Cfinico Chimica Acta, 17.

5 (1988) 175-182
Elsevier
CCA 04196

A more simple, rapid and sensitive ~uo~~etri~

method for the det~~ati~~
of isoniazid
and acetylisoniazid in serum.
App~cation for acetylator ~~enot~i~~
Pinelopi

C. Ioannou

Laboratory ofAna!~ticalChem&tiy, Chemistry Department, Universi@ ofAthens, Athens (Greece)

(Received 24 September 1987; revision received 26 3jebrwq
accepted after revision 22 March 1988)
Ke,vwords: Isoniazid; Acetylisoaiazid; Fkorescent

1988;

analysis; Fluorescent ch~ate-ph~ot~~~

test

A simple, rapid and sensitive fluorimetric method for isoniazid determination in

serum is described. The method is based on the reaction of isoniazid with 2&yboxy-l-naph~~dehyde
in acidic medium in the presence of excess of scandium.
Wit-an
precision (Cv) was lS%, 1.0% and 1.2% at mean isoniazid concentration
in serum of 0.244, 1.94, and 25.9 mg/l respectively (n = 10); between-run precision
(CV) was 3.055, 2.6%, and 1.9% at mean isoniazid concentration of 0.265, 1.93, and
26.2 mg/l respectively (n = 10). The linearity of the method extended over the range
of O-300 mg isonia.zid/l serum. The detection limit (defined as three times the SD of
the mean blank) for the method is 0.008 mgfl of serum. Samples from 80
tuberculous patients treated with isoniazid were analysed by the proposed method
(y) and by the modified spectrofluorimetric method of Miceli et al (x). Linear
regression analysis of the results yielded the equation y = 0.98x + 0.05 fr = 0.986,
Sxy = 0.22).

Isoniazid (isonicotinic acid hydrazide, INK) is considered to be the most effective

of the commonly administered antituberculosis drugs. INH, like sulfamethazine, is

Correspondence to: P.C. Ioatxnou, Laboratory of Analytical Chemistry, Chemistry Department, Wniversity of Athens, 104 Solonos St., Athens 10680, Greece.
~9-~981/88/$0~.50

6 1988 Elsevier Science Publishers B.V. (Biomedical Division)

176

acetylated and inactivated in the liver by the acetyl-transferase enzyme [1,2]: the
rate of acetylation is genetically controlled.
Several methods have been used to measure INH and acetylisoniazid (AcINH) in
serum, including calorimetry [3], spectrophotometry
[4-61, fluorescence analysis
[4,7], and high performance liquid chromatography (HPLC) [8,9]. Calorimetric and
spectrophotometric methods are time-consuming, not sensitive enough to measure
the relatively lower concentrations of INH and AcINH in serum, and are not free
from interferences. On the other hand, analysis by HPLC is more specific, but has
the laborious steps of separation from biological material and derivatisation prior to
the injection of the sample into the chromatograph to improve the sensitivity of the
method. Fluorimetric methods for INH determination are the most sensitive,
especially the method of Scott and Wright [7], which was improved by Miceli et al
[lo], and was modified and adapted for AcINH determination by Olson et al [ll].
However, this method too is time-consuming (the procedure for INH determination
takes more than an hour), and laborious for use in clinical practice. For clinical
purposes a simple, rapid and sensitive method for INH and AcINH determination
would be advantageous.
Recently, I reported a new fluorimetric method for the determination of INH in
aqueous solutions [12]. The method was based on the reaction of INH with
2-hydroxy-l-naphthaldehyde
(HNA) in acidic medium in the presence of scandium
excess with the consequent formation of a strongly fluorescent complex (A, = 430
nm, A,,= 510 nm) between the hydrazone formed and scandium in weakly acidic
medium. For INH measurement two procedures were proposed, a kinetic procedure
(reaction time 1 min), which was less time-consuming with fewer manipulation
steps, and an equilibrium one (reaction time 10 min), which allowed the avoidance
of the precise timing control.
The present study represents an application of this method for the determination
of INH in serum using the equilibrium procedure. To avoid the interference from
serum components, removal of proteins by treatment with acetonitrile is required.
Concentration values are calculated against an aqueous standard curve, which is
similar in slope to curves constructed with data on serum supemates. The method is
adapted for the determination of AcINH in serum after gentle acidic hydrolysis to
INH, and for acetylator phenotyping by estimating the ratio of AcINH (expressed
as INH) to the apparent INH.
Materials and methods

Instrumentation
A Model 512 fluorescence spectrophotometer (Perkin-Elmer Corp., Norwalk, CT
USA) with a 150 W xenon lamp, equipped with a magnetic stirrer under the cell
holder, was used. The instrument settings were as follows: ratio mode, excitation
wavelength 430 nm, with a band width of 20 run; emission wavelength 510 run, with
a band width of 20 mn. A constant temperature of 25.0 o C in the 1 Ooo-cm sample
cell was maintained with a thermostated water bath. The sonicator used was from
Struers Co., Model Metason C., H 50-60. Finpipette microsyringes for transfer of
small sample and reagent volumes were used.

