Pharmacokinetics andPharmacodynamics of InsulinGlargine Given in the Evening asCompared With in the Morningin Type 2 DiabetesDiabetes Care 2015;38:503–512 | DOI: 10.2337/dc14-0649

OBJECTIVE

To compare pharmacokinetics (PK) and pharmacodynamics (PD) of insulin glarginein type 2 diabetes mellitus (T2DM) after evening versus morning administration.

RESEARCH DESIGN AND METHODS

Ten T2DM insulin-treated persons were studied during 24-h euglycemic glucoseclamp, after glargine injection (0.4 units/kg s.c.), either in the evening (2200 h) orthe morning (1000 h).

RESULTS

The 24-h glucose infusion rate area under the curve (AUC0–24h) was similar in theevening andmorning studies (1,0586 571 and 9956 691mg/kg3 24 h, P = 0.503),but the first 12 h (AUC0–12h) was lower with evening versus morning glargine(3576 244 vs. 5936 374mg/kg3 12 h, P = 0.004), whereas the opposite occurredfor the second 12 h (AUC12–24h 7006 396 vs. 4036 343 mg/kg3 24 h, P = 0.002).The glucose infusion rate differences were totally accounted for by different ratesof endogenous glucose production, not utilization. Plasma insulin and C-peptidelevels did not differ in evening versus morning studies. Plasma glucagon levels(AUC0–24h 1,533 6 656 vs. 1,120 6 344 ng/L/h, P = 0.027) and lipolysis (free fattyacid AUC0–24h 7.56 1.6 vs. 8.96 1.9 mmol/L/h, P = 0.005; b-OH-butyrate AUC0–24h6.8 6 4.7 vs. 17.0 6 11.9 mmol/L/h, P = 0.005; glycerol, P < 0.020) were overallmore suppressed after evening versus morning glargine administration.

CONCLUSIONS

The PD of insulin glargine differs depending on time of administration. Withmorning administration insulin activity is greater in the first 0–12 h, while withevening administration the activity is greater in the 12–24 h period followingdosing. However, glargine PK and plasma C-peptide levels were similar, as wellas glargine PD when analyzed by 24-h clock time independent of the time ofadministration. Thus, the results reflect the impact of circadian changes in insulinsensitivity in T2DM (lower in the night-early morning vs. afternoon hours) ratherthan glargine per se.

Department of Medicine, University of PerugiaMedical School, Perugia, Italy

Corresponding author: Geremia B. Bolli, geremia.bolli@unipg.it.

Received 12 March 2014 and accepted 10November 2014.

Clinical trial reg. no. EudraCT2010-019368-35,http://eudract.ema.europa.eu/.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc14-0649/-/DC1.

This is an independent, investigator-designedproject, neither shared with nor supported byany pharmaceutical company.

© 2015 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered.

Francesca Porcellati, Paola Lucidi,

Patrizia Cioli, Paola Candeloro,

Anna Marinelli Andreoli,

Stefania Marzotti, Maura Ambrogi,

Geremia B. Bolli, and Carmine G. Fanelli

Diabetes Care Volume 38, March 2015 503

EMER

GINGTEC

HNOLO

GIES

ANDTH

ERAPEU

TICS

The insulin dynamic of the postabsorptiveand interprandial state of nondiabeticsubjects (basal insulin) is best repro-duced in people with diabetes requiringinsulin administration by the long-actinginsulin analogs compared with NPH insu-lin (1). This more physiological substitu-tion translates into clinical benefitscompared with NPH, primarily a reducedrisk of nocturnal hypoglycemia for similarA1C levels (1–7).NPH insulin is most often injected at

bedtime to limit the risk of nocturnalhypoglycemia due to its peak of action4–6 h postinjection (1–3). In contrast, thesmoother activity of the long-acting insulinanalog glargine allows more flexibility indosing, and its administration is less strictlybound to the time of injection. Hence, in-sulin glargine is approved for administra-tion at any time of day, provided it is at thesame time each day (8).However, only a few studies have com-

pared the effects of glargine given in theevening comparedwith other times ofdayboth in type 1 diabetes mellitus (9–11) aswell as in type 2 diabetes mellitus (T2DM)(12,13), with results basically consistentin the noninferiority of morning versusevening dosing. Yet, the large majority ofpeople with diabetes inject glargine in theevening, often at bedtime, in keepingwiththe tradition of NPH administration.One reasonwhy the potentially greater

flexibility of glargine in the treatmentof T2DM has not translated into a rec-ognized option of administration at adifferent time of day compared withNPH, might be the lack of head-to-head studies of pharmacokinetics (PK)and pharmacodynamics (PD), compar-ing glargine given in the evening versusthe morning. Yet, an indirect compari-son of PK and PD between studies,where glargine was given in the evening(14) compared with the morning (15–17), suggests important differences inaction profiles, although in such a com-parison the contribution of circadianchanges in insulin sensitivity should betaken into account (18,19).The aim of the current study was to

perform a head-to-head comparison ofthe PK and PD of insulin glargine in per-sons with T2DM, after either evening ormorning administration.

RESEARCH DESIGN AND METHODS

After approval by the local ethics commit-tee and after written informed consent

was obtained, 10 T2DM persons receiv-ing insulin therapy (all glargine as basalwith/without mealtime rapid-acting in-sulin and oral hypoglycemic agents)were recruited (Table 1) and studied ac-cording to the guidelines of the HelsinkiDeclaration and good clinical practicerequirements. Subjects were free ofany detectable microangiopathic andmacroangiopathic complications and ofany major illness other than diabetes,as indicated by medical history, physicalexamination, electrocardiogram, or rou-tine laboratory tests. All of the subjectswere receiving insulin glargine with theevening meal.

