HPLC Analysis of Reduced and Oxidized Coenzyme Q10 in Human Plasma

Abstract

The percentage of reduced coenzyme Q(10) (CoQ(10)H(2)) in total coenzyme Q(10) (TQ(10)) is decreased in plasma of patients with prematurity, hyperlipidemia, and liver disease. CoQ(10)H(2) is, however, easily oxidized and difficult to measure, and therefore reliable quantification of plasma CoQ(10)H(2) is of clinical importance.
Venous blood was collected into evacuated tubes containing heparin, which were immediately placed on ice and promptly centrifuged at 4 degrees C. The plasma was harvested and stored in screw-top polypropylene tubes at -80 degrees C until analysis. After extraction with 1-propanol and centrifugation, the supernatant was injected directly into an HPLC system with coulometric detection.
The in-line reduction procedure permitted transformation of CoQ(10) into CoQ(10)H(2) and avoided artifactual oxidation of CoQ(10)H(2). The electrochemical reduction yielded 99% CoQ(10)H(2). Only 100 microL of plasma was required to simultaneously measure CoQ(10)H(2) and CoQ(10) over an analytical range of 10 microg/L to 4 mg/L. Intra- and interassay CVs for CoQ(10) in human plasma were 1.2-4.9% across this range. Analytical recoveries were 95.8-101.0%. The percentage of CoQ(10)H(2) in TQ(10) was approximately 96% in apparently healthy individuals. The method allowed analysis of up to 40 samples within an 8-h period.
This optimized method for CoQ(10)H(2) analysis provides rapid and precise results with the potential for high throughput. This method is specific and sufficiently sensitive for use in both clinical and research laboratories.

Full text

HPLC Analysis of Reduced and Oxidized
Coenzyme Q
10
in Human Plasma
Peter H. Tang,
1*
Michael V. Miles,
1
Antonius DeGrauw,
1
Andrew Hershey,
1
and
Amadeo Pesce
2
Background: The percentage of reduced coenzyme Q
10
(CoQ
10
H
2
) in total coenzyme Q
10
(TQ
10
) is decreased in
plasma of patients with prematurity, hyperlipidemia,
and liver disease. CoQ
10
H
2
is, however, easily oxidized
and difficult to measure, and therefore reliable quanti-
fication of plasma CoQ
10
H
2
is of clinical importance.
Methods: Venous blood was collected into evacuated
tubes containing heparin, which were immediately
placed on ice and promptly centrifuged at 4 °C. The
plasma was harvested and stored in screw-top polypro-
pylene tubes at ⴚ80 °C until analysis. After extraction
with 1-propanol and centrifugation, the supernatant was
injected directly into an HPLC system with coulometric
detection.
Results: The in-line reduction procedure permitted
transformation of CoQ
10
into CoQ
10
H
2
and avoided
artifactual oxidation of CoQ
10
H
2
. The electrochemical
reduction yielded 99% CoQ
10
H
2
. Only 100
␮
L of plasma
was required to simultaneously measure CoQ
10
H
2
and
CoQ
10
over an analytical range of 10
␮
g/L to 4 mg/L.
Intra- and interassay CVs for CoQ
10
in human plasma
were 1.2–4.9% across this range. Analytical recoveries
were 95.8 –101.0%. The percentage of CoQ
10
H
2
in TQ
10
was ⬃96% in apparently healthy individuals. The
method allowed analysis of up to 40 samples within an
8-h period.
Conclusions: This optimized method for CoQ
10
H
2
anal-
ysis provides rapid and precise results with the poten-
tial for high throughput. This method is specific and
sufficiently sensitive for use in both clinical and re-
search laboratories.
© 2001 American Association for Clinical Chemistry
Coenzyme Q
10
is an essential cofactor in the mitochon-
drial respiratory chain responsible for oxidative phos-
phorylation (1 ). Furthermore, coenzyme Q
10
has a pri-
mary function as an antioxidant and is carried mainly by
lipoproteins in the circulation (2). Approximately 60% of
coenzyme Q
10
is associated with LDL, 25% with HDL, and
15% with other lipoproteins (2). When LDL is subjected to
oxidative stress in vivo (3), the reduced form of CoQ
10
(CoQ
10
H
2
)
3
functions as an antioxidant. It has been pos-
tulated that CoQ
10
H
2
prevents lipid peroxidation in
plasma lipoproteins and biological membranes (4). The
antioxidative activity of CoQ
10
H
2
depends not only on its
concentration, but also on its redox status. Recent reports
(5–14) have suggested that the percentage of CoQ
10
H
2
in
total CoQ
10
(CoQ
10
H
2
:TQ
10
) may be lower in patients with
certain conditions, including Parkinson disease (5), pre-
maturity (6), hemodialysis (7 ), chronic active hepatitis
(8), liver cirrhosis (8 ), hepatocellular carcinoma (8), hy-
perlipidemia (9, 10 ), heart disease (11, 12),
␤
-thalassemia
(13), and DNA damage (14 ). Therefore, CoQ
10
H
2
may be
a useful marker of oxidative stress, and the measurement
and function of CoQ
10
H
2
are of considerable interest.
Several investigators have reported analytical tech-
niques for measurement of CoQ
10
H
2
(15–25). In these
publications, electrochemical (EC) detection was pre-
ferred for measurement of CoQ
10
H
2
because of its high
sensitivity. The EC reactions were measured at electrodes,
which detected the current produced by the reduction of
oxidized CoQ
10
(CoQ
10
) or by the oxidation of CoQ
10
H
2
(Fig. 1).
Investigation of CoQ
10
H
2
in clinical studies has been
hampered by instability during sample handling, storage,
and processing (15–21). According to several investiga-
tors (22–25 ), the concentration of CoQ
10
H
2
decreases
rapidly within 1 h after phlebotomy. At room tempera-
ture, it is oxidized at a rate of ⬃3 nmol/L per min in the
1
Division of Pediatric Neurology, The Children’s Hospital Medical Center,
3333 Burnet Ave., Cincinnati, OH 45229-3039.
2
Department of Pathology and Laboratory Medicine, College of Medicine,
University of Cincinnati, 231 Bethesda Ave., Cincinnati, OH 45267-0559.
*Address correspondence to this author at: Clinical Neuropharmacology
Laboratory, Division of Pediatric Neurology, The Children’s Hospital Medical
Center, 3333 Burnet Ave., Cincinnati, OH 45229-3039. Fax 513-636-6359; e-mail
Tangp0@chmcc.org.
Received August 8, 2000; accepted November 30, 2000.