177

Reagents

All solutions were prepared in deionised, distilled water from reagent-grade

materials, unless otherwise stated.
HNA solution, 10.0 mmoE/I, in acetonitrile. Prepared from HNA obtained from
Fluka. HNA solution is stable at room temperature for several weeks.
Scandium sol~io~, 20.0 rnrn~~~l. Prepared from scandium oxide (specpure,
Johnson, Mattey and Co.) according to [12]. A 10.0 mmol,/l working solution with
pH 1.1 prepared by appropriate dilution with water and few drops of saturated
sodium hydroxide solution.
Stock buffer solution (0.1 mol/l sodium acetate, 10.0 g/l hydroxylammonium
chloride), pH 6.3. Working buffer solution prepared by mixing one volume of the
stock buffer solution with one volume of acetonitrile.
Standard INH solution, 1000 mg/l, in water, Prepared from INH obtained
from Sigma. Stored at 4C this solution is stable for several weeks. Working
solutions were prepared by appropriate dilution daily.
Standard AcINH solution, I~.0 mg/l.
Prepared according to the method of
Eidus at al [13].
Sodium hydroxide, I mol/l.
Hydrochloric acid 1 mol/l.
Pooled sera. A normal serum pool was prepared from blood drawn from healthy
volunteers and stored in 18 x 118 mm polypropylene tubes (Kartell, Binasco, MI,

USA) at 4OC.
Tuberculous patients samples. The patients were given orally a single dose of
300 mg INH, and blood specimens were obtained 4 h later. Blood was allowed to
clot and was centrifuged for 10 min at 1500 X g. Serum specimens for INH
dete~ation
can be stored at 4 o C for several weeks without deprote~tion.
For
AcINH determination the removal of proteins within 24 h after the serum were
collected is essential because of the instability of AcINH in serum [ll].
Metho&
Sample preparation.
To remove protein, 100 pl of serum was mixed with 200 ~1
of acetonitrile in a polypropylene test tube, stirred for 60 s on a vortex-type mixer,
and centrifuged for 5 min at 1500 x g.
INH measurement.
100 ~1 of the serum supernate was transferred to a test tube,
100 ~1 of scandium working solution added along with 50 ~1 of HNA solution. The
mixture was stirred for 10 s and agitated vigorously in an ultrasonic bath for 10 min.
2.00 ml of the working buffer solution was added, the mixture stirred and the
fluorescence intensity measured. A standard curve based on aqueous INH standards
was prepared by mixing one volume of the INH standard solutions with two
volumes of acetonitrile. The procedure for INH was followed, and the specimen
con~ntration was calculated from this curve.

178

Total hydrazides (INH plus AcINH) measurement.

100 ~1 of serum supemate
was transferred to a test tube, 20 pl of 1.0 mol/l hydrochloric acid added, the tube
capped (cap must be pierced), and the mixture heated in a water bath at 80C for
an hour. After cooling the mixture, 20 ~1 of 1.0 mol/l of sodium hydroxide was
added and the measurement performed as above.
Ana~tica~-recoue~
experiments.
200 ~1 of pooled serum were pipetted to each of
twelve polypropylene tubes, 20 ~1 of aqueous INH standards were added and mixed
thoroughly. After protein precipitation, INH measurements were performed threefold for each sample.
Results and discussion

