Simple Measures to Monitor b -Cell Mass and Assess Islet Graft Dysfunction

Transcription

1 American Journal of Transplantation 2007; 7: Blackwell Munksgaard C 2006 The Authors Journal compilation C 2006 The American Society of Transplantation and the American Society of Transplant Surgeons doi: /j x Simple Measures to Monitor b -Cell Mass and Assess Islet Graft Dysfunction R. N. Faradji a,b, K. Monroy a, S. Messinger a,c, A. Pileggi a,d, T. Froud a,d, D. A. Baidal a, P. E. Cure a, C. Ricordi, a,d L. Luzi a,e and R. Alejandro a,b, a Clinical Islet Transplant Program at the Diabetes Research Institute, University of Miami Leonard M. Miller School of Medicine, Miami, United States Departments of b Medicine, c Epidemiology and d Surgery, Univ. Miami Leonard M. Miller School of Medicine, Miami, Florida e Diabetes Research Institute Milan, Istituto Scientifico H. San Raffaele, University of Milan, Milan, Italy Corresponding author: Rodolfo Alejandro, ralejand@med.miami.edu The aim of this study was to develop a simple test for the assessment of islet graft dysfunction based on measures involving fasting C-peptide. Calculations were made to account for the dependence of C-peptide secretion on glucose concentration (Cpeptide/glucose ratio [CP/G]) and adjusted for renal function by calculating the C-peptide/glucosecreatinine ratio (CP/GCr). Values from 22 recipients were analyzed at different times post-last islet infusion. Receiver operating characteristic curves were used to determine which of these measures best predicts high 90-minute glucose (90 min-glc; >10 mmol/l) after a Mixed Meal Tolerance Test (MMTT). In this initial analysis, CP/G was found to be superior predicting high 90 min-glc with a larger area under the ROC curve than C-peptide (p = 0.01) and CP/GCr (p = 0.06). We then correlated C-peptide and CP/G with islet equivalents-ieq/kg infused, 90 min-glc after MMTT and clinical outcome (b -score). C-peptide and CP/G in the first 3 months post-last islet infusion correlated with IEQ/kg infused. CP/G correlated with 90 min-glc and b -score. C-peptide and CP/G are good indicators of islet mass transplanted. CP/G is more indicative of graft dysfunction and clinical outcome than C-peptide alone. The ease of calculation and the good correlation with other tests makes this ratio a practical tool when monitoring and managing islet transplant recipients. Key words: b-cell mass, C-peptide, graft dysfunction, insulin independence, islet transplantation Received 26 May 2006, revised 18 September 2006 and accepted for publication 27 September 2006 Introduction There is no consensus on which test is best to monitor b-cell mass and function in patients with diabetes and after islet transplantation (1). Most tests give variable results, are time consuming and difficult to perform. Therefore, there is an increasing need for simple metabolic tests for the assessment of islet graft function over time. The relatively low variability, high reproducibility and close relationship of C-peptide measurements in the systemic circulation to endogenously secreted insulin in the portal system makes this a suitable assay for monitoring b-cell function (2). C-peptide is currently performed before and after islet transplantation to document islet graft survival (3). b-cell response to meals is evaluated by the Mixed Meal Tolerance Test (MMTT) (3). This test is being widely used to study the response following therapeutic interventions in new onset T1DM (4) and to monitor the function of transplanted islets (3,5). Evaluation of the 90-min glucose (90 min-glc) after MMTT is considered a good acute measure of b-cell reserve following islet transplantation (3,5), as well as the homeostatic ability of the individual to handle a secretagogue load. Levels of 10.0 mmol/l (180 mg/dl) indicate good graft function (6). Plasma C-peptide levels depend on glycemic values. Any given plasma C-peptide value may indicate good graft function if glucose levels are normal, but may be inappropriately low if glucose levels are high. Calculations involving C-peptide measurements accounting for glucose concentrations in type 2 diabetes, as well as islet and pancreas transplant models have been reported for the last three decades