`JOGmalof
`Clinical
`Pharmacolo y
`
`.
`
`AUROBINDO EX. 1005, 1
`
`
`
`Edited for the British Pharmacological
`Society by
`
`A. Richens (Secretary)
`C.T. Dollery (Chairman)
`Annmari e Hedges (Press Editor)
`
`T.B. Binns
`A. Breckenridge
`W . Watson Buchanan
`D.B. Cain e
`T. Chard
`J . Crooks
`D.A.P. Evans
`C.F. George
`A. Goldberg
`D. Grahame-Smith
`I.W.B. Grant
`R.L. Hodge
`M .H. Lader
`
`A.F. Lant
`G. Mawer
`D. G. McDevitt
`A N. Nicholson
`F. O'G rady
`J.P. Payne
`B.N .C. Prichard
`M . Rawl ins
`J. Reid
`R.G . Shanks
`F.O. Simpson
`,J.R. Trounce
`D.W . Vere
`O.L. Wade
`
`Papers will be considered for publication on all aspects of drug action in man .
`
`Manuscripts (two copies) should be sent to Professor A. Richens, Departm ent of
`Pharmacology and Materia Medica, Welsh Nation al School of Medicine, Heath Park,
`Cardiff CF4 4XN . The instructions to authors are printed in vo lume 1 (pages 3 to 6) of the
`journal. Details of the Sf symbols for units can be found in t he January 1974 issue
`(volume 50, pages 6 to 23) of the British Journal of Pharmacology.
`
`British Journal of Clinical Pharmacology is published monthly by Scient ific & Med ical
`Division, Macmillan Publishers Ltd, Houndmills, Basingstoke, Ham pshire RG21 2XS .
`Telephone: Basingstoke 29242. ISSN 0306-5251 .
`
`Annual subscription: UK £75.00; USA & Canada, US$185; Rest of world, £89; Airmail
`£102 (or equivalent in any other currency). (Orders m ust be accompanied by
`remittance. Prices correct at time of publication .)
`
`Cheques should be made payable to Macmillan Journals Limited: Subscription
`Department, Brunei Road, Basingstoke, Hampshire RG21 2XS, England. Where
`appropriate, overseas subscribers may make payments into UK Post Office Giro
`Account No. 519 2455. Full details must accompany th e payment.
`
`Enquiries concerning advertising space or rates should be addressed to : Advertisement
`Department, Macmillan Journals Limited, 4 Little Essex St reet London WC2R 3LF.
`Telephone: 01 -836 6633.
`
`Copyright © 1981 by the British Journal of Clinical Pharmacolog y. All rights of
`reproduction are reserved in respect of all papers, articles, illustrations, etc., published
`in this journal in all countries of the world.
`
`Second class postage paid at New York, N.Y. US Mailing Agent: Expediters of the
`printed Word, Inc., Suite 1217, 527 Madison Ave., New York, N.Y. 10022.
`
`AUROBINDO EX. 1005, 2
`
`
`
`Br. J. clin. Pharmac. (1981), 12, 235-246
`
`METFORMIN KINETICS IN HEALTHY SUBJECTS AND IN PATIENTS
`WITH DIABETES MELLITUS
`
`G.T. TUCKER, C. CASEY, P.J. PHILLIPS*, H. CONNOR, J.D. WARD
`&H.F. WOODS
`University Department of Therapeutics,
`The Royal Hallamshire Hospital, Sheffield, S 10 2JF
`
`1 The kinetics of metforinin were studied after i.v. and oral administration in four healthy subjects and
`after oral administration in twelve maturity onset (Type II) diabetic patients.
`2 After i.v. administration most of the dose was rapidly eliminated but with a mean 'terminal' Tr of
`4 h measured up to 12 h in plasma and of 16 h measured up to 60 h from the urinary excretion rate. On
`average, 80% of the dose was recovered as unchanged drug in the urine with none detected in the
`faeces.
`3 After single oral doses (0.5 and 1.5 g), maximum plasma concentrations and urinary excretion rates
`were observed at about 2 h with urinary recoveries of unchanged drug of 3 5-5 0% and faecal recoveries
`of about 30%. Urinary recoveries were significantly lower after the higher dose. Absolute oral
`bioavailability was 50-60% of the dose.
`4 Deconvolution analysis showed that after a short lag-time, the available oral dose was absorbed at
`an exponential rate over about 6 h. Implications for the design of prolonged release dosage forms are
`discussed.
`5 Plasma metformin concentrations measured throughout the seve.nth and fourteenth days of
`continuous 0.5 g twice daily treatment were accurately predicted from single dose data, although a
`discrepancy between observed and predicted trough levels reflected the existence of a slow elimination
`phase. Implications of the latter for a gradual accumulation of metformin in peripheral tissues and a
`possible association with lactic acidosis are discussed.
