The described method for quantitative determination of griseofulvin

in plasma offers several advantages over the spectrofluorometric (2,7,

8 ) and GLC (8) methods. The HPLC procedure is simple and rapid, with
a turnaround time of &6 min required for each sample. The analysis uses
only a small volume of plasma and is specific and reproducible; commonly
used drugs do not interfere. The method is well suited for the clinical
monitoring of plasma griseofulvin concentrations or for pharmacokinetic
studies.

REF EREN CES

(1) “The Pharmacological Basis of Therapeutics,” L. S. Goodman
and A. Gilman, Eds., Macmillan, New York, N.Y., 1975.
(2) M. Rowland, S. Riegelman, and W. L. Epstein, J . Pharm. Sci.,
57,984 (1968).
(3) W. L. Chiou and S. Riegelman, ibid., 60,1376 (1971).
(4) S. Riegelman, M. Rowland, and W. L. Epstein, J. Am. Med.
Assoc., 213,426 (1970).

r J l , l , z r
JQ 0.01
(5) D. Busfield, K. J. Child, R. M. Atkinson, and E. G. Tomich,
Lancet, 2,1042 (1963).
(6) S. Riegelman and W. Sadee, “Clinical Pharmacokinetics: A
a 1 2 3 4 5 6 Symposium,” APhA Academy of Pharmaceutical Sciences, Washington,
HOURS D.C., 1974, p. 169.
(7) V. P. Shah, S. Riegelman, and W. L. Epstein, J . Pharm. Sci.,61,
Figure 4-Plasma griseofulvin concentration as a function of time in 634 (1972).
a 4-kg rabbit giuen 16 mg ofgriseojuluin intravenously ouer I min. (8) H. J. Schwarz, B. A. Waldman, and V. Madrid, ibid., 65, 370
(1976).
(9) W. L. Chiou, M. A. F. Gadalla, and G . W. Peng, ibid., 67, 182
jected into the liquid chromatograph, indicating that the peak was not (1978).
due to another component or an impurity in the griseofulvin. It seems (10) J. W. Nelson, A. L. Cordry, C. G. Aron, and R. A. Bartell, Clin.
most likely that the peak was due to a metabolite of griseofulvin, although Chem., 23,124 (1977).
the structure of this compound is unknown. The time course of plasma (11) I. Nilsson-Ehle, T. T. Yoshikawa, J. E. Edwards, M. C. Schotz,
griseofulvin concentrations in the rabbit is shown in Fig. 4. and L. B. Gum, J . Inject. Dis., 135,414 (1977).

Stability of Furosemide in Aqueous Systems

A. G. GHANEKAR, V. DAS GUPTA x, and CHARLES W. GIBBS, Jr. *

Received February 28,1977, from the College o f Pharmacy, University of Houston, Houston, T X 77004. Accepted for publication September
28, 1977. ‘Present address: Harris County Hospital District Pharmacy, Taub Loop, Houston, T X 77030.