in urine (7) and recently in serum (8 10). One of them, the HOMA C-peptide b-cell index-2, requires the use of a specific computer program for its assessment and has not been validated in islet transplant recipients or in patients using exogenous insulin therapy (10). We propose the C-peptide to glucose ratio (CP/G), which corrects for glycemic values. Since C-peptide is cleared through the kidney, we also present a formula that accounts for renal function: the ratio of C-peptide to the product of glucose and serum creatinine (CP/GCr) (11). Our objective was to assess and compare each measure involving C-peptide (C-peptide alone, CP/G and CP/GCr) in 303

2 Faradji et al. order to determine which best serves as a simple test for predicting high levels (>10 mmol/l) of 90 min-glc after MMTT (graft dysfunction) in islet transplant recipients. We then looked for the association between these measures and islet mass transplanted, 90 min-glc after MMTT, and clinical outcome as assessed by b-score (6). We present data showing that CP/G is the best predictor for high 90 min-glc and can identify graft dysfunction. Since this ratio is easy to perform on basal peripheral blood samples we propose its use as a practical tool for the monitoring of islet graft function. Methods Islet transplantation We performed a retrospective analysis on 22 T1DM recipients of allogeneic islet grafts. Patients were enrolled in three different islet transplantation protocols: 14 belonged to the Islet Transplantation Alone (IA) (12), four to the Islet After Kidney (IAK, stable renal grafts for 6 months prior to islet transplantation) and four to islet transplantation with concomitant administration of donor-derived, CD34 + -enriched, bone marrow cells (BMC) (13). The studies were approved by the Institutional Review Board. Islets were isolated from human pancreata obtained from deceased, multiorgan donors utilizing a modification of the automated method (14) followed by density gradient purification (15). Islets were transplanted as previously described (12,13). Steroid-free immunosuppressive regimen consisting of induction with daclizumab (Zenapax, Roche Nutley, NJ) and maintenance with tacrolimus (Prograf, Fujisawa Deerfield, IL) and sirolimus (Rapamune, Wyeth-Ayerst Madison, NJ), as described (12,13,16). In the IA + BMC recipients, immunosuppression was discontinued 1 year after transplantation. Metabolic assays Serum glucose concentrations were determined by the hexokinase method. Plasma C-peptide was measured by double antibody radioimmunoassay (detection limits: ng/ml, inter- and intra-assay variation coefficient <10%, cross-reactivity with insulin and proinsulin: 20%). Serum creatinine reference range was mg/dl with detection limits of mg/dl. Serum insulin was measured by solid phase I 125 radio-immunoassay in a gamma counter. MMTT was performed after 8 to 12 h of fasting. Glucose and C-peptide were measured before (basal) and 90 min after ingesting 360 ml BOOST High-protein (Novartis; 360 calories, 9 g fat, 49.5 g carbohydrate, 22.5 g protein). This test was performed pre-transplant, and every 3 months up to 18 months post-last islet infusion. Intravenous Glucose Tolerance Test (IVGTT) (17) was performed posttransplant at the same time points as MMTT, in 18 subjects (IA and IAK recipients). The acute insulin response to glucose (AIRglu) was obtained from mean insulin values (3, 4 and 5 min) post-glucose infusion minus the basal value (17). Correction of C-peptide for glucose and renal function was performed using formula 2: C peptide/glucose creatinine ratio C peptide (ng/ml) 100 = Glucose (mg/dl) Creatinine (mg/dl). All calculations were performed using measures for C-peptide, glucose and serum creatinine obtained from the same fasting blood sample for each time point. Experimental design Our objective was to compare each measure involving C-peptide (C-peptide alone, CP/G and CP/GCr) in order to determine which best serves as a simple test for predicting high 90 min-glc (>10mmol/L) after MMTT in islet transplant recipients. Additionally, we examined the relationship between these measures with islet mass (total IEQ/kg infused, at 7 days, 1 and 3 months post-transplant), and clinical outcome given by the