- Email: [email protected]
Spleno-adrenal shunt: A novel alternative for portosystemic decompression in children with portal vein cavernous transformation
Journal of Pediatric Surgery (2012) 47, 2189–2193
www.elsevier.com/locate/jpedsurg
Spleno-adrenal shunt: A novel alternative for portosystemic decompression in children with portal vein cavernous transformation Song Gu a,⁎, Shirong Chang b, Jun Chu a, Min Xu a, Zhilong Yan a, Donald C. Liu b,1,2, Qimin Chen a a
Shanghai Children's Medical Center, Shanghai Jiao Tong University, School of Medicine, Shanghai, China University of Chicago, Comer Children's Hospital, Chicago, IL 60637, USA
b
Received 24 August 2012; accepted 1 September 2012
Key words: Spleno-adrenal shunt; Portal hypertension; Children; Portal vein cavernous transformation
Abstract Purpose: Children with portal vein cavernous transformation (PVCT) can develop life-threatening variceal hemorrhage from progressive portal hypertension. While spleno-renal shunt±splenectomy is the most common portosystemic decompression surgery performed in children, we have adopted a modified spleno-adrenal (SA) shunt for complicated PVCT. We describe our 10 year experience focusing on technique evolution and treatment efficacy. Methods: Between 2001 and 2011, 15 children (9 girls and 6 boys, ages 3–11years, median: 6years) with PVCT, portal hypertension, and hypersplenism were treated with SA shunt with splenectomy in Shanghai Children's Medical Center. All children in the study had endoscopy proven active esophageal variceal bleeding requiring multiple transfusions (mean: 4.2 units) with failed sclerotherapy (mean: 2.6 times). Greater omental vein pressure (GVP) approximating portal venous pressure was measured pre- and post-SA shunt. Pre- and postoperative ammonia levels were obtained. Follow-up ranged from 6months to 10years (mean: 4.2±2years). Results: Intra-operative adrenal vein diameter and length ranged from 0.7 to 1.8 cm and 2 to 3 cm, respectively. Intra-operative GVPs pre-and post-SA shunt were (30 ±11) and (22 ±7) mmHg, respectively (pb0.01). On follow-up, there have been no recurrences of GI bleeding. Liver function tests remained normal in all children with the exception of elevated post-operative mean blood ammonia levels [Pre (18±7) mmol/L, post (60±17) mmol/L (pb0.05)] in all children. Ammonia levels normalized in all cases on outpatient follow-up. There have been no cases of hepatic encephalopathy, and all have normal age appropriate neurodevelopment (Bayley's assessment). Barium swallow and/or upper endoscopy showed interval resolution of esophageal varices in all children, and vascular ultrasound showed patent shunt anastomosis without stricture in 14 (93%). Conclusions: The left adrenal vein is a viable conduit for effective selective portosystemic decompression. Similar to the more traditional spleno-renal shunt, SA appears also to have the advantage of preventing hepatic encephalopathy preserving neurodevelopment, although the rise in post-
⁎ Corresponding author. Tel.: +1 773 702 6175; fax: +1 773 702 1192. E-mail addresses: [email protected], [email protected] (S. Gu). 1 Co-senior author. 2 Deceased. 0022-3468/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2012.09.007
2190
S. Gu et al. operative ammonia levels was unexpected. Longer follow-up is needed to look for late signs of encephalopathy assessing neurodevelopment long term. © 2012 Elsevier Inc. All rights reserved.
Extrahepatic portal vein obstruction is an important cause of portal hypertension among children [1-3]. It is characterized mainly by portal vein cavernous transformation (PVCT) which is defined as the formation of venous channels within or around a previously thrombosed portal vein and is detected in 40% of children with upper gastrointestinal bleeding caused by esophageal varices. Approximately 80% of children diagnosed with portal vein thrombosis will have at least one episode of variceal bleeding which is often life threatening. In spite of the advancement in technology and improvement of therapeutic modalities in pharmacology, endoscopy (sclerotherapy or band ligation), and interventional radiology (transjugular intrahepatic portosystemic shunt); surgical decompressive shunts still play an important role in the treatment of variceal bleeding caused by portal hypertension [4,5]. The success rate of bleeding control with endoscopy and pharmacological agents is still poor at only fifty percent [6]. Although children with PVCT characteristically have normal liver function without cirrhosis, they still are at significant risk for severe hypersplenism and life threatening variceal hemorrhage secondary to progressive portal hypertension. Despite the advent of nonsurgical methods to control portal hypertensive bleeding and further refinements of liver transplantation in children, a significant role for portosystemic shunting for pediatric portal hypertension continues to evolve. Distal splenorenal shunt (DSRS) and/or selective portocaval shunt are still considered the preferable surgical procedures in children with preserved liver function and absent or minimal cirrhosis, but have significant medically refractory variceal hemorrhage secondary to progressive portal hypertension [7-10]. Often concomitant significant splenomegaly secondary to hypersplenism develops requiring splenorenal shunt with splenectomy. The technique of DSRS was first illustrated by Warren in 1967 describing an end-to-side anastomosis between the splenic vein and left renal vein [11]. Technical difficulties of DSRS mainly entail difficulties in establishing a favorable anastomosis due to several factors which mainly include splenic vein diameter, discrepancy in diameter caliber between the two veins and often an "awkward" angle between the two veins for a suitable anastomosis. A few reports in the literature have described the use of the adrenal vein as a viable, potentially more favorable conduit to achieve effective selective portal decompression compared to DSRS [12-14]. In this report, we describe the use of the spleno-adrenal (SA shunt) as our preferred technique for selective portal decompression in a series of 15 consecutive children with portal vein cavernous transformation (PVCT) suffering life threatening variceal hemorrhage from progressive portal hypertension.
