| Objective: The euglycemic clamp is the gold standard for assessing the pharmacokinetics(PK)and pharmacodynamics(PD)of insulin preparations.When applying the euglycemic clamp,many problems always arise: e.g.(1)how to evaluate the qualified quality of the euglycemic clamp which is directly related to the accuracy of the pharmacokinetic and pharmacodynamic parameters of the tested insulin;(2)the euglycemic clamp to assess the PK/PD of the insulin preparation has been modified to be performed without the establishment of hyperinsulinemia which was different from the original and the classical method that was accompanied by artificial hyperinsulinemia,and whether this change would affect the PK/PD of the tested insulin or not is unknown;(3)PD parameters are always derived from the GIR versus time profile that is unsmoothed due to frequent adjustment of 20% dextrose according to glucose variability,extremely rough in automatic clamps resulting in containing noise even wrong information.Locally weighted regression(LOESS)is always used in automatic clamps to reduce the bias.In a manual clamp,the GIR-Time curve is much less unsmoothed,therefore,is it necessary to smooth the GIR-Time curve?;(4)when the euglycemic clamp is performed in the healthy volunteers,the endogenous insulin secretion should be inhibited to ensure that PK/PD of the tested insulin is free from the interference of the endogenous insulin.We evaluate the interference of endogenous insulin by monitoring whether the C-peptide level is below the baseline or not.However,the phenomenon that C-peptide could fluctuate upon and below the baseline in the clinical research is not occasional.Whether the rise of C-peptide from baseline indicates the unsuppressed endogenous insulin secretion and would affect the PK/PD of the tested insulin is not clear;(5)Postdosing C-peptide may be higher than the baseline level during euglycemic clamp studies.The effect of inappropriately increased postdosing C-peptide on PK parameters could be ignored by a specific detection method for the tested insulin(e.g.high-pressure liquid mass spectrometry method),while the effect on PD parameters is impossible to be neglected at present.It is not clear that the extent to which the postdosing C-peptide increases from the baseline could significantly affect the accuracy of the PD values of the studied insulin preparation.To address the above issues,the study aimed(1)to establish the quality control indexes and evaluate whether these indexes are suitable for the quality assessment of the euglycemic clamps;(2)to evaluate the effect of artificial hyperinsulinemia on the assessment of PK/PD of insulin preparations in euglycemic clamps;(3)to assess the effect of the LOESS smoothing method on the PD parameters in the manual euglycemic clamp;(4)to verify the effect of inappropriately increased postdosing C-peptide on PK/PD parameters and(5)to explore the extent to which the postdosing C-peptide increases from the baseline could significantly affect the accuracy of the PD.Materials and Methods:(1)As for the quality control of euglycemic clamps,five parameters as follows were additionally included for the quality control besides the world-wild used parameter--coefficient of variation of the blood glucose concentrations(CVBG):(1)the percentages of glucose excursion from target range(GEFTR),(2)the duration of GEFTR,(3)the area under the curve(AUC)of GEFTR,(4)the mean value of excursion from target glucose(GEFT)and(5)the AUC of GEFT.Data collected from 80 euglycemic glucose clamp studies performed from 2014 to 2017 were divided into 4 groups according to CVBG: group A(CVBG≤4.5%),group B(4.5%<CVBG≤5.0%),group C(5.0%<CVBG≤5.5%),and group D(CVBG>5.5%).GEFTR,the duration of GEFTR,the AUC of GEFTR,the mean value of GEFT,and the AUC of GEFT were calculated and compared among them.The receiver operating characteristic(ROC)curve was drawn and the sensitivity and specificity of each quality control index were calculated.(2)To explore the effect of hyperinsulinemia on PK/PD of the test insulin,a total of 40 healthy male volunteers aged 18~45 years old in West China Hospital between 2015 and 2017 who underwent a euglycemic clamp were included according to the inclusion criteria.They were divided into euglycemic-hyperinsulinaemic clamp(A)group and euglycemic clamp(B)group according to the enrolling sequence.The tested short-acting human insulin(Humulin R,0.2 IU/kg)was given subcutaneously at the steady-state of the clamp after infusion of short-acting insulin in group A while in group B Humulin R was given subcutaneously without the establishment of artificial hyperinsulinemia on the testing day.The blood glucose was maintained within the target range during the whole euglycemic clamp.Blood samples were collected at the predefined time after subcutaneous injection of the tested insulin for the analysis of blood glucose,insulin concentration,and C-peptide levels.The PK/PD parameters were calculated.