ª Springer Science+Business Media New York 2014
Abdominal Imaging
Abdom Imaging (2014) DOI: 10.1007/s00261-014-0164-3
Multi detector computed tomography (MDCT) for the diagnosis of early complications after pancreas transplantation M. Vincent,1 O. Morla,1 J. Branchereau,2 G. Karam,2 B. Dupas,1 E. Frampas1 1
Central Department of Radiology and Medical Imaging, Hoˆtel-Dieu, CHU Nantes, 1 Place Alexis Ricordeau, 44093 Nantes Cedex 1, France 2 Department of Urology, Hoˆtel-Dieu, CHU Nantes, 1 Place Alexis Ricordeau, 44093 Nantes Cedex 1, France
Abstract Purpose: Solitary Pancreas (SPT) and simultaneous kidney-pancreas (SPKT) transplants carry a high risk of surgical complications that may lead to the loss of the pancreas graft and impact later kidney function. The purpose of this study was to investigate the role of MDCT in the diagnosis of early complications and its impact on kidney function. Methods: All patients receiving SPT or SPKT over 5 years were retrospectively included. Complications that occurred within the first 15 days were registered and MDCT data analyzed. Data regarding donor, transplant, and recipient characteristics as well as transplantation procedures were analyzed according to the occurrence of early complications. Kidney function at day 3 following MDCT was evaluated. Results: One hundred and forty-one patients were included (85 men, 56 women; mean age 40.1 years, SD 7.7) with 119 SPKT and 22 SPT. Sixty-four complications were registered in 50 patients. Partial (P-) or complete venous thrombosis (C-VT) occurred in 12.1 % (n = 17), arterial thrombosis (AT) in 1.4 % (n = 2), and hemorrhage in 8.5 % (n = 12) of all patients. For venous thrombosis, the predominant risk factor was body mass index (BMI) for either recipients (P < 0.05) or donors (P < 0.01). Median time for venous thrombosis diagnosis with MDCT was 4 days. Kidney function was not altered following MDCT. Fourteen pancreatectomies were necessary. All patients with C-VT and AT had to undergo graftectomy. Conclusion: Vascular complications occurred early following grafting. Systematic early-enhanced MDCT at day 2–3 should be adequate to detect early thrombosis,
Correspondence to: E. Frampas; email: [email protected]
especially if risk factors have been identified, without induced kidney function alteration. Key words: MDCT—Pancreas transplant—Surgical complications—Graft venous thrombosis—Graft failure
In patients with insulin-requiring diabetes, pancreas transplantation is considered as a successful treatment to normalize glucose metabolism, and to improve and prevent secondary diabetic complications. Since 1966, more than 37000 pancreas transplants have been reported to the International Pancreas Transplant Registry [1]. Most of them are simultaneous pancreatic and kidney transplants (SPKT), representing 75 % of cases, compared to those given after kidney transplants (PAKT), or to solitary pancreas transplants (SPT). Improvements in surgical techniques, postoperative management, and immunosuppressive protocols have led to improved patient and graft survival, with pancreas graft survival reaching 86 % at one year for SPK. Reported 1-year patient survival varies between 92 %–98 % and pancreatic transplantation offers considerable survival benefits compared to the mortality of patients on waiting list which approaches ~30% [2]. Nevertheless, pancreatic transplantation remains a complex surgical procedure with a high risk of surgical complications, and the need for relaparotomy in up to 43 % of cases which significantly affects pancreas graft survival [3]. The most frequent complications are fluid collections while venous thrombosis is responsible for the highest rate of pancreatectomy [4]. Diagnosis of early surgical complication is crucial as this can impact kidney graft survival. Lower kidney graft survival rates have been reported in patients with early pancreas loss as
M. Vincent et al.: Multi detector computed tomography
compared to recipients without pancreas failure (71.4 % and 59.5 % at one and three years vs. 86 % and 82 %) [5]. The purpose of this study was to determine the risk factors that predispose to early complications within the 15 first days following pancreas transplantation, and evaluate the use of MDCT in the early postoperative course.
Materials and methods Study design
against thrombosis with low-dose heparin during the hospital stay followed by oral antiplatelet therapy with aspirin and dipyridamole.