3
Nonstandard abbreviations: CoQ
10
H
2
, reduced coenzyme Q
10
; CoQ
10
,
oxidized coenzyme Q
10
;TQ
10
, total coenzyme Q
10
; EC, electrochemical; CoQ
9
,
oxidized coenzyme Q
9
; and CoQ
9
H
2
, reduced coenzyme Q
9
.
Clinical Chemistry 47:2
256–265 (2001)
Enzymes and Protein
Markers
256

hexane extract of human plasma (23). Sample preparation
may have a profound effect on the redox status of CoQ
10
,
and the utmost care is required to ensure reliable quanti-
fication of CoQ
10
H
2
. Recently, investigators recom-
mended that plasma samples be individually thawed,
extracted, and analyzed immediately as a continuous
process to minimize CoQ
10
H
2
oxidation (22–24 ). This is
very impractical for analyzing even small numbers of
biological specimens. In earlier studies, biological fluid
samples were converted into either CoQ
10
by use of an
oxidizing reagent such as ferric chloride, or into CoQ
10
H
2
by a reducing agent such as sodium tetrahydroborate
(NaBH
4
) or sodium dithionite (Na
2
S
2
O
4
; Table 1). How-
ever, these methods are also inefficient, are susceptible to
preanalytical degradation, and have increased potential
for analytical error because of the lability of CoQ
10
H
2
.
We therefore developed a simple and rapid HPLC
procedure with coulometric detection for simultaneous
determination of CoQ
10
and CoQ
10
H
2
in human plasma.
Materials and Methods
materials
CoQ
10
and coenzyme Q
9
(CoQ
9
) were obtained from
Sigma. Methanol, ethanol, 1-propanol, 2-propanol, hex-
ane, sodium acetate, and glacial acetic acid were obtained
from Fisher Scientific. All chemicals were HPLC grade
and were used without further purification. Dade
®
immu-
noassay comprehensive tri-level controls were from Dade
International.
apparatus
The HPLC-EC system configuration is depicted in Fig. 2.
The system consists of an Model 582 Solvent Delivery
Module (ESA) equipped with a double plunger recipro-
cating pump, an AS3000 variable-loop autosampler
(Thermo Separation Products), an analytical column,
an ESA CouloChem II Model 5200A EC detector, and a
Dell Pentium II 350 Mz computer/controller with
ChromQuest software (Thermo Separation Products). The
system consisted of two cells (pre- and postcolumn) and
an analytical cell (Fig. 2). One carbon filter was placed
before the precolumn cell and another between the ana-
lytical column and the postcolumn cell. Both pre- and
postcolumn cells (E1 and E2) were coulometric electrodes
(ESA Model 5020). The postcolumn cells were configured
Fig. 1. Diagram depicting the EC reactivity of CoQ
10
H
2
and CoQ
10
.
Table 1. Comparison between current and previous studies on CoQ
10
H
2
analysis.
Ref.
Blood
sample
Sample
size,
␮
L
Extraction
solvent(s)
Precolumn
reduction
Internal
standard
HPLC run
time, min
CoQ
10
H
2
:TQ
10
,
a
%% Detection technique
15
Heparin 1000 Ethanol/hexane NaBH
4
CoQ
9
10 51.1 ⫾ 4.2 Amperometric detector
16
Heparin 200 Hexane Na
2
S
2
O
4
CoQ
9
14
b
NA
c
Amperometric detector
17
Heparin 300 1-Propanol NaBH
4
Diethoxy-
CoQ
10
11
b
NA Coupled-column/EC-UV
b
18
Heparin 200 Hexane NaBH
4
CoQ
9
16
b
87.0 ⫾ 1.0 Amperometric detector
19
EDTA 1000 Hexane NA CoQ
9
10 NA Column-switching/EC
20
Serum 100 Hexane NaBH
4
CoQ
9
15
b
65.8 ⫾ 4.2 Amperometric detector
21
Heparin 10 Hexane Na
2
S
2
O
4
CoQ
9
20
b
⬃95 Coulometric detector
22
EDTA 300 1-Propanol NaBH
4
None 13 94.3 ⫾ 0.7 Postcolumn valve
switching/EC
23
Heparin 50 Methanol/hexane NaBH
4
CoQ
9
22
b
95.6 ⫾ 1.6 Amperometric detector
24
Heparin
EDTA
50 Methanol/hexane NA CoQ
7
, CoQ
9
15–20 88.6 ⫾ 1.0 Coulometric detector
25
EDTA 100 Ethanol NaBH
4
None 13 ⬃93 Amperometric detector
Current
study
Heparin 100 1-Propanol EC CoQ
9
8 96.3 ⫾ 2.0 Coulometric detector
a
Values of CoQ
10
H
2
:TQ
10
are mean ⫾ SD.
b
Analytes in addition to CoQ
10
H
2
and CoQ
10
were included in this method.
c
NA, data not provided; UV, ultraviolet.
Clinical Chemistry 47, No. 2, 2001 257

in series as described by Edlund (17 ). The analytical cell
(ESA Model 5010) consisted of a series of two coulometric
electrodes and was connected in series to the postcolumn
cell; the first electrode (E3) was for reduction of CoQ
10
,
and the second electrode (E4) was for detection of
CoQ
10
H
2
.
The analytical column was a reversed-phase Mi-
crosorb-MV column (4.6 mm ⫻ 15 cm; 5
␮
m bead size;
Rainin). A reversed-phase C
18
guard column (4.6 ⫻ 10
mm; 5
␮
m bead size) was used to protect the analytical
column. The AS3000 injector was set at a needle height of
1.5 mm, and the injection volume was set at 20
␮
L for each
sample. The cooling temperature of the autosampler was
set at 0 °C. The mobile phase for the isocratic elution of
CoQ
10
was prepared as follows: sodium acetate trihydrate
(6.8 g), 15 mL of glacial acetic acid, and 15 mL of
2-propanol were added to 695 mL of methanol and 275
mL of hexane. The mobile phase was filtered through a 0.2
␮
m (47 mm diameter) nylon or analogous filter. The pH of
the mobile phase was ⬃6, and the flow rate was 1
mL/min.
preparation of calibrators
All sample preparation work was carried out under a dim
light to avoid photochemical decomposition of CoQ
10
and
CoQ
9
. To prepare a 5 mg/L working solution of CoQ
10
,
we dissolved 10 mg of CoQ
10
in 10 mL of hexane and
diluted this solution to 100 mL with 1-propanol. The
solution was thoroughly vortex-mixed until complete
dissolution. A working solution was then prepared by
dilution with 1-propanol to 5 mg/L. The concentration of
the working solution was then calculated by reading the
absorbance on a spectrophotometer (275 nm wavelength;
1-cm quartz cuvette), using a molar absorptivity (
⑀
)of
14 200. A series of calibration solutions was then prepared
with the appropriate volume of 1-propanol to final CoQ
10
concentrations of 10, 100, 500, 1000, 2000, and 4000
␮
g/L.
