A Review of Transsplenic Access for Portal Vein Interventions

History

Transsplenic portography was one of the first procedures performed in the developing specialty of interventional radiology. The technique was first described by Abeatici and Campi in 1951. ¹ The procedure continued with various techniques until a standardized approach to these procedures was developed by Kreel in 1970 to increase the safety of the procedure and consistency of results. ² Transsplenic portography had a relatively high rate of complications but was an important diagnostic tool until computed tomography (CT) rendered the procedure obsolete for diagnostic purposes. ³ Portography from this approach allowed the characterization of portal anatomy, direct measurement of portal pressure, characterization of varices, treatment of variceal bleeding, evaluation of infections, and diagnosis of hepatic tumors. The general procedure involved blind insertion of a 19-gauge needle into the 9th to 10th intercostal space with a slightly upward and posterior trajectory. The trajectory was adjusted until blood dripping from the needle indicated intravascular placement. A splenic pressure measurement was then taken (normal: 8–14 mm Hg from a transsplenic 19-gauge needle). 50 mL of 50% contrast was then injected and imaging was performed with cut-film. At the end of the procedure, the needle was removed without any maneuvers to aid hemostasis (as these embolic techniques were relatively unknown at that time). There was significant bleeding risk with these early procedures, and as such, the procedure was largely abandoned in the late 1970s with the wide adoption of abdominal CT. It was not until the early 2000s that renewed interest in transsplenic access led to series entering the literature with a new focus on therapeutic interventions from a splenic vein approach. Of note, these procedures had a significantly lower complication rate compared with the early studies in part owing to the ability to use advanced hemostatic maneuvers. Since 2009, there have been several small studies demonstrating safe and effective transsplenic interventions. In a series of 46 patients who underwent endovascular portal vein interventions via a transsplenic approach, technical success was achieved in 44 procedures, major bleeding complicated 3 procedures, and 6 patients experienced minor bleeding. ⁴ In a series of 22 pediatric patients who underwent successful transsplenic procedures, there was 1 major bleed and 1 minor bleed. ⁵ In another series on pediatric patients, 12 of 44 had significant bleeding. ³ In a series of 11 adults in which transsplenic access was performed to allow portal venous reconstruction (PVR) with transjugular intrahepatic portosystemic shunt (TIPS) placement in chronically thrombosed portal veins, 11 of 11 had TIPS placed successfully with no bleeding complications. ⁶ The available series are small but do give useful data on method of access and hemostasis given the variability in outcomes. In the series with the highest bleeding complications, 12 of 44 (27%), gel foam alone was used for tract embolization. ³ In Habib's TIPS-PVR series in which coils were utilized, there were no bleeding complications. ⁶

Although the transjugular and transhepatic approaches are more widely applied in clinical practice, developing skills with transsplenic approaches allows treatment to a broader set of patients.

Procedural Basics

In our practice, we have developed a standardized approach to these patients to achieve safe and consistent results. This is what we are describing below, and this approach has been utilized consistently in more than 100 patients.

Current indications for splenic vein access include:

• Portal vein recanalization in chronic or acute thrombosis.

• Variceal embolization.

• Portal/mesenteric/splenic venoplasty and/or stenting.

• TIPS declot when not possible from a jugular approach.

• Portal vein declot.

