Reduction in new-onset diabetes mellitus after renal transplant with erythropoietin-stimulating agents: a retrospective cohort study

Background Studies have shown that erythropoietin-stimulating agents (ESAs) protect mice against the development of diabetes through direct effects on pancreatic ß cells. However, the effect of ESAs on the incidence of diabetes in humans has not been well studied. It is unknown whether exposure to ESAs is associated with a reduced incidence of new-onset diabetes after transplant (NODAT). Objective The objective of this study is to examine the relationship between ESA exposure post-renal transplant and the development of NODAT. Design We performed a single center, retrospective cohort analysis. Patients We compared patients who received a first live or deceased donor renal allograft, with any exposure to an ESA vs. those without such exposure and who developed NODAT and who did not. Patients with a prior history of diabetes mellitus or multi-organ transplant, including a second renal transplant were excluded. Measurements and methods NODAT was defined based on the 2008 Canadian Diabetes Association criteria. Multivariate logistic regression analysis was performed to determine factors independently associated with NODAT. Results One hundred thirty-two (29 %) patients were exposed to an ESA, four of which developed NODAT compared to 128 who did not develop NODAT (p < 0.0001). Of those not exposed to an ESA, 15 % (48/319) developed NODAT. By Fisher’s exact test, exposure to an ESA at any time post-transplant reduced the risk of developing NODAT; odds ratio (OR) = 0.08, 95 % confidence interval (CI) (0.018–0.352), p = 0.0008. Older age; OR = 1.41, 95 % CI (1.036–1.933), p < 0.02, higher random blood sugar at discharge; OR = 1.30, 95 % CI (1.077–1.57), p < 0.006 and deceased donor; OR 2.18 CI (1.009–4.729), p = 0.04 were associated with an increased risk of NODAT. Limitations The limitations of this study include its retrospective nature, single center, and homogenous population; thus, generalizability of the results must be approached with caution. Conclusion ESA exposure may be associated with a reduced incidence of NODAT in the post-renal transplant population. The role of ESA in preventing NODAT requires further investigation.


What was known before
We know in animal studies that erythropoietinstimulating agents (ESAs) protect against the development of diabetes through direct effects on pancreatic ß cells and have been associated with improvement in glucose tolerance. Furthermore, patients with end-stage renal disease on hemodialysis treated with ESAs have been shown to have improvement in insulin sensitivity. However, the effect of ESA on the risk of diabetes in post-transplant patients has not been well studied.

Introduction
New-onset diabetes mellitus after renal transplantation (NODAT), a common post-transplant event, adversely affects graft and patient survival and quality of life [1,2]. Evidence suggests that transplant recipients who develop NODAT are at greater risk of graft loss, rejection, and infection as well as diabetes-associated microvascular and macrovascular complications [3]. About one-third of nondiabetic renal transplant recipients develop persistently impaired glucose metabolism by 6 months posttransplantation [4][5][6]. However, the prevalence of NODAT at our institution is approximately 11.5 %. Risk factors such as advanced age, obesity, male gender, nonwhite ethnicity, a family history of diabetes, and the prescription of corticosteroids and calcineurin-inhibitors are well described [7,8].
The prevalence of anemia post-transplant is dependent on the time of observation after transplantation [9,10]. Although highly prevalent in the pre-transplant population with chronic kidney disease (CKD), anemia is a common event following renal transplantation as well. Studies report variable prevalence of 20 to 53 % at 1-year post transplant [11,12]. Risk factors for post-transplant anemia include iron deficiency, graft rejection or dysfunction, erythropoietin (EPO) deficiency, viral infection, immunosuppression, infection prophylaxis medications, use of angiotensin converting enzyme inhibitors/angiotensin receptor blockers, low serum albumin, and older kidney donors [11,12]. However, anemia in the immediate posttransplant period is almost always due to lower than normal hemoglobin targets in CKD and end-stage renal disease population on an ESA, surgical blood loss, and frequent phlebotomy [11]. Approximately, 10-15 % of post-renal transplant patients require an erythropoietinstimulating agent (ESA) for the treatment of posttransplant anemia.
Erythropoietin (EPO), a hematopoietic growth factor, is largely produced in the kidney and has several biological functions, including the regulation of glucose metabolism [13]. Recombinant EPO forms or ESAs have been used to treat anemia in chronic kidney disease and other conditions [13,14]. Patients with end-stage renal disease on hemodialysis treated with recombinant EPO have been shown to have improvement in insulin sensitivity [14][15][16][17]. ESA has been associated with improvement in glucose tolerance, both at the peripheral and skeletal tissue in mice models [13,16,18]. Furthermore, inhibition of gluconeogenesis and inflammation in the liver were observed with ESA in high-fat diet-fed mice [13]. Choi et al. and others demonstrated that EPO has a direct effect on pancreatic beta cells by increasing β-cell mass through anti-apoptotic, proliferative, and angiogenic mechanisms in mice models with type 1 and type 2 diabetes [14,15,19]. In Choi's study, EPO was given three times weekly for 4 weeks compared to a saline control group. In both mice models, EPO lowered fasting blood glucose and slowed the loss of β-cell mass. In addition, by binding to erythropoietin receptors found in pancreatic cells, EPO significantly inhibited apoptosis induced by cytokines in vitro [20]. These studies suggest that EPO may reduce or prevent diabetes as it has protective effects on the beta cells as well as reducing insulin resistance.
A recent study by Shah et al., found no association with endogenous erythropoietin and glucose homeostasis in post-partum women with varying degrees of glucose intolerance [21]. However, the effect of exogenous ESA on the risk of diabetes in humans has not been well studied. It is unknown whether exposure to ESAs is associated with a reduced incidence of NODAT, or if ESAs play a role in preventing the development of NODAT. We therefore sought to examine the relationship between ESA exposure post-renal transplant and the development of NODAT.

