Early Relaparotomy After Simultaneous Pancreas-Kidney Transplantation

Abstract

Background 

Simultaneous pancreas-kidney transplantation (SPKT) is a promising therapy for type 1 diabetes mellitus with chronic kidney disease. Although the long-term outcome of SPKT is extensively documented, the incidence of early complications within the first weeks after the surgery is less described. The aim of this study was to assess the incidence, causes, and risk factors of early relaparotomy after SPKT.

Methods 

All SPKT performed in the university hospital between 2005 and 2008 were enrolled. The primary endpoint was defined as the need for at least one relaparotomy after SPKT within the initial hospital stay. The secondary endpoints were the incidence of vascular graft thrombosis, postoperative sepsis, patient, and graft survival.

Results 

Sixty-one patients were included. During their initial hospital stay, 27 (44.3%) SPKT recipients required at least one relaparotomy. The main causes of relaparotomy were hemorrhage (59.3%) and vascular graft thrombosis (22.2%). First relaparotomy occurred at a median postoperative time of 1 day (interquartile range, 1–6). Pretransplant dialysis and nontraumatic cause of donor brain death were identified as independent risk factors for early relaparotomy. Thirty-two patients (52.4%) experienced a symptomatic or asymptomatic vascular graft thrombosis.

Conclusions 

The early postoperative period remains a high-risk phase for relaparotomy. The selection of recipients before initiation of long-term dialysis and of donors deceased from traumatic causes may reduce the rate of these early complications after SPKT. Vascular graft thrombosis and bleeding are two major issues that arise during this critical period, suggesting the importance an adequate management of postoperative anticoagulation and hemostasis.

Since 1966, pancreas transplantation has been developed as an endocrine replacement therapy designed to achieve an insulin-independent euglycemic state (¹). Simultaneous pancreas-kidney transplantation (SPKT) is proposed for patients with insulin-dependent diabetes mellitus and end-stage renal disease. Improvement in the quality of life and successful long-term outcomes have been highlighted by many publications (^2–4). SPKT has been reported to be superior to dialysis or renal transplant alone, regarding diabetic patient survival (^{5, 6}). Nevertheless, the long-term benefits of SPKT must be weighed against the potential risk of early mortality and morbidity associated with the procedure.

During these last decades, there have been many improvements in the management of SPKT, especially in the area of surgical techniques and immunosuppressive regimens (¹). Despite this progress, SPKT is still associated with significant initial morbidity, and the transplantation failure rate remains significant (⁵). The pancreas graft survival rate is 85% at 1 year according to the International Pancreas Transplant Registry (⁷). Surgical complications, such as vascular graft thrombosis and hemorrhage, are important issues because they can lead to reintervention, graft loss, and delayed hospital discharge (⁸). Gruessner and colleagues (⁹) showed an increase in the mortality hazard during the first weeks after pancreas transplantation as compared with patients who remained on the waiting lists for pancreas transplant. The incidence of early relaparotomy in the days and weeks after SPKT is less known because only very few studies have been focused on this period (¹⁰).

A better knowledge of the factors involved in these early complications could help improve immediate postoperative care. The aim of our prospective observational study was to analyze these early relaparotomies and identify risk factors to improve postoperative care with prompt diagnosis and prophylactic measures.

RESULTS

Demographic Data

Sixty-one patients were included. The demographic characteristics of these patients are presented in Table 1. In all cases, type I diabetes was the cause of pancreas insufficiency. At baseline, the prevalence of long-term complications of diabetes was 98.4% for hypertension, 93.4% for retinopathy, 19.7% for gastroparesis, and 14.8% for coronaropathy. None of them showed a history of stroke or extremity amputation. Sixty-one percent of patients were on long-term dialysis.

