We thank Choukem and Gautier (1) for their comment, in which constructive criticism relating to the evaluation of glucagon-like peptide-1 (GLP-1) secretion is raised, concerning our study (2).
Secretion of hormones following different stimuli are difficult to estimate from peripheral levels of hormone concentrations. Therefore, hormone responses as measured by concentration changes in the peripheral circulation are often used as indicators of the actual amount of secretion.
Choukem and Gautier are right that when the computation of the incremental area under curve (AUC) is based on the difference between the first and the subsequent measures, a possible error of that first measurement will accumulate (3). Therefore, in our studies, we use the average of three baseline values (time points −15, −10, and 0 min) when computing incremental AUCs to minimize the risk of errors at baseline. The incremental AUC values were chosen as a result of the relatively small intact GLP-1 responses observed. We anticipated that less pronounced differences between the responses to the two stimuli (50-g oral glucose tolerance test [OGTT] vs. isoglycemic intravenous glucose infusion) would be easier to detect using the incremental values.
In the study in question, we measured both total GLP-1 concentrations (intact and active peptide plus the inactive metabolite) and intact levels (active peptide only) of GLP-1. The intact levels are indicators of the concentration of active hormones in the peripheral circulation and therefore its signaling via the endocrine (blood-borne) pathway. The total GLP-1 concentrations are indicators of the overall levels of secretion and therefore reflect the total L-cell secretion including the amount of hormone that is degraded by dipeptidyl peptidase 4 but that had a chance to interact with neural pathways and receptors in the portal system (4). The latter pathway may be the most important for the physiological actions of GLP-1.
First, looking at the incremental AUCs in the group of patients with type 2 diabetes, we actually do find a difference between the responses of total GLP-1 to OGTT and isoglycemic intravenous glucose infusion, indicating a higher overall level of GLP-1 secretion during the former compared with the latter. However, with respect to intact GLP-1, there was no significant difference between the two responses estimated from the incremental AUC. We offered the explanation that, because GLP-1 is subject to degradation by dipeptidyl peptidase 4 almost immediately upon its release (5), only 10–15% of intact GLP-1 actually reaches the systemic circulation (4), thereby reducing the possibility of detecting a small difference between the responses in the peripheral circulation. However, when the intact GLP-1 responses in the group of patients with type 2 diabetes are estimated using the AUC “with respect to ground” (i.e., from zero; total AUCs), as suggested by Choukem and Gautier, we do actually find a significant difference between the responses to OGTT and isoglycemic intravenous glucose infusion (711 ± 126 vs. 435 ± 91, respectively; P = 0.039). The same is true for the remaining groups except for the group of patients with chronic pancreatitis and normal glucose tolerance, in which no significant difference is observed (622 ± 122 vs. 452 ± 126; P = 0.2).
As mentioned, incremental AUCs are of interest when responses are not very pronounced, as is the case for intact GLP-1 responses in our study. Nevertheless, in this situation, we agree with Choukem and Gautier that using the total AUC (AUC with respect to ground) results in a better description of what actually occurs.
Choukem and Gautier additionally suggest that another way to evaluate the hormone responses is to use the incremental AUC, but to stop at the point where the hormone concentration returns to baseline level because what occurs after that time leads to negative values and might therefore not be physiologically relevant. This can definitely be of interest in many situations, but since the one stimulus (isoglycemic intravenous glucose infusion) in the study in question is not expected to elicit any GLP-1 response, values below zero can be expected already from the first poststimulus measurement as a result of variation. Therefore, this method is impracticable in this particular situation.
Last, Choukem and Gautier mention that our study does not provide evidence for a defective GLP-1 secretion in type 2 diabetes, contrary to what has been previously reported. GLP-1 responses to mixed-meal tests have consistently been shown to be attenuated in patients with type 2 diabetes (6,7), and recently this has been confirmed in another study from our group (8). In the study in question, we do not find compromised GLP-1 responses during OGTT in our patients with type 2 diabetes. One of the reasons for this may be the use of an OGTT rather than a mixed-meal test. OGTTs have previously been found to produce a normal or even elevated (total) GLP-1 response in patients with type 2 diabetes compared with healthy control subjects (9). However, a more important explanation may be the fact that we here studied lean patients with type 2 diabetes (to allow a comparison with the patients with chronic pancreatitis).
Several investigators have shown that GLP-1 responses following orally ingested nutrients are reduced among obese subjects (6,10–12), and a recent study by Muscelli et al. (13) describes separate impacts of obesity and reduced glucose tolerance, respectively, on the secretion of GLP-1. In fact, in patients with morbid obesity, GLP-1 may not be measurable at all (14). Interestingly, one of the studies mentioned above (Orskov et al. [9]), describing increased GLP-1 responses during OGTT in patients with type 2 diabetes (9), included patients with an average BMI of 28.2 kg/m2, i.e., not obese but slightly overweight patients. Thus, obesity appears to be one of the factors responsible for impaired secretion of GLP-1 in type 2 diabetes (6,13).
In conclusion, we agree with Choukem and Gautier that hormone responses can be calculated in a number of ways and that investigators should be careful in their choice of computation in order to obtain the most physiologically relevant results. With regard to GLP-1 responses, these must be considered in relation to the stimulus used and to the level of obesity in the patients. Last, we would like to thank Choukem and Gautier for their relevant and very helpful comments.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