Methods
St. Michael's Hospital is a university-affiliated tertiary facility that provides post-transplant care to approximately 1600 adult renal transplant recipients. As shown in Fig. 1, we identified a cohort of all de novo renal transplant recipients who received a living or deceased donor allograft between the period of January 1, 2005 and December 31, 2010 and had a minimum of 6-month post-transplant follow-up. Data was obtained from an electronic renal transplant database. Data was collected retrospectively for demographic and baseline attributes and to identify recipients who were prescribed an ESA. ESA exposure measurement included time of the first dose of ESA post-transplant within < 3, 3-6, and >6 months from transplant, as well as total duration of treatment post-transplant <3, 3-6, and >6 months. NODAT was identified through routine post-transplant blood glucose screening and defined by a fasting blood glucose (FBG) ≥7.0 mmol/L and/or random blood glucose (RBG) ≥11.1 mmol/L on two separate occasions in the absence of acute illness as per the 2008 Canadian Diabetes Association diagnostic criteria [22]. Of note, there were no female participants who became pregnant and developed gestational diabetes. Univariate comparisons for transplant demographic/baseline characteristics and outcome variables between ESA exposure or no ESA exposure and those who developed NODAT or did not were done by chi-square analysis or Fisher's exact test as appropriate. Multivariate logistic regression models were then constructed using ESA, age, donor source, random blood glucose (RBG) at discharge, acute rejection (AR), and body mass index (BMI) as independent variables and univariate factors with p < 0.20 as other independent variables, and NODAT as the dependent variable. Immunosuppressive profile was based on the initial hospital discharge summary. Patients with any prior history of diabetes or glucose intolerance or had multi-organ transplant including a second renal transplant or more were excluded. Institutional Review Board approval was obtained (Research ethics board #10-313). SAS was the statistical software used for the analysis.

Results
Six hundred fifteen recipients met initial criteria of which 164 were excluded. Of those excluded, 147 had a previous history of diabetes mellitus prior to transplant, two had impaired glucose tolerance, one patient had follow-up less than 6 months post-transplant, four patients had a second renal transplant, one patient had a failed renal transplant, and nine patients had NODAT prior to ESA initiation. In the remaining 451 recipients who were transplanted between January 1, 2005 and December 31, 2010, 132 were exposed to an ESA at any time post-transplant and 319 were not exposed. Of those exposed to an ESA, there was a significant reduction in the development of NODAT with four developing NODAT compared to 128 who did not develop NODAT There are two ESAs currently available in Canada, namely epoetin alfa (Eprex) and darbepoetin (Aranesp). Of the 132 recipients exposed to an ESA, 75 % received darbepoetin (Aranesp) and 25 % received epoetin alfa (Eprex).