TABLE 1: Donor and recipient baseline characteristics

Relaparotomy

During their initial hospital stay, 27 patients (44.3%) required at least one relaparotomy. One- and three-month relaparotomy rates were 39.3% and 44.3%, respectively. Twenty-one patients experienced one surgical procedure, four patients required two procedures, and two patients needed three repeat laparotomies. The leading cause of the first relaparotomy was hemorrhage (n=16, 59.3% of relaparotomies), followed by vascular graft thrombosis (n=6, 22.2%), intra-abdominal sepsis (n=4, 14.8%), and another cause is intestinal occlusion (n=1, 3.7%). A vascular graft anastomotic leak was identified as the cause of bleeding for five patients (two superior mesenteric arterial anastomoses on the arterial Y-graft, two venous anastomoses, and one common iliac arterial anastomosis). Graft pancreatectomy was performed in seven patients (26%) who underwent relaparotomy. First relaparotomy occurred at a median postoperative time of 1 day (interquartile range, 1–6). Regarding outcome, the length of hospital stay was 31 (17) days in the relaparotomy group and 19 (9) days in the group without relaparaomy (P<0.0001). Intensive care unit (ICU) length of stay was 6 (8) days in the relaparotomy group and 3 (2) days in the group without relaparotomy (P=0.002).

The results of the univariate analysis are presented in Table 2. In univariate analysis, therapeutic anticoagulation was not associated with relaparotomy (P=0.90). Table 3 shows the multivariate analysis of independent risk factors for early relaparotomy after SPKT. The two independent risk factors were nontraumatic cause of donor death and pretransplant dialysis. Cold ischemia duration, recipient, and donor age were not either identified as risk factors for early relaparotomy.

TABLE 2: Univariate analysis of risk factors for experiencing relaparotomy after SPKT

TABLE 3: Multivariate analysis of risk factors for experiencing relaparotomy after SPKT

Other Events

The overall vascular graft thrombosis rate was 52.4% (32 patients), which was diagnosed either by tomography or clinically during reoperation. Only one case of thrombosis concerned the kidney graft. Localization of vascular graft thrombosis was arterial (n=13, 40.6% of thromboses), venous (n=12, 37.5% of thromboses), or both (n=7, 21.8% of thromboses). Most of the arterial or venous thromboses were partial (n=25, 78.1% of thromboses). Seven patients developed a complete vascular graft thrombosis that required graft removal (six pancreases and one kidney). Twelve patients required heparin-based therapeutic anticoagulation for partial vascular graft thrombosis, depending on localization and extension. Median number of red blood cell units transfused during the ICU stay was 0 (0–2) in the group without relaparotomy versus 4 (3–8) in the group with relaparotomy (P<0.0001).

Seventeen patients (27.9%) developed sepsis; 14 were intra-abdominal sepsis, and 3 were nosocomial pneumoniae. Four cases of intra-abdominal sepsis led to repeat laparotomies. One patient experienced graft pancreatitis. Isolated abdominal pathogens are presented in Table 4.

TABLE 4: Cause of intra-abdominal sepsis and pathogens isolated

Patient and Graft Survival

The follow-up of patients for graft loss or death was 1 year. No patient died, and the pancreas graft survival rate was 85% at 1 month. Kidney graft survival rate was 98% at 1 month (Fig. 1). The pancreas and kidney 1-year survival rate were 82% and 98%, respectively. Although it did not reach statistical significance, there was a trend toward a higher rate of pancreas graft loss for patients who underwent relaparotomy versus those without relaparotomy, with a graft survival rate of 77.8% and 91.2% at one month, respectively (P=0.09) (Fig. 2).

FIGURE 1: Graft survival rates for simultaneous pancreas-kidney transplantation recipient.

FIGURE 2: Pancreas graft survival rates for simultaneous pancreas-kidney transplant recipient with vs. without relaparotomy.

The causes of pancreas graft loss at 1 month in the group with relaparotomy included the following: 3 combined arterial and venous thromboses, 1 ischemia without thrombosis, 1 graft vascular mycotic aneurysm, and 1 graft pancreatitis with enteric anastomotic leak. The causes at 1 month in the group without relaparotomy included the following: 2 complete arterial thromboses and 1 venous thrombosis.

