In reply to the letter of Dr Penny Watson, we would like to make several comments. The general remark is that Dr Watson oversimplifies the subject, which is not entirely justified and does not fit the nuances that are important when addressing the complex problem of hepatic encephalopathy. We share the opinion given in Dr Watson's comment that there is no proof that strict protein restriction is required for dogs with congenital portosystemic shunts. However, moderate protein reduction is still indicated.
Dr Watson makes no distinction between different forms of portosystemic shunting. There are fundamental differences between different types of shunts. The surgically induced portocaval shunt (Eck fistula) constructed in adult dogs provides only a partial representation of the actual situation in congenital shunts or acquired shunting in case of chronic portal hypertension (eg, cirrhosis, congenital portal vein hypoplasia). An abundance of literature indicates that clinical hepatic encephalopathy occurs only when there is a combination of portosystemic shunting and impaired functional liver mass. This occurs in congenital portosystemic shunting because in such cases the liver does not receive stimuli such as growth factors in portal blood that are necessary for normal growth, development, and function of the liver. Dogs with congenital shunts gradually develop small livers in relation to body size, because liver growth lags far behind total body growth. This time-dependent convergence causes clinical signs of hepatic encephalopathy usually when dogs are approximately 6 months of age and also later in life. Surgically created shunts in mature dogs (Ecks fistulas) cause only a slight decrease in liver mass, and such dogs do not develop hepatic encephalopathy unless given highly ammoniagenic proteins. In cirrhosis, the degree of portosystemic shunting is usually much less than in dogs with congenital shunts, and the degree of liver dysfunction is very variable ranging from moderate to severe. In such dogs and humans, the severe portal hypertension with ascites and associated complications form an integral part of the pathophysiology. Of all these pathologies, congential portosystemic shunts are associated with the most severe (nearly 100%) portosystemic shunting, and, in contrast to other types of pathology, congenital shunts are not complicated by all of the consequences of portal hypertension. Congenital shunts may be considered the best-known example in human and veterinary medicine of pure hepatic encephalopathy.
Our manuscript addressed congenital portosystemic shunts in dogs, and it would be best to restrict the discussion to this limited condition without confusing the readers with data from experimentally created shunts in adult dogs or the human literature on cirrhosis, including the referred article in the Lancet.
It has been long known that the intermediate metabolism of ammonia occurs in many different tissues. Ammonia is removed by different cell types by incorporation into glutamine, in addition to urea production in the liver, and both the small intestines and the kidneys can produce large amounts of ammonia. The net outcome depends very much on the quality of liver function and metabolic factors such as catabolism and pH regulation. We have previously published on the effects of protein sources that contain beneficial branched chain rather than aromatic amino acids. These diets performed less well in preventing hepatic encephalopathy because control dogs preferred the more palatable control food with lower branched chain/aromatic amino acid ratio and were not catabolic. Dr Watson stresses the negative effects of very low protein diets (eg renal diets) in dogs with congenital portosystemic shunts. We agree completely with this comment. It also is our experience that dogs with congenital shunts when kept on very low protein diets develop severe hypoalbuminemia and ascites after 6–12 months. However, the diets tested in our study were only mildly restricted (ie, 16% of calories from protein as compared with 12% in commonly prescribed renal diets) with respect to protein content. Not only quantity of protein but also quality (eg, digestibility, balance of amino acids) of protein matters. We have never seen adverse effects with such formulations. The composition also prevented any chance of a catabolic state developing. In addition, it is unjustified to draw far-reaching conclusions from the ranges given in Table 2. It is impossible to follow individual dogs in this table, and conclusions should be based on the statistics of the entire group in such a relatively small study.
We think it would be an oversimplification to state that colonic ammonia production is unimportant, and that the small intestine is the only intestinal source of ammonia. The less digestible source of protein in our study (ie, poultry-based) resulted in higher plasma ammonia concentrations. Of course, both aspects are important, but the principal nitrogen source remains the uptake of nitrogenous constituents from the intestinal lumen. Later, many different tissues, including the small intestinal wall, take part in the complex intermediate metabolism of ammonia.
We think that the situation is more complex than represented by Dr Watson in her comments. In light of the current knowledge, consensus should be based on moderately protein-restricted diets, and then finding the best dietary composition. Our controlled, randomized, and blinded study was an attempt to contribute to the understanding of the medical management of dogs with portosystemic shunts. What is needed is a randomized, long-term study in dogs with congenital portosystemic shunts to see whether dietary support or surgery produces the best long-term results for survival and quality of life.
On behalf of the authors, Sarah Proot, Jan Rothuizen, and Vincent Biourge.