Acta pharmacol. et toxicol. 1977, 41, 254-262.
From the Institute of Pharmacology, University of Aarhus, DK-8000Aarhus, Denmark
Simultaneous Determination of Furosemide and Two of Its Possible Metabolites in Biological Fluids BY E. Mikkelsen and F. Andreasen (Received January 25, 1977; Accepted March 1, 1977)
Abstract: By the means of thin-layer chromatography (TLC) prior to fluorometry it was possible to determine furosemide (F)in plasma and urine with improved specificity. Antranilic acid (A) and 4-chlor-5-sulphamoyl-antranilic acid (CSA) could be determined simultaneously with F in plasma. In urine, only F and A could be determined quantitatively by this method. The sensitivity is 0.1 @ n l for F and 0.15-0.20 p g h l for A and CSA. By means of TLC prior to fluorometry the blank value for F was reduced in normal subjects and in different groups of patients, and its diurnal variation was eliminated. The average values of the plasma concentrations of F after the intravenous injection of 40 mg into 12 normal subjects were at all times found to be lower by the TLC method than by the fluorometric method without TLC. A higher plasma clearance and a lower volume of distribution were found for F when the values obtained by the TLC method were used for the calculation. Key-words: Furosemide method.
-
metabolites - man
-
urine
-
blood - determination
The fluorometric method for the determination of furosemide in plasma (HkTDiT & HAUSSLER1964; HXUSSLER & lUmCr 1964; ANDREASEN & JAKOBSEN 1974) is very sensitive and, particularly, in in vitro experiments sufficiently specific for many purpolses. If the fate of furosemide in the organism is to be followed and its extensive binding to the plasma proteins studied, an improved specificity is needed (ANDREASEN & MIKKELSEN 2974 & 1977). This need is obvious if the drug is given in large doses to patients with decreased ability to excrete the drug (LANCET1971; Postgraduate Medical Journal 1971). A quantitative fluorometric thin-layer chromatography (TLC) method for the determination of furosemide and two of its possible metabolites (cf. HAUSLER& WICHA1965) is described. The TLC method is compared with the fluorometric method mentioned above.
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
255
Materials and Methods Our procedure for the TLC determination of furosemide is a modification of the T SCHMID (1970). For urine, only the modified TLC method method described by K ~ D & was used, but for plasma the TLC method was compared with the fluorometric method & JAKOBSEN (1974). described by ANDREASEN Reagents. Furosemide (F), antranilic acid (A) and 4-chloro-5-sulphamoyl-antranilic acid (CSA) was kindly supplied by Hoechst. All other reagents were of analytical grade and obtained from Merck, Darmstadt, Germany. The diethyl ether was Diathylather fur die Spectroscopie (Merck). The precoated TLC plates Silica Gel 60 without fluorescence indicator were also obtained from Merck. They were 20 X 20 cm in size and the thickness of the layer was 0.25 mm. Before use they were activated for 20 minutes at 110". TLC method. In samples of both urine and plasma, attempts were made to perform simultaneous determination of the concentrations of F, A and CSA. Urine. Amounts of 0.4 ml 0.2 M H,PO(KH,PO, buffer (pH 2) and 5 ml diethyl ether were added to 1 ml of urine; after 10 minutes of vigorous shaking the samples were centrifuged for 10 minutes at 3000 r.p.m. in the usual laboratory centrifuge, and 4 ml of the ether phase was transferred to a rotating pear-shaped bottle and evaporated to dryness at reduced pressure at 26". 0.5 ml methanol was added. 100 p1 of the methanol was applied to the plate, and the chromatogram was developed (height 10 cm) in darkness using a mixture of chloroform, methanol and glacial acetic acid (89-6-5). After direct application of solutions of F, A and CSA the Rf values were 0.51, 0.78 and 0.27, respectively. A standard solution of each of these substances in methanol was applied to indicate the exact height of a possible fluorescent zone in the unknown sample. Areas of equal size (0.5 X 2.0 cm) corresponding to the localization of the standard substances were scraped off. The scraped-off powders were treated differently. For F, they were transferred to 2 ml of a phosphate buffer (pH 2), and after shaking and centrifuging the reading on the spectrophotofluorometer (AmincoBowman) was done at Eexcjt. 279 and Esmiea.416 nm. For A, the solvent was a borate buffer (pH 9) and Eclcit. was 320 and Eemiss.400. For CSA, a phosphate buffer (pH 7) was used and Ertxoit.was 273 and Eemis,. 