`6 Renal clearance of metformin was highly correlated with creatinine clearance. However, a weaker
`relationship between total oral clearance of the drug and creatinine clearance suggests that the latter
`may not always be a reliable indicator of potential metformin accumulation owing to variability in
`absorption and possibly non-renal clearance of the drug,
`
`Introduction
`
`following the restrictions placed upon the prescrib(cid:173)
`ing of phenformin (phenylethylbiguanide) in some
`countries, and its removal from the market in others,
`metformin (Nt, N 1-dimethylbiguanide) and bufor(cid:173)
`min (N1-butylbiguanide) are now the most common(cid:173)
`ly prescribed oral hypoglycaemic drugs of the
`biguanide class. Metformin has been recommended
`as the drug of choice because the risk of developing
`lactic acidosis during treatment is less than that re(cid:173)
`sulting from the use of phenformin (British Medical
`Journal, 1977; Phillips, Thomas & Harding, 1977;
`Bergman, Boinan & Wiholm, 1978).
`* On leave from the Institute of Medical and Veterinary
`Science, Adelaide, Australia. Present address: Flinders
`University, Medical Centre, Adelaide, Australia.
`
`In spite of its widespread use, little is known about
`the pharmacokinetics of metformin. Two recent pap(cid:173)
`ers have described the kinetics in normal man (Sirtori
`et al., 1978; Pentikainen, Neuvonen & Penttila,
`1979) and in one of these (Sirtori eta/., 1978) special
`emphasis was placed upon the relationship between
`impairment of renal function and the resulting slow(cid:173)
`ing of metformin clearance. The literature is deficient
`in information concerning the kinetics of metformin
`in patients with diabetes mellitus.
`In this paper we describe the pharmacokinetics of
`metformin.administered via the oral and intravenous
`routes in normal subjects and in patients with diab(cid:173)
`etes mellitus.
`
`0306-5251/81/080235-12$01.00
`
`© Macmillan Publishers Ltd 1981
`
`AUROBINDO EX. 1005, 3
`
`
`
`236
`
`G.T. TUCKER, C. CASEY, P.J. PHILLIPS, H. CONNOR, J.D. WARD & H.F. WOODS
`
`Methods
`
`Protocol
`
`Three groups of subjects were studied (Table 1).
`
`Group I consisted of f9ur healthy male volunteers
`who were given single intravenous and oral doses of
`metformin HCI on separate occasions at least 2
`weeks apart, according to a cross-over design. The
`intravenous dose was 0.25 g given by constant-rate
`infusion over the course of 15 min and the oral doses
`were 0.5 g and 1.5 g, respectively, in the form of
`Glucophage® tablets from a single batch. The in(cid:173)
`travenous solution and tablets were analysed and
`found to contain between 98 and 100% of the stated
`dose. The oral doses were taken with breakfast. Drug
`concentrations were measured in serial samples of
`whole blood, plasma, urine and faeces. Blood sam(cid:173)
`ples, obtained by venepuncture without stasis, were
`collected up to 12 h (i.v. study) and 24 h (oral study);
`urine samples were collected up to 72 h and faecal
`samples to 5 days.
`The purpose of these experiments was to assess: (I)
`the recovery of unchanged drug; (II) the rate and
`extent of oral bioavailability of metformin; (III) any
`dose-dependence in its kinetics and (IV) the distribu(cid:173)
`tion of the drug between plasma and blood cells.
`
`Group II consisted of four newly diagnosed maturity
`onset (Type II) diabetic patients. They received a
`single 1.0 g oral dose (as Glucophage® tablets) fol-
`
`lowed 3 days later by 0.5 g (p.o.) twice a day. The
`drug was taken with meals as is normally advised.
`Metformin concentrations were determined in ser(cid:173)
`ial plasma samples up to 24 h and in serial urine
`samples up to 72 h after the first dose and in plasma
`samples during days 7 and 14 of continuous twice
`daily dosing.
`The purpose of this experiment was: (I) to com(cid:173)
`pare metformin kinetics in the patients with that in
`healthy subjects (Group I); (II) to assess the accumu(cid:173)
`lation of the drug and the degree to which this could
`be predicted from single dose data.
`
`Group III was composed of eight maturity onset
`(Type II) diabetic patients four of whom were taking
`chlorpropamide (Table 1). They were all given a
`single 1.0 g oral dose in the form of Glucophage®
`tablets. Metformin concentrations were measured in
`serial plasma samples up to 24 h and in serial urine
`samples up to 72 h.
`The purpose of this experiment was to assess the
`relationship between metformin kinetics and renal
`function .. Combining the data from all three groups
`gave information about metformin clearance over a
`range of creatinine clearance from 4 7 to 179 ml
`min- 1
`• (Table 1.)
`These studies were approved by the local hospital
`Ethics Committee.
