EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Postoperative hospital readmissions constitute a significant burden on the health care system; readmissions are associated with an increase in health care costs, patient morbidity, and sometimes even increased mortality.¹ Early hospital readmission rates among general surgery patients have been reported to range from 3% to 26%.^2-4 Transplant recipients, due to their multiple comorbidities and immunocompromised status, are at even higher risk for hospital readmission. Readmission rates after kidney transplant have been estimated to be around 31% in the first 30 days posttransplant and can be as high as 46% in deconditioned patients.^5,6 Outcomes of pancreas transplant have witnessed significant improvements since the procedure was first described in 1966 by Kelly and Leillehei.^7,8 With advances in surgical technique, perioperative care, organ preservation, and immunosuppression, 5-year patient and graft survival rates are now above 93% and 73%, respectively.⁹ Despite these advances, readmission rates for pancreas transplant recipients are still high and reported as anywhere between 55% and 74%.^10-12

Recognizing the magnitude of this problem, we investigated the potential factors contributing to the high readmission rate in patients receiving pancreas transplant. We specifically looked at the time when the operation started and hypothesized that patients who undergo deceased pancreas transplant in the morning will have lower readmission rates than patients who undergo transplant in the afternoon or at night.

Materials and Methods

Institutional review board approval was obtained to retrospectively review all patients who underwent deceased donor pancreas transplant at a single high-volume center between January 2017 and December 2018. We included all patients who had simultaneous pancreas and kidney (SPK) transplant, pancreas after kidney (PAK) transplant, and pancreas transplant alone (PTA). There was a single case of a simultaneous genetically living unrelated kidney (husband and wife) and deceased donor pancreas transplant, which was included in the SPK group.

Patient demographics and comorbidities were reviewed. Donor organ variables, including cold ischemic time (CIT), kidney donor profile index (KDPI), organ origin (local or imported), and donation status (donation after circulatory death [DCD] or donation after brain death), were examined. Cold ischemia time was defined as the time from cross clamp at the donor operation to organ reperfusion in the recipient. Local organs were those procured from donors within our institution’s organ procurement organization (OPO), whereas imported organs were those shipped from OPOs across the country. Operative data such as venous drainage of the pancreas allograft (portal vs systemic), drainage of the pancreatic enzymes (enteric vs bladder), intraoperative estimated blood loss (EBL), use of drains, and start time of the operation were analyzed. Operations that started between 6 AM and 3 PM comprised the “morning group,” whereas operations that started between 3 PM and 7 PM and 7 PM and 6 AM comprised the “afternoon group” and “evening group,” respectively. The group divisions were based on shift changes of the operating room staff and anesthesia team. Evaluated postoperative laboratory parameters included peak lipase (U/L), serum creatinine (mg/dL), and lipase at 30 days posttransplant (U/L). Peak lipase was defined as the highest serum lipase value within 7 days of the transplant. The number of imaging studies obtained within 30 days of transplant, including pancreas allograft ultrasonograph and computerized tomography scans, was also evaluated.

The primary outcome evaluated was readmission rate within 30 days of discharge from the index hospitalization. Secondary outcomes included rates of reoperation, percutaneous drain placement for symptomatic fluid collection or hematoma, development of pancreatic leak or graft failure, and length of hospital stay. Pancreatic leak was defined as the presence of a peripancreatic fluid collection with a drain amylase level at least 3 times the serum level. Delayed graft function (DGF) of the renal allograft for patients with a concurrent kidney transplant was defined as any form of dialysis posttransplant within 7 days of operation. Graft failure was defined as death with a functioning graft, graft pancreatectomy, or return to insulin dependence.

Operative course
All pancreas transplants were from deceased donors and were done by 2 surgical teams in our institution; 1 team performed the backbench preparation and the second team performed the implantation. The 2 teams worked simultaneously in order to decrease the operative time. To minimize CIT, the cases were scheduled as soon as the organs arrived at the hospital. The backbench preparation consisted of splenectomy, shortening and oversewing of both duodenal ends, oversewing of the small bowel mesentery, and creation of the Y-graft. The implantation began with a midline laparotomy and mobilization of the right colon and small bowel mesentery to expose the inferior vena cava and the aortic bifurcation. For systemically drained allografts, the pancreas was anastomosed to the distal inferior vena cava and right common iliac artery. For portally drained allografts, the recipient’s superior mesenteric vein was used as an outflow. A handsewn, 2-layered, side-to-side duodenojejunostomy approximately 40 cm distal to the ligament of Treitz was used for exocrine drainage in all our transplant patients. For SPK transplants, the kidney was placed intraperitoneally on the left side. Prophylactic transabdominal drains in the retroperitoneum around the pancreas were placed at the discretion of the operating surgeon. After pancreas reperfusion, the mean arterial pressure was maintained above 75 mm Hg, and point-of-care glucose checks were performed every 10 minutes.