1. Methods Under IRB approval, we examined our computerized database to identify fifteen consecutive children (9 girls and 6 boys, ages 3–11 years, median: 6 years) with portal vein cavernous transformation (PVCT) who underwent SA shunt in a ten year period between 2001 and 2011. Subjects were followed for periods ranging between six months and ten years (mean: 4.2±2 years) and all operations were carried out as elective procedures. Medical records were reviewed for demographic data; cause of portal hypertension (PVCT in all cases; 12-history of umbilical catheterization; 3-idiopathic); thirty day operative mortality; peri- and post-operative morbidities including rebleeding episodes, ascites, and clinical grade of portosystemic encephalopathy; short and long term liver function and shunt patency (Doppler ultrasonography); and overall survival. Indications for SA shunt with splenectomy operation in all children were major gastrointestinal bleeding and clinically significant hypersplenism. All children in the study had endoscopically proven active esophageal ±gastric variceal bleeding requiring multiple units of transfusions (mean 4.2; range 2–9) with failed number of attempts of sclerotherapy (mean 2.6). Medical management involved a stepwise approach of volume resuscitation with crystalloids and blood products ±octreotide or vasopressin infusion in the acute setting. All patients underwent early endoscopy after gastroesophageal lavage to determine the extent of gastroesophageal varices and the need for sclerotherapy. Further evaluation of the cause of the portal hypertension included abdominal Doppler ultrasonography (US) evaluation of the hepatic and portal vasculature and percutaneous liver biopsy. No cases of cirrhosis were found. Splenic vein diameter was assessed via US in all cases. Left adrenal vein diameter was measured intra-operatively. Magnetic resonance angiography (MRA) was used as the noninvasive method of evaluating the splenic and portal veins in 7/15 patients. Follow-up ranged from 6 months to 10 years (Mean: 4.2 ±2 years). US was performed routinely for shunt patency surveillance post-operatively in all children and MRA in 7/15. Neurocognitive function was evaluated with Bayley's neurodevelopment scale [15]. Statistical analysis was via student t test and Fisher exact analysis when necessary. All of our patients have clinically significant hypersplenism with significant splenomegaly; in many the spleen occupied the whole left side of the abdominal cavity. All had developed persistent thrombocytopenia with platelet counts b50,000/μL that require regular multiple transfusions.
Spleno-adrenal shunt for portosystemic decompression Therefore, total splenectomy with spleno-adrenal shunt was performed in all cases. All patients in the study received age appropriate pneumococcal vaccine, H. influenza and quadrivalent meningococcal vaccine at least two weeks prior to surgery and routine penicillin prophylaxis post-operatively.
1.1. Surgical technique In children, we and other colleagues have often noted the presence of an enlarged left adrenal vein in patients with portal hypertension [13]. In our technique for spleno-adrenal shunt, the patient is placed in a supine position on the operating table with the left flank elevated 15° on a folded sheet. A long, left subcostal incision is carried well into the mid-axillary line and is extended across the midline. The transverse colon is retracted superiorly, and the small bowel is eviscerated to the patient's right. The ligament of Treitz is divided, and the duodenum is reflected superiorly. The retroperitoneum is incised anterior to the abdominal aorta and medial to the inferior mesenteric vein. Dissection is then continued cephalad along the aorta, until the left renal and adrenal veins are identified and isolated, importantly dissecting free the left adrenal vein but not ligating it. The caliber of the adrenal vein is measured and visualized to determine its suitability for spleno-adrenal shunt. The splenic vein is identified by knowing its relationship to inferior mesenteric vein, and dissected free from the undersurface of pancreas through the retroperitoneal incision. The rows of small, fragile pancreatic branches of the splenic vein are carefully isolated, ligated, and divided; such care is necessary to avoid bleeding. When the splenic, renal and adrenal veins are completely mobilized, vascular clamps are placed on the splenic and renal veins. The enlarged left
2191 adrenal vein is ligated as far distal as possible from the renal vein to achieve the greatest length. The splenic vein is transected near its confluence with the superior mesenteric vein to provide sufficient length to allow tension-free end-toend anastomosis between its proximal end and the adrenal vein and also to best avoid abnormal angulation or kinking. Due to the significant hypersplenism in all cases with severe splenomegaly, total splenectomy with ligation of short gastric vessels was performed in all cases. The final endto-end splenic vein to adrenal vein anastomosis is completed with a running non-absorbable suture without tension. (Fig. 1) Pre- and post-portal pressures were estimated by measuring greater omental venous pressure (GVP) via catheterization of the greater omental vein.
