Viken Douzdjian

Multivariate analysis of donor risk factors for pancreas allograft failure after simultaneous pancreas-kidney transplantation

BACKGROUND Donor and recipient selection criteria for pancreas allograft are not standardized and may vary from center to center. METHODS Simultaneous pancreas-kidney transplantations performed between April 1988 and June 1994 were reviewed (n = 61), and univariate and multivariate analyses of factors that affect pancreas graft survival were performed. Analysis of all cases and cases excluding early thrombosis were performed separately. RESULTS Pancreas graft survival when early thrombosis was excluded and in the overall group was 76% and 70%, respectively, at 1 year. Although blood group and donor gender were weak predictors of graft survival by univariate analysis, neither affected graft survival in the multivariate model. Risk factors for graft failure as determined by Cox regression analysis and in descending order of significance were (1) duration of brain death before procurement, (2) length of donor admission, and (3) donor age of 40 years or older. The risk of graft failure for each of these factors was increased 2.2-, 3.2-, and 4-fold, respectively. Prolonged brain death was the only risk factor in the overall group, suggesting an association with early thrombosis. CONCLUSIONS Center-specific donor risk factors for pancreas graft survival after simultaneous pancreas-kidney transplantation were identified in this study, the importance of which need to be better defined.

Renal allograft and patient outcome after transplantation pancreas-kidney versus kidney-alone transplants in type 1 diabetic patients versus kidney-alone transplants in nondiabetic patients

Despite recent advances and improved outcome, pancreas transplantation remains controversial. The purpose of this review was to study renal allograft outcome after simultaneous pancreas-kidney transplants (SPK, n = 61), kidney-alone transplants in type I diabetic patients (KA-D, n = 63), and kidney-alone transplants in nondiabetic patients (KA-ND, n = 80). Patients were matched for donor age, donor gender, donor race, interval from donor admission to procurement, DR mismatch, and recipient gender. The mean renal allograft cold ischemic time and recipient age were lower in the SPK group. Patient survival was highest in the KA-ND group (99% and 86% at 1 and 5 years, respectively), intermediate in the SPK group (90% and 78% at 1 and 5 years, respectively), and lowest in the KA-D group (89% and 66% at 1 and 5 years, respectively) (P = 0.004). similarly, renal allograft survival was higher in the KA-ND (89% and 63% at 1 and 5 years, respectively) and SPK (82% and 69% at 1 and 5 years, respectively) groups compared with the KA-D group (76% and 49% at 1 and 5 years, respectively) (P = 0.07). This difference disappeared when renal graft survival was censored for death, which probably reflects the selection bias. Actuarial pancreas graft survival was 76% and 62% at 1 and 5 years, respectively. Acute rejection (AR) was more frequent in the SPK group than in the KA-D and KA-ND groups (41% v 16% v 29%; P = 0.007). Delayed graft function (DGF), on the other hand, occurred more frequently in the KA-D group than in the KA-ND and SPK groups (66% v 55% v 38%; P = 0.08). Death as a result of a cardiovascular event occurred more frequently in the KA-D group. Cardiovascular death and renal graft failure occurred earlier in the SPK group. Cox regression analysis revealed a 1.6 and 1.8 times higher risk of renal graft failure in the SPK group when the donor was > or = 40 years old or female and a five times higher risk of graft failure in the KA-ND group in the presence of AR. Graft survival in patients with AR/DGF was lower than that in patients with no AR/no DGF in both the KA-D (71% and 63% v 100% and 100% at 1 and 5 years, respectively; P = 0.03) and KA-ND (90% and 56% v 100% and 100% at 1 and 5 years, respectively; P = 0.001) groups. Acute rejection did not affect graft survival in the SPK group. In the absence of AR, DGF had no effect on graft survival in any of the groups. Although the selection bias in favor of pancreas transplantation does not allow for definitive conclusions, our results show that outcome after SPK transplantation is acceptable and factors that influence the outcome after this procedure may be different from the ones affecting KA-D recipients.