CT Imaging When clinically suspected, complications were investigated with multidetector CT (MDCT). Patients were scanned with 16 or 64-detector CT (Sensation 16, Siemens Healthcare, or Lightspeed VCT 64, General Elec-
Following institutional review, this retrospective study included all adult pancreas transplants over a five-year period from January 2004 to February 2009 in a single institution. Data regarding donor, transplant and recipient characteristics, graft and patient outcomes were obtained from the electronic medical database of the institution (Clinicom, Intersystems). Considering the retrospective nature of our study, our institutional review board judged it to be exempt from informed consent. All complications were registered and considered as early when they occurred during the 15 days following the graft.
Patient and clinical data One hundred and forty one consecutive patients (85 men and 56 women; mean age 40.1 years; range 21–60) grafted for type 1 diabetes from a single center have been included. Transplantations consisted in 119 SPKP and 22 SPT (Table 1) with portal venous drainage for 114 patients and systemic drainage for 19 (missing data for eight patients) (Figs. 1, 2, and 3). All were exocrine enteric drainage. Postoperative care consisted in an induction therapy with a polyclonal antilymphocyte agent and the maintenance of immunosuppression with tacrolimus, mycophenolate mofetil, and prednisone. This was associated with an antibacterial prophylaxis with a broadspectrum antibiotic (Ceftriaxone and metronidazole) for 5 days. Recipients also underwent standard prophylaxis
Fig. 1. Normal pancreas graft (white arrowheads). Enhanced coronal MDCT.
Table 1. Patient characteristics Characteristics Recipients (n) Sex (M/F) Mean age ± SD (years) BMI: mean ± SD Donor Sex (M/F) Mean age ± SD (years) BMI: mean ± SD Transplantation procedures Kidney-pancreas / Pancreas alone (n) Exocrine enteric drainage Venous graft: Portal/systemic/ missing data Associated venous graft (n) Cold ischemia time (h) (mean ± SD)
141 85/46 40.1 (7.7) 22.4 ± 3.3 78/63 35.0 ± 10.9 23.1 ± 2.8 119/22 141 114/19/8 29 (missing data 20) 13.3 ± 3.0 (range 5.2–21.5)
Fig. 2. Pancreatic graft with systemic venous drainage. Right iliac arterial anastomosis (white arrow) and cava venous anastomosis (black arrow). Maximal Intensity Projection coronal enhanced MDCT.
M. Vincent et al.: Multi detector computed tomography
Table 2. Early complications Early complications
Venous thrombosis Partial Complete Arterial thrombosis Pancreatitis Fluid collection Fistula Bowel obstruction Hemorrhage Eventration
n
% of complications
% of total effective
6 11 2 6 21 2 3 12 1
9.4 17.2 3.1 9.4 32.8 3.1 4.7 18.8 1.5
4.3 7.8 1.4 4.3 14.9 1.4 2.1 8.5 0.7
Median time of diagnosis with MDCT (days, [range]) 4 [1–12] 5 [3–11] 10 11 [1–14] 13.5 [1–18] 7.5 [1–14] 10 [2–13] 6 [1–17] 16
Statistical considerations
Fig. 3. Pancreatic graft with portal venous drainage (black arrow). Maximal Intensity Projection coronal enhanced MDCT.
tric, Milwaukee, WI, USA). Respective parameters were: section reconstructions 2 and 1.25 mm, section interval 2 and 1.25 mm, section rotation speed 0.51 and 0.6 sec; 120 kV, automatic tube current, Field of View 40 cm and 512 9 512 matrix size. After a pre-contrast phase, arterial phase images were acquired using bolus tracking, after intravenous injection of non-ionic iodinated contrast (350, 1.5 mL/kg, rate 3 mL/s using a power injector). This was followed by a portal venous phase acquired at 60–70 s from the time of start of contrast injection. All series covered the entire abdomen from the diaphragm to the pelvis. Forty-one patients benefited of complete contrast-enhanced imaging of their abdomen and pelvis. Isolated unenhanced MDCT was performed for 10 patients due to acute renal failure. Data are missing for one patient. No oral contrast was administered. Multiplanar reconstructions (MPR), Maximum Intensity Projection (MIP), Volume Rendering (VR), and vascular reconstructions were performed on dedicated workstations. Diagnosis of early complications was based on initial clinical reports and retrospective review after consensus evaluation by two radiologists (M.V., 5 years of experience in radiology, E.F, 10 years of experience in gastrointestinal radiology). Partial venous thrombosis was considered as nonocclusive venous thrombosis involving the splenic or mesenteric vein, without parenchymal enhancement abnormality. Renal function was evaluated before and 3 days following injection of iodine contrast using the Cockcroft method.