2. Results Intra-operative adrenal vein diameter and length ranged from 0.7 to 1.8 cm and 2.0 to 3.0 cm, respectively. Intraoperative GVP pre-SA shunt (30 ±11 mm Hg) decreased (22±7 mm Hg) post-SA shunt (p b0.05). On follow-up (range: 6 months to 10 years; average 4.2± 2 years), there have been no recurrences of clinical or subclinical GI bleeding and no patients have required further transfusions. Patients were followed weekly for the initial three months after surgery with lab draws, and then at 6 month intervals with Doppler ultrasound. Leukocyte and platelet counts returned to normal levels in all patients. Liver function tests remained normal in all children with the exception of elevated mean blood ammonia level Pre- (18±7 mmol/L), Post(60±17 mmol/L), (pb0.05) in the immediate post-operative period which was not associated with encephalopathic changes. Elevated ammonia levels persisted for a mean period of 4.2 weeks (range 2–6 weeks) but eventually normalized in all patients. All patients were initially placed on lactulose therapy until resolution of ammonia levels. All patients have continued to have normal ammonia levels on follow-up and are off lactulose. None of our patients developed overwhelming postsplenectomy sepsis at the time of this study. All patients underwent barium swallow and/or upper endoscopy at six months follow-up showing interval resolution of esophageal varices in all children. No one developed encephalopathy in our studies and all 15 patients have normal age-appropriate neuro-cognitive function by Bayley's assessment. US at 6 month interval demonstrated a patent shunt anastomosis without stricture in 14/15 (93%) of patients. Liver function remains normal in all patients on follow-up. All patients currently are alive.
3. Discussion Fig. 1
Intra-operative photo of end-to-end spleno-adrenal shunt.
In children, surgical shunt procedures for portal decompression are generally reserved for those with well-
2192 compensated chronic liver disease or extra-hepatic portal vein obstruction with preserved liver function who continue to have variceal hemorrhage despite endoscopic treatment with either banding or sclerotherapy. The main principles in shunt surgery in these children are notably 1) to provide selective portal decompression where pedal flow to the liver is preserved to prevent encephalopathy and ascites and 2) create a tension-free and wide anastomosis with a sufficient caliber into the appropriate venous conduit to prevent obstruction of the shunt and maintain long term patency. To this end and as first described by Warren et al. in 1967, the distal spleno-renal shunt has become the preferred and most commonly performed selective portal decompression procedure in children with excellent short and long-term outcomes [11]. Concomitant splenectomy is often necessary as these patients also commonly suffer from significant hypersplenism. Often however, one may encounter anatomical obstacles including excessive fibrotic tissue in the pancreatic bed housing the splenic vein, inadequate diameter of the splenic vein or an odd anatomic angulation of the splenic vein making the splenic vein to left renal vein anastomosis difficult, unsatisfying and more likely to fail. Sometimes the splenic vein may be so large creating a less than ideal angle of anastomosis with the renal vein. These factors are thought to increase the short- and long-term risk of shunt thrombosis. While shunt procedures, such as selective portocaval shunt with autologous or prosthetic graft and the Rex shunt (mesenteric to left portal vein shunt) or devascularization surgeries such as the Sugiura procedure have been described as viable alternatives, most if not all of the above procedures are relatively more complex and appear to have a higher morbidity with at best equal short term, but possibly worse long term outcomes [16-18]. In 1989, Mazariego and Reyes in Pittsburgh described a modification of the distal spleno-renal shunt technique using the adrenal vein as the inflow vessel into the renal vein to accomplish selective portal decompression in children [12]. In their study which thus far is the first published series in children (12 patients), a technical modification of the distal splenorenal shunt using an end-to-end spleno-adrenal anastomosis allowed for effective selective decompression of portal hypertension in children and was accomplished without peri-operative mortality and minimal morbidity with long term shunt patency. Our current study shows similar successful outcomes in preventing recurrent variceal bleeding and progression to hepatic encephalopathy while preserving liver function long term. We also demonstrated that intra-operative portal pressures were significantly lowered and long-term shunt patency rates were excellent (93% compared to their study of 83%). There were however, interesting differences between the two studies. All of our patients as opposed to none in the Pittsburgh study also had significant hypersplenism requiring splenectomy in addition to SA shunt decompression. 100% of our patients had portal vein thrombosis with cavernous transformation, whereas only one such patient