Procedural Setup and Performance

It is often helpful to put a wedge under the left side of the patient to allow access to a wider range of intrasplenic vein tributaries. We perform the procedure with direct ultrasound visualization of the splenic veins for needle puncture. It is possible to perform the procedure with only fluoroscopic guidance when there is poor splenic visualization (which can be seen in postoperative patients); our general technique would be to obtain a very clear idea of how deep the spleen extends beyond the abdominal wall, either with a cone beam CT on the table or preprocedure conventional helical CT. The aim is to minimize traversing the capsule with needle passes. Under fluoroscopic guidance, a 21-gauge needle is extended through the 9th to 10th rib space posterior and superior and then slowly retracted with gentle suction on a syringe filled with contrast until there is a significant flow of blood. The needle is then injected to confirm the branch accessed is not an artery and is central enough to allow wire passage. For ultrasound technique, which is our preference, central veins need to be targeted as the veins become significantly more tortuous as they progress through the splenic parenchyma. We avoid traversing the capsule of the spleen past the hilum with the needle, and only advance the needle with real-time ultrasound visualization of the tip. The arteries have a wall that is echogenic relative to the veins and tend to be on the cephalad side of the veins ( Fig. 1 ). A 21-gauge needle is advanced into a central vessel (but still within the parenchyma). If possible, the echogenic wall should be avoided to prevent accessing through the artery ( Fig. 1 ). Generally, a transitional dilator system is utilized to access the main splenic vein to allow initial access with a 21-gauge needle and increase the safety of the procedure. We perform venography at this point to confirm we have not traversed the vein and accessed an artery. In such cases, venography will demonstrate flow toward the splenic periphery as opposed to centrally and away from the spleen ( Fig. 2 ). After upsizing to an 0.035″ wire, we place a 6-Fr 30-cm braided sheath as our working access for stability. A 6-Fr sheath accommodates balloons up to 12 mm and nitinol self-expanding stents up to 14 mm. This also allows us to perform single-session thrombolysis using rheolytic devices or rotatory devices. As the spleen can expand and contract, we do not perform overnight TPA drip thrombolysis from a transsplenic approach, as the access site can expand and allow bleeding. After completing the procedure, the sheath is retracted into the splenic parenchymal tract and a 4- or 5-Fr radiopaque-tipped catheter is extended to the end of the sheath. We utilize 0.035″ 4-mm diameter 14-cm nester coils (Cook Medical, Bloomington, IN). It generally requires one to two coils to embolize the entire tract. The spleen is spongy in texture, so we coil in the parenchyma until we see the coil constrained by the capsule of the spleen. This is possible because the 4-mm coil is slightly oversized relative to the 6-Fr sheath and as such within the parenchyma of the spleen the coil will form slightly larger than the sheath but at the capsule it will be constrained to the size of the 6-Fr sheath ( Fig. 3 ). With this method of tract hemostasis, we have an excellent safety profile with hemodynamically significant bleeding uncommon even in patients with significant portal hypertension, coagulopathy, or very young patients. It is possible to immediately anticoagulate the patients after successful tract closure in this way, even in the setting of ascites.

Fig. 1.

( a, b ) The appearance of the splenic vein with the adjacent artery near the hilum of the spleen. Note the echogenic walls of the arterial side (arrow— a ).

Fig. 2.

Venographic appearance of artery opacifying during venography—of note the flow is toward the spleen not away and the vessels are smaller and more delicate. In this case, this was traversal of a small arterial branch while accessing the vein. The vein was appropriately accessed, and the patient had no complication from this finding after completion of the procedure. Blue arrow—the entry vein. Yellow arrows—arterial radicals opacified by the venous puncture that crossed arterial vessels.

Fig. 3.

Appearance of coil within the spleen with coil constraint when the coil is in the capsule of the spleen.

Discussion and Example Cases

Even in very small children, it is possible to perform procedures from a transsplenic approach. The spleen is oversized in infants relative to adults (this is also true for the liver). In this case, a 3.6-kg baby postoperative from a failed Kasai procedure developed portal vein thrombosis. The portal vein was accessed via a 4-Fr low-profile sheath and the main portal vein was recanalized and ballooned with improvement in appearance, without bleeding complication, and improvement clinically following the procedure ( Fig. 4 ). In follow-up, the portal vein remained patent over the next year and the patient eventually was transplanted when he grew enough for this to be clinically and anatomically possible. In this pediatric case, the wire went straight to the right when inserted through the splenic access sheath. Varices often do not take a straight trajectory, and as we suspected the canalized vessel may not have been a varix, we placed a 4-Fr catheter over the wire which confirmed placement within the portal vein ( Fig. 5 ). Care should be made not to retract the wire when a straight course is observed during a portal recanalization case, as this trajectory is indicative more often of the native portal vein than a varix.

Fig. 4.

Portal recanalization in a 3.6-kg infant post Kasai procedure. The portal vein is not visualized; however, the trajectory of the wire suggests portal vein placement. ( a ) Splenic appearance prior to puncture. ( b ) 0.018″ wire (arrows) taking the expected trajectory of the splenic vein. ( c ) Initial venogram from the splenic side demonstrating significant varices (arrow) and nonopacification of the portal vein. Note the wire extending in a straight trajectory to the right (arrowhead). ( d ) Injection of a 4-Fr catheter which was carefully placed over the microwire due to the trajectory suggesting portal vein placement. Injection confirmed that the wire was within the native portal vein. ( e ) Balloon venoplasty (arrow) of the portal vein at the area of narrowing. ( f ) Completion venogram demonstrating that the portal vein now opacifies following the balloon treatment of the portal vein. Arrow—main portal vein.

Fig. 5.