The study has a randomized, single-dose, open-label, crossover design (Sup-plementary Fig. 1). After a 2-week run-inperiod, during which subjects with dia-betes continued their treatment, andthe dose of basal insulin glargine wasoptimized to achieve a fasting plasmaglucose (PG) concentration of #100mg/dL while avoiding nocturnal hypo-glycemia (PG ,72 mg/dL), subjectswere randomized to two 24-h euglyce-mic clamp studies, after either evening(2200 h) or morning (1000 h) subcuta-neous injection of insulin glargine (0.4units/kg). Each occasion was separatedby an interval of 3–4 weeks, duringwhich the usual insulin regimen of therun-in period was continued (glargine atdinnertime).

After randomization, subjects wereswitched in a timely manner from

dinnertime to either morning (1000 h,N = 5) or evening (2200 h, N = 5) glargineadministration 9 days (day 29) prior tothe study day (day 0). For morning dos-ing, on day29, subjects injected 50% ofthe usual glargine dose at dinnertime,followed by the remaining 50% thenext morning at 1000 h (day 28). Onday 27 and thereafter until the studyday, the full glargine dose was injectedat 1000 h every day. For evening dosing,subjects injected glargine at 2200 h ev-ery day from day29 until the study day.With either morning or evening dosing,basal insulin concentration was titratedas in the run-in period, and treatmentwith mealtime rapid-acting insulin wascontinued, if patients had been receiv-ing such treatment.

On both evening and morning studyoccasions, subjects had the last subcu-taneous glargine injection 24 h beforethe time they received glargine on thestudy day. An equivalent interval of fast-ing was observed before initiation ofeach clamp study (14 h), with the lastmeal (688 kcal, 54% carbohydrate, 30%protein, and 16% lipids) consumed ei-ther at 0800 h the same day of the study(glargine evening) or at 2000 h the daybefore the study (glargine morning).Care was taken by subjects to avoid hy-poglycemia (PG,72 mg/dL) with bloodglucose self-monitoring in the 48 h be-fore the clamp studies. For the entireduration of the studies, all persons mon-itored blood glucose levels by means

Table 1—Demographic and baseline clinical characteristics of the subjects

Gla at 1000–2200 ha Gla at 2200–1000 ha

P valueN = 10b,c N = 5 N = 5

SexMale 7 3 4Female 3 2 1

Age (years) 61.6 6 4.5 62.2 6 3.4 60.8 6 4.1 0.573

BMI (kg/m2) 28.0 6 3.6 28.5 6 4.3 27.5 6 4.0 0.713

Known diabetesduration (years) 16.5 6 8.0 16.4 6 3.2 16.6 6 11.5 0.971

A1C 0.801% 7.2 6 0.6 7.1 6 0.5 7.2 6 0.7mmol/mol 55 6 8.1 54 6 7.9 55 6 9.0

Daily insulin dose(units/kg)c

Total 0.53 6 0.31 0.46 6 0.29 0.60 6 0.34 0.503Basal 0.28 6 0.11 0.27 6 0.12 0.30 6 0.12 0.703

Data are mean6 SD, unless otherwise indicated. aSubjects split by sequence of treatments. bAll 10subjects were receiving basal insulin glargine, 7 on basal bolus (glargine plus prandial rapid-actinganalog); 8 subjectswerealso receivingoral agents (metformin, 5 subjects onbasal bolus and2 subjectsreceiving a basal insulin regimen; 1 subject was receiving a basal insulin regimen plus 1 mgglimepiride). cPrandial (if any, rapid-acting insulin analog) plus basal (glargine insulin) regimen.

504 Glargine Evening vs. Morning in T2DM Diabetes Care Volume 38, March 2015

of a reflectometer (OneTouch Vita; Life-Scan Italy [part of Johnson & JohnsonMedical S.p.A.], Milan, Italy).

Euglycemic Clamp ProcedurePersons with diabetes were admitted inthe fasting state to the Clinical ResearchUnit of the Department of Medicine,Section of Internal Medicine, Endocri-nology and Metabolism of the Univer-sity of Perugia Medical School, 4 hprior to study initiation (i.e., at either0600 or 1800 h, respectively, for themorning and evening studies), and re-mained fasting for the subsequent 28 h(Supplementary Fig. 1). Theywere put atrest in a bed and studied with the eugly-cemic clamp technique (20). In brief,two venous lines were inserted, an ante-cubital vein on one arm (for infusions)and a dorsal vein (cannulated retro-gradely with a 20-gauge butterfly nee-dle) on the contralateral hand (kept at658C in a hot box) for intermittent sam-pling of arterialized venous blood. Anintravenous feedback insulin infusionwas initiated to achieve and maintainPG at 100 6 5 mg/dL until time 0 (i.e.,1000 or 2200 h), as previously described(21). At 2120 min, a primed sterile,pyrogen-free constant infusion (0.222mmol/kg/min) of [6,6-2H2]-glucose(Cambridge Isotope Laboratories, Cam-bridge, MA) was started and maintainedthroughout the experiment to deter-mine glucose kinetics. A variable rateof a 20% dextrose solution enriched to4% with [6,6-2H2]-glucose was used toclamp the PG concentration at the de-sired target of 100 mg/dL for 24 h (14)and to avoid non–steady-state errors inthe measurement of glucose turnover,as described previously (22). At 1000 or2200 h, a subcutaneous injection of 0.4units/kg insulin glargine (Lantus 100units/mL) was administered by meansof injection with a 0.5-mL insulin syringeinto the abdominal area. On both occa-sions, the clamp procedure was termi-nated at 24 h after the subcutaneousinjection of glargine.

