The observability of nucleoside triphosphate (NTP) by 31P NMR spectroscopy was studied in the isolated rat liver during hypothermic perfusion and a subsequent 4-h cold ischemia. The influence of hypothermia (4 deg.C) was examined because of its delaying effects on cell injury induced by the ischemic conditions. The viability of the liver after hypothermic ischemia was assessed by measuring the recovery of the βB-NTP resonance after reperfusion. In 4-h cold ischemic liver, recovery was found to be in the range of 90-100% and consequently NTP visibility was studied under these conditions. Because the individual purine (or pyrimidine) NTPs are not distinguishable in the liver on the basis of their 31P NMR chemical shifts, the contributions of UTP and GTP were investigated by HPLC. The changes in liver NTP content measured either by NMR on isolated liver or by HPLC after perchloric acid extraction from the same organ are not significantly different. The total NTP level in normothermic perfused liver is 7.6±0.2 ±umol NTP/g liver dry wt as determined by NMR. In such a liver, ATP + GTP + UTP and ATP contents measured by HPLC are, respectively, 7.9±1.0 and 6.3±0.9 ±mol/g liver dry wt. This indicates that all NTP is detected by NMR and that a 20% contribution of the signal occurs from UTP + GTP. Under 4-h cold ischemic conditions, NTP visibility remains unchanged, furthermore the UTP+GTP contribution reaches 32% of the whole NTP content. Further experiments in 6-h cold ischemic livers show that βB-NTP signals are at the limit of detection whereas the ATP+GTP+UTP level, measured by HPLC, still represents 20% of the basal value. This result suggests that NTP remains fully detectable by NMR in liver as long as cellular integrity is maintained.