The impact of race on metabolic, graft, and patient outcomes after pancreas transplantation

Abstract

Background:

Racial disparities following pancreas transplantation (PTX) are poorly defined.

Methods:

This was a large-scale, single-center, longitudinal cohort study including adult PTX recipients. Patients were grouped by race to allow for comparisons.

Results:

287 PTX recipients were included; 125 (43.5%) were African American (AA). At baseline, AAs had a significantly higher proportion of T2DM (19.4% vs. 5.7%, p = 0.001), were younger, and more likely to be female. AAs experienced significantly higher rates of pancreatic leaks and post-operative bleeding. PTX rejection was comparable, however, kidney rejection tended to be higher among AA SPKs. Long-term mean HgbA1C levels were significantly higher among AAs (6.9% vs. 6.3%, p = 0.039). Patient and graft survival was comparable between groups, but early patient survival tended to be lower in AAs.

Conclusions:

This study demonstrated significant perioperative health disparities among AA PTX recipients, including poorer glycemic control and more early deaths, despite similar long-term patient and graft survival.

Introduction

Since pancreas transplantation was first introduced as a treatment modality in the late 1960s, significant advancements in immunosuppression, coupled with improved diagnostic testing and new interventions for infectious complications have ushered in a new era marked by dramatic improvements in perioperative outcomes.¹ However, racial disparities among pancreas transplant (PTX) recipients have likely persisted despite these improvements, precluding African-American (AA) PTX recipients from experiencing the same benefits in short and long-term outcomes.

Although the influence of race on health outcomes following kidney transplantation (KTX) has been widely documented over the past several decades, racial disparities following PTX are poorly defined.^2,3 Several small-scale studies have attempted to define the disparity; however, the evidence is largely conflicting. Luan et al. sought to address the inconsistency among the literature by analyzing national data provided by the Organ Procurement and Transplantation Network (OTPN)/Scientific Registry of Transplant Recipients (SRTR). In a cohort of 6,585 adult simultaneous kidney-pancreas (SPK) transplant recipients, AAs had significantly higher risk for late death-censored pancreas graft failure, as compared to white patients.⁴ These findings were congruent with the observations of the majority of race-based SPK studies. However, some small-scale studies have contested that recipient race influences perioperative outcomes, noting no significant differences between survival for AA and non-AA patients.^5–7 The aim of this study was to enhance our understanding of the impact of race on perioperative and long-term outcomes in PTX recipients at a major pancreas transplant program with a large volume of AA patients.

Methods

Study population

This was a large-scale, single-center, retrospective longitudinal cohort study including adult PTX conducted following institutional review board approval (Pro00083620) between 2001 and 2012 (n = 287). Patients less than 18 years of age and multi-organ transplant recipients were excluded. Patients were followed from time of transplant to death, loss to follow-up, or end of the study. The overall patient cohort was divided into two groups: AA (n = 125) and non-AA PTX (n = 162) recipients to allow for comparisons by race. Detailed electronic and manual chart abstraction was conducted to gain granular baseline and follow-up data for novel comparisons.

Recipient and donor characteristics

Recipient characteristics included PTX type (SPK or PTA), age (years), sex, ethnicity (AA or non-AA), weight (kg), body mass index, and prior transplant. Additional variables collected included history of diabetes, daily insulin requirement post-transplant (total U/day), C-peptide levels (ng/ml), HgbA1c (mg/dl), glucose (mg/dl), lipase (U/L), and weight change over follow-up (kg). Donor characteristics, including age and sex, pancreas transplant cold and warm ischemia time, and postoperative surgical complications were also recorded.

Complications after pancreas transplant

Patient’s electronic medical records were abstracted for peri-operative complications and outcomes including kidney or pancreas rejection, graft failure, postoperative insulin requirements, acute pancreatitis, reoperation, episodes of bleeding, wound infection and other surgical complications (bladder anastomosis, arterial thrombosis, venous thrombosis, leak, small bowel obstruction, abscess). Biopsy-proven acute rejection was defined as either using Banff criteria. Graft failure was defined as re-transplant or death.