Abstract 0 A stability-indicating assay for furosemide based on high- USP (1)indicates that the pH of the injection should be
pressure liquid chromatography was developed. The method is sensitive, between 8.9 and 9.3.
accurate, and precise. The standard deviation based on six injections of The present investigations were undertaken to develop
the standard solution was f1.37%. This method was used to study furo- a stability-indicating assay for furosemide using high-
semide stability in various aqueous solutions and dosage forms. Stability
tests were conducted a t room temperature as well as a t higher tempera- pressure liquid chromatography (HPLC) and to study
tures (45,65, and 85”) at various pH values and with different vehicles. furosemide stability in aqueous systems.
Some decomposition products were identified.
EXPERIMENTAL
Keyphrases o Furosemide-high-pressure liquid chromatographic
analysis, stability in aqueous solutions and dosage forms 0 High-pressure Chemicals and Reagents-All chemicals and reagents were ACS, NF,
liquid chromatography-analysis, furosemide in aqueous solutions and or USP quality and were used without further purification.
dosage forms Stability-furosemide in aqueous solutions and dosage Pr ep ar at i o n of Solutions a n d Dosage Forms-The solutions and
forms, high-pressure liquid chromatographic analysis 0 Diuretics- dosage forms prepared are given in Table I.
furosemide, high-pressure liquid chromatographic analysis, stability in Apparatus-A high-pressure liquid chromatograph1 equipped with
aqueous solutions and dosage forms a UV detector (254 nm), a recorder*, and an integrator? was used.
Column-A column4 (30 cm X 4 mm i.d.) of a very nonpolar packing
material, consisting of octadecyltrichlorosilane permanently bonded by
Furosemide is extensively used as a diuretic. Very little silicon-carbon bonds, was used.
information is available concerning its stability, possibly Chromatographic Conditions-The chromatographic mobile phase
because of the lack of a stability-indicating assay. The USP was 0.01 M (NH4)2HP04in water containing 25% (v/v) methanol. The
method (1) for furosemide in injections and tablets is based
on UV absorption. A fluorometric method for furosemide Model ALC 202, equipped with U6K universal chromatograph injector, Waters
in urine and serum also was reported (2), but any other Associates, Milford, Mass.
2 Omniscribe 5213-12, Houston Instruments, Austin, Tex.
fluorescent substances interfere with the assay. Furo- Autolah Minigrator,Spectra-Physics,Santa Clara, Calif.
semide was reported to be unstable in acidic media (3). The VBondapak Cis, Catalog No. 27324, Waters Associates, Milford,Mass.

808 I Journal of Pharmaceutical Sciences

Table I-Solutions an d Dosage Forms P r e p a r e d
Furosemide
Concentration, Other Ingredients
Solution Vehicle mdml DH (If Anv)
1 0.01
. .~~N N a-~~~
~
OH n.lia -
2 O.1NNaOH 1.0 13
3 0.1 N HC1 1.0 1.2
4 0.04 M (NHd)sHP04 1.0 8
5 Phosphate buffers (0.05 M) 0.5 See Table IV -h.c
6 Water 0.5 4.2 -h , c
(adjusted with acetic acid)
7 Water 1.0h See Table 111 Various percentages of sugar
8 Commercial vehicle' 1.0 and 7.4d
2.0 7.5d
9 50% (v/v) glycerin in water 1.0 5.6 Methylparaben, 0.005%
Propylparaben, 0.002%
10 70% sorbitol in water 1.o 6.8 Methylparaben, 0.005%
Propylparaben, 0.002%
11 50% sorbitol in water 1.0 8.5 Methylparaben, 0.005%
(adjusted with 0.1 N NaOH) Propylparaben, 0.002%
12 50% sorbitol in water with 20% (v/v) alcohol 1.0 8.5 Methylparaben, 0.005%
Propylparaben, 0.002%
13 50% sorbitol in water with 10%(v/v) alcohol 1.0 8.5 Methylparaben, 0.005%
Propylparaben, 0.00%
14 85% sugar in water LO* 8.2 -h
(adjusted with 4% NaHCOa)
*
For a standard solution, it was further diluted with water to a concentration of 50 Gglml. Prepared from furosemide injection. Syrpalta syrup, Emerson Laboratories,
Dallas, Tex. The syrup also contained dye(s), flavor(s),and sodium benzoate (as confirmed during these investigations).An 8.0-ml quantity of 4% NaHC03/50 ml of the
syrup was added to raise the pH to keep furosemide in solution. The pH values are of the recompounded samples; pH values of the originai samples were not deter-
mined