S. Gu et al. who had the diagnosis of PVCT in the Pittsburgh study, which was the only patient in that study who had portal vein thrombosis of the native liver. Interestingly, three of the four patients in their study with portal vein thrombosis did not have PVCT which may be due to the fact that in these three patients, the portal vein thrombosis developed after the liver transplantation. While 25% in the Pittsburgh study had elevated ammonia levels postoperatively, 100% of our patients had elevated ammonia levels during the initial post-operative period. In their study, the only patient with PVCT also developed post-operative ammonia level elevation suggesting that the elevated ammonia levels may be unique to PVCT diagnosis, and not a technique-related issue. Since elevated ammonia levels are more commonly seen in patients with poorly compensated cirrhotic liver disease, perhaps our patients had subclinical liver dysfunction not yet manifested in liver function test abnormalities or liver biopsy, which would be an alternative explanation for the elevated ammonia levels seen in our patients. Importantly, in our study group the elevated post-op ammonia levels were not associated with near or long term mental dysfunction/ encephalopathy and resolved over time with short term medical management. We believe this report illustrates that children with noncirrhotic portal hypertension secondary to PVCT can expect to have excellent long term liver function and hemorrhage free survival after SA shunt comparable to those who have undergone splenorenal shunt. Although the SA shunt is the primary option for our group, we believe that at minimum especially when the splenorenal shunt becomes technically difficult, the SA shunt procedure should be considered as an alternate method with reasonable expectation of comparable success in preventing recurrent variceal hemorrhage, preventing progression to hepatic encephalopathy and preservation of long term liver function in children with progressive portal hypertension. Although SA shunt appears to have the same advantage in preventing hepatic encephalopathy preserving neurodevelopment in our report with only small number of patients, longer follow-up is needed to look for late signs of encephalopathy assessing neurodevelopment long term. Furthermore, a prospective randomized trial comparing the splenorenal shunt and spleno-adrenal shunt should be conducted in the future.
References [1] Sorrentino D, et al. Cavernous transformation of the portal vein associated to multiorgan developmental abnormalities. Liver Int 2004;24(1):80-3. [2] Li Y, et al. Clinical analysis of childhood cavernous transformation of the portal vein. Zhongguo Dang Dai Er Ke Za Zhi 2006;8(1):75-6. [3] Warren WD, et al. Noncirrhotic portal vein thrombosis. Physiology before and after shunts. Ann Surg 1980;192(3):341-9. [4] Orloff MJ, Orloff MS, Rambotti M. Treatment of bleeding esophagogastric varices due to extrahepatic portal hypertension:
Spleno-adrenal shunt for portosystemic decompression
[5]
[6]
[7] [8] [9] [10] [11]
results of portal-systemic shunts during 35 years. J Pediatr Surg 1994;29(2):142-51 [discussion 151-4]. Rikkers LF, et al. Shunt surgery versus endoscopic sclerotherapy for variceal hemorrhage: late results of a randomized trial. Am J Surg 1993;165(1):27-32 [discussion 32-3]. Rosch J, Keller FS. Transjugular intrahepatic portosystemic shunt: present status, comparison with endoscopic therapy and shunt surgery, and future prospectives. World J Surg 2001;25(3):337-45 [discussion 345-6]. Shun A, et al. Portosystemic shunting for paediatric portal hypertension. J Pediatr Surg 1997;32(3):489-93. Reyes J, et al. The role of portosystemic shunting in children in the transplant era. J Pediatr Surg 1999;34(1):117-22 [discussion 122–3]. Botha JF, et al. Portosystemic shunts in children: a 15-year experience. J Am Coll Surg 2004;199(2):179-85. Marmon LM, et al. Successful splenorenal shunt for emergent control of variceal bleeding in a small child. J Pediatr Surg 1992;27(8):1049-50. Warren WD, Zeppa R, Forman JS. Selective trasplenic decompression of gastroesophageal varcies by distal splenorenal shunt. Ann Surg 1967;166:437-45.
2193 [12] Mazariegos GV, Reyes J. A technique for distal splenoadrenal shunting in pediatric portal hypertension. J Am Coll Surg 1998; 187(6):634-6. [13] Kulkarni VM, Nagral SS, Mathur SK. Use of adrenal vein conduit for splenorenal shunts: a case report. Hepatogastroenterology 1999; 46(27):2033-4. [14] Pujahari AK. Lieno-adrenal shunt. Trop Gastroenterol 2006;27(3): 136-7. [15] Cook MJ, et al. An examination of the stability of the Bayley scales of infants development with high-risk infants. J Early Interv 1989;13(1): 45-9. [16] Bismuth H, Franco D, Alagille D. Portal diversion for portal hypertension in children. The first ninety patients. Ann Surg 1980; 192(1):18-24. [17] Sigalet DL, Mayer S, Blanchard H. Portal venous decompression with H-type mesocaval shunt using autologous vein graft: a North American experience. J Pediatr Surg 2001;36(1):91-6. [18] Feng LS, et al. Triplex operation for portal hypertension with esophageal variceal bleeding: report of 140 cases. Hepatobiliary Pancreat Dis Int 2004;3(4):534-7.