Analytical MethodsPG concentration was measured atthe bedside using the YSI 2300 glucoseanalyzer (YSI 2300 STAT; Yellow SpringsInstruments, Yellow Springs, OH). Theplasma free insulin concentration wasmeasured by radioimmunoassay afterpolyethylene glycol extraction of anti-bodies from the plasma (23). Plasma

free fatty acid (FFA), glycerol, b-OH-butyrate, C-peptide, and glucagonconcentrations were measured bypreviously described methods (14).Glucose enrichment was determinedon its penta-acetate (penta-O-acetyl-b-D-glucopyranose) derivative by gaschromatography–mass spectrometry(HP 6890 II GC, HP 5973A GC/MS;Hewlett-Packard, Palo Alto, CA) inelectron impact ionization mode mon-itoring the ions 200 and 202 for theunlabeled and [6,6-2H2]-glucose, respec-tively (24).

Calculations and Statistical Analysis

Calculations of glucose fluxes are basedon a non–steady-state assumption andthe total Ra and Rd values were calcu-lated using amodified form of the Steeleequation to account for the addition ofstable labeled tracer to the exogenousglucose infusate (24). Endogenous glu-cose production (EGP) was obtained asthe difference between Ra and the ex-ogenous glucose infusion rate (GIR) dur-ing the clamp. Because the isotopicsteady state might have not beenachieved during the first hours of thestudy (25), in order to have reliable val-ues of glucose fluxes, calculations wereperformed by excluding the first 6 h ofthe study.

The linear trapezoidal rule was usedto calculate the concentration-timearea under the curve (AUC) between0 and 24 h (AUC0–24h) for GIRs, plasmafree insulin, and nonglucose substrates.The maximum plasma concentration(Cmax) and the time to reach Cmax

(Tmax) for the same variables were readdirectly from the plasma concentration-time data for each subject. The determi-nations of Cmax and Tmax for the GIRwere derived from a smoothed 3-pointrunning average GIR curve for eachsubject in order to provide reliabledata for calculation.

The primary analysis of the PK andPD parameters was performed usingANOVA, which allowed for variationof subjects nested within sequence, pe-riod, and treatment. Tmax variables wereanalyzed nonparametrically (Wilcoxonrank sum test and Hodges-Lehmannestimates of the treatment with 95%confidence limits).

The primary end point of the studywasthe GIR over the first 12 h (AUC0–12h)and the GIR over the second 12 h

(AUC12–24h). Secondary end pointswere PG concentration; glucose fluxes(AUC7–24h); plasma C-peptide, glucagon,and nonglucose substrate concentra-tions; intravenous insulin infusion ratesprior to and immediately after subcu-taneous glargine insulin injection;and plasma insulin concentrations. Asample size of 10 subjects was chosenin order to achieve 80% power todetect a mean of paired differencesof 40% with an estimated SD of differ-ences of 40% between the GIR for glar-gine infused during both the first 12 h(AUC0–12h for glargine at 1000 h vs. glar-gine at 2200 h) and the second 12 h(AUC12–24h for glargine at 1000 h vs. glar-gine at 2200 h). All tests of statisticalhypothesis were carried out at the 5%level of significance and comparisonswere two sided. Data in the text are ex-pressed as the mean 6 SD and medianwith 25th and 75th percentiles as appro-priate, and in figures as the mean 6 SE.Statistical analysis was usually per-formed using PASS (NCSS StatisticalSoftware, Kaysville, UT).

RESULTS

All 10 subjects enrolled completed thetwo studies.

Glycemic Control and Insulin Doses inthe Week Prior to StudiesAfter randomization, home daily glycemiccontrol (mean PG from self-monitoringdata) during the 7 days before theclamp studies was not different in themorning study (139 6 11 mg/dL) versusthe evening study (142 6 7 mg/dL, P =0.221). The dose of glargine over the 7days prior to the studies was slightly, butsignificantly, higher in the morning ver-sus the evening study (0.31 6 0.03 vs.0.29 6 0.03 units/kg, P = 0.002). Also,total daily insulin doses were greater inthe morning (0.56 6 0.3 units/kg) com-pared with the evening study (0.55 60.3 units/kg), although the differencedid not reach statistical significance(P = 0.070).

Rates of Intravenous Insulin InfusionPrior to Subcutaneous InsulinInjection, Plasma Insulin and GlucoseConcentrations, GIRs, Rates ofEndogenous Glucose Production, andPeripheral Glucose UtilizationThe amount of insulin infused to achieveand maintain euglycemia during thefeedback period (time 22 to 2 h before

care.diabetesjournals.org Porcellati and Associates 505

and after time 0 h) was greater withglargine given in the morning (0.3 60.2 mU/kg/min) compared with theevening (0.2 6 0.2 mU/kg/min), butthe difference did not reach statistical

significance (P = 0.190). After thesubcutaneous injection of glargineat time 0 h, the rate of intravenousinsulin infusion decreased in bothtreatments. However, the withdrawal

of intravenous insulin infusion tendedto be faster with evening (time 21 61 h) compared with morning glarginedosing (time 0.3 6 1.3 h) (P , 0.065)(Fig. 1 and Table 2).