398 nm. The calculations of the concentrations of the three substances in unknown samples were made from standard graphs produced by carrying solutions of known strength through the entire procedure. The recovery (in per cent) was calculated by relating the number of microyrammes which theoretically could be carried through the entire procedure to the number of microgrammes which were actually found in the final solution prepared for the reading in the spectrofluorometer. Plasma. Amounts of 100 pl 1 N-NaOH and 5 ml CHCl, were added to a 2 ml sample. After vigorous shaking and centrifuging, 1.5 ml of the supernatant was transferred to another tube, and 50 pl concentrated HCl and 5 ml diethyl ether were added. 4 ml of the ether was evaporated as described above for the urine assay, and the subsequent procedures were the same as described for urine. Plasma and urine samples. Plasma samples for the determination of possible variations in the blank values during the day for furosemide were obtained (1) from five healthy volunteers (aged 25-35 years) working in the laboratory, and (2) from seven bedridden patients with cardiac disease (mean age 69 years, range 50-78 years) who had not been treated with furosemide. Blank values were also determined by the two methods in plasma samples from five patients (mean age 57 years, range 45-70 years) under long-term treatment with furosemide. The samples were taken 18 hours after the last furosemide administration. Twelve volunteers (mean age 57 years, range 28-76
256
E. MIKKELSEN A N D F. ANDREASEN
years) with apparently normal hepatic, renal and cardiac function were informed of the procedures involved and given 40 mg of furosemide intravenously at 8 a.m. Blood samples were then drawn into heparinized glass tubes, a t 5, 15, 30, 45, 60, 90, 120, 180, 240 and 360 minutes, and the furosemide concentration in each sample was determined by both methods. A blood sample drawn from each patient prior to the injection was used as blank. For the elaboration of the urine assay, only in vitro addition to urine from normal volunteers was used. Calculations. The disappearance of furosemide from the plasma was described by a two-compartment open model, A Hewlett Packard computer (HP 30) was used, and & MIKKEL~EN the procedure was the same as that described previously (ANDREASEN 1977). Only (Vd)P (GIBALDI et a/. 1969; GIBALDI& ~ R R I E R1972), (V&%, v, and the et al. 1968; THOMPSON et al. (1973) were plasma clearance (RIGGS 1963; REGELMANN used for the comparison of the pharmacokinetic data obtained by the two methods.
Results Fig. 1 shows how the blank values for furosemide obtained by the two
--_J
8
10
I2
i
10
12
i4 16 TIME OF THE DAY
14 16 TIME OF THE DAY
Fig. 1. Diurnal variation in the blank value for the furosemide analysis measured in five normal volunteers. The upper curves show the results obtained by the method without TLC and the lower curves those obtained if TLC was performed prior to the reading on the fluorometer.
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
257
Table I. Diurnal variation in the plasma blank values, for the two furosemide analyses, in seven bedridden cardiac patients. They had not been treated with furosemide, but had received several other drugs. The relative fluorescence intensity (RFI) as well as the corresponding average number of microgrammes per ml are indicated for the two methods. Method
Spectrophotofluorornctty without
TLC
Patient
12.00
14.00
16.00
08.00
1 2 3 4 5 6 7
3.35 0.94 2.16 1.13 2.32 0.96 1.60 1.78 ? 0.89 1.73 f 0.85
3.05 0.86 1.76 1.05 2.48
2.62 1.03 I .54 0.87 2.33 0.78 0.83 1.43 k 0.76 1.38 f 0.73
4.80 2.45 1.32 1.66 2.21 0.82 0.88 2.03 k 1.36 1.98 k 1.31
Mean
f S.D. pLdm1
f S.D.
Spectrophotofluorornetry after TLC
1 2 3 4 5
6 7 Mean f S.D. Pkml
f S.D.
0.074 0.058 0.065 0.062 0.052 0.092 0.083 0.0694 k 0.0143 0.244 k 0.051
1.13
1.72 5 0.88 1.67 f 0.85 0.085 0.054 0.059 0.069 0.052 0.143
-
0.0770 k 0.0345
0.271
f 0.122
0.103 0.050 0.083 0.068
0.059 0.103 0.089 0.0793 k 0.0209 0.279 f 0.074
0.062 0.087 0.062 0.076 0.052 0.073 0.092 0.0720 k 0.0144 0.254 k 0.051
analytical methods changed during the day in five normal volunteers working in the laboratory. The blank values for the fluorometric method without TLC increased for all the individuals. The average increase from 08:OO to 16:OO hrs was 50 Vo of the morning value. The average blank values for the TLC method were stable during the day, and the number of microgrammes per ml represented by the blanks were reduced by 20-40 olo. The blank values in the bedridden cardiac patients who were not treated with furosemide, but with several other drugs (nitrazepam, penicillin, phenobarbital, digoxin) are shown in Table l. The values are higher than the blank values in the normal volunteers. Here no systematic change during the day is seen. The use of TLC reduces the blank values very considerably and eliminates most of the variation over the day. However, the values are still higher than for the blanks found in the normal subjects.
258
E. MIKKELSEN AND F. ANDREASEN
Table 2. Blank values expressed as pg furosemide per ml in plasma withdrawn before breakfast and morning medication from cardiac patients under long-term treatment with furosemide (120-160 mg daily).
