We tested the hypothesis that the sedative, euphoric, and analgesic ef
fects of intravenous fentanyl would distinguish intravenous from epidu
ral administration. One hundred RSA I and II labouring parturients rec
eived 100 mu g fentanyl either iv or via an epidural catheter in a dou
ble-blind, randomized, crossover fashion. Nineteen anaesthetists (8 st
aff and 11 residents) participated and correctly guessed the route of
administration of the fentanyl in 61/66 intravenous doses and in 69/75
epidural doses yielding a sensitivity of 92.4%, a specificity of 92.0
%, a positive predictive value of 91.0%, and a negative predictive val
ue of 93.2%. Of the 41 patients that were crossed over, 38 (92.7%) wer
e able to detect a difference between the routes of administration. Mo
st patients experienced prompt, short-lived symptoms with iv fentanyl
but no important differences in fetal heart rate pattern or in materna
l desaturation were seen between the groups. This study suggests that
subjective symptoms will accurately distinguish intravenous from epidu
ral fentanyl administration in labouring parturients (P < 0.001), and
should serve as a safe and reliable intravenous test dose for epidural
anaesthesia in the obstetric population.