The low control was prepared by diluting pooled plasma
containing 0.45 mg/L CoQ
10
with distilled water to a final
concentration of 75
␮
g/L. The middle and high controls
were prepared by adding working solutions containing
1.2 and 3.0 mg/L CoQ
10
to pooled plasma samples to final
concentrations of 1.65 and 3.45 mg/L, respectively. The
calibrators and controls were stored in 1.8-mL polypro-
pylene tubes (Sarstedt) without addition of argon or
nitrogen at ⫺80 °C and used throughout the study. CoQ
9
was chosen as the internal standard. To prepare a CoQ
9
solution, we dissolved 2 mg of CoQ
9
in 100 mL of
1-propanol. The CoQ
9
solution was thoroughly vortex-
mixed until complete dissolution. A working solution of
CoQ
9
was then prepared by dilution with 1-propanol to 2
mg/L. All solutions were stored in 1.8-mL polypropylene
tubes at ⫺80 °C and used throughout the study.
preparation of plasma samples
Venous blood was collected into a Vacutainer
®
Tube
(Becton Dickinson) containing heparin as anticoagulant
and mixed by gentle inversion 5– 6 times. The Vacutainer
Tube was not opened to ambient air and was placed in ice
or kept refrigerated before processing. Blood samples
were processed within4hofcollection and centrifuged at
2000g for 10 min at 4 °C. Plasma was collected, placed in
a capped polypropylene tube, and immediately stored
without addition of argon or nitrogen at or below ⫺80 °C
until analysis.
liquid-liquid extraction
Under our experimental conditions, we optimized the
extracting procedure of Edlund (17 ) and compared the
efficiency of different mixtures of organic solvents for
liquid-liquid extraction of CoQ
10
and CoQ
9
from human
plasma. Quantitative recoveries (⬃100%) of these com-
pounds were obtained with two solvents: 1-propanol and
a mixture of ethanol-hexane (2:5 by volume). The 1-pro-
panol extraction procedure was used for subsequent
studies.
coulometric detection
The hydrodynamic voltammograms were obtained by
repeated injections into the HPLC system of a mixture of
CoQ
9
(1 mg/L) and CoQ
10
(4 mg/L) in water-1-propanol
(1:9 by volume). The detector potential was increased by
0.05 V in each subsequent run. Anodic currents for
CoQ
10
H
2
and cathodic currents for CoQ
10
reached maxi-
mum response at applied voltages of ⫹0.35 V and ⫺0.65
V, respectively. On the basis of the assessed hydrody-
namic voltammogram, the E2 cell potential was always
set at ⫹0.7 V to oxidize any electrochemically active
eluates. The E3 and E4 cell potentials were set at ⫺0.65 V
and ⫹0.45 V, respectively. When the E1 cell potential was
set to ⫺0.7 V for the precolumn reduction mode, all CoQ
10
was reduced to CoQ
10
H
2
before column separation. Total
CoQ
10
H
2
was then measured, and a calibration curve of
Fig. 2. Schematic diagram of the HPLC-EC system.
This system can be operated in three different modes, as described in the text.
258 Tang et al.: HPLC Analysis of Coenzyme Q
10

CoQ
10
H
2
was established. For the precolumn oxidation
mode, the E1 cell potential was set at ⫹0.7 V. All CoQ
10
H
2
was oxidized to CoQ
10
, and TQ
10
was measured. A
calibration curve of CoQ
10
was thus obtained. For simul-
taneous determination of CoQ
10
H
2
and CoQ
10
, the E1 cell
was turned off. Because no current flowed into the cell, all
compounds remained at their original state.
assessment of possible interfering substances
To explore possible sources of interference, we processed
several lyophilized products from human blood, highly
purified chemicals, and biochemicals (included in the
Dade high control) according to the developed method.
Briefly, Dade high control, which contains 45 drugs and
endogenous substances, was supplemented with 20 com-
monly prescribed drugs at concentrations exceeding clin-
ically relevant values (Table 2); 100-
␮
L aliquots of the
supplemented control were then placed in 1.8-mL capped
polypropylene tubes, processed, and analyzed.
To assess the possible interference of endogenous
CoQ
9
or other substances in patient plasma samples,
blood samples were collected from 25 patients (ages 1–18
years) in the neurology clinic at the Children’s Hospital
Medical Center, Cincinnati, OH. These patients were
diagnosed with a variety of neurological disorders. An
additional 25 specimens were obtained from apparently
healthy individuals (ages 0.2– 65 years). Informed consent
was obtained from all adults and from the parents (or
guardians) of all minors.
sample analysis
We simultaneously processed samples in batches of 20,
which is the capacity of our centrifugation instrument.
Each frozen sample was thawed at room temperature,
and then a 100-
␮
L aliquot of the sample was placed in a
1.8-mL capped polypropylene tube containing 50
␮
Lof
internal standard solution. All tubes were kept in an ice
Table 2. Lyophilized products from human blood, highly purified chemicals and biochemicals, and commonly prescribed
drugs tested for potential interference with HPLC-EC.
Material Concentration Material Concentration
Acetaminophen 155 mg/L Human growth hormone (hGH) 10.5
␮
g/L
n
-Acetylprocainamide
a
25 mg/L Human luteinizing hormone (hLH) 85 IU/L
Aldosterone 468 ng/L IgE 340 kIU/L
␣
-Fetoprotein (AFP) 332
␮
g/L Imipramine
a
500
␮
g/L
Amikacin 30 mg/L Insulin 97 mIU/L
Amitriptyline
a
500
␮
g/L Lamotrigine
a
10 mg/L
Carbamazepine 15 mg/L Lidocaine 8 mg/L
10,11-Carbamazepine epoxide
a
5 mg/L Lithium 2.8 mEq/L
Carcinoembryonic antigen (CEA) 110
␮
g/L Lorazepam
a
5 mg/L
Chlordiazepoxide
a
5 mg/L Methsuximide
a
5 mg/L
Clonazepam
a
100
␮
g/L Oxazepam
a
5 mg/L
Cortisol 320
␮
g/L Phenobarbital 53 mg/L
Desipramine
a
500
␮
g/L Phenytoin 30 mg/L
n
-Desmethyldiazepam
a
5 mg/L Primidone 20 mg/L
n
-Desmethylmethsuximide
a
5 mg/L Procainamide 15 mg/L
Diazepam
a
5 mg/L Progesterone 18
␮
g/L
Digoxin 3.5
␮
g/L Prolactin 80
␮
g/L
trans
-10,11-Dihydroxycarbamazepine
a
5 mg/L Prostate-specific antigen (PSA) 38
␮
g/L
Disopyramide 6 mg/L Prostatic acid phosphatase (PAP) 53
␮
g/L
Doxepin
a
500
␮
g/L Quinidine 6 mg/L
Estradiol 1
␮
g/L Salicylate 570 mg/L
Estriol 267
␮
g/L Testosterone 9
␮
g/L
Ethosuximide 120 mg/L Theophylline 26 mg/L
2-Ethyl-2-phenylmalonamide
a
5 mg/L Thyroid uptake/T
3
uptake 0.35 TU
Felbamate
a
250 mg/L Thyroxine (total T
4
) 150
␮
g/L
Ferritin 670
␮
g/L Tobramycin 9 mg/L
Folate 13
␮
g/L Triazolam
a
5 mg/L
Follicle-stimulating hormone (FSH) 70 IU/L Triiodothyronine (total T
3
)4
␮
g/L
Free T
3
20 ng/L Thyroid-stimulating hormone (TSH) 37 mIU/L
Free T
4
50 ng/L Valproic acid 133 mg/L
Gentamicin 7.5 mg/L Vancomycin 73 mg/L
5-(4-Hydroxyphenyl)-5-phenylhydantoin
a
10 mg/L Vitamin B
12
735 ng/L
Human chorionic gonadotropin (hCG) 400 IU/L
a
Drugs added to the commercial Dade姞 immunoassay control prior to extraction.
Clinical Chemistry 47, No. 2, 2001 259

bath. The sample was then mixed with 850
␮
L of cold
1-propanol. All tubes were vortex-mixed for 2 min on a
mechanical vortex-type mixer and centrifuged for 10 min
at 21 000 g and 0 °C. The resulting supernatant was
separated from the precipitate and transferred to a glass
autosampler vial. Sample vials were immediately placed
in the autosampler tray at 0 °C. A batch of 20 samples was
analyzed immediately in a single run sequence. A 20-
␮
L
aliquot of 1-propanol extract from a vial was injected
immediately into an automated HPLC. Peak height and
area measurements for each injection were obtained by
the ChromQuest software. The CoQ
10
:CoQ
9
peak-height
ratios were used (peak area was optional) to obtain
least-squares linear regression equations, which were
used to calculate the CoQ
10
concentrations of the frozen
control samples and patient samples. If an error occurred
in the system, the sample vials were resealed and imme-
diately restored at ⫺80 °C or below for further investiga-
tion. A single technician could complete the analysis of a
20-specimen batch routinely within 4 h.
preliminary reference interval data
To evaluate CoQ
10
H
2
:TQ
10
reference intervals, blood sam-
ples were obtained from 25 apparently healthy individu-
als (5 males and 20 females; age range, 12– 64 years) after
obtaining their consent. Individuals were carefully
screened and excluded if taking any medication chroni-
cally, had any history of acute or chronic illness, or were
taking any form of coenzyme Q
10
as a supplement.
Results
efficiency of reduction cell
CoQ
10
was converted to CoQ
10
H
2
electrochemically by the
reduction cell. The efficiency of conversion was measured
by comparing the peak heights of CoQ
10
H
2
and CoQ
10
per
injected amount of CoQ
10
. A series of solutions containing
0.01, 0.1, 1.0, 1.5, 2.0, 3.0, and 4.0
␮
g/L of CoQ
10
in
water/1-propanol (1:9 by volume) were injected in dupli-
cate into the HPLC system. Quantitative conversion rates
of 99.4% ⫾ 0.5% were obtained.
chromatographic analysis
As seen in Fig. 3, CoQ
10
and CoQ
10
H
2
could be measured
in the same HPLC run. CoQ
9
and CoQ
10
eluted at ⬃5.5
and ⬃6.9 min, respectively (Fig. 3A). Two peaks were
observed for the reduced form (CoQ
9
H
2
)ofCoQ
9
(inter-
nal standard) and CoQ
10
H
2
at ⬃3.6 and ⬃4.1 min, respec-
tively (Fig. 3B).
calibration curves and linearity
Calibration curves for CoQ
10
H
2
and CoQ
10
are shown in
Fig. 4. An excellent linear relationship was observed
between the peak-height ratios of each compound vs
CoQ
9
over a wide concentration range from 10
␮
g/L to 4
mg/L. The regression equations were: y ⫽ 1.151x ⫹ 0.003
(r
2
⫽ 0.999) for CoQ
10
H
2
; and y ⫽ 0.846x ⫹ 0.001 (r
2
⫽
0.998) for CoQ
10
. The detection limits of CoQ
10
H
2
and
CoQ
10
were ⬃5
␮
g/L (signal-to-noise ratio ⫽ 3).
extraction efficiency
Under our experimental conditions, quantitative recover-
ies of CoQ
10
and CoQ
9
for the current method using
1-propanol were compared with previously published
extraction methods (Table 1). With the 1-propanol
method, the mean recoveries were 99% ⫾ 3% for CoQ
10
and 100% ⫾ 2% for CoQ
9
(n ⫽ 6). Comparison with other
extraction solvents (n ⫽ 6 replicates each) produced the
following mean recoveries: 2-propanol, 89% ⫾ 5%; etha-
nol, 88% ⫾ 4%; n-butanol, 85% ⫾ 5%; acetone, 71% ⫾ 8%;
methanol-hexane (0.2:2.5 by volume), 64% ⫾ 10%; hexane,
52% ⫾ 9%; acetonitrile, 19% ⫾ 11%; and methanol, 19% ⫾
10%.
stability of CoQ
10
H
2
in stored whole blood
Venous blood was collected from five healthy adults in
tubes containing sodium heparinate. Blood samples,
Fig. 3. HPLC chromatograms of CoQ
10
and CoQ
9
calibrators.
(
A
), chromatogram showing two oxidation peaks for CoQ
10
(8 ng on column) and
CoQ
9
(2 ng) calibrators at ⬃5.5 and ⬃6.9 min, respectively. (
B
), chromatogram
showing two oxidation peaks for CoQ
9
H
2
and CoQ
10
H
2
calibrators at ⬃3.6 and
⬃4.1 min, respectively. Precolumn reduction was performed to transform CoQ
10
and CoQ
9
to CoQ
10
H
2
and CoQ
9
H
2
, respectively.
260 Tang et al.: HPLC Analysis of Coenzyme Q
10

which were kept on ice or in refrigerated at 4 °C, were
processed identically at hourly intervals up to 8 h after
collection. Plasma from each blood specimen was sepa-
rated and frozen at ⫺80 °C until analysis. The results
showed that CoQ
10
H
2
in whole blood stored at 4 °C was
stable for at least8hwithaCV⬍5%. The mean (SD) ratio
of CoQ
10
H
2
:TQ
10
in 25 heparinized whole blood speci-
mens was 95.3% (⫾ 1.8%) 8 h after blood collection (4 °C).
On the basis of these findings, we recommend that blood
for CoQ
10
H
2
analysis be refrigerated to ensure sample
stability for up to 8 h after collection.
precision and accuracy
The analytical recoveries of CoQ
10
in human plasma
controls are shown in Table 3. The inter- and intraday
assay CVs were ⬍5% over four concentrations of CoQ
10
.
Because CoQ
10
H
2
is oxidized rapidly, no control tests for
CoQ
10
H
2
were performed. To verify the reproducibility of
the CoQ
10
H
2
analysis, human plasma samples from 10
healthy individuals were examined (5 replicates each;
Table 4). The reproducibility of the analysis is presented
in Table 4. The CVs for the ratio of CoQ
10
H
2
:TQ
10
were
ⱕ1.0%, which shows the excellent reproducibility of anal-
ysis.
interference studies
Testing of the supplemented Dade control indicated that
a few unidentified electroactive compounds and sub-
stances (Table 2) eluted from the column within the first
3.5 min (data not shown). These compounds and sub-
stances may have been more hydrophilic than CoQ
10
H
2
because an organic solvent-based mobile phase was used.
Only one unknown compound eluted at ⬃6.9 min, which
corresponded exactly with the CoQ
10
elution time. Be-
cause the Dade control is plasma-based, this peak most
likely was residual CoQ
10
in this control. None of the
lyophilized products of human blood, highly purified
chemicals, biochemicals, and medications added to the
Dade High Control produced interference in the analysis.
To assess the possible interference of endogenous
CoQ
9
with that added as internal standard, 50 plasma
samples, including 25 from apparently healthy individu-
als and 25 from patients, were extracted without adding
the internal standard, CoQ
9
. Only one plasma sample,
from a patient with a rare glycogen storage disease (type
I), was found to have ⬃25
␮
g/L CoQ
9
, which corre-
sponded to ⬃2.5% of the internal standard concentration.
Measurable CoQ
9
was not detected in the plasma samples
from the remaining 25 healthy individuals or 24 patients.
On the basis of these findings, we conclude that interfer-
ence from endogenous CoQ
9
is very unlikely to cause
significant analytical error with our method, and thus is a
very suitable internal standard (Fig. 5).
preliminary reference interval results
To provide preliminary data for establishing the reference
interval for the ratio of CoQ
10
H
2
:TQ
10
, plasma specimens
were collected from 25 apparently healthy individuals.
The CoQ
10
H
2
:TQ
10
ratio was 96.3% ⫾ 2.0% (mean ⫾ SD).
The mean plasma concentrations of CoQ
10
H
2
and TQ
10
were 803 (⫾ 264) and 835 (⫾ 276)
␮
g/L, respectively.
Discussion
Various HPLC-EC methods have been described in the
past that attempted to measure CoQ
10
and CoQ
10
H
2
(Table 1). Because CoQ
10
is somewhat insensitive to EC
detection, in-line postcolumn reduction of CoQ
10
to
CoQ
10
H
2
by a coulometric method (21, 22, 24) or a reduc-
tion column (23 ) has been reported recently, which allows
for the sensitive detection of CoQ
10
by an EC electrode.
Historically, these methods used chemical reagents to
obtain CoQ
10
H
2
by reducing the commercially available
CoQ
10
. This process typically involves the addition of an
excess amount of the reducing reagent to convert CoQ
10
into CoQ
10
H
2
and to maintain the reduced form. Addi-
tional care is also required to avoid oxidation by ambient
oxygen, usually by storage under argon or nitrogen.
Fig. 4. Calibration curves for CoQ
10
(E) and CoQ
10
H
2
(䡺).
CoQ
10
calibrators were dissolved in 1-propanol.
Table 3. Precision results for CoQ
10
analysis.
Intended concentration,
mg/L
Measured
concentration
(mean ⴞ SD), mg/L CV, %%
Mean
recovery,
%%
Within-day precision
(n ⫽ 6)
0.075
a
0.074 ⫾ 0.003 3.6 99.3
0.45 0.431 ⫾ 0.011 2.7 95.8
1.2
b
1.171 ⫾ 0.014 1.2 97.6
3
c
3.010 ⫾ 0.095 1.7 100.0
Day-to-day precision
(n ⫽ 30, 5 days)
0.075
a
0.076 ⫾ 0.004 4.9 100.7
0.45 0.437 ⫾ 0.019 4.3 97.1
1.2
b
1.174 ⫾ 0.028 2.4 97.8
3
c
3.027 ⫾ 0.070 2.0 101.0
a
Diluted from pooled plasma containing 0.45 mg/L CoQ
10
.
b
Pooled plasma was fortified with 1.2 mg/L CoQ
10
to a final concentration of
1.65 mg/L.
c
Pooled plasma was fortified with 3 mg/L CoQ
10
to a final concentration of
3.45 mg/L.
Clinical Chemistry 47, No. 2, 2001 261

Preparation of calibrators and controls is particularly
problematic to protect CoQ
10
H
2
from oxidation. The cur-
rent method uses in-line precolumn EC reduction to
convert CoQ
10
into CoQ
10
H
2
and avoids the artifactual
oxidation that frequently occurs during the chemical
process to produce CoQ
10
H
2
. Coulometric detectors effi-
ciently convert ⬃99% of CoQ
10
to CoQ
10
H
2
. This improve-
ment is important because it dispenses with the need for
a reducing agent and additional sample clean-up steps.
The first study to substantially improve earlier meth-
ods was reported by Grossi et al. (19), who introduced a
precolumn oxidation cell for the CoQ
10
H
2
study. How-
ever, their quantitative measurement of CoQ
10
H
2
was
unsuccessful (Table 1).
Finckh et al. (21) developed a micromethod for simul-
taneous measurement of several lipophilic antioxidants
using HPLC with coulometric EC detection (Table 1).
Postcolumn EC detectors in the reduction-reduction-oxi-
dation mode, as described by Grossi et al. (19), were used.
According to their procedure, 5 or 10
␮
L of sample was
extracted with ethanol,
␤
-hydroxytoluene, and hexane.
After centrifugation, the hexane phase was evaporated to
dryness under a stream of argon and redissolved in a
mixture of ethanol and methanol. Poor recoveries of 54%
⫾ 37% and 76% ⫾ 36% were reported for CoQ
10
H
2
and
CoQ
10
, respectively, without internal standardization by
CoQ
9
H
2
and CoQ
9
. To correct this problem, the authors
added internal standardization with CoQ
9
H
2
and CoQ
9
.
This improved the accuracy and precision, i.e., recoveries
were 105% ⫾ 21% and 97% ⫾ 11%, respectively, for
CoQ
10
H
2
and CoQ
10
. However, the imprecision of the
CoQ
10
H
2
analysis was excessive. The instability of hex-
ane-extracted CoQ
10
H
2
after drying has been reported by
other investigators (23 ) and may contribute to this prob-
lem. Again, it should be noted that this procedure is
relatively tedious and complex.
Lagendijk et al. (22) also reported a rapid HPLC-EC
procedure for the determination of CoQ
10
H
2
and CoQ
10
in
1-propanol extracts (Table 1). They used the postcolumn
EC electrodes in the oxidation-reduction-oxidation mode
as described by Edlund (17). Their extraction and analysis
procedures without internal standardization were also
used to obtain the ratio between CoQ
10
H
2
and CoQ
10
.To
prevent a coulometric overload with an 80-
␮
L injection of
sample, a sophisticated switching valve was used to
ensure that only the compounds of interest were chan-
neled through the coulometric cells. Although they used a
1-propanol extraction similar to the one used in the
current method, their sample and solvent volumes (300
␮
L and 1 mL, respectively) were much greater than the
current method (100 and 900
␮
L, respectively), and their
Table 4. Reproducibility of the analysis of coenzyme Q
10
in human plasma.
Sample
Mean ⴞ SD (CV)
CoQ
10
H
2
, mg/L
(n ⴝ 5)
TQ
10
, mg/L
(n ⴝ 5)
CoQ
10
H
2
:TQ
10
ratio (%)
(n ⴝ 5)
1 1.338 ⫾ 0.029 (2.2%) 1.383 ⫾ 0.033 (2.4%) 96.8 ⫾ 0.8 (0.8%)
2 0.553 ⫾ 0.020 (3.6%) 0.573 ⫾ 0.021 (3.7%) 96.5 ⫾ 0.2 (0.2%)
3 0.640 ⫾ 0.002 (0.4%) 0.661 ⫾ 0.005 (0.8%) 96.9 ⫾ 0.5 (0.5%)
4 0.853 ⫾ 0.024 (2.8%) 0.906 ⫾ 0.031 (3.4%) 94.2 ⫾ 0.9 (1.0%)
5 0.715 ⫾ 0.027 (3.8%) 0.750 ⫾ 0.027 (3.7%) 95.3 ⫾ 0.8 (0.9%)
6 0.690 ⫾ 0.006 (0.8%) 0.721 ⫾ 0.010 (1.4%) 95.7 ⫾ 0.6 (0.7%)
7 0.856 ⫾ 0.011 (1.2%) 0.880 ⫾ 0.017 (1.9%) 97.3 ⫾ 0.8 (0.8%)
8 0.696 ⫾ 0.022 (3.2%) 0.726 ⫾ 0.025 (3.4%) 95.9 ⫾ 0.5 (0.5%)
9 0.827 ⫾ 0.018 (2.2%) 0.880 ⫾ 0.018 (2.0%) 93.9 ⫾ 0.3 (0.4%)
10 0.555 ⫾ 0.018 (3.2%) 0.586 ⫾ 0.017 (2.9%) 94.7 ⫾ 0.7 (0.8%)
Fig. 5. Chromatograms of extracts of a patient’s plasma containing
1.693 mg/L CoQ
10
H
2
and 43
␮
g/L CoQ
10
with (
B
) or without (
A
) CoQ
9
internal standard (2 ng on column).
262 Tang et al.: HPLC Analysis of Coenzyme Q
10

injection volume was fourfold increased (80
␮
Lvs20
␮
L).
In addition, the current method uses in-line precolumn
reduction, an autosampler, and cooling of the samples to
0 °C. Although the method used by Lagendijk et al. (22)
may accurately measure the CoQ
10
H
2
:CoQ
10
ratio
(⬃16.7:1), it may also be prone to analytical variation
because it does not use internal standard for quantifying
CoQ
10
H
2
and CoQ
10
. According to the authors’ recom-
mendations for reliable results, the time span from collec-
tion to analysis must be within 15 min. The instability of
CoQ
10
H
2
limits the practical application of their method
because only 8 –10 samples can be analyzed per day.
Yamashita and Yamamoto (23 ) reported a HPLC-EC
procedure using single extraction with methanol-hexane
(1:2 by volume; Table 1). To prevent the air oxidation of
CoQ
10
H
2
, they incorporated an immediate and direct
injection step into their procedure. Their results clearly
indicated that the hexane extract should be analyzed
immediately after extraction, and the analysis of one
sample at a time was emphasized. In addition, Finckh et
al. (21) and Wang et al. (25) reported that hexane is not an
efficient extraction solvent. Our data (unpublished) also
indicate the poor recovery of CoQ
10
H
2
(52–64%) with the
use of hexane.
Kaikkonen et al. (24) reported a method similar to the
one described by Finckh et al. (21 ) for measuring
CoQ
10
H
2
, but their results indicated a lower mean (⬃88%)
and broader range for the CoQ
10
H
2
:TQ
10
ratio (80.9 –
90.9%) than the current study (mean, 96.3% ⫾ 2%; Table
1). Their method also used a complex sample preparation
procedure and evaporation under nitrogen, which may
explain their decreased recovery of CoQ
10
H
2
: The long
pretreatment, extraction, and evaporation procedures re-
quired for their method may have allowed the oxidation
of a significant portion of CoQ
10
H
2
. Their method was
very tedious and slow, and according to their own de-
scription was capable of analyzing only one sample at a
time (24 ).
Wang et al. (25) recently reported a gradient HPLC
method with automated precolumn reduction to assess
CoQ
10
H
2
and TQ
10
concentrations in plasma. Their
method uses chemical reduction of CoQ
10
H
2
. As a result,
each clinical specimen requires duplicate injections to
complete the analysis of CoQ
10
H
2
and TQ
10
. In addition,
CoQ
10
H
2
and TQ
10
determinations must be performed
before and after each sample is mixed with reducing
agent before HPLC analysis. Because the reducing reagent
is very unstable, a fresh and adequate amount of reducing
reagent must be prepared every three samples. Previous
experience by the current investigators found that excess
reducing agent may overload the EC electrode and
shorten the life-span of the detector cells (unpublished
data). Additionally, the method of Wang et al. (25 ) may be
prone to analytical variation because they do not use an
internal standard. According to the authors, their method
is also limited to the analysis of a maximum of one sample
per hour and requires continuous effort by a technician.
Although the current procedure requires 100
␮
Lof
plasma, the sample size could be further reduced to 25 or
50
␮
L depending on the detection of trace amounts of
CoQ
10
. Because CoQ
10
H
2
and CoQ
10
are measured simul-
taneously, the total analysis time is substantially shorter
than those for other methods. The method described
herein makes it possible to analyze up to 40 samples
within an 8-h period. Although other, longer methods
have included additional analytes (16 –18, 20, 21, 23 ), the
current method has been optimized to measure CoQ
10
H
2
and CoQ
10
as rapidly, simply, accurately, and precisely as
possible. Tables 3 and 4 summarize the excellent repro-
ducibility of the analysis. The individual CVs for
CoQ
10
H
2
,TQ
10
, and the CoQ
10
H
2
:TQ
10
ratio were ⱕ3.8%,
ⱕ3.7%, and ⱕ1.0%, respectively (Table 4). The current
method uses single extraction with 1-propanol as solvent
to disrupt lipoproteins and efficiently solubilize CoQ
10
H
2
and CoQ
10
. This eliminates the necessity of the repeated
extraction procedures that frequently were required by
earlier procedures in which mixtures of either methanol
or ethanol and hexane were used. In contrast to earlier
methods, no evaporation step and no additional cleanup
of the 1-propanol extract are needed. Although ⬎1000
samples have been injected into the current system over
the previous 6 months, the in-line filters have been
replaced only once. The current system also avoids the
need for complex system configurations, such as coupled
columns with column-switching valves and postcolumn
two-way valves.
Some controversy exists concerning the use of CoQ
9
as
an internal standard for CoQ
10
H
2
analysis. Evidence of
endogenous CoQ
9
in some individuals was cited by some
investigators (26 ), but not others (27 ). Our current results
and considerable experience indicate that CoQ
9
is rarely
found in measurable quantities in human plasma and
thus is a suitable internal standard for this procedure.
The ranges for the percentage of CoQ
10
H
2
in TQ
10
from
previous reports are quite variable (Table 1). The percent-
ages of CoQ
10
H
2
in TQ
10
reported recently, i.e., ⬃95%
(21), ⬃94% (22), ⬃96% (23 ), and ⬃93% (25 ), agree well
with the results of the current study (⬃96%). Other
methods may accurately measure CoQ
10
H
2
; however,
they generally are more labor-intensive and more prone
to error than the current method.
Accurate determination of CoQ
10
H
2
makes it a possible
marker for assessing the presence of oxidative stress in
many pathologic states. Although significant differences
in the plasma CoQ
10
H
2
:TQ
10
ratio between controls and
patients with atherosclerosis, coronary artery disease, and
Alzheimer disease have not been observed by some
investigators (10, 28), other researchers have reported
decreased CoQ
10
H
2
concentrations associated with certain
disease processes. Hara et al. (6 ) suggested that the
CoQ
10
H
2
:TQ
10
ratio is a good marker of oxidative stress in
infants with asphyxia (6 ). Hemodialysis patients have
also been found to have significantly lower concentrations
of plasma CoQ
10
H
2
than healthy controls (7 ). According
Clinical Chemistry 47, No. 2, 2001 263

to one report (7 ), a single hemodialysis session causes a
30% decrease in mean plasma CoQ
10
H
2
concentrations.
Plasma CoQ
10
H
2
was also found to be significantly lower
in hyperlipidemic patients and in patients with liver
disease (10 ). In 64 patients with chronic active hepatitis,
liver cirrhosis, and hepatocellular carcinoma, significantly
increased CoQ
10
and decreased CoQ
10
H
2
were observed
(8). Paloma¨kietal.(12) observed that lovastatin treatment
diminishes the CoQ
10
H
2
concentration in the LDL of
hypercholesterolemic patients with coronary heart dis-
ease. There are also concerns that patients could experi-
ence deleterious effects as a result of long-term therapy
with hydroxymethylglutaryl-CoA reductase inhibitors or
“statin” therapy. Monitoring of the effects of statin ther-
apy on CoQ
10
H
2
may be useful for diagnosing CoQ
10
H
2
deficiency in many patient populations. These are but a
few of a growing numbers of studies that suggest that
CoQ
10
H
2
deficiency may be related to pathophysiologic
mechanisms.
Recent studies have reported new findings related to
CoQ
10
H
2
that may lead to a better understanding of the
cellular function of CoQ
10
H
2
. One study in patients with
␤
-thalassemia showed that severely depleted CoQ
10
H
2
concentrations (⫺62.5%) are associated with increased
plasma concentrations of lipoperoxidation byproducts
and urinary concentrations of catecholamine metabolites
and azelaic acid (13 ). These changes may indicate both
neurological and lipoperoxidation stress (13 ). Gotz et al.
(5) also reported that platelet CoQ
10
H
2
:CoQ
10
ratios were
significantly decreased in patients with Parkinson disease.
An altered redox state of platelet coenzyme Q
10
may
reflect a change in membrane electron transport and the
effectiveness of defense against toxic reactive oxygen
species, such as hydrogen peroxide and superoxide (5 ).
Another recent study suggested that CoQ
10
enrichment
may decrease oxidative DNA damage in human lympho-
cytes (14 ). Additional studies are needed to understand
the function and protective role of CoQ
10
H
2
in these and
other diseases.
In conclusion, we developed a simple, rapid, and isocratic
HPLC method for the determination of CoQ
10
H
2
and
CoQ
10
in human plasma. An extraction process using
1-propanol as solvent allows rapid and simple sample
extraction and minimizes oxidation of CoQ
10
H
2
during
sample processing. An in-line precolumn reduction cell is
used to convert CoQ
10
into CoQ
10
H
2
. The EC reduction
yields 99% CoQ
10
H
2
and avoids the artifactual oxidation
that frequently occurs with CoQ
10
H
2
produced through
the chemical reduction process. This optimized method
provides excellent sensitivity, precision, and accuracy for
relatively high-throughput assessment of CoQ
10
H
2
and
CoQ
10
in human plasma. This method is suitable for
research and can be easily adapted for clinical testing
purposes. Studies are in progress to establish reference
intervals and to evaluate the clinical significance of
plasma and cerebrospinal fluid concentrations of
CoQ
10
H
2
in several patient populations.
We thank Dr. Paul Steele for valuable suggestions, and
Gail Chuck and Laura Schroer for assistance in collecting
patient specimens and information for this study.
References
1. Ernster L, Dallner G. Biochemical, physiological and medical
aspects of ubiquinone function. Biochim Biophys Acta 1995;
1271:195–204.