Patient
Method SpectrophotoSPectroPhotofluorometry f 1uorom etry after TLC
Additional medication
pg/ml ~~
Digoxin Oxazepam Digoxin Quinidine Digoxin Aldactone Digoxin Aldactone Glibenclamide Digoxin
0.468
0.335
0.550
0.219
0.750
0.394
1.337
0.405
0.357
0.405
The highest blank values for furosemide were foand in bedridden cardiac patients who had been under long-term treatment with furosemide as well as other drugs. As seen from Table 2, the blanks were markedly reduced when the TLC method was used, but they were still higher than the values obtained in normal subjects and the patients who had not been treated with furosemide. In Table 2, the sensitivity, recovery and reproducibility of the TLC method are shown for furosemide as well as for A and CSA. A quantitative deter-
Table 3. The sensitivity, recovery and reproducibility for the TLC determination of furosemide (F), antranilic acid (A) and 4-chloro-5-sulphamoyl-antranilic acid (CSA). Plasma
F
A
0.1 50.0
0.2 11.3
Urine CSA
F
A
0.1 57.0
0.1 38.0
CSA
~~
Sensitivity Recovery Added (n = 7-10) Found ( f S.D.)
pg/ml % pg/ml Idm1
2.5 2.40 k 0.10
1.0 1.19 f 0.63
0.15 37.0 1.0 1.31 f 0.88
10.0 10.0 f 0.35
l.o 1.02
f 0.23
Only qualitative
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
” c1
E
7
259
X TOTRL METHOD I T L C VETHOD
E
+
THE
DIFFERENCE
TOTRL-TLC
M I NUTES 68
128
180
21.10
308
368
1.128
L(E8
Fig. 2. The average plasma concentration of furosernide in 12 normal subjects after intravenous injection of 40 mg. The two curves were drawn using the results obtained by the two different analytical methods as indicated.
mination of CSA in the plasma was possible by this method, but in urine, the blank values were found to be highly variable. Accordingly, no CSA values are shown for urine. The sensitivity is highest folr furosemide. Fig. 2 illustrates the average plasma concentrations of furosemide in 12 normal subjects after intravenous administration of 40 mg. In each plasma sample, the furosemide concentration was determined in duplicate by both methods. The averages of the 10 plasma concentrations determined from 5 minutes to 6 hours after the administration were all higher when the method without TLC was used. The difference between the results obtained by the two methods is indicated as &ml furosemide. The data from fig. 2 were used to calculate the pharmacokinetical parameters listed in Table 4. The values obtained by the TLC method seem to result in a shorter half-life for the distribution and elimination phases, a smaller volume of distribution and a higher plasma clearance.
Discussion
The fluorometric method for the determination of furosemide -6 & HAUSSLER1964; and ANDREASEN & JAKOBSEN1974) is more sensitive than as well the GLC-method described by LINDSTROM& MOLANDER (1974) as the HPLC method described by MACDQUGALL et al. (1975). However the specificity is not always satisfactory for clinical pharmacukinetic studies. The
260
E. MIKKELSEN AND F. ANDREASEN
specificity is markedly improved by the chromatographic methods mentioned above as well as by the TLC method described here. The variation in the blank values for fluorometric analysis of urinary extracts was emphasized by FORREYet al. (1974), but variations in the plasma blank values over the day were usually disregarded when fluorometric analyses are performed. We have shown that the plasma blank values for the fluorometric analysis of furosemide increased during the day in normal subjects working in the laboratory. In bedridden patients under long-term treatment with furosemidz and/or other drugs no fixed diurnal rhythm was found, but the plasma blanks were higher, and both the intra- and the interindividual variations were more pronounced. By the use of TLC prior to the fluorometric reading, the blanks were reduced in all categories of subjects; the interindividual variation was reduced considerably, and the variations over the day were eliminated. However, the blanks were still higher in the cardiac patients than in the normal subjects. Fluorometric determination preceded by TLC thus has a higher specificity than the fluorometric method without TLC,but the sensitivity is lower. The TLC method can be recommended for in vivo determination of furosemide, while the more sensitive but less specific fluorometric method should be useful in in vitro experiments. The average furosemide concentrations in plasma after intravenous administration of 40 mg in 12 volunteers were invariably lower when the TLC method was used for the determination. No attempts were made to determine the two metabolites in these plasma samples, and it cannot be established whether the difference observed between the two methods is due to
Table 4. Pharmacokinetic parameters for furosemide ( ? S.D.) calculated from the average plasma concentrations (cf. fig. 2) in 12 normal subjects after intravenous administration af 40 mg. The furosemide concentration was determined without TLC and with TLC. Half-life DistnEliminabution tion phase phase Alpha Beta phase phase min. min. Without
10.5
76.2
With
litres 6.82
k 0.38
TLC
TLC
VO
8.0
57.2
7.20
f 0.41
('d)~
litres 14.46 rt 5.78 13.81 rt 6.32
Wd)B litres
Plasma clearance ml/min.
