In Reply to: Re: Case 1097_02 -- Portal vein thrombosis in the setting of acute pancreatitis: treatment with portal reconstruction posted by Eric Benson on October 21, 1997 at 18:12:09:
In the last five years I've treated a fair number of patients with portal system thromboses, due to benign or malignant causes, with clots in the portal, splenic and mesenteric veins (>30+). I used to approach all of them from a semi-lunar approach: groin (angiogram), side, then neck. In the last several years I've switched to a purely transjugular approach. Its worked in all cases.
The technical tradeoffs include the difficulty and necessity of puncturing an occluded portal branch, recognizing it, and recanalizing with a colapinto needle for stability. On the other hand, with practice and some hydrophilic wires or L-R torque wires, this works well and is ultimately Much faster for me. It also avoids the difficulty of having to puncture a peripherally patent, occasionally tiny portal branch (sometimes using a tandem Chiba needle technique), making sure its angle is suitable for working down into the main pv and finally moving to the neck to marry it all up with a TIPS.
I have found that only a small select group of patients actually have no component of cirrhotic liver disease and no portal hypertension. If you've got that rare patient, then a portal stent alone may be enough. Remember that if the clot extends deep into the liver, though the liver is not cirrhotic, portal pressures may remain elevated; there is still presinusoidal portal hypertension.
More often than not, the flow and pressures are still elevated and the risk of stent thrombosis is non-trivial. Also, a TIPS provides an access for stent revision if needed. If pressures are marginally increased (mild portal htn), leave behind a narrow TIPS, and cease to revise it if the portal stent proves durable. I've done this successfully in cases of mesenteric venous thrombolysis. In some cases, the artificially created low pressure outflow of the TIPS has contributed to more rapid lysis of residual clot. The latter is a phenomenon I have witnessed many times.
In patients not treated for bleeding, I'd strongly consider a course of anticoagulation (perhaps 1-2 months maximum) to prevent early thrombosis. There is a randomized Dutch trial that supports this.
Finally, if in doubt as to where to extend your stents (as stents should generally extend from patent vessel to patent vessel through the clot), run them into the splenic vein. This provides through the shunt decompression of the typical bleeding source (gastroesophageal varices) and is also fine for ascites patients. This will not interfere with a surgeon's later desire to transplant the patient and perform a mesoportal anastomosis. One does not need to run the shunt into the SMV to "preserve shunt flow and patency."
The exception is the rare case of a patient bleeding from intestinal varices wherein the shunts should be extended into the SMV. Remeber that even in this case, the upper SMV lies within the pancreas and stenting the intrapancreatic portion of the SMV (and not below it) should not interfere with later mesoportal anastomoses. If the clot extended below the level of the left colic vein, then often the whole SMV becomes thrombosed.
Last week I performed a follow up shunt venogram on a patient whose "TIPS" stents runs from splenic hilum to hepatic vein. He had had portal and splenic vein thrombosis that was over a year old and absolutely massive ascites. The shunt is widely patent, his ascites is non-existent, his nutrition is greatly improved (presumably due to a sense of well-being, decreased bowel wall edema, hyperinsulinism, etc). He's back to driving in the demolition derby.