| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Departments of Diagnostic Imaging (R.J.T.O., K.O., M.C.M.), Medicine (E.A.R., B.W.P.), and Surgery (D.L.B.) and the Surgical-Medical Research Institute and Clinical Islet Transplant Program (J.R.T.L., N.M.K., G.S.K., R.V.R., A.M.J.S.), University of Alberta Hospital, 8440 - 112th St, Edmonton, Alberta, Canada T6G 2B7. Received December 2, 2002; revision requested January 30, 2003; revision received February 21; accepted April 11. Supported by a clinical centre grant from the Juvenile Diabetes Research Foundation, the Alberta Foundation for Diabetes Research, a donation from the Roberts family, a grant from the Canadian Institute for Health Research, the Alberta Health Services Innovation Fund, and institutional support by the University of Alberta Hospitals, Capital Health, and Province Wide Services of the Government of Alberta. Address correspondence to R.J.T.O. (e-mail: drrichardowen@shaw.ca).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Between March 1999 and May 2002, 34 patients underwent 68 islet cell transplantation procedures. Patients with C-peptide–negative type 1 diabetes were selected on the basis of poor metabolic control (hypoglycemia or lability) despite compliance with optimal medical therapy. Islet cells were isolated from brain-dead donors. Access to the portal vein was gained from a right percutaneous transhepatic approach, and islet cells were infused with intermittent pressure monitoring. Twenty patients underwent two transplantations, seven patients underwent three transplantations, and seven patients underwent one transplantation. Complications during and after the procedure and postprocedural diabetic status were monitored.
RESULTS: Successful portal vein cannulation and islet cell infusion were achieved in all cases. Fluoroscopy was used as the primary guidance modality in 58 of 68 (85%) procedures, and ultrasonography was used in 10 of 68 (15%). Total recorded fluoroscopy time varied from 0.6 to 103 minutes, with a median of 6.9 minutes. Potentially serious complications occurred in six of 68 (9%) procedures. Two patients developed portal venous thrombosis, and with subsequent anticoagulation therapy, one of the two developed an expanding hepatic hematoma that required surgery. Clinically important hemorrhage occurred in four patients, three of whom required blood transfusions. Of 26 patients who received completed transplants, all became insulin independent, and 81% (21 of 26) remained insulin free at 1 year.
CONCLUSION: The percutaneous transhepatic approach for the implantation of islet cells into the portal vein is a safe procedure, and together with use of current cell separation techniques and an immunosuppressive regimen, offers a marked advance in the treatment of type 1 diabetes mellitus.
© RSNA, 2003
Index terms: Diabetes mellitus • Interventional procedures, complications, 957.1264, 947.442 • Pancreas, transplantation, 770.1267 • Portal vein, therapeutic embolization, 957.1264
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Successful engraftment of isolated pancreatic islet cells into the livers of rats with diabetes via the portal veins was performed as long ago as 1973 (8,9). The techniques used to isolate and purify islet cells in rodents proved unsuccessful in the fibrous human pancreas, however, and the use of impure fractions carried substantial risk (10). With improved cell separation techniques that were developed in the early 1990s, islet cell transplantation resumed with initial high hopes of success (11). Use of the portal vein for islet cell engraftment was described in a series of patients undergoing total pancreatectomy in which surgical exposure was used (12). The first reports of radiologically gained access to the portal vein for the purpose of islet cell transplantation were issued by Weimar and colleagues, who used a combination of computed tomography (CT) and fluoroscopy (13).
Recent advances in glucocorticoid-free immunosuppression and cell separation techniques led to remarkable enhancement in success, with insulin independence rates increasing from 12% to 100% in a preliminary series of seven patients (14). One of the keys to the minimal morbidity of islet cell transplantation in our study has been the routine use of a percutaneous transhepatic approach for intraportal infusion of the isolated islet cells. There are very few published data that describe the radiologic technique and technical aspects of the procedure. The purpose of our study was to report our experience with percutaneous transhepatic pancreatic islet cell transplantation in patients with type 1 diabetes mellitus.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
All patients had type 1 diabetes with a mean duration of 25.5 years ± 1.9 (range, 5–50 years). Median duration of follow-up was 21.5 months (25th through 75th percentiles, 10–32 months; maximum follow-up, 39 months). Median hospital inpatient stay was 1 day (25th through 75th percentiles, 1–2 days; range, 0.6–16 days).