19.33
175.83
k 3.61
f 30.85
17.13 k 3.79
207.56 k 34.13
-
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
261
degradation products or metabolic products of furmemide (A, CSA or others). However, such compounds can be extracted, and if they are present, they will contribute to the fluorescence. Thus the existence of such products would explain the (non-significant) difference between the pharmacokinetic data listed in Table 4: (1) The volume of distribution - (Va)p or (V& - is higher for the “total” method. Furosemide is more strongly bound to protein than the two possible metabolites (ANDREASEN, HUSTED& MIKKELSEN, unpublished results) - cf. MARTIN(1965). (2) Furosemide is excreted by tubular secretion, while this is probably not (or at least to a much smaller extent) the case for its metabolic or degradation products, This would explain the shorter half-life and the higher plasma clearance obtained by the TLC method. (3) The slightly higher initial volume of distribution found by the TLC method may well be coincidental (cf. the S.D. values). However, a more intense specific binding of the unchanged furosemide molecules to the tubular cells during the first few passages of blood through the kidneys might cause a relatively larger reduction in the first two plasma concentrations found by the TLC method. In this way, a “first-pass effect” of the kidney could cause the apparently larger V, found with the TLC method.
REFERENCES Andreasen, F. & P. Jakobsen: Determination of furosemide in blood plasma and its binding to proteins in normal plasma and in plasma from patients with acute renal failure. Acia pharmacol. et toxicol. 1974, 35, 49-57. Andreasen, F. & E. Mikkelsen: Protein binding of furosemide in the plasma from patients with cardiac disease and in normal plasma. Acta pharmacot et roxicol. 1974, 35, suppl. I, 18. Andreasen, F. & E. Mikkelsen: Distribution, elimination and effect of furosemide in normal subjects and in patients with heart failure. Eur. J . clin. Pharmacol. 1977, in press. Forrey, A. W., B. Kimpel, A. D. Blair & R. E. Cutler: Furosemide concentrations in serum and urine, and its binding by serum proteins as measured fluorometrically. Clin. Chem. 1974,20, 152-158. Gibaldi, M.,R. Nagashima & G. Levy: Relationship between drug concentration in plasma or serum and amount of drug in the body. J . pharm. Sci. 1969, 58, 193197. Gibaldi, M. & D. Perrier: Drug elimination and apparent volume of distribution in multicompartment systems. J . pharm. Sci. 1972, 61,952-953. Hajdii, P. & A. Haussler: Untersuchungen rnit dem Salidiureticum 4-Chlor-N-(2furylmethyl)-5-sulfamyl-anthranil-saurs.Arzneimittelforsch. 1964, 14, 709-710. Haussler, A. & P. Hajdk Untersuchungen mit dem Salidiureticum 4-Chlor-N(2furylmethyl)-5-sulfamyl-mthranilsaure.ArzneimitteZforsch. 1964, 14,71&713. Haussler, A. & H. Wicha: Untersuchungen mit dem Salidiureticum 4-Chlor-N-(furylmethyl)-5-sulfamyl-anthranilsaure.Arzneimittelforsch. 1965, 15, 81-83. Kindt, H. & E. Schmid: Uber die Harnausscheidung von Furosemid bei Gesunden und Kranken mit Lebercirrhose. Pharmacol. Clin. 1970,2,221-226.
E. MIKKELSEN AND F. ANDREASEN
262
Lancet: Editorial, 1971,2,803-804. Lindstrom, B. & M. Molander: Gas chromatographic determination of furosemide in plasma using an extractive alkylation technique and an electron capture detector. J . Chroinat. 1974,101,219-221. MacDougall, M. L., D. W. Shoeman & D. L. Azarnoff: Separation and quantitative analysis of furosemide and 4-chloro-5-sulfamoylanthranilicacid (CSA) by high pressure liquid chromatography. Res. C o m m . Chem. Path. Pharm. 1975, 10, 285-
292. Martin, B. K.: Potential effect of the plasma protein on drug distribution. Nature 1965, 207,274-276. Postgraduate Medical J . 1971,47, suppl., 1-58. Riegelman, S., J. Loo & M. Rowland: Concept of a volume of distribution and possible errors in evaluation of this parameter. J . pharm. Sci. 1968,57, 128-133. Riggs, D.S.: The mathentatical approach to physiological problems. William and Wilkins, Baltimore, Md., 1963,pp. 193-217. Thomson, P. D., K. L., Melmon, J. A. Richardson, K. Cohn, W. Steinbrunn, R. Cucihee & M. Rowland: Lidocaine pharmacokinetics in advanced heart failure, liver disease, and renal failure in humans. Ann. intern. Med. 1973,78, 499-508.