The pancreas graft survival rate censored for graft removed for thrombosis was not different either between patient who underwent relaparotomy and those without relaparotomy (P=0.07) (Fig. 3). No pancreas graft presented primary nonfunction, and one presented chronic rejection at 8 months. Only two kidneys with delayed graft function requiring postoperative dialysis were observed.

FIGURE 3: Pancreas graft survival rates censored for graft removed for thrombosis, for simultaneous pancreas-kidney transplant recipient with vs. without relaparotomy.

DISCUSSION

In this observational study, we found that 44% of SPKT recipients who were admitted to our institution between 2005 and 2008 underwent at least one relaparotomy. This finding demonstrates that pancreas transplantation remains a major surgical procedure with a significant early morbidity. Thus, initial postoperative care may be an important period regarding this morbidity after SPKT.

Various reoperation rates have been reported, and the impact on long-term outcome is debated (¹⁰). In a series of 441 consecutive, whole organ, bladder-drained pancreas transplantations performed between 1986 and 1994, 32% of recipients required relaparotomy (¹¹). Michalak and coworkers (¹²) reported a reoperation rate of 45% after SPKT with bladder or enteric drainage during a 16-year period. Humar and colleagues (⁸) identified a significant decrease of relaparotomy rate between two eras, from 1985 to 1994 (relaparotomy rate of 32%) and from 1994 to 1997 (relaparotomy rate of 19%). Despite important changes in SPKT techniques such as enteric drainage, morbidity remains still high. For example, in a recent prospective study of enteric-drained SPKT, an incidence of relaparotomy of 26% has been reported (¹³). Overall, most studies report relaparotomy rates ranging from 10% to 65% (^14–16). However, these rates, generally calculated during a follow-up period ranging from 3 months to 1 year, do not reflect short-term morbidity. Our results focus on the immediate postoperative period and pinpoint the high morbidity that occurs during the first few days after SPKT.

Knowing the risk factors of these severe complications is crucial because it could impact the selection of donors and recipients for SPKT. For example, the recruitment of the recipients before dialysis (i.e., preemptive transplantation) rather than at the time of dialysis could be an interesting approach. Indeed, recipients receiving dialysis before transplantation were found to classically have a poorer 1-year survival rate compared with those undergoing preemptive SPKT (¹⁷). In another study, Pruijm et al. (¹⁸) failed to demonstrate any benefit of being dialyzed before transplantation, on the pancreas, kidney, or in patient long-term survival. Paradoxically, dialysis was shown to improve platelet functional abnormalities related to uraemia and to reduce the risk of hemorrhage, the main cause of relaparotomy identified in our study (^{19, 20}). However, the deleterious effect of dialysis before transplantation has also been reported in kidney alone transplantation with more risk of acute rejection (^{21, 22}). Nevertheless, a histopathologic analysis of explanted pancreases for graft thrombosis suggested that the cause of thrombosis was rejection in one-third of the cases (²³). This suggests that the deleterious effect of dialysis on relaparotomy rates could be at least partially explained by changes in immune reactivity induced by long-term dialysis. These changes could increase the rate of immunologically related graft thrombosis. However, the implication of an immunologic mechanism rather than ischemia-reperfusion injury on early graft thrombosis is questionable. Nontraumatic cause of death in the donor was the second risk factor identified in our multivariate analysis. This relationship between pancreas graft thrombosis and a cerebrovascular cause of donor death has already been identified (^{24, 25}).

The impact of this morbidity on transplant survival is debated (²⁶). In this study, pancreas and kidney graft 1-year survival rates were 84% and 98%, respectively. These survival rates are in accordance with results published by other investigators (^{7, 27, 28}). The cumulative graft survival rate analysis demonstrates that most of the pancreatectomies occurred during the first 3 days after the transplantation. In our study, the second surgical procedure performed within a short amount of time saved the transplants in most of the cases.