`
`Drug analysis
`
`Concentrations of unchanged metformin were meas-
`
`Table 1 Clinical details of normal subjects and diabetic patients
`
`Group
`
`Subject
`
`Age
`(years)
`
`Sex
`
`Weight
`(kg)
`
`Height
`(em)
`
`ClcRa
`(mlmin- 1
`)
`
`Other"
`drugs
`
`II
`
`III
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`
`34
`30
`30
`36
`62
`68
`70
`46
`70
`67
`68
`81
`73
`57
`70
`59
`
`M
`M
`M
`M
`M
`M
`F
`M
`M
`M
`F
`F
`F
`M
`M
`M
`
`66
`68
`64
`83
`95
`74
`68
`90
`73
`58
`78
`57
`64
`84
`81
`82
`
`108
`178
`175
`178
`173
`160
`152
`179
`168
`168
`173
`155
`127
`176
`168
`184
`
`113
`106
`145
`179
`120
`85
`86
`168
`51
`97
`85
`47
`57
`116
`72
`107
`
`M,N
`N
`
`P,C
`
`M
`c
`A,C
`M
`·c
`
`a creatinine clearance, determined by the Jaffe method using an Autoanalyser.
`
`atenolol;
`b A
`N = Navidrex K;
`
`chlorpropamide;
`C
`P = prochlorperazine
`
`M = a-methyldopa;
`
`AUROBINDO EX. 1005, 4
`
`
`
`ured using a specific gas chromatographic method
`(Lennard et al., 1978). Coefficients of variation of
`the assay were ±9% and ±5% at 50ngml- 1 and
`2 J.Lg ml- 1
`, respectively. All samples were assayed in
`duplicate.
`
`Plasma binding
`
`The binding of metformin in plasma samples from
`healthy subjects, spiked with drug concentrations of
`0.1 and 10J,Lgml- 1 was determined using the
`Dianorm® equilibrium dialysis apparatus (Weder,
`Schildknecht & Kesselring, 1971). Samples were
`dialysed for 3 h at 37°C against phosphate buffer,
`pH7.4, using 1.0 ml half-cells and a cellulose acetate
`membrane.
`
`Pharmacokinetic analysis
`
`Intravenous administration (Group I) Post-infusion
`plasma drug concentrations (Cpost) were fitted by a
`triexponential equation:
`
`I -)lit
`3
`Cpost = I C .e
`
`i=l
`
`(1)
`
`Initial estimates of the coefficients C~ and Ai were
`obtained graphically by the method of residuals. The
`values were then refined using the Gauss-Newton
`iterative procedure incorporated in a modification of
`the IGPHARM package (Gomeni & Gomeni, 1978).
`Experimental data points were weighted according to
`the square of their reciprocal values. Assessment of
`the goodness of fit of computed data to observed data
`was based on plots of weighted residuals against time
`and the coefficient of determination (Boxenbaum,
`Riegelman & Elashoff, 1974).
`
`Values of the coefficients c: were corrected to
`
`those expected following an instantaneous bolus in(cid:173)
`jection (Ci) using equation 2:
`
`(2)
`
`where -r is the infusion time.
`Totai plasma clearance (Cl) was calculated from:
`
`Cl=_Q_
`AUC
`
`(3)
`
`where Dis the dose and AUC is the area under the
`plasma drug concentration-time curve extrapolated
`to infinite time and given by:
`
`3 G
`AUC= I . -1
`i=t X.i
`
`(4)
`
`Renal clearance ( ClR) was obtained using equation 5:
`
`METFORMIN KINETICS
`
`237
`
`Cl = Ae (12)
`R AUC(12)
`
`(5)
`
`where Ae(12) is the amount of unchanged drug
`excreted in the urine up to 12 hand AUC(12) is the
`area under the plasma drug concentration-time curve
`up to 12 hand given by:
`AUC (12) = AUC-C(l2)
`A3
`where C(12) is the estimated plasma drug concentra(cid:173)
`tion at 12 h after a bolus injection.
`The fraction of the dose excreted as unchanged
`drug (fe) was calculated from:
`
`(6)
`
`fe
`
`Ae(72)
`D
`
`(7)
`
`where Ae(72) is the urinary recovery of unchanged
`drug at 12h.
`The volume of distribution at pseudoequilibrium
`during the terminal log-linear phase (V) was calcu(cid:173)
`lated from:
`
`V=_Q_
`A3
`Urinary excretion rates were also fitted by a triex(cid:173)
`ponential equation using non-linear least squares
`regression with weighting by the square of reciprocal
`values.
`
`(8)
`
`Oral administration
`(a) Single dose
`(Groups I-III) Plasma drug
`concentration-time curves were fitted graphically by
`a triexponential equation with one negative and two
`positive terms in a similar manner to the i. v. data.
`When subjected to non-linear least squares regres(cid:173)
`sion analysis the solutions for many of the data sets
`converged on an equation consisting of a negative
`and a positive exponential term plus a constant term.
`The latter reflected the slow disappearance of met(cid:173)
`formin between 12 and 24 h and the lack of data
`points between these times. In view of this deficiency
`in data collection an approximate 'terminal' T 1 is
`reported, based upon the initial graphical estimates.
`Oral bioavailability was calculated in two ways,
`using plasma data (Fp) by equation 9 and using urine
`data (Fur) by equation 10:
`
`[AUC(12)po]. D
`[AUC(12)]. Dpo
`
`(9)
`
`where Dpo in the oral dose and AUC(12)po is the area
`under the plasma drug concentration-time curve up
`to 12 h after the oral dose calculated by the trapezoi(cid:173)
`dal rule. D and AUC(12) refer to intravenous ad(cid:173)
`ministration, as above.