From the Institute of Pharmacology, University of Aarhus, DK-8000Aarhus, Denmark
Simultaneous Determination of Furosemide and Two of Its Possible Metabolites in Biological Fluids BY E. Mikkelsen and F. Andreasen (Received January 25, 1977; Accepted March 1, 1977)
Abstract: By the means of thin-layer chromatography (TLC) prior to fluorometry it was possible to determine furosemide (F)in plasma and urine with improved specificity. Antranilic acid (A) and 4-chlor-5-sulphamoyl-antranilic acid (CSA) could be determined simultaneously with F in plasma. In urine, only F and A could be determined quantitatively by this method. The sensitivity is 0.1 @ n l for F and 0.15-0.20 p g h l for A and CSA. By means of TLC prior to fluorometry the blank value for F was reduced in normal subjects and in different groups of patients, and its diurnal variation was eliminated. The average values of the plasma concentrations of F after the intravenous injection of 40 mg into 12 normal subjects were at all times found to be lower by the TLC method than by the fluorometric method without TLC. A higher plasma clearance and a lower volume of distribution were found for F when the values obtained by the TLC method were used for the calculation. Key-words: Furosemide method.
-
metabolites - man
-
urine
-
blood - determination
The fluorometric method for the determination of furosemide in plasma (HkTDiT & HAUSSLER1964; HXUSSLER & lUmCr 1964; ANDREASEN & JAKOBSEN 1974) is very sensitive and, particularly, in in vitro experiments sufficiently specific for many purpolses. If the fate of furosemide in the organism is to be followed and its extensive binding to the plasma proteins studied, an improved specificity is needed (ANDREASEN & MIKKELSEN 2974 & 1977). This need is obvious if the drug is given in large doses to patients with decreased ability to excrete the drug (LANCET1971; Postgraduate Medical Journal 1971). A quantitative fluorometric thin-layer chromatography (TLC) method for the determination of furosemide and two of its possible metabolites (cf. HAUSLER& WICHA1965) is described. The TLC method is compared with the fluorometric method mentioned above.
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
255
Materials and Methods Our procedure for the TLC determination of furosemide is a modification of the T SCHMID (1970). For urine, only the modified TLC method method described by K ~ D & was used, but for plasma the TLC method was compared with the fluorometric method & JAKOBSEN (1974). described by ANDREASEN Reagents. Furosemide (F), antranilic acid (A) and 4-chloro-5-sulphamoyl-antranilic acid (CSA) was kindly supplied by Hoechst. All other reagents were of analytical grade and obtained from Merck, Darmstadt, Germany. The diethyl ether was Diathylather fur die Spectroscopie (Merck). The precoated TLC plates Silica Gel 60 without fluorescence indicator were also obtained from Merck. They were 20 X 20 cm in size and the thickness of the layer was 0.25 mm. Before use they were activated for 20 minutes at 110". TLC method. In samples of both urine and plasma, attempts were made to perform simultaneous determination of the concentrations of F, A and CSA. Urine. Amounts of 0.4 ml 0.2 M H,PO(KH,PO, buffer (pH 2) and 5 ml diethyl ether were added to 1 ml of urine; after 10 minutes of vigorous shaking the samples were centrifuged for 10 minutes at 3000 r.p.m. in the usual laboratory centrifuge, and 4 ml of the ether phase was transferred to a rotating pear-shaped bottle and evaporated to dryness at reduced pressure at 26". 0.5 ml methanol was added. 100 p1 of the methanol was applied to the plate, and the chromatogram was developed (height 10 cm) in darkness using a mixture of chloroform, methanol and glacial acetic acid (89-6-5). After direct application of solutions of F, A and CSA the Rf values were 0.51, 0.78 and 0.27, respectively. A standard solution of each of these substances in methanol was applied to indicate the exact height of a possible fluorescent zone in the unknown sample. Areas of equal size (0.5 X 2.0 cm) corresponding to the localization of the standard substances were scraped off. The scraped-off powders were treated differently. For F, they were transferred to 2 ml of a phosphate buffer (pH 2), and after shaking and centrifuging the reading on the spectrophotofluorometer (AmincoBowman) was done at Eexcjt. 279 and Esmiea.416 nm. For A, the solvent was a borate buffer (pH 9) and Eclcit. was 320 and Eemiss.400. For CSA, a phosphate buffer (pH 7) was used and Ertxoit.was 273 and Eemis,. 398 nm. The calculations of the concentrations of the three substances in unknown samples were made from standard graphs produced by carrying solutions of known strength through the entire procedure. The recovery (in per cent) was calculated by relating the number of microyrammes which theoretically could be carried through the entire procedure to the number of microgrammes which were actually found in the final solution prepared for the reading in the spectrofluorometer. Plasma. Amounts of 100 pl 1 N-NaOH and 5 ml CHCl, were added to a 2 ml sample. After vigorous shaking and centrifuging, 1.5 ml of the supernatant was transferred to another tube, and 50 pl concentrated HCl and 5 ml diethyl ether were added. 