Energy Metabolism in Diabetic and Nondiabetic Heart Transplant Recipients

Patients underwent heart transplantation at the Cardiac Surgery Unit of the Ospedali Riuniti of Bergamo and were classified as diabetic (D-Tx) and nondiabetic (Tx) subjects by means of a standard 75-g OGTT according to the 1997 American Diabetes Association criteria. In a total population of 480 heart transplant patients (with a 6% prevalence of type 2 diabetes), 18 patients undergoing heart transplantation for post-ischemic heart disease or idiopathic cardiomyopathy (12 with type 2 diabetes and 6 without diabetes) were selected, and 5 healthy volunteers served as control subjects (CON). Two D-Tx were on insulin therapy (29± 8 units/day), four were on oral hypoglycemic agents, and six were on diet. All patients were in stable clinical and nutritional condition at the time of study. All transplant patients (with the exception of four patients on 2.5–5 mg/day of prednisone) were on a steroid-free immunosuppressive regimen including cyclosporin A (4.2± 0.6 mg/day) and azathoprine (0.8± 0.3 mg/day). Diabetic patients were evaluated for the severity of nephropathy (microalbuminuria), retinopathy (fundus oculi), and neuropathy (nerve conduction velocity). Subjects were fully informed of the possible risks of the study and gave their consent. The experimental protocol was approved by the Ethical Committee of the Istituto Scientifico H San Raffaele.

All subjects underwent an OGTT and a euglycemic-hyperinsulinemic clamp. In the 2 weeks preceding the study, all the subjects assumed an isocaloric diet containing at least 250 g carbohydrate and 70–90 g protein per day. On day 1 at 5 p.m., the patients were admitted to the Department of Internal Medicine I of the Istituto Scientifico H San Raffaele for the OGTT (75 g) and the euglycemic insulin clamp. D-Tx on insulin received the last doses of intermediate- and short-acting insulin 24 and 12 h, respectively, before the OGTT; patients on sulfonylureas discontinued the medication at least 3 days before the study. After a 12-h overnight fast, the OGTT or insulin clamp began at 8:00 a.m. of day 2 (the OGTT or the clamp was randomly performed as the first test protocol). After the first test, plasma glucose was adjusted in diabetic patients by means of subcutaneous injections of 2–4 units of short-acting insulin aimed at keeping the blood glucose level between 100 and 180 mg/dl; these injections were given until 11:00 p.m. of day 2. At 8:00 a.m. in the morning of day 3, the second test was performed. H ¹-NMR spectroscopy was performed on day 4 to assess muscle triglyceride content in five D-Tx, four Tx, and five CON.

Twelve D-Tx, six Tx, and five CON were studied (Table 1). The study was performed to assess the insulin secretory response after an oral glucose load. It was performed at 8:00 a.m. according to the 1997 American Diabetes Association recommendations (oral glucose load, 75 g) (11). A Teflon catheter was inserted into an antecubital vein for blood sampling, and samples were obtained for glucose, free insulin, and C-peptide measurements in the basal period and during the OGTT sampling at 30-min intervals for 180 min.

All subjects underwent the insulin clamp to assess the whole-body glucose disposal and the endogenous glucose production during physiological hyperinsulinemia. The clamp was performed as previously described (10,12). Continuous indirect calorimetry (Sensor Medics Italia, model 2900, Milan) was performed during the last 45 min of the study as previously described (10).

¹H-NMR spectroscopy was performed on a Signa 1.5 Tesla scanner (General Electric Medical Systems, Milwaukee, WI) using a conventional linear extremity coil. This analytical procedure was previously described (13). Only four Tx (out of six) and five D-Tx (out of twelve) underwent the ¹H-NMR spectroscopy study because of technical problems (pacemakers) or claustrophobia.

All blood samples were placed on ice until the plasma or serum was prepared by centrifugation at 4°C (within 1.5 h of sampling). All plasma and serum aliquots were frozen at −60°C until later analysis. The d₂-glucose enrichment was measured by gas chromatography–mass spectronomy (GCMS) after preparation of the butyl-boronate derivative (14). Samples were injected into a GCMS instrument (Hewlett-Packard, model 5970; Hewlett-Packard, Palo Alto, CA) operated using electron impact ionization. The ion masses at 297 and 299 m/z were selectively monitored for unlabeled glucose and d₂-glucose, respectively. After addition of the appropriate internal standards ([²H₇]leucine, [ ²H₂]phenylalanine, and ketocaproate), amino and keto acids were isolated from 0.5-ml aliquots of plasma using cation-exchange columns as previously described (15). The α-ketoisocaproic (KIC) acid enrichment and concentration were measured after eluants from the columns were derivatized to the trimethylsilyl-quinoxalinol derivatives (16). Plasma glucose was measured at the bedside with a glucose analyzer (Beckman Instruments, Fullerton, CA) (10) and free insulin, glucagon, C-peptide growth hormone, and cortisol with a radioimmunoassay (10). The determination of epinephrine was measured with a high-performance liquid chromatography as previously described (16). Lactate, β-O-hydroxybutyrate (β-OHB), pyruvate, free fatty acid (FFA), and glycerol were measured as previously described (17,18). Glucose and amino acid kinetics were calculated using Steele’s equation for the non–steady state (19,20).

Data are expressed as means ±SEM. Comparison between the basal and insulin-stimulated state were performed with the Student’s t test for paired data. Comparisons among groups were performed using ANOVA for repeated measures.

The underlying disease leading to heart allograft was heart failure due to idiopathic cardiomyopathy (13 cases) or ischemic heart disease (5 cases). The allograft successfully cured heart failure in all cases. Of 12 patients with diabetes after transplantation, 7 became diabetic after transplantation and 5 were already diabetic. No heart transplant patient with normal glucose tolerance had biochemical or clinical signs of nephropathy, retinopathy, or neuropathy. In contrast, in the 12 heart transplant patients with overt diabetes (5± 1 years after transplantation), 2 showed clinically significant microalbuminuria, 4 had retinopathy, and 6 had reduced nerve conduction velocity (Table 1).

Fasting plasma glucose was increased in D-Tx when compared with the CON. In contrast, the fasting plasma glucose was within the normal range in Tx. The plasma glucose level at 120 min was significantly higher in D-Tx during the OGTT that in CON. In contrast, in Tx, the plasma glucose concentration at 120 min was slightly higher than that in CON (but always within values either of normality or of impaired glucose tolerance). D-Tx were characterized by a defective insulin response during the first hour after the glucose challenge, whereas Tx showed a response similar to normal subjects. D-Tx showed higher postabsorptive and glucose-stimulated C-peptide concentrations than CON. At 120 min after glucose ingestion, the C-peptide value was significantly higher in all transplant patients with respect to normal subjects. Nonetheless, the concentration of C-peptide at 120 min was lower in D-Tx than in Tx (P < 0.05).

In the postabsorptive state, plasma glucagon levels tended to be higher in D-Tx than in Tx and CON. In contrast, the postabsorptive plasma concentrations of growthhormone, cortisol, epinephrine, and norepinephrine were comparable in the three groups. Plasma FFAs were significantly higher in D-Tx (P < 0.05 vs. Tx and CON) in the basal state. Plasma lactate and glycerol concentrations also showed a trend to be higher in D-Tx (P=0.26 vs. CON and P < 0.05 vs. Tx). No difference in the basal state was observed for pyruvate, β-OHB, and alanine among the three groups. During the last hour of the euglycemic-hyperinsulinemic clamp, the free insulin level increased similarly, and the C-peptide concentration was suppressed equally in the three groups. No statistically significant variation from basal was observed during insulin infusion for the growth hormone, cortisol, epinephrine, and norepinephrine concentrations. During insulin infusion, both lactate and pyruvate levels tended to increase (lactate: P=0.17 in D-Tx, P=0.03 in Tx, and P=0.004 in CON; pyruvate: P=0.09 in D-Tx, P=0.19 in Tx, and P=0.34 in CON, with respect to basal). Plasma glycerol concentration during euglycemic hyperinsulinemia was significantly higher in Tx versus CON and D-Tx (P < 0.05), indicating a defective anti-lipolytic action of insulin in diabetic recipients. In contrast, the insulin-mediated inhibition of β-OHB was similar in the three groups. Finally, insulin did not cause variations of alanine concentration in any group (Table 2).

The postabsorptive endogenous glucose production and its inhibition by insulin was not statistically different among the three groups. In contrast, D-Tx showed a marked defect in the insulin-mediated stimulation of glucose oxidation and nonoxidative glucose disposal. Therefore, both the total glucose disposal and the total glucose disposal corrected for the ambient glucose concentration (namely the metabolic clearance rate of glucose) was markedly impaired in D-Tx compared with Tx and CON (Table 3).