In the first publication from our group concerning the islet cell transplantation program, we described the findings in the initial seven patients treated with the Edmonton protocol (14). The clinical complications and metabolic monitoring were described in two further publications in 2001 (12 patients) (15) and 2002 (30 patients) (16). Portal venous pressure changes that occur after sequential islet cell transplantation were reported in 2002 (26 patients) (17). In the current article, we report on an expanded number of patients that were also included in these previous publications; although the clinical protocols have evolved and changed over time, the radiologic access procedure has remained relatively constant across protocols.
Immunosuppressive Regimen
All patients received steroid-free immunosuppression therapy, which commenced immediately prior to islet cell embolization. The regimen, as described previously (14,16), consisted of daclizumab (Zenapax; Roche, Mississauga, Ontario, Canada), sirolimus (Rapamune; Wyeth-Ayerst, Markham, Ontario, Canada), and tacrolimus (Prograf; Fujisawa, Markham, Ontario, Canada). In addition, patients were also given standard prophylactic antibiotics for the procedure.
Islet Cell Preparation
Pancreas organs were obtained from brain-dead donors after informed consent was obtained from relatives. The islet cells were isolated by using a combination of enzymatic and mechanical dissolution and were prepared in xenoprotein-free medium (18–20). Islet cell transplantation proceeded if more than 4,000 purified islet cell equivalents were prepared in a packed cell volume of less than 10 mL, if ABO blood group compatibility matched, and if the Gram stain was negative and the endotoxin content was less than 5 endotoxin units per kilogram. The quantification of islet cells is expressed in terms of islet cell equivalents to account for variation in islet cell volume, with a standard islet cell measuring 150 µm (21).
Transplantation Procedure
Thirty-four consecutive patients underwent 68 transhepatic portal islet cell embolization procedures. Each was performed by one of nine interventional radiologists.
All patients were sedated with intravenous midazolam and fentanyl (Sabex; Boucherville, Quebec, Canada). Oxygen was administered via the nasal cannula at 6 L/min. A right-sided percutaneous approach was used with patients positioned supine. The point of hepatic puncture (anterior or midaxillary line) was determined by using fluoroscopy, ultrasonography (US), or a combination of the two. Total fluoroscopic time was recorded for each procedure. The number of hepatic punctures required to gain access to the portal vein was not recorded consistently.
Aseptic technique was used, and the subcutaneous tissues and hepatic capsule were infiltrated with local anaesthetic. A 22-gauge (0.022-inch) Chiba needle was advanced into a branch of the right portal vein with fluoroscopic or US guidance. A second- or third-order branch was selected in most cases. An 18-gauge (0.018-inch) guidewire was then advanced into the main portal vein. By using standard technique, the sheath from the Neff percutaneous access set or stiffened micropuncture set (Cook Canada, Stouffville, Ontario, Canada) was advanced over the guidewire into the portal vein and positioned just proximal to the portal confluence. In some cases, this was exchanged for a 5-F Kumpe (Cook Canada) or H1 catheter for islet cell infusion. A portal venogram was used to confirm optimal catheter placement (Figure). Baseline portal venous pressure (in millimeters of mercury) was recorded at this stage by using an indirect pressure transducer (Medex, Hillard, Ohio).
|
Islet cell preparation was suspended in 120 mL of supplemented media (M199; Mediatech, Herndon, Va) that contained heparin and 20% human albumin. Heparin (35 U/kg) was added when the packed islet cell volume was less than 5 mL, and the amount of heparin was increased to 70 U/kg if the packed cell volume exceeded 5 mL. The islet cells were initially administered over approximately 10 minutes by using a 60-mL syringe. Portal venous pressure was recorded after the first 50-mL aliquot and after subsequent 50-mL aliquots. More recently (August 2001), a gravity-based closed infusion bag system has been used for the purpose of minimizing the shear forces on the islet cells to provide an alternative indirect method of continuous portal venous pressure monitoring and to reduce the risk of preparation contamination during islet cell delivery. Portal venous pressure was recorded after administration of the first 5 mL of packed cell volume, after administration of each subsequent milliliter of packed cell volume, and at the end of the procedure. The procedure was terminated if portal venous pressure was higher than 20 mm Hg at the outset or if it increased to twice the baseline value or to higher than 22 mm Hg during the procedure.
Of the 34 patients who underwent transplantation, seven underwent one islet cell infusion, 20 underwent two infusions, and seven underwent three infusions. An average of 376,954 islet cell equivalents ± 17,891 were injected in the first procedure, with each patient who completed the protocol receiving 857,468 ± 37,496 during the course of two to three infusions. The criterion for success was gaining satisfactory portal venous access and infusing all of the prepared islet cells.
