Czabak-Garbacz, Roza*†; Schneditz, Daniel*; Zierler, Edda*; Eichmann, Eva*; Harter, Gerson*; Hafner-Giessauf, Hildegard‡; Obermayer-Pietsch, Barbara§
Protein-energy wasting characterized by a loss of appetite, body weight, and muscle mass is considered to be one of the most important factors to determine the higher morbidity and mortality seen in dialysis patients.1 To minimize caloric losses and to protect patients from hypoglycemia, dialysate glucose ranges between 100 and 200 mg/dL (5.5 and 11.1 mmol/L).1–4 Dialysate glucose concentrations as high as 1,600 mg/dL historically used to support ultrafiltration across low-flux dialysis membranes essentially for the same reason glucose is used in peritoneal dialysis have been discontinued. Administration of excess glucose may cause prolonged hyperglycemia, which by itself can be harmful.2,5,6 However, glucose-free dialysate that improves potassium removal during hemodialysis7 bears a considerable risk of hypoglycemia especially in diabetic patients.3,8–10 Thus, the benefits of different levels of glucose in dialysate have been debated.3,8,9,11,12 However, the increasing number of patients suffering from diabetes13–15 or becoming diabetic or prediabetic being on dialysis has once more brought this question to the attention of nephrologists.16
Although several studies have dealt with the problem of dialysate with or without glucose, little is known about the resulting insulinemia. This is of special interest because the delivery of glucose through the dialysate and through extracorporeal blood flow returning to the circulation bypasses the alimentary tract, thereby representing a parenteral mode of glucose administration. Parenteral administration of glucose, however, is known to cause a sluggish insulin response17,18 essentially because the costimulation of insulin secretion by enteric hormones such as glucagon-like peptide-1 (GLP-1) is bypassed. High-glucose dialysate is therefore expected to induce an impaired insulinemia for a given change in plasma glucose concentrations. Notwithstanding, and in contrast to all expectations, an increased insulinemia secondary to intradialytic parenteral nutrition has been reported in a recent study.19
Therefore, we decided to investigate the insulin response to high-glucose dialysate in stable, nondiabetic hemodialysis patients during their regular hemodialysis treatment and to compare the response of parenteral administration using dialysate with the oral administration of glucose.
Materials and Methods
Stable, chronic, and nondiabetic hemodialysis patients from the outpatient dialysis program of the Division of Nephrology, Department of Medicine, Medical University of Graz, were asked to participate in the study approved by the Internal Review Board. Participants provided written informed consent and were asked to fasten for at least 6 hours before and throughout the first part of four regular midweek hemodialysis treatments. The administration of glucose given orally or using the dialysate was started after having established constant treatment conditions during a baseline phase of 30 minutes, and the patient response was assessed for a subsequent observation phase of up to 2 hours.
For delivery using the dialysate, the glucose concentration in the dialysate was varied between 0 mmol/L during the first 30 minutes of baseline phase, followed by 11.1 mmol/L for the duration of 60 minutes, and 0 mmol/L for another 30 minutes. After 90 minutes of intervention, regular hemodialysis was continued using the standard glucose concentration of 5.5 mmol/L. Thus, with plasma glucose levels in the range of 4.5 mmol/L and exposure to high-glucose dialysate of 11.1 mmol/L for the duration of 60 minutes and with an expected glucose clearance of 130 mL/min,20 ∼9 g (50 mmol) of glucose were delivered parenterally to the patient.
In the same patient, but on a different treatment day, glucose was also administered orally during hemodialysis using a standard dialysate glucose concentration of 5.5 mmoL/L. For oral administration, 75 g (417 mmoL) glucose (Glucoral; Unipack, Wiener Neustadt, Austria) was dissolved in 300 mL of water and the solution was ingested within 5 minutes in a sitting body position. The glucose and insulin responses resulting from both parenteral and oral glucose administrations were measured and compared. All tests were repeated once.
Xenium (Baxter, Zurich, Switzerland) and Polyflux (Gambro AB, Lund, Sweden) dialyzers and Gambro dialysis machines (Gambro AB, Lund, Sweden) were used to deliver bicarbonate hemodialysis or online hemodiafiltration at a dialysate flow of 500 mL/min. Ultrafiltration rates and electrolytes were adjusted and delivered as prescribed: 32–34 mmol/L for HCO3−, 134–138 mmol/L for Na+, 1–4 mmol/L for K+, 1.25 mmol/L for Ca2+, and 0.75 mmol/L for Mg2+. Arterial blood pressures and heart rates were measured by cuff-technique and oscillometric detection.
Blood samples were taken from the arterial line of the extracorporeal circulation. Glucose concentrations (cG, in mmol/L) were measured by glucose oxidase using the cobas b211 (Roche Diagnostics GmbH, Graz, Austria). Insulin concentrations (cI, in mU/L) were measured by two-side enzyme immunoassay (DRG Insulin ELISA EIA 1825; DRG Instrument GmbH, Marburg/Lahn, Germany). Plasma concentrations were corrected for plasma water content of 93%.20
The response of insulin to a given glucose stimulus was quantified by the slope of the linear regression of paired insulin and glucose concentrations. This measure has been shown to be independent of extracorporeal clearance by ongoing hemodialysis21 and has been termed insulinogenic index (IG, in U/mol) because of the mathematical equivalence to the IG introduced by Seltzer et al.22
The homeostasis model assessment (HOMA, in mmol · mU/L2) index of insulin resistance was calculated as fasting glucose concentrations (in mmol/L) multiplied by fasting insulin (in mU/L) divided by 22.5.23,24
The correlations between different variables and the relationships between identical variables obtained in subsequent measures were examined by analysis of variance (ANOVA) and linear regression analysis. A probability p < 0.05 was considered significant to reject the null hypothesis. Data are presented as means ± standard deviation.
Complete sets of duplicate data done with both parenteral and oral glucose administration were available in 10 patients (6 female patients) so that 40 studies entered final analysis (Table 1). Six subjects received online hemodiafiltration (HDF) with a mean replacement volume of 31 ± 5.7 L (Table 2). Five patients were treated by a central-venous access. In both experimental procedures, fasting plasma glucose and insulin concentrations and the corresponding HOMA index of insulin resistance indices were almost identical and within normal physiological range. No complications or meaningful changes in mean arterial blood pressures and heart rates were observed during the study.
At the end of the baseline phase, plasma glucose and insulin concentrations slightly dropped to 4.8 ± 0.7 mmol/L and 1.1 ± 1.3 mU/L because of 30 min of hemodialysis against glucose-free dialysate (Table 3). Next (t = 0), dialysate was changed to 11.1 mmol/L for the duration of 1 hour, which led to a continuous increase of both glucose and insulin concentrations above baseline levels (Figure 1). The maximum concentrations of 7.2 ± 0.7 mmol/L for glucose and 8.6 ± 3.4 mU/L for insulin were obtained at the end of that phase (t = 60). At that time, dialysis was continued with glucose-free dialysate for a period of 30 min during which plasma glucose and insulin concentrations dropped until the end of the observation phase (t = 90). The glucose perturbations and the resulting insulin responses were highly reproducible (Figure 2).
After 30 minutes of hemodialysis against regular dialysate, plasma glucose and insulin concentrations reached values of 5.7 ± 0.5 mmol/L and 3.8 ± 3.2 mU/L, respectively (Figure 3). At that time (t = 0), 417 mmol of glucose was ingested which led to a significant and sustained increase of plasma glucose and insulin concentrations (Table 3). Maximum glucose and insulin concentrations of 10.8 ± 1.6 mmol/L and 53.5 ± 22.5 mU/L were reached 1 hour (t = 60) after oral glucose administration. The glucose perturbations and the resulting insulin responses were highly reproducible (Figure 4).
Oral Versus Parenteral Administration
Paired glucose and insulin concentrations showed a strong linear relationship for both modes of glucose administration characterized by the IG. Such a powerful correlation was seen for average concentrations measured in all studies split for the mode of glucose administration (Figure 5). The IG was fairly reproducible within individual patients but different between patients and significantly different between the two modes of glucose delivered to the same patient (Figure 6, Table 3). The IG was 3.0 ± 1.1 U/mol with parenteral administration and always lower (p < 0.001, Table 3) and only 34 ± 13% of that seen with oral delivery (9.3 ± 2.6 U/mol). Conversely, the change in glycemia normalized to the change in insulinemia was larger with parenteral (0.38 ± 0.21 mol/U) compared with oral administration (0.12 ± 0.05 mol/U) in all studies (p < 0.001, Table 3).
In this study, the response of plasma glucose and insulin to oral or parenteral administration of glucose using a high glucose concentration in the dialysate was studied in stable nondiabetic end-stage renal disease patients during their regular hemodialysis treatment. The main findings of this study are as follows: a) the perturbation produces significant and reproducible changes in plasma glucose and insulin concentrations; b) the response is different for the two modes of glucose administration; and c) the change in insulin for a given change in glucose is much larger with oral than with parenteral administration of glucose.
The examination of the glucose/insulin system during hemodialysis is uncommon, probably because hemodialysis interferes with glucose and insulin balance. Both glucose and insulin are cleared by hemodialysis, and the interpretation of concentrations is difficult because concentrations are not unambiguously related to the condition of the patient studied.20 However, if the contribution of the extracorporeal system is measured, the effects caused by dialysis can be accounted for. In a previous study dealing with this question, it was also observed that the ratio of insulin to glucose changes was independent of extracorporeal clearance and ongoing hemodialysis.21 This makes the IG suitable for comparing the glucose/insulin system in hemodialysis patients during their regular dialysis treatment.
Although the time course of plasma glucose and insulin concentrations was different between modes of glucose delivery, both oral and parenteral administration produced significant and reproducible changes in plasma glucose and insulin concentrations. A comparison of glucose levels (Table 3) shows that the change following parenteral administration (2.3 ± 0.6 mmol/L) was about half of that induced by oral glucose delivery (5.2 ± 1.6 mmol/L). This difference essentially can be explained by the smaller amount of glucose delivered with parenteral (50 mmol, 9 g) compared with oral administration (417 mmol, 75 g).
In comparison, the change in insulin levels (7.5 ± 3.2 mU/L) with glucose delivery using the dialysate was only about one-seventh of that (49.6 ± 22.1, Table 3) seen after oral administration. Although reduced insulinemia is expected with parenteral delivery (see below), such a comparison has to account for differences in glycemia imposed by the two protocols. One possibility to explain the differences in glucose concentrations is based on relating the changes in insulin to the changes in glucose, or in other words, by comparing the insulinogenic indices. This approach provides a normalization of the insulin response to particular changes of glycemia. However, even after normalization, the insulin response with parenteral administration was only 34 ± 13% of that seen with oral delivery of glucose. The IG observed with parenteral administration in this study is in line with that of 6.06 ± 5.04 U/mol using a much larger intravenous glucose dose (38.5 ± 9.1 g) in a previous study,21 as both indices were lower than that seen with oral glucose delivery (9.5 ± 2.7 U/mol, Table 3).
The insulinemia observed in this study is consistent with the so-called incretin effect. With oral administration and the subsequent release of enteric hormones such as GLP-1 from the small intestine, insulin secretion from β-cells is enhanced in excess of the release controlled by glucose alone. In normal subjects, incretins are responsible for 20–70% of insulin release increasing both hepatic glucose uptake and peripheral glucose utilization and thereby blunting the level and the duration of hyperglycemia.17,18,25 With parenteral administration of glucose bypassing the small intestine and in absence of costimulation by incretins, insulinemia is blunted so that the level and the duration of hyperglycemia is potentially enhanced. This effect may be important in peritoneal dialysis where glucose is used as an osmotic agent for fluid removal.26
However, the insulin response observed in this investigation is at odds with data from a recent study19 in which the authors concluded that the incretin effect was abolished during hemodialysis. However, the infusate used for parenteral nutrition in that study contained large amounts of insulin (6 U delivered at a rate of 1.5 U/h) and the blunted insulinemia expected with parenteral administration in absence of incretin action seems to have been masked by the infusion of insulin.
Hyperglycemia, again, is known to impair β-cell function, to strain the mitochondria and to act proinflammatory,2,5,27 all of which could also be harmful in subjects treated with hemodialysis. Therefore, parenteral administration of glucose including the use of high-glucose dialysate solutions should be avoided in hemodialysis patients as long as the enteral mode of administration is feasible.
In conclusion, hemodialysis with a dialysate glucose concentration of 11.1 mmol/L (200 mg/dL) leads to a small but significant influx of glucose into the patient and stimulates a significant albeit inadequate release of insulin, as incretin-induced stimulation of β-cells is bypassed by this parenteral mode of glucose administration. As a consequence, glycemia per unit insulin concentration is increased. This is suspected to provide proinflammatory stress and should be avoided with hemodialysis. To which degree these observations can be extended to diabetic and prediabetic hemodialysis patient remains to be studied.
The authors thank the patients for participating in this study, Gabriele Koch and her staff at the dialysis unit for the collaborative effort, Juliane Buchgraber and Michaela Eichinger, PhD, from the nephrology and endocrinology laboratories for experimental help.
1. Dukkipati R, Kalantar-Zadeh K, Kopple JD: Is there a role for intradialytic parenteral nutrition? A review of the evidence. Am J Kidney Dis 55: 352–364, 2010.
2. Sharma R, Rosner MH: Glucose in the dialysate: Historical perspective and possible implications? Hemodial Int 12: 221–226, 2008.
3. Takahashi A, Kubota T, Shibahara N, et al: The mechanism of hypoglycemia caused by hemodialysis. Clin Nephrol 62: 362–368, 2004.
4. Smolle KH, Kaufmann P, Holzer H, Druml W: Intradialytic parenteral nutrition in malnourished patients on chronic haemodialysis therapy. Nephrol Dial Transplant 10: 1411–1416, 1995.
5. Deacon CF: Incretin-based treatment of type 2 diabetes: Glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors. Diabetes Obes Metab 9: 23–31, 2007.
6. Schetz M, Vanhorebeek I, Wouters PJ, et al: Tight blood glucose control is renoprotective in critically ill patients. J Am Soc Nephrol 19: 571–578, 2008.
7. Zehnder CE, Gutzwiller J-P, Huber A, et al: Low-potassium and glucose-free dialysis maintains urea but enhances potassium removal. Nephrol Dial Transplant 16: 78–84, 2001.
8. Jackson MA, Holland MR, Nicholas J, et al: Hemodialysis-induced hypoglycemia in diabetic patients. Clin Nephrol 54: 30–34, 2000.
9. Burmeister JE, Scapini A, da Rosa Miltersteiner D, et al: Glucose-added dialysis fluid prevents asymptomatic hypoglycaemia in regular haemodialysis. Nephrol Dial Transplant 22: 1184–1189, 2007.
10. Biolo G, Stulle M, Bianco F, et al: Insulin action on glucose and protein metabolism during L-carnitine supplementation in maintenance haemodialysis patients. Nephrol Dial Transplant 23: 991–997, 2008.
11. Raimann JG, Kruse A, Thijssen S, et al: Fatigue in hemodialysis patients with and without diabetes: Results from a randomized controlled trial of two glucose-containing dialysates. Diabetes Care 33: e121, 2010.
12. Schneditz D: Glucose-added dialysis fluid prevents asymptomatic hypoglycaemia in regular haemodialysis. Nephrol Dial Transplant 23: 1066–1067, 2008.
13. Bouffard Y, Tissot S, Delafosse B, et al: Metabolic effects of hemodialysis with and without glucose in the dialysate. Kidney Int 43: 1086–1090, 1993.
14. Mahnensmith RL, Zorzanello M, Hsu Y-H, Williams ME: A quality improvement model for optimizing care of the diabetic end-stage renal disease patient. Semin Dial 23: 206–213, 2010.
15. Kovesdy CP, Park JC, Kalantar-Zadeh K: Glycemic control and burnt-out diabetes in ESRD. Semin Dial 23: 148–156, 2010.
16. Simic-Ogrizovic S, Backus G, Mayer A, et al: The influence of different glucose concentrations in haemodialysis solutions on metabolism and blood pressure stability in diabetic patients. Int J Artif Organs 24: 863–869, 2001.
17. Perley MJ, Kipnis DM: Plasma insulin responses to oral and intravenous glucose: Studies in normal and diabetic subjects. J Clin Invest 46: 1954–1962, 1967.
18. Nauck MA, Homberger E, Siegel EG, et al: Incretin effects of increasing glucose loads in man calculated from venous insulin and C-peptide responses. J Clin Endocrinol Metab 63: 492–498, 1986.
19. Fernández-Reyes MJ, Sánchez R, García L, et al: Acute responses of gastrointestinal hormones to both oral and parenteral intradialytic nutrition. Am J Nephrol 32: 272–278, 2010.
20. Schneditz D, Hafner-Giessauf H, Holzer H, Thomaseth K: Intracorporeal glucose disposal during hemodialysis after a standardized glucose load. ASAIO J 56: 204–209, 2010.
21. Schneditz D, Hafner-Giessauf H, Thomaseth K, et al: Insulinogenic index in non-diabetics during haemodialysis. Nephrol Dial Transplant 25: 3365–3372, 2010.
22. Seltzer HS, Allen EW, Herron AL, Brennan MT: Insulin secretion in response to glycemic stimulus: Relation of delayed initial release to carbohydrate intolerance in mild diabetes mellitus. J Clin Invest 46: 323–335, 1967.
23. Kanauchi M, Kimura K, Akai Y, Saito Y: Insulin resistance and pancreatic beta-cell function in patients with hypertensive kidney disease. Nephrol Dial Transplant 19: 2025–2029, 2004.
24. Muniyappa R, Lee S, Chen H, Quon MJ: Current approaches for assessing insulin sensitivity and resistance in vivo: Advantages, limitations, and appropriate usage. Am J Physiol Endocrinol Metab 294: E15–E26, 2008.
25. Dalla Man C, Micheletto F, Sathananthan A, et al: A model of GLP-1 action on insulin secretion in nondiabetic subjects. Am J Physiol Endocrinol Metab 298: E1115–E1121, 2010.
26. Galach M, Waniewski J, Axelsson J, et al: Mathematical modeling of the glucose-insulin system during peritoneal dialysis with glucose-based fluids. ASAIO J 57: 41–47, 2011.
27. Vanhorebeek I, Langouche L: Molecular mechanisms behind clinical benefits of intensive insulin therapy during critical illness: glucose vs. insulin. Best Pract Res Clin Anaesthesiol 23: 449–459, 2009.