Background: Cyclooxygenase (COX) inhibitors are one of the most often ingested drugs during pregnancy. Unlike general toxicity data, their prenatal toxic effects were not extensively studied before. The aim of the experiment was to evaluate the developmental toxicity of the non-selective (piroxicam) and selective (DFU; 5,5-dimethyl-3-(3-fluorophenyl)-4-(4-methylsulphonyl) phenyl-2(5H)-furanon) COX-2 inhibitors. Methods: Drugs were separately, orally once daily dosed to pregnant rats from day 8 to 21 (GD1 =plug day). Doses were set at 0.3, 3.0 and 30.0 mg/kg for piroxicam and 0.2, 2.0 and 20.0 mg/kg for DFU. Fetuses were delivered on GD 21 and routinely examined. Comprehensive clinical and developmental measurements were done. The pooled statistical analysis for ventricular septal (VSD) and midline (MD) defects was performed for rat fetuses exposed to piroxicam, selective and non-selective COX-2 inhibitor based on present and historic data. Results: Maternal toxicity, intrauterine growth retardation, and increase of external and skeletal variations were found in rats treated with the highest dose of piroxicam. Decrease of fetal length was the only signs of the DFU developmental toxicity observed in pups exposed to the highest compound dose. Lack of teratogenicity was found in piroxicam and DFU-exposed groups. Prenatal exposure to non-selective COX inhibitors increases the risk of VSD and MD when compared to historic control but not with selective COX-2 inhibitors. Conclusion: Both selective and non-selective COX-2 inhibitors were toxic for rats fetuses when administered in the highest dose. Unlike DFU, piroxicam was also highly toxic to the dams. Prenatal exposure to selective COX-2 inhibitors does not increase the risk of ventricular septal and midline defects in rat when compared to non-selective drugs and historic control. (c) 2005 Elsevier Ireland Ltd. All rights reserved.