All statistical analyses were performed with ‘‘Statistical Package for the Social Sciences’’ SPSS software for Windows (version 16.0; SPSS Inc. Chicago, IL). Continuous variables were summarized as mean (±SD), and categoric variables were summarized as frequency and percentage. A paired nonparametric Wilcoxon’s test was used to compare renal function before and after MDCT. The relationship between clinical data and early postoperative complications or venous thrombosis was carried out with a multivariate analysis (binary logistic regression). The threshold P value was 0.05.
Results Complications Fifty patients (35.5 %) suffered from 64 postoperative complications (Table 2), and 12 of them suffered more than one complication. MDCT was available for 41 of these patients. Nine patients with clinical postoperative complications were not investigated with MDCT (3 complete venous and 1 complete arterial thrombosis diagnosed with ultrasonography, 1 pancreatitis and 4 patients with early acute hemorrhage). Peripancreatic fluid collections (PPFC) were the most frequent (21 patients, 32.8 % of all complications) with a median period of 13.5 days (range 1–18) (Fig. 4). Five patients required percutaneous drainage and one a surgical approach. Acute thrombosis represented nearly a third of early complications, venous thrombosis being the most frequent (26.6 % of complications) (Figs. 5, 6, and 7). Thrombosis complicated 13.5 % of transplants, with a median diagnosis time of 5 days for partial venous thrombosis (range 3–11 days), four days for complete thrombosis (range 1–12 days), and three days for arterial thrombosis. Considering venous thrombosis, all partial venous thrombosis (6/6) were diagnosed following MDCT.
M. Vincent et al.: Multi detector computed tomography
Fig. 4. Peripancreatic fluid collection from duodenal leakage needing surgical reintervention. Enhanced Axial MDCT.
Fig. 6. Partial venous thrombosis with marginal thrombosis of the venous portal graft (white arrow). Normal enhancement of the graft (parenchyma white arrowheads and duodenum black arrowhead). Portal enhanced axial MDCT.
Fig. 5. Partial venous thrombosis of the graft splenic vein. Portal enhanced axial MDCT. Intravenous defect (white arrow) without pancreatic parenchymal enhancement abnormality (white arrowhead). Peripancreatic fluid collection (black arrow).
Hemorrhage represented 18.8 % of complications (n = 12). In four patients, acute hemorrhage occurred within the first day, corresponding to vascular anastomotic leak and was treated surgically without imaging. Secondary hemorrhage occurred with a median time of six days (range 1–17), surrounding the pancreas graft. Fourteen patients had to undergo pancreatectomy: all patients with CVT (n = 11), patients with arterial thrombosis (n = 2), and one patient with pancreatitis and hemorrhage. All complete thrombosis were confirmed operatively. Small bowel obstructions (3 patients, 4.7 % of complications) were due to postoperative adherences, as
Fig. 7. Complete venous thrombosis (white arrow) with lack of pancreatic parenchyma enhancement (black arrows). Normal renal graft in the left iliac fossa (white arrowheads). Coronal MPR enhanced MDCT.
demonstrated by the presence of a transition zone on MDCT. One patient required surgery.
Renal function Renal function before and three days after MDCT was evaluated with the Cockroft formula in 34 patients.
M. Vincent et al.: Multi detector computed tomography
Table 3. Multivariate analysis of risk factors for venous thrombosis of the pancreas graft Risk factor Recipient BMI Donor BMI Donor age Transplantation Venous drainage Venous graft Cold ischemia time
‡25 kg/m2 (vs.