temperature was ambient, and the flow rate was 2.0 ml/min. The detector pH was adjusted to 8.2 with sodium bicarbonate gave 26.1% by the de-
was set a t a sensitivity of 0.16 absorbance unit for full-scale deflection, veloped method (Fig. 2C) versus 82.1% by the USP method. Therefore,
and the chart speed was 30.5 cm/hr. stability studies on the dosage forms and solutions were conducted using
P roced u r e for Furosemide-In Solution and Liquid Dosage the developed method.
Forms-An appropriate quantity of the solution or dosage form was di- Stability Studies-Solution in Commercial Vehicle Containing 2.0
luted with water to a 50-pg/ml concentration, and 40 pl was injected. For mg of Furosemidelml (Table [)-This solution was divided into various
purposes of comparison, an identical volume of the standard solution was portions and stored at 45,65, and 85' as well as at room temperature. The
injected after the assay solution was eluted. portions were assayed frequently (Table 11). A solution of furosemide in
I n Tablets-One tablet was ground and mixed with 4 ml of methanol. the commercial vehicle (1.0 mg/ml) when stored a t room temperature
Then enough 0.05 M sodium bicarbonate was added to bring the solution showed more decomposition than the 2.0-mglml solution (Table 11).
to 100.0 ml. The solution was filtered (if necessary), the clear filtrate was Solutions Containing Sugar-The solutions of furosemide (1.0 mg/ml)
diluted with water to a concentration of 50 pg of furosemide/ml, and 40 in water containing various percentages of sucrose were stored a t 65" and
p1 was injected as already described. assayed after 3 weeks (Table 111).
Calculations-Since preliminary investigations indicated that the Aqueous Solutions in Phosphate Buffers, Acetic Acid, Hydrochloric
peak area of furosemide was directly related to the concentration, the Acid, and Sodium Hydroxide-These solutions were stored a t 65' for
results were calculated using the following equation: 30 days and assayed (Table 111).
Dibasic Ammonium Phosphate-An aqueous solution in dibasic
A,
- X 100 = % of label claim (Eq. 1) ammonium phosphate was stored for 24 hr a t room temperature and
AS
where A , is the peak area of the assay sample and A , is the peak area of
the standard sample. The results on commercial tablets and injection I I

1,
were 101.8 and 99.6% of the label claim, respectively. The standard de-
viation based on six injections of the standard solution was f1.37%.
A sample chromatogram is presented in Fig. 1A.
The investigations indicated (Fig. 1B) that the developed method was
stability indicating versus the USP method ( I ) , which was not. For ex-
ample, a 42-month-old sample in an 85% solution of sugar in water whose

4
v)

I
I
I z 4

0 5
MINUTES
Figure 2-Chromatograms of furosemide from dosage forms. A: i n 50%
aqueous solution of sorbitol after 3 weeks of storage at 65'. Peak 3 is from
furosemide, peak 4 is f r o m methylparaben, and peaks 1 and 2 are from
the solvent and decomposition products of furosemide. R: same as A
except that 20% ( v h ) alcohol was substituted for 20% water. C: i n
simple syrup after 3.5 years of storage at room temperature. Peak 6 is
MINUTES from furosemide; peaks 1 and 2 are from sugar, solvent, levulinic acid,
Figure 1-Chromatograms of furosemide from solutions. Peak 1 is from and 4-chloro-5-sulfamoylanthranilic acid; peaks 3and 5 are similar t o
furosemide Key: A , f r o m a standard solution; B, from a solution in 0.1 the peaks obtained from a solution of furfuryl alcohol i n water after 3
N HC'I after 30 days of storage at 6Fi'; and c, from a solution i n 0.1 N weeks of storage at room temperature; and peak 4 is from furfuryl al-
NaOH after 30 days of storage at 65'. cohol.

Vol. 67, No. 6, June 19781 009

Table 11-Assay Results o n Furosemide (with 8.0 ml of 4% Sodium Bicarbonate) in Conmercial Vehicle at Various Tem p e r a t ur e s

Storage Original Percent of Furosemide Retained after

Temperature Concentration, 7 14 21 28 30 42 49 56 60 87 90
(*I0) mg/ml Days Days Days Days Days Days Days Days Days Days Days
450 2 - 94.1b - 92.9 - 91.2 - - - 84.9 -
65' 2 - 91.4b - 91.7 - - 87.1 84.7 - 70.9* -
85" 2 92.5 87.5 81.9 68.gb - - - - - - -
24O 2 - - - - 93.8 - - - 88.OC - 87.5'
24' 1 - - - - 95.7 - - - 89.6 - 86.2
Lasix injection was used to make these solutions. These results may not be reliable due to minor mechanical problems in the integrator,but they do not affect the
conclusions of this study. These results may not be reliable due to the growth of fungus, but they do not affect the conclusions of this study.