2. Alleva R, Tomasetti M, Bompadre S, Littarru P. Oxidation of LDL
and their subfractions: kinetic aspects and CoQ
10
content. Mol
Aspects Med 1997;18:s105–12.
3. Stocker R, Bowry VW, Frei B. Ubiquinol-10 protects human low
density lipoprotein more efficiently against lipid peroxidation than
does
␣
-tocopherol. Proc Natl Acad SciUSA1991;88:1646–50.
4. Ernster L, Forsmark-Andre´e P. Ubiquinol: an endogenous antioxi-
dant in aerobic organisms [Review]. Clin Investig 1993;71:S60–5.
5. Gotz ME, Gerstner A, Harth R, Dirr A, Janetzky B, Kuhn W, et al.
Altered redox state of platelet coenzyme Q
10
in Parkinson’s
disease. J Neural Transm 2000;107:41–8.
6. Hara K, Yamashita S, Fujisawa A, Ishiwa S, Ogawa T, Yamamoto
Y. Oxidative stress in newborn infants with and without asphyxia
as measured by plasma antioxidants and free fatty acids. Bio-
chem Biophys Res Commun 1999;257:244– 8.
7. De Cavanagh EM, Ferder CG, Carrasquedo F, Serivo D, Wasser-
mann A, Fraga CG, Inserra F. Higher levels of antioxidant defense
in enalapril-treated versus non-enalapril-treated hemodialysis pa-
tients. Am J Kidney Dis 1999;34:445–55.
8. Yamamoto Y, Yamashita S. Plasma ubiquinone to ubiquinol ratio
in patients with hepatitis, cirrhosis, and hepatoma, and in patients
treated with percutaneous transluminal coronary reperfusion. Bio-
factors 1999;9:241– 6.
9. Kontush A, Reich A, Baum K, Spranger T, Finckh B, Kohlschutter A,
Beisiegel U. Plasma ubiquinol-10 is decreased in patients with
hyperlipidaemia. Atherosclerosis 1997;129:119–26.
10. Kontush A, Schippling S, Spranger T, Beisiegel U. Plasma ubiqui-
nol-10 as a marker for disease: is the assay worthwhile? Biofac-
tors 1999;9:225–9.
11. Lagendijk J, Ubbink JB, Delport R, Vermaak WJ, Human JA.
Ubiquinol/ubiquinone ratio as marker of oxidative stress in coro-
nary artery disease. Res Commun Mol Pathol Pharmacol 1997;
95:11–20.
12. Paloma¨ki A, Malminiemi K, Metsa¨-Ketela¨ T. Enhanced oxidizability
of ubiquinol and
␣
-tocopherol during lovastatin treatment. FEBS
Lett 1997;410:254 – 8.
13. De Luca C, Filosa A, Grandinetti M, Maggio F, Lamba M, Passi S.
Blood antioxidant status and urinary levels of catecholamine
metabolites in
␤
-thalassemia. Free Radic Res 1999;30:453– 62.
14. Tomasetti M, Littarru GP, Stocker R, Alleva R. Coenzyme Q
10
enrichment decreases oxidative DNA damage in human lympho-
cytes. Free Radic Biol Med 1999;27:1027–32.
15. Takada M, Ikenoya S, Yuzuriha T, Katayama K. Simultaneous
determination of reduced and oxidized ubiquinones. Methods
Enzymol 1984;105:147–55.
16. Lang JK, Packer L. Quantitative determination of vitamin E and
oxidized and reduced coenzyme Q by HPLC with in-line ultraviolet
and electrochemical detection. J Chromatogr 1987;385:109 –17.
17. Edlund PO. Determination of coenzyme Q
10
,
␣
-tocopherol and
cholesterol in biological samples by coupled-column liquid chro-
matography with coulometric and ultraviolet detection. J Chro-
matogr 1988;425:87–97.
264 Tang et al.: HPLC Analysis of Coenzyme Q
10

18. Okamoto T, Fukunaga Y, Ida Y, Kishi T. Determination of reduced and
total ubiquinones in biological materials by liquid chromatography
with electrochemical detection. J Chromatogr B 1988;430:11–9.
19. Grossi G, Bargossi AM, Fiorella PL, Piazzi S. Improved high-
performance liquid chromatographic method for the determination
of coenzyme Q
10
in plasma. J Chromatogr 1992;593:217–26.
20. Wakabayashi H, Yamato S, Nakajima M, Shimada K. Simulta-
neous determination of oxidized and reduced coenzyme Q and
␣
-tocopherol in biological samples by high performance liquid
chromatography with platinum catalyst reduction and electro-
chemical detection. Biol Pharm Bull 1994;17:997–1002.
21. Finckh B, Kontush A, Commentz J, Hubner C, Burdelski M,
Kohlschutter A. Monitoring of ubiquinol-10, ubiquinone-10, caro-
tenoids, and tocopherols in neonatal plasma microsamples using
high-performance liquid chromatography with coulometric electro-
chemical detection. Anal Biochem 1995;232:210– 6.
22. Lagendijk J, Ubbink JB, Delport R, Hayward WJ, Human JA.
Measurement of the ratio between the reduced and oxidized
forms of CoQ
10
in human plasma as a possible marker of
oxidative stress. J Lipid Res 1996;37:67–75.
23. Yamashita S, Yamamoto Y. Simultaneous detection of ubiquinol
and ubiquinone in human plasma as a marker of oxidative stress.
Anal Biochem 1997;250:66–73.
24. Kaikkonen J, Nyyssonen K, Salonen JT. Measurement and stabil-
ity of plasma reduced, oxidized and total coenzyme Q
10
in human.
Scand J Clin Lab Invest 1999;59:457– 66.
25. Wang Q, Lee BL, Ong CN. Automated high-performance liquid
chromatographic method with pre-column reduction for the deter-
mination of ubiquinol and ubiquinone plasma. J Chromatogr B
1999;726:297–302.
26. Artuch R, Moreno J, Quintana M, Puig RM, Vilaseca MA. Serum
ubiquinone-10 in a pediatric population. Clin Chem 1998;44:
2378 –9.
27. Zierz S, Jahns G, Jerusalem F. Coenzyme Q in serum and muscle
of 5 patients with Kearns-Sayre syndrome and 12 patients with
ophthalmoplegia plus. J Neurol 1989;236:97–101.
28. Cleary J, Mohr D, Adams MR, Celermajer DS, Stocker R. Plasma
and LDL levels of major lipophilic antioxidants are similar in
patients with advanced atherosclerosis and age-matched con-
trols. Free Radic Res 1997;26:175–82.
Clinical Chemistry 47, No. 2, 2001 265