. The aim of this study was to evaluate DES expansion in patients with diabetes. Ninety-five patients with diabetes were randomized to Cypher Select (n = 48) or Taxus Express-2 (n = 47) stent implantation. Intravascular ultrasound was performed after stent implantation. Stent expansion was defined as the ratio of measured to predicted minimum stent diameter. There was a trend for lower stent expansion in the Cypher Select stent group (0.74 +/- 0.08 vs 0.78 +/- 0.11 in the Taxus Express-2 stent group, p = 0.061). Cypher Select stents achieved a final minimal stent cross-sectional area of 5.5 +/- 1. 8 mm2, compared with 6.4 +/- 1.9 mm2 for Taxus Express-2 stents (p = 0.015). For stents with nominal diameters > or =2.75 mm (Cypher Select n = 40, Taxus Express-2 n = 38), 42.5% of the Cypher Select stents and 10.5% of the Taxus Express-2 stents did not achieve a final minimum stent area of 5 mm2 (p = 0.002). Insulin treatment (relative risk 0.31, 95% confidence interval 0.10 to 0.95, p = 0.041) and stent type (relative risk 0.15, 95% CI 0.04 to 0.53, p = 0.003) were independent predictors of not achieving a minimum stent area >5.0 mm2. In conclusion, an important percentage of DES in patients with diabetes fail to achieve the manufacturers' predicted final minimal stent diameter. Cypher Select stent and insulin treatment were independent predictors of not achieving a minimum stent area >5.0 mm2.