(3)To evaluate the necessity of the LOESS smoothing method in the manual euglycemic clamp,a total of 68 healthy volunteers who underwent a manual euglycemic glucose clamp in West China Hospital of Sichuan University from 2017 to 2019 were enrolled.24 of them in group A received a subcutaneous injection of regular human insulin,24 of them in group B received a subcutaneous injection of insulin aspart and 20 of them in group C received a subcutaneous injection of insulin glargine.Euglycemia was maintained during the entire clamp.The longest period of the clamp lasts 10 hours for regular insulin and insulin aspart,24 hours for insulin glargine.PD parameters included the maximum of glucose infusion rate(GIRmax),time to GIRmax(GIR-Tmax),the early and late time to the half of GIRmax(GIR-Tearly50% and GIR-Tlate50%),and the area under the GIR-Time cure from 0 h to the end of the clamp for regular insulin,insulin aspart and insulin glargine,additionally area under the GIR-Time cure from 0 h to 6 h and 12 h were calculated for insulin glargine.Compare the differences of the PD parameters in each group derived from the GIR-Time curve before and after fitted with a LOESS.The smoothing parameter was 0.25.(4)To verify the effect of the rise of postdosing C-peptide on PK/PD,this study included 33 males who underwent a manual euglycemic clamp with a subcutaneous injection of insulin aspart(IAsp,0.2 IU/kg).Time-profiles of whole blood glucose,human insulin,glucose infusion rate(GIR),and C-peptide(CP)were recorded.They were divided into two groups: group A [(CPpostdosing)max > CPbaseline],group B [(CPpostdosing)max ≤ CPbaseline] with an allocation ratio of 2:1.Total endogenous insulin was roughly equal to the measurement of human insulin or calculated by C-peptide.The relationship between human insulin and C-peptide concentrations during clamp was examined in two groups and the factors that affect the C-peptide concentrations were analyzed.The insulin aspart PK/PD values were also analyzed.(5)To explore the oscillations of C-peptide in the euglycemic clamp and the extent to which the postdosing C-peptide increases could affect the pharmacodynamic assessment of insulin preparations,first,10 healthy males underwent a 10-h euglycemic clamp without exogenous insulin administration to obtain a reference interval(RI)for the ratio of baseline C-peptide(CP0)to C-peptide at each time point thereafter(CPt).Then,data of pharmacokinetics and pharmacodynamics of insulin aspart(IAsp)were analyzed,and 70 eligible euglycemic clamps were grouped by CPt/CP0: group A([CPt/CP0]max>upper limit of RI),group B(1<[CPt/CP0]max ≤ upper limit of RI),and group C([CPt/CP0]max≤1).The differences in basal and clamped blood glucose,CPt/CP0,and IAsp’s PK/PD were compared among groups,and the relationship between elevated CPt and the accuracy of pharmacodynamics was analyzed.Results:(1)There were 37,14,9,and 20 in groups A,B,C,and D respectively.The median values of CVBG were 3.80%,4.76%,5.27%,and 6.07% in groups A,B,C,and D respectively.The percentages of GEFTR,the duration of GEFTR,the AUC of GEFTR,the mean value of GEFT,and the AUC of GEFT in group A were all less than those of other groups(P<0.05),and these indexes were similar between group B and C and higher in group D.CVBG was positively correlated with other quality control indexes(correlation coefficient r was 0.770-0.805).The cut-off points of the percentages of GEFTR,the duration of GEFTR,the AUC of GEFTR,the mean value of GEFT,and the AUC of GEFT were 5.8%,14.6 min,22.8 mg/d L×min,3.23 mg/d L,216 mg/d L×min/h respectively according to the cut-off point 5% of CVBG,with a sensitivity of 79.3%~100% and a specificity of 74.5%~89.7%.Applying these indexes to evaluate the quality of these euglycemic glucose clamps,8.11% of euglycemic glucose clamps were found to be of less good quality in group A,while 66.67% of euglycemic glucose clamps were found to be with acceptable quality in group C.(2)No differences were detected in age,height,weight,body mass index(BMI),the dosage of Humulin R,and the ‘clamped’ glucose between the two groups.After injection of tested insulin,maximum insulin concentration(667±141 vs.267 ± 68 pmol/L,P<0.01)and area under the curve(AUC)of insulin concentration(152±32 vs.57±7 nmol/L×min,P<0.01)in A group were higher while maximum glucose infusion rate(GIR)(3.70±0.70 vs.7.66±2.11 mg/kg/min,P<0.01)and AUC of GIR from 0 to 8 hours(931±272 vs.1920±452 mg/kg,P<0.01)were lower compared to B group.The serum C-peptide levels were lower to a different extent in both groups after administration of insulin compared with the basal levels(group A 30.9%±16.4%,group B 49.2%±10.8%,P<0.01).