NODAT vs. no NODAT
Baseline demographics at the time of transplant (N = 451) of older age (p = 0.0001), hypertension as a cause of end-stage renal disease (p = 0.003), prednisone (p = 0.02), and ASA (p < 0.0001) were statistically significant between the NODAT and no NODAT groups, where the risk of NODAT was higher (see Table 1-not all data is shown). In the 399 patients who had no NODAT, immunosuppression consisted of tacrolimus (311), tacrolimus extended release (73), cyclosporine (55), MMF (299), Myfortic (53), prednisone (345), azathioprine [16], and Rapamune [23]. Prednisone dose was not significant at discharge (31.6 mg ± 12 with NODAT and 29.4 mg ±16.8 with NoNODAT, p = 0.26), at 3 months (7.3 mg ± 5 with NODAT and 7.1 mg ± 7.1 with no NODAT, p = 0.88) and 6 months (6.3 mg ± 4.4 with NODAT and 5.5 mg ± 2.7 with no NODAT, p = 0.10) post transplant between the two groups. Two patients were on a prednisone avoidance immunosuppressive regimen, neither of which was on an ESA. One patient was not given prednisone due to the risk of GI bleed and the other was due to the strong family history of type 2 diabetes. Despite not being on prednisone, the patients still got NODAT.  Table 1.
We examined NODAT rates 6 months after transplant and compared those exposed to an ESA vs. those who were not exposed to an ESA by chi-square analysis. Of the 132 patients who were exposed to an ESA, two developed NODAT. Of the 319 patients who were not exposed to an ESA, 15 developed NODAT. However, this was not statistically significant (p = 0.08). Therefore, early exposure to ESA does not seem to protect against NODAT at 6 months post transplant.
As shown in Table 2 There were 29 patients from the original cohort that were no longer followed in our clinic by the end of the study period. Twelve patients were deceased, 14 returned to dialysis due to graft failure, and 3 were discharged from our clinic/transferred to another facility for their post-transplant care. Of the 29 patients who were no longer followed, 22 were exposed to an ESA, 19 of which did not have NODAT.

Discussion
Exposure to an ESA at any time post transplant, particularly if given for up to 6 months duration, was associated with a reduced risk of developing NODAT compared to those who were not exposed. The mechanism of this effect is unclear. It is also unclear whether exposure to ESA for greater than 6-month duration impacts the risk of NODAT. In our study and in other analyses, the highest risk of NODAT was within the first 12 months posttransplant.
ESAs have decreased the need for blood transfusions and improved the quality of life in patients with CKD and anemia compared to no treatment; however, the risks and benefits of ESA therapy as well as cost effectiveness is unclear in renal transplant recipients due to the limited evidence available [24][25][26]. The Correction of Anemia and Progression of Renal Insufficiency in Transplanted (CAPRIT) patients study, a European, multicenter, randomized trial of renal transplant recipients assessed the effect of normalization of anemia (130-150 g/L) compared to partial correction (105-115 g/L) with ESA on the progression of graft dysfunction. The CAPRIT authors found after 2 years of follow-up that, normalization of hemoglobin was associated with improved quality of life, and less progression of chronic allograft nephropathy without an increase risk of cardiovascular events [23]. However, in a retrospective study, an increase in mortality was seen in renal transplant recipients who had hemoglobin levels lower or higher than 125 g/L on an ESA (U-shaped curve) compared to those not on an ESA, where the relationship between hemoglobin and mortality was linear [26]. Studies targeting normal or higher hemoglobin levels in patients with CKD showed an increased risk of adverse outcomes including cerebrovascular events, vascular access thrombosis and hypertension, and an increase in cost compared to a lower hemoglobin target [25,27,28]. It has also been postulated that higher ESA doses may increase the risk of mortality; however, this has yet to be confirmed [29]. Therefore, weighing potential benefits of ESA in reducing the risk of NODAT must be weighed against the potential adverse effects particularly when the hemoglobin is near normal and at higher doses of ESA.
Limitations of this study include its retrospective nature, single center, and homogenous population; thus, generalizability of the results must be approached with caution. Although several confounders are included in the multivariate regression model, we did not include other potentially important confounders such as gender, ethnicity, hepatitis C, and steroid use in the analysis and therefore residual confounding may exist. In addition, immunosuppression may have been altered by the prescriber in response to blood sugar levels to minimize risk of NODAT, thus leading to selection bias.

Conclusions
The incidence of NODAT and its effects in renal transplant recipients is significant, including greater risk of graft loss and reduced patient survival. In our study, we found that older age, higher RBG at discharge, and deceased donor were associated with an increased risk of NODAT; however, even accounting for these factors, ESA exposure remained protective. Reducing the risk of NODAT may improve the overall health and quality of life of patients and their renal transplant. Being on an ESA may have a role in preventing NODAT, particularly if given for up to 6 months duration post-transplant,