Changani et al. correctly state that it is conventional to reference the resonances in a 31P-nuclear magnetic resonance spectrum relative to the phosphocreatinine (PCr) resonance, which is-if present-set at 0 ppm. When, however, no PCr is present, another spectral peak has to be used for reference. As Changani et al. mention, one can use the α-ATP for reference, which is quite convenient when using experimental setups with high signal-to-noise ratios, for example, when working at field strengths of 11.7 Tesla. This is even more convenient when the investigator can use freshly harvested rat livers, tissue extracts, or cell cultures. Clinical liver transplantation, however, is different: Most of the livers that we studied came from long distances with accordingly long cold ischemia times. As is well known, the longer the cold ischemia time, the lower the total adenine nucleotide (TAN) levels. In most of the livers that we studied, there was hardly any α-ATP to be detected. In contrast, NADH was present in all livers and could be used for reference, because the NADH peak position is pH-independent and the peak is stable over longer periods of time. The use of NADH as a reference peak is adapted from previous experiments on rat livers by other authors (1,2). We have successfully used this method to study tissue pH changes in cold stored human livers during cold preservation (3). Changani et al. furthermore ask what could be the origin of the small but distinct peak attributed (by us) to PCr. Their statement that PCr levels drop very rapidly after the onset of ischemia is absolutely correct; this has been observed in (too?) many animal experiments. Their suggestion that PCr peaks may originate from muscle cells in the extrahepatic biliary tree and great vessels is interesting. However, if that were the case, we think all livers in our study should have shown the PCr peak, which was not observed. The spectrum that Changani et al. provide to convince us that a liver does not contain spectroscopically detectable PCr is recorded from a perchloric acid extract of normal human liver (presumably a liver biopsy). Our spectra were recorded from cold stored whole human livers after in situ preservation perfusion and several hours of cold ischemia. These events have probably had a large impact on liver adenine (and energy) metabolism, and this is probably the basic reason for the presence of a peak at the PCr site. Indeed, we did not prove (for example, by means of high-performance liquid chromatography or high-field strength nuclear magnetic resonance) whether this peak was actually PCr, since that was not the purpose of our study; thus, the true origin of the peak at the PCr site remains to be determined. Changani et al. furthermore remark that the sequential spectra show better resolution without indication of a second inorganic phosphate pool. As the figure to which they refer was taken from a previous publication(4) (as the figure legend mentions), different spectral processing parameters were employed which account for the optical difference between the respective spectra. Their remark that the concentration of inorganic phosphate in University of Wisconsin preservation solution is significantly higher than the intracellular phosphate level is true. For exactly that reason, we positioned the donor liver as close as possible against the precooled spectroscopy coil to prevent "drowning" of the hepatic inorganic phosphate signal in the University of Wisconsin inorganic phosphate signal. We agree with Changani et al.'s final conclusions that 31P magnetic resonance spectroscopy is the only rapid nonivasive means of acquiring biochemical information prior to transplantation. Based on its ability to monitor on-line energy metabolism in cold stored organs, this technique may help to develop better preservation solutions. On the basis of our results, however, we are not convinced that 31P magnetic resonance spectroscopy is yet a reliable clinical instrument to provide data that can convince a clinician to proceed to transplant the liver or to discard the liver because of an "abnormal spectrum." R.F.E. Wolf Department of Radiology; University Hospital; 9700 RB Groningen, The Netherlands