`
`AUROBINDO EX. 1005, 5
`
`
`
`238
`
`G.T. TUCKER, C. CASEY, P.J. PHILLIPS, H. CONNOR, J.D. WARD & H.F. WOODS
`
`[Ae(72)po].D
`[Ae (72)].Dpo
`
`(10)
`
`where Ae(72)po and Ae(72) are the urinary re(cid:173)
`coveries of unchanged drug at 72 h after oral and i.v.
`administration, respectively.
`Oral clearance (CLpo) was calculated from:
`
`CLpo
`
`Dpo
`AUCpo
`
`(11)
`
`The fraction of the oral dose unabsorbed as a
`function of time was ca,Iculated by deconvolution of
`the corresponding i.v. and oral plasma drug concent(cid:173)
`ration v time data using both the Point-Area Method
`. (Vaughan & Dennis, 1978) and a linear interpolation
`method analogous to the Loo-Riegelman procedure
`(Loo & Riegelman, 1968). It was found possible to
`describe the fraction unabsorbed data accurately by
`assuming a short lag-time followed by a monoexpo(cid:173)
`nential input of the available fraction of the dose.
`Using this input function to drive the disposition
`function derived from the i. v. data, the plasma drug
`concentrations observed after oral administration
`were recovered satisfactorily, thereby providing a
`check on the deconvolution routines.
`(b) Multiple dosing (Group II) Plasma drug con(cid:173)
`centrations during days 7 and 14 of continuous treat(cid:173)
`ment in Group II patients were predicted from their
`initial dose data by application of the Superposition
`Principle (Wagner, 1975).
`The area under the plasma drug concentration(cid:173)
`time curve extrapolated to infinite time after the
`initial dose in each subject was compared to the areas
`under the curves during days 7 and 14.
`The latter were calculated by the trapezoidal rule.
`
`These areas should be identical if the kinetics of the
`drug are linear and steady-state has been reached
`during continuous dosing.
`
`Statistical analysis
`A paired t-test was used to assess differences in the
`urinary and faecal recovery of metformin at the two
`dose levels in Group I subjects. Pearson product
`moment correlation coefficients were calculated to
`assess the strength of linear correlations between
`metformin and creatinine clearances.
`Statistical significance was assumed when P was
`<0.05.
`
`Results
`
`Intravenous administration (Group I)
`
`Plasma metformin concentrations decreased rapidly
`after the end of infusion and fell below the limit of the
`assay after 12 h (Figure 1). Over this period the data
`were accurately described by equation 1. Values of
`the parameters of equation 1 are listed in Table 2, as
`are values of clearance and volume of distribution.
`The 'terminal' half-life over the period of observa(cid:173)
`tion ranged from 2.5 to 7.0 h, with an harmonic mean
`of3.8h.
`The urinary excretion rate of the drug could be
`measured up to 60 hand the half-life calculated from
`the latter part of these data ranged from 13-34, with
`an harmonic mean value of 16 h (Table 2, Figure 1).
`An average of 79% of the dose was recovered as
`unchanged drug in the urine after 72 h, 95% of this
`total appearing in the first 8 h (Figure 2). Renal
`clearance was about four times the creatinine clear-
`
`Table 2 Pharmacokinetic parameters describing the disposition of metformin in Group I subjects after
`intravenous administration of 0.25 g metformin HCI
`
`Parameter
`
`1
`
`2
`
`3
`
`Subject
`4
`
`Meanb
`
`s.d.
`
`Ct(J.Lgml-1)a
`Cz(J.&.g mJ-1 )a
`C3(J.Lg mJ-:-1)a
`1\.t(h-1)
`1\.z(h- 1)
`1\.3(h-1)
`Ttz(plasma) (hY
`Ttz(urine) (h)d
`Cl (ml min- 1
`)
`fe
`CIR(ml min- 1
`V(l)
`
`)
`
`22.99
`3.91
`0.442
`27.17
`1.079
`0.225
`3.08
`14.2
`657
`0.737
`462
`175
`
`3.33
`2.85
`0.081
`4.26
`0.649
`0.129
`5.39
`12.9
`718
`0.839
`600
`333
`
`13.27
`2.52
`0.372
`6.68
`1.033
`0.276
`2.51
`14.9
`721
`0.818
`578
`156
`
`10.46
`2.04
`0.104
`4.85
`0.815
`0.098
`7.04
`34.2
`728
`0.763
`537
`443
`
`12.51
`2.83
`0.245
`10.74
`0.894
`0.182
`4.50
`19.0
`706
`0.789
`544
`276
`
`8.14
`0.79
`0.18
`11.00
`0.20
`0.083
`2.10
`10.11
`33
`0.047
`61
`136
`
`a as metformin HCI
`b arithmetic mean
`c
`'terminal' half-life determined up to 12 h
`'terminal' half-life determined from urinary excretion rate up to 60 h
`d
`
`AUROBINDO EX. 1005, 6
`
`
`
`100
`
`0.01 1 . . - J . . . . . . .L - - - 1 . . - - . l . . - - - - - - ' - - - J . . . - - - - l
`. 36
`24
`60
`48
`0 4 8 12
`Time (h)
`
`Figure 1 Decay of mean plasma concentrations
`<• j.Lg/ml) and urinary excretion rate (0, mg/h) of met(cid:173)
`formin {as HCl) post-i.v. infusion of 0.25 g over 15 min
`in Group I subjects (bars indicate±s.d.).