b-score (6). This score ranges from 0 (no graft function) to 8 (good clinical outcome) (6). Time points considered for the metabolic and clinical outcomes included pre-transplant and 3, 6, 9, 12, 15 and 18 months post-last islet infusion. Definitions Insulin independence: C-peptide-positive recipients that were able to maintain HbA1c 6.5%, without exogenous insulin administration, and their capillary fasting and 2-h post-prandial blood glucose levels were 7.8 mmol/l (140 mg/dl) and 10mmol/L (180 mg/dl), respectively. Graft dysfunction: C-peptide-positive recipients that presented with fasting capillary glucose >7.8 mmol/l and/or 2-h post-prandial capillary glucose >10.0 mmol/l in three or more occasions in 1 week, requiring the reintroduction of exogenous insulin therapy, confirmed by a 90-min-Glc >10.0 mmol/l after MMTT. Statistical analysis Data from this study are clustered: each patient has measures taken at baseline and at every follow-up time point. Analyses for clustered data are used for all comparisons where appropriate. We compared the diagnostic accuracy of the measures (C-peptide alone, CP/G and CP/GCr) using Receiver operating characteristic (ROC) curves and test statistics presented by Obuchowski (18) for the comparison of correlated ROC curves arising from clustered data. ROC curve methodology provides a good measure of test accuracy, incorporating both sensitivity and specificity and accounting for the inherent trade-offs between them in decision making. Multiple logistic regression for repeated measures with generalized estimating equations (GEE) were used to estimate the associations of CP/G to 90 min-glc, b-score and insulin use. Linear regression methods for repeated measures were used to assess the relationship between C-peptide measures, 90 min-glc, AIRglu and b-score. Simple linear regression was used to assess the relationship between C-peptide measures and islet mass at 7 days, 1 and 3 months post-transplant. Two-sample t-tests were used to compare measures involving C-peptide, 90 min-glc and b-score between insulin-independent and insulin-dependent patients at 15 months post-transplant. P-values <0.05 were considered statistically significant. Calculations using C-peptide Correction of C-peptide (ng/ml) for glucose levels was performed using formula 1: C peptide/glucose ratio = C peptide (ng/ml) 100. Glucose (mg/dl) Results Demographics Twenty-two patients were studied (11 males, 11 females), mean age 43.4 ± 7.9 years, mean diabetes duration American Journal of Transplantation 2007; 7:

3 Monitoring b -Cell Mass and Function Table 1: Serum creatinine values (median and mean) for all patients at different time points Median (mg/dl) Mean Count (Range) (mg/dl) ± SD Pre-transplant ( ) 0.96 ± months ( ) 0.97 ± months ( ) 0.99 ± months ( ) 0.95 ± months ( ) 0.96 ± months ( ) 0.97 ± months ( ) 0.93 ± 0.25 ± 11.8 years. Six patients received only one islet allograft, while the remaining 16 were recipients of 2 sequential islet infusions. Study subjects received a mean of 7,808 ± 2,315 IEQ/kg with the first infusion and those who had two infusions received a total of 13,815 ± 2,784 IEQ/kg. All patients were on tacrolimus and sirolimus maintenance therapy. No patient received oral hypoglycemic agents or insulin sensitizers for the duration of the study. All patients had stable kidney function (Table 1) and none had clinical evidence of gastroparesis. Comparison of C-peptide alone, CP/G and CP/GCr as indicators of islet dysfunction assessed by 90 min-glc Measures involving C-peptide represent good diagnostic indicators of islet dysfunction assessed by 90 min-glc, as illustrated by their elevated areas under their respective ROC curves (Figure 1). For C-peptide alone, CP/G and CP/GCr, these values were 0.863, and respectively. CP/G was superior to C-peptide alone, illustrated by a larger area under the ROC curve (p = 0.01) (Figure 1). The area under the CP/G ROC curve was greater than the CP/GCr curve (p = 0.06). Comparing the ROC curves of CP/GCr versus C-peptide alone suggests that CP/GCr yields no significant improvement over C-peptide alone (p = 0.25). These results suggest that CP/G performs as the best test predicting high 90 min-glc levels for this