Eur J Clin Chem Clin Biochem 1996; 34:651-654 © 1996 by Walter de Gruyler · Berlin · New York
Interferences of Glycerol, Propylene Glycol, and other Diols
in the Enzymatic Assay of Ethylene Glycol1)
Henri Malandain and Yves Cano
Laboratoire de Biochimie, Centre Hospitalier Chubert, Vannes, France
Summary: As an alternative to gas chromatography, the enzymatic
UV assay of ethylene glycol is often used by emergency laborato-
ries. Many variants of this technique have been published, all based
on the reaction between NAD+ and ethylene glycol in the presence
of glycerol dehydrogenase (EC 188.8.131.52). We show that other a-
diols interfere in this reaction. Some of them, like 2,3-butanediol,
give false positive reactions; whereas other diols, e. g. glycerol and
propylene glycol, interfere only when ethylene glycol is present in
the sample and lower the ethylene glycol response; these interfer-
ents are of particular concern because some parenteral drugs used
in emergency situations contain glycerol or propylene glycol in
their vehicle. This drawback has hitherto been largely underesti-
mated, and we think that ethylene glycol results obtained with these
enzymatic techniques should be interpreted with caution, even if
the sample is pre-treated with glycerokinase (EC 184.108.40.206); this
pre-treatment effectively corrects the glycerol interference but not
that of propylene glycol.
Ethylene glycol causes severe poisoning in man. In suspected
cases, an emergency ethylene glycol assay is necessary but gas
chromatography is not available in every laboratory. Hanson &
Masson (1) and Standefer & Blackwell (2) recently popularized
simple UV enzymatic assays of ethylene glycol using glycerol de-
hydrogenase (EC 220.127.116.11) from Enterobacter aerogenes:
+ NADH + H+
These authors found that glycerol and propylene glycol reacted
very quickly and thus did not interfere if absorbance readings be-
gan after a delay sufficient for their complete conversion (fig. 1).
These authors also noted an inhibitory effect of the reaction pro-
duct, glycolaldehyde; but this inhibition was of no clinical impor-
tance because glycolaldehyde, like other aldehydes, is rapidly
cleared from the circulation and does not accumulate sufficiently
to affect the ethylene glycol measurement. Nilsson & Jones (3)
later reported 2,3-butanediol as an additional substrate. 2,3-Bu-
tanediol reacts at a rate similar to that of ethylene glycol (fig. 1),
and it interferes in the ethylene glycol measurement irrespective of
the absorbance reading times.
In view of the poor specificity of glycerol dehydrogenase, we de-
cided to test some other compounds of toxicological interest. We
also performed these tests in the presence of ethylene glycol to
reveal any competitive or inhibitory effects.
Materials and Methods
All chemicals were of analytical grade.
Experiments were done on a Boehringer Hitachi 717 at 30 °C.
Sample: 10 μΐ.
Reagent 1: 250 μΐ of 15 mmol/1 NAD+ in Tris buffer 0.1 mol/1 pH
9.0 (Sigma ref. 545-2).
Reagent 2: 50 μΐ of 45 kU/1 glycerol dehydrogenase from Entero-
bacter (Boehringer ref. 258-255) in the same buffer.
Reagent 2 was added 5 min after reagent 1.
Two 340 nm absorbance readings were taken 2 min and 5 min after
reagent 2 addition (these reading times correspond to the 35th and
50th Hitachi 717 cycles, respectively).
Concentrations were calculated from the difference between these
two absorbance readings.
Calibration curves were prepared with 0, 5, 10, 20, and 40 mmol/1
ethylene glycol standards prepared in 0.154 mol/1 NaCl.
Reagent 1 was modified by the addition of glycerokinase, ATP and
Mg2"1" to catalyse the conversion of glycerol to glycerol-3-phos-
phate before the addition of glycerol dehydrogenase:
}) This work was presented in part at the 47th A. A. C. C. Annual
Meeting, July 16-20 1995, Anaheim (CA), USA
Time after reagent 2 addition [min]
Fig. 1 Reaction kinetics of ethylene glycol and of some interfer-
Glycerol + ATP
Glycerol-3-phosphate + ADP
Modified reagent 1 was prepared as follows: to 15 ml of normal
reagent 1 were added 1.4 mmol of ATP · 2Na (Sigma rcf. A7699),
100 μΐ of glycerokinase EC 18.104.22.168 (Sigma ref. G6278), and 100
μΐ of 1 mol/1 MgCl2.
If not otherwise stated, possible interferents were assayed at 50
mmol/1 in 0.154 mol/1 NaCl in the absence of ethylene glycol and
in the presence of 20 mmol/1 ethylene glycol. A compound did not
significantly interfere when it produced a result < 0.5 mmol/1 in
the absence of ethylene glycol and within 20 ± 1 mmol/1 in the
presence of 20 mmol/1 ethylene glycol.
Glycerol and propylene glycol interferences were studied more
thoroughly: 0.5, 1,5, and 20 mmol/l of the interferent were assayed
in the presence of 0, 1, 5, 20, and 40 mmol/1 of ethylene glycol.
With the normal procedure most of the compounds we tested did
not interfere (tab. 1), but:
- a false positive result was observed with 2,3-butanediol which
gave a response equivalent to 49% ofthat of ethylene glycol on a
molar basis. This effect was quantitatively of the same magnitude
in the absence as in the presence of ethylene glycol in the sample;
- a negative effect on the ethylene glycol response was observed
with glycerol, propylene glycol, glycolaldehyde, 1,2-butanediol
and glycolic acid. These compounds produced no significant re-
sponse as long as ethylene glycol was absent from the sample; but
when it was present the expected ethylene glycol response was
lowered by an amount which depended on the nature and on the
concentration of the interferent; as an example, glycerol and pro-
pylene glycol interferences are presented in figure 2. The effect of
1,2-butanediol was quantitatively the same as that of propylene
glycol. The interferences observed with glycolaldehyde and gly-
Tab. 1 Interference results (see text for details). Possible interfer-
ents were assayed at 50 or 40 (*) mmol/1
These compounds interfered:
— by giving false positive
— by depressing the ethylene
These compounds did not interfere significantly:
Ascorbic acid (180 mg/1)
5 Ethylene glycol
Fig. 2 The effect of propylene glycol and of glycerol on the re-
sponse of ethylene glycol.
colic acid were only -15% and -9.5% on a molar basis, respec-
Glycerol no longer interfered with the ethylene glycol measure-
ment in the modified procedure, as glycerol was converted into
glycerol-3-phosphate before the addition of reagent 2 (tab. 2). This
trapping procedure, however, did not work with propylene glycol
or with other interfering compounds.
It would be advantageous to measure ethylene glycol without the
need of a specific instrument like a gas Chromatograph. This is
possible with enzymatic techniques like those of Hanson & Masson
or Standefer & B lackwall (1,2). These authors claimed that glyc-
erol and propylene glycerol did not interfere if the absorbance read-
ings were taken after a sufficient delay. Unfortunately, these au-
thors did not assay glycerol and propylene glycol in the presence
of ethylene glycol.
Tab. 2 Ethylene glycol response in the presence of glycerol after
pre-treatment with glycerokinase
Ethylene glycol (mmol/1)
0 1 5
The weak specificity of glycerol dehydrogenase from Enterobacter
aerogenes was noted as early as 1982 (4, 5), when it was used as
the basis of a screening test for ethylene giycol poisoning using
the DuPont aca triglyceride UV reagent packs (6). Ryder and co-
authors tried to improve their original technique by pretreating the
sample with glycerokinase (7); they nevertheless observed that pro-
pylene giycol could also react (6). Blandford & Desjardins recently
suggested a similar procedure (a "triglyceride gap" between aca
and triglyceride-PAP results) and noted that propylene giycol inter-
fered (8). Malhy et al. adapted the Hanson & Masson technique to
a Cobas Mira and noted that glycerol interfered, but did not check
the propylene giycol effect (9). Most of these authors did not un-
derline the clinical importance of such interferences.
Our present study shows that the ethyiene giycol response is
greatly affected by the presence of other α-diols, even with a delay
before the first absorbance reading. The glycerol and propylene
giycol interferences are clinically serious because these molecules
are present in the vehicle of some parenteral drugs in common
use (tab. 3).
Many studies have shown high plasma propylene giycol levels after
parenteral drug use: up to 9.3 mmol/1 after phenytoin and trimetho-
prim/sulfamethoxazole (10, 11), 52.6 mmol/1 after etomidate (12),
up to 122 mmol/1 after a multivitamin drug (13), and, on average,
30.3 mmol/1 after nitroglycerol (14). Among these drugs, anticon-
vulsants are sometimes used during the management of ethylene
giycol poisoning (15). The avoidance of propylene giycol as in-
ternal standard in the ethylene giycol gas Chromatographie assay
has been practised for a long time (10, 16); for instance, the pres-
ence of propylene giycol was detected in 35 out of 138 plasma
samples sent for ethylene giycol analysis (17).
The presence of glycerol in parenteral drug formulations did not
receive as much attention as that of propylene giycol. Glycerol is
present in fewer parenteral drugs than propylene giycol (tab. 3).
Some authors however consider that the level of free glycerol in
plasma can be unpredictably high, even in the ambulatory patient
(18); stress (19) and ketotic situations (18) are possible causes of
glycerol accumulation. In addition, during alcoholic and diabetic
ketoacidosis abnormal glycerol and propylene giycol levels can be
simultaneously present (20, 21). It should be possible to overcome
the glycerol interference with the addition of glycerokinase to the
first reagent, as in the Ryder et al. method (7) or in the modified
procedure we describe. However, this means that the reagent prepa-
ration is more complicated. Moreover, propylene giycol still in-
Other interferents are of less importance in practice:
- glycolaldehyde and glycolic acid are produced in vivo during
the catabolism of ethylene giycol but glycolaldehyde, like other
aldehydes, is rapidly cleared from the blood circulation and does
not accumulate sufficiently to affect the ethylene giycol measure-
ment; glycolic acid slightly lowers the ethylene giycol response;
but as much as 21 mmol/1 of glycolic acid would be necessary to
lower the ethylene giycol result of 2 mmol/l;
- among the different diols tested, only α-diols interfered: 2,3-
butanediol cannot be differentiated from ethylene giycol because
of their similar reaction rates; however, 2,3-butanediol circulating
levels are usually lower than 1 mmol/1 even in some pathological
situations like alcoholism or diabetes (22); appreciable levels of
2,3-butanediol (up to 10 mmol/1) are seen only in anecdotal cases
like genetic acidaemias (22) or after ingestion of alcohol denatur-
ated with 2-butanone (3). Finally, the presence of 1,2-butanediol is
very unlikely owing to the limited use of this compound.
In conclusion, the results of ethylene giycol analysis with glycerol
dehydrogenase need cautious interpretation. Nevertheless, this en-
zymatic method could remain a valuable screening test for most
Tab. 3 Parenteral drugs containing propylene giycol or glycerol in their vehicle (commercial drugs
available in France in 1996)
containing propylene giycol
Hydroxy-vitamin D,l alpha
Sulfadoxin + pyrimethamine
Trimethoprim + sulfamethoxazole
IV lipid emulsions
Voldal injectable 75 mg
Voltarene inj. 75 mg
Hypnomidate 2 mg/ml
Un-alfa 1 and 2 μg
Lenitral 3 and 15 mg
Feldene 20 mg
Mediatensyl 25 and 50 mg
We thank Professeur Claude Bohuon of the Institut Gustave-
Roussy, Villejuif, France, for critically reviewing the manuscript.
1. Hansson P, Masson P. Simple enzymatic screening assay for
ethylene glycol in serum. Clin Chim Acta 1989; 182:95-102.