Our study has several limitations. Regarding the choice of the primary endpoint, a composite endpoint was defined as the first complication leading to a reoperation, regardless of the cause of the relaparotomy. This endpoint was chosen to assess nonspecific morbidity of the early postoperative period. However, the total number of surgical complications was slightly underestimated by this method. Another limitation concerns the relatively small size of our population. This may explain the failure to detect some risk factors previously reported in the literature. For example, recipient age of more than 45 years, previously reported as a risk factor of poorer graft survival rate, was not identified in our study as a risk factor of early morbidity (^{1, 4, 8}). Recipient obesity could also not be identified as a risk factor because of the low incidence of this disease in our population (²⁹). Although reported as a risk factor in several studies, donor age was not even identified as a pejorative factor in this study (^{4, 8, 30}).

Although bleeding was the most frequent cause of relaparotomy in our series, partial or complete vascular graft thrombosis was also identified in many cases. The high rate of partial vascular graft thrombosis may be related to the systematic sonography or computed tomography performed in postoperative care, which leads to diagnoses of asymptomatic thrombosis. These hemorrhagic and thrombotic issues leading to relaparotomies have previously been reported in other studies (^{8, 13, 31}). Multiple factors may be associated with pancreas graft thrombosis, problems related to the donor or the recipient, the procurement and preservation injuries, or the transplantation procedure (^{25, 29}). Diabetes-related hypercoagulable state has already been described (³²). A low blood flow through the splenic and superior mesenteric veins of the pancreas graft increases the risk of thrombosis (^{29, 33}). Ischemia-reperfusion–induced microvascular dysfunction of the pancreas may also play an important role in thrombotic events after SPKT (^{34, 35}). A better knowledge of molecular pathways involved in these microvascular disorders could help in proposing specific therapies to prevent graft thrombosis.

There is no consensus for graft thrombosis prophylaxis after SPKT and no prospective study supporting its benefit (²⁹). Although platelet inhibitors (mainly aspirin) are used in most transplant centers during the postoperative months, the ideal prophylactic anticoagulation regimen is still unknown (^{33, 36}). The elevated hemorrhage rate identified in our study strongly suggests that heparin could have played a role in the occurrence of bleeding. Heparin should therefore be manipulated with caution in this early postoperative period. A thromboelastogram (TE) performed either during the surgical procedure or the early postoperative period could represent one strategy to rapidly diagnose hypercoagulability or hemostatic disorders and therefore guide the anticoagulant administration (³⁷). Based on TE parameters, Burke and colleagues (³⁸) demonstrated that patients undergoing SPKT were significantly more susceptible to hypercoagulability than nondiabetic kidney transplant recipients. In another analysis of 74 SPKT, TE was able to diagnose hypercoagulability in 34% of patients, leading to anticoagulation administration (³³). This data suggests that TE could improve postoperative anticoagulation management by monitoring coagulation changes occurring after revascularisation of the pancreas and kidney grafts.

In conclusion, hemorrhage and vascular graft thrombosis are the main causes of early relaparotomy. Pretransplant dialysis and nontraumatic cause of donor brain death have been identified as risk factors of relaparotomy. Better selection criteria including these risk factors, such as proposing SPKT to potential recipients before dialysis or avoiding donors deceased from cerebrovascular cause, could improve early morbidity. Moreover, our findings suggest that prophylactic heparin could increase the risk of bleeding and highlight better diagnosis and that rapid treatment of coagulation disorders are the cornerstones of immediate postoperative management of SPKT. The usefulness of TE for SPKT should be assessed by other studies.

MATERIALS AND METHODS

Patients

All patients who underwent SPKT at the transplant center of our university hospital in Lyon, France, between January 1, 2005, and December 31, 2008, were included. Patients who received either pancreas transplantation alone or pancreas after kidney transplantation were excluded. The severity of clinical parameters on admission was assessed by the simplified acute physiology score II and the sequential organ failure assessment. The regional ethics committee of Lyon approved this observational study.