Sphincter of oddi dysfunction after liver transplantation

To The Editor: We read with great interest the study by Douzdjian et al (1) presenting five patients with sphincter of Oddi dysfunction (SOD) following orthotopic liver transplantation (OLT). Many aspects of this peculiar entity remain obscure including its pathophysiology, incidence, and treatment. The study by Douzdjian et al (1), showing abnormal sphincter of Oddi manometry in patients with clinical evidence of SOD, provides further support for papillary dyskinesia as one cause of biliary problems following OLT. We concur with Douzdjian et al (1) that denervation and devascularization of the papilla as suggested by Stieber et al (2) is only a partial explanation of the disease. The fact that some patients develop SOD and others do not while the same surgical technique is applied is intriguing. Our recent experience suggests that a dilated recipient common bile duct (CBD) represents a significant risk factor for developing posttransplant SOD. Four patients with dilated CBD (>15 mm in diameter) had liver transplantation with choledochocholedochostomy with T-tube drainage. The discrepancy in size between donor and recipient CBD was minimal. Two of these patients had cryptogenic cirrhosis, one had autoimmune hepatitis, and one hepatitis C cirrhosis. At the time of T-tube damping, significant increases in bilirubin and alkaline phosphatase levels were observed in each patient. T-tube cholangiograms showed no evidence for anastomotic strictures or leaks, but increased or persistent diffuse dilatation of the CBD was present in all cases. The liver enzyme levels normalized within one to three days following T-tube reopening. Two patients had repeat attempts at T-tube clamping with similar results. None of these patients had evidence of rejection. Suspecting SOD, an ERCP was performed in each patient with insertion of a stent (8-10 French, 5 cm in length) through the papilla without sphincterotomy. The T tube could be clamped within 24 hr following the endoscopic procedure with no subsequent elevation of serum liver enzymes. Stents were removed endoscopically four to six weeks later without further biliary complications. Use of CC for anastomosis in recipients with dilated CBD is uncommon since hepatojejunostomy is usually preferred, particularly if there is marked discrepancy between donor and recipient CBD. This experience with four cases strongly suggests that CBD dilation is a important risk factor for posttransplant SOD. While the mechanism is unclear, preexisting papilla dyskinesia may be a contributing factor. However, resolution of SOD over time without specific intervention argues for a contribution from peritransplant factors, eg, denervation or devascularization. We are currently selecting hepatojejunostomy reconstruction in cases involving large recipient CBD, even in the absence of size discrepancy. Evaluation of other risk factors is warranted since posttransplant SOD can also occur in the presence of small CBD. We would also stress the importance of early diagnosis of SOD to permit appropriate therapy. Short stents placed endoscopically through the papilla appear to be curative. Endoscopic sphincterotomy has not been necessary in our patients. Recognition of this entity may prevent misdiagnoses, such as interpreting anastomotic narrowing as a significant stricture resulting in unnecessary endoscopic or surgical intervention. The bedside manometry technique through the T tube as described by Sherman et al (3) and proposed by Douzdjian et al (1) as a screening test for SOD may be useful. However, the specificity and sensitivity of this test is still unclear. For instance, in the study by Douzdjian et al (1), one of the five patients with SOD had low basal pressure. Similarly, there is no information about false positive manometry, which may result in unnecessary treatments. We consider the presence of diffuse dilatation of the CBD with elevation of the liver enzymes after T-tube clamping and which resolves promptly after reopening as sufficient to warrant a presumptive diagnosis of SOD and to proceed with treatment. PIERRE-ALAIN CLAVIEN, 1 MD, PHD CARLOS A. CAMARGO, JR, 1 M D JOHN BAILLIE, 2 M D J. GREGORY FITZ, 2 M D Liver Transplantation Program Department of Surgery I and Medicine 2 Duke University Medical Center Durham, North Carolina

Primary enteric drainage of the pancreas allograft revisited

BACKGROUND Historically, primary enteric drainage (ED) of exocrine secretions in pancreas allografts was associated with a poor outcome, mostly as a result of infectious complications. On the other hand, bladder drainage (BD), which is presently used in the majority of institutions, is associated with substantial urologic morbidity. The aim of this study is to reassess the role of primary ED by reviewing our experience with ED versus BD in simultaneous pancreas-kidney transplantations. STUDY DESIGN The records of all pancreas-kidney transplantations performed between October 1990 and September 1996 were reviewed (n = 42). Enteric drainage was used in the last 16 (38%) and BD in the first 26 (62%). The BD and ED groups were comparable with respect to donor and recipient characteristics. RESULTS Length of stay for the transplantation (mean +/- standard deviation) was significantly shorter with ED than with BD (12.9 +/- 5.6 versus 20.4 +/- 9.6 days, p = 0.007). The total number of readmissions (1.7 +/- 1.5 versus 1.2 +/- 1.2 days, p = 0.2) and the length of hospital stay in the first 6 months after discharge (13.7 +/- 16.2 versus 10 +/- 11.3 days, p = 0.4) were similar between BD and ED. Complications requiring admission were distributed as follows in BD and ED recipients: recurrent/persistent urinary complications (46% versus 6%, p = 0.01), dehydration (27% versus 6%, p = 0.05), symptomatic graft pancreatitis (8% versus 6%, p = 0.9), gastrointestinal disturbance (27% versus 12%, p = 0.1), and wound infection (12% versus 19%, p = 0.5). The duration of the operative procedure was shorter in ED than in BD (4.3 +/- 0.9 versus 5.4 +/- 0.8 hours, p = 0.01). Reoperation during the initial transplantation stay was necessary in 23% of the patients having BD, compared with none having ED (p = 0.04). Similarly, fewer ED patients underwent reoperations compared with BD patients in the first 6 months after discharge (38% versus 69%, p = 0.04). Hospital charges for ED were lower than for BD for the initial admission (

The impact of midline versus transverse incisions on wound complications and outcome in simultaneous pancreas‐kidney transplants: a retrospective analysis

73,458 +/- 17,103 versus

Effect of race on outcome after kidney and kidney-pancreas transplantation in type 1 diabetic patients

107,193 +/- 32,965, p = 0.001). Actuarial patient (96% versus 94%, p = 0.6), kidney (85% versus 87%, p = 0.9), and technically successful pancreas (90% versus 85%, p = 0.6) survival rates at 1 year were similar for BD and ED. CONCLUSIONS Our results indicate that, compared with BD, ED is associated with less morbidity and shorter hospitalization without compromising outcome. Primary ED is a viable alternative to BD in simultaneous pancreas-kidney transplantation. More clinical experience with careful cost-effectiveness analysis is needed to better assess the implications of primary ED.