4 ml of the ether was evaporated as described above for the urine assay, and the subsequent procedures were the same as described for urine. Plasma and urine samples. Plasma samples for the determination of possible variations in the blank values during the day for furosemide were obtained (1) from five healthy volunteers (aged 25-35 years) working in the laboratory, and (2) from seven bedridden patients with cardiac disease (mean age 69 years, range 50-78 years) who had not been treated with furosemide. Blank values were also determined by the two methods in plasma samples from five patients (mean age 57 years, range 45-70 years) under long-term treatment with furosemide. The samples were taken 18 hours after the last furosemide administration. Twelve volunteers (mean age 57 years, range 28-76
256
E. MIKKELSEN A N D F. ANDREASEN
years) with apparently normal hepatic, renal and cardiac function were informed of the procedures involved and given 40 mg of furosemide intravenously at 8 a.m. Blood samples were then drawn into heparinized glass tubes, a t 5, 15, 30, 45, 60, 90, 120, 180, 240 and 360 minutes, and the furosemide concentration in each sample was determined by both methods. A blood sample drawn from each patient prior to the injection was used as blank. For the elaboration of the urine assay, only in vitro addition to urine from normal volunteers was used. Calculations. The disappearance of furosemide from the plasma was described by a two-compartment open model, A Hewlett Packard computer (HP 30) was used, and & MIKKEL~EN the procedure was the same as that described previously (ANDREASEN 1977). Only (Vd)P (GIBALDI et a/. 1969; GIBALDI& ~ R R I E R1972), (V&%, v, and the et al. 1968; THOMPSON et al. (1973) were plasma clearance (RIGGS 1963; REGELMANN used for the comparison of the pharmacokinetic data obtained by the two methods.
Results Fig. 1 shows how the blank values for furosemide obtained by the two
--_J
8
10
I2
i
10
12
i4 16 TIME OF THE DAY
14 16 TIME OF THE DAY
Fig. 1. Diurnal variation in the blank value for the furosemide analysis measured in five normal volunteers. The upper curves show the results obtained by the method without TLC and the lower curves those obtained if TLC was performed prior to the reading on the fluorometer.
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
257
Table I. Diurnal variation in the plasma blank values, for the two furosemide analyses, in seven bedridden cardiac patients. They had not been treated with furosemide, but had received several other drugs. The relative fluorescence intensity (RFI) as well as the corresponding average number of microgrammes per ml are indicated for the two methods. Method
Spectrophotofluorornctty without
TLC
Patient
12.00
14.00
16.00
08.00
1 2 3 4 5 6 7
3.35 0.94 2.16 1.13 2.32 0.96 1.60 1.78 ? 0.89 1.73 f 0.85
3.05 0.86 1.76 1.05 2.48
2.62 1.03 I .54 0.87 2.33 0.78 0.83 1.43 k 0.76 1.38 f 0.73
4.80 2.45 1.32 1.66 2.21 0.82 0.88 2.03 k 1.36 1.98 k 1.31
Mean
f S.D. pLdm1
f S.D.
Spectrophotofluorornetry after TLC
1 2 3 4 5
6 7 Mean f S.D. Pkml
f S.D.
0.074 0.058 0.065 0.062 0.052 0.092 0.083 0.0694 k 0.0143 0.244 k 0.051
1.13
1.72 5 0.88 1.67 f 0.85 0.085 0.054 0.059 0.069 0.052 0.143
-
0.0770 k 0.0345
0.271
f 0.122
0.103 0.050 0.083 0.068
0.059 0.103 0.089 0.0793 k 0.0209 0.279 f 0.074
0.062 0.087 0.062 0.076 0.052 0.073 0.092 0.0720 k 0.0144 0.254 k 0.051
analytical methods changed during the day in five normal volunteers working in the laboratory. The blank values for the fluorometric method without TLC increased for all the individuals. The average increase from 08:OO to 16:OO hrs was 50 Vo of the morning value. The average blank values for the TLC method were stable during the day, and the number of microgrammes per ml represented by the blanks were reduced by 20-40 olo. The blank values in the bedridden cardiac patients who were not treated with furosemide, but with several other drugs (nitrazepam, penicillin, phenobarbital, digoxin) are shown in Table l. The values are higher than the blank values in the normal volunteers. Here no systematic change during the day is seen. The use of TLC reduces the blank values very considerably and eliminates most of the variation over the day. However, the values are still higher than for the blanks found in the normal subjects.
258
E. MIKKELSEN AND F. ANDREASEN
Table 2. Blank values expressed as pg furosemide per ml in plasma withdrawn before breakfast and morning medication from cardiac patients under long-term treatment with furosemide (120-160 mg daily).