population.once we determined that CP/G is the best test to predict high 90 min-glc and since all of our patients had normal and stable kidney function (Table 1), further analysis was concentrated on C-peptide alone and CP/G ratio. Relationship between C-peptide and CP/G to 90 min-glc No significant correlation existed between 90 min-glc and C-peptide at early times post-transplant (3, 6, 9, 12 and 15 months). We observed a significant correlation at 18 months post-transplant possibly due to graft failure in IA + BMC recipients, after discontinuing immunosuppression at 12 months. The CP/G had a significant negative correlation with 90 min-glc at all time points (Table 2). Relationships using data from all patients at all time points (n = 140 values) comparing 90 min-glc to C-peptide and CP/G (Figure 2) showed significant correlation for all values (r = 0.65, p < 0.001; and r = 0.762, p < 0.001, respectively). Multiple logistic regression for repeated measures, considering all time points, revealed that CP/G <1.0 is associated with having higher 90 min-glc levels (>10 mmol/l, OR = 14.02, p < ). Having a higher b-score (>6) carried a positive association with having higher CP/G (>1.0) (OR = 14.43, p < 0.001). The odds of having lower CP/G (<1.0) are about 4 times greater with exogenous insulin use (OR = 4.65, p < 0.001). The results described above for the association between CP/G and 90 min-glc were adjusted for insulin use and b-score. Figure 1: ROC curve to determine which measure best predicts a 90 min glucose >10 mmol/l. CP/G was found to be a superior test associated with a larger area under the ROC curve compared to C-peptide alone (p = 0.01) and CP/GCr (p = 0.06). American Journal of Transplantation 2007; 7:

4 Faradji et al. Figure 2: Correlation between 90min- glu after MMTT and C-peptide/glucose ratio in 140 tests performed on 22 subjects before and after islet transplantation. Relationship between CP/G, 90 min-glc and AIRglu Relationships using data from the IA and IAK recipients (n = 18) at all time points (n = 83 values) comparing AIRglu to 90 min-glc and CP/G showed a significant correlation (r = 0.58, p < 0.001; and r = 0.28, p < 0.01; respectively). C-peptide alone did not correlate with the AIRglu (r = 0.171, p = 0.12). Relationship between infused islet mass and C-peptide and CP/G C-peptide and CP/G had a positive correlation with total IEQ infused at 7 days (r = 0.551, p < 0.01; r = 0.502, p = 0.02, respectively), 1 month (r = 0.599, p < 0.01; r = 0.574, p < 0.01, respectively) and 3 months (r = 0.459, p = 0.02; r = 0.55, p < 0.01, respectively) post-last islet infusion. An increase in 1000 IEQ/kg infused was associated with a 0.08 units in the CP/G (p < 0.01) at 3 months post-last islet infusion. Thus, infusing a total of 14,000 IEQ/kg is associated with an average CP/G of 1.12 at 3 months post-last islet infusion in subjects treated with this immunosuppression regimen. Relationship between C-peptide and CP/G to b -score No correlation was found at any time point between C-peptide and b-score. The CP/G was found to be significantly associated with b-score at 12 months (r = 0.71, p < 0.001) and 15 months (r = 0.666, p = 0.001) posttransplant. When the values for all patients at all time points (n = 140) were analyzed for C-peptide and CP/G a correlation was found between each of these measurements and the b-score (r = 0.627, p < 0.001; r = 0.723, p < 0.001, respectively). Insulin independence at 15 months post-transplant Significant differences were observed when comparing patients maintaining insulin independence at 15 months after the last islet infusion to those requiring exogenous insulin, based on CP/G (1.14 ± 0.21 vs ± 0.41, respectively; p < 0.01), 90 min-glc after MMTT (6.98 ± 1.16 vs ± 3.21 mmol/l, respectively; p < 0.01) and b-score (7.40 ± 0.55 vs ± 1.73, respectively; p < 0.001). Table 2: Correlation between C-peptide measures with 90-minute glucose post-last islet infusion at different time points C-peptide alone C-peptide/glucose ratio 90 min-glc vs. time after last infusion n r P r p 3 months < months < months < months < months < months < <0.001 Statistically significant. 306 American Journal of Transplantation 2007; 7:

5 Monitoring b -Cell Mass and Function Discussion Development of simple methods to assess islet graft dysfunction and assist with the management of transplanted patients is highly desirable. We evaluated calculations using fasting C-peptide such as CP/G and CP/GCr to assess graft dysfunction. Toward this aim, we assessed which one of these measures best serves as a simple diagnostic test for predicting high levels of 90 min-glc (>10 mmol/l). The CP/G was found to be a superior diagnostic tool to detect graft dysfunction by predicting a high 90 min-glc when compared to C-peptide (p = 0.01) and to CP/GCr (p = 0.06). CP/G and C-peptide, correlated with the total mass of islets infused when assessed at 7 days and 1 and 3 months post-transplant, suggesting that these values are useful in the assessment of islet engraftment. CP/G correlated with the proposed measures for b-cell function after islet transplantation (90 min-glc after MMTT; Figure 2, AIRglu after IVGTT) and with the clinical scoring system (b-score) in islet transplantation (6), suggesting that it represents a valuable surrogate marker of islet graft dysfunction that can be easily monitored in transplanted patients. It is conceivable that CP/GCr would be better to assess islet graft dysfunction in patients with renal disease; however, this hypothesis was not formally addressed in our study, since all subjects had normal and stable renal function. Larger studies in patients with progressive nephropathy will be of assistance to address this issue. The CP/G is a simple test that can be estimated at any time point during the post-transplant period, given the ease of its measurement taken from a single fasting blood sample. It can be used as a parameter of insulin secretion on the basal state and can have an immediate application in the assessment of graft dysfunction after islet transplantation. A high CP/G value provides reassurance that a given patient has adequate islet graft function. A lower value could indicate early graft dysfunction and the requirement of more complicated metabolic and immunologic work-up to guide prompt implementation of anti-rejection and graft rescue therapy. A limitation of this study is the absence of control groups (non-obese, non-diabetic subjects; type 2 diabetes patients with and without kidney disease; kidney pancreas recipients; non-diabetic kidney alone or other solid organ recipients) that could help dissect the effects of kidney disease and immunosuppressive therapy on C-peptide levels. One limitation of the CP/G ratio is that it does not account for the overall patient metabolic state since it does not focus directly on exogenous insulin use or HbA1c levels. Another limitation is that it does not correct for the degree of insulin sensitivity or resistance. To assess the glucose tolerance state, it is best to calculate insulin secretion, insulin sensitivity and the disposition index by the euglycemic hyperinsulinemic glucose clamp (1), the insulin modified frequently sampled IVGTT (19) or the oral glucose tolerance test described by Cobelli et al. (20). Although the HOMA indices have been used to correlate and extrapolate insulin secretion and sensitivity data obtained from the tests mentioned above, in normal controls and in patients with type 2 diabetes, its use has not been validated on subjects using exogenous insulin or in islet transplant recipients (10). Although HOMA C-peptide b-cell index (HOMA-%B) may be superior to CP/G, both measures correlate remarkably well (r = 0.86, p < 0.001) since they contain the same variables with different constants. One advantage of the HOMA b-cell index over the CP/G is the presence of a threshold glucose level of 3.5 mmol/l (63 mg/dl). Calculation of the CP/G is simpler and is superior to the use of C-peptide alone for the detection of graft dysfunction. Further analysis incorporating measures of insulin resistance into these ratios may improve their overall sensitivity and extend their applications to other clinical settings requiring monitoring of b-cell mass and function. In conclusion, our data show that the CP/G is a simple test, easy to perform, which correlates with the mass of transplanted islets, and which is more indicative of graft dysfunction than using CP/GCr and