The final question, in terms of efficacy, is the potential benefit _ of a persistent but 
incomplete correction of hypercholesterolemia in improving the morbidity and mortality due to 
CAD. There is a huge volume of epidemiologic data that supports a relationship between LDL 
cholesterol and the development of CAD in the general population (117-120). The disease FH 
also confirms this concept in that homozygotes with LDL-cholesterol levels in the 500 to 1000 
mg/dl range have much worst disease than heterozygotes with LDL-cholesterol levels in the 
300 mg/dl range (2). It seems logical to assume that a 33% decline in LDL-cholesterol from 
800 mg/dl to 560 mg/dl, for example, may be therapeutic. Genetic heterogeneity in the FH 
population has provided some insight into the possible relationships between modest differences 
in serum cholesterol, such as that expected from gene therapy, and the natural history of the 
disease. Brewer and colleagues have examined this relationship in a series of studies performed 
at the NIH. They showed a negative correlation between residual LDL receptor activity (2-29% 
of normal) and serum LDL-cholesterol (838-304 mg/dl) and a positive correlation between 
LDL receptor activity and age of onset of angina (41). A separate study categorized these 
patients with respect to presence or absence of CAD (43). The CAD symptomatic group had 
substantially greater LDL cholesterols (817 +/- 62 mg/dl) than the asymptomatic group 
(561 +/- mg/dl). Four of seven patients died in the symptomatic group during the course of 
the study. 
B. Risks of Hepatic Resection and Hepatocyte Infusion 
Resection of the left lateral segment of the liver is a fairly straightforward surgical 
procedure that is associated with some risk. The liver resection should be accomplished without 
the necessity of a transfusion. Mortality as a result of this procedure should be low; the most 
likely cause of mortality would be a cardiac event occurring in the perioperative period. The 
incidence of this should be approximately 1% (95,96). The patient will undergo extensive 
preoperative evaluation and intraoperative hemodynamic monitoring to minimize these risks. 
Possible noncardiac morbidity could result from the liver resection and catheter placement in 
the portal circulation. These include 1) excessive bleeding as a result of the resection, 2) 
infection at the site of the hepatectomy or via the indwelling catheter, and 3) leakage of bile 
causing peritonitis. This procedure might make subsequent orthotopic liver transplantation 
more difficult but it wouldn't preclude a successful organ transplant. Removal of the catheter 
should be associated with little morbidity. Removal will involve no further anesthetic, and can 
be done at the bedside. 
Infusion of the hepatocytes via the indwelling Broviac catheter into the portal 
circulation is also associated with risks. The most morbid complication of this procedure is 
portal vein thrombosis. There are no data in humans on the incidence of portal vein thrombosis 
after the intraportal administration of autologous hepatocytes, but when pancreatic islet preps 
are used, the incidence was about 4% (109). We expect a lower incidence of portal vein 
thrombosis with autologous hepatocytes because the cells are autologous rather than allogeneic 
and the infusate contains a suspension of disaggregated cells rather than tissue fragments. 
Complete thrombosis with organized clot adherent to the endothelium of the portal vein may 
preclude subsequent liver transplantation. In addition, the sequalae of portal vein thrombosis, 
portal hypertension with the development of ascites or variceal hemorrhage may cause clinical 
deterioration in these patients with diminished cardiovascular reserves. The studies done in the 
WHHL rabbit have not shown portal vein thrombosis to date. Furthermore, our experience in 
the baboon model indicates that portal vein thrombosis or portal vein hypertension may not be a 
problem. Treatment of the complications of portal hypertension are relatively standard. 
Additional complications of the hepatocyte infusion are 1) anaphylaxis, and 2) infection of the 
Broviac catheter. Removal will involve no further anesthetic, and can be done at the bedside. It 
is a common procedure. 
Recombinant DNA Research, Volume 15 
[193] 