Following the infusion procedure, the catheter was removed with embolization of the tract by using gelatin sponge (Gelfoam; Pharmacia & Upjohn, Mississauga, Ontario, Canada) pledgets in most of the first 30 procedures. With the development of a stiffened 4-F micropuncture catheter (Cook Canada), tract embolization is no longer performed routinely after islet cell infusion.
Complications were documented from data recorded at the time of the procedure, at clinical follow-up, and at confirmatory review of the radiology chart and images (three authors in consensus: R.J.T.O., E.A.R., A.M.J.S.) and were considered serious if they posed a threat to the patient’s life.
Postprocedural Care
Postprocedural patient care involved a standard protocol for posthepatic interventions, including bed rest in a right recumbent position and close observation for 4 hours. US of the liver with Doppler examination of the portal vein was performed within 24 hours of transplantation. Postprocedural diabetic status was monitored by discontinuing insulin after transplantation. It was not resumed unless the plasma glucose level increased to more than 10.0 mmol/L postprandially or more than 8.0 mmol/L in the fasting state.
Subsequent Transplantations
Limitations of islet cell yield from donors mean that only in a minority of patients were enough islet cells isolated to allow insulin independence after a single procedure. On average, only 5,000 islet cell equivalents per kilogram were isolated and transplanted per procedure, with the average number required to reach insulin independence approximately 13,000 islet cell equivalents per kilogram (16). Supplemental transplantations were therefore required in most cases to provide an adequate islet cell engraftment mass, as described previously (14). In addition, two patients underwent supplemental islet cell infusions; they discontinued insulin after a second transplantation but had recently required insulin therapy. Twenty patients underwent a second and seven patients a third islet cell transplantation procedure. Seven patients in this series had received only one transplant, with most waiting for a subsequent infusion on the basis of donor availability at the time of analysis. Twenty-six patients had completed the protocol at the time of the writing of this article.
Statistical Analysis
Data were presented as the median, with the 25th through 75th percentiles and the mean ± standard error calculated with the Mann-Whitney test or the Student t test for comparison where appropriate (Sigmastat version 3.0 for Windows; SPSS, Chicago, Ill).
Outcome Measures
Fluoroscopic time to gain portal venous access, portal venous pressure levels pre- and postinfusion, complications that occurred during or after the procedure, and postprocedural diabetic status were evaluated.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In 58 procedures, fluoroscopy was used as the primary imaging and guidance modality (in seven of these procedures, US was used for localization), and in 10 procedures, US was used for localization and guidance. Total fluoroscopic time varied from 0.6 to 103 minutes, with times recorded in 66 of 68 (97%) procedures (mean, 12.4 minutes ± 2.1; median, 6.9 minutes). The two longest recorded times (103 and 88.2 minutes) were recorded in the first 10 procedures, with a mean of 29.5 minutes ± 10.6. The subsequent 56 procedures required a mean fluoroscopic time of 9.3 minutes ± 1.3. In those procedures (10 of 68) in which US was used as the primary guidance modality, a mean fluoroscopic time of 5.5 minutes ± 0.9 was recorded.
Portal Venous Pressure
Portal venous pressure levels were recorded after catheter placement in the portal vein prior to islet cell infusion and during and after transplantation. The median pressure was 11.0 mm Hg (25th through 75th percentiles, 7.8–13.3) prior to the procedure and 13.0 mm Hg (25th through 75th percentiles, 11.1–17.0) afterward, indicating a significant difference (P < .001).
Procedure-related Complications
Procedure-related complications occurred in 13 of 68 (19%) procedures. Potentially serious complications occurred in six of 68 (9%) procedures. Early in the series, four patients experienced substantial perihepatic bleeding that required transfusion in three; the other patient was treated conservatively. This prompted initiation of posttransplantation tract embolization with gelatin sponge pledgets. With the development of the 4-F stiffened micropuncture kit, routine catheter tract embolization ceased.
Two patients developed segmental portal vein occlusion. The first patient underwent anticoagulation therapy with heparin and experienced no further complications. The other patient developed right-branch portal venous thrombosis after the third islet cell transplantation. In this patient, portal venous pressure increased to 28 mm Hg after islet cell infusion, and the patient experienced mild abdominal pain and nausea. Although symptoms settled, follow-up US demonstrated right-branch portal venous thrombosis, and after subsequent anticoagulation therapy with low-molecular-weight heparin, an expanding hepatic hematoma developed that required surgical decompression and wedge resection. This patient made a full recovery and continues to have detectable C-peptide. At this patient’s most recent follow-up examination, the right portal vein had recanalized and remained patent.
