• Potential conflict of interest: Nothing to report.

We appreciate the interest by Zheng et al. in our recent paper.[1] However, we have serious concerns regarding their suggestion that venous blood ammonia and cerebral blood flow (CBF; mL blood/mL brain tissue/min) are correlated in hepatic encephalopathy (HE) and that magnetic resonance imaging (MRI) can replace positron emission tomography (PET) for studies of cerebral ammonia metabolism.

First, the main purpose of our study was to test the hypotheses of decreased CBF and cerebral metabolic rate of oxygen (CMRO2; μmol/mL brain tissue/min) and increased arterial blood ammonia and cerebral metabolic rate of ammonia (CMRA; μmol/mL brain tissue/min) during HE in a paired study. Our findings showed that CBF and CMRO2 were significantly lower during HE than after recovery and that arterial blood ammonia was significantly higher, whereas CMRA was unchanged.

Zheng et al. suggest that, based on observations in a single patient before and after transjugular intrahepatic portosystemic shunt (TIPS), CBF may be negatively correlated with venous blood ammonia concentration. To this point, as seen in Fig. 4C in our study,[1] our paired observations in seven subjects examined during and after recovery from HE showed no significant relationship between changes in CBF and arterial ammonia blood concentration. In addition, in another paired study from our group,[2] individual changes in CBF after TIPS (n = 9) did not correlate with changes in arterial blood ammonia.

An important issue in this context is that measurements of ammonia concentration in venous blood cannot replace measurements in arterial blood for pathophysiological studies, because it is the arterial blood concentration to which the brain is exposed, and the relationship between arterial and venous blood concentrations of ammonia is unpredictable. In fact, venous blood ammonia can be normal in spite of arterial hyperammonemia.[3]

In addition, even when using the arterial blood concentration of ammonia (A; μmol/L arterial blood), CMRA cannot be calculated as CBF · A (i.e., the delivery of ammonia to the brain). CMRA is equal to the A − V difference across the brain multiplied by CBF (V; μmol/L venous blood) and the PET-estimated metabolic clearance of ammonia (Kmet; mL blood/mL brain tissue/min) multiplied by A[4]:

display math

which means that

display math

Thus, CMRA is always lower than CBF · A, but not systematically.[1] Measurement of CMRA thus requires measurements of arterial blood ammonia concentration in conjunction with either measurements of V (blood samples) and simultaneous CBF (PET or MRI) or measurement of Kmet (PET).

Accordingly, we contend that MRI and venous blood ammonia measurements cannot replace PET and arterial ammonia measurements for studies of CMRA.

  • Gitte Dam M.D., Ph.D.1

  • Susanne Keiding M.D., M.D.Sc.1,2

  • Ole L. Munk Ph.D.1

  • Peter Ott M.D., M.D.Sc.2

  • Hendrik Vilstrup M.D., M.D.Sc.2

  • Lasse K. Bak Ph.D.3

  • Helle S. Waagepetersen Ph.D.3

  • Arne Schousboe D.Sc.3

  • Michael S⊘rensen M.D., Ph.D.1,2

  • 1Department of Nuclear Medicine & PET Center, Aarhus University Hospital, Aarhus, Denmark

  • 2Department of Hepato-Gastroenterology, Aarhus University Hospital, Aarhus, Denmark

  • 3Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark