Until recently, the main interest in glucose during hemodialysis has been to prevent hemodialysis-induced hypoglycemia and the loss of calories when using glucose-free dialysate.1,2 Meanwhile, there is growing concern regarding uncontrolled hyperglycemia.3 Furthermore, impaired glucose control is considered to play a major role in the high morbidity and mortality of hemodialysis patients. Classic markers of glucose control such as HbA1c and fructosamin are of limited value in diabetic hemodialysis patients, and there is increased interest in continuous glucose monitoring.4–6 We, therefore, speculated whether characteristics of glucose control could be assessed during hemodialysis in the presence of ongoing extracorporeal clearance.
The extracorporeal system used in hemodialysis offers an almost unique and under-recognized access to the circulation of the patient. First, solutes such as glucose can be delivered to the patient at high rates by injection into the venous line of the extracorporeal circulation. Second, the response of the patient is easily measured in the blood entering the extracorporeal circulation and in the dialysate outflow. And third, such tests can be done during regular hemodialysis, possibly from measurements in dialysate alone.
It was the aim of this study to quantify the contribution of extracorporeal clearance on overall disposal of glucose after a standardized load delivered into the extracorporeal blood stream during hemodialysis, to correct for the contribution of extracorporeal clearance, to estimate intracorporeal clearance, and to explore whether intracorporeal disposal could be assessed without blood sampling.
Materials and Methods
Studies were done in stable, non-diabetic maintenance hemodialysis patients during their regular dialysis treatment as approved by the Ethics Committee of the Medical University of Graz. Patients were asked to refrain from eating and drinking >3 h before starting hemodialysis and for the duration of the test.
Hemodialysis was delivered using a 4008H dialysis machine (Fresenius Medical Care, Schweinfurt, Germany) delivering bicarbonate dialysate at a dialysate flow Qd of 500 mL/min. The composition of dialysate was 134 to 138 mmol/L (Na+), 32 to 34 mmol/L (HCO3−), 1 to 4 mmol/L (K+), 1.25 mmol/L (Ca2+), 0.75 mmol/L (Mg2+), and 5.5 mmol/L (glucose). Ultrafiltration rate Qu was set at a constant rate and delivered as prescribed.
Approximately 30 min after having started hemodialysis and after having measured dialysis recirculation using the blood temperature monitor (Fresenius Medical Care, Bad Homburg, Germany), a 33% solution of glucose was injected into the venous drip chamber at a constant rate of 1 mL/s (Angiomat 6000, Liebel-Flarsheim, Cincinnati, OH) at time t = 0. The infusion volume was chosen to deliver a glucose load of 0.5 g/kg target body mass (so-called dry weight) comparable with the standard intravenous glucose tolerance test. Samples were taken from the arterial blood line at t = −30, −15, 0, 3, 6, 10, 14, 20, 30, 40, and 60 min, from the dialysate outlet at t = 0, 3, 6, 10, 14, 20, 30, 40, and 60 min, and from the dialysate inlet at t = 0, 10, and 20 min. Plasma was separated from blood samples by centrifugation. Plasma and dialysate samples were analyzed for glucose concentration by glucose oxidase and amperometric measurement of resulting hydrogen peroxide (cobas b221, Roche Diagnostics, Graz, Austria). The precision for the whole range of glucose concentrations examined in this study was better than 3%. Glucose concentration in plasma was corrected for a plasma water fraction of fpw = 0.93.7
Blood Side Analysis
In the single compartment model with constant solute clearance K, in the absence of solute generation, and under the assumption of a constant volume V, the concentration ct of a solute at time t is given as
The constants A and k of this relationship are obtained from linear regression analysis of the logarithms of solute concentrations measured at successive time points. The rate constant k in Equation 1 is related to clearance K by
where the solute distribution volume V is obtained from M, the mass of solute injected into the system, and from the concentration c0 according to
In the experimental situation, and to account for the duration of solute injection, the concentration c0 was determined from Equation 1 for the time point after having injected half of the solute.