Four procedures were complicated by puncture of the biliary system (one puncture of the intrahepatic bile duct, three punctures of the gallbladder), but all such complications alleviated with conservative treatment. In one of these cases, acute discomfort during the procedure, which was triggered by distention of the gallbladder, led to completion of the procedure with use of a general anesthetic. Biliary tract puncture did not occur in the 17 of 68 (25%) procedures in which US was used for localization.
Two patients had vasovagal episodes, with one requiring intravenous atropine. One patient developed severe retching, which caused the catheter to dislodge partially and require repeated repositioning. Subsequent US evaluation showed a small amount of intraabdominal fluid thought to represent blood related to capsular trauma.
Postprocedural Diabetic Status
Twenty-six patients received completed transplants, and all became insulin independent, with 21 of 26 (81%) remaining insulin free at 1 year. Three subjects who were insulin independent have become C-peptide negative, indicating loss of islet cell function. More detailed clinical and metabolic outcomes have been reported recently on this patient series (17).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The portal vein was selected as the site for islet cell implantation because of the ease of technical access combined with cumulative data that indicated that this route provided the safest and most durable route for transplantation (11). Results of previous studies have shown that percutaneous splenic embolization has a high complication rate and low success rate (22). Likewise, subcapsular renal placement, omental pouch implantation, subcutaneous delivery, and celiac artery infusion have been attempted with very limited success in large animal models and in clinical trials (23–26). The most promising method appears to be portal vein embolization, either surgically by means of omental vein embolization or percutaneously, as in our study.
Empirically, embolization into the portal vein is appealing, with a large vascular bed available for distribution, anatomy favorable for percutaneous access, and previous reports of portal vein embolization (albeit for a different reason) that suggest procedural safety and perhaps some minimal advantages over other techniques (27). Selective transhepatic catheterization of the portal vein was first used as a diagnostic tool and as a means of acquiring access to bleeding esophageal varices (28). Subsequently, the technique has been used to induce hypertrophy of the anticipated liver remnant prior to major liver resection. Prior to surgery, embolization is performed by using gelatin sponge particles, embolization coils, or polyvinyl alcohol particles to induce hypertrophy in the nonembolized segments. This technique has a complication rate of less than 5% (27), which supports our belief that this is a safe route for islet cell transplantation. We also know that the portal venous blood supply is adequate to maintain hepatic viability after hepatic artery embolization with particulate or chemoembolic agents (29). Finally, percutaneous liver access for transhepatic cholangiography, biliary drainage procedures, and liver biopsy is practiced widely, with a proven safety record.
Delivery of the islet cells into the portal vein was initially performed by using either a 5-F diagnostic catheter or the 7-F sheath in the percutaneous biliary access set. To reduce catheter size and thereby reduce hemorrhage risk, the 4-F inner dilator was used in some cases; however, the inner diameter of this catheter is only 0.018 inch (457 µm). Islet cell size is up to 500 µm, and therefore, this was felt to subject considerable shear stress to the larger islet cells. In association with Cook Canada, we developed a stiffened 4-F micropuncture kit with a central metallic stiffener, a 3-F inner dilator (that would accept a 0.018-inch guidewire), and an outer 4-F dilator that would accept a 0.038-inch (965-µm) guidewire. In a further attempt to reduce shear forces on the islet cells, a gravity-based system for infusion is now used.
Of the potential complications that arise from percutaneous islet cell transplantation, those relating to rejection, drug regimens, long-term viability, and potential long-term hepatic parenchymal changes are beyond the scope of this article. Complications that relate to the procedure itself are, however, worthy of discussion here. Acute hemorrhage that resulted from capsular puncture occured in four patients early in the study, which prompted the introduction of tract embolization with gelatin sponge particles. To avoid this event, the transjugular route could be used for portal venous cannulation. However, reports on transjugular intrahepatic portosystemic shunts, or TIPS, indicate that there are still substantial risks of biliary puncture, capsular puncture, and extrahepatic portal vein puncture (29–31). It is also worth noting that in our patient who developed an expanding intraparenchymal hematoma that required surgical decompression, this complication could still have occurred if a TIPS approach had been used. The number of individual hepatic punctures used during the procedure was not recorded in our study, and therefore, we cannot relate this to bleeding complications. Intuitively, one might expect a higher number of punctures to be associated with a higher risk of hemorrhage, and an approach involving both CT and fluoroscopic guidance is purported to allow catheterization of the portal vein by using a reduced number of punctures (13).