Abdominal Imaging
Abdom Imaging (2014) DOI: 10.1007/s00261-014-0164-3
Multi detector computed tomography (MDCT) for the diagnosis of early complications after pancreas transplantation M. Vincent,1 O. Morla,1 J. Branchereau,2 G. Karam,2 B. Dupas,1 E. Frampas1 1
Central Department of Radiology and Medical Imaging, Hoˆtel-Dieu, CHU Nantes, 1 Place Alexis Ricordeau, 44093 Nantes Cedex 1, France 2 Department of Urology, Hoˆtel-Dieu, CHU Nantes, 1 Place Alexis Ricordeau, 44093 Nantes Cedex 1, France
Abstract Purpose: Solitary Pancreas (SPT) and simultaneous kidney-pancreas (SPKT) transplants carry a high risk of surgical complications that may lead to the loss of the pancreas graft and impact later kidney function. The purpose of this study was to investigate the role of MDCT in the diagnosis of early complications and its impact on kidney function. Methods: All patients receiving SPT or SPKT over 5 years were retrospectively included. Complications that occurred within the first 15 days were registered and MDCT data analyzed. Data regarding donor, transplant, and recipient characteristics as well as transplantation procedures were analyzed according to the occurrence of early complications. Kidney function at day 3 following MDCT was evaluated. Results: One hundred and forty-one patients were included (85 men, 56 women; mean age 40.1 years, SD 7.7) with 119 SPKT and 22 SPT. Sixty-four complications were registered in 50 patients. Partial (P-) or complete venous thrombosis (C-VT) occurred in 12.1 % (n = 17), arterial thrombosis (AT) in 1.4 % (n = 2), and hemorrhage in 8.5 % (n = 12) of all patients. For venous thrombosis, the predominant risk factor was body mass index (BMI) for either recipients (P < 0.05) or donors (P < 0.01). Median time for venous thrombosis diagnosis with MDCT was 4 days. Kidney function was not altered following MDCT. Fourteen pancreatectomies were necessary. All patients with C-VT and AT had to undergo graftectomy. Conclusion: Vascular complications occurred early following grafting. Systematic early-enhanced MDCT at day 2–3 should be adequate to detect early thrombosis,
Correspondence to: E. Frampas; email: [email protected]
especially if risk factors have been identified, without induced kidney function alteration. Key words: MDCT—Pancreas transplant—Surgical complications—Graft venous thrombosis—Graft failure
In patients with insulin-requiring diabetes, pancreas transplantation is considered as a successful treatment to normalize glucose metabolism, and to improve and prevent secondary diabetic complications. Since 1966, more than 37000 pancreas transplants have been reported to the International Pancreas Transplant Registry [1]. Most of them are simultaneous pancreatic and kidney transplants (SPKT), representing 75 % of cases, compared to those given after kidney transplants (PAKT), or to solitary pancreas transplants (SPT). Improvements in surgical techniques, postoperative management, and immunosuppressive protocols have led to improved patient and graft survival, with pancreas graft survival reaching 86 % at one year for SPK. Reported 1-year patient survival varies between 92 %–98 % and pancreatic transplantation offers considerable survival benefits compared to the mortality of patients on waiting list which approaches ~30% [2]. Nevertheless, pancreatic transplantation remains a complex surgical procedure with a high risk of surgical complications, and the need for relaparotomy in up to 43 % of cases which significantly affects pancreas graft survival [3]. The most frequent complications are fluid collections while venous thrombosis is responsible for the highest rate of pancreatectomy [4]. Diagnosis of early surgical complication is crucial as this can impact kidney graft survival. Lower kidney graft survival rates have been reported in patients with early pancreas loss as
M. Vincent et al.: Multi detector computed tomography
compared to recipients without pancreas failure (71.4 % and 59.5 % at one and three years vs. 86 % and 82 %) [5]. The purpose of this study was to determine the risk factors that predispose to early complications within the 15 first days following pancreas transplantation, and evaluate the use of MDCT in the early postoperative course.
Materials and methods Study design
against thrombosis with low-dose heparin during the hospital stay followed by oral antiplatelet therapy with aspirin and dipyridamole.