`
`100
`
`10
`
`0.1
`
`Group I (n = 4)
`0.5 g p.o. (normalised to 1.0 g)
`~ 1.5 g p.o. (normalised to 1.0 g)
`
`0
`
`Urine
`
`Ill
`• Plasma
`
`METFORMIN KINETICS
`
`239
`
`Urine (79%)
`
`100
`
`80
`
`Q) 60
`!/)
`0 -c
`'#. 40
`
`20
`
`8 12
`
`24
`
`36
`Time (h)
`
`48
`
`60
`
`72
`
`Figure 2 Cumulative mean urinary excretion of met(cid:173)
`formin after 0.25 g i.v. in Group I subjects (bars
`indicate±s.d.).
`
`ance (Tables 1 and 2). No metformin was detected in
`faecal samples.
`
`Oral administration
`(a) Single dose (Groups I-III) Although peak plas(cid:173)
`ma metformin concentrations were slightly higher
`and peak urinary excretion rates were lower in the
`patients of Groups II and III than in the subjects of
`Group I, when normalized for dose, the kinetic pro(cid:173)
`files in plasma and urine were similar (Figure 3,
`Tables 3, 5, 6).
`
`Group II ( n = 4)
`~ 1.0 g p.o.
`
`Group Ill ( n
`~ 1.0 g p.o.
`
`8)
`
`0
`
`II
`
`• Plasma
`
`0
`0
`
`Urine
`
`0.001 1 . . - . 1 . . . - . l . . - . l . . - - - -L - - - - ' - - - - - - ' - - - . . . . . l
`0 4 8 12
`0 4 8 12
`24
`36
`48
`60
`Time (h)
`Figure 3 Comparison of mean plasma concentrations and urinary excretion rates after oral administration of
`metformin HCl in healthy subjects and diabetic patients.
`
`60
`
`24
`
`36
`
`48
`
`AUROBINDO EX. 1005, 7
`
`
`
`Table 3 Parameters describing the kinetics of metformin after oral administration to Group I subjects
`
`Parameter
`
`Dose(g)
`
`1
`
`2
`
`3
`
`4
`
`Mean
`
`s.d.
`
`0.5
`
`1.5
`
`0.5
`
`1.5
`
`0.5
`
`1.5
`
`0.5
`
`1.5
`
`0.5
`
`1.5
`
`0.5
`
`1.5
`
`Subject
`
`tmax(h)D
`Cmax(J.Lgml- 1
`)b
`AUC(24)po(J.Lg ml- 1h)
`fepoC
`fepd
`
`Fp
`
`Fur
`CIR(ml min- 1
`)
`CLpo( ml min -I)
`T!z(plasma) (hl
`Tt z( urine) (h )1
`
`2.5
`0.59
`4.52
`0.316
`0.204
`0.34
`0.43
`382
`1843
`7.54
`
`2.0
`3.37
`18.61
`0.288
`0.210
`0.45
`0.39
`316
`1343
`5.33
`23
`
`2.5
`1.30
`7.91
`0.510
`0.190
`0.63
`0.62
`550
`1053
`4.90
`20
`
`1.5
`3.18
`24.86
`0.433
`0.213
`0.75
`0.52
`387
`1006
`5.40
`20
`
`2.0
`1.29
`8.47
`0.595
`0.250
`0.69
`0.73
`486
`984
`4.10
`20
`
`1.0
`4.03
`20.70
`0.423
`0.447
`0.57
`0.52
`445
`1208
`5.00
`20
`
`2.0
`0.91
`5.96
`0.562
`0.434
`0.54
`0.74
`683
`1398
`5.20
`-
`
`1.5
`1.81
`9.42
`0.373
`0.448
`0.24
`0.49
`929
`2654
`8.20
`16
`
`2.2
`1.02
`6.71
`0.496
`0.269
`0.55
`0.63
`525
`1319
`5.43
`
`1.5
`3.10
`18.40
`0.379
`0.329
`0.50
`0.48
`519
`1552
`5.98
`
`0.3
`0.34
`1.82
`0.125
`0.113
`0.15
`0.14
`125
`393
`1.47
`
`0.4
`0.93
`6.52
`0.066
`0.136
`0.21
`0.06
`278
`347
`1.49
`
`a = Time of maximum plasma drug concentration
`b =Maximum plasma drug concentration (as metformin HCI)
`c = Fraction of dose excreted in urine as unchanged drug
`d = Fraction of dose excreted in faeces as unchanged drug
`
`e = 'Terminal' half-life determined up to 24 h
`f = 'Terminal' half-life approximated from urinary excretion rate between 36-60 h.
`Other symbols as defined in the text.