www.elsevier.com/locate/jpedsurg
Spleno-adrenal shunt: A novel alternative for portosystemic decompression in children with portal vein cavernous transformation Song Gu a,⁎, Shirong Chang b, Jun Chu a, Min Xu a, Zhilong Yan a, Donald C. Liu b,1,2, Qimin Chen a a
Shanghai Children's Medical Center, Shanghai Jiao Tong University, School of Medicine, Shanghai, China University of Chicago, Comer Children's Hospital, Chicago, IL 60637, USA
b
Received 24 August 2012; accepted 1 September 2012
Key words: Spleno-adrenal shunt; Portal hypertension; Children; Portal vein cavernous transformation
Abstract Purpose: Children with portal vein cavernous transformation (PVCT) can develop life-threatening variceal hemorrhage from progressive portal hypertension. While spleno-renal shunt±splenectomy is the most common portosystemic decompression surgery performed in children, we have adopted a modified spleno-adrenal (SA) shunt for complicated PVCT. We describe our 10 year experience focusing on technique evolution and treatment efficacy. Methods: Between 2001 and 2011, 15 children (9 girls and 6 boys, ages 3–11years, median: 6years) with PVCT, portal hypertension, and hypersplenism were treated with SA shunt with splenectomy in Shanghai Children's Medical Center. All children in the study had endoscopy proven active esophageal variceal bleeding requiring multiple transfusions (mean: 4.2 units) with failed sclerotherapy (mean: 2.6 times). Greater omental vein pressure (GVP) approximating portal venous pressure was measured pre- and post-SA shunt. Pre- and postoperative ammonia levels were obtained. Follow-up ranged from 6months to 10years (mean: 4.2±2years). Results: Intra-operative adrenal vein diameter and length ranged from 0.7 to 1.8 cm and 2 to 3 cm, respectively. Intra-operative GVPs pre-and post-SA shunt were (30 ±11) and (22 ±7) mmHg, respectively (pb0.01). On follow-up, there have been no recurrences of GI bleeding. Liver function tests remained normal in all children with the exception of elevated post-operative mean blood ammonia levels [Pre (18±7) mmol/L, post (60±17) mmol/L (pb0.05)] in all children. Ammonia levels normalized in all cases on outpatient follow-up. There have been no cases of hepatic encephalopathy, and all have normal age appropriate neurodevelopment (Bayley's assessment). Barium swallow and/or upper endoscopy showed interval resolution of esophageal varices in all children, and vascular ultrasound showed patent shunt anastomosis without stricture in 14 (93%). Conclusions: The left adrenal vein is a viable conduit for effective selective portosystemic decompression. Similar to the more traditional spleno-renal shunt, SA appears also to have the advantage of preventing hepatic encephalopathy preserving neurodevelopment, although the rise in post-
⁎ Corresponding author. Tel.: +1 773 702 6175; fax: +1 773 702 1192. E-mail addresses: [email protected], [email protected] (S. Gu). 1 Co-senior author. 2 Deceased. 0022-3468/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2012.09.007
2190
S. Gu et al. operative ammonia levels was unexpected. Longer follow-up is needed to look for late signs of encephalopathy assessing neurodevelopment long term. © 2012 Elsevier Inc. All rights reserved.
Extrahepatic portal vein obstruction is an important cause of portal hypertension among children [1-3]. It is characterized mainly by portal vein cavernous transformation (PVCT) which is defined as the formation of venous channels within or around a previously thrombosed portal vein and is detected in 40% of children with upper gastrointestinal bleeding caused by esophageal varices. Approximately 80% of children diagnosed with portal vein thrombosis will have at least one episode of variceal bleeding which is often life threatening. In spite of the advancement in technology and improvement of therapeutic modalities in pharmacology, endoscopy (sclerotherapy or band ligation), and interventional radiology (transjugular intrahepatic portosystemic shunt); surgical decompressive shunts still play an important role in the treatment of variceal bleeding caused by portal hypertension [4,5]. The success rate of bleeding control with endoscopy and pharmacological agents is still poor at only fifty percent [6]. Although children with PVCT characteristically have normal liver function without cirrhosis, they still are at significant risk for severe hypersplenism and life threatening variceal hemorrhage secondary to progressive portal hypertension. Despite the advent of nonsurgical methods to control portal hypertensive bleeding and further refinements of liver transplantation in children, a significant role for portosystemic shunting for pediatric portal hypertension continues to evolve. Distal splenorenal shunt (DSRS) and/or selective portocaval shunt are still considered the preferable surgical procedures in children with preserved liver function and absent or minimal cirrhosis, but have significant medically refractory variceal hemorrhage secondary to progressive portal hypertension [7-10]. Often concomitant significant splenomegaly secondary to hypersplenism develops requiring splenorenal shunt with splenectomy. The technique of DSRS was first illustrated by Warren in 1967 describing an end-to-side anastomosis between the splenic vein and left renal vein [11]. Technical difficulties of DSRS mainly entail difficulties in establishing a favorable anastomosis due to several factors which mainly include splenic vein diameter, discrepancy in diameter caliber between the two veins and often an "awkward" angle between the two veins for a suitable anastomosis. A few reports in the literature have described the use of the adrenal vein as a viable, potentially more favorable conduit to achieve effective selective portal decompression compared to DSRS [12-14]. In this report, we describe the use of the spleno-adrenal (SA shunt) as our preferred technique for selective portal decompression in a series of 15 consecutive children with portal vein cavernous transformation (PVCT) suffering life threatening variceal hemorrhage from progressive portal hypertension.