Table 111-Percentages of Furosemide Retained in Various stability. The pH is apparently indirectly responsible for furosemide
Aqueous Solutions a f t e r 30 Days (21 Days f o r S u g a r Solutions) decomposition in sugar solutions (Table 111) since they changed consid-
of Storage a t 65" erably and were acidic (Table 111).The pH changes in sorbitol solutions
Solution Percent were not as wide as in sugar or glycerin (Tables 111and IV), so sorbitol
Containing Initial a H Final oH Retained solutions were studied further. Alcohol appears to have a stabilizing effect
(Table IV). Since more than 10% (v/v) alcohol is a reasonable upper limit
10%sugar 8.5 4.9 82.3 in pediatric dosage forms, alcoholic concentrations were kept to a mini-
20% sugar 8.1 4.3 60.5
30% sugar 8.2b 4.3 55.5 mum. Only two concentrations (10 and 20% v/v) were studied.
50%sugar 7.9 4.0 43.9 The samples prepared in the commercial vehicle (2.0 mg/ml) that
60%sugar 7.8 4.0 44.1 contained sodium benzoate as the preservative showed fungus growth
70% sugar 7.7 3.9 40.3 after about 30 days. This growth was expected since sodium benzoate is
85% sugar 7.6 3.9 40.3 not a good preservative in a basic medium. I t is the undissociated benzoic
Phosphate buffer 5.0 88.6 acid that prevents fungus growth. No such problem was noticed with
Phosphate buffer 4.8 76.8 parabens as the preservatives (0.005% methylparaben and 0.002% pro-
Phosphate buffer 4.65 72.1 pylparaben). These preservatives did not interfere with the assay (Figs.
Phosphate buffer 4.55 67.1
Acetate 4.2 11.5 2A and 2B).
Hydrochloric acid 1.2 0.0 The decomposition of furosemide (I) appears to follow Scheme I as
Sodium hvdroxide 13 99.5 reported by Kovar et al. (3). The presence of decomposition products in
Initial pH values were higher since the samples were made using furosemide
a 3.5-year-old sample (Fig. 2C) was confirmed by injecting solutions of
injection whose pH was 8.9. * The pH of simple syrup was approximately6.5. furfuryl alcohol (11)(0.1 m1/250 ml of water), levulinic acid (111) (1 m1/250

Table IV-Percentages of Furosemide Retained when Solution Contained Either Glycerin, Sorbitol, o r Sorbitol with Alcohol
Percent Retained after
Stored a t 21 30 61 120 155 182
Solution in PH (*lo) Days Days Days Days Days Days
50% glycerin 5.7 65" 78.9 (4.6)O - - - - -
70% sorhitol 6.8 65' 88.4 (5.7) - - - - -
50% sorbitol 8.56 65' 81.7 (4.9) - - - - -
50% sorbitol with 20% alcohol 8.56 65' 93.3 (4.7) - 77.4 (4.7) - - -
50% sorhitol with 20% alcohol 8.Sb 24' 99.5 (7.1) - 99.0 (6.7) - - -
50% sorhitol with 10%alcohol 8.Sh 65' 90.8 - 81.1 (4.8) 21.7 (4.0) 10.4 (3.9) 0 (3.9)
50% sorbitol with 10%alcohol 8.5b 24" 99.3 99.2 (6.1) - 97.5c (5.8) 99.2 (5.6) 99.3 (5.5)
0 Observed pH. h The pH was adjusted with 0.1 N %OH solution in water. C It appears that there is an error in this reading.

assayed to determine if furosemide, like etharrynic acids, undergoes fast ml of water), and 4-chloro-5-sulfamoylanthranilic acid (IV) (20 mg/100
decomposition in the presence of ammonium ions. No decomposition of ml of methanol). The solution of furfuryl alcohol in water decomposed
furosemide was noted. quite fast, even at room temperature. The decomposition of furfuryl al-
Aqueous Solutions in Glycerin and Sorbitol-These solutions were coho1 appears to be complex (Fig. 2C) and not a simple decomposition
to levulinic acid as reported by Kovar e t al. (3).