Figure 1—Rates of insulin and glucose infusion, production and utilization of glucose, and PG and insulin concentrations after subcutaneous injectionof insulin glargine (Gla) at either 2200 or 1000 h in 10 subjects with T2DM. Because of an insufficient isotopic steady state, the fragment of the curvescorresponding to the initial 6 h of the studies has been deleted for both production and utilization of glucose.

506 Glargine Evening vs. Morning in T2DM Diabetes Care Volume 38, March 2015

At time 0 min, the plasma insulinconcentration tended to be higher inthe morning versus the evening study(P = 0.082). The overall plasma free in-sulin concentration after subcutaneousglargine injection did not differ be-tween the two treatments (Fig. 1 andTable 2).PG concentration decreased similarly

during intravenous insulin infusion priorto glargine injection on the two occa-sions (data not shown), and at time0 h it was not different in the two stud-ies (102 6 4.3 and 100 6 2.2 mg/dL,respectively, for morning vs. eveningstudy, P = 0.061). After glargine injec-tion, the mean PG over the 24-h studyperiod was superimposable in themorning and evening studies. Duringthe first 12 h of the study, the two treat-ment groups had similar mean PG con-centrations, in the second half of thestudy period (time 12–24 h), the meanPG concentration was slightly, but signifi-cantly, higher with morning compared

with evening glargine dosing. The PGconcentration at the end of the clampwas higher with morning glargine com-pared with evening injection (1086 8 vs.1006 2 mg/dL, P = 0.033) (Fig. 1).

The GIR needed to maintain euglyce-mia was significantly different betweenthe two studies at time 0 h: positive(0.45 6 0.55 mg/kg/min) after theevening compared with morning dos-ing, where it was basically turned off(P = 0.038). The mean GIR for the 24-hstudy period (AUC0–24h) was equivalentin the two studies. However, over theinitial 12-h period, the GIR was greaterin the morning compared with eveningstudy, whereas the opposite was ob-served for the second 12-h period. Al-though the maximum GIR values weresimilar with both treatments, the GIRTmax occurred earlier after morning in-jection than after evening dosing (Fig. 1and Table 2).

EGP AUC7–24h, calculated during thelast 17 h of the study at isotopic steady

state, was not different with eveningcompared with morning dosing. How-ever, the EGP exhibited opposite pro-files between day and night, indicatinggreater insulin sensitivity during the daycompared with the night. In fact, EGPdecreased from a peak value of 1.14 60.57 mg/kg/min at 0800 h to the lowestvalue of 0.146 0.18mg/kg/min at 1800 h(evening insulin injection), whereas it in-creased steadily from 2200 h (0.20 60.21 mg/kg/min) over the night hoursto achieve a peak value of 0.96 6 0.30mg/kg/min between 0800 and 1000 h(morning insulin injection).

Glucose utilization (GU) was not stim-ulated and overall was not differentamong insulin treatments (Fig. 1 andTable 2).

Plasma C-Peptide, Glucagon, FFA, and

b-OH-Butyrate Concentrations

Overall, the plasma C-peptide concen-tration was similarly suppressed withboth treatments (P = 0.104) (Fig. 2 andTable 2).

Plasma glucagon concentrations weremore suppressed with evening comparedwith morning dosing. However, over theinitial 12-h period (AUC0–12h), the degreeof suppression was equivalent, whereasin the second half of observation periodit was less suppressed with morning glar-gine dosing (Fig. 2).

Plasma FFA concentrations weremore suppressed with evening versusmorning glargine dosing over the 24-hstudy period. During the last 12 h withboth treatments, FFA plasma concentra-tions tended to increase compared withthe initial 12 h of the clamp; however,this increase was greater with morningglargine dosing (Fig. 2).

Similarly, plasma b-OH-butyrate con-centrationswere lowerwith evening glar-gine dosing over the whole study period(0–24 h) (P = 0.005). Indeed, the plasmab-OH-butyrate concentration increasedto 1.43 6 0.99 mmol/L by the end ofstudy with the morning dosing, but onlyto 0.426 0.32mmol/L with evening glar-gine dosing (P = 0.003) (Fig. 2).

After initial suppression, plasma glyc-erol concentrations increased in bothtreatments, and remained significantlylower after evening comparedwithmorn-ing dosing (P = 0.020) (data not shown).

Lactate and alanine concentrationswere not different between treatments(data not shown).

Table 2—Summary of the main study results

Morninginjection

Eveninginjection P value

PG (mg/dL)0–24 h 101 6 3.0 101 6 2.0 0.5510–12 h 100 6 1.0 102 6 3.0 0.10412–24 h 103 6 4.0 100 6 1.0 0.028

FIRI (mmol/L/h)AUC0–24h 1,636 6 440 1,605 6 504 0.836AUC0–12h 898 6 241 828 6 272 0.367AUC12–24h 739 6 204 777 6 245 0.625

PDGIR-AUC0–24h (mg/kg 3 24 h) 995 6 691 1,058 6 571 0.503GIR-AUC0–12h (mg/kg 3 12 h) 593 6 374 357 6 244 0.004GIR-AUC12–24h (mg/kg 3 12 h) 403 6 343 700 6 396 0.002Maximum GIR (mg/kg/min) 1.4 6 0.8 1.5 6 0.6 0.533GIR Tmax (h)

a 9 (8.1, 10.7) 18 (16.5, 20.4) 0.002

Glucose fluxesEGP-AUC7–24h (mg/kg 3 17 h) 595 6 348 495 6 267 0.351GU-AUC7–24h (mg/kg 3 17 h) 1,260 6 349 1,395 6 368 0.141

