Glucose Levels at the Site of Subcutaneous Insulin Administration and Their Relationship to Plasma Levels

The study was conducted in five healthy male adult volunteers. The age of the subjects was 29.8 ± 4.0 years (range 19–44) and the BMI was 23.0 ± 0.9 kg/m² (20.7–26.0). Written informed consent was obtained after the purpose, nature, and potential risks of the study were explained to the subjects. The study was approved by the ethics committee of the Medical University of Graz.

In the morning after an overnight fast, study subjects were admitted to the clinical research center. At ∼7:30 a.m., an intravenous catheter was placed into an arm vein for glucose infusion. Another catheter was inserted into a vein of the contralateral hand for blood sampling. The forearm with the sampling catheter was then placed in a thermoregulated box (55°C) to ensure arterialization of the venous samples. Subsequently, a 24-gauge microperfusion (MP) catheter (7,8) and a microdialysis (MD) catheter (CMA60, modified as described below; CMA/Microdialysis, Solna, Sweden) were inserted into the periumbilical subcutaneous adipose tissue, and peristaltic pumps (Minipuls 3; Gilson, Villiers-le-Bel, France) were attached to the inflow and outflow tubing of each catheter (Fig. 1). Both catheters were then sequentially perfused with an insulin-free solution (5% mannitol; Fresenius Kabi, Graz, Austria) for 2 h, a standard human insulin preparation (100 units/ml Actrapid; Novo Nordisk, Bagsvaerd, Denmark) for 6 h, and, again, an insulin-free solution (5% mannitol) for 2 h. During the 6-h insulin perfusion period, the insulin delivered to the tissue by the MP catheter was at a rate of ∼0.33 units/h over the first 2 h, ∼0.66 units/h over the subsequent 2 h, and ∼1.00 units/h over the final 2 h. Similar insulin delivery rates, but in reverse chronological order, were administered via the MD catheters. As the outflow rates of the two catheters were maintained at a constant value (∼0.45 μl/min) throughout the experiment, the insulin delivery rate of each catheter was adjusted by simply adjusting the inflow rate of the catheter (dual-pump operation mode) (9) (Fig. 1). During the experiment, catheter effluents were continuously collected in 30-min fractions in vials (Microvial; CMA/Microdialysis) kept on ice and covered to avoid fluid evaporation. In parallel, glucose concentrations in blood plasma were measured frequently (i.e., every 5–30 min), and, to prevent a fall in the blood glucose concentration during and after the subcutaneous insulin delivery period, glucose (0.1 g/ml; Fresenius Kabi) was intravenously infused at a variable rate.

MP and MD catheters applied were of concentric design with a cylindrical inner and outer tube (Fig. 1,B and C). The outer tube of the MP catheter (7,8) consisted of a conventional, intravenous 24-G cannula (shaft length: 19 mm, Neoflon; Becton Dickinson, Helsingborg, Sweden) in which 27 perforations (each 0.3 mm in diameter) were formed in the cannula wall using an Excimer laser (LZH, Hannover, Germany). Two 750-mm lengths of Tygon tubing (outer diameter: 2.0 mm, inner diameter: 0.19 mm; Cole-Parmer, Vernon Hills, IL) were used to connect catheter inlet and outlet with perfusate reservoir and sampling vial, respectively. In the case of the MD catheter (CMA60), part of the outer tube was made from a polyamide membrane with a molecular weight cutoff of 20 kDa (membrane length: 30 mm, pore sizes: ∼1 μm). To operate MD catheters in the dual pump mode (Fig. 1), the original inflow and outflow tubing of the catheters were replaced by two 750-mm lengths of Tygon tubing. In addition, to increase the insulin delivery efficiency of MD catheters, a fluid communication between the distal inner tube end and the outer surface of the catheter tip was established by forming a perforation into the catheter tip using a 30-G syringe needle (point length: 1.2 mm; Microlance, Becton Dickinson, Drogheda, Ireland). To monitor the outflow rate of the MP and MD catheters and to determine the exact sampled effluent volume, the sample vials were weighed before and after sample collection. The effluent sampling delay time introduced by the dead space volume of the catheter outflow tubing was taken into account when the sample collection was begun.

Plasma glucose concentrations were measured at the bedside using a Beckman Glucose Analyzer II (Beckman Instruments, Fullerton, CA) with a coefficient of variation (CV) of 2%. The glucose concentrations in catheter effluents were determined using an automated CMA600 analyzer (CMA/Microdialysis) with a within-run CV of 2%. Separate in vitro experiments have shown that preservatives contained in the regular human insulin solution used to perfuse the catheters (Actrapid) do not compromise the accuracy of the CMA600 glucose analyzer. The conductivities in the plasma, perfusate, and catheter effluent samples were measured using a contactless conductivity detector (TraceDec; I.S.T., Strasshof, Austria). The electrical conductivity was determined with a within-run CV of <1%.

The ISF glucose concentration was calculated (7,8) as the glucose concentration in the catheter effluent sample divided by the glucose recovery of the catheter (R). The recovery was determined for each sampling period as R = (C_out − C_in)/(C_pl − C_in), where C_in, C_out, and C_pl are the measured electrical conductivity in the perfusate, the effluent sample, and the corresponding plasma sample, respectively. Application of this technique (ionic reference technique) (7,8) was possible because the electrical conductivity in the used catheter perfusates was either negligible (mannitol) or low compared with that in blood plasma (Actrapid: ∼1.7% of average C_pl). To facilitate comparison of glucose time courses observed in the tissue ISF with those determined in the blood plasma, we calibrated the MP-derived and MD-derived ISF glucose values against the mean plasma glucose concentration observed during the 2-h baseline period (basal calibration). These glucose values (termed tissue glucose concentrations) and the corresponding plasma glucose levels were then used to calculate the time courses of the tissue-to-plasma glucose ratio (TPR). The differences of mean glucose and TPR results were examined by one-factor repeated-measures ANOVA. If significance was achieved, post hoc comparison of means with Dunnett's test was performed. A P value <0.05 was considered to indicate statistical significance. Normality of data were assessed using normal probability plots. Data are presented as means ± SE. Data analysis was performed using MATLAB (MathWorks, Natick, MA) and SPSS (SPSS, Chicago, IL) software packages.

Figure 2,A shows the time courses of the glucose concentration in arterialized blood plasma and in the tissue surrounding the MP and MD catheters. Commencement of insulin delivery with the MD and MP catheters induced a fall in the tissue glucose concentration for ∼60 min (P < 0.05). However, during the rest of the 6-h period of variable insulin delivery, the tissue glucose concentration paralleled the glucose concentration observed in plasma, thereby indicating that insulin's effect on the tissue glucose concentration saturated within 60 min after the start of insulin administration. After switching back from insulin to mannitol perfusates, tissue glucose concentration increased slowly and reattained plasma glucose levels at the end of the experiments (Fig. 2 A).

Figures 2 B and C depict the TPR for the MD and MP catheters as a function of time. As can be seen, during the first ∼60 min of insulin delivery, the TPR declined to a level of 83.8 ± 3.6 and 78.5 ± 5.0% for MD and MP catheters, respectively, and then remained at these levels until the end of the 6-h period of variable insulin delivery. During the final 5 h of the insulin delivery period, the TPR levels for MD and MP catheters averaged 83.2 ± 3.1 and 77.1 ± 4.8%, respectively.

Taken together, the temporal pattern of change in tissue and plasma glucose and TPR values suggests that within 60 min after exposing adipose tissue to a standard 100 units/ml insulin preparation, insulin's effect on the tissue glucose concentration saturates and a stable ratio between the tissue and plasma glucose concentration is attained.

At the used outflow rate of ∼0.47 μl/min, the transport efficiency (recovery) of glucose for the MP catheter was found to be 25.1 ± 5.4% during the first 2 h of insulin delivery and remained unchanged during the subsequent 2 h and final 2 h of insulin delivery (24.6 ± 4.4 and 24.8 ± 4.9%, respectively; P = 0.93, one-factor repeated-measures ANOVA). Similarly, recovery for the MD catheter was 96.8 ± 1.2% during the first 2 h of insulin delivery and remained at this level during the subsequent 2 h (98.0 ± 1.3%) and final 2 h (98.1 ± 1.3%) of insulin delivery (P = 0.20, one-factor repeated-measures ANOVA). The lower glucose recovery for MP catheters compared with MD catheters may be mainly attributable to the different size of the exchange area of the applied catheters (MP perforations distributed over an 11-mm shaft length versus MD membrane length of 30 mm). Figure 2 B and C show the average time course of the total insulin delivery rate (i.e., the sum of the convective and diffusive portions of the insulin delivery rates) for the MD and MP catheters (see also online appendix supplementary methods).

The time course of the intravenous glucose infusion rate is shown in online appendix Fig. 2. As can be seen, the glucose infusion rate needed to maintain euglycemia gradually increased during the second half of the 6-h insulin delivery period and reached a peak value of 1.75 ± 0.31 mg · kg⁻¹ · min⁻¹ 1 h after termination of insulin delivery. Glucose infusion was required until the end of the experiments.

The present investigation was undertaken to assess insulin's effect on the tissue glucose concentration at the site of subcutaneous insulin administration. For this reason, we subcutaneously administered insulin at variable rates and simultaneously sampled ISF glucose directly from the administration site. To carry out ISF glucose sampling and simultaneous insulin administration at the same adipose tissue site, we utilized MP and MD catheters together with standard insulin preparations as catheter perfusates and used two peristaltic pumps to operate each catheter (Fig. 1). This technique of catheter operation allowed the withdrawal of ISF at a constant rate as well as the simultaneous adjustment of the insulin delivery rate during the experiments.

When insulin delivery to the adipose tissue of the healthy humans was started at basal rates similar to those used in the replacement therapy of diabetic patients, there was an initial delay of ∼60 min before the effect of insulin on the tissue-to-blood glucose concentration gradient at this delivery site reached its maximum (Fig. 2). This observed delay in insulin action (activation time) may be a reflection of the time required for the convectional and diffusional transport of insulin from the catheter to the surface of the insulin-sensitive fat cells surrounding the catheter as well as for the subsequent recruitment and activation of glucose transport proteins stimulating cell glucose uptake (10). After this activation time, despite changes in the insulin delivery rates, the tissue–to–blood glucose concentration gradient remained stable, thereby supporting our hypothesis that insulin's effect on the tissue glucose concentration at the insulin delivery site is saturated and attains steady-state values. This steady-state condition in insulin action at the insulin delivery site in turn allowed us to reliably estimate blood glucose concentrations from glucose concentrations measured in the ISF of this tissue site (11). After switching back from insulin to mannitol perfusates, the tissue glucose concentration and the tissue–to–blood glucose concentration gradient increased slowly and reattained preinsulin delivery levels by the end of the 2-h mannitol perfusion period (Fig. 2), thereby indicating that insulin's effect on the tissue glucose concentration vanished within 2 h after termination of the insulin delivery. Overall, the observed activation and deactivation kinetics of insulin action on the tissue–to–blood glucose concentration gradient in adipose tissue are compatible with previous findings in humans showing similar activation and deactivation kinetics of insulin's action on the forearm (12), leg (4), and peripheral (13) glucose uptake.

Conclusions in the present study differ from the recent study of Hermanides et al. (14), who reported no influence of insulin on the glucose concentration in the adipose tissue located near the insulin infusion site. The disparity between our results and those of Hermanides et al. may be attributable to the different experimental procedure used to assess the effect of insulin on the tissue glucose at the insulin administration site. Whereas the MP and MD catheters used in the present study functioned simultaneously as insulin delivery and ISF sampling instruments (single-port approaches), Hermanides et al. (14) used a dual-port approach consisting of a subcutaneously inserted insulin infusion catheter and an MD catheter, which had a membrane length of 25 mm (14,15) and which was inserted in subcutaneous tissue at a closest mean distance of 9 ± 2 mm from the infusion catheter. Thus, in contrast to our study in which ISF glucose was withdrawn from the insulin-exposed tissue layer surrounding the MD and MP catheter, Hermanides et al. (14) performed ISF glucose sampling from tissue regions that were at considerable distances from the insulin administration site. A previous study by Linde and Philip (16), using radiography of the radioactivity distribution of subcutaneously injected ¹²⁵I-insulin, indicated that when a bolus amount of 0.25 ml of a 40 unit/ml insulin solution (10 units) is administered subcutaneously, the maximal volume of distribution of the administered insulin may have an average cross-sectional area of ∼170 mm² or an average radius of ∼7 mm (when a near-spherical shape is assumed) (17). Furthermore, our recent study in diabetic subjects (11), using MP catheters for insulin delivery and simultaneous glucose sampling, showed that a subcutaneous administration of a bolus amount of ∼0.08 ml of a 100 unit/ml insulin solution (∼8 units) is causing a dilution of the ISF at the insulin delivery site, thereby decreasing the glucose concentration around the insulin delivery catheter during and shortly after the bolus delivery period. In comparison, in the study of Hermanides et al. (14), a bolus amount of ∼0.11 ml of a 100 unit/ml insulin solution (∼11 units) was administered. Apparently, no decrease in the effluent glucose during the bolus administrations and, thus, no dilution effect of the insulin solvent were observed. Therefore, a possible reason for not observing a significant effect of insulin and/or insulin solvent on the local tissue glucose concentration in their study may be that, due to the considerable length of the microdialysis membrane used (25 mm), the whole or most parts of the glucose-exchanging membrane, especially the membrane part near the catheter outlet, may not have been positioned in the insulin-exposed tissue layer surrounding the insulin infusion catheter. There have been other studies evaluating the effect of insulin on the glucose concentration in human adipose tissue by using either continuous glucose sensors (18,19) or microdialysis-based glucose sensing (19,–21). However, whereas we and Hermanides et al. (14) assessed the effect of supraphysiological insulin levels (i.e., 100 units/ml) on the adipose tissue glucose concentration, these studies evaluated the effect of physiological insulin levels (<150 μU/ml) on this glucose concentration. Furthermore, in contrast to the study of Hermanides et al. (14) and our study, where insulin was locally introduced into the ISF of subcutaneous adipose tissue, these studies invariably increased the insulin levels in the central circulation by intravenous insulin infusions (18,,–21) and/or by enhancing endogenous insulin secretion using intravenous glucose infusions (18,19). The results of these studies indicated that physiological changes in the insulin levels affect (19,21) or do not affect (18,20) the ISF-to-plasma gradient in human adipose tissue. The divergent findings may be partially attributable to technical and procedural differences in the performance of the studies (18).

Besides the effect of insulin on the fat cell glucose uptake, an additional mechanism potentially influencing the glucose concentration at the site of insulin delivery may be the local dilution of the ISF by the insulin solvent. We reasoned that if an increase in the perfusate flow fraction directed to the tissue (Fig. 1, thin straight arrows) is causing a dilution of the ISF surrounding the catheter, then there will be a decrease in the efficiency of exchange of solutes between the ISF and the perfusates of the two catheters. We found that changes in the tissue-directed flow rates did not cause changes in the exchange efficiency (recovery) of the two catheters, thereby suggesting that insulin solvent delivery at basal rates is not diluting the ISF in the tissue surrounding the catheters. This observation is in agreement with the results of our recent study in diabetic subjects (11), in which after changes in the basal delivery rates, no changes in the catheter recovery were observed. The relatively high blood flow per unit adipose tissue weight (20–30 μl · min⁻¹ · g⁻¹) (22) compared with the amount of insulin solution infused during basal delivery (0.1–0.2 μl/min) may be the likely reason for not observing a local dilution of the ISF when basal insulin is delivered.

Our present and recent study (11) may also provide the basis for the pursuit of a principal concept to the design of a single-port treatment system. This concept involves the integration of a continuous glucose sensor directly onto the shaft wall of an insulin infusion catheter. When connected to an insulin pump, the system may then permit simultaneous insulin infusion and glucose sensing. Available information on solute transport in ISF surrounding infusion catheters (which is governed by a combination of convection, diffusion, and tissue clearance) (17,23,24) may be used to guide the exact placement of the glucose sensor onto the shaft of an insulin infusion catheter. For example, when a typical bolus amount of 0.1–0.2 ml (10–20 units) of a standard insulin solution is administered through a conventional infusion catheter (diameter: ∼0.5 mm; shaft length: ∼10 mm), there may be a significant backflow of infusate from the catheter tip along the catheter shaft (24) toward the skin, so that the initial area of distribution of infused insulin (initial depot volume) (17,23) may correspond to a cylindrical tissue layer surrounding the catheter shaft (24) and having a width of a few millimeters (17). Shortly after bolus administration, the relative rates of diffusion and clearance may then produce first an expansion of the insulin-exposed tissue layer about the catheter shaft to a width of several millimeters (e.g., ∼7 mm) (16) and, as soon as clearance prevails, a decrease of the size of this layer, so that 6–10 h after bolus administration, all the insulin is cleared from the tissue site of insulin administration (17,21). Furthermore, in the case of insulin delivery at basal rates (∼0.01 ml/h = ∼1 unit/h, usually administered as microboluses with 0.025–0.1 units per pump stroke) (25), backflow distances may be shorter and hence the volume of insulin-exposed tissue about the catheter shaft may be smaller than after bolus administrations (24). Therefore, when a basal-bolus insulin regimen is applied, the best placement location of a continuous glucose sensor on the catheter shaft may be close to the catheter tip because at this site the tissue surrounding the sensor may be permanently exposed to insulin. In the case of suspension of the insulin pump treatment, a recalibration of the glucose sensor may be required when the length of pump suspension exceeds the deactivation time of insulin action. The deactivation time, however, may largely depend on the actual insulin depot size at the delivery site, which can be very high after bolus administrations or lower after prolonged basal insulin delivery (two to five times the hourly infusion rates) (25).

In summary, our results show that within 60 min after exposing adipose tissue to a standard 100 units/ml insulin preparation, insulin's effect on the tissue glucose concentration saturates and a stable ratio between the tissue and plasma glucose concentration is attained, thus indicating that insulin delivery and glucose sensing may be performed simultaneously at the same adipose tissue site via a single tissue catheter. This single-port treatment approach may hold great promise to simplify and improve glucose management in diabetes.

This work was supported in part by funding from Science Park Graz and the Federal Ministry of Economics and Labor of the Republic of Austria.

T.R.P. is a cofounder and shareholder of Smart*Med and a consultant of Novo Nordisk and sanofi-aventis. M.E. is a consultant of Sensile Medical and BBraun and a cofounder and shareholder of Smart*Med. L.S. and H.K. are cofounders and shareholders of Smart*Med. W.R., L.S., G.K., M.E., and T.R.P. have filed patent applications relating to the methodology described in this report. No other potential conflicts of interest relevant to this article were reported.

Parts of this study were presented in abstract form at the 2nd European Diabetes Technology and Transplantation Meeting, Innsbruck, Austria, 27–29 January 2008.

We are grateful to A. Wutte for her expert assistance in conducting the study, M. Suppan for performing insulin assays, A. Groselj-Strele for her assistance with statistical analysis, and all volunteers for participating in the study.