Postabsorptive leucine (129.1± 9.9 to 89.2± 6.8 mmol/l in D-Tx, 111.5± 3.1 to 68.0± 5.1 mmol/l in Tx, and 126.8± 6.2 to 79.8± 5.0 mmol/l in CON) and phenylalanine (59.1± 3.6 to 49.1± 3.2 mmol/l in D-Tx, 51.5± 2.1 to 38.0± 3.4 mmol/l in Tx, and 52.4± 2.1 to 41.8± 2.1 mmol/l in CON) concentrations decreased less in D-Tx that in Tx and CON during euglycemic hyperinsulinemia. Postabsorptive KIC level was similar in the three groups and also decreased similarly during hyperinsulinemia (41.0± 6.6 to 25.8± 5.5, 37.9± 5.9 to 22.0± 2.0, and 44.5± 1.8 to 24.1± 2.4 mmol/l in D-Tx, Tx, and CON, respectively) (Fig. 1). Finally, the insulin-mediated suppression of exogenous leucine flux (proteolysis) was impaired in D-Tx (13%) compared with Tx (20%) and CON (20%) (102.1± 4.6 to 89.1± 2.5, 126.2± 15.2 to 100.3± 12.2, and 117.1± 2.9 to 94.8± 2.4 mmol · kg ⁻¹ · h ⁻¹ in D-Tx, Tx, and CON, respectively, indicating a defect of the anti-proteolytic action of insulin in D-Tx (Fig. 2).

Plasma lipid profile was abnormal in both D-Tx and Tx when compared with CON (Tables 1 and 3). Plasma FFA level was increased in D-Tx when compared with both Tx (P < 0.05) and CON (P < 0.05). During insulin infusion, the plasma FFA level showed a trend (Table 3, P=0.08) toward a defective suppression in D-Tx compared with Tx and CON. Lipid oxidation, measured by means of indirect calorimetry was comparable among all study groups in the postabsorptive condition (0.9± 0.1, 1.0± 0.2, and 0.9± 0.1 mg/[kg·min] in D-Tx, Tx, and CON, respectively). During the clamp, insulin-dependent inhibition of whole-body lipid oxidation was also similar (0.47± 0.11, 0.27± 0.11, and 0.40± 0.09 mg/[kg·min] in D-Tx, Tx, and CON, respectively). IMCL content in the soleus muscle was higher in D-Tx (103± 18 arbitrary units [AU]) than in Tx (71± 19 AU; P=0.09) and CON (45± 11 AU; P=0.02). IMCL content in the tibialis anterior muscle was significantly increased in D-Tx (24± 3 AU) compared with both Tx (11± 3 AU, P=0.01) and CON (10± 2 AU, P=0.01). The IMCL content of the soleus (R²=0.52, P=0.03) and the tibialis anterior (R²=0.49, P=0.05) showed a significant association with whole-body insulin sensitivity when data of the three groups were pooled.

To assess the stage of diabetic complications, all transplant patients underwent the measurement of nerve conduction velocity (neuropathy), the assessment of microalbuminuria (nephropathy), and an ophthalmologic examination (retinopathy). In Tx, no renal, retinal, and neurological alterations were found. In contrast, 2 of 12 D-Tx had microalbuminuria, 6 had a reduction of nerve conduction velocity, and 4 had retinopathy (Table 1).

We studied the glucose, leucine, and lipid metabolism in the postabsorptive and insulin-stimulated conditions in heart transplant patients with and without diabetes with a 5-year average graft duration. Patients were classified as diabetic (D-Tx) and nondiabetic (Tx) according to the 1997 AmericanDiabetes Association recommendations. Besides the higher incidence of pre- and postoperative complications, diabeticpatients had worse metabolic control than nondiabetic transplant patients. Additional problems were caused by thenecessity to perform life-long immunosuppressive treatment. In fact, cyclosporine A is known both to impair insulin secretion and to decrease insulin action (21,22). The avoidance of steroidal immunosuppressive drugs dramatically reduced the derangements of glucose metabolism. This was shown by us in other populations of organ transplant patients, such as liver (9) and kidney recipients (10). Interestingly, in patients already diabetic before transplant and with poor metabolic control, heart transplantation did not cause any additional worsening of either the triglyceride or cholesterol plasma levels (Table 4). This result was obtained at the expense of higher insulinization. In contrast, nondiabetic patients undergoing heart transplantation and immunosuppression demonstrated an increase of fasting glucose (P=0.29), triglyceride (P=0.04), and cholesterol (P=0.01) concentrations with respect to the condition before the transplantation. In fact, of the 12 diabetic patients studied in this protocol, only 5 were already diabetic before the transplantation, whereas 7 became diabetic after the transplantation. It is evident that both the genetic background (predisposing to type 2 diabetes) and the immunosuppressive drugs are responsible for the hyperglycemia and other metabolic derangements seen after heart transplantation. The major conclusions of the present work are based on the comparison between diabetic and nondiabetic heart transplant patients because control groups of type 2 diabetic patients (not transplant patients) on and without immunosuppressive drugs are not available.