`
`~
`
`0
`~
`d (j
`~
`,?:1
`0
`(j > en
`~
`'"d
`~
`~
`~
`~
`~
`(j
`
`~
`
`0 z z
`0 to
`~
`~
`§
`
`R<>
`:I:
`~
`~
`0
`0 en
`
`AUROBINDO EX. 1005, 8
`
`
`
`Table4 Absorption lag times ( tlag) and rate constants (ka)
`in Group I subjectsa
`
`Subject
`
`Dose
`(g)
`
`t[ag
`(h)
`
`ka
`(h-1)
`
`,2
`
`2
`
`3
`
`4
`
`Mean±s.d.
`
`0.5
`1.5
`0.5
`1.5
`0.5
`1.5
`0.5
`1.5
`0.5
`1.5
`
`0.20
`0.00
`0.22
`0.40
`0.45
`0.25
`0.47
`0.33
`0.33±0.14
`0.24±0.17
`
`0.993
`0.994
`0.994
`0.993
`0.992
`0.997
`0.996
`0.999
`
`0.259
`0.379
`0.317
`0.351
`0.296
`o.368
`0.293
`0.430
`0.291±0.024
`0.382±0.034
`
`a A monoexponential equation was fitted to a plot of the
`differences between the bioavailability (Fp) and the
`fraction absorbed as a function of time. The r2 values
`refer to the regression of the log transformed data v
`time; the lag time was taken as the intercept on the
`time-axis.
`
`METFORMIN KINETICS
`
`241
`
`of the dose, of which 50% appeared in the urine and
`27% in the faeces (Figure 4, Table 3). Following the
`1.5 g dose the mean total recovery was 71%, of which
`38% was in the urine and 33% in the faeces (Table
`3). The difference between the urinary recoveries at
`the two dose levels was statistically significant
`(P< 0.05).
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`C1l
`1/)
`0
`"0
`'#.
`
`.....-------lr--.....----
`Total (77%)
`
`04 812
`
`24
`
`48
`36
`Time (h)
`
`60
`
`Figure 4 Cumulative mean urinary and faecal excre(cid:173)
`tion of metformin after 0.5 g p.o. in Goup I subjects
`(bars indicate± s.d.).
`
`Peak plasma drug concentrations and urinary ex(cid:173)
`cretion rates were observed at about 2 h in all
`Groups. Plasma data collected up to 24 h indicated a
`'terminal' half-life of 4-8 h, whereas urinary excre(cid:173)
`tion rates measured up to 60 h disclosed the presence
`of a further elimination phase with a half-life of about
`20h.
`After the 0.5 g dose the mean total recovery of
`unchanged metformin in Group I subjects was 77%
`
`Estimates of oral bioavailability showed good ag(cid:173)
`reement using plasma data and urinary recoveries
`and averaged 50-60% of the dose (Table 3).
`Renal clearances after oral dosage were similar to
`those observed after intravenous administration (Ta(cid:173)
`bles 2 and 3).
`Deconvolution analysis showed that most of the
`oral absorption of metformin takes place over 6 h.
`Beyond this time the absorption rate is negligible
`
`Table 5 Parameters describing the kinetics of metformin after oral administration of the
`initiall.O g dose to Group IT subjects
`
`Parameter
`
`5
`
`6
`
`7
`
`Subject
`
`8
`
`mean
`
`s.d.
`
`tmax(h)
`Cmax(h)
`AUC(24)po(j.Lg ml- 1h)
`fepo
`CIR(ml min- 1
`)
`Clpo(ml min -I)
`T1 z(plasma)(h)
`
`2~0
`3.54
`16.67
`0.482
`344
`1000
`5.00
`
`3.0
`3.36
`26.43
`0.372
`217
`630
`4.70
`
`2.5
`4.20
`24.75
`0.266
`178
`673
`5.60
`
`1.0
`1.89
`11.22
`0.403
`548
`1485
`5.85
`
`2.1
`3.25
`19.77
`0.381
`322
`947
`5.25
`
`0.8
`0.97
`7.11
`0.089
`166
`394
`0.61
`
`Table6 Parameters describing the kinetics of metformin after oral administration of 1.0 g to Group III subjects
`
`Parameter
`
`9
`
`10
`
`11
`
`12
`
`13
`
`Subject
`14
`
`15
`
`16
`
`Mean
`
`s.d.