Posttransplant care and immunosuppression
Patients were admitted to the intensive care unit for 24 to 48 hours for close monitoring. Mean arterial pressure was maintained at >75 mm Hg. Laboratory variables, including complete blood count, chemistry, serum lactate, serum amylase, and lipase, were checked regularly every 6 hours for the first 24 hours. Point-of-care glucose checks were performed hourly. Nasogastric decompression was maintained for 24 hours posttransplant, and a clear liquid diet was initiated on posttransplant day 3. Aspirin (81 mg) was started on posttransplant day 1. Antibiotics were continued for 48 hours posttransplant, and antifungal coverage was continued for 5 days. At least 1 pancreas allograft Doppler ultrasonograph was performed before discharge. Dialysis needs were assessed daily during multidisciplinary rounds with the transplant surgeon and the nephrologist. Drain amylase was checked at the discretion of the transplant surgeon based on clinical suspicion for presence of a pancreatic leak. Drains were usually removed at the first outpatient follow-up visit if there were no suspicions of a pancreatic leak. After discharge, all patients were followed closely in our institution’s transitional care clinic. They were evaluated twice per week for 2 weeks and then weekly for 4 weeks. During these visits, drains and surgical sites were evaluated, need for dialysis was assessed, and laboratory values and immunosup-pression levels were evaluated.

For induction immunosuppression, alemtuzumab was used in all patients except for those with human immunodeficiency virus infection, in which case thymoglobulin was used. Patients on therapeutic immunosuppression received basiliximab induction. Triple therapy regimen consisting of tacrolimus (goal of 9-11 ng/mL), mycophenolic acid, and a steroid course tapered over 21 days was used as maintenance immunosuppression.

Statistical analyses
Clinical variables for patients in the morning, afternoon, and evening groups were compared. Statistics such as medians, interquartile ranges (IQR), and percentages were compiled for all measured variables. Binary data were analyzed using Fisher exact test. The Kruskal-Wallis test was used to compare continuous variables among the 3 groups and, if significant, was followed by the Wilcoxon rank sum test for paired comparisons. Statistical significance was determined at the 2-sided α = 0.05 level (P ≤ .05) for all tests. All statistical analyses were performed with R software (version 3.5.1).¹³

We identified 83 patients who underwent pancreas transplant during the study period. Median age at transplant was 45 years (IQR, 40-52 years); 45 patients (54.2%) were males and 46 patients (55.4%) were white. Type 1 diabetes mellitus was the indication for transplant in 66 patients (79.5%). Fifty transplant procedures (60%) were performed in the morning (43 SPK, 4 PAK, and 3 PTA), 25 (30%) were performed in the afternoon (17 SPK, 6 PAK, and 2 PTA), and 8 (10%) were performed in the evening (8 SPK). The 3 groups were similar in terms of comorbidities and demographics without any significant difference (Table 1).

Regarding the transplanted organs, 41 (49.4%) were procured from local OPOs. The median KDPI of the transplanted kidneys was 25% (IQR, 12%-40%). The median CIT for the pancreas allograft was 8.6 hours (IQR, 5.5-14.3 hours), and the median CIT for the kidney was 10.6 hours (IQR, 7.0-16.1 hours). Eight organs were procured from DCD donors. There were no statistically significant differences among the 3 groups in terms of organ origin, KDPI, CIT, or DCD status (Table 2).

With regard to the distribution of the cases, 68 patients (81.9%) underwent SPK transplant, 10 patients (12.0%) underwent PAK transplant, and 5 patients (6.0%) underwent PTA transplant. All transplants required systemic venous drainage except for 3 cases where portal drainage was performed. Prophylactic transabdominal drains were used in 30 patients (36%). The average intraoperative EBL was 522 mL; there was a trend toward a lower average of intraoperative EBL in the morning group, but it did not reach significance (486 mL vs 516 mL vs 769 mL; P = .07). No statistical difference was identified in the postoperative laboratory work and number of imaging modalities obtained among the 3 groups (Table 3).