(3)The overall value of the coefficient of variation of blood glucose was ranging from 3.15% to 4.18% among the three groups.No difference was detected in PD parameters mentioned above of regular human insulin,insulin aspart,and insulin glargine derived from the GIR-Time curve before and after LOESS.(4)Basal BG concentration and CVBG were comparable between the two groups.The ‘clamped’ BG concentrations were 99.7%±7.1%(group A)and 94.9%±5.1%(group B)of baseline respectively.Insulin aspart PK values were consistent in the two groups while AUCGIR,0~8h was slightly higher in group A than group B(1815±551 vs.1327±306 mg/kg,P=0.01).The serum C-peptide(308 vs.299 pmol/L,P=0.75)and human insulin(24.0±7.4 vs.29.5±10.2 pmol/L,P=0.09)levels were of no significant difference at baseline between the two groups but group B had a much lower serum C-peptide concentration(168 vs.309 pmol/L,P<0.01)and human insulin level(17.3 vs.26.0 pmol/L,P<0.01)when compared with group A after dosing.The method using serum C-peptide concentration to predict endogenous insulin secretion had a sensitivity of 85.19%(138/162)and a specificity of 79.20%(316/399).CVBG and the extent of serum C-peptide after dosing higher than baseline were positively correlated(r=0.51,P=0.01).The AUC of clamped BG higher than the target BG and the AUC of serum human insulin higher than baseline shared the same trend.The change of the ratio of serum human insulin to C-peptide made sense in group A while it was of no significant difference in group B.Predicting the endogenous insulin secretion by C-peptide was significantly different from that by human insulin(80.6±24.2 vs.91.0±28.4 ng/m L×min,P=0.033)in group A,while no statistical difference was detected in group B(54.0±21.1 vs.63.0±27.3 ng/m L×min,P=0.14).(5)The RI of CPt/CP0 was 22.7%~152.1% in euglycemic clamps without exogenous insulin injection;1.5×baseline might be the ceiling for the rise of CPt under stable conditions.There were 9,45,16 euglycemic clamps in groups A[(CPt/CP0)max > 1.5],B[1 <(CPt/CP0)max ≤ 1.5],and C[(CPt/CP0)max ≤ 1] respectively.The AUCIAsp,0~10h were 502±116,511±105,and 517±88 ng/m L×min in groups A,B,and C respectively(P=0.958).The maximum glucose infusion rate(GIR)tended to be higher in group A than group B or C(Pfor trend=0.033).Although no significant difference(P=0.070)was detected in AUCGIR,0~8h among the groups,AUCGIR,0~8h tended to be higher accompanied by an increase in CPt(Pfor trend=0.022).The AUCGIR,0~10h was 1983±650,1682±454,and 1479±440 mg/kg in groups A,B,and C(P=0.047).No intergroup difference was detected in clamped glucose,IAsp dose,or body mass index.Conclusion:(1)CVBG less than 4.5% indicates good quality,and the above-mentioned quality control indexes especially the AUC of GEFT(cut-off point: 216 mg/d L×min/h)should be evaluated when CVBG is from 4.5% to 5.5%.False high quality and false low-quality euglycemic clamps might be detected and a more precise estimation of quality assessment could be made with the help of these indexes.(2)Applying a euglycemic-hyperinsulinaemic clamp to evaluate the PK/PD of insulin preparations might overestimate the PK parameters and underestimate the PD parameters of insulin preparations.The euglycemic glucose clamp without artificial hyperinsulinemia might obtain a more accurate value of the PK/PD parameter with a simple and convenient procedure.(3)The smoothing method for the GIR-Time curve in the manual euglycemic glucose clamps with good quality might have no significant effect on PD parameters.(4)Inappropriate elevated serum C-peptide concentrations after exogenous insulin administration indicate insufficient inhibition of endogenous insulin secretion in a euglycemic glucose clamp.The BG during euglycemic glucose clamp was preferably controlled below the baseline level to sufficiently inhibit endogenous insulin secretion of the subject.Additionally,the high fluctuation of BG in the clamp might also be one of the factors leading to inadequate inhibition of endogenous insulin.The estimated insulin PK values obtained from the C-peptide correction method might not be very accurate if C-peptide level was higher than the baseline after dosing during the euglycemic glucose clamp,and if C-peptide was continuously inhibited the method using C-peptide to correct the endogenous insulin would still be valid even in the clamp evaluating the rapid-action insulin.(5)A postdosing C-peptide over 1.5×baseline would compromise the pharmacodynamics of insulin preparation. |
PDF Full Text Request