1. Methods Under IRB approval, we examined our computerized database to identify fifteen consecutive children (9 girls and 6 boys, ages 3–11 years, median: 6 years) with portal vein cavernous transformation (PVCT) who underwent SA shunt in a ten year period between 2001 and 2011. Subjects were followed for periods ranging between six months and ten years (mean: 4.2±2 years) and all operations were carried out as elective procedures. Medical records were reviewed for demographic data; cause of portal hypertension (PVCT in all cases; 12-history of umbilical catheterization; 3-idiopathic); thirty day operative mortality; peri- and post-operative morbidities including rebleeding episodes, ascites, and clinical grade of portosystemic encephalopathy; short and long term liver function and shunt patency (Doppler ultrasonography); and overall survival. Indications for SA shunt with splenectomy operation in all children were major gastrointestinal bleeding and clinically significant hypersplenism. All children in the study had endoscopically proven active esophageal ±gastric variceal bleeding requiring multiple units of transfusions (mean 4.2; range 2–9) with failed number of attempts of sclerotherapy (mean 2.6). Medical management involved a stepwise approach of volume resuscitation with crystalloids and blood products ±octreotide or vasopressin infusion in the acute setting. All patients underwent early endoscopy after gastroesophageal lavage to determine the extent of gastroesophageal varices and the need for sclerotherapy. Further evaluation of the cause of the portal hypertension included abdominal Doppler ultrasonography (US) evaluation of the hepatic and portal vasculature and percutaneous liver biopsy. No cases of cirrhosis were found. Splenic vein diameter was assessed via US in all cases. Left adrenal vein diameter was measured intra-operatively. Magnetic resonance angiography (MRA) was used as the noninvasive method of evaluating the splenic and portal veins in 7/15 patients. Follow-up ranged from 6 months to 10 years (Mean: 4.2 ±2 years). US was performed routinely for shunt patency surveillance post-operatively in all children and MRA in 7/15. Neurocognitive function was evaluated with Bayley's neurodevelopment scale [15]. Statistical analysis was via student t test and Fisher exact analysis when necessary. All of our patients have clinically significant hypersplenism with significant splenomegaly; in many the spleen occupied the whole left side of the abdominal cavity. All had developed persistent thrombocytopenia with platelet counts b50,000/μL that require regular multiple transfusions.
Spleno-adrenal shunt for portosystemic decompression Therefore, total splenectomy with spleno-adrenal shunt was performed in all cases. All patients in the study received age appropriate pneumococcal vaccine, H. influenza and quadrivalent meningococcal vaccine at least two weeks prior to surgery and routine penicillin prophylaxis post-operatively.
1.1. Surgical technique In children, we and other colleagues have often noted the presence of an enlarged left adrenal vein in patients with portal hypertension [13]. In our technique for spleno-adrenal shunt, the patient is placed in a supine position on the operating table with the left flank elevated 15° on a folded sheet. A long, left subcostal incision is carried well into the mid-axillary line and is extended across the midline. The transverse colon is retracted superiorly, and the small bowel is eviscerated to the patient's right. The ligament of Treitz is divided, and the duodenum is reflected superiorly. The retroperitoneum is incised anterior to the abdominal aorta and medial to the inferior mesenteric vein. Dissection is then continued cephalad along the aorta, until the left renal and adrenal veins are identified and isolated, importantly dissecting free the left adrenal vein but not ligating it. The caliber of the adrenal vein is measured and visualized to determine its suitability for spleno-adrenal shunt. The splenic vein is identified by knowing its relationship to inferior mesenteric vein, and dissected free from the undersurface of pancreas through the retroperitoneal incision. The rows of small, fragile pancreatic branches of the splenic vein are carefully isolated, ligated, and divided; such care is necessary to avoid bleeding. When the splenic, renal and adrenal veins are completely mobilized, vascular clamps are placed on the splenic and renal veins. The enlarged left
2191 adrenal vein is ligated as far distal as possible from the renal vein to achieve the greatest length. The splenic vein is transected near its confluence with the superior mesenteric vein to provide sufficient length to allow tension-free end-toend anastomosis between its proximal end and the adrenal vein and also to best avoid abnormal angulation or kinking. Due to the significant hypersplenism in all cases with severe splenomegaly, total splenectomy with ligation of short gastric vessels was performed in all cases. The final endto-end splenic vein to adrenal vein anastomosis is completed with a running non-absorbable suture without tension. (Fig. 1) Pre- and post-portal pressures were estimated by measuring greater omental venous pressure (GVP) via catheterization of the greater omental vein.