Glucagon (ng/L/h)AUC0–24h 1,533 6 656 1,120 6 344 0.027AUC0–12h 739 6 322 576 6 160 0.067AUC12–24h 794 6 342 544 6 200 0.012

FFA (mmol/L/h)AUC0–24h 8.9 6 1.9 7.5 6 1.6 0.005AUC0–12h 4.0 6 0.7 3.6 6 0.7 0.042AUC12–24h 4.8 6 1.2 3.8 6 1.0 0.006

b-OH-butyrate (mmol/L/h)AUC0–24h 17.0 6 11.9 6.8 6 4.7 0.005AUC0–12h 7.1 6 6.4 2.7 6 1.6 0.035AUC12–24h 9.9 6 6.4 4.1 6 3.2 0.001

Data are mean 6 SD, unless otherwise indicated. FIRI, free immunoreactive insulin. aMedianwith 25 and 75 percentiles.

care.diabetesjournals.org Porcellati and Associates 507

Insulin Activity Analyzed by 24-h Clock

Time, Independently From the Time of

Glargine Injection

When insulin activity (as indicated byGIR, EGP, and GU profiles) was analyzedby 24-h clock time, comparing daytime(1000–2200 h) versus nighttime (2200–1000 h) periods from the last 12 h ofeach study, the GIR was greater duringthe day versus the night independent ofthe time of glargine injection (AUC12–24h7016 396 vs. 412 6 331 mg/kg 3 12 hday vs. night, respectively, P = 0.002)(Fig. 3). EGP changes were consistentwith differences in GIR requirements(AUC12–24h 290 6 204 vs. 412 6 202mg/kg3 12 h day vs. night, respectively,P = 0.041) (Fig. 3). GU did not change andwas not different between treatments(P = 0.079 and P = 0.406, respectively,

for time and treatment3 time interac-tion, from repeated-measures ANOVA)(Fig. 3).

CONCLUSIONS

This is the first study directly comparingin T2DM subjects the PK and PD of basalinsulin glargine administered at differ-ent times of the day (i.e., in the eveningand in the morning). The question is rel-evant not only in light of the potentialclinical translation of the findings, butalso considering the great emphasis re-cently given to the concept of flexibilityof dosing time, which has been advo-cated mainly for the second-generationlong-acting insulin analogs (26).

The results indicate that overall 24-hinsulin activity was not different whether

glargine was injected in the evening or inthemorning. However, relevant differencesemerged when considering time actionprofiles and distribution of metabolic ef-fect in the first 12-h period comparedwith the second 12-h period. Glargineadministration in the morning exertedgreater activity in the first 12 h, in con-trast to evening dosing, which exhibitedhigher potency in the second 12 h. Inaddition, the latter was more effectivein suppressing overall lipolysis andplasma glucagon concentrations, pri-marily over the last 12 h of the studies.

In the current study, morning glargineadministration induced a robust increasein GIR during the first 12 h,whereas in thesecond 12 h of the study a progressivedecline in GIR occurred, as the resultof a rebound in EGP, and PG tended to

Figure 2—Plasma C-peptide, glucagon, FFAs, and b-OH-butyrate concentrations after subcutaneous injection of 0.4 units/kg insulin glargine (Gla) ateither 2200 or 1000 h in 10 subjects with T2DM.

508 Glargine Evening vs. Morning in T2DM Diabetes Care Volume 38, March 2015

increase (Fig. 1). This action profile is re-markably in line with previous reports(15–17) in which glargine PD was studiedafter morning injection. In contrast, withevening glargine the distribution of themetabolic effect was quite different. Inthe first 12 h, GIRwas nearly flat, whereasin the second 12 h, GIR increased notablyas the result of the suppression of EGP.The superior insulin activity at the end ofthe 24-h period after evening glargine ad-ministration is also indirectly suggested

by the lower rate of intravenous insulininfusion and greater GIR required toreach euglycemia prior to clamp initiationbefore insulin injection at 2200 h, com-pared with the rates infused before themorning clamp.

None of the two glargine dosing regi-mens increased glucose disposal. This isin line with the physiological conceptthat when basal insulin is replaced attherapeutic doses to target euglycemia,it exerts its effects primarily at the liver

and adipose tissue level (suppression ofglucose production and lipolysis) with-out promoting muscle glucose uptake(17,27).

Morning and evening glargine admin-istration had differential effects on 24-hlipolysis and plasma glucagon concen-trations. The evening glargine adminis-tration, more effectively than morning,restrained lipolysis and ketogenesis,and suppressed plasma glucagonconcentrations.

Figure 3—Rates of glucose infusion, production, and utilization presented by time of day (24-h clock time) as result of the last 12 h of the eveningstudy (left panel, 1000–2200 h) and the last 12 h of the morning study (right panel, 2200–1000 h). Gla, glargine.