The optimum experimental

conditions
of the HNAINH-hydrazone
and
HNAINH-SC chelate formation (optimum pH, reagents concentration, standing
time, ~terfer~ces,
etc.) have been studied [12]. According to the method described,
the fluorescence intensity of the HNAINH-SC chelate, which was measured after
the buffer addition, decreased by 15% during the first 2 min and then remained
stable. In the present study the nature of this phenomenon was examined in order to
eliminate this decrease and to simplify the measurements. It was found that this
decrease was due to the presence of Fe3+ traces in solutions, which caused the
quenching effect. Because of the non-interference from the Fe2+, the addition of
some reducing agents, such as ascorbic acid and hydroxylammonium chloride, was
examined. The best results (no decrease in fluorescence intensity) were obtained
after the addition of hydroxyl~o~~
chloride to the acetate buffer at a
~on~n~ation of 10.0 g/l and pH 6.3 in the final buffer solution. It was also found
that the pH of the a~e~t~hydroxyl~o~~
solution remained stable after
acetonitrile addition.
In order to choose the better technique for the measurements, I calculated the
mean slope for aqueous standard curves (relative fluorescence intensity, F vs INH
concentration, mg/l), which was found to be 6.44 F. (mg/l)- (SD = 0.03, n = S),
and the mean slope for standard addition curves with serum supemates after protein
precipitation, which was found to be 6.38 F. (mg/l)-1 (SD = 0.26, n = 20). The
mean slope in both cases correlated well showing no interference from the serum
matrix, so INH ~n~ntration
values were calculated against an aqueous standard
curve. Linearity of the method extended from O-300 mg/l using the recommended
procedure, and can be increased to the concentration above 300 mg/l by reducing
the supemate volume taken for the measurement. The detection limit (defined as
three times the SD of the mean blank) for the method is 0.008 mg/l of serum, or
0.005 mg/l using 200 ~1 of the serum supemate for the measurement. To determine
the within-run and between-run precision, serum pools containing different INH
concentration were measured. For within-run precision assessment, three samples,
after protein removing, were measured 10 times each. For between-run precision, I
performed protein precipitation 10 times for each of the samples and measured the
INH con~ntration
in triplicate for each. Table I shows the precision results for

179
TABLE I
Precision data
Between-nm

Within-run
a(mg/l)
sD (mg/l)
cv @)
n = 10,

0.244
0.004
1.6

1.94
0.02
1.0

25.9
0.3
1.2

0.265
0.008
3.0

1.93
0.05
2.6

26.2
0.5
1.9

throughout.

serum samples of different INH concentration. The CV values obtained confirm the
high reproducibi~ty of the proposed method. Analytical recovery experiments on
pooled serum samples, which were performed for the method evaluation, are shown
in Table II. Recovery ranged from 90.0-105.28 (mean 98.7%) for three INH
concentration ranges and indicated that using acetonitrile for protein precipitation,
INH could be obtained practically quantitatively in serum supernate. Comparison
data for INH dete~nation
on 80 patients samples, as assayed by the proposed
and by Miceli methods, yielded the following linear regression equation: y = a + bx
(a = 0.05, SD = 0.04; b = 0.98, SD = 0.02); the standard error of estimate = 0.22;
r = 0.986 (x-axis, reference method; y-axis, proposed method). The INH concentration in plasma varied from 0.20-5.9 mg/l. To adapt the proposed method for
AcINH determination and afterwards for phenotyping of patients, the hydrolysis of
AcINH on pooled serum was studied, and recovery experiments on pooled serum
spiked with AcINH and INH were performed. As it was found, the hydrolysis of
AcINH to INH in serum supemate in the presence of hydrochloric acid solution is
completed in an hour at 80C. Recovery for INH measured in aliquot A, total

TABLE II
Analytical recovery data of INH added to serum
INH added (mg,/l)
0.250
0.500
0.750
1.00

INH measured (mg/l)

0.242
0.464
0.789
1.01

2.50
5.00
7.50
10.0

2.30
5.18
7.48
10.1

20.0
30.0
50.0
60.0

18.0
31.5
46.8
62.0

INH recovery (%)

96.8
92.8
105.2
101.0
Mean 98.9
92.0
103.6
99.7
101.0
Mean 99.1
90.0
105.0
93.6
103.4
Mean 98.0

180
TABLE III
Analytical recovery data of INH and AcINH (expressed as INH) added to serum
INH (mg/l) a

AcINH (mg,/l) b

Added

Measured

2.0
4.0

2.0
3.8
5.9
8.1

6.0
8.0

Recovery
100.0
95.0
98.3
101.3
Mean

98.7

Added

Measured

3.0
6.0
9.0
12.0

2.8
6.2
8.4
11.2

Total (mg/l)
Recovery
93.3
103.3
93.3
93.3
Mean

Added

Measured

5.0
10.0
15.0
20.0

4.8
10.0
14.3
19.3

95.8

Recovery
96.0
100.0
95.3
96.5
Mean

97.0

Measured in aliquot A.
b Calculated as (total INH in aliquot B-MH in aliquot A).
Measured in aliquot B.