`
`2.0
`tmax(h)
`Cmax(!-Lg ml- 1
`4.76
`)
`AUC(24)po(I-Lg ml- 1h) 32.22
`fepo
`0.395
`ClR(mlmin- 1
`192
`)
`Clpo(ml min- 1
`517
`)
`
`3.0
`3.02
`29.73
`0.350
`183
`561
`
`3.0
`6.10
`33.75
`0.330
`173
`494
`
`3.0
`2.00
`19.32
`0.181
`155
`863
`
`2.0
`2.53
`18.28
`0.291
`259
`912
`
`3.0
`1.79
`15.87
`0.322
`328
`1050
`
`2.0
`2.74
`25.55
`0.382
`235
`652
`
`1.0
`2.99
`24.01
`0.388
`269
`694
`
`2.4
`3.24
`24.84
`0.347
`224
`718
`
`0.7
`1.46
`6.68
`0.070
`58
`203
`
`AUROBINDO EX. 1005, 9
`
`
`
`242
`
`G.T. TUCKER, C. CASEY, P.J. PHILLIPS, H. CONNOR, J.D. WARD & H.F. WOODS
`
`(Figure 5). The two methods of deconvolution used
`gave similar results and only those obtained with the
`Point-Area method are reported. Representative
`data for one subject illustrating how the drug input
`profile may be described by a short lag-time plus a
`monoexponential phase operating upon the bioavail(cid:173)
`able fraction of the dose are shown in Figure 6. Table
`4 lists the lag-times and input rate constants obtained
`at each dose level in the four subjects of Group I. The
`difference between the input rate constant at the two
`dose levels was significant (P< 0.05).
`In the patients of Groups II and III urinary re(cid:173)
`coveries of unchanged metformin averaged 38% and
`35% of the 1.0 g dose, respectively (Tables 5 and 6).
`(b) Multiple dosing (Group II) Overall agreement
`between· observed and predicted plasma drug con(cid:173)
`centrations on days 7 and 14 of continuous dosage
`was excellent (Figure 7). In individual subjects the
`ratio of the area under the curve during a day of
`continuous treatment to that under single dose data
`ranged from 0.92-1.06forday 7 and from 1.05-1.22
`for day 14. Nevertheless, observed trough concentra(cid:173)
`tions on continuous dosing were about 95% higher
`than predicted, reflecting the existence of an addi(cid:173)
`tional elimination phase or phases undetected from
`24 h single dose plasma data. It was also evident that
`accumulation was still proceeding slowly at 14 days
`since the trough levels at this time were higher than
`those at 7 days (Figure 7).
`a
`
`Plasma binding
`
`Meformin was not bound to plasma protein.
`
`Distribution of metformin between plasma and blood
`cells
`
`A consistent finding was that although blood drug
`concentrations were initially lower than those in plas(cid:173)
`ma, they crossed over at about 8 h and subsequently
`remained much higher. Figure 8 shows representa(cid:173)
`tive data from an individual subject. A 'terminal'
`half-life of 17 h in blood was similar to that recorded
`from urinary excretion rate data.
`In vitro experiments with blood samples spiked
`· with metformin indicated that the value of the drug
`blood/plasma concentration ratio is time-dependent
`rather than concentration-dependent.
`
`Metformin clearance and renal function ·
`
`On combining data from all three Groups a highly
`significant linear correlatjon was observed between
`the renal clearances of metformin and creatinine
`(Figure 9a). However, the relationship between total
`oral clearance of the drug and creatinine clearance
`was much weaker (Figure 10).
`
`b
`
`"C
`Q)
`..c
`0
`..c
`co
`c::
`:::J
`
`(/)
`
`Q)
`(/)
`0
`"C
`'+-
`0
`
`c:: s (.)
`
`~
`u..
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`2
`
`4
`
`6
`
`8
`
`10
`
`12
`Time (h)
`
`2
`
`4
`
`6
`
`8
`
`10
`
`12
`
`Figure 5 Fraction of oral dose remaining to be absorbed as a function of time in Group I subjects, after (a) 0.5 g
`dose and (b) 1.0 g dose. 0 subject 1, e subject 2, 0 subject 3, Ill subject 4.
`
`AUROBINDO EX. 1005, 10
`
`
`
`METFORMIN KINETICS
`
`243
`
`100
`
`1.0
`
`1.0
`
`-o
`Q)
`..0
`0
`Ill
`..0 ro c:
`
`:J
`~
`
`Q)
`
`0.1
`
`0.01
`
`~
`-o
`Q)
`::0
`~
`"(ij
`> ro
`0
`0
`j_
`~
`0 -o
`0
`c:
`0 :g
`~
`u.. 0.001 L----l.--L..-...1.--L----L---'
`2
`12
`8
`6
`10
`4
`Q
`Time (h)
`
`..9-
`I
`E
`0>
`2:
`c:
`.E
`0
`~
`E
`ro
`E
`Ill ro a:
`
`0.1
`~
`
`0.01
`
`Figure 6 Plasma concentrations of metformin and its
`rate of absorption after an oral dose of 0.5 gin subject 2.
`(The smooth line drawn through the plasma data was
`obtained by combining the fitted input function and the
`disposition function from corresponding i. v. data.).
`
`10
`
`0.1
`
`E
`........
`0>
`3-
`c: "§
`
`0
`';!:::
`Q)
`E
`E
`Ill
`(1l a:
`
`(1l
`
`0.01 ~----~--~~·~~----~---r-~----~
`15
`' 3
`7
`8
`14
`0
`t Time (days)
`i
`
`1.0 g
`
`0.5 g twice daily
`
`Figure 7 Plasma concentrations of metformin (asH Cl)
`after the first dose and during days 7 and 14 of continu(cid:173)
`ous treatment in Group II patients (bars indicate± s.d.).
`e observed,- predicted values.
`
`24
`
`36
`Time (h)
`
`48
`
`60
`
`72
`
`Figure 8 Plasma (0) and whole blood (Ill) concentra(cid:173)
`tions and urinary excretion rate (e) of metformin after
`1.5 g p.o. in subject 3.
`
`Discussion
`
`Apart from the work reported here two other groups
`have studied the kinetics of metformin after intraven(cid:173)
`ous administration to normal subjects (Sirtori et a/.,
`1978; Pentikainen eta/., 1979). A comparison of the
`findings is summarized in Table 6. There are some
`discrepancies between clearance and half-life values
`which, in part, may be related to differences in sampl(cid:173)
`ing periods and in curve-fitting procedures. The most
`important difference is seen when figures for the
`fraction of the dose recovered unchanged in the urine
`(fe) are. compared. Using (14C]-labelled metformin
`Pentikainen eta/. ( 1979) recovered all of the dose as
`intact drug in the urine whereas we did not account
`for 20% of the dose and must presume that this
`represents drug which has been metabolised. These
`observations are difficult to reconcile, firstly because
`animal studies have failed to detect any metabolites
`of the drug (Beckmann, 1969) and secondly because
`Pentikainen eta/. (1979) appear to have established,
`by
`thin-layer autoradiography,
`that all of the
`radiolabel in their urine samples was in the form of
`metformin. Nevertheless, our findings are supported
`by those of Sirtori et al. (1978). Although they con(cid:173)
`cluded that metformin is not metabolised, inspection
`of their data reveals that mean renal clearance was
`only 77% of mean total plasma clearance (Table 7),
`in accordance with our results. Incubation of our
`
`AUROBINDO EX. 1005, 11
`
`
`
`244
`
`G.T. TUCKER, C. CASEY, P.J. PHILLIPS, H. CONNOR, J.D. WARD & H.F. WOODS
`
`58 + 3.8 X
`y
`r
`0.85
`p < 0.001
`
`0
`
`b
`
`61 + 3.8 X
`Y =
`r
`0.88
`p < 0.001
`
`400
`
`200
`
`0
`
`0
`
`•
`
`•
`• •
`
`A
`
`0
`
`A G
`
`•
`
`0
`
`4
`t;.
`
`-
`
`.;....
`c::
`.E
`E 600
`
`Q)
`
`0 c::
`~
`co
`Q)
`0
`c;;
`c::
`~
`c::
`. E
`0
`
`-Q)
`
`~
`
`0
`
`A
`
`0
`
`0
`
`20
`
`60
`
`80 100 120 140 160 180
`
`Creatinine clearance (ml/min)
`
`Figure 9 Relationship between the renal clearance of metformin and creatinine clearance (a) Data from this study
`(0 Group I subjects-mean values for the two oral doses; A Group II patients; e Group ill patients). (b)As(a) but
`including data (.,6.) reported by Sirtori etal. (4).
`
`y = 294 + 6x
`r = 0.66
`p < 0.01
`
`0
`
`0
`
`0
`
`•
`
`0
`
`0
`
`•
`
`Ill
`
`•
`
`0
`
`A
`
`•• •
`•
`
`c: 1600
`
`.E -E
`
`; 1200
`~
`.£
`Q)
`0
`c:: 800
`~
`ctl
`Q)
`C3
`c::
`·E 4oo / /
`0
`i
`2
`
`0
`
`100
`140
`60
`20
`Creatinine clearance (ml/min)
`
`180
`
`Figure 10 Relationship between oral clearance of
`metformin-and creatinine clearance (symbols as in
`Figure 9).
`
`urine samples with a glucuronidase-sulphatase mix(cid:173)
`ture failed to increase the yield of metformin, indicat(cid:173)
`ing that simple conjugates were not formed.
`Since no drug was recovered in the faeces after
`intravenous administration net secretion of metfor(cid:173)
`min from blood into the gut lumen was negligible.
`Therefore, faecal recoveries of about 30% of an oral
`dose must represent unabsorbed drug. The estimated
`oral bioavailability of 50-60% agrees with the find(cid:173)
`ings of others (Sirtori etal., 1978; Pentikainen etal.,
`
`1979) and the difference between faecal recovery
`and bioavailability may reflect pre-systemic metabol(cid:173)
`ism of drug.
`Lower urinary recoveries and higher faecal re(cid:173)
`coveries after the 1.5 g compared to the 0.5 g oral
`dose in Group I subjects (Table 3) suggest that
`increasing dosage is accompanied by a decrease in the
`absorption of metformin. A similar conclusion was
`made by Noel (1979) who found that the fraction of
`the dose recovered in the urine fell from 0.86 to 0.42
`as the dose was raised from 0.25 g to 2.0 g.
`A longer plasma half-life after oral administration
`compared to that observed after intravenous ad(cid:173)
`ministration (Tables 2 and 3) indicates that slow
`absorption of metformin rate-limits its disposition
`over a significant period of time; an observation also
`made by Pentikainen et al. (1979). The calculations
`indicating that the majority of the dose which is going
`to be absorbed has done so by about 6 h (Figure 5)
`have implications for th