Patient
Method SpectrophotoSPectroPhotofluorometry f 1uorom etry after TLC
Additional medication
pg/ml ~~
Digoxin Oxazepam Digoxin Quinidine Digoxin Aldactone Digoxin Aldactone Glibenclamide Digoxin
0.468
0.335
0.550
0.219
0.750
0.394
1.337
0.405
0.357
0.405
The highest blank values for furosemide were foand in bedridden cardiac patients who had been under long-term treatment with furosemide as well as other drugs. As seen from Table 2, the blanks were markedly reduced when the TLC method was used, but they were still higher than the values obtained in normal subjects and the patients who had not been treated with furosemide. In Table 2, the sensitivity, recovery and reproducibility of the TLC method are shown for furosemide as well as for A and CSA. A quantitative deter-
Table 3. The sensitivity, recovery and reproducibility for the TLC determination of furosemide (F), antranilic acid (A) and 4-chloro-5-sulphamoyl-antranilic acid (CSA). Plasma
F
A
0.1 50.0
0.2 11.3
Urine CSA
F
A
0.1 57.0
0.1 38.0
CSA
~~
Sensitivity Recovery Added (n = 7-10) Found ( f S.D.)
pg/ml % pg/ml Idm1
2.5 2.40 k 0.10
1.0 1.19 f 0.63
0.15 37.0 1.0 1.31 f 0.88
10.0 10.0 f 0.35
l.o 1.02
f 0.23
Only qualitative
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
” c1
E
7
259
X TOTRL METHOD I T L C VETHOD
E
+
THE
DIFFERENCE
TOTRL-TLC
M I NUTES 68
128
180
21.10
308
368
1.128
L(E8
Fig. 2. The average plasma concentration of furosernide in 12 normal subjects after intravenous injection of 40 mg. The two curves were drawn using the results obtained by the two different analytical methods as indicated.
mination of CSA in the plasma was possible by this method, but in urine, the blank values were found to be highly variable. Accordingly, no CSA values are shown for urine. The sensitivity is highest folr furosemide. Fig. 2 illustrates the average plasma concentrations of furosemide in 12 normal subjects after intravenous administration of 40 mg. In each plasma sample, the furosemide concentration was determined in duplicate by both methods. The averages of the 10 plasma concentrations determined from 5 minutes to 6 hours after the administration were all higher when the method without TLC was used. The difference between the results obtained by the two methods is indicated as &ml furosemide. The data from fig. 2 were used to calculate the pharmacokinetical parameters listed in Table 4. The values obtained by the TLC method seem to result in a shorter half-life for the distribution and elimination phases, a smaller volume of distribution and a higher plasma clearance.
Discussion
The fluorometric method for the determination of furosemide -6 & HAUSSLER1964; and ANDREASEN & JAKOBSEN1974) is more sensitive than as well the GLC-method described by LINDSTROM& MOLANDER (1974) as the HPLC method described by MACDQUGALL et al. (1975). However the specificity is not always satisfactory for clinical pharmacukinetic studies. The
260
E. MIKKELSEN AND F. ANDREASEN
specificity is markedly improved by the chromatographic methods mentioned above as well as by the TLC method described here. The variation in the blank values for fluorometric analysis of urinary extracts was emphasized by FORREYet al. (1974), but variations in the plasma blank values over the day were usually disregarded when fluorometric analyses are performed. We have shown that the plasma blank values for the fluorometric analysis of furosemide increased during the day in normal subjects working in the laboratory. In bedridden patients under long-term treatment with furosemidz and/or other drugs no fixed diurnal rhythm was found, but the plasma blanks were higher, and both the intra- and the interindividual variations were more pronounced. By the use of TLC prior to the fluorometric reading, the blanks were reduced in all categories of subjects; the interindividual variation was reduced considerably, and the variations over the day were eliminated. However, the blanks were still higher in the cardiac patients than in the normal subjects. Fluorometric determination preceded by TLC thus has a higher specificity than the fluorometric method without TLC,but the sensitivity is lower. The TLC method can be recommended for in vivo determination of furosemide, while the more sensitive but less specific fluorometric method should be useful in in vitro experiments. The average furosemide concentrations in plasma after intravenous administration of 40 mg in 12 volunteers were invariably lower when the TLC method was used for the determination. No attempts were made to determine the two metabolites in these plasma samples, and it cannot be established whether the difference observed between the two methods is due to
Table 4. Pharmacokinetic parameters for furosemide ( ? S.D.) calculated from the average plasma concentrations (cf. fig. 2) in 12 normal subjects after intravenous administration af 40 mg. The furosemide concentration was determined without TLC and with TLC. Half-life DistnEliminabution tion phase phase Alpha Beta phase phase min. min. Without
10.5
76.2
With
litres 6.82
k 0.38
TLC
TLC
VO
8.0
57.2
7.20
f 0.41
('d)~
litres 14.46 rt 5.78 13.81 rt 6.32
Wd)B litres
Plasma clearance ml/min.