In terms of outcomes, there was a significantly lower readmission rate in the group of patients who had surgery in the morning compared with results shown in the afternoon and evening groups (48% vs 84% vs 87.5%; P = .004). The most common indications for readmission among the study population were fever (21/53, 40%), intractable vomiting (19/53, 36%), hyperglycemia (6/53, 11%), and peripheral or nonocclusive venous thrombosis (3/53, 5%). The rate of DGF was zero in the afternoon group, which was statistically lower than the rate of DGF in the morning and the evening group, respectively (0.0% vs 25.6% vs 25.0%; P = .04). There were no differences among the 3 groups in terms of reoperation, graft loss, or placement of postoperative percutaneous drains (Table 4).

Despite the significant advances in pancreas transplant outcomes over previous decades, readmission still constitutes a major problem with significant health care costs.^1-4 Identifying possible risk factors for readmission after pancreas transplant, especially those that can be modifiable, can potentially ameliorate the burden of this problem and decrease health care spending. In this study, we investigated the impact of timing of surgery on readmission rates after pancreas transplant. During the 2 years of the study, our institution performed 83 pancreas transplant procedures. Most procedures (81.9%) were SPK transplants, and 60% of the cases were performed during the morning. Despite the similarities in comorbidities and donor characteristics, we demonstrated a significant reduction in read-mission rate in the patients who had their operations in the morning compared with rate of readmission shown in the afternoon and evening groups (48% vs 84% vs 87.5%; P = .004). However, there were no statistical differences in reoperation, graft loss, or postoperative drain placement among the 3 groups.

At our institution, we have a dedicated transplant operating room team and anesthesia team who are familiar with the procedure. Unfortunately, this team is only available during the morning shift, with the on-call nursing staff covering any transplants that occur over the remainder of the day. In addition, the afternoon and evening cases coincide with shift changes between the anesthesia and nursing teams, creating additional sign outs among different teams. Given the complex nature of pancreas transplant, full attention and familiarity with the case among every member in the operating room are required. Furthermore, the intraoperative management of the patient’s hemodynamics after pancreas reperfusion is very important; we rely on the anesthesia team to provide adequate resuscitation to the patient and maintain a mean arterial pressure >75 mm Hg. We believe that these shift changes and the lack of a dedicated team during the evening and at night can potentially affect patient outcomes, as seen with the increased readmission rate. Previous studies have evaluated the impact of team familiarity on surgical complications among operating room staff.^14-16 Most surgical teams are created ad hoc, comprising different members from day to day. These settings challenge the adaptive capacity of the surgical team,¹⁵ and lack of knowledge about one’s colleagues has been shown to increase the likelihood of miscommunication and interruptions during surgical procedures.¹⁶ The quality of the surgical team is thus rooted in not only their knowledge and technical skills but also in their communication and familiarity with other team members, most notably during unexpected surgical challenges.

Relational coordination, defined as communicating and relating for the purpose of task integration, has been used to describe the effect of a team’s familiarity on quality treatment and patient safety.¹⁷ Relational coordination is composed of shared goals and knowledge and respect for other team members that allow for frequent and accurate problem solving. It is a form of complex coordination that is expected to impact performance most significantly under conditions of task interdependence, uncertainty, and time constraints.¹⁴ Relational coordination has been associated with lower postoperative pain, higher postoperative function, decreased hospital-related infections, and shorter hospital lengths of stay.¹⁸

To minimize the gravity of this problem, one solution can be the careful selection of donor factors at the time of organ arrival. This can potentially decrease the number of transplants performed in the evening or at night. Having said that, this could affect ischemic times and further increase the discard rate of usable organs. The other option is to have a dedicated transplant operating room team that is available 24 hours, but this will increase health care expenses. All this should be weighed against increased health expenditures with readmissions.

There are several limitations of this study. First, this is a retrospective study that cannot draw a causal relationship. Second, the small number of patients transplanted in the afternoon and night could have potentially underestimated the differences in posttransplant complications among the different groups. Third, we grouped all patients who underwent pancreas transplant in this study, potentially increasing the heterogeneity of the cohort; however, in our institution, the pathways we use for intraoperative and postoperative care of these patients share many similarities. Fourth, we failed to factor the operative difficulty of each case; there was a trend for higher intraoperative blood loss in the afternoon and night cases, but this did not reach statistical significance. This could have potentially created a bias toward having easier cases in the morning. Additionally, this was a single-center study; therefore, our results may not be generalizable to other transplant centers.

Further prospective multicenter studies are needed to better identify the relationship between the time of day that surgery occurs and the readmission rate after pancreas transplant. Our findings suggest that patients undergoing pancreas transplant in the morning have a lower readmission rate compared with patients who receive transplants in the afternoon and at night. We believe that this is likely due to the presence of a dedicated operating nursing team in the morning shift.

Volume: 20 Issue: 1 January 2022

FULL TEXT