With hemodialysis, and in case of a constant glucose concentration in dialysate inflow cd,in, the concentrations ct and c0 are conveniently replaced by the glucose gradient (cart − cd,in). The average blood to dialysate glucose gradient is obtained from the area under the curve (AUC) and from the observation time t. The average glucose flux from blood into dialysate is then given as
Effective extracorporeal clearance Ke was derived from dialyzer clearance Kd and from the magnitude of combined access and cardiopulmonary recirculation R and extracorporeal blood flow Qb according to the following relationship described elsewhere8–10:
Dialyzer clearance Kd was determined by the standard relationship
where cd,in and cd,out refer to concentrations in dialysate in- and outflow, and where cart refers to equilibrated plasma concentration in blood inflow. Dialysate inflow Qd and ultrafiltration rate Qu were read from the settings of the dialysis machine. In presence of simultaneous intra- and extracorporeal clearance (Ki, Ke) and because the effect of clearance is additive, the overall clearance K and the overall rate constant k during hemodialysis is given as
respectively, where ki = Ki/V and ke = Ke/V. Thus, with the information on Ke obtained from extracorporeal measurements and V (Equation 3) intracorporeal clearance Ki as well as the intracorporeal rate constant ki are given as
To account for differences in clearance and distribution volumes, it is useful to normalize Ki and to introduce the fraction of intracorporeal clearance FK as
Dialysate Side Analysis
The extracorporeal flux of glucose Jd into the dialysate both by diffusive and convective transport is given as
The amount of glucose removed extracorporeally (Me) within the observation phase is then obtained by integration of the flux
The difference between the total amount of glucose injected (M) and Me describes the intracorporeal disposal of glucose as
To account for different amounts of glucose injected it is useful to normalize Mi and to define the fraction of glucose disposal FM as
Notice that no blood samples are required for the measurement of Je, Me, or FM.
The relationship between different variables as well as between identical variables obtained in subsequent measurements was examined by linear regression analysis, calculation of Pearson correlation coefficients, and Fisher's r to z transformation using StatView 4.5 software (Abacus Concepts Inc., Berkeley, CA). A p value <0.05 was considered significant to reject the null hypothesis. Data are presented as means ± standard deviation.
Test results from duplicate measurements were available in four female and five male patients so that data from 18 treatments entered the final analysis. The patient (N = 9) and treatment (N = 18) data are summarized in Table 1.
Blood Side Data
The average time course of arterial glucose concentration (cart) before and after the injection of glucose measured is shown in Figure 1. In the observation phase that followed the infusion, the gradient between arterial line and dialyzer inlet concentrations (cart − cd,in) followed an almost perfect exponential decay with a rate constant of 0.047 ± 0.018 min−1. The average glucose gradient was 7.3 ± 1.8 mmol/L providing an average solute flux Jb of 0.96 ± 0.31 mmol/min (Table 2).
The glucose distribution volume V was determined as 10.3 ± 2.8 L. The total clearance K of glucose was therefore derived as 458 ± 142 mL/min. Extracorporeal clearance Ke was measured as 131 ± 19 mL/min and correlated with target body mass (r = 0.51, p < 0.05) (Table 3). Thus, 30.9 ± 9.4% of total clearance was attributed to extracorporeal clearance. Intracorporeal clearance Ki (Equation 7) was then computed as 327 ± 137 mL/min, or 69.1 ± 9.4% of total clearance (Table 2). The intracorporeal rate constant ki was 0.033 ± 0.017 min−1. Overall clearance K, intracorporeal clearance Ki, intracorporeal rate constant ki, and the fraction of intracorporeal clearance FK were independent of extracorporeal clearance, body mass, or body mass index (Table 3).
Dialysate Side Data
The average time course of glucose concentration in dialysate outflow (cd,out) is shown in Figure 2. Although 204 ± 51 mmol of glucose were injected at time t = 0, 80 ± 20 mmol of glucose were removed extracorporeally (Me) during the following hour of dialysis (Equation 13) (Table 4). The difference of 124 ± 41 mmol was retained intracorporeally (Mi, Equation 14). Me (r = 0.88, p < 0.001) and Mi (r = 0.96, p < 0.001) were measured with high reproducibility. The absolute amount removed Me correlated with target body mass (r = 0.63, p < 0.01) and body mass index (r = 0.75, p < 0.001) (Table 3) because glucose load delivered to the patients was scaled to target body mass (see Material and Methods). For the same reason, the amount retained Mi also correlated with target body mass (r = 0.92, p < 0.001) and body mass index (r = 0.76, p < 0.001). Therefore, to account for variable glucose dose, the amount retained was normalized (FM, Equation 15) and determined as 60.1 ± 10.5%. This measure was reproducible (r = 0.79, p < 0.01) and independent of extracorporeal clearance Ke, target body weight, or body mass index (Table 3).
Blood Versus Dialysate Side Data
The amount of glucose appearing in the dialysate Me was largely determined by the average glucose gradient cart − cd,in (Figure 3). The relationship between Me and solute removal was further improved when accounting for effects of extracorporeal clearance Ke and measuring the average solute flux Jb leaving the blood (Equation 4, Figure 4). Fraction of clearance and fraction of glucose disposal covered a considerable range from 0.50 to 0.81 and 0.35 to 0.72, respectively. However, comparison of these measures evaluated within the same test showed a strong linear relationship (FK = 0.19 + 0.84 × FM 0.19, r2 = 0.88, p < 0.001) (Figure 5).