2. Results Intra-operative adrenal vein diameter and length ranged from 0.7 to 1.8 cm and 2.0 to 3.0 cm, respectively. Intraoperative GVP pre-SA shunt (30 ±11 mm Hg) decreased (22±7 mm Hg) post-SA shunt (p b0.05). On follow-up (range: 6 months to 10 years; average 4.2± 2 years), there have been no recurrences of clinical or subclinical GI bleeding and no patients have required further transfusions. Patients were followed weekly for the initial three months after surgery with lab draws, and then at 6 month intervals with Doppler ultrasound. Leukocyte and platelet counts returned to normal levels in all patients. Liver function tests remained normal in all children with the exception of elevated mean blood ammonia level Pre- (18±7 mmol/L), Post(60±17 mmol/L), (pb0.05) in the immediate post-operative period which was not associated with encephalopathic changes. Elevated ammonia levels persisted for a mean period of 4.2 weeks (range 2–6 weeks) but eventually normalized in all patients. All patients were initially placed on lactulose therapy until resolution of ammonia levels. All patients have continued to have normal ammonia levels on follow-up and are off lactulose. None of our patients developed overwhelming postsplenectomy sepsis at the time of this study. All patients underwent barium swallow and/or upper endoscopy at six months follow-up showing interval resolution of esophageal varices in all children. No one developed encephalopathy in our studies and all 15 patients have normal age-appropriate neuro-cognitive function by Bayley's assessment. US at 6 month interval demonstrated a patent shunt anastomosis without stricture in 14/15 (93%) of patients. Liver function remains normal in all patients on follow-up. All patients currently are alive.
3. Discussion Fig. 1
Intra-operative photo of end-to-end spleno-adrenal shunt.
In children, surgical shunt procedures for portal decompression are generally reserved for those with well-
2192 compensated chronic liver disease or extra-hepatic portal vein obstruction with preserved liver function who continue to have variceal hemorrhage despite endoscopic treatment with either banding or sclerotherapy. The main principles in shunt surgery in these children are notably 1) to provide selective portal decompression where pedal flow to the liver is preserved to prevent encephalopathy and ascites and 2) create a tension-free and wide anastomosis with a sufficient caliber into the appropriate venous conduit to prevent obstruction of the shunt and maintain long term patency. To this end and as first described by Warren et al. in 1967, the distal spleno-renal shunt has become the preferred and most commonly performed selective portal decompression procedure in children with excellent short and long-term outcomes [11]. Concomitant splenectomy is often necessary as these patients also commonly suffer from significant hypersplenism. Often however, one may encounter anatomical obstacles including excessive fibrotic tissue in the pancreatic bed housing the splenic vein, inadequate diameter of the splenic vein or an odd anatomic angulation of the splenic vein making the splenic vein to left renal vein anastomosis difficult, unsatisfying and more likely to fail. Sometimes the splenic vein may be so large creating a less than ideal angle of anastomosis with the renal vein. These factors are thought to increase the short- and long-term risk of shunt thrombosis. While shunt procedures, such as selective portocaval shunt with autologous or prosthetic graft and the Rex shunt (mesenteric to left portal vein shunt) or devascularization surgeries such as the Sugiura procedure have been described as viable alternatives, most if not all of the above procedures are relatively more complex and appear to have a higher morbidity with at best equal short term, but possibly worse long term outcomes [16-18]. In 1989, Mazariego and Reyes in Pittsburgh described a modification of the distal spleno-renal shunt technique using the adrenal vein as the inflow vessel into the renal vein to accomplish selective portal decompression in children [12]. In their study which thus far is the first published series in children (12 patients), a technical modification of the distal splenorenal shunt using an end-to-end spleno-adrenal anastomosis allowed for effective selective decompression of portal hypertension in children and was accomplished without peri-operative mortality and minimal morbidity with long term shunt patency. Our current study shows similar successful outcomes in preventing recurrent variceal bleeding and progression to hepatic encephalopathy while preserving liver function long term. We also demonstrated that intra-operative portal pressures were significantly lowered and long-term shunt patency rates were excellent (93% compared to their study of 83%). There were however, interesting differences between the two studies. All of our patients as opposed to none in the Pittsburgh study also had significant hypersplenism requiring splenectomy in addition to SA shunt decompression. 100% of our patients had portal vein thrombosis with cavernous transformation, whereas only one such patient
S. Gu et al. who had the diagnosis of PVCT in the Pittsburgh study, which was the only patient in that study who had portal vein thrombosis of the native liver. Interestingly, three of the four patients in their study with portal vein thrombosis did not have PVCT which may be due to the fact that in these three patients, the portal vein thrombosis developed after the liver transplantation. While 25% in the Pittsburgh study had elevated ammonia levels postoperatively, 100% of our patients had elevated ammonia levels during the initial post-operative period. In their study, the only patient with PVCT also developed post-operative ammonia level elevation suggesting that the elevated ammonia levels may be unique to PVCT diagnosis, and not a technique-related issue. Since elevated ammonia levels are more commonly seen in patients with poorly compensated cirrhotic liver disease, perhaps our patients had subclinical liver dysfunction not yet manifested in liver function test abnormalities or liver biopsy, which would be an alternative explanation for the elevated ammonia levels seen in our patients. Importantly, in our study group the elevated post-op ammonia levels were not associated with near or long term mental dysfunction/ encephalopathy and resolved over time with short term medical management. We believe this report illustrates that children with noncirrhotic portal hypertension secondary to PVCT can expect to have excellent long term liver function and hemorrhage free survival after SA shunt comparable to those who have undergone splenorenal shunt. Although the SA shunt is the primary option for our group, we believe that at minimum especially when the splenorenal shunt becomes technically difficult, the SA shunt procedure should be considered as an alternate method with reasonable expectation of comparable success in preventing recurrent variceal hemorrhage, preventing progression to hepatic encephalopathy and preservation of long term liver function in children with progressive portal hypertension. Although SA shunt appears to have the same advantage in preventing hepatic encephalopathy preserving neurodevelopment in our report with only small number of patients, longer follow-up is needed to look for late signs of encephalopathy assessing neurodevelopment long term. Furthermore, a prospective randomized trial comparing the splenorenal shunt and spleno-adrenal shunt should be conducted in the future.
References [1] Sorrentino D, et al. Cavernous transformation of the portal vein associated to multiorgan developmental abnormalities. Liver Int 2004;24(1):80-3. [2] Li Y, et al. Clinical analysis of childhood cavernous transformation of the portal vein. Zhongguo Dang Dai Er Ke Za Zhi 2006;8(1):75-6. [3] Warren WD, et al. Noncirrhotic portal vein thrombosis. Physiology before and after shunts. Ann Surg 1980;192(3):341-9. [4] Orloff MJ, Orloff MS, Rambotti M. Treatment of bleeding esophagogastric varices due to extrahepatic portal hypertension:
Spleno-adrenal shunt for portosystemic decompression
[5]
[6]
[7] [8] [9] [10] [11]
results of portal-systemic shunts during 35 years. J Pediatr Surg 1994;29(2):142-51 [discussion 151-4]. Rikkers LF, et al. Shunt surgery versus endoscopic sclerotherapy for variceal hemorrhage: late results of a randomized trial. Am J Surg 1993;165(1):27-32 [discussion 32-3]. Rosch J, Keller FS. Transjugular intrahepatic portosystemic shunt: present status, comparison with endoscopic therapy and shunt surgery, and future prospectives. World J Surg 2001;25(3):337-45 [discussion 345-6]. Shun A, et al. Portosystemic shunting for paediatric portal hypertension. J Pediatr Surg 1997;32(3):489-93. Reyes J, et al. The role of portosystemic shunting in children in the transplant era. J Pediatr Surg 1999;34(1):117-22 [discussion 122–3]. Botha JF, et al. Portosystemic shunts in children: a 15-year experience. J Am Coll Surg 2004;199(2):179-85. Marmon LM, et al. Successful splenorenal shunt for emergent control of variceal bleeding in a small child. J Pediatr Surg 1992;27(8):1049-50. Warren WD, Zeppa R, Forman JS. Selective trasplenic decompression of gastroesophageal varcies by distal splenorenal shunt. Ann Surg 1967;166:437-45.
2193 [12] Mazariegos GV, Reyes J. A technique for distal splenoadrenal shunting in pediatric portal hypertension. J Am Coll Surg 1998; 187(6):634-6. [13] Kulkarni VM, Nagral SS, Mathur SK. Use of adrenal vein conduit for splenorenal shunts: a case report. Hepatogastroenterology 1999; 46(27):2033-4. [14] Pujahari AK. Lieno-adrenal shunt. Trop Gastroenterol 2006;27(3): 136-7. [15] Cook MJ, et al. An examination of the stability of the Bayley scales of infants development with high-risk infants. J Early Interv 1989;13(1): 45-9. [16] Bismuth H, Franco D, Alagille D. Portal diversion for portal hypertension in children. The first ninety patients. Ann Surg 1980; 192(1):18-24. [17] Sigalet DL, Mayer S, Blanchard H. Portal venous decompression with H-type mesocaval shunt using autologous vein graft: a North American experience. J Pediatr Surg 2001;36(1):91-6. [18] Feng LS, et al. Triplex operation for portal hypertension with esophageal variceal bleeding: report of 140 cases. Hepatobiliary Pancreat Dis Int 2004;3(4):534-7.