19.33
175.83
k 3.61
f 30.85
17.13 k 3.79
207.56 k 34.13
-
FUROSEMIDE AND METABOLITES IN BLOOD AND URINE
261
degradation products or metabolic products of furmemide (A, CSA or others). However, such compounds can be extracted, and if they are present, they will contribute to the fluorescence. Thus the existence of such products would explain the (non-significant) difference between the pharmacokinetic data listed in Table 4: (1) The volume of distribution - (Va)p or (V& - is higher for the “total” method. Furosemide is more strongly bound to protein than the two possible metabolites (ANDREASEN, HUSTED& MIKKELSEN, unpublished results) - cf. MARTIN(1965). (2) Furosemide is excreted by tubular secretion, while this is probably not (or at least to a much smaller extent) the case for its metabolic or degradation products, This would explain the shorter half-life and the higher plasma clearance obtained by the TLC method. (3) The slightly higher initial volume of distribution found by the TLC method may well be coincidental (cf. the S.D. values). However, a more intense specific binding of the unchanged furosemide molecules to the tubular cells during the first few passages of blood through the kidneys might cause a relatively larger reduction in the first two plasma concentrations found by the TLC method. In this way, a “first-pass effect” of the kidney could cause the apparently larger V, found with the TLC method.
REFERENCES Andreasen, F. & P. Jakobsen: Determination of furosemide in blood plasma and its binding to proteins in normal plasma and in plasma from patients with acute renal failure. Acia pharmacol. et toxicol. 1974, 35, 49-57. Andreasen, F. & E. Mikkelsen: Protein binding of furosemide in the plasma from patients with cardiac disease and in normal plasma. Acta pharmacot et roxicol. 1974, 35, suppl. I, 18. Andreasen, F. & E. Mikkelsen: Distribution, elimination and effect of furosemide in normal subjects and in patients with heart failure. Eur. J . clin. Pharmacol. 1977, in press. Forrey, A. W., B. Kimpel, A. D. Blair & R. E. Cutler: Furosemide concentrations in serum and urine, and its binding by serum proteins as measured fluorometrically. Clin. Chem. 1974,20, 152-158. Gibaldi, M.,R. Nagashima & G. Levy: Relationship between drug concentration in plasma or serum and amount of drug in the body. J . pharm. Sci. 1969, 58, 193197. Gibaldi, M. & D. Perrier: Drug elimination and apparent volume of distribution in multicompartment systems. J . pharm. Sci. 1972, 61,952-953. Hajdii, P. & A. Haussler: Untersuchungen rnit dem Salidiureticum 4-Chlor-N-(2furylmethyl)-5-sulfamyl-anthranil-saurs.Arzneimittelforsch. 1964, 14, 709-710. Haussler, A. & P. Hajdk Untersuchungen mit dem Salidiureticum 4-Chlor-N(2furylmethyl)-5-sulfamyl-mthranilsaure.ArzneimitteZforsch. 1964, 14,71&713. Haussler, A. & H. Wicha: Untersuchungen mit dem Salidiureticum 4-Chlor-N-(furylmethyl)-5-sulfamyl-anthranilsaure.Arzneimittelforsch. 1965, 15, 81-83. Kindt, H. & E. Schmid: Uber die Harnausscheidung von Furosemid bei Gesunden und Kranken mit Lebercirrhose. Pharmacol. Clin. 1970,2,221-226.
E. MIKKELSEN AND F. ANDREASEN
262
Lancet: Editorial, 1971,2,803-804. Lindstrom, B. & M. Molander: Gas chromatographic determination of furosemide in plasma using an extractive alkylation technique and an electron capture detector. J . Chroinat. 1974,101,219-221. MacDougall, M. L., D. W. Shoeman & D. L. Azarnoff: Separation and quantitative analysis of furosemide and 4-chloro-5-sulfamoylanthranilicacid (CSA) by high pressure liquid chromatography. Res. C o m m . Chem. Path. Pharm. 1975, 10, 285-
292. Martin, B. K.: Potential effect of the plasma protein on drug distribution. Nature 1965, 207,274-276. Postgraduate Medical J . 1971,47, suppl., 1-58. Riegelman, S., J. Loo & M. Rowland: Concept of a volume of distribution and possible errors in evaluation of this parameter. J . pharm. Sci. 1968,57, 128-133. Riggs, D.S.: The mathentatical approach to physiological problems. William and Wilkins, Baltimore, Md., 1963,pp. 193-217. Thomson, P. D., K. L., Melmon, J. A. Richardson, K. Cohn, W. Steinbrunn, R. Cucihee & M. Rowland: Lidocaine pharmacokinetics in advanced heart failure, liver disease, and renal failure in humans. Ann. intern. Med. 1973,78, 499-508.