hydrazides measured in aliquot B, and AcINH calculated as (INH in aliquot B-INH

in aliquot A) ranged from 93.3-103.38 (Table III). For the classification of the
patients into slow or fast inactivators, the ratio of AcINH (expressed as INH) to the
apparent INH was estimated:
AcINH-Index

(total INH in aliquot B-INH in aliquot A) (mg/l)

(INH in aliquot A) (mg/l)

All the 80 samples were phenotyped by the proposed method, and according to
[8,14] 52 of them (65%) with an average AcINH-index of 0.71 (0.10-1.30) were
classified as slow and 28 of them (35%) with an average AcINH-index of 3.10
(1.70-11.30) as fast inactivators. Several drugs, such as rimfampicin, p-aminosalicylic acid, theophylline, sulfamerazine, sulfanilbenzamide, that may be used in
conjuction with INH, were tested to see if they interfere in the assay. When toxic
concentrations of each drug listed were added to standards of INH, no quenching,
interference, or increase in fluorescence intensity was found.
The proposed method is more simple and less laborious than the other fluorimetric methods, very sensitive, and permits the use of as little as 100 ~1 of serum for
both INH and AcINH measurements. The deproteinization with acetonitrile is
found to be more effective than by other methods, gives excellent recoveries and
eliminates the effect of serum matrix, especially metal ions, such as Fe3+ and Cu2+,
which cause quenching effect. The method is suitable for routine use in clinical
laboratories, and can be automated (kinetic procedure) using deproteinized samples
with various analysers, such as flow injection, air-segmented continuous flow, or
discrete analysers with a fluorimetric detection.
Acknowledgements

I thank Ms. Ana~ostopo~ou,

Mr. Marinis and his colleagues at the Microbiology Laboratory, Hospital Sotiria for Thorax Diseases {Athens, Greece) for
supplying the patients samples.

181

References
1 Evans DAP, White TA. Human acetylation
polymorphism.
J Lab Clin Med 1964;63:394-403.
2 Boxenbaum
HG, Riegelman
S. Determination
of isoniazid and metabolites
in biological
fluids. J
Pharmacol Sci 1974;63:1191-1197.
3 Hashmi
MH, Adil AS, Malik FR, Ajmal AI. Calorimetric
determination
of isonicotinic
acid
hydrazide. Mikrochim Acta 1969;4:772-777.
4 Lever M. Rapid fluorometric
or spectrophotometric
determination
of isoniazid.
Biochem Med
1972;6:65-71.
5 Eidus L, Hamanansingh
MT. A more sensitive spectrophotometric
method for determination
of
isoniazid in serum or plasma. Clin Chem 1971;17:492-494.
6 Eidus L, Hamilton
EJ. A new method for the determination
of N-acetyl isoniazid
in urine of
ambulatory
patients. Am Rev Resp Dis 1964;89:587-588.
7 Scott EM, Wright RC. Fluorometric
determination
of isonicotinic
acid hydrazide
in serum. J Lab
Clin Med 1967;70:355-360.
8 Saxena SJ, Stewart JT, Honigberg
IL, Washington
JG, Keene GR. Liquid chromatography
in
pharmaceutical
analysis. VIII: Determination
of isoniazid and acetyl derivative in plasma and urine
samples. J Pharmacol
Sci 1977;66:813-816.
9 Svensson JO, Muchtar A, Ericsson 0. Ion-pair high-performance
liquid chromatographic
determination of isoniazid and acetylisoniazid
in plasma and urine. Application
for acetylator phenotyping.
.I
Chromatogr
1985;341:193-197.
10 Miceli JN, Olson WA, Weber WW. An improved micro spectrofluorometric
assay for determining
isoniazid in serum. Biochem Med 1975;12:348-355.
11 Olson WA, Dayton
PG, Israili ZH, Pruitt AW. Spectrofluorometric
assay for isoniazid
and
acetylisoniazid
in plasma adapted to pediatric studies. Clin Chem 1977;23:745-748.
12 Ioannou PC. A simple and rapid fluorimetric
method for the microdetermination
of isonicotinic
acid
hydrazid. Talanta 1987;34:857-860.
13 Varughese
P, Hamilton
ET, Eidus L. Mass phenotyping
of isoniazid
inactivators
by automated
determination
of acetylisoniazid
in urine. Clin Chem 1974;20:639-641.
14 Eidus L, Ling GM, Hamanansingh
AMT. Isoniazid excrection in fast and slow inactivators
and its
practical aspect for phenotyping.
Arzneim Forsch 1971;11:1696-1699.