C-peptide alone. The CP/G correlate well with 90 min-glc and with the clinical b-score, which are widely used in islet transplantation. The ease of calculation of this ratio and its good correlation with other tests assessing graft dysfunction make it a valuable complement for the monitoring and management of recipients of allogeneic islets. Acknowledgments We express gratitude to the staff of the Clinical Islet Transplant Program, the Islet Cell Processing Center, the General Clinical Research Center and the Informatics Core for their continuous support. We also thank Dr. Andrea Caumo, Dr. Norma S. Kenyon, Dr. Paul Casanova-Romero and Dr. Jay Sosenko for their valuable and stimulating discussion. This study was supported by grants from the National Institutes of Health/NIDDK (R01 DK52802), NCRR (GCRCMO1RR16587; U42RR016603), Juvenile Diabetes Foundation International-JDRFI ( ), State of Florida and Diabetes Research Institute Foundation. References 1. Ferrannini E, Mari A. Beta cell function and its relation to insulin action in humans: a critical appraisal. Diabetologia 2004; 47: Palmer JP, Fleming GA, Greenbaum CJ et al. C-peptide is the appropriate outcome measure for type 1 diabetes clinical trials to preserve beta-cell function: report of an ADA workshop, October Diabetes 2004; 53: Alejandro R, Lehmann R, Ricordi C et al. Long-term function (6 years) of islet allografts in type 1 diabetes. Diabetes 1997; 46: Greenbaum CJ, Harrison LC. Guidelines for intervention trials in subjects with newly diagnosed type 1 diabetes. Diabetes 2003; 52: American Journal of Transplantation 2007; 7:

6 Faradji et al. 5. Ryan EA, Lakey JR, Paty BW et al. Successful islet transplantation: continued insulin reserve provides long-term glycemic control. Diabetes 2002; 51: Ryan EALJ, Paty BW et al. Beta-score An assessment of betacell function after islet transplantation. Diabetes Care 2005; 28: Secchi A, Pontiroli AE, Traeger J et al. A method for early detection of graft failure in pancreas transplantation. Transplantation 1983; 35: Matsumoto S, Yamada Y, Okitsu T et al. Simple evaluation of engraftment by secretory unit of islet transplant objects for living donor and cadaveric donor fresh or cultured islet transplantation. Transplant Proc 2005; 37: Yamada Y, Fukuda K, Fujimoto S et al. SUIT, secretory units of islets in transplantation: an index for therapeutic management of islet transplanted patients and its application to type 2 diabetes. Diabetes Res Clin Pract 2006; 74: Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care 2004; 27: Zerbini G, Mangili R, Luzi L. Higher post-absorptive C-peptide levels in Type 1 diabetic patients without renal complications. Diabet Med 1999; 16: Froud T, Ricordi C, Baidal DA et al. Islet transplantation in type 1 diabetes mellitus using cultured islets and steroid-free immunosuppression: miami experience. Am J Transplant 2005; 5: Pileggi ARC, Kenyon NS, Froud T et al. Twenty years of clinical islet transplantation at the Diabetes Research institute-university of Miami. In: Cecka JM, Terasaki PI (eds). Clinical transplants. Los Angeles: UCLA Immunogenetics Center, 2004: Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW. Automated method for isolation of human pancreatic islets. Diabetes 1988; 37: Ichii H, Pileggi A, Molano RD et al. Rescue purification maximizes the use of human islet preparations for transplantation. Am J Transplant 2005; 5: Shapiro AM, Lakey JR, Ryan EA et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 2000; 343: Teuscher A, Kendall D, Smets Y, Leone J, Sutherland D, Robertson R. Successful islet autotransplantation in humans: functional insulin secretory reserve as an estimate of surviving islet cell mass. Diabetes 1998; 47: Obuchowski NA. Nonparametric analysis of clustered ROC curve data. Biometrics 1997; 53: Bergman RN, Ader M, Huecking K, Van Citters G. Accurate assessment of beta-cell function: the hyperbolic correction. Diabetes 2002; 51(Suppl 1): S Dalla Man C, Caumo A, Basu R, Rizza R, Toffolo G, Cobelli C. Minimal model estimation of glucose absorption and insulin sensitivity from oral test: validation with a tracer method. Am J Physiol Endocrinol Metab 2004; 287: E637 E American Journal of Transplantation 2007; 7: