Ann Avron, Merana Tamir, Treatment Effect in a Type 1 Diabetes … · Dana Elias, 4Ann Avron, Merana Tamir, Rachel Eren, 4Shlomo Dagan, and Irun R. Cohen5 1384 Diabetes Care Volume

Evaluation of Long-TermTreatment Effect in a Type 1Diabetes Intervention Trial:Differences After StimulationWith Glucagon or a Mixed Meal

OBJECTIVE

Endogenous insulin secretion, measured by C-peptide area under the curve (AUC),can be tested using both the glucagon stimulation test (GST) and the mixed-mealtolerance test (MMTT). This study compares these two stimulation methods usinglong-term data from patients newly diagnosed with type 1 diabetes or with latentautoimmune diabetes.

RESEARCH DESIGN AND METHODS

A recently completed phase 3 intervention study with DiaPep277 demonstratedimproved glycemic control and a significant treatment effect of glucagon-stimulatedC-peptide secretion. Unexpectedly, MMTT failed to detect differences between thetreated and control groups. Data from 343 patients in two balanced-randomized,double-blind, placebo-controlled, parallel-group trials of DiaPep277 were used tocompare and correlate betweenGST- andMMTT-derived C-peptide AUC. Pearson’scorrelations were calculated for absolute C-peptide AUC at baseline and 12 and 24months and for long-term changes in AUC (ΔAUC).

RESULTS

The absolute AUC values obtained at any single time point by the two tests werewell correlated in both data sets (r = 0.74–0.9). However, the correlations betweenthe ΔAUC were much weaker (r = 0.39–0.58). GST-stimulated C-peptide secretionwas stable over the fasting glucose range permitted for the test (4–11.1 mmol/L),but MMTT-stimulated C-peptide secretion decreased over the same range, im-plying differences in sensitivity to glucose.

CONCLUSIONS

Measurement of long-term changes in stimulated C-peptide, reflecting endoge-nous insulin secretion, during the course of intervention trials may be affected bythe method of stimulation, possibly reflecting different sensitivities to the phys-iological status of the tested subject.Diabetes Care 2014;37:1384–1391 | DOI: 10.2337/dc13-1392

1Department of Endocrinology and Diabetes,University Campus Bio-Medico, Rome, Italy2Center of Diabetes, Barts and the London Schoolof Medicine, Queen Mary University of London,London, United Kingdom3Diabetes Unit, Department of InternalMedicine, Hadassah University Hospital,Jerusalem, Israel4Andromeda Biotech Ltd., Yavne, Israel5Department of Immunology, WeizmannInstitute of Science, Rehovot, Israel

Corresponding author: Itamar Raz, [email protected].

Received 12 June 2013 and accepted 9 October2013.Clinical trial reg. nos. NCT00615264 andNCT00058981, clinicaltrials.gov.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-1392/-/DC1.

© 2014 by the American Diabetes Association.See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

See accompanying articles, pp. 1173and 1392.

Paolo Pozzilli,1,2 Itamar Raz,3 Dana Peled,4

Dana Elias,4 Ann Avron,4 Merana Tamir,4

Rachel Eren,4 Shlomo Dagan,4 and

Irun R. Cohen5

1384 Diabetes Care Volume 37, May 2014

EMER

GINGTECHNOLO

GIESANDTH

ERAPEU

TICS

© © © © © ©

The hallmark of type 1 diabetes isdeterioration of endogenous insulinsecretion. Since insulin and C-peptideare cosecreted by b-cells on anequimolar basis, stimulated secretion ofC-peptide is widely accepted as ameasure of b-cell function (1–4).Moreover, the Diabetes Control andComplications Trial demonstrated thatlevels of stimulated C-peptide as low as0.2 nmol/L are clinically meaningful andcould prevent or delay long-termdiabetes complications (5,6). C-peptidesecretion is commonly stimulated by theintravenous administration of glucagonin the glucagon stimulation test (GST) orby the ingestion of a mixed-mealformula in the mixed-meal tolerancetest (MMTT) (7,8). Rapid infusion ofglucagon induces insulin secretionimmediately by directly stimulatingb-cells (9). By contrast, the mixed-mealstimulates insulin secretion from b-cellsindirectly after ingestion ofcarbohydrates, fats, and proteins (10).

The use of the MMTT wasrecommended for the purpose ofstandardization between interventionstudies involving patients with newlydiagnosed type 1 diabetes. Investigatorswho have chosen to use the GST havebeen advised to perform also the MMTTat least at baseline and annually in orderto obtain comparative data (8). Acomparison between the MMTT andthe GST conducted by the Type 1Diabetes TrialNet Research Group andthe European C-Peptide Trial StudyGroup (11) found that both procedureswere valid for the determination ofb-cell function. Although theMMTTwasreported to be more sensitive andreproducible, both tests weredemonstrated to be highly reproducibleand well correlated when performedwithin a month. However, the abovestudy did not investigate the use of thetwo procedures to measure long-termchanges in b-cell function.

In this report, we compare and correlatethe absolute values for the C-peptidearea under the curve (AUC) and thelong-term changes from baseline in AUC(ΔAUC) at specific time points obtainedusing the GST and MMTT procedures.The data are from individual patientswith autoimmune diabetes participatingin the DIA-AID 1 or LADA intervention

trials. These two trials were conductedto evaluate the efficacy of DiaPep277,as a treatment modality for thepreservation of b-cell function. The endpoint was defined as the change frombaseline to study end in stimulated C-peptide AUC secretion and wasmeasured by both the GST and MMTT.

Data obtained from the DIA-AID 1 studydemonstrated significantly preservedC-peptide secretion in the DiaPep277treatment arm compared with theplacebo arm when measured by GSTbut not by MMTT (12). This wasaccompanied by a significantimprovement in glycemic control.

Herein, the discrepancy between theGST and MMTT data obtained by long-term follow-up in a clinical setting isfurther investigated and discussed.

RESEARCH DESIGN AND METHODS

Clinical StudiesData from two trials registered withClinicalTrials.gov, namely, DIA-AID 1(protocol 901; trial numberNCT00615264)and LADA (protocol 702; trial numberNCT00058981) were used. All patientssigned informed consent forms. Theprotocol and consent documents wereapproved by an independent ethicscommittee or institutional review board ateach participating center (DIA-AID 1, 46international centers; LADA, 7U.S. centers).

DIA-AID 1 (Protocol 901) is a phase 3,multinational, balanced-randomized,double-blind, placebo-controlled,parallel-group study that was designedto evaluate the clinical safety andefficacy of DiaPep277 as an additionaltreatment in newly diagnosed patientswith type 1 diabetes who were beingtreated with insulin. For a fulldescription of this study, see reference(12). Briefly, participants consisted ofadolescents and adults (aged 16–45years) who had been diagnosed withtype 1 diabetes, as defined by theAmerican Diabetes Association or theWorld Health Organization, and whopossessed residual b-cell function asevidenced by fasting C-peptide levels$0.22 nmol/L at the screening visit.Patients received 1 mg DiaPep277 orplacebo by subcutaneous injection atmonths 0, 1, 3, 6, 9, 12, 15, 18, and 21 ofthe study, with follow-up continuingfor a further 3 months.

The change in stimulated C-peptidesecretion from baseline to study end(month 24) was measured as the AUCobtained by the GST (primary end point)and the MMTT (secondary end point).Clinical end points included theproportion of patients who, at studyend, maintained HbA1c treat-to-targetlevels of #7% (#53 mmol/mol), insulindose at study end, and hypoglycemicevent rate (12).

The current study used data from allDIA-AID 1 study patients for whom GSTand MMTT data were available both atbaseline and at study end (N = 297,designated the subgroup population).Within this subgroup population, 146patients were treated with DiaPep277and 151 with placebo.

The LADA study (protocol 702) was aphase 2, randomized, double-blind,placebo-controlled, parallel-groupclinical study to investigate the safetyand tolerability as well as theimmunological and clinical efficacy ofmultiple subcutaneous doses ofDiaPep277 in patients with latentautoimmune diabetes. Participantsconsisted of adult patients (aged 30–70years) who had been diagnosed withdiabetes for up to 5 years and who weresubsequently defined as LADA based onpresence of GAD autoantibodies atscreening. Patients had fastingC-peptide levels of $0.3 nmol/L andreceived 1 mg DiaPep277 or placebo bysubcutaneous injection at months 0, 1,3, 6, 9, 12, 15, and 18. The study wasterminated prematurely following aninterim analysis showing that it was notpowered to demonstrate statisticallysignificant efficacy, so the majority ofpatients were only followed up tomonth 12.

The change from baseline to month 12in stimulated C-peptide secretion wasmeasured as the AUC obtained by theGST and the MMTT. The current studyused data from all LADA study patientsfor whom GST and MMTT data wereavailable both at baseline and atmonth 12 (N = 46, 22 in the DiaPep277group and 24 in the placebo group).Albeit an abbreviated study, long-termchanges in C-peptide measured by GSTandMMTT in individual patients could becompared.

care.diabetesjournals.org Pozzilli and Associates 1385

©

GST and MMTT ProceduresIn the DIA-AID 1 study, C-peptide levelswere measured after stimulation byglucagon at month 1 (GST baseline),month 12, and month 24. MMTT-stimulated C-peptide secretion wasmeasured at month 0 (MMTT baseline)and at months 6, 12, 18, and 24. The GSTat months 12 and 24 was performedeither 3 weeks before or up to 2 weeksafter the MMTT.

In the LADA study, C-peptide levels weremeasured first by MMTT and two weekslater by GST at baseline and month 12.

Fasting and stimulated C-peptide datafrom baseline and month 12 of bothstudies, and from month 24 of the DIA-AID 1 study, were analyzed. The testswere performed in the morning after anovernight fast of 8–10 h only if thefasting plasma glucose level was withinthe permitted range of 4 to 11.1 mmol/L(72 to 200 mg/dL) (8).

For the GST, 1 mg glucagon (GlucaGenHypoKit, Novo Nordisk) was rapidlyadministered intravenously at time 0.Blood samples for the determination ofC-peptide were drawn at 25 min, attime 0 (immediately before glucagonadministration), and at 2, 6, 10, and20 min postadministration.

For the MMTT, each patient ingested amixed-meal (Ensure, Abbott, for the DIAAID 1 study and Boost, Nestle, for theLADA study) whose volume wascalculated as 6 mL/kg body weightwith a maximum dose of 360 mL. Bloodsamples for the determination ofC-peptide were drawn at 210 min, attime 0 (immediately before mixed-mealingestion), and at 10, 20, 30, 60, 90, and120 min postingestion.

Fasting C-peptide was defined as theaverage of the concentrations measuredat time25min and time0 (for theGST) orat210 min and time 0 (for the MMTT).Where a blood sample was collected atonly one of these time points, that valuewas considered time 0.

Laboratory TestsAll laboratory tests were performed incentral laboratories. Samples forC-peptide analysis were shipped frozenfrom the clinical sites to the LKF GmbHclinical laboratory (Schwentinental,Germany) for the DIA-AID 1 study and to

the Northwest Lipid ResearchLaboratories (Seattle, WA) for the LADAstudy. C-Peptide was measured usingradioimmunoassay kits (Human C-peptideRIA kit, LINCO). The analytical range of theassay is 0.0331–1.655 nmol/L.

Data AnalysesC-peptide AUC was calculated using thetrapezoidal rule. In order to comparethe areas under the GST- and MMTT-stimulated C-peptide curves, each AUCwas normalized by dividing it by thetime interval (20 min for GST or 120 minfor MMTT), and thus all data arereported in units of nanomoles per literper minute.

Changes in C-peptide AUC (DAUC) werecalculated from baseline to month 12and, for data from the DIA-AID 1 study,also frombaseline tomonth 24 and frommonth 12 to month 24, using the mixed-effects model for repeatedmeasurements (13,14). Relativetreatment effect was defined as theratio between the DAUC values of theDiaPep277-treated group and theplacebo group and is expressed as apercentage. Pearson’s correlationcoefficients (r) were calculated inMicrosoft Excel 2007.

Since two different valid methods formeasuring a single parameter are beingcompared, a strong correlation of atleast r = 0.8 would be expected.

To determine the impact of fastingglucose on the C-peptide response tostimulation, patients were groupedaccording to their fasting glucoseconcentration as measured at the startof each test. A stimulation index wascalculated for each patient at each visitas Cmax/fasting C-peptide (where Cmax isthe maximum stimulated C-peptideconcentration measured post-testadministration) to obtain a measure ofstimulation in relation to the absolutevalues of fasting C-peptide.

RESULTS

Stimulated C-Peptide and theDiaPep277 Treatment EffectThe DIA-AID 1 clinical study hasdemonstrated that DiaPep277treatment significantly improvedglycemic control while using similar oreven lower insulin dose, indicatingpreservation of b-cell function. This was

accompanied by significant differencesin the change of C-peptide levelsbetween the DiaPep277-treated groupand the placebo group as measured byGST but not by MMTT (12).

In an attempt to explain the discrepancybetween the results of the twostimulation methods, we studied thecorrelation between long-term changesin stimulated C-peptide AUC obtainedby GST and MMTT. Data from 297patients of the subgroup populationwho had C-peptide levels measured byboth stimulation methods at baselineand at study end were analyzed.

The demographic and baselineparameters of the subgroup populationwere similar to those of the patientswho were analyzed in the modifiedintent-to-treat (mITT) population of theDIA-AID 1 study (12) (Table 1).

The treatment effect observed in thesubgroup population was similar to thatof the mITT population when C-peptidewas stimulated by glucagon. Likewise,there was no treatment effect whenC-peptide was stimulated by the mixedmeal.

Correlation analyses were performed onthe data obtained from patients in eachtreatment arm of the subgrouppopulation (Supplementary Figs. 1 and2). A good positive correlation wasobserved between the absoluteC-peptide AUC values obtained by theGST andMMTTmethods at specific timepoints. However, the ΔAUC valuesobtained by the two stimulationmethods were only weakly correlated.This was true in both the placebo arm(Supplementary Fig. 1) and theDiaPep277-treated arm (SupplementaryFig. 2), indicating that the observeddiscrepancy did not result from theDiaPep277 treatment. Therefore, thedata from the two treatment arms werepooled in all subsequent correlationanalyses.

The AUC of the stimulated C-peptidesecretion levels obtained by the twomethods were compared and correlatedin each individual patient in thesubgroup population (Fig. 1).

The Pearson’s correlation coefficientsbetween the absolute C-peptide AUCsobtained by GST andMMTT at individual

1386 Change in C-Peptide AUC Measured by GST or MMTT Diabetes Care Volume 37, May 2014

©

time points were good as expected, r =0.74, 0.82, and 0.89 at baseline, month12, and month 24, respectively (Fig. 1A–C). The correlations between the DAUCstimulated by GST and MMTT weremuch weaker over the course of thestudy, indicating that the change inC-peptide levels measured by oneprocedure cannot predict the changemeasured by the other procedure. Thecorrelation coefficients were r = 0.58across the whole study period(baseline–month 24), r = 0.41 during thefirst year of the study (baseline–month12), and r = 0.39 during the second yearof the study (months 12–24) (Fig. 1D–F).It should be noted that, already in thefirst year of the study, the DAUC valuesmeasured by the two methods were onlyweakly correlated. The weak correlationsfor DAUC from months 12–24 indicatethat the discrepancies between themethods cannot have resulted frommeasurements at baseline, despite the1-month difference between the GST andMMTT baselines.

Additional analyses by Bland–Altmanplots further support the Pearson’scorrelation of the test results (seeSupplementary Fig. 3).

A similar discrepancy was observedusing the smaller data set from the LADAstudy (protocol 702). While there wasgood correlation between the absolutevalues of AUC measured by the MMTTand GST procedures at baseline and atmonth 12 (r = 0.9 and 0.85, respectively;

Fig. 2A and B), a much weakercorrelation (r = 0.48) was foundbetween the ΔAUC values over thestudy time interval (Fig. 2C).

Fasting C-Peptide and Glucose LevelsPrior to C-Peptide StimulationFasting glucose was shown tosignificantly affect the C-peptideresponse to a mixed meal (11,15).Therefore, we assessed the influence ofglucose levels on the extent of C-peptidestimulation (regardless of the state ofthe disease). The levels of C-peptidesecretion stimulated by glucagon andby a mixed meal were compared andgrouped according to the fastingglucose range within which the testswere performed (4–11.1 mmol/L). Weused a stimulation index, defined asCmax/fasting C-peptide, to compensatefor differences in fasting C-peptidelevels. The indices obtained by theglucagon procedure (1.6–1.8) werestable over the glucose range permittedfor the tests. In contrast, the indicesobtained by MMTT were affected by thefasting glucose level: 4.5 at glucoselevels of 4–5 mmol/L that decreased to1.9 at glucose levels of 10–11.1 mmol/L(Fig. 3A).

Fasting C-peptide and glucose weremeasured prior to applying theC-peptide stimulation procedures atbaseline and at study end (Fig. 3B).Similar median fasting C-peptide valueswere obtained prior to the MMTT andGST at baseline (0.33 and 0.34 nmol/L,

respectively) and at study end (0.175and 0.177 nmol/L, respectively) in thesubgroup population. For fastingglucose, the median values were alsosimilar prior to these procedures atbaseline (MMTT, 5.96 mmol/L; GST,6.13 mmol/L) and at study end (MMTT,6.91 mmol/L; GST, 7.14 mmol/L). Asexpected from the natural history of thedisease, the median fasting C-peptidelevels decreased by month 24 and themedian fasting glucose levels increased,reflecting a decline in endogenousinsulin secretion (Fig. 3B).

CONCLUSIONS

Stimulated C-peptide provides a goodapproximation of endogenous insulinsecretion (and thus residual b-cellfunction). This parameter, while not adirect measure of b-cell mass, is thecurrently accepted primary efficacyend point parameter in immuneintervention studies in type 1 diabetes(3). Clinical benefits are generallyevaluated as secondary end pointsto support the primary efficacy endpoint.

The DIA-AID 1 study indicated significantpreservation of b-cell functionsupported by clinical benefits when thestimulation of C-peptide secretion wasevaluated by GST. However, changes inC-peptide secretion stimulated byMMTT were not significant and did notcorrelate with the observed clinicalbenefits (12).

Table 1—Baseline and demographic parameters of the subgroup population compared with the mITT populationfrom the DIA-AID 1 study

Parameter Subgroup population mITT population from the DIA-AID 1 study

Number of patients 297 422

Sex (male/female) 206/91 276/146

Age (years) 27.55 6 8.00 (16–45) 26.6 6 7.74 (16–45)

Time from diagnosis (months) 2.74 6 1.2 (0.7–7.6) 2.8 6 1.2 (0.5–7.6)

BMI (kg/m2) 22.64 6 2.90 (17.3–33.2) 22.67 6 3.0 (17–33.2)

Fasting C-peptide (nmol/L) 0.44 6 0.21 (0.21–1.84)/[0.28–0.54]* 0.45 6 0.23 (0.21–2.28)/[0.3–0.54]*

HbA1c (%) 7.28 6 1.56 (4.0–14.6) 7.4 6 1.7 (4–16.4)

HbA1c (mmol/mol) 56.1 6 17.01 (20.2–136.1) 57.5 6 18.7 (20.2–155.7)

Insulin dose (units/kg/day at 3 months) 0.38 6 0.19 (0.02–1.21) 0.38 6 0.26 (0.02–2.02)

Autoantibodies (% positive)IA-2A 62.6% 61%IA 73.7% 74%GADA 87.5% 87%

Data are average 6 SD (range). IA-2A, IA-2 protein tyrosine phosphatase; IA, insulin antibody; GADA, glutamic acid decarboxylase. *Interquartilerange.

care.diabetesjournals.org Pozzilli and Associates 1387

©

The current analysis was undertaken toelucidate this inconsistency using twoindependent data sets from the DIA-AID1 and LADA trials. It revealed that theabsolute values of stimulated C-peptidemeasured by GST and MMTT were

reproducible and well correlated whenevaluated at each individual time point,so confirming the findings of previousreports (7,11,16). However, the ΔAUCsobtained by the two methods over thecourse of the studies were only weakly

correlated. This discrepancy is ofparticular importance since it is ΔAUCs(rather than absolute AUCs) thatindicate change over time from baselineand thus reveal the dynamics of diseaseprogression. As such, it is only ΔAUCs

Figure 1—Comparison of C-peptide stimulation methods in the DIA-AID 1 study subgroup population (N = 297). The correlations between the AUCsfor stimulated C-peptide concentration obtained using the GST and the MMTT at (A) baseline, (B) month 12, and (C) month 24. The correlationbetween the changes in stimulated C-peptide AUC (DAUC)measured by GST andMMTT (D) from baseline tomonth 24 (whole study period), (E) frombaseline to month 12 (first year of the study), and (F) from month 12 to month 24 (second year of the study). Each data point represents a singlepatient.

1388 Change in C-Peptide AUC Measured by GST or MMTT Diabetes Care Volume 37, May 2014

©

that should be used to calculatetreatment effect and preservation ofresidual b-cell function in interventionstudies. This inconsistency between theGST- and MMTT-stimulated C-peptideresults has not been reported in otherlong-term intervention trials because noother trials have evaluated patients byboth procedures. In fact, these are thefirst long-term intervention trials to useboth procedures and so enable such anobservation.

Several mechanisms could contribute tothe differences in stimulated C-peptidesecretion after glucagon or a mixed-meal administration.

Glucagon stimulates the first-phaseinsulin release, which occurs within,10min, while a slow rise in blood glucosefollowing an ingested meal willstimulate the second phase of insulin

secretion. It is possible that the differentstorage compartments within the b-cellare differently affected by theprogression of diabetes and possibly bythe interventional treatment (10,17).

The incretin effect plays a role in theMMTT, but not in the GST (11). Indeed,the contribution of the incretinresponse to the MMTT was recentlydiscussed in type 1 diabetes (18).Moreover, the MMTT elicits a responsein insulin secretion because of thecombined action of gastric inhibitory

polypeptide and glucagon-like peptide-1 (GLP-1); the GST, by contrast, has adirect effect on the b-cell (19).

Endogenously secreted GLP-1 affectsthe regulation of postprandial glucoseexcursions in type 1 diabetes bymodulating glucagon levels, gastricemptying rate, and b-cellresponsiveness to glucose (18).Approximately 50% of type 1 diabeticpatients may have abnormalities ingastric emptying (without correlationwith symptoms) that could confound

Figure 2—Comparison of C-peptidestimulation methods using data from theLADA study (N = 46). The correlationsbetween the AUCs for stimulated C-peptidemeasured using the GST and MMTT at (A)baseline and (B) month 12. (C) Thecorrelation between changes in AUC (DAUC)measured by GST and MMTT from baselineto month 12. Each data point representsa single patient.

Figure 3—C-peptide and glucose levels from the DIA-AID 1 study subgroup population (N = 297).(A) C-peptide stimulation index calculated from the MMTT and GST stimulations and groupedaccording to fasting glucose levels prior to administration of the test. The numbers below thegraph represent the number of observations. (B) Fasting C-peptide (left) and fasting glucose(right) prior to C-peptide stimulation using the MMTT (white circles) or the GST (black circles) atbaseline and at study end. The horizontal bars indicate median values.

care.diabetesjournals.org Pozzilli and Associates 1389

©

interpretation of C-peptideconcentrations post-MMTT (20–22).

Thus the MMTT may be affected bygastrointestinal mechanisms triggeredby an oral nutrient load, which can bevariable during the peridiagnosis periodof type 1 diabetes (23,24).

A recent study has demonstrated thattype 1 diabetic patients in partialremission have higher levels ofproinsulin and lower levels of GLP-1 andglucagon and that GLP-1 differedsignificantly between patients inremission and not in remission (25). Itwas also shown that high proinsulin andlow GLP-1 levels measured close todiagnosis could predict remission after 1year, indicating that the actions of GLP-1influence partial remission. Theseobservations are in agreement with aprevious study showing that GLP-1 canreverse the antiproliferative effectmediated by inflammatory cytokinessuch as IL-6, TNF-a, and IFN-g in ratislets (26). Thus it could be consideredthat cytokine inflammation may be lessaggressive in patients in remission andthat the GLP-1 level as a response to thatis lower. Therefore, we may speculatethat in patients who have responded totherapy such as those treated withDiaPep277, stimulation with glucagonmay elicit a higher response ofC-peptide while stimulation with mixed-meal may elicit a lower response, thusmasking the treatment effect.

Stimulation of C-peptide is dependenton the blood glucose level (11,15,27),and this is the basis for the requirementthat the test be performed only withincertain glucose limits (4–11.1 mmol/L).Here, stimulation by the GST was stableover the permitted glucose range; incontrast, stimulation by MMTTshowed a marked decrease at higherglucose levels, similar to that shown byMadsbad et al. (27). It could bespeculated that MMTT-stimulatedC-peptide may be affected by dailyfluctuations in fasting glucose levels thatare common in patients with type 1diabetes (28,29). Regardless of anyactual change in b-cell function, asubject with increased fasting glucoselevel could manifest an attenuatedC-peptide response byMMTT but not byGST.

Both GST and MMTT are valid andaccepted methods for C-peptidestimulation (11). The MMTT is affectedby the physiological status of thepatients, which is very variable duringthe first year of type 1 diabetes. On theother hand, the GST directly stimulatesthe endogenous insulin secretionregardless of the physiological status ofthe patient.

While further research is required, theanalysis presented here suggests thatsince different stimuli measure differentoutcomes, both GST and MMTT shouldbe used for evaluating changes in insulinsecretion in newly diagnosed type 1diabetic patients.

This strategy is being implemented in ourongoing second phase 3 interventionstudy with DiaPep277 (DIA-AID 2).

Duality of Interest. Andromeda Biotechfunded the clinical trials. P.P., I.R., and I.R.C. areconsultants to Andromeda Biotech. D.P., D.E.,A.A., M.T., R.E., and S.D. are employees ofAndromeda Biotech. D.E. and I.R.C. are also theinventors of DiaPep277. No other potentialconflicts of interest relevant to this article werereported.

Author Contributions. P.P., I.R., M.T., andI.R.C. contributed to the discussion andreviewed and edited the manuscript. D.P., R.E.,and S.D. researched data and wrote themanuscript. D.E. and A.A. researched data,contributed to the discussion, and reviewed andedited the manuscript. All authors had fullaccess to the data, contributed to datainterpretation, and reviewed and approved thefinal version of this article. I.R. is the guarantorof this work and, as such, had full access to allthe data in the study and takes responsibility forthe integrity of the data and the accuracy of thedata analysis.

Prior Presentation. Parts of this study werepresented in abstract form at the followingscientific meetings: the 12th InternationalCongress of the Immunology of DiabetesSociety, City of Victoria on Vancouver Island,Canada, 15–19 June 2012; the 48th AnnualMeeting of the European Association for theStudy of Diabetes, Berlin, Germany, 1–5October 2012; and the 73rd Scientific Sessionsof the American Diabetes Association, Chicago,IL, 21–25 June 2013.

References1. Rubenstein AH, Clark JL, Melani F, Steiner

DF. Secretion of proinsulin C-peptide bypancreatic b cells and its circulation inblood. Nature 1969;224:697–699

2. Cernea S, Raz I, Herold KC, et al.; D-CureWorkshop. Challenges in developing

endpoints for type 1 diabetes interventionstudies. Diabetes Metab Res Rev 2009;25:694–704

3. Palmer JP, Fleming GA, Greenbaum CJ, et al.C-peptide is the appropriate outcomemeasure for type 1 diabetes clinical trials topreserve beta-cell function: report of anADA workshop, 21-22 October 2001.Diabetes 2004;53:250–264

4. Hendriksen C, Faber OK, Drejer J, Binder C.Prevalence of residual B-cell function ininsulin-treated diabetics evaluated bythe plasma C-peptide response tointravenous glucagon. Diabetologia 1977;13:615–619

5. The Diabetes Control and ComplicationsTrial Research Group. Effect of intensivetherapy on residual beta-cell function inpatients with type 1 diabetes in thediabetes control and complications trial.A randomized, controlled trial. Ann InternMed 1998;128:517–523

6. Steffes MW, Sibley S, Jackson M, ThomasW. Beta-cell function and the developmentof diabetes-related complications in thediabetes control and complications trial.Diabetes Care 2003;26:832–836

7. Faber OK, Binder C. C-peptide response toglucagon. A test for the residual beta-cellfunction in diabetes mellitus. Diabetes1977;26:605–610

8. Greenbaum CJ, Harrison LC; Immunology ofDiabetes Society. Guidelines forintervention trials in subjects with newlydiagnosed type 1 diabetes. Diabetes 2003;52:1059–1065

9. Samols E, Marri G, Marks V. Promotion ofInsulin Secretion by Glucagon. Lancet 1965;2:415–416

10. Cobelli C, Toffolo GM, Dalla Man C, et al.Assessment of beta-cell function inhumans, simultaneously with insulinsensitivity and hepatic extraction, fromintravenous and oral glucose tests. Am JPhysiol Endocrinol Metab 2007;293:E1–E15

11. Greenbaum CJ,Mandrup-Poulsen T,McGeePF, et al.; Type 1 Diabetes Trial NetResearch Group; European C-Peptide TrialStudy Group. Mixed-meal tolerance testversus glucagon stimulation test for theassessment of beta-cell function intherapeutic trials in type 1 diabetes.Diabetes Care 2008;31:1966–1971

12. Raz I, Ziegler AG, Linn T, et al.; DIA-AID 1Writing Group. Treatment of recent-onsettype 1 diabetic patients with DiaPep277:results of a double-blind, placebo-controlled, randomized phase 3 trial.Diabetes Care 2014;37:1392–1400

13. Mallinckrodt CH, Lane PW, Schnell D, Peng Y,Mancuso JP. Recommendations for theprimaryanalysis of continuous endpoint in longitudinalclinical trials. Drug Inf J 2008;42:303–319

14. Siddiqui O, Hung HM, O’Neill R. MMRM vs.LOCF: a comprehensive comparison based

1390 Change in C-Peptide AUC Measured by GST or MMTT Diabetes Care Volume 37, May 2014

©

on simulation study and 25 NDA datasets.J Biopharm Stat 2009;19:227–246

15. Besser RE, Shields BM, Casas R, HattersleyAT, Ludvigsson J. Lessons from the mixed-meal tolerance test: use of 90-minute andfasting C-peptide in pediatric diabetes.Diabetes Care 2013;36:195–201

16. Faber OK, Madsbad S, Kehlet H, Binder C.Pancreatic beta cell secretion duringoral and intravenous glucoseadministration. Acta Med Scand Suppl1979;624:61–64

17. Grodsky GM. Kinetics of insulin secretion. InDiabetes Mellitus: A Fundamental andClinical Text. LeRoith D, Taylor SI, OlefskyJM, Eds. Philadelphia, Lippincott Williams &Wilkins, 2000, p. 2–11

18. Kielgast U, Holst JJ, Madsbad S. Antidiabeticactions of endogenous and exogenousGLP-1 in type 1 diabetic patients with andwithout residual b-cell function. Diabetes2011;60:1599–1607

19. Samols E, Marri G, Marks V.Interrelationship of glucagon, insulin andglucose. The insulinogenic effect ofglucagon. Diabetes 1966;15:855–866

20. Fraser RJ, Horowitz M, Maddox AF, HardingPE, Chatterton BE, Dent J. Hyperglycaemiaslows gastric emptying in type 1 (insulin-

dependent) diabetes mellitus. Diabetologia1990;33:675–680

21. De Block CE, De Leeuw IH, Pelckmans PA,Callens D, Maday E, Van Gaal LF. Delayedgastric emptying and gastric autoimmunityin type 1 diabetes. Diabetes Care 2002;25:912–917

22. Schvarcz E, Palmer M, Aman J, Lindkvist B,Beckman KW. Hypoglycaemia increases thegastric emptying rate in patients with type1 diabetes mellitus. Diabet Med 1993;10:660–663

23. Sosenko JM, Palmer JP, Rafkin LE, et al.;Diabetes Prevention Trial-Type 1 StudyGroup. Trends of earlier and laterresponses of C-peptide to oral glucosechallenges with progression to type 1diabetes in diabetes prevention trial-type 1participants. Diabetes Care 2010;33:620–625

24. Sosenko JM, Palmer JP, Rafkin-Mervis L,et al. Glucose and C-peptide changes in theperionset period of type 1 diabetes in theDiabetes Prevention Trial-Type 1. DiabetesCare 2008;31:2188–2192

25. Kaas A, Andersen ML, Fredheim S, et al.;Hvidoere Study Group on childhooddiabetes. Proinsulin, GLP-1, and glucagonare associated with partial remission

in children and adolescents with newlydiagnosed type 1 diabetes. PediatrDiabetes 2012;13:51–58

26. Blandino-Rosano M, Perez-Arana G,Mellado-Gil JM, Segundo C, Aguilar-Diosdado M. Anti-proliferative effect ofpro-inflammatory cytokines in culturedbeta cells is associated with extracellularsignal-regulated kinase 1/2 pathwayinhibition: protective role of glucagon-likepeptide -1. J Mol Endocrinol 2008;41:35–44

27. Madsbad S, Sauerbrey N, Møller-Jensen B,Krarup T, Kuhl C. Outcome of theglucagon test depends upon theprevailing blood glucose concentration intype I (insulin-dependent) diabeticpatients. Acta Med Scand 1987;222:71–74

28. Schmidt MI, Hadji-Georgopoulos A, RendellM, Margolis S, Kowarski A. The dawnphenomenon, an early morning glucoserise: implications for diabetic intradayblood glucose variation. Diabetes Care1981;4:579–585

29. Bolli GB, Perriello G, Fanelli CG, De Feo P.Nocturnal blood glucose control in type Idiabetes mellitus. Diabetes Care 1993;16(Suppl. 3):71–89

care.diabetesjournals.org Pozzilli and Associates 1391

©

SUPPLEMENTARY DATA  

©2013 American Diabetes Association. Published online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-1392/-/DC1

Evaluation of Long-Term Treatment Effect in an Intervention Trial in Type 1 Diabetes: Differences after Stimulation with Glucagon or a Mixed-Meal Pearson’s Correlations between GST and MMTT measurements in the DiaPep277® and Placebo patients. Pearson’s correlations between the absolute C-peptide AUC values at specific time points as obtained using the GST and the MMTT methods were performed separately on patients treated with DiaPep277® or placebo in the Subgroup population. In the placebo group, the correlation coefficients (r) for C-peptide AUC values obtained by the GST and MMTT methods at individual time points were 0.73, 0.81, and 0.86 at baseline, month 12, and month 24, respectively (Figures S1A, B, C). In DiaPep277®-treated patients, the corresponding correlation coefficients were r = 0.77, 0.80, and 0.92 at baseline, month 12, and month 24, respectively (Figures S2 A, B, C) Correlations between the two stimulated C-peptide measurement methods were also obtained across time periods (ΔAUC). The correlations in the placebo group were: r= 0.57 across the whole study period (baseline–month 24); r = 0.35 during the first year of the study (baseline–month 12); and r = 0.39 during the second year (month 12–month 24) (Supplemental Figure S1D, E, F). In the DiaPep277® group, the corresponding correlations were: r= 0.59 across the whole study period (baseline–month 24); r= 0.42 during the first year of the study (baseline–month12); and r = 0.49 during the second year (month 12–month 24) (Supplemental Figure S2 D, E, F). Bland-Altman plots were used to assess homogeneity of error variation over the range of subject values. The difference between the percent change in AUC by MMTT and that by GST was plotted against the average percent change of the two methods for each subject (Bland JM and Altman DG. Statistical methods for assessing agreement between two methods of clinical measurements. Lancet 8476: 307-10, 1986) (Supplemental Figure S3). The Kappa coefficient agreement is 0.53.  

SUPPLEMENTARY DATA  

©2013 American Diabetes Association. Published online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-1392/-/DC1

r=0.81

0

0.5

1

1.5

2

2.5

3

3.5

0 0.5 1 1.5 2 2.5

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC GST(nmol/l/min)

r = 0.86

0

0.5

1

1.5

2

2.5

3

3.5

0 0.5 1 1.5 2 2.5

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC by GST (nmol/l/min)

r = 0.35

-2

-1.5

-1

-0.5

0

0.5

1

-1.6 -1.3 -1 -0.7 -0.4 -0.1 0.2 0.5 0.8

AUC by GST (nmol/l/min)

A

UC

by M

MT

T (

nm

ol/

l/m

in)

r = 0.39

-2

-1.5

-1

-0.5

0

0.5

1

-1.6 -1.3 -1 -0.7 -0.4 -0.1 0.2 0.5 0.8

AUC by GST (nmol/l/min)

A

UC

by M

MT

T (

nm

ol/

l/m

in)

r=0.57

-2

-1.5

-1

-0.5

0

0.5

1

-1.6 -1.3 -1 -0.7 -0.4 -0.1 0.2 0.5 0.8

AUC by GST (nmol/l/min)

A

UC

by

MM

TT

(n

mo

l/l/

min

)

Supplementary Figure 1. Placebo-treated patients in the Subgroup Population. Correlations between the GST and MMTT methods for obtaining the area under the curve (AUC) for stimulated C-peptide secretion at: (A) baseline; (B) month 12; and (C) month 24. Correlations between the GST and MMTT methods for obtaining changes in the AUC values (∆AUC): (D) from baseline to month 24; (E) from baseline to month 12; (F) from month 12 to month 24. Each data point represents a single patient.  

r =0.73

0

0.5

1

1.5

2

2.5

3

3.5

0 0.5 1 1.5 2 2.5

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC by GST (nmol/l/min)

SUPPLEMENTARY DATA  

©2013 American Diabetes Association. Published online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-1392/-/DC1

Supplementary Figure 2. DiaPep277®-treated patients in Subgroup Population.Correlations between the GST and MMTT methods for obtaining the area under the curve (AUC) for stimulated C-peptide secretion at: (A) baseline; (B) month 12; and (C) month 24 . Correlations between the GST and MMTT methods for obtaining changes in the AUC values (∆AUC): (D) from baseline to month 24; (E) from baseline to month 12; (F) from month 12 to month 24. Each data point represents a single patient.

0

0.5

1

1.5

2

2.5

3

3.5

4

0 0.5 1 1.5 2

r=0.92

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC by GST (nmol/l/min)

0

0.5

1

1.5

2

2.5

3

3.5

4

0 0.5 1 1.5 2

r=0.8

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC by GST (nmol/l/min)

0

0.5

1

1.5

2

2.5

3

3.5

4

0 0.5 1 1.5 2

r=0.77

AU

C b

y M

MT

T (

nm

ol/

l/m

in)

AUC by GST (nmol/l/min)

-1.5

-1

-0.5

0

0.5

1

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

r=0.49

A

UC

by

MM

TT

(n

mo

l/l/

min

)

AUC by GST (nmol/l/min)

-1.5

-1

-0.5

0

0.5

1

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

r=0.42

A

UC

by

MM

TT

(n

mo

l/l/

min

)

AUC by GST (nmol/l/min)

-1.5

-1

-0.5

0

0.5

1

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

r=0.59

A

UC

by

MM

TT

(n

mo

l/l/

min

)

AUC by GST (nmol/l/min)

SUPPLEMENTARY DATA  

©2013 American Diabetes Association. Published online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-1392/-/DC1

Supplementary Figure 3. Bland-Altman Plot: Average and difference of AUC Relative Change (%) GST and MMTT.