The defect of insulin action in diabetic transplant patients is common in the oxidative and nonoxidative pathway (mainly glycogen synthesis, under the present experimental conditions of euglycemic hyperinsulinemia). It is noteworthy that transplant patients without diabetes have a normal postabsorptive and insulin-mediated glucose metabolism (Table 3). Similarly, leucine metabolism is impaired in diabetic heart recipients (mainly where insulin-mediated inhibition of proteolysis is concerned) but is normal in nondiabetic heart recipients. Finally, the insulin-mediated inhibition of glycerol and FFAs is altered in diabetic recipients but is normal in nondiabetic heart transplant patients (Table 2). In contrast, triglyceride and cholesterol levels are not different in diabetic and nondiabetic heart recipients, being higher than normal in both groups (Table 1). Taken together, the present results indicate a generalized defect of insulin action in diabetic recipients, probably consequent to the concomitance of diabetes and immunosuppressive treatment. In contrast, nondiabetic recipients show normal glucose, leucine, and FFA metabolism, with a persistent impairment of lipid profile, probably due to immunosuppressive treatment. This is probably consequent to a modest worsening of peripheral (not hepatic) insulin resistance (due to immunosuppressive treatment) with respect to the pretransplant condition. In a previous work, we showed that cyclosporine alone (or combined with azathioprine) causes no major metabolic derangement (10). It is also important to note that type 2 diabetic patients without immunosuppression show a defect of insulin-mediated glucose metabolism similar to that of diabetic heart recipients studied in this work (23). Nonetheless, type 2 diabetes not associated with a marked increment of body weight is commonly characterized by normal insulin action on protein metabolism (23). This more generalized defect of insulin action in transplant patients was also identified by us in other organ transplantation populations (10,24).

A further consideration must be made on the pathogenic role of hyperglycemia and metabolic derangements on diabetic complications. Recent clear-cut data (the Diabetes Control and Complications Trial and the U.K. Prospective Diabetes Study) have indicated that the intensive insulin treatment leading to a near-normalization of glucose control and of glycosylated hemoglobin may prevent or at least delay the development of microvascular complications (25,26). In contrast, the effect of strict metabolic control via insulin on macrovascular complications is still debated (25). This means that strict glycemic control may prevent the development or progression of diabetic complications in diabetic heart recipients, but the overinsulinization consequent to intensive insulin treatment may even accelerate (theoretically) the atherosclerotic disease in transplanted hearts (27). Also cyclosporin may play a dual role in the progression of diabetic complications: in fact, it was shown to induce alteration of lipid profile (28,29). The IMCL content correlates well with insulin action. Its rise in diabetic heart recipients is an additional index of poor metabolic control. A number of previous studies showed an impaired insulin action in patients with coronary vessel atherosclerosis. The ¹H-NMR spectroscopy of calf muscles allows a noninvasive repeatable way to assess and follow up on insulin resistance in heart transplant recipients (30).

In conclusion, we presently show that diabetic heart recipients have altered glucose, leucine, and lipid metabolism, which is known to be associated with a high prevalence of diabetic complications. Diabetes may precede or be a consequence of heart transplantation and immunosuppressive treatment and is characterized by both reduced insulin secretion and action. In contrast, nondiabetic heart recipients have a nearly normal metabolic picture. The present results have a strong clinical impact: 1) diabetic heart recipients require an intensive and aggressive management of hyperglycemia to prevent micro- and macrovascular complications, and 2) nondiabetic heart recipients need close monitoring of major parameters of insulin action (IMCL content, plasma lipid profile, and glucose-insulin responses during an oral glucose challenge). A steroid-free immunosuppressive treatment is advisable in diabetic and nondiabetic heart recipients to minimize metabolic derangements.

This work was supported by grants from the Italian Ministry of Health (030.5/RF96.305 and 030.5/RF98.49) and Italian National Research Council (CNR 97.00485.CT04).

We wish to thank Van Chuong Phan, Paola Sandoli, and Sabrina Costa for skilled work with radioimmunoassay measurements, and Antonella Scollo, RN, of the Metabolic Unit of the Istituto Scientifico H San Raffaele for nursing assistance.