Immunosuppression

Immunosuppression regimens were based on institution specific protocols which evolved during the study. Patients received induction therapy including either antithymocyte globulin (Thymoglobulin; Genzyme, Cambridge, MA, USA), alemtuzumab (Campath-1H; ILEX, San Antonio, TX, USA), interleukin 2 receptor antagonists (basiliximab) (Simulect; Novartis, Basel, Switzerland), or daclizumab (Zenapax; Roche, Basel, Switzerland) depending on their immunologic risk. Similarly, maintenance immunosuppression protocols changed throughout the study. Regimens consisted of a calcineuin inhibitor (Sandimmune; Sandoz, Holzkirchen, Germany), cyclosporine (Neoral; Novartis) or tacrolimus (Prograf; Astellas, Tokyo, Japan), plus an antiproliferative agent (azathioprine) (Imuran; GlaxoSmithKline, London, UK) or mycophenolic acid (CellCept; Roche), and low-dose corticosteroids. The current standard of care includes tacrolimus, mycophenolic acid, and prednisone.

Statistical analysis

For descriptive statistics, data are displayed as means ± standard deviation (SD) for continuous variables and proportions (%) for categorical data. For comparative inferential statistics, patients were divided into two groups, AA and non-AA, for the initial univariate analysis of baseline characteristics, post-transplant complications and outcomes. Chi-square or Fisher’s two-sided exact tests were conducted to make group comparisons for categorical data and odds-ratios were utilized for comparison between cohorts. Independent t-tests were performed for continuous data. Kaplan-Meier survival analysis was used to compared survival between the two cohorts.

Results

A total of 287 PTX recipients were included in this analysis (125 AAs, 162 non-AAs). Baseline characteristics of donors and recipients compared by race are displayed in Table 1. Compared to non-AA PTX recipients, AAs were younger at the time of transplant (40 vs. 43 years, p = 0.004) and less likely to be female (31.2% vs. 45.1%, p = 0.017). Baseline transplant characteristics revealed AAs had 2.8 higher odds of receiving a simultaneous kidney (p = 0.001) and experienced longer OR times than non-AA patients (327.5 ± 100.5 vs. 292.1 ± 91.3 min, p = 0.032). Donor age and sex was similar between cohorts. A significantly higher proportion of AAs had a history of T2DM (19.4% vs. 5.7%, p = 0.001). Average follow-up time was comparable across the cohort (5.1 vs. 6.3 years for AA and non-AAs, respectively).

Table 1.

Baseline characteristics by race.

Characteristic	Non-AA (n = 162)	AA (n = 125)	p-value
Female	45.1%	31.2%	.017
Age at transplant (yr)	42.9 ± 8.2	39.9 ± 9.0	.004
BMI	24.5 ± 3.1	26.1 ± 11.6	.119
CIT (mins)	742.5 ± 235.0	753.7 ± 222.7	.687
WIT (mins)	32.3 ± 10.6	33.3 ± 7.9	.398
OR time (mins)	292.1 ± 91.3	327.5 ± 100.5	.032
Donor age	22.1 ± 7.8	23.6 ± 8.8	.233
Female donor	28.6%	24.0%	.479
SPK	67.9%	85.6%	.001
T2DM	5.7%	19.4%	.001
Prior transplant	10.2%	8.9%	.759

Data are means ± SD unless indicated otherwise.

Abbreviations: AA, African American. CIT, cold ischemia time. SPK, simultaneous kidney-pancreas. T2DM, type 2 diabetes mellitus. WIT, warm ischemia time.

Peri-operative outcomes

Peri-operative outcomes compared between AAs and non-AAs is displayed in Table 2. Acute rejection of the pancreas was comparable between AAs and non-AAs (16.8% vs 14.2%, p = 0.544). Rejection of the kidney in SPKs tended to be higher in AAs (13.6% vs. 8.0%, p = 0.126). Similar rates of insulin use post-transplant and acute allograft pancreatitis (Amylase >300) observed across groups, however, data missingness was high (≥20%) across these outcomes. AAs experienced more perioperative complications requiring reoperation, although this difference was not significant (24.5% vs.16.9%, p = 0.143). AAs were 2.2 times more likely to experience pancreatic leaks and had significantly higher rates of post-operative bleeding (10.4% vs. 3.1%, p = 0.012), as compared to non-AAs. Other peri-operative complications, including thrombosis and wound infection, were similar between groups (Table 2).

Table 2.

Clinical outcomes by race.

Clinical Outcomes	Non-AA	AA	p-value
	(n = 162)	(n = 125)
Surgical Complications
Kidney rejection	8.0%	13.6%	.126
Pancreas rejection	14.2%	16.8%	.544
Bladder anastomosis	0.6%	0.8%	.849
Insulin use post-tx	27.3%	29.1%	.783
Acute allograft pancreatitis	7.8%	8.9%	.790
Return for ex-lap	16.9%	24.5%	.143
Return for bleed	6.3%	8.8%	.464
Thrombosis			.197
Arterial	3.7%	0.8%
Venous	8.0%	10.4%
Leak	8.1%	16.0%	.037
SBO	5.0%	4.8%	.948
Abscess	4.3%	9.6%	.077
Bleeding	3.1%	10.4%	.012
Wound infection	2.5%	3.2%	.716
Other	4.3%	14.4%	.003
Laboratory values and follow-up parameters
C-peptide (ng/ml)	5.1 ± 3.4	4.6 ± 3.6	.439
Hgb A1c (mg/dl)	6.3 ± 1.5	6.9 ± 2.3	.039
Glucose (mg/dl)	108.6 ± 33.4	117.1 ± 45.0	.163
Lipase (U/L)	40.3 ± 30.0	49.9 ± 53.3	.160
Weight change overtime (kg)	12.7 ± 9.2	13.8 ± 8.8	.391
Weight (kg)	71.9 ± 13.4	73.1 ± 13.1	.516
Follow-up (yr)	6.3 ± 10.4	5.1 ± 4.2	.183

Data are means ± SD unless indicated otherwise.

Abbreviations: AA, African American. Hgb A1c, hemoglobin A1c. SBO, small bowel obstruction. Tx, transplant.

Laboratory values

Laboratory values and follow-up parameters are outlined at the bottom of Table 2. C-peptide, glucose, and lipase levels were comparable across both groups. Long-term mean hemoglobin A1C levels were significantly higher among AAs (6.9% vs. 6.3%, p = 0.039); however, 41.4% of non-AAs and 39.2% of AAs did not have these measures recorded after transplant. During follow-up, there was no difference observed in weight or weight change over time between groups (Table 2). Of note, data missingness was fairly high across all laboratory values. Further detail on the degree of data missingness can be found in the supplemental material (Supplemental Table 1).

Patient and graft survival

Fig. 1 displays two Kaplan-Meier analyses comparing pancreas graft and patient survival between AAs and non-AAs. Graft survival was similar across AAs and non-AAs, although slightly higher in non-AAs (top Fig. 1; p = 0.515). Patient survival was also comparable between groups; however, early post-transplant patient survival tended to be lower in AA recipients (bottom Fig. 1; p = 0.256).

Fig. 1.

Pancreas graft and patient survival after transplantation.

Discussion

This large-scale, single-center experience demonstrated significant disparities in perioperative health outcomes and glycemic control in the AA PTX population as compared to their non-AA counterparts. To our knowledge, this is the largest single-center experience with 44% of patients being AA. AA PTX tended to experience reduced early post-transplant patient survival, despite comparable long-term patient and graft survival between groups.

Previous research on racial disparities in the PTX is scarce and findings are inconsistent. One small, single-center study by Rogers et al. retrospectively reviewed 96 consecutive SPK recipients. Results from the study demonstrated significantly worse 1-, 3-, and 5-year pancreas graft survival in AA SPKs compared to Caucasians, despite similar rates of long-term kidney graft survival. Higher rates of early death with a functioning graft, early graft loss due to acute rejection, and late chronic rejection contributed to suboptimal pancreas graft outcomes in AA SPKs. The authors also reported reduced metabolic control in AA recipients. Similar to other studies, the authors speculated that the higher incidence of resistance to contemporary immunosuppression agents may also contribute to inferior graft outcomes in the AA population.^4,8,9 In contrast to previous studies and our findings, Luan et al. performed a large-scale analysis of national data from OTPN/SRTR and found no racial disparities in early post-transplant SPK outcomes. However, similar to Rogers et al. investigators observed significantly worse long-term outcomes in AA SPKs. Like previous investigators, Luan et al. cited enhanced immunologic reactivity and lower socioeconomic status as potential contributors to poorer health outcomes in AA SPKs.⁴