CT Imaging When clinically suspected, complications were investigated with multidetector CT (MDCT). Patients were scanned with 16 or 64-detector CT (Sensation 16, Siemens Healthcare, or Lightspeed VCT 64, General Elec-
Following institutional review, this retrospective study included all adult pancreas transplants over a five-year period from January 2004 to February 2009 in a single institution. Data regarding donor, transplant and recipient characteristics, graft and patient outcomes were obtained from the electronic medical database of the institution (Clinicom, Intersystems). Considering the retrospective nature of our study, our institutional review board judged it to be exempt from informed consent. All complications were registered and considered as early when they occurred during the 15 days following the graft.
Patient and clinical data One hundred and forty one consecutive patients (85 men and 56 women; mean age 40.1 years; range 21–60) grafted for type 1 diabetes from a single center have been included. Transplantations consisted in 119 SPKP and 22 SPT (Table 1) with portal venous drainage for 114 patients and systemic drainage for 19 (missing data for eight patients) (Figs. 1, 2, and 3). All were exocrine enteric drainage. Postoperative care consisted in an induction therapy with a polyclonal antilymphocyte agent and the maintenance of immunosuppression with tacrolimus, mycophenolate mofetil, and prednisone. This was associated with an antibacterial prophylaxis with a broadspectrum antibiotic (Ceftriaxone and metronidazole) for 5 days. Recipients also underwent standard prophylaxis
Fig. 1. Normal pancreas graft (white arrowheads). Enhanced coronal MDCT.
Table 1. Patient characteristics Characteristics Recipients (n) Sex (M/F) Mean age ± SD (years) BMI: mean ± SD Donor Sex (M/F) Mean age ± SD (years) BMI: mean ± SD Transplantation procedures Kidney-pancreas / Pancreas alone (n) Exocrine enteric drainage Venous graft: Portal/systemic/ missing data Associated venous graft (n) Cold ischemia time (h) (mean ± SD)
141 85/46 40.1 (7.7) 22.4 ± 3.3 78/63 35.0 ± 10.9 23.1 ± 2.8 119/22 141 114/19/8 29 (missing data 20) 13.3 ± 3.0 (range 5.2–21.5)
Fig. 2. Pancreatic graft with systemic venous drainage. Right iliac arterial anastomosis (white arrow) and cava venous anastomosis (black arrow). Maximal Intensity Projection coronal enhanced MDCT.
M. Vincent et al.: Multi detector computed tomography
Table 2. Early complications Early complications
Venous thrombosis Partial Complete Arterial thrombosis Pancreatitis Fluid collection Fistula Bowel obstruction Hemorrhage Eventration
n
% of complications
% of total effective
6 11 2 6 21 2 3 12 1
9.4 17.2 3.1 9.4 32.8 3.1 4.7 18.8 1.5
4.3 7.8 1.4 4.3 14.9 1.4 2.1 8.5 0.7
Median time of diagnosis with MDCT (days, [range]) 4 [1–12] 5 [3–11] 10 11 [1–14] 13.5 [1–18] 7.5 [1–14] 10 [2–13] 6 [1–17] 16
Statistical considerations
Fig. 3. Pancreatic graft with portal venous drainage (black arrow). Maximal Intensity Projection coronal enhanced MDCT.
tric, Milwaukee, WI, USA). Respective parameters were: section reconstructions 2 and 1.25 mm, section interval 2 and 1.25 mm, section rotation speed 0.51 and 0.6 sec; 120 kV, automatic tube current, Field of View 40 cm and 512 9 512 matrix size. After a pre-contrast phase, arterial phase images were acquired using bolus tracking, after intravenous injection of non-ionic iodinated contrast (350, 1.5 mL/kg, rate 3 mL/s using a power injector). This was followed by a portal venous phase acquired at 60–70 s from the time of start of contrast injection. All series covered the entire abdomen from the diaphragm to the pelvis. Forty-one patients benefited of complete contrast-enhanced imaging of their abdomen and pelvis. Isolated unenhanced MDCT was performed for 10 patients due to acute renal failure. Data are missing for one patient. No oral contrast was administered. Multiplanar reconstructions (MPR), Maximum Intensity Projection (MIP), Volume Rendering (VR), and vascular reconstructions were performed on dedicated workstations. Diagnosis of early complications was based on initial clinical reports and retrospective review after consensus evaluation by two radiologists (M.V., 5 years of experience in radiology, E.F, 10 years of experience in gastrointestinal radiology). Partial venous thrombosis was considered as nonocclusive venous thrombosis involving the splenic or mesenteric vein, without parenchymal enhancement abnormality. Renal function was evaluated before and 3 days following injection of iodine contrast using the Cockcroft method.