Portal recanalization in a patient with a 15-year history of portal thrombosis following a liver transplant. From the splenic side, the portal anastomosis was treated with venoplasty and stenting, with complete resolution of symptoms following this treatment. ( a ) Initial venogram demonstrating hepatofugal flow of all portal blood through a very large direct splenorenal shunt. ( b, c ) Navigation with a catheter and glide wire with access to the occluded portal vein (arrow, b ) which was confirmed with opacification of native portal radicals (arrows, c ). ( d ) Following balloon venoplasty, marked enlargement of the vein and hepatopetal flow was noted; however, because of the waist on the anastomotic site a self-expanding nitinol stent was placed to prevent restenosis. ( e ) The large variceal connection to the renal was treated with coils and gel foam to favor blood flow toward the newly opened portal vein. ( f ) Follow-up CT demonstrating patency of the stent (arrow) and flow in the intrahepatic portal vein which was not previously visible on CT.

In a patient with longstanding portal vein thrombosis, access via a transhepatic approach can be technically difficult or impossible secondary to cavernous transformation and difficulty in selecting the native portal vein. The next example is of a patient with an orthotopic liver transplant that developed portal thrombosis 15 years prior to presentation. He was suffering from refractory encephalopathy and failure to thrive despite no evidence of cirrhosis on biopsy. In spite of the long interval of occlusion, the portal vein was able to be treated from a transsplenic approach. The native portal vein was selected and the vein was treated with balloon venoplasty and a self-expanding nitinol stent which allowed hepatopetal flow into the liver which, in conjunction with embolization of large splenorenal varices, resulted in complete resolution of the patient's encephalopathy and failure to thrive ( Fig. 5 ).

In patients with (TIPS) thrombosis, it is often possible to recanalize the stents from a transjugular approach. In some cases, the cranial (hepatic vein) aspect of the stent is not accessible from the transjugular approach. In the last case example, the patient presented following three prior unsuccessful transjugular TIPS recanalizations. This is despite the prior attempt utilizing intracardiac echo probe to facilitate visualization. The patient was accessed from a transsplenic approach and the TIPS was successfully accessed from the portal side. After traversing the clot and getting through-and-through wire access from the splenic access to the right internal jugular vein, the stent was able to be successfully thrombolysed and a stent was utilized to extend the TIPS more centrally as seen in Fig. 6 . This redemonstrates the wire going straight into the TIPS from the portal (splenic access) side as well as the mechanical advantage going from the spleen as opposed to a transhepatic approach which places access at an angle that confers mechanical disadvantage because of the angle required to access the portal vein. The patient's ascites resolved, and esophageal varices decompressed at next endoscopy.

Fig. 6.

A transjugular intrahepatic portosystemic shunt (TIPS) was unsuccessfully recanalized from a jugular approach after three attempts. Transsplenic access was obtained, and the TIPS was able to be recanalized with resolution of the patient's symptoms. ( a ) Despite the difficulties of the prior attempts, the wire from a transsplenic approach immediately tracked to the portal end of the TIPS (arrow). ( b ) Initial venogram showing TIPS thrombosis (circle) with hepatopetal flow in non-TIPS vessels. ( c ) The hepatic vein was able to be accessed from the portal side with a catheter and wire (arrow). ( d ) Using a combination of jugular and transsplenic approach the occlusion was crossed with a catheter and through-and-through access was obtained. ( e ) Balloon venoplasty improved flow through the TIPS (arrows). ( f ) After extending the central (portal-side) aspect of the TIPS, good flow (arrow) and resolution of the high portal gradient was achieved.

Conclusion

Portal vein interventions are becoming increasingly common due to the innovation of more endovascular procedures to treat portal pathology, as well as the background of complex surgical procedures being performed, the safety of the procedures, and the growing awareness of referring providers for these cases. Adding therapeutic procedures from a transsplenic approach to an interventional radiologist's repertoire can significantly expand the interventionalist's ability to treat complex portal pathology without significant increase in the risks if careful technique is utilized. This is particularly useful in the patient with intrahepatic portal vein malignancy located along the trajectory of potential transhepatic access, in patients with postsurgical portal vein thrombosis, patients with a small transhepatic window, or when the TIPS and direct intrahepatic portocaval shunt procedures simply are not possible. There is also relative mechanical advantage as compared with transhepatic approaches, particularly if both the left and right portal veins need to be accessed for a therapeutic procedure. Bleeding, while much less common with modern ultrasound-guided puncture, 21-gauge needle access, and coil tract embolization, remains a known complication but has at present reached an acceptable threshold even in the end-stage liver or immediate postoperative patient for this to become a standard approach for an advanced interventional radiology practice.