A wide variety of locations have been used for the implantation of hepatocytes in various 
animal models. To date, however, no human studies have been performed. We propose the direct 
re-infusion of autologous human hepatocytes into a tributary of the portal venous system. There 
are benefits as well as drawbacks to this approach. The ability to leave an indwelling Broviac- 
type catheter at the time of hepatocyte harvest allows the administration of hepatocytes without 
the need for further surgery or anesthesia. Patients with FH have premature atherosclerotic 
cardiovascular disease predisposing them to increased cardiovascular morbidity. The ability to 
reinfuse cells without the extra stress brought about by a second laparotomy for reinfusion is 
therefore useful. The patients can be studied radiologically immediately prior to hepatocyte 
reinfusion to insure that the vessel is open and that hepatoportal flow is present. 
There are also theoretical disadvantages to the direct infusion of hepatocytes into the 
portal venous system. The most devastating complication of this approach would be portal vein 
thrombosis. Portal vein thrombosis might complicate a subsequent attempt at orthotopic liver 
transplantation and could lead to variceal hemorrhage and ascites. In a review of the incidence 
of portal vein thrombosis after the intraportal infusion of pancreatic islets, 3 of 57 patients 
developed portal vein thrombosis (109). The cells infused in most cases are portions of 
pancreatic tissue acutely digested and reinfused. The incidence of portal vein thrombosis 
following infusion of autologous hepatocytes should be lower because the cells are autologous 
rather than allogeneic, and the infusate contains a suspension of disaggregated cells rather than 
tissue fragments. 
A number of other forms of hepatocyte delivery have been considered and rejected for a 
variety of reasons. The most direct methods would be the infusion of cells by interventional 
radiologists. There are three possible approaches: 1) hepatic arterial reinfusion via femoral 
puncture (110); 2) percutaneous transhepatic/transplenic injection into the portal vein 
(111); and 3) retrograde injection into the portal vein by a transjugular hepatic venous 
puncture (112). There are risks shared by all three. Bleeding, either overt or occult may be a 
result of these procedures. Also, the risk of portal venous, hepatic arterial or other major 
vascular thrombosis may occur, complicating the future performance of a hepatic transplant. 
Lastly, these procedures require several hours of sedation and in children would almost 
certainly require the use of a general anesthetic. The additional physiologic burden of a second 
anesthetic in a patient population at risk for adverse cardiovascular events cannot be 
overestimated. 
There are several points to consider in the use of interventional techniques which have 
led us not to abandon the concept entirely. This patient population does not have a coagulopathy 
such as is seen in patients with fulminant hepatic failure, so that percutaneous puncture is 
relatively safe. In fact, this will be our preferred route of delivery in case of catheter 
thrombosis prior to reinfusion. Also, the cells can be directed into a subsegmental branch of the 
portal vein and cell migration can be further minimized by the placement of coils or gelfoam. 
The main theoretical advantage of this technique is the possibility of decreasing main portal vein 
thrombosis to a bare minimum by avoiding diffuse showering of cells throughout the portal 
distribution. 
Laparoscopic delivery of hepatocytes into the patient is an attractive but currently 
unfeasible approach. Laparoscopy would also require a second anesthetic, and there are no 
standardized techniques for the isolation of a reasonable branch of the portal venous circulation. 
Such approaches will no doubt be worked out in the near future. 
There are potential problems with this proposed method of hepatocyte introduction such 
as occlusion of the catheter prior to cell infusion. The vessels available for use in children are 
relatively small and are not easily kept patent once cannulated. To minimize the risks of 
thrombosis, the Broviac catheter will be flushed every eight hours with a heparin flush 
[ 186 ] 
Recombinant DNA Research, Volume 15 