Other studies have found no differences in perioperative outcomes between AA and non-AA PTX recipients. One multicenter, prospective cohort study of 297 SPKs, found no differences in 1-year pancreas graft survival between AA and non-AA patients. Furthermore, rates of rejection and incidence of adverse events were comparable across groups. Of note, observed mean HgbA1C levels were slightly higher in AAs 1-year post-transplant, but this difference was not significant.⁵ Another large review found no difference in long-term patient or graft survival between AA and non-AA SPKs. Although, this report was limited by the inconsistency of immunosuppressive regimens of the patients included.¹⁰ Young et al. recently published a report of a 15-year, single-center experience regarding the influence of ethnicity on SPK outcomes in T1DM patients, demonstrating similar outcomes among AA and non-AA SPK recipients at a center overrepresented by AAs. These results contrasted those observed in a contemporaneous national cohort and those reported by Luan et al.¹

Overall, findings from studies investigating the impact of race on outcomes following PTX remain largely incongruent. However, our observations support previous research suggesting AA PTXs experience worse early post-transplant outcomes, including higher rates of rejection and inferior long-term graft survival. Additionally, our findings suggest history of T2DM and post-transplant glycemic control may play an important role in peri-operative outcomes following PTX. We observed higher rates of peri-operative complications in a cohort of AA PTXs with a higher proportion of T2DM compared to a group of non-AA patients. Of note, in the particularly large number of T2DM and male patients included in this study, increased surgical complications may be related to increased BMI and body habitus. Previous literature has associated T2DM diagnosis with an increased risk for inferior peri-operative outcomes. In a study cohort of predominantly T2DM patients undergoing either orthopedic or general surgery, Wang et al. demonstrated 7.7% of patients experienced post-operative adverse events, including healing abnormalities and infection.¹¹ Peri-operative complications contribute to longer hospital lengths of stay, higher economic burden, higher health care resource utilization, and increased risk of morbidity and mortality as compared to nondiabetic patients.¹⁴ Tighter glycemic control has been shown to improve peri-operative outcomes, including stroke and death.^12,13 Transplant clinicians should consider recipient history of diabetes and the potential impact on peri-operative outcomes when determining plans for glycemic control. Additionally, peri-operative HgbA1C and glucose levels should be closely monitored.

In contrast to PTX, the influence of diabetes on racial disparities in the KTX population are well-defined. Taber et al. demonstrated that AA KTX recipients were more likely to have preexisting DM and experienced poorer graft outcomes. Graft survival was found to be significantly lower in AA patients and AA with DM were considerably less likely to achieve tight postoperative glycemic control, as compared to white KTX recipients. In both univariate and multivariate analyses, DM was a significant contributor to graft failure. However, this relationship appeared to be attenuated with tighter glycemic control.² The resolution of DM significantly contributes to diminishing racial disparities and improving long-term KTX outcome. Interestingly, our findings may support this pattern in the PTX population as well.

There are several limitations in our study. First, the retrospective nature of this study precluded the inclusion of all potentially confounding variables, such as socioeconomic status or compliance with medical care. It is possible that variables we were unable to capture may have influenced our results. Second, several data were missing from various outcomes of interest, which may have impacted the results of our analyses. While we reported a high degree of missingness for HgbA1c, one of our highlighted outcome disparities, the missingness was fairly equally distributed between groups. Additionally, we were unable to adjust for diabetes type or pancreas transplant type. It was not feasible to conduct multivariable analysis due to small relative sample size and event rates. Lastly, complications treated at an outside hospital were not captured unless they were recorded in the patient’s medical record. Therefore, the incidence of long-term issues may be underreported in our analysis.

Conclusion

In conclusion, this study provided evidence to support an ongoing disparity in AA PTX recipients for peri-operative outcomes, with similar rates of long-term patient and graft survival. These peri-operative issues may be related to higher rates of T2DM and reduced glycemic control. Immunologic risk leading to higher rates of rejection do not appear to be a significant problem in our cohort of AA PTX recipients. Future studies are warranted to investigate the underlying mechanisms of this disparity, perhaps related to baseline diabetes control, as well as to provide insight on potential interventions to reduce peri-operative disparities in AA PTX recipients.