In this study, a standardized glucose load was delivered during hemodialysis to quantify the absolute and relative magnitudes of intra- and extracorporeal clearance of glucose in non-diabetic hemodialysis patients. It was shown that intracorporeal clearance accounted for ∼70% of total glucose clearance. This fraction was highly correlated to the fraction of glucose disposed intracorporeally derived from dialysate measurements without taking blood samples. Although there have been several studies on glucose kinetics done before and after hemodialysis,11,12 this is the first study done during hemodialysis.
The decline in glucose concentration after the infusion of a glucose load is caused by different concomitant processes: the distribution of glucose across two or more compartments, insulin-dependent and insulin-independent uptake into cells, insulin-mediated reduction of hepatic glucose production, and renal clearance.13–15 Thus, the description by a monoexponential decline and a single parameter can only provide a crude estimate of glucose tolerance. During hemodialysis, however, an additional process has to be considered.
There is substantial extracorporeal clearance during hemodialysis expected to accelerate the elimination of glucose that follows the injection of the glucose bolus. Indeed, the rate constant of arterial glucose decay k was in the range of 0.047 ± 0.023 min−1. When corrected for the contribution of extracorporeal clearance (Equation 10), the rate constant was reduced to ki = 0.033 ± 0.017 min−1. This was comparable with results obtained by Ferrannini et al.11 where a similar approach was used to analyze glucose assimilation in uremic patients before (0.047 ± 0.006 min−1) and after (0.035 ± 0.007 min−1) hemodialysis, thereby confirming our approach. The trend for a small reduction of intracorporeal glucose disposal observed in our study could be explained by decreased splanchnic blood flow known to occur during hemodialysis and ultrafiltration.16,17
The initial distribution volume of glucose is assumed to represent the extracellular fluid volume of central and highly perfused organs. In a recent study done in 150 intensive care patients, this volume was estimated as 7.24 ± 1.63 L.18 The larger value of 10.3 ± 2.8 L found in our study is consistent with increased values with thoracic fluid accumulation19 and could be explained by the volume excess in dialysis patients at the beginning of the dialysis treatment. Volume excess is an important issue in end stage renal disease and acute renal failure and the initial distribution volume of glucose could be of additional interest in this group of patients.
The rate constants and coefficients of glucose assimilation obtained in this study were derived from arterial concentrations. This is characteristic for hemodialysis when using an arteriovenous access8 and different from the usual situation where rate constants and coefficients of glucose assimilation refer to venous concentrations. Arterial glucose concentrations measured during hemodialysis refer to mixed venous outflow from all tissues and are likely to differ from venous concentrations taken from a peripheral vein representing the outflow of a specific tissue.
Renal clearance of glucose is negligible or totally absent in patients with end stage renal disease. Even with a residual renal clearance of 15 mL/min considered a threshold to commence dialysis, the contribution of renal clearance will be one order of magnitude smaller than extracorporeal clearance. During hemodialysis, however, glucose is cleared by the extracorporeal system, and the removal bears similarities with renal glucose handling. For example, a high plasma glucose concentration relative to the concentration in the dialysate provides a positive glucose gradient and causes a glucose flux Jb from the blood into the dialysate (Equation 4, Figure 3). In this case, the loss of glucose into the dialysate is akin to glucose spillover in the kidney when exceeding the tubular maximum. Thus, with arterial glucose concentrations reaching 25 mmol/L the maximum extracorporeal glucose flux was in the range of 2.5 mmol/min or 0.5 g/min. This is a significant fraction of overall glucose disposal. On the other hand, when plasma glucose is lower than dialysate glucose, there will be an influx of glucose into the patient. Substantial glucose influx into the normoglycemic patient is also expected with a dialysate glucose concentration of 11 mmol/L, still customary in many dialysis units. The resulting influx can be expected to have a major effect on intra- and extracorporeal disposal should a glucose load be applied under such conditions. It is therefore important to notice that the results reported in this study were obtained with a dialysate glucose concentration of 5.5 mmol/L. Also, notice that for measuring the proper glucose gradient across the dialyzer membrane, arterial plasma concentrations have to be corrected for a blood water fraction of fpw = 0.93.
The amount of glucose appearing in the dialysate over some period of time is an indirect measure of the average plasma concentration (Equation 4, Figures 3 and 4). An excessive and/or prolonged elevation of arterial glucose concentrations secondary to impaired intracorporeal glucose disposal is therefore likely to increase extracorporeal glucose flux Jb and extracorporeal glucose elimination. We hypothesize that the fraction of glucose recovered in the dialysate will be increased under such circumstances and that this fraction could be an indirect measure of impaired glucose tolerance in hemodialysis patients. This fraction was independent of extracorporeal clearance but it remains to be studied whether this result can be extrapolated to treatments with much higher efficiencies.