Transplantation Technique

Kidneys and pancreases were obtained from brain-dead multiorgan deceased donors. A classical reversed Y-graft from the donor iliac artery was sutured onto the donor superior mesenteric and splenic arteries of the pancreas graft. General anesthesia with tracheal intubation and neuromuscular blockade was performed. A central venous catheter and an oesophageal doppler were inserted. Ceftriaxone and metronidazole were used as antimicrobial prophylaxis for 2 days. At our institution, all patients who undergo SPKT are admitted to the ICU for the postoperative period.

Exocrine duct management using an enteric drainage technique was implemented in all SPKT cases. A systemic venous drainage of the pancreas allograft was performed. Five surgeons performed all the transplantations. All recipients received quadruple immunosuppressive induction therapy, which was composed of thymoglobulin (1.5 mg/kg/day for 10 days), methylprednisolone, mycophenolate mofetil, and tacrolimus. Thymoglobulin was suspended if platelet count decreased below 70×10⁹/L. Immunosuppressive maintenance therapy was based on tacrolimus, mycophenolate mofetil, and prednisolone. In the absence of postoperative bleeding, 100 U/kg per day of unfractionated heparin were administered within 8 to 12 hr after transplantation with an activated partial thromboplastin time goal of 40 to 50 sec to prevent vascular graft thrombosis. Therapeutic anticoagulation (400 U/kg/day of unfractionated heparin with an activated partial thromboplastin time goal of 60 to 80 sec) was administered in selected cases for partial vascular graft thrombosis, depending on localization and extension of the thrombus.

Outcomes

Data related to the type of surgical and medical complications were recorded. The primary endpoint was defined as the need for a relaparotomy after transplantation within the initial hospital stay. The reasons for the first relaparotomy were classified as hemorrhage, vascular graft thrombosis, intra-abdominal infection, or other cause.

The secondary endpoints were the occurrence of symptomatic or asymptomatic vascular thrombosis of pancreas or kidney graft, postoperative sepsis, patient, and graft survival. Vascular graft thrombosis was identified by abdominal sonography performed routinely within the 3 days after transplantation, confirmed by computed tomography when necessary or clinically identified during reoperation. In this work, bleeding was defined as a hemorrhage that could not be controlled and solved with transfusions and correction of coagulopathy disorders. General complications analyzed was sepsis. Sepsis was defined as a systemic inflammatory syndrome in response to infection, according to the American College of Chest Physicians/Society of Critical Care Medicine Consensus (³⁹).

Patient and graft survival were also assessed. Renal allograft loss was defined as transplant nephrectomy. Pancreas graft loss was defined as a return to insulin therapy.

Statistics

Data are described as mean (SD) or relative frequencies. Time to first relaparotomy and units of red blood cell transfusion in ICU are shown as median interquartile range (nonparametric data). Categorical data were compared using the chi-square test; and continuous variables, using the Mann-Whitney U test.

Potential risk factors for relaparotomy after SPKT with P<0.2 on univariate analysis were included in the multivariate logistic regression model. Studied covariates were recipient sex, age (≤45 vs. >45 years), mean duration of diabetes mellitus (≤30 vs. >30 years), pretransplant dialysis status (yes vs. no), history of tabagism, diabetic retinopathy, diabetic neuropathy, pancreas cold ischemia time (≤12 vs. >12 hr), pancreas warm ischemia time (≤30 vs. >30 min), donor age (≤40 vs. >40 yr), traumatic or nontraumatic cause of donor death, desmopressin administration to the donor 48 hr before transplantation, and donor ICU length of stay (≤3 vs. >3 days). A P value of ≤0.05 was considered statistically significant. Pancreas and kidney graft survival rates were calculated according to the Kaplan-Meier method and compared using the log-rank test. SPSS statistical software (version 17; SPSS Inc., Chicago, IL) was used for data management and analysis.

ACKNOWLEDGMENTS

The authors thank Kathryn Arbogast and Rita Oliveira for their technical assistance.