| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
3rd Medical Clinic (T.L., B.F., N.S., M.E., R.G.B.) and Institute of Nutrition (T.L., J.E., C.K.), Justus Liebig University Giessen, 35385 Giessen, Germany
Address all correspondence and requests for reprints to: Professor Thomas Linn, Clinical Research Unit, 3rd Medical Clinic and Policlinic, Rodthohl 6, 35385 Giessen, Germany. E-mail: thomas.linn{at}innere.med.uni-giessen.de.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Methods: In 10 type 2 diabetic patients, euglycemic clamps were performed, including [6,6']2H2 glucose, to study the rate of disappearance (Rd) and endogenous production (EGP) of glucose during the night. On separate days at bedtime (2200 h), patients received a sc injection of insulin glargine, NPH insulin, or saline in a randomized, double-blind fashion.
Results: Similar doses of both insulins had different metabolic profiles. NPH insulin had a greater effect on both Rd and EGP in the night compared with insulin glargine. By contrast, in the morning, insulin glargine was more effective, increasing Rd by 5.8 µmol/kg–1·min–1 (95% confidence interval 4.7–6.9) and reducing EGP –5.7 (–5.0 to –6.4) compared with NPH insulin. Nearly 80% of the glucose lowering effect in the morning was due to insulin glargine’s reduction of EGP. Its injection was associated with one-third lower morning glucagon levels compared with NPH insulin (P = 0.021).
Conclusion/Interpretation: Nocturnal variations of EGP and Rd explain the reduced incidence of hypoglycemia and lower fasting glucose levels reported for insulin glargine compared with human NPH insulin.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Insulin action defined as the maximum glucose infusion rate (GIRmax) was reached about 4 h later with glargine compared with NPH insulin in healthy subjects (11, 12). In patients with type 1 diabetes (T1DM), there was also a trend for a later maximum effect, but with large variations (13, 14). In T2DM, the physiological effects of sc administered long-acting insulin is subject to greater variability owing to insulin resistance and poor absorption (15).
Studies indicate that insulin glargine has a flatter postinjection profile compared with intermediate or NPH insulin, however, data are inconclusive owing to the lack of placebo groups controlling for variability (16, 17). If given at bedtime, the peak insulin activity would be reached at approximately 0300 h with NPH insulin, resulting in a high risk of nocturnal hypoglycemia. Therefore, it is important to determine whether there are differences between insulin glargine and NPH insulin overnight after bedtime administration.
The majority of studies have indirectly examined the effect of insulin on glucose disappearance, characterized by changes in glucose infusion rates (GIRs) to maintain blood glucose (BG) concentration at physiological concentrations (11, 12, 13, 14, 15, 16). However, these studies tend to overlook changes in the rate of glucose appearance (Ra), disappearance (Rd), and endogenous glucose production (EGP), which is normally suppressed by insulin, but enhanced in the fasting state.
Two reports in healthy male subjects have directly addressed the impact of insulin glargine on EGP and glucose balance (18, 19). These two studies showed that iv insulin glargine and regular human insulin have equipotent effects on glucose balance and suppressing EGP. They conclude that the specific effects of insulin glargine on glucose turnover are completely dependent on its absorption kinetics. Therefore, it remains to be investigated whether or not EGP changes at all after a single sc injection of insulin glargine. This is of particular interest in people with T2DM who frequently experience increased EGP in the morning (20).
With any insulin therapy, a clinically important issue is the prevention of hypoglycemia during the night. However, there are no studies available on the overnight action profile of NPH insulin or insulin glargine. Therefore, the aim of this study was to investigate the nocturnal metabolic effects in terms of glucose turnover after the administration of glargine or NPH insulin in patients with T2DM.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This was a randomized, placebo-controlled, double-blind, three-way cross-over study to investigate the metabolic effects of glargine compared with NPH insulin in patients with T2DM. The Ethics Committee of the University Hospital at Giessen approved the protocol, and the study was conducted according to the Declaration of Helsinki and the principles of Good Clinical Practice. All patients provided written informed consent before study entry.
Patients with T2DM treated with oral antidiabetic agents (OADs) alone or in combination with insulin and glycosylated hemoglobin (HbA1c) less than 10% were eligible for this study. Exclusion criteria included smoking, concomitant illnesses, taking concomitant medication other than that for treatment of diabetes, or hypersensitivity to any of the study medications or excipients.
At the first visit, patients were screened for eligibility, and baseline characteristics were recorded. Informed consent was obtained. Nurses or physicians provided assistance to the patients to help them optimize their glycemic control. In four patients the dosage of OADs or insulin was changed.
At visit 2, approximately 4 wk later, the patient’s glycemic control was reviewed, and the patients were randomized to their treatment sequence, and visit 3 was scheduled.
For visits 3–5, euglycemic-hyperinsulinemic clamps were performed according to their predefined randomization procedure. The patients and investigators were kept blind to the study medication, which was prepared and injected by a person who was not involved in the study design or analysis.
OADs were discontinued 1 d before the clamp. Patients were admitted to hospital at 1800 h, where they stayed for about 24 h. They remained fasted until completion of the clamp. They were allowed access to water throughout the study.
Euglycemic clamp procedure
The patients were fitted with different iv cannulae at the forearms to allow infusion of glucose/insulin and blood sampling. At 2000 h, the euglycemic clamp was initiated with a primed infusion of insulin (0.2 mU/kg–1·min–1; 1.4 pmol/kg–1·min–1) (Fig. 1
). Variable glucose infusion was started to maintain BG at 6 mmol/liter in conjunction with a primed, constant infusion (2 mg/min; 11.1 µmol/min) of [6,6']2H2 glucose (Campro Scientific, Berlin, Germany). During the insulin infusion, plasma glucose was measured at 5-min intervals and adjusted by a concomitant infusion of 20% dextrose, which was enriched with 4 mmol/mol [6,6]2H2 glucose to minimize changes in plasma glucose enrichments.
|
Plasma enrichment with [6,6]2H2 glucose was measured by gas-chromatography mass spectroscopy as previously described (22).
At 2200 h a single dose of 0.9% NaCl solution (saline), NPH insulin, or glargine (0.4 U/kg each) was administered by sc injection into the abdominal region. The same region was used for all three visits to avoid changes in absorption and distribution of insulin (23, 24) and minimize within-subject variability, even though rates of absorption are similar in the arm, leg, and abdomen (25). The variable GIR was continued to maintain BG at approximately 6 mmol/liter.
On the next day and at the end of the clamp, the patients were allowed to recover under observation before they left the hospital. A 2-wk washout period was mandatory before the patients were readmitted for the next stage of the cross-over study.
Safety
During the study, eight adverse events were reported, of which four were dislocation of the needle used for drawing blood, and four were headache/sleeplessness.
Data analyses and statistics
Insulin action over time was calculated using the BG concentrations measured every 5–15 min for the duration of the clamps: GIRmax, time to GIRmax (Tmax), and area under the GIR curve (AUCGIR). Correction for the placebo effect was performed by subtracting the measured parameter (GIR, T, or AUC) under the placebo condition at each time point from the parameter measured under the experimental conditions (NPH insulin or glargine) at the same time point.
Areas under the GIR profiles were calculated by the trapezoidal rule. A polynomial function (fifth grade) was fitted to the individual GIRs to allow for calculation of GIRmax and Tmax. The best-fit values of the constants of the polynomial function were proved to approximate a normal distribution.
We used nonsteady-state equations for stable isotopes, i.e. GIR in µmol/min divided by the substrate plasma enrichment minus the tracer infusion rate Ra = F/(plasma enrichment – F) (26), where F is the infusion rate of the labeled glucose. Glucose Ra and Rd were calculated using the equations described by Finegood et al. (27). EGP was calculated as Ra – F.
We have previously determined EGP in healthy subjects to be 11.6 ± 1.7 µmol/kg–1·min–1 (22). Assuming 
= 0.05, β = 0.85, intersubject CV 25% of maximum, and an expected difference between means of more than or equal to 2.3 µmol/kg–1·min–1 (20% of control), a minimum of eight was calculated per treatment arm of the study.
ANOVA for repeated measurements was used to determine differences among the three treatment groups, with treatment entered as a fixed variable. Group differences were tested by Bonferroni’s multiple comparison test. Calculations and statistical analyses, including tests for normal distribution of data sets, were performed using Prism (GraphPad Software Inc., San Diego, CA). All results are given as means ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
A total of 10 patients (four males, six females; mean age: 56 ± 8.5 yr) with T2DM (mean duration: 7.7 ± 4.9 yr) completed all three arms of the cross-over study; their baseline characteristics are presented in Table 1 (they are also shown in the supplemental data, which are published as supplemental data on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org).
|
Clamp profiles
Euglycemia was maintained throughout all clamps. Plasma glucose levels were stable between 2200 and 1000 h, with mean concentrations of 6.2 ± 0.5 mmol/liter (placebo/saline), 6.3 ± 0.5 mmol/liter (NPH insulin), and 6.1 ± 0.5 mmol/liter (glargine) (Table 2). The mean of the individual CVs for glucose during the clamp studies was 4.6 ± 1.3% for saline, 7.5 ± 2.4% for NPH insulin, and 7.3 ± 2.1% for glargine. One hundred twenty minutes before the sc injection of study medication according to the randomization schedule, glucose infusion was commenced to maintain steady BG concentrations (Fig. 2A).
|
|
and Table 2 demonstrate that GIRmax was 15.3, 35.0, and 34.6 µmol/kg–1·min–1 with saline, NPH insulin, and insulin glargine, respectively. After adjustment for saline, GIRmax was more than doubled with NPH insulin vs. glargine (P = 0.014). Similarly, the corrected AUCGIR for the total 12-h clamp was significantly greater with NPH insulin vs. glargine (5.6 vs. 3.9 M/min; P = 0.018). GIRmax was reached 130 min later with glargine (592 min) compared with NPH insulin (462 min) or saline (159 min). Adjusting for the effect of saline injection Tmax was still significantly later with glargine vs. NPH insulin (597 vs. 431 min; P = 0.019). Intersubject variability in GIR profiles was highly variable with both NPH insulin (41 ± 8%) and glargine (36 ± 8%). Hormone concentrations
Overall insulin concentrations for the duration of each clamp were 136 ± 27 pmol/liter for saline, and similar for NPH insulin and glargine treatment (NPH insulin, 166 ± 31 pmol/liter; insulin glargine, 160 ± 36 pmol/liter). Insulin profiles differed considerably as expected (Fig. 3A). Glucagon levels in the morning and at the end of the clamp were significantly (P < 0.021) higher with NPH insulin (36 ± 5 pmol/liter) compared with glargine (24 ± 4 pmol/liter) (Fig. 3B), whereas cortisol concentrations did not show a difference. As expected, C-peptide concentrations (Fig. 3C) were decreased in each experimental arm starting with insulin infusion, however, no significant difference of the magnitude of reduction of C-peptide levels was observed after sc saline or insulin injection.
|
The total amount of glucose infused to maintain euglycemia was nearly identical for NPH insulin and insulin glargine, demonstrating the equipotency of the insulin doses (Table 2). There was no difference in the time to starting glucose infusion, and GIRsmax were different only after adjustment for iv insulin. The profile of EGP was constant in the placebo group, allowing for the analysis of the effect of NPH insulin and insulin glargine on EGP.
The statistical analysis of the mean rate of glucose Rd of the insulin preparations is demonstrated in Table 3. In particular, time courses made differences evident (Fig. 4, A and B). At night, glucose Rd was more efficiently influenced by NPH insulin [2300–0200 h, 60–240 min glargine 28.4 ± 2.7 vs. NPH insulin 31.0 ± 2.5; difference –2.6; confidence interval (CI) –1.9 to –3.3 µmol/kg–1·min–1; P < 0.034]. During the morning hours, the situation was completely vice versa. Insulin glargine increased Rd significantly compared with NPH insulin (0700–1000 h, 540–720 min insulin glargine 33.0 ± 6 vs. NPH insulin 27.2 ± 5; difference 5.8; CI 4.7–6.9 µmol/kg–1·min–1; P < 0.018).
|
|
), with an increase in the morning hours. By contrast, glargine injection resulted in lower glucose production in the morning compared with NPH insulin (8.0 ± 1.2 vs. 13.7 ± 1.7; difference 5.7; CI 5.0–6.4 µmol/kg–1·min–1; P = 0.002). EGP time courses paralleled plasma glucagon levels, indicating that the suppression of EGP was partly mediated by glucagon. Of glargine’s mean glucose lowering effect in the morning, i.e. the proportion of infused glucose needed to overcome each of Rd and EGP, 78 ± 5% was due to the decrease in EGP, and 22 ± 3% was due to increase of Rd. The corresponding values for NPH insulin action in the morning were 9 ± 3 and 91 ± 8%, respectively (both P = 0.013), demonstrating that glargine’s effect on EGP was more pronounced than on Rd, whereas NPH insulin had a greater effect on Rd. Again this indicates a major difference in the metabolic effect of each insulin preparation.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
On the following morning, NPH insulin injected the night before resulted in an increase of EGP compared with glargine. The AUCsGIRs and the total amount of glucose infused were similar for both insulin preparations, indicating that their overall effect on glucose levels was equivalent. The difference was revealed by the time action profile. GIRmax was reached more than 2 h later with glargine compared with NPH insulin; insulin glargine was associated with greater glucose Rd compared with NPH insulin in the morning, but not at midnight. Possible reasons may be the increase of insulin antagonist hormones and/or a diurnal circadian mechanism directed by the autonomous nervous system in the early morning hours. In our experimental design, we observed that plasma glucagon after NPH insulin injection increased to levels higher than in the saline group, whereas glucagon concentrations were decreased subsequent to the application of insulin glargine. Presupposing glucagon supported the fasting plasma glucose level, its suppression after insulin glargine application contributed to the lower morning EGP compared with NPH insulin (31, 32). Other insulin antagonists such as GH or cortisol were increased overnight in T2DM, resulting in elevation of postabsorptive BG (33).
Before initiating this study, we assessed different insulin infusion doses because decrements of glucose production by actual insulin levels and hypoglycemic events might influence the results of the clamping tests. As demonstrated in the supplemental figure (which is published as supplemental data on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org), insulin infusion rates up to 0.8 mU/kg–1·min–1 (5.8 pmol/kg–1·min–1) were feasible to measure rates of EGP in T2DM in conjunction with bedtime insulin injection. Mean serum insulin concentrations of 160 pM adding sc insulin to an insulin infusion rate of 0.2 mU/kg–1·min–1 (1.4 pmol/kg–1·min–1) during the night allowed the detection of significant changes of EGP. Insulin infusion alone reduced C-peptide and free fatty acid concentrations (data not shown) by 30% in these insulin-resistant T2DM patients. Therefore, we cannot exclude that our model correctly reflects changes but underestimates absolute rates of glucose kinetics after sc injection of insulin. The main action of either of the long-acting insulins in practice is on basal glycemia, and with hyperinsulinemia resulting from insulin infusion overlaid on their action, that action is minimized, and other effects, such as enhancement of glucose uptake in muscle and suppression of fatty acid release in adipose tissue, are increased. Along these lines the hyperinsulinemic clamp as a model to study sc injected insulin has inherent limitations.
EGP is dependent on hepatic or splanchnic glucose turnover, whereas glucose fluxes of skeletal muscle mainly result in changes of Rd. In the presence of insulin glargine, Rd increased continuously. After injection of NPH insulin, the Rd profile had a flat maximum 2 h after midnight. Compared with insulin glargine, NPH insulin had a more pronounced effect on Rd between 2300 and 0200 h.
The EGP profiles disclosed an inverted picture. After insulin glargine injection, EGP resulted in a slow decline. By contrast, injection of NPH insulin showed a minimum of the EGP profile by midnight. In the morning there seemed to be a rebound phenomenon because EGP tended to be higher after NPH insulin injection than in the placebo group.
Interestingly, in the morning between 0600 and 1000 h, significant differences of Rd and EGP occurred without any change of the serum insulin levels, demonstrating the time-dependent variability of insulin effectiveness. Insulin glargine was more capable than NPH insulin to regulate both hepatic and peripheral glucose turnover in the morning hours. The glucose lowering effect of insulin glargine in the morning was largely caused by the suppression of liver and/or splanchnic glucose production and through reduced glucagon levels.
We used a three-way cross-over study of 10 T2DM patients who received saline, NPH insulin, and insulin glargine according to the randomization schedule. A cross-over study design is important in studies of T2DM patients owing to interpatient variability in endogenous insulin secretion, which is related to progression of β-cell dysfunction and peripheral insulin sensitivity in these patients. Therefore, we were able to correct for placebo effects, and, thus, calculate the proportional effects of insulin glargine and NPH insulin on GIR, Rd, and EGP.
In contrast to earlier reports on glucose kinetics (11, 12, 13, 14, 16, 18, 29), we have performed a head-to-head comparison of glargine with NPH insulin in T2DM. Most studies were performed on healthy volunteers or patients with T1DM. In a study by Luzio et al. (16), insulin glargine was compared with biphasic insulin aspart, whereas a study by Klein et al. (30) compared insulin glargine with insulin detemir. Pharmacodynamic properties of insulin glargine in T2DM were described, but no placebo control was included.
Because the duration of our clamps was 12 h, we followed the nocturnal duration of action of a single dose of insulin glargine in T2DM patients. Tmax reported here is broadly consistent with studies of healthy subjects (11, 12, 18, 29) and patients with T1DM (13, 14). In a study of patients with T2DM, which compared insulin glargine with biphasic insulin aspart, maximum concentration was reached at 6–16 h, and GIRmax was reached at 12 h after administration in the morning (16). In our study, GIRmax was reached 10 h after administration at bedtime. This is a fair congruity given the interindividual differences in absorption kinetics and study design.
The direct comparison of insulin glargine compared with NPH insulin allows some insight into the mechanism of the lower risk of nocturnal hypoglycemia described for glargine. GIRmax for NPH insulin was reached 6–7 h after administration at 2200 h, corresponding to 0400–0500 h in the morning, whereas other studies have reported that GIRmax is reached more quickly, at 4–6 h after administration (11, 12, 14, 30). In contrast to ours, these studies used morning injections.
Recently, it was demonstrated that glucose production in T2DM displays diurnal rhythms that result from factors antagonistic to insulin action triggered by the brain (34, 35). Because GIR reflects the current biological activity of insulin, diurnal changes in insulin effectiveness will modulate the metabolic profile of insulin, when injected in the morning or at bedtime.
From a clinical perspective, GIRmax with insulin glargine was attained almost 10 h after administration, corresponding to 0800 h in the morning, a time when the patient is more likely to be awake and having breakfast, or is more able to take appropriate corrective measures against potential hypoglycemia. In addition, because the placebo-corrected GIRmax was lower, and rates of EGP are higher with insulin glargine during the night, a precipitous decline in BG is prevented.
Appropriate expression of insulin receptors in -cells is required for their glucose-dependent glucagon release into the blood (36). Insulin exerts paracrine control on pancreatic -cells by modulating its KATP channel sensitivity, which becomes distorted in T2DM, leading to dysregulated glucagon secretion (37). Although binding of insulin glargine to the insulin receptor is identical to human insulin, intracellular degradation of glargine is reduced compared with regular insulin, resulting in a modulation of protein synthesis in cultured cells (38). Human insulin could be less effective at inhibiting glucagon secretion compared with the analog in T2DM. If this occurred in a clinical situation, i.e. beyond the confines of a euglycemic clamp with constant regular insulin infusion, enhanced glucagon secretion would be even more pronounced, and the reduction of glucagon concentrations by glargine could be even greater.
In conclusion, insulin glargine had a delayed onset of peak action compared with NPH insulin in patients with T2DM. We have directly demonstrated that insulin glargine was associated with improved control of morning EGP compared with NPH insulin, when administered at bedtime. Together, these properties of insulin glargine would be expected to contribute to the reduced risk of nocturnal hypoglycemia and lower FBG compared with NPH insulin, both of which were consistently reported in clinical trials.
|
Footnotes |
|---|
Selected data from this manuscript have been presented as a poster at the 66th Scientific Sessions of the American Diabetes Association, Washington, D.C., June 9–13, 2006 [Diabetes 55(Suppl 1):A111 (Abstract 466-P)].
This trial is registered at ClinicalTrials.gov ID No. NCT00468364.
Disclosure Statement: T.L. has received a research grant from Sanofi-Aventis. R.G.B. served as a consultant to or gave lectures organized by Lilly, Novo-Nordisk, and Sanofi-Aventis.
First Published Online July 8, 2008
Abbreviations: AUCGIR, Area under the glucose infusion rate curve; BG, blood glucose; CI, confidence interval; CV, coefficient of variation; EGP, endogenous glucose production; FBG, fasting blood glucose; GIR, glucose infusion rate; GIRmax, maximum glucose infusion rate; HbA1c, glycosylated hemoglobin; NPH, neutral protamine Hagedorn; OAD, oral antidiabetic agent; Ra, rate of appearance; Rd, rate of disappearance; Tmax, time to maximum glucose infusion rate; T1DM, type 1 diabetes; T2DM, type 2 diabetes.
Received December 31, 2007.
Accepted July 1, 2008.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
-cells. J Biol Chem 280:33487–33496This article has been cited by other articles:
 |
|
 |
|
  R. G. Bretzel, M. Eckhard, W. Landgraf, D. R. Owens, and T. Linn Initiating Insulin Therapy in Type 2 Diabetic Patients Failing on Oral Hypoglycemic Agents: Basal or prandial insulin? The APOLLO trial and beyond Diabetes Care, November 1, 2009; 32(suppl_2): S260 - S265. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