We believe that US guidance offers a more readily available technique to reduce both the time required for portal cannulation and the number of capsular punctures. The extent to which US was used in this study varied between operators, with some using US guidance for all but guidewire manipulation and others using surface anatomy and fluoroscopic guidance only. A reduced fluoroscopic time was observed in the 10 procedures in which US was used, however, compared with that in the 56 procedures in which fluoroscopy was used as the primary imaging modality. Even more important may be catheter size. After introduction of the 4-F delivery system, no further episodes of major hemorrhage occurred, despite abandonment of tract embolization as routine. An additional benefit of US may be reduction of the incidence of gallbladder puncture. In the 17 patients in whom US was used for localization of the hepatic puncture site, there were no cases of biliary puncture.
The most serious complication we encountered was a case of right-branch portal venous thrombosis following islet cell transplantation. This patient developed an expanding subcapsular hematoma following anticoagulation therapy and ultimately required partial hepatic resection. After this event, the heparin dose administered with the islet cell preparation was increased to 35 U/kg, and portal venous pressure levels were recorded after the first 5 mL of packed islet cell volume were administered and after subsequent aliquots of 1-mL packed cell volume. The procedure was terminated if portal venous pressure was higher than 20 mm Hg at the outset or increased to twice the baseline value or to higher than 22 mm Hg. The rationale for both of these prophylactic measures relates to the precipitating factors that lead to thrombosis, including vascular damage, presence of thrombogenic material, and reduced portal venous flow.
We also encountered an asymptomatic peripheral segmental portal vein occlusion identified at US. This may have been related to embolization with use of gelatin sponge slurry, some of which may have escaped into the portal vein along the catheter tract. This patient was treated conservatively with temporary anticoagulation therapy for 3 months.
Complete thrombosis of the main portal vein that leads to serious consequences has been described rarely and only in situations in which impure or partially purified islet cell preparations were used (32). Data from animal models have shown that occlusion of the vascular bed progressively increases the vascular resistance, which will eventually lead to elevation of portal venous pressure (33). Increased flow in the hepatic artery and reduced flow in the portal vein have been documented following embolization of the portal vein with particulate emboli; this phenomenon is explained by simple fluid mechanics (34). Islet cell clumps act as particulate emboli and may occlude arterioles up to 500 µm in diameter. Therefore, it may be that the portal venous pressure changes we demonstrated reflected an alteration in vascular resistance. Portal venous pressure elevation has been shown to be more dramatic with subsequent procedures (14), which suggests a diminution of hepatic reserve. A significant pressure elevation in the portal vein following islet cell embolization has also been demonstrated in relation to the number of islet cells transplanted, the packed cell volume, and the number of procedures performed per patient (14). We allow a maximum of 10 mL of packed cell volume and prefer to have less than 5 mL. We also noticed that there appeared to be an inverse relationship between cell purity and portal venous pressure increase. We believe this may be due to inevitable contaminants, predominantly connective tissue and exocrine pancreas remnants.
If islet cells need to be transplanted into the liver, then use of the percutaneous route is straightforward and safe in the hands of an experienced interventional radiologist. As for the portability of the procedure and reproducibility of results, we propose that the procedure is well within the capability of all dedicated interventionists and that with adequate cell separation and purification, the procedure and clinical results should be reproducible.
In conclusion, the percutaneous transhepatic approach for the implantation of islet cells into the portal vein is a safe procedure that allows most patients to remain free of insulin and maintain excellent glycemic control.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Onaca, B. Naziruddin, S. Matsumoto, H. Noguchi, G. B. Klintmalm, and M. F. Levy Pancreatic Islet Cell Transplantation: Update and New Developments Nutr Clin Pract, October 1, 2007; 22(5): 485 - 493. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.M. J. Shapiro, C. Ricordi, B. J. Hering, H. Auchincloss, R. Lindblad, R. P. Robertson, A. Secchi, M. D. Brendel, T. Berney, D. C. Brennan, et al. International Trial of the Edmonton Protocol for Islet Transplantation. N. Engl. J. Med., September 28, 2006; 355(13): 1318 - 1330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Venturini, E. Angeli, P. Maffi, P. Fiorina, F. Bertuzzi, M. Salvioni, F. De Cobelli, C. Socci, L. Aldrighetti, C. Losio, et al. Technique, Complications, and Therapeutic Efficacy of Percutaneous Transplantation of Human Pancreatic Islet Cells in Type 1 Diabetes: The Role of US Radiology, February 1, 2005; 234(2): 617 - 624. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Harlan Gene-Altered Islets for Transplant: Giant Leap or Small Step? Endocrinology, February 1, 2004; 145(2): 463 - 466. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