Renal scintigraphy of an infarction in an en bloc transplantation of a horseshoe kidney

Intraperitoneal placement of the pancreas allograft, usually through a midline incision, has so far achieved the best results in pancreas transplantation. The usefulness and safety of a transverse incision has not been previously reported. The purpose of this study was to compare midline and transverse incisions, with respect to wound complications and outcome, in simultaneous pancreas-kidney transplant recipients with intraperitoneal placement of the pancreatic graft. The incidence of deep abscess formation, superficial abscess formation, wound leak, and fascial dehiscence, as well as graft survival, were retrospectively compared in 41 bladder-drained simultaneous pancreas-kidney recipients with a midline incision and in 15 with a transverse incision. The overall incidence of wound complications was similar (34% vs 20%, P=NS) in the two groups. Deep abscess formation occurred more frequently in the midline group (27% vs 0%, P=0.02). Staphylococcus epidermidis and Candida albicans were the most common microbial isolates from deep abscesses. Multivariate logistic regression analysis revealed donor age 40 years or older (P=0.04), the occurrence of a bladder leak (P=0.05), and a peak serum amylase in the 1st week of 1000 IU/l or greater (P=0.02) to be independent risk factors for the development of wound complications. The type of incision, however, was not found to be an independent risk factor. Patient (90% vs 83%, P=NS), pancreas allograft (78% vs 82%, P=NS), and kidney allograft (83% vs 70%, P=NS) survival rates were similar for the midline and transverse groups. We conclude that the transverse incision is a reasonable alternative to the midline incision in simultaneous pancreas-kidney transplantation and it is presently the incision of choice at our institution. It offers excellent exposure and is associated with a similar wound complication rate and outcome when compared to the midline incision.

Renal retransplants : Effect of primary allograft nephrectomy on early function, acute rejection and outcome

OBJECTIVE The racial impact on graft outcome is not well defined in diabetic recipients. The purpose of this study is to analyze our experience with kidney-alone (KA) and kidney-pancreas (KP) transplantation in type 1 diabetic recipients and evaluate the impact of racial disparity on outcome. RESEARCH DESIGN AND METHODS The records of 217 kidney transplants (118 KA, 99 KP) performed on type 1 diabetic patients between 1985 and 1995 at the Medical University of South Carolina and the University of Texas Medical Branch were reviewed. RESULTS A total of 53 (31%) white patients and 15 (33%) black patients experienced at least one episode of biopsy-proven acute rejection of the renal graft (NS). Patient survival at 1, 2, and 5 years was similar in white (92, 87, 69%) and black (91, 91, 69%) patients (NS). Kidney graft survival at 1, 2, and 5 years in the KA group was 72, 62, and 42% in blacks, compared with 79, 76, and 53% in whites (NS). Kidney graft survival at 1, 2, and 5 years in the KP group was 92, 92, and 74% in blacks, compared with 83, 77, and 58% in whites (NS). Pancreas graft survival at 1, 2, and 5 years was 81, 81, and 81% in blacks, compared with 81, 75, and 62% in whites (NS). Cox regression analysis revealed that donor age ≥ 40 years increased the risk of renal graft failure 6.2-fold (P = 0.0001), whereas the addition of a pancreas transplant to a kidney and a living-related transplant decreased the risk of failure of the kidney graft 0.2 (P = 0.005) and 0.1 times (P = 0.005). CONCLUSIONS Our results suggest that when compared with whites, there may be a trend toward an improved kidney and pancreas graft outcome in blacks undergoing KP transplants. These findings suggest that diabetes may override the risk factors that account for the pronounced disparity in outcome observed between nondiabetic white and black recipients.

Renal allograft failure after simultaneous pancreas-kidney transplantation: Univariate and multivariate analyses of donor and recipient risk factors

A 30-year-old woman with end stage renal disease received a cadaveric horseshoe kidney transplant. The horseshoe kidney embryologically represents the fusion of the lower poles of the kidneys in 95% of occasions. Another characteristic is the abnormal blood supply thought to be persistent from the embryonic period. Because each renal artery supplies its own area, there is no collateral circulation between the areas ― an artery ligated by accident will always cause an infarction. This often happens in the fusion bridge, which receives its separate blood vessels from the distal aorta, or even the iliac artery