All statistical analyses were performed with ‘‘Statistical Package for the Social Sciences’’ SPSS software for Windows (version 16.0; SPSS Inc. Chicago, IL). Continuous variables were summarized as mean (±SD), and categoric variables were summarized as frequency and percentage. A paired nonparametric Wilcoxon’s test was used to compare renal function before and after MDCT. The relationship between clinical data and early postoperative complications or venous thrombosis was carried out with a multivariate analysis (binary logistic regression). The threshold P value was 0.05.
Results Complications Fifty patients (35.5 %) suffered from 64 postoperative complications (Table 2), and 12 of them suffered more than one complication. MDCT was available for 41 of these patients. Nine patients with clinical postoperative complications were not investigated with MDCT (3 complete venous and 1 complete arterial thrombosis diagnosed with ultrasonography, 1 pancreatitis and 4 patients with early acute hemorrhage). Peripancreatic fluid collections (PPFC) were the most frequent (21 patients, 32.8 % of all complications) with a median period of 13.5 days (range 1–18) (Fig. 4). Five patients required percutaneous drainage and one a surgical approach. Acute thrombosis represented nearly a third of early complications, venous thrombosis being the most frequent (26.6 % of complications) (Figs. 5, 6, and 7). Thrombosis complicated 13.5 % of transplants, with a median diagnosis time of 5 days for partial venous thrombosis (range 3–11 days), four days for complete thrombosis (range 1–12 days), and three days for arterial thrombosis. Considering venous thrombosis, all partial venous thrombosis (6/6) were diagnosed following MDCT.
M. Vincent et al.: Multi detector computed tomography
Fig. 4. Peripancreatic fluid collection from duodenal leakage needing surgical reintervention. Enhanced Axial MDCT.
Fig. 6. Partial venous thrombosis with marginal thrombosis of the venous portal graft (white arrow). Normal enhancement of the graft (parenchyma white arrowheads and duodenum black arrowhead). Portal enhanced axial MDCT.
Fig. 5. Partial venous thrombosis of the graft splenic vein. Portal enhanced axial MDCT. Intravenous defect (white arrow) without pancreatic parenchymal enhancement abnormality (white arrowhead). Peripancreatic fluid collection (black arrow).
Hemorrhage represented 18.8 % of complications (n = 12). In four patients, acute hemorrhage occurred within the first day, corresponding to vascular anastomotic leak and was treated surgically without imaging. Secondary hemorrhage occurred with a median time of six days (range 1–17), surrounding the pancreas graft. Fourteen patients had to undergo pancreatectomy: all patients with CVT (n = 11), patients with arterial thrombosis (n = 2), and one patient with pancreatitis and hemorrhage. All complete thrombosis were confirmed operatively. Small bowel obstructions (3 patients, 4.7 % of complications) were due to postoperative adherences, as
Fig. 7. Complete venous thrombosis (white arrow) with lack of pancreatic parenchyma enhancement (black arrows). Normal renal graft in the left iliac fossa (white arrowheads). Coronal MPR enhanced MDCT.
demonstrated by the presence of a transition zone on MDCT. One patient required surgery.
Renal function Renal function before and three days after MDCT was evaluated with the Cockroft formula in 34 patients.
M. Vincent et al.: Multi detector computed tomography
Table 3. Multivariate analysis of risk factors for venous thrombosis of the pancreas graft Risk factor Recipient BMI Donor BMI Donor age Transplantation Venous drainage Venous graft Cold ischemia time
‡25 kg/m2 (vs.
