Assessment of liver function: what are we measuring?

Insufficient liver function is the single, most important determinant of outcome after a major liver resection. At the same time, there is not one liver test that reflects all components of the diverse spectrum of liver function.¹ Two papers^2,3 in this issue of HPB deal with the prediction of post-hepatectomy liver failure (PHLF), a dramatic event that is difficult to manage as effective substitution of liver function by means of liver assist devices including the bioartificial liver, is still lacking. PHLF should be the equation of failure to predict accurately liver functional reserve pre-operatively, given the extent of the resection considered necessary to remove all of the tumour and the quality of liver parenchyma. The reason that pre-operative assessment of liver function might fail to predict PHLF is that pre-existing liver injury may be underestimated on the basis of CT volumetric studies. Furthermore, many patients referred for a resection have undergone extensive chemotherapy with consequent steatotic or microvascular changes of their livers. Even when applying dynamic, quantitative liver function tests in the pre-operative work-up, intra-operative circumstances such as deviation of the operative plan or inadvertent massive blood loss, are factors influencing post-operative liver function that cannot be taken into account pre-operatively. The risk assessment of PHLF in the pre-operative setting and a prediction of PHLF after a resection, are therefore distinct. Whereas pre-operative risk assessment is used to select patients amenable to resection, prediction of PHLF is used to anticipate treatment, albeit limited to largely supportive care. A salient point is that patients with PHLF unresponsive to therapy seldom are candidates for salvage (cadaveric) liver transplantation, as the tumour burden underlying these extensive resections usually by far exceeds accepted criteria for liver replacement.

In both papers, from Leeds² and Warsaw,³ PHLF prediction is based on serum liver function tests (LFTs). The term LFT is confusing in the present context as these parameters do not all reflect liver function. The serum transaminase levels [aspartate transaminase (AST) and alanine transaminase (ALT)] are liver damage parameters, and although the extent of hepatocellular necrosis will obviously impact liver functional reserve, they are not liver function tests per se. Perhaps for this reason, AST came out as significant predictor of 90-day mortality in the analysis of the Warsaw group. The International Normalized Ratio (INR) on the other hand is a functional parameter as plasma levels of coagulation factors are correlated with synthetic activity of the liver. Plasma bilirubin, alone or in combination with INR, was identified as a strong predictor of PHLF. These blood tests are quite familiar to us as they appear in various, widely used clinical scoring systems such as the Child–Pugh score⁴ and the MELD (model for end-stage liver disease) score⁵ in the assessment of patients with cirrhosis. Also in the ‘50-50 criteria’ proposed by the Beaujon group in Paris⁶ to predict PHLF, plasma bilirubin and INR are the key parameters along with a time dimension, i.e. the fifth post-operative day.

The plasma bilirubin concentration is a function of uptake, conjugation and excretion of bilirubin by the liver and increased plasma levels indicate serious malfunction of the liver.¹ The authors of both papers have elegantly shown that plasma bilirubin, alone or in combination with INR, is an early predictor of PHLF.^2,3 Although specificity is low, elevated bilirubin on the first postoperative day was shown to be a harbinger of impending PHLF, as defined by the International Study Group for Liver Surgery.⁷ As such, it will provide a valuable factor in the clinical picture of a patient who after a major liver resection, develops drowsiness with derangement of coagulation parameters and increased plasma lactate levels. Except for intensive care unit admission, supplementation of deficient substances as coagulation factors and albumin, as well as broad antibiotic coverage to compensate for the loss of Kupffer cell function, there is little that can be offered in the treatment of these patients. In a series of patients with PHLF treated with the Molecular Adsorbent Recirculating System (MARS) device, our group observed transient clinical and biochemical improvement, but eventually, all patients died because of septic complications.⁸ In my view, active interventions in PHLF are started too late to exert any benefit as recovery of the patient mainly relies on the ability of the remnant liver to regenerate, thereby catching up function. In spite of the strong trigger for regeneration after extensive liver resection, hepatocellular function and proliferation are suppressed by the toxic load with which the already small liver remnant has to cope. Many years spent in the development of a bioartificial liver has taught us that hepatocytes fail to thrive and to function optimally when exposed to the plasma of patients in liver failure.⁹ Early treatment with liver support devices to reduce the toxic load to facilitate regeneration of the remnant liver therefore seems crucial. Although ideally, one might want to use true quantitative functional liver tests such as the indocyanine green clearance test or hepatobiliary scintigraphy to define liver functional reserve after extensive liver resection, measurement of plasma bilirubin on the first post-operative day provides a simple laboratory test to identify patients at an increased risk of developing PHLF.

Conflicts of interest

None declared.