2. Standefer J, Blackwell W. Enzymatic method for measuring
ethylene glycol with a centrifugal analyser: Clin Chem 1991;
3. Nilsson L, Jones AW. 2,3-Butanediol: a potential interfering
substance in the assay of ethylene glycol by an enzymatic
method. Clin Chim Acta 1992; 208:225-9.
4. Cano Y, Malandain H. Interference du propylene glycol dans
le dosage enzymatique des triglycerides. Ann Biol Clin 1982;
5. Click MR, Enockson CB, Taylor DK, Trent JL. Sample in-
teraction (carryover) on the aca when glycol-stabilized serum
is used. Clin Chem 1982; 28:1821.
6. Ryder KW, Click MR, Jackson SA. Emergency screening for
ethylene glycol in serum. Clin Chem 1986; 32:1574-7.
7. Ochs ML, Click MR, Ryder KW, Moorehead WR. Improved
method for emergency screening of ethylene glycol in serum.
Clin Chem 1988; 34:1507-8.
8. Blandford DE, Desjardins PR. A rapid method for the mea-
surement of ethylene glycol. Clin Biochem 1994; 27:25-30.
9. Malhy M, Lardet G, Vallon JJ. Automated Cobas Mira kinetic
enzymatic assay for ethylene glycol applied to emergency situ-
ations. J Anal Toxicol 1994; 18:269-71.
10. LeGatt DF, Tisdell RH. Ethylene glycol quantification: avoid
propylene glycol as an internal standard. Clin Chem 1990;
11. Keiner NJ, Bailey DN. Propylene glycol as a cause of lactic
acidosis. J Anal Toxicol 1985; 9:40-2.
12. Bedichek E, Kirschbaum B. A case of propylene glycol toxic
reaction associated with etomidate infusion. Arch Intern Med
13. Glasgow AM, Boeckx RL, Miller MK, MacDonald MG, Au-
gust GP, Goodman SI. Hyperosmolality in small infants due
to propylene glycol. Pediatrics 1983; 72:353-5.
14. Demey HE, Daelemans RA, Verpooten GAj de Broe ME,
Campenhout CM, Lakiere FV, et al. Propylene glycol-induced
side effects during intravenous nitroglycerin therapy. Intens
Care Med 1988; 14:221-6.
15. Ellenhorn MJ, Barceloux DG. Ethylene glycol. In: Ellenhorn
MJ, Barceloux DG, editors. Medical toxicology - diagnosis
and treatment of human poisoning. New York: Elsevier Sci-
16. Apple FS, Googins MK, Resen D. Propylene glycol interfer-
ence in gas^chromatographic assay of ethylene glycol. Clin
Chem 1993; 39:167.
17. Miceli JN, Vroon DH, Harris C, Poor J. Ethylene glycol: a
continuing problem. Clin Chem 1992; 38:995.
18. Naito HK. Triglycerides. In: Pesce AJ, Kaplan LA, editors.
Methods in clinical chemistry.
19. Klein S, Peters EJ, Shangraw RE, Wolfe RR. Lipolytic re-
sponse to metabolic stress in critically ill patients. Crit Care
Med 1991; 19:776-9.
20. Braden GL, Strayhorn CH, Germain MJ, Mulhem JG,
Skutches CL. Increased osmolal gap in alcoholic acidosis.
Arch Intern Med 1993; 153:2377-80.
21. Reichard GA Jr, Skutches CL, Hoeldtke RD, Owen OE. Ace-
tone metabolism in humans during diabetic ketoacidosis. Dia-
betes 1986; 35:668-74.
22. Casazza JP, Song BJ, Veech RL. Short chain diol metabolism
in human disease states. Trends Biochem Sei 1990; 15:26-30.
St. Louis: Mosby,
Received January 19/May 9, 1996
Corresponding author: Henri Malandain, Laboratoire de
Biochimie, Centre Hospitalier Chubert, 20 Boulevard General
Maurice Guillaudot, BP 555, F-56017 Vannes, France
Eur J Clin Chem Clin Biochem 1996; 34:655-657 © 1996 by Walter de Gruyter · Berlin · New York
Audit in Laboratory Medicine
Mario Plebani1 and Maria Laura Chiozza2
1 Department of Laboratory Medicine
2 Department of Pediatrics
University-Hospital, Padova, Italy
Summary: Laboratory medicine is at a crossroads that provides
great opportunities to improve the efficacy of the service and its
role in healthcare. Many of the new roles for laboratory staff will
be outside the boundaries of the traditional laboratory. These roles
include increased emphasis on consultative activities, participation
in interdisciplinary teams, and efforts to assure appropriate test
utilization. Audit in laboratory medicine may be defined as a pro-
cess of review and assessment of laboratory performance, and its
purpose should be to improve patient care by enhancing laboratory
performance and making better use of resources. Here we discuss
the rationale of audit in laboratory medicine, its goals as well as
its topics. The suggested conduct for an audit and the involvement
of personnel are also reviewed.
In Italy, as well as in other European countries, there is wide debate
regarding the laboratory accreditation process (1, 2). The concept
of laboratory accreditation is defined by ISO/IEC as a formal re-
cognition that a testing laboratory is competent in carrying out spe-
cific tests or specific types of tests (3). However, from a profes-
sional viewpoint, the accreditation programmes need to confirm
that a laboratory is providing quality service and that clinicians,
patients and managers are satisfied with the service. Much evidence
exists that quality audit is an important tool for reviewing and im-
proving the quality of service. Continuing audit of the service pro-
vided should be a fundamental part of any accreditation scheme,
as it is an important element in the laboratory quality system. Audit
in laboratory medicine may be defined as a process of review and
assessment of laboratory performance, and its purpose should be
to improve patient care by enhancing laboratory performance and
making better use of resources. The audit process should, more-
over, contribute to the education of laboratory staff and clinicians
The need for audit
The main role of the clinical laboratory is to provide information
to assist in the diagnosis of illness, and to promote patient care.
However, as yet, quality assessment programmes in clinical
laboratories have dealt with the integrity of the analytical process
that produces laboratory test results, without considering whether
the tests enhance patient care. In the past, the main focus was on
improving the analytical reliability of laboratory tests, but the new
standard to be adopted should be to establish how appropriate tests
are for patient care. Evidence shows that patient protection cannot
be achieved merely through accuracy and precision in the analyti-
cal phase of the testing process. In their investigation, Ross and
Boone found that mistakes made in laboratory testing were distrib-
uted as follows: 46% preanalytical, 7% analytical and 47% postan-
alytical (7). Non-laboratory personnel were responsible for 28.6%
of these mistakes. Similar results were described by Bachner and
colleagues in a College of American Pathologists Q-Probe study
on blood-bank quality assurance practice (8). Mistakes in the pre-
analytical and postanalytical phases of testing seem to compromise
the usefulness of laboratory results in patient care more often than
poor performance in the analytical phase of testing. This observa-
tion stresses the need to evaluate all steps in the entire testing pro-
cess and to promote the quality of the total testing process itself
(9, 10). Laboratory competence, therefore, calls for the ability to
conduct test ordering, specimen collection, transport, storage,
analysis and result reporting in an accurate, timely, and cost-effec-
tive way (11). Moreover, the rationale for the enactment of Clinical
Laboratory Improved Amendments (CLIA '88), the accreditation
programme for clinical laboratories in USA, was born out of an
understanding of "the critical role played by laboratory testing in
the delivery of health services and in maintaining good health"
(12). This programme is testimony to the government's shift: rather
than protecting physicians from poor quality laboratory testing, it
attempts to protect patients from poor-quality outcomes associated
with testing errors. Patients should thus be protected from all errors
in the testing process, whatever their origin. Again, the protection
of the patient's interests and the safeguarding of public health calls
for a departure from the conventional view of the laboratory, which
focuses upon the quality control of analytical activities within the
laboratory, and promotion of quality control of the total testing
process. This necessitates close cooperation between laboratory and
clinicians, creating for the future "a laboratory without walls" (13).
Experts in the total testing process stress the following key factors,
which are of fundamental importance:
a) the formulation of the clinical question,
b) the interpretation of the laboratory