#
stored at 65' and at room temperature for varying periods and reassayed
(Table IV). Since the investigations indicated that sorbitol was better
than glycerin (from a stability point of view) and that alcohol had a sta-
bilizing
replacedeffect,
with alcohol
dosage were
formsalso
in sorbitol
investigated
in which
(Table
part
IV).
of the water was
N H ;"20+ C H , ~ I t
DISCUSSION H,NO,S
The method developed for the quantitative determination of furo- CI
semide appears to be stability indicating (Figs. 1A, 1B, and 2A-2C). It I
is accurate, precise (standard deviation based on six injections of the
standard solution was f1.37%),and sensitive. The peak area was directly
related to concentration (range tested was 0.5-3 fig of furosemide) with

Ir&?
a correlation coefficient of 0.99996. The USP method ( l ) ,based on UV
absorption, is not stability indicating since a 3.5-year-old sample in simple
syrup gave 82.1%versus only 26.1%by the developed method. In the two
commercial dosage forms (injection and tablets), no decomposition was
NH, BnroH I1
HY
noted in samples that were a t least 3 years old. Furosemide was very
unstable in acidic media and very stable in basic media (Table 111).
Sugar appears to have an adverse effect (Table 111) on furosemide

The study on etharrynic acid will be presented in a separate report.

H,NO,S

4 c1
IV
products

Scheme I
pp
CH ,COCH,CH,COOH
111

8 10 I Journal of Pharmaceutical Sciences

 Considering this discussion and the results, it is possible to formulate REFERENCES
a liquid dosage form with limited stability containing sorbitol, alcohol (1) ‘<TheUnited States Pharmacopeia,” 19th rev., ~~~k Publishing
(10-2@%v h ) , and preservatives (methylparahen and propylparahen). c0., Easton, p a , , 1975, pp. 213,214,
The pH should be adjusted to at least 8.5 with sodium hydroxide in water. (2) A. W. F ~B, ~ ~ i ~A. ,~ E,, cutler, clin,Chem,,
l ~i R.
~ D.~ ~ ~~l ~and
The shelflife of the dosage form may be extended for more than 1 year 20,152 (1974).
by adding an overage of about 5% of furosemide. (3) K. A. Kovar, G. P. Wojtovicz, and H. Auterhoff, Arch. Pharm., 307,
The data were not treated mathematically because the pH values were 657 (1974).
not constant (Table 111) because of the hydrolysis of sugar. The com-
mercial vehicle also contained high concentration of sugar. No attempt ACKNOWLEDGMENTS
was made to buffer the dosage forms heavily t o prevent p H changes. If
a biologically safe buffer can he used to maintain the pH value a t around T h e authors thank Hoechst-Roussel Pharmaceuticals Inc. for a gen-
8.5, it would certainly stabilize the dosage form further. erous supply of furosemide powder and injections.

Quantitation of Cocaine and Its Principal Metabolite,

Benzoylecgonine, by GLC-Mass Spectrometry Using
Stable Isotope Labeled Analogs as Internal Standards

SATYA P. JINDAL and PER VESTERGAARD

Received dune 27, 1977, from the Rockland Research Institute, Orangeburg, N Y 10962. Accepted for publication Octoher 12,1977.