care.diabetesjournals.org Porcellati and Associates 509

It is tempting to speculate about themechanisms by which glargine whengiven in the morning has differentialmetabolic effects compared with even-ing administration. One explanationmight be different PK in the first 12-hperiod compared with the second 12-hperiod (Fig. 1). However, plasma insulinconcentrations were relatively stablethroughout the 24-h study, withoutany difference between first and second12-h periods, after both morning andevening dosing. This applies to the sup-pression of plasma C-peptide as well(Fig. 2).The alternative hypothesis might be

that the results are primarily related todifferential responses to insulin at dif-ferent times of day (i.e., insulin sensi-tivity might be greater during daytimehours compared with nighttime hours).In fact, when the data are analyzed by24-h clock time (Fig. 3), it appears thatEGP, the driver of basal insulin effectson glucose metabolism (i.e., GIR), is lesssuppressed during the night (2200–1000 h) than during the day (1000–2200 h), and to a similar extent inmorn-ing and evening dosing studies. At theend of the night, at 1000 h, EGP surpris-ingly reached similar values after bothmorning and evening dosing. However,the incremental increase in EGP wasgreater with morning glargine dos-ing compared with evening glarginedosing, possibly due to more acceler-ated lipolysis and greater plasma gluca-gon concentrations. Conversely, duringthe day (1000–2200 h), insulin activityincreases progressively as a resultof the suppression of EGP. The latter,again surprisingly, reaches similarvalues at 2200 h after either morningor evening glargine administration.Taken together, these differencesspeak in favor of the circadian rhythmof insulin sensitivity, with greatersensitivity during the day comparedwith the night (19). Thus, the effectsof glargine administered in the morn-ing or evening only appear to differ.They are remarkably reproduciblewhen analyzed by absolute time ofday, not the relative number of hourspostinjection.The results of the current study

closely reproduce those observed byRadziuk and Pye (28), and represent ad-ditional evidence of the dawn phenom-enon in T2DM (29–31).

The existence of diurnal rhythmicityin insulin sensitivity driving changes inPG concentration in normal subjectsand in T2DM patients has been reportedby older (32,33) andmore recent studies(18,19,34,35). Boden et al. (19) de-scribed a robust and well-definedrhythm in insulin sensitivity in T2DM pa-tients, with lower sensitivity in themorn-ing and higher sensitivity in the evening.Similar to the current study, the circa-dian changes in GIR were completely ac-counted for by changes in EGP, whereasthe Rd did not change (19). The elegantwork of Radziuk and Pye (28) provided adeeper and fascinating insight with thehypothesis that the observed diurnaltemporal pattern in EGP might resultfrom factors counteracting insulin actionand triggered by the brain through cen-tral regulatory pathways. Nowadays, theconcept of diurnal rhythmicity in insulinsensitivity finds novel relevance in lightof the more extensive concept of a “mo-lecular clock,” which has been shown toregulate the expression and functionof a variety of enzymes, transport sys-tems, and nuclear receptors involvedin lipid and carbohydrate metabolism(36–38).

Regardless of the mechanisms, thedifferential insulin activity observed inthe current study after morning glar-gine administration compared withevening glargine administration is rele-vant to clinical practice. Indeed, insulinglargine provides a rather flat activityprofile at night whether given in themorning or in the evening. This explainsthe reported lower risk of nocturnal hy-poglycemia with glargine comparedwith “peak insulin” such as NPH (4–7).On the other hand, the present studiesindicate similarly greater insulin activityduring the daytime after glargine ad-ministration in either the morning orthe evening. This may predict potentia-tion of the risk of hypoglycemia in theafternoon, for example, when prandialglucose-lowering drugs are combinedwith basal insulin. Whether dosingat a more usual breakfast time (0700–0800 h) would result in circadian PDeffects similar to those observed inthe current study cannot be directlyderived from our results. However,the time curve of GIR indicates lowerinsulin sensitivity in the interval 0200–1000 h, suggesting that insulin dosingwithin this time frame would produce

lesser effects until 1000 h and greatereffects thereafter.

Since the PD results of the presentstudies observed with glargine mightbe interpreted not on the basis of PK,but rather on the circadian rhythm ofinsulin sensitivity, it is likely that simi-lar effects may occur also with otherlong-acting insulin formulations. In arecent study (14), NPH and detemir ex-hibited greater activity in the after-noon compared with morning hoursafter evening injection in T2DM pa-tients. However, a direct comparisonwith morning and evening dosing ofother long-acting insulins (degludec,already on market outside the US, andglargine 300 units as well as pegylatedlispro, which has not yet been approved)is needed.

A limitation of our study is the rela-tively small number of subjects studied.However, the homogeneity of the dataobserved and the crossover design rein-force the findings.

A strength of our study is its design,which mimicked real-life conditions asclosely as possible: an insulin dose closeto that used by the insulin-treated sub-jects was administered, and the studywas performed after multiple dosing ofglargine (steady state).

Finally, the results of the currentstudy are relevant to the understandingof PD in clamp studies after subcutane-ous injection of long-acting insulin ana-logs in T2DM (39,40). Distribution of themetabolic effect may be quite differentif the insulin to be studied is injectedin the morning compared with theevening. Therefore, it would not be cor-rect to assume that the greater meta-bolic activity of glargine, observed 6–8 hafter morning dosing (15–17), is repre-sentative of the true action profile sincein the current study this appears quitedifferent when glargine is administeredin the evening.

In conclusion, total insulin activity onglucose metabolism is similar with even-ing or morning glargine administration.However, with evening glargine adminis-tration, the suppression of nocturnal EGP,lipolysis, and glucagon concentration aremore consistent. As result, targeting fast-ing euglycemia appears more convenientwith evening glargine dosing comparedwith morning glargine dosing. Con-versely, morning dosing may be prefera-ble whenever greater protection against

510 Glargine Evening vs. Morning in T2DM Diabetes Care Volume 38, March 2015

the risk for nocturnal hypoglycemia isneeded.

