Glucose Clamping in Clinical Trials

By John Illingworth, Managing Director, Illingworth Research 


The glucose clamp technique is the standard method for determining the time and action profile of new insulin preparations, and the measurement insulin and beta-cell sensitivity (1). The European Medicines Agency (EMA) requires data on the time-action profiles of new insulin preparations, using the euglycaemic clamp technique, for registration of new insulins in the treatment of diabetes mellitus (2). This article will focus on the euglycaemic clamp in the assessment of new insulin preparations in clinical trials.

History of Glucose Clamping

The isoglycaemic clamp technique was developed to measure tissue sensitivity to insulin. Glucose clamp methodology was introduced in 1979 by DeFronzo et al (3), who describe the hyperglycaemic and euglycaemic glucose clamp techniques. The hyperglycaemic clamp involves elevating the blood glucose concentration above normal basal levels using a glucose infusion. Blood glucose concentrations are maintained, or “clamped”, using an adjustable glucose infusion. The amount of IV glucose needed to maintain this glucose concentration at the target level, is equal to the glucose uptake of all tissues (3). The glucose infusion rate (GIR) is a measure of glucose metabolism, and as insulin starts to take effect, the glucose infusion rate is slowly increased to counterbalance the drop in blood glucose. DeFronzo et al also described the euglycaemic insulin clamp technique in the measure of tissue sensitivity to exogenous insulin. This is where insulin is infused and the plasma glucose held at a constant basal level by a glucose infusion. The GIR equals the glucose uptake by all the tissues in the body and is a measure of tissue sensitivity.

Patient and Healthy Subject Studies

When testing potential new insulin for efficacy, it is not possible to administer the insulin to a healthy trial subject and observe the results, as the subject’s blood sugar will drop significantly and the subject will become hypoglycaemic. In order to permit evaluation of a new insulin, the subject must be administered a glucose infusion, to counteract the glucose-lowering effects of the insulin. The subject’s blood sugar is maintained at a constant level, for example in the case of a euglycaemic clamp, the blood sugar level will be clamped at the blood glucose level of a typical fasting healthy person which would usually be in the range of 4.1 to 6.1 mmol/L. The clamp clinical study protocol will always specify the blood glucose level which the study technicians should maintain the glucose levels at.

After admission to the hospital or research clinic the day prior to the clamp, all subjects are assessed to ensure eligibility and provided with a meal. Subjects are then required to fast overnight. Another blood glucose measurement is taken prior to dosing on the day of the clamp to ensure eligibility. Subjects will usually be cannulated in both arms, one for blood sampling and intravenous saline, and the other for a glucose infusion. The subject is then administered the protocol-specified dose of insulin, for example, using an insulin pen and administering a 0.2U/kg dose, or using a needle with a specified volume of insulin.

Blood sampling to determine blood glucose is then performed at regular intervals, usually 5 minutes, for the duration of the clamp. A number of glucose reports will be generated by the glucose reading machines, which require real-time interpretation and adjustment of the GIR. For example, if the blood glucose has dropped since the previous measurement, the GIR will be increased. The GIR and the time the GIR was changed are documented.

The length of the clamp is determined by the type of insulin under study. For example, a short acting insulin like Insulin Aspart (NovoRapid®) (5) which has a rapid onset of action of 10-20 minutes and a duration of action of 3-5 hours, will require a short clamp of perhaps 8 to 10 hours. This period would allow sufficient time to assess the efficacy and onset of the insulin.

A long acting basal insulin, like Insulin glargine (Lantus®), has a mean duration of action of 24 hours (6), and would therefore require a longer duration clamp, for example up to 30 hours.

Glucose Clamping Figure 1Diagram 1. Representation of Euglycaemic Glucose Clamping. The shape in the centre represents the subject, with saline, insulin and glucose being infused / administered at various time points throughout the clamp, and then blood sampling at frequent time points to measure blood glucose levels. The GIR is changed depending on the blood glucose measurement.



Data Monitoring and Data Management in Clamp Studies

During the study set-up phase, a decision to use a paper or electronic case report form (eCRF) for the purposes of data capture, must be made. Depending on which method of data capture is chosen, this may change the data monitoring aspects of the study. For example, a paper CRF may be used as source documentation in the clinic during the clamp, whereas an eCRF may not be the ideal data capture source for real-time entry during the clamp. In this case, a source document workbook may be created to capture all the specific clamp data, including glucose readings and averages (mmol/L), time of blood draws (hh:mm), GIR (mg/kg/min or mL/Hr), GIR time change (hh:mm:ss). These data are then entered into the eCRF once the clamp has completed.

Prior to enrolment of the first subject, it is important to establish and define the source data for the study and which data are to be captured in the eCRF. Source data will usually consist of the real-time glucose reports, source data workbook, laboratory reports, ECGs and screening data workbook. It is also important to discuss the percentage source data verification (SDV) that will be required, i.e. the percentage of eCRF data that is verified against the source data. Most sponsors will usually require 100% SDV for phase I studies, but due to the huge amount of data generated in glucose clamp studies, it may be necessary to reduce this, especially for 30h clamps with 5 minute interval blood sampling. 100% SDV may be necessary to ensure accuracy and consistency of the data as partial SDV runs the risk of unclean data being analysed and affecting the validity of results.

eCRF completion guidelines are also hugely important, which must be produced by the study team prior to the study start. These are often accessed in the eCRF web portal and can be referred to while entering data. The completion guidelines must document the format and requirement of the CRF data, including how to complete the eCRF fields, in particular missing or incomplete data, and how to avoid system generated queries.

Data monitoring requirements must be decided by the study team prior to the study start, and should be fully documented in a monitoring plan or similar document. Data monitoring may be a time consuming and sometimes costly process, but the benefits will be seen when the clean data are analysed and interpreted.

Glucose Clamping Figure 2Diagram 2. Data Flow. The above diagram gives a summary of the stages involved in the data flow during a typical clamp using an eCRF, from generation and completion of source data workbooks, completion of data in the eCRF of data in the clinic during the clamp, to data monitoring




Results and Interpretation of Data

The GIR plays an essential role in describing the pharmacodynamic effects of the insulin under study. When the GIR is plotted over the time period of a typical clamp, the curve will describe the action of the study insulin. This principle is used in all glucose clamp studies. When looking at basal insulins, the main factors to be assessed are the duration of action, flatness and variability (6). The duration of action is the difference between onset of action and the end of action, terms that are normally defined in the clinical trial protocol, and is approximately 24 hours for insulin glargine. Insulin glargine, a human insulin analogue, forms micro-precipitates after injection into subcutaneous tissues. Small amounts of insulin glargine are continuously released from the micro-precipitates, which provide a smooth and peakless, concentration/time profile (8), with duration of action of around 24 hours. There is no standard definition of ‘flatness of curve’, although the term ‘peaklessness’ is also used. Pharmaceutical companies want smooth, peakless curves, with the aim of minimising the possibility of nocturnal hypoglycaemia. Within-subject and between subject variability are also very important in assessing the pharmacodynamic properties of new insulins. Within-subject variability will affect the extent to which blood glucose levels fluctuate in individual patients from one day to another (6).


There are a number of different types of glucose clamp techniques, used to assess the sensitivity of human tissues to insulin, and to also assess the pharmacodynamic properties of new insulin preparations. The assessment of new insulin preparations by the euglycaemic clamp technique is now a requirement by the EMA for the registration of insulins in diabetes mellitus.

Note – The author wishes to acknowledge the considerable input of Lee Woodward in the preparation of this article.


1. Glucose Clamp Algorithms and Insulin Time-Action Profiles. Bequette, Wayne. 5, 2009, Journal of Diabetes Science and Technology, Vol. 3, pp. 1005-1013.

2. (EMA), European Medicines Agency. Note for Guidance on Clinical Investigation of Medicinal Products in the Treatment of Diabetes Mellitus. London : s.n., 2002.

3. Glucose clamp technique: a method for quanitifying insulin secretion. DeFronzo, Ralph A., Tobin, Jordan D., and Andres, Reubin. 3, 1979, American Journal of Physiology and Endocrinology Metabolism, Vol. 273.

4. Early Pharmacokinetic and Pharmacodynamic Effects of Mixing Lispro With Glargine Insulin. Cengiz, E., Tamborlane, W.V., Martin-Fredericksen, M., Dziura, J. and Weinzimer, S.A. 2010, Diabetes Care, Vol. 33, pp. 1009-1012.

5. Summary of Product Characteristics for NovoRapid. 2011.

6. Towards peakless, reproducible and long-acting insulins. An assessment of the basal analogues based on isoglycaemic clamp studies. Pieber, T., and Heise T.R. 2007, Diabetes, Obesity and Metabolism, Vol. 9, pp. 648-659.

7. The interpretation of glucose clamp studies of long-acting insulin analogues: from physiology to marketing and back. Swinnen, S. G. H. A., Holleman, F. and DeVries, J. H. Amsterdam : Springer, 2008, Diabetologia, Vol. 51, pp. 1790-1795.

8. Summary of Product Characteristics for Lantus. 2011.