benzoylecgonine, ecgonine, and possibly other metabolites (10) of cocaine

Abstract 0 A quantitative GLC-mass spectrometric assay was devel- are converted by the procedure (9) to the same derivative. Therefore, the
oped for the determination of cocaine and its principal metabolite, GLC-measured cocaine concentration cannot he the true concentration
henzoylecgonine, in human urine. The assay utilizes selective ion focusing of cocaine alone.
t o monitor in a GLC effluent the molecular ions of cocaine and henzo- Recently, a GLC procedure for cocaine and henzoylecgonine in the
ylecgonine generated by electron-impact ionization. Cocaine-d3 and urine of cocaine users was reported (11).The method involves methyl-
~)enzoylecgonine-da were the internal standards. The assay can measure ation of urinary benzoylecgonine t o cocaine by treatment with diazo-
2 ng of cocaine/ml and 5 ng of benzoylecgonine/ml with about 5%preci- methane, thereby measuring combined henzoylecgonine and cocaine.
sion. The curves relating the amounts of cocaine and henzoylecgonine Original cocaine was determined in a separate extract without the
added c p r s u s the amounts found over a large range of cocaine and hen- methylation step. Thus, benzoylecgonine could he determined by dif-
zoylecgonine concentrations were straight lines with nearly zero inter- ference. Cocaine is metabolized in the human body rather rapidly (12),
cepts and slopes of 0.98 + 0.01 and 0.97 f 0.01, respectively. The method benzoylecgonine being the major metabolite (13,14).Consequently, most
was used for the analysis of urinary cocaine and benzoylecgonine in co- urine samples of cocaine users contain trace amounts of unchanged drug
caine addicts. Assay specificity was confirmed by complete identity of and large concentrations of benzoylecgonine. Furthermore, cocaine and
the mass spectra of cocaine and henzoylecgonine with those of authentic henzoylecgonine have rather widely different and variable extraction
materials. efficiencies, and one ends up dealing with differences of rather large and
Keyphrases Cocaine-GLC-mass spectrometric analysis in human very small numbers. As a result, this assay is of limited usefulness.
urine Benzoylecgonine-GLC-mass spectrometric analysis in human This paper describes a GIA-mass spectrometric assay of cocaine and
urine GLC-mass spectrometry-analyses, cocaine and henzoylecgonine henzoylecgonine in human urine. Selective ion monitoring, the technique
in human urine Narcotics-cocaine and benzoylecgonine, GLC-mass built on combined GLC-mass spectrometry with selective focusing on
spectrometric analyses in human urine suitable fragments of the molecular ion (mass fragmentography) or the
molecular ion itself, is a well-estahlished technique, used widely in
pharmacology (15, 16). This technique was used to develop a sensitive
Cocaine abuse (1-5) has prompted considerable interest and specific assay for cocaine and benzoylecgonine in human urine with
in methods for the detection and quantitation of the parent site-specific deuterium-labeled cocaine and henzoylecgonine as internal
standards.
drug and its metabolites in biological fluids and tissues.
TLC and the enzyme multiplied immunoassay technique EXPERIMENTAL
(EMIT) are the most frequently employed screening Materials-Analytical grade cocaine hydrochloride’, henzoylecgon-
methods for the detection of cocaine in human urine. ine2, norcocaine hydrochloride?, methyl iodide-d j4, and N-ethyl-N-
However, these methods are inherently only semiquanti- nitro-N-nitrosoguanidine5 were used without further purification. All
tative at best. Although cocaine may be determined by solvents were analytical grade6. Silanized tubes7 (10 ml) with screw caps”
GLC, the amount of unchanged cocaine excreted in human were used for urine extraction; final solvent evaporation was performed
in 5-ml glass-stoppered centrifuge tuhes9. I’asteur pipets with hand-
urine ( 2 , 6) is generally below the method’s limits of de-
tection (7, 8).
’ Techrnan
Mallinckrodt Chemical Works, New York. N.Y.
Inc., Park Forest South. Ill.
BACKGROUND Norcocaine hydrochloride was a gift from Dr. Sally, Chemistry Department,
IJniversity of Wisconsin.
A rather sensitive GLC procedure (9) for cocaine is based on the re- International Chemical & Nuclear Corp., Irvine, Calif.
duction of cocaine with lithium aluminum hydride, derivatization of the Pfaltz & Bauer, Flushing, N.Y.
Fisher Scientific Co., Pittsburgh, Pa.
reduced product with pentafluoropropionic anhydride, and subsequent Kimble, Owens-Illinois,Toledo, Ohio.
detection by electron capture. This method is sensitive h u t unspecific, 8 Lined with Teflon (DuPont).
and it is not applicable to cocaine in biological specimens. Besides cocaine, Pyrex 8084.

Vol. 67,No. 6, June 19781 811