In conclusion, the intracorporeal clearance of glucose during hemodialysis after a standardized glucose load can be determined from the difference of overall and extracorporeal glucose clearance determined by standard techniques. When the amount of glucose removed by dialysis is compared with the total load, ∼40% of glucose administered intravenously is removed by hemodialysis, whereas 60% are disposed intracorporeally. This parameter is measured with high reproducibility and without blood sampling and could be helpful to identify patients with impaired glucose metabolism and reduced intracorporeal glucose disposal. Such measurements could be done during the regular hemodialysis with little interference on treatment and patient routine.
1. 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.
2. 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.
3. Ritchie-McLean S, Kirwan C, Levy JB: Is there a role for intensive insulin therapy in patients with kidney disease? Am J Kidney Dis
50: 371–378, 2007.
4. Uzu T, Hatta T, Deji N, et al
: Target for glycemic control in type 2 diabetic patients on hemodialysis: Effects of anemia and erythropoietin injection on hemoglobin A1c. Ther Apher Dial
13: 89–94, 2009.
5. Riveline JP, Teynie J, Belmouaz S, et al
: Glycaemic control in type 2 diabetic patients on chronic haemodialysis: Use of a continuous glucose monitoring system. Nephrol Dial Transplant
24: 2866–2871, 2009.
6. Kazempour-Ardebili S, Lecamwasam VL, Dassanyake T, et al
: Assessing glycemic control in maintenance hemodialysis patients with type 2 diabetes. Diabetes Care
32: 1137–1142, 2009.
7. Depner TA: Refinements and application of urea modeling, in Depner TA (ed), Prescribing Hemodialysis: A Guide to Urea Modeling
. Boston/Dordrecht/London, Kluwer Academic Publishers, 1991, pp. 167–194.
8. Schneditz D, Kaufman AM, Polaschegg HD, et al
: Cardiopulmonary recirculation during hemodialysis. Kidney Int
42: 1450–1456, 1992.
9. Schneditz D: Glucose-added dialysis fluid prevents asymptomatic hypoglycaemia in regular haemodialysis. Nephrol Dial Transplant
23: 1066–1067, 2008.
10. Schneditz D, Hafner-Giessauf H, Holzer H, Thomaseth K: Intra- and extracorporeal clearance of glucose following an intra-dialytic glucose load, in Dössel O, Schlegel WC (eds), IFMBE Proceedings
. Munich, the World Congress on Medical Physics and Biomedical Engineering, 2009, 25(VII), pp. 855–858.
11. Ferrannini E, Pilo A, Tuoni M: The response to intravenous glucose of patients on maintenance hemodialysis: Effects of dialysis. Metabolism
28: 125–136, 1979.
12. Allegra V, Mengozzi G, Martimbianco L, Vasile A: Glucose-induced insulin secretion in uremia: Effects of aminophylline infusion and glucose loads. Kidney Int
38: 1146–1150, 1990.
13. Mari A, Stojanovska L, Proietto J, Thorburn AW: A circulatory model for calculating non-steady-state glucose fluxes. Validation and comparison with compartmental models. Comput Meth Programs Biomed
71: 269–281, 2003.
14. Thomaseth K, Pavan A, Berria R, et al
: Model-based assessment of insulin sensitivity of glucose disposal and endogenous glucose production from double-tracer oral glucose tolerance test. Comput Meth Programs Biomed
89: 132–140, 2008.
15. Brehm A, Thomaseth K, Bernroider E, et al
: The role of endocrine counterregulation for estimating insulin sensitivity from intravenous glucose tolerance tests. J Clin Endocrinol Metab
91: 2272–2278, 2006.
16. Jakob SM, Ruokonen E, Vuolteenaho O, et al
: Splanchnic perfusion during hemodialysis: evidence for marginal tissue perfusion. Crit Care Med
29: 1393–1398, 2001.
17. Rigalleau V, Baillet L, Lasseur C, et al
: Splanchnic tissues play a crucial role in uremic glucose intolerance. J Renal Nutr
13: 212–218, 2003.
18. Hirota K, Ishihara H, Tsubo T, Matsuki A: Estimation of the initial distribution volume of glucose by an incremental plasma glucose level at 3 min after i.v. glucose in humans. Br J Clin Pharmacol
47: 361–364, 1999.
19. Ishihara H, Suzuki A, Okawa H, et al
: Comparison of initial distribution volume of glucose and plasma volume in thoracic fluid-accumulated patients. Crit Care Med
29: 1532–1538, 2001.