Euglycemic Infusion of Insulin Detemir Compared With Human Insulin Appears to Increase Direct Current Brain Potential Response and Reduces Food Intake While Inducing Similar Systemic Effects

According to a single-blind, within-subject comparison, 15 healthy, normal-weight men (mean age ± SE, 28.5 ± 1.0 years; BMI, 23.1 ± 0.5 kg/m²) participated in two hyperinsulinemic-euglycemic clamp experiments (human insulin, detemir) spaced apart at least 1 week. The order of conditions was balanced across subjects. All subjects gave written informed consent to the experiments, which were approved by the local ethics committee. In both experimental conditions, an insulin bolus injection (17.75 mU/kg body wt [= 0.1065 nmol/kg body wt] human insulin, Insulin Actrapid; Novo Nordisk, Bagsværd, Denmark; 90 mU/kg body wt [= 2.16 nmol/kg body wt] detemir, Insulin Levemir; Novo Nordisk) was given followed by a steady 90-min infusion of 1.0 mU · kg⁻¹ · min⁻¹ (= 0.006 nmol · kg⁻¹ · min⁻¹) human insulin versus 2.0 mU · kg⁻¹ · min⁻¹ (= 0.048 nmol · kg⁻¹ · min⁻¹) detemir. To compensate for the relatively slower onset of the action of detemir, a higher dosage was chosen (15,17) to induce comparable peripheral effects of both compounds as assessed by rates of glucose infusion necessary to keep blood glucose levels in the euglycemic range.

Volunteers reported to the laboratory at 0800, after an overnight fast of 10 h, and were prepared for electroencephalographic recordings, insulin infusions, and blood samplings. We inserted venous cannulas into the subject's arms and connected them to tubes enabling infusion and blood sampling from an adjacent room without awareness of the subject. The arm from which samples were taken was positioned in a heated box (55°C) to enable drawing of arterialized venous blood. During recordings, subjects sat in a reclining chair in a sound-attenuated room of constant temperature, with their heads stabilized by a cushion. They were instructed to relax and not to move during recordings and to fixate their gaze on the wall in front of them. Subjects pressed a button every estimated 30 s to maintain a constant state of mental activity and to not doze off.

Recordings of DC potentials started at 0940 with a baseline phase of 20 min followed by the bolus injection of human insulin and detemir, respectively (t = 0). Subsequent insulin infusion lasted for 90 min, ending at 1130 when EEG recordings were also stopped. Arterialized blood was drawn at 15-min intervals during baseline, at 5-min intervals during insulin infusion, and at ∼15-min intervals thereafter to monitor blood glucose concentration (HemoCue B-Glucose-Analyzer, HemoCue AB, Angelholm, Sweden). During and after insulin administration, subjects intravenously received a 20% glucose solution at a variable rate to maintain normal plasma glucose levels. Blood samples for the determination of hormonal parameters were repeatedly collected, and routine assays were used to determine concentrations of serum C-peptide, plasma ACTH, serum cortisol (all Immulite; DPC, Los Angeles, CA), plasma glucagon (RIA; Adaltis, Montreal, Quebec, Canada), and serum leptin (RIA; Linco Research, St. Charles, MO).

At 1150 (i.e., 20 min after insulin infusion had stopped), a standardized buffet of around 4,650 kcal was offered from which subjects were allowed to eat ad libitum during the subsequent 50 min (Table 1). Subjects were kept unaware of hypothesized treatment effects on food intake and were not aware that their food intake was measured by weighing buffet components before and after food intake. In addition, to prevent overeating, subjects were allowed to take with them any remaining food afterward. Before (at 0915) and after (at 1135) recordings and at the end of the session at ∼1300, subjects rated their hunger, thirst, and tiredness on 10-point scales and completed a questionnaire assessing alertness and autonomic symptoms on 5-point bipolar scales of 20 contrasting adjective pairs (e.g., activated-inert and sweating-shivery) (22). They also filled in a checklist of 161 adjectives assessing mood on 14 dimensions (23).

Standard recordings of DC potentials, electro-oculogram, and electromyogram were performed as described previously (24,25). DC-potential recordings were obtained from left and right frontal (F3, F4), frontocentral (FC3, FC4), and central (C3, C4) electrodes referenced to linked electrodes at the mastoids. A BrainAmp DC amplifier (Brain Vision, London, U.K.; low-pass filter: 30 Hz, sampling rate: 200 Hz) was used. DC-potential drifts with short-circuited input were constantly <5 μV/h, and electrode impedance, measured before and after recordings, never exceeded 5 kΩ. Average DC-potential values were determined offline for subsequent 5-s intervals. Linear potential drifts during the 20-min baseline period extending into the 90-min insulin infusion period were removed using a linear regression method. Periods where electromyogram or electro-oculogram indicated increased muscular activity or eye movements were excluded from analysis. The average DC potential during baseline was set to 0 μV, and potential shifts during treatment were expressed as difference values.

Differences in DC-potential values between conditions were evaluated first on an exploratory basis by point-wise comparisons using t tests to identify time ranges with most consistent differences (26). The time range selected for analysis covered 21–90 min of insulin infusion. Values for this time interval were then subjected to ANOVA, including the repeated-measures factors Treatment (human insulin vs. detemir) and Topography (electrode locations). Degrees of freedom were corrected using the Greenhouse-Geisser procedure. Post hoc contrasts were used to specify significant ANOVA main effects and interactions. For the DC-potential analysis, data from four subjects had to be excluded because of technical failures and artifacts of apparent nonbiological origin. Behavioral measures and hormonal parameters were analyzed with ANOVA and paired t tests as appropriate. P ≤ 0.05 was considered significant.

Rates of glucose infusion to maintain euglycemia were similar between conditions (Fig. 1,A; P > 0.14 for all comparisons), resulting in identical total amounts of energy supplied via glucose infusion (human insulin, 240.8 ± 23.3 kcal; detemir, 239.5 ± 25.0 kcal; P > 0.93). Correspondingly, blood glucose concentrations were comparable throughout the experiment (Fig. 1,B; P > 0.34 for all comparisons) as well as when restricting analysis to the period of insulin infusion (P > 0.13). Serum C-peptide concentrations did not differ between conditions during insulin infusion (P > 0.25) but rose to slightly elevated levels in the human insulin compared with the detemir condition at the end of experiments (Fig. 1,C; F[3,46] = 3.77, P < 0.02 for Treatment × Time). Concentrations of serum glucagon (Fig. 1 D) as well as leptin did not differ between conditions (all comparisons, P > 0.40). Concentrations of plasma ACTH (P > 0.57) and serum cortisol (P > 0.71) were likewise similar in both conditions, with cortisol levels showing the expected meal-related increase (F[3,47] = 4.19, P < 0.008).

The DC potential in the detemir condition showed a marked negative shift shortly after insulin injection that reached maximum values exceeding −600 μV toward the end of the recording epoch (Fig. 2). This strong negative DC shift was generally absent in the human insulin condition. Analyses of average DC-potential levels during the relevant interval from 21 min after the start until the end of insulin infusions confirmed a distinctly more negative potential level in the detemir than human insulin condition (F[1,10] = 7.03, P = 0.02; Table 2), with this effect displaying an even topographic distribution (F[2,24] = 0.59, P > 0.59, for Treatment × Electrode location). Comparisons with preinjection baseline levels confirmed significance for the negative DC potential in the detemir condition during 21–90 min of insulin infusion (F[1,10] = 8.53, P = 0.02; F[3,27] = 1.04, P > 0.39 for Electrode location). Although DC potentials in the human insulin condition appeared to shift slightly toward positive values over central positions, respective analyses did not yield significant differences from baseline values (F[1,10] = 0.17, P > 0.43; Table 2).

Table 3 summarizes treatment effects on food intake. Detemir in comparison with human insulin significantly reduced food consumption by 303 ± 135.7 kcal. Macronutrient comparisons suggested this effect was particularly pronounced for protein and, to a lesser extent, for carbohydrate intake, but there was no significant statistical interaction between the factors Treatment and Macronutrients (F[2,24] = 1.63, P > 0.22). The reduction in total energy consumption and also carbohydrate intake by detemir in comparison with human insulin was also observed in analyses taking into account the energy received in the form of intravenous glucose (Table 3). Differences in carbohydrate and protein intake between conditions correlated significantly with respective differences in DC-potential shifts (averaged over all recording sites) from 21 to 90 min of infusions (r = 0.74, P < 0.04, and r = 0.87, P < 0.01, respectively, bivariate Pearson coefficients). The respective correlation with total energy intake failed to reach significance (r = 0.61, P = 0.11; r = 0.21, P > 0.62 for fat intake).

Ten-point scale hunger ratings did not differ between conditions, showing the expected decline from baseline (detemir, 4.87 ± 0.68; human insulin, 5.93 ± 0.54; P > 0.11) and post-EEG recording values (5.87 ± 0.62 vs. 5.93 ± 0.56, P > 0.91) to low postingestion levels (1.47 ± 0.43 vs. 1.00 ± 0.24, P > 0.33; P > 0.12 for Treatment × Time). A comparable pattern without significant differences between conditions was observed for thirst (all P > 0.12) and tiredness (P > 0.33) ratings. According to the bipolar questionnaire, subjects in the detemir compared with the human insulin condition felt more critical (critical-comfortable, 2.87 ± 0.13 vs. 3.27 ± 0.18, P < 0.009) and reported increased wakefulness (sleepy-awake, 3.33 ± 0.23 vs. 2.73 ± 0.23, P < 0.03) and hunger (full-hungry, 4.20 ± 0.20 vs. 3.53 ± 0.22, P < 0.01) immediately after infusions. The remaining dimensions and the post–food intake assessment were not affected. The mood adjective checklist did not yield significant differences between conditions for any of the subscales.

Administration of insulin to the central nervous system reduces food intake (3,5) and body weight (4,6). We demonstrate that while eliciting comparable peripheral effects, euglycemic infusion of insulin detemir compared with human insulin triggers a distinct negative shift in EEG DC-potential recordings and reduces calorie uptake in healthy men, supporting our hypothesis that detemir affects brain functions to a greater extent than human insulin and induces stronger anorexigenic effects on central nervous networks that control food intake. This outcome suggests that enhanced catabolic insulin signaling to the brain may be an important mechanism behind the limitation of weight gain observed in diabetic patients receiving detemir treatment (9,11,,–14).

In accordance with other investigators (15,17), we administered insulin doses that were considerably higher in the detemir than in the human insulin condition to compensate for the delayed onset that human insulin–equimolar detemir dosages would display. Accordingly, timing and strength of the effects of detemir and human insulin on systemic glucose homeostasis as reflected by the rates of glucose infusion as well as blood glucose and serum glucagon concentrations were identical. C-peptide concentrations likewise did not differ between conditions during insulin infusion, merely showing a slight increase in the human insulin condition after the test buffet that may have been due to greater food intake in this compared with the detemir condition. Congruent peripheral effects were also indicated by comparable serum leptin concentrations that are known to respond to insulin infusion (27,28).

On the background of equipotent systemic effects, detemir elicited a marked brain response as indicated by a widespread negative shift of scalp-recorded DC potentials that started around 15 min after detemir bolus injection and exceeded potential levels in the control condition that remained unaffected by human insulin infusion. In the human insulin condition, subjects received roughly the same total amount of insulin that, when administered in single bolus form, induced a negative DC-potential shift in foregoing experiments (21). Slowly infused over the course of 90 min, this dose obviously was too weak a stimulus to evoke DC-potential responses to human insulin in the present experiments. In contrast, the 90-min infusion of a detemir dose equivalent in terms of systemic action triggered a sustained negative DC-potential shift comparable with the previously reported effect of human insulin administered in high-dose bolus form (21). This pattern indicates that although central nervous detemir effects may be mimicked by disproportionally high doses of human insulin, the relative impact of detemir on brain functions is considerably greater when both insulins are administered at doses with similar peripheral impact.

The mechanisms behind the strong effect of detemir on scalp-recorded DC potentials cannot be derived from our data. Brain DC-potential shifts of this amplitude most likely reflect changes in extracellular ionic concentrations stemming from potential shifts at glial membranes that are endowed with receptors for insulin and IGF (21,29,–31). The assumption of a widespread effect on cerebral cellular networks also fits with the global nature and long duration of the DC-potential shifts induced by detemir infusion. Our observations corroborate previous findings of increased brain responses to detemir in comparison with human insulin in animals (32) and humans (15,–17). Superior central nervous efficacy of detemir may be due to improved permeation of the lipophilic molecule into the brain compartment, with enhanced receptor-mediated blood-brain barrier transport of albumin-bound detemir adding to this effect (15,32). The detemir-induced negative DC-potential shift that is presumably of primary glial origin thus may be reinforced by electrical potentials generated in the course of receptor-mediated detemir transport across the blood-brain barrier (33).

A most remarkable finding of our study is the reduction of ad libitum food intake by around 300 kcal in the detemir compared with the human insulin condition in the presence of identical peripheral actions of both insulins. This pattern renders the contribution of systemic mediators to this effect highly unlikely, rather suggesting that enhanced central nervous insulin signaling in the detemir condition resulted in decreased caloric intake. The concept of insulin providing catabolic feedback on the body's energy resources to brain networks that control energy homeostasis has been well established in animals (2,3) as well as in humans (5,6). Thus, the decrease in food intake after detemir compared with human insulin infusion suggests that the enhanced effect on brain functions indicated by the negative DC-potential shift particularly impacts the central nervous control of food intake. Although because of methodological constraints, DC-EEG could not be recorded during food intake proper (34), this interpretation is supported by the strong correlation between the reduction in calorie intake and the antecedent DC-potential effect elicited by detemir. It is also in line with animal experiments in which intravenous detemir compared with human insulin injections were associated with enhanced insulin receptor phosphorylation in hypothalamic and cerebrocortical tissue in conjunction with increased EEG-assessed cortical activity, whereas the activation of the insulin receptor signaling cascade was similar in muscle tissue and liver (32). Interestingly, detemir compared with human insulin infusion increased rather than decreased self-rated hunger before test buffet presentation, which may have been due to a biasing influence of enhanced wakefulness and activation after detemir infusion (35). Alternatively, this finding may also indicate that central nervous insulin exerts its anorexigenic effects via meal-related signals that contribute to the termination of a meal, but not by affecting hunger motivation per se (5).

Weight gain is a frequent side effect when blood glucose levels of diabetic patients are normalized by insulin (36,37), but is less pronounced in patients undergoing detemir therapy (9,11,,–14). The assumption that the use of detemir limits weight gain because its favorable safety profile decreases defensive snacking to prevent hypoglycemia was not supported by comparative clinical studies showing that insulin glargine, which like detemir reduces the risk of hypoglycemia, is associated with greater weight gain than detemir (38,39). Against this background, the anorexigenic brain impact of detemir found in our study rather suggests the contribution of central nervous mechanisms to the weight-sparing effect of detemir in the clinical context.

Supported by Deutsche Forschungsgemeinschaft (KFO 126/B5) and Novo Nordisk. The funding source had no input in the design and conduct of this study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the article.

M.H. and B.S. have received honoraria for speaking engagements by Novo Nordisk. W.K. has received grant support, honoraria for speaking engagements, and consulting fees for participating on scientific boards of Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.

We thank Katrin Dieckmann and Nina Eggers for their expert assistance and Ingrid von Lützau, Martina Grohs, and Heidi Ruf for their invaluable laboratory work.