Acknowledgments. The authors thank Dr.Gianluca Curti, Department of Medicine, Uni-versity of Perugia Medical School, Perugia, Italy,for his skillful technical assistance in performingthe gas chromatography–mass spectrometryassay for glucose kinetics. This article is dedi-cated to the subjects with type 2 diabetes whohave volunteered for the studies.Duality of Interest. F.P. has received honorariafor speaker fees and/or travel grants fromSanofi, Eli Lilly & Co, Bristol-Myers Squibb, andMerck & Co. P.L. has received travel grants forscientific meetings from Sanofi and Menarini.G.B.B. has received honoraria from Sanofi,MannKind, and Eli Lilly & Co. for scientific advis-ing and consulting. C.G.F. has served on scien-tific advisory panels for Sanofi and has receivedhonoraria for speaker fees and/or travel grantsfrom Bristol-Myers Squibb, Merck & Co., andMenarini. No other potential conflicts of inter-est relevant to this article were reported.Author Contributions. F.P. enrolled patients,performed clamps, analyzed data, and wrotethe manuscript. P.L. performed clamps andglucose turnover measurements, analyzeddata, and reviewed and edited the manuscript.P.Ci. performed clamps and laboratory assaysand reviewed and edited the manuscript. P.Ca.,A.M.A., and S.M. performed clamps and re-viewed and edited the manuscript. M.A. per-formed glucose turnover measurements andreviewed and edited the manuscript. G.B.B.provided the study concept and design, super-vised the protocol development and the re-search, contributed to discussion, and reviewedand edited the manuscript. C.G.F. performedclamps, analyzed data, performed the statisticalanalysis, contributed to discussion, and re-viewed and edited the manuscript. F.P. is theguarantor of this work and, as such, had fullaccess to all the data in the study and takesresponsibility for the integrity of the data andthe accuracy of the data analysis.Prior Presentation. Parts of this study werepresented in abstract form at the 73rd ScientificSessions of the American Diabetes Association,Chicago, IL, 21–25 June 2013.

References1. Owens DR, Bolli GB. Beyond the era of NPHinsulin–long-acting insulin analogs: chemistry,comparative pharmacology, and clinical appli-cation. Diabetes Technol Ther 2008;10:333–3492. LeporeM, Pampanelli S, Fanelli C, et al. Phar-macokinetics and pharmacodynamics of subcu-taneous injection of long-acting human insulinanalog glargine, NPH insulin, and ultralente hu-man insulin and continuous subcutaneous in-fusion of insulin lispro. Diabetes 2000;49:2142–21483. Heise T, Nosek L, Rønn BB, et al. Lowerwithin-subject variability of insulin detemir incomparison to NPH insulin and insulin glarginein people with type 1 diabetes. Diabetes 2004;53:1614–16204. Porcellati F, Rossetti P, Pampanelli S, et al.Better long-term glycaemic control with the

basal insulin glargine as compared with NPH inpatients with Type 1 diabetes mellitus givenmeal-time lispro insulin. Diabet Med 2004;21:1213–12205. Riddle MC, Rosenstock J, Gerich J; InsulinGlargine 4002 Study Investigators. The treat-to-target trial: randomized addition of glargineor human NPH insulin to oral therapy of type 2diabetic patients. Diabetes Care 2003;26:3080–30866. Hermansen K, Davies M, Derezinski T,Martinez Ravn G, Clauson P, Home P. A 26-week, randomized, parallel, treat-to-target trialcomparing insulin detemir with NPH insulin asadd-on therapy to oral glucose-lowering drugsin insulin-naive people with type 2 diabetes.Diabetes Care 2006;29:1269–12747. Vague P, Selam JL, Skeie S, et al. Insulindetemir is associated with more predictableglycemic control and reduced risk of hypogly-cemia than NPH insulin in patients with type 1diabetes on a basal-bolus regimen with pre-meal insulin aspart. Diabetes Care 2003;26:590–5968. European Medicines Agency: Find medicine/Human medicine/Lantus, insulin glargine.Available from http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/000284/human_med_000882.jsp&mid=WC0b01ac058001d124. Accessed 1December 20149. Hamann A, Matthaei S, Rosak C, Silvestre L;HOE901/4007 Study Group. A randomized clin-ical trial comparing breakfast, dinner, or bed-time administration of insulin glargine inpatients with type 1 diabetes. Diabetes Care2003;26:1738–174410. Rossetti P, Pampanelli S, Fanelli C, et al. In-tensive replacement of basal insulin in patientswith type 1 diabetes given rapid-acting insulinanalog at mealtime: a 3-month comparison be-tween administration of NPH insulin four timesdaily and glargine insulin at dinner or bedtime.Diabetes Care 2003;26:1490–149611. Ashwell SG, Gebbie J, Home PD. Optimal tim-ing of injection of once-daily insulin glargine inpeople with Type 1 diabetes using insulin lisproat meal-times. Diabet Med 2006;23:46–5212. Fritsche A, Schweitzer MA, Haring HU; 4001Study Group. Glimepiride combined with morn-ing insulin glargine, bedtime neutral protaminehagedorn insulin, or bedtime insulin glargine inpatients with type 2 diabetes. A randomized,controlled trial. Ann Intern Med 2003;138:952–95913. Standl E, Maxeiner S, Raptis S, Karimi-Anderesi Z, SchweitzerMA; HOE901/4009 StudyGroup. Good glycemic control with flexibility intiming of basal insulin supply: a 24-week com-parison of insulin glargine given once daily in themorning or at bedtime in combination withmorning glimepiride. Diabetes Care 2005;28:419–42014. Lucidi P, Porcellati F, Rossetti P, et al. Phar-macokinetics and pharmacodynamics of thera-peutic doses of basal insulins NPH, glargine, anddetemir after 1 week of daily administration atbedtime in type 2 diabetic subjects: a random-ized cross-over study. Diabetes Care 2011;34:1312–131415. Klein O, Lynge J, Endahl L, Damholt B, NosekL, Heise T. Albumin-bound basal insulin

analogues (insulin detemir and NN344): compa-rable time-action profiles but less variabilitythan insulin glargine in type 2 diabetes. Diabe-tes Obes Metab 2007;9:290–29916. Luzio S, Dunseath G, Peter R, Pauvaday V,Owens DR. Comparison of the pharmacokineticsand pharmacodynamics of biphasic insulin as-part and insulin glargine in people with type 2diabetes. Diabetologia 2006;49:1163–116817. Wang Z, Hedrington MS, Gogitidze Joy N,et al. Dose-response effects of insulin glarginein type 2 diabetes. Diabetes Care 2010;33:1555–156018. Jarrett RJ. Rhythms in insulin and glucose.In Endocrine Rhythms. Krieger DT, Ed. New York,Raven, 1979, p. 247–25819. Boden G, Chen X, Urbain JL. Evidence for acircadian rhythm of insulin sensitivity in pa-tients with NIDDM caused by cyclic changes inhepatic glucose production. Diabetes 1996;45:1044–105020. Andres R, Swerdoff T, Pozefsky T,Coleman D. Manual feedback technique forthe control of blood glucose concentration.In Automation in Analytical Chemistry.Skeggs LT Jr, Ed. New York, Mediad, 1966, p.486–49121. Porcellati F, Rossetti P, Busciantella NR,et al. Comparison of pharmacokinetics anddynamics of the long-acting insulin analogsglargine and detemir at steady state in type 1diabetes: a double-blind, randomized, crossoverstudy. Diabetes Care 2007;30:2447–245222. Gastaldelli A, Coggan AR, Wolfe RR. Assess-ment of methods for improving tracer estima-tion of non-steady-state rate of appearance. JAppl Physiol (1985) 1999;87:1813–182223. Kuzuya H, Blix PM, Horwitz DL, Steiner DF,Rubenstein AH. Determination of free and totalinsulin and C-peptide in insulin-treated dia-betics. Diabetes 1977;26:22–2924. Wolfe RR. Radioactive and Stable IsotopeTracers in Biomedicine: Principles and Practiceof Kinetic Analysis. New York, Wiley-Liss, 1992,p. 425–42625. Radziuk J, Pye S. Quantitation of basal en-dogenous glucose production in Type II diabe-tes: importance of the volume of distribution.Diabetologia 2002;45:1053–108426. Meneghini L, Atkin SL, Gough SC, et al.;NN1250-3668 (BEGIN FLEX) Trial Investigators.The efficacy and safety of insulin degludecgiven in variable once-daily dosing intervalscompared with insulin glargine and insulindegludec dosed at the same time daily:a 26-week, randomized, open-label, parallel-group, treat-to-target trial in individuals withtype 2 diabetes. Diabetes Care 2013;36:858–86427. Gerich JE. Physiology of glucose ho-meostasis. Diabetes Obes Metab 2000;2:345–35028. Radziuk J, Pye S. Diurnal rhythm in en-dogenous glucose production is a major con-tributor to fasting hyperglycaemia in type2 diabetes. Suprachiasmatic deficit or limit cy-cle behaviour? Diabetologia 2006;49:1619–162829. Bolli GB, Gerich JE. The “dawn phenome-non”da common occurrence in both non-insulin-dependent and insulin-dependent diabetesmellitus. N Engl J Med 1984;310:746–750

care.diabetesjournals.org Porcellati and Associates 511

30. Porcellati F, Lucidi P, Bolli GB, Fanelli CG.Thirty years of research on the dawn phenome-non: lessons to optimize blood glucose control indiabetes. Diabetes Care 2013;36:3860–386231. Monnier L, Colette C, Dejager S, Owens D.Magnitude of the dawn phenomenon and itsimpact on the overall glucose exposure in type 2diabetes: is this of concern? Diabetes Care 2013;36:4057–406232. Hatlehol R. Paradoxical rise of the bloodsugar concentration in diabetes mellitus. ActaMed Scand Suppl 1924;8:211–26633. Izzo JL. Diurnal (24-hour) rhythm in diabe-tes mellitus; diurnal variations in levels of

glucose in blood and urine. Proc Am DiabetesAssoc 1949;9:247–27334. Schmidt MI, Hadji-Georgopoulos A,Rendell M, Margolis S, Kowarski A. The dawnphenomenon, an early morning glucose rise:implications for diabetic intraday blood glucosevariation. Diabetes Care 1981;4:579–58535. Van Cauter E, Polonsky KS, Scheen AJ. Rolesof circadian rhythmicity and sleep in humanglucose regulation. Endocr Rev 1997;18:716–73836. Staels B. When the clock stops ticking, met-abolic syndrome explodes. Nat Med 2006;12:54–55

37. Yang X, Downes M, Yu RT, et al. Nuclear re-ceptor expression links the circadian clock to me-tabolism. Cell 2006;126:801–81038. Prasai MJ, George JT, Scott EM. Molecularclocks, type 2 diabetes and cardiovascular dis-ease. Diab Vasc Dis Res 2008;5:89–9539. Porcellati F, Bolli GB, Fanelli CG. Pharmaco-kinetics and pharmacodynamics of basal insu-lins. Diabetes Technol Ther 2011;13(Suppl. 1):S15–S2440. Heinemann L, Anderson JH Jr. Mea-surement of insulin absorption and insulinaction. Diabetes Technol Ther 2004;6:698–718

512 Glargine Evening vs. Morning in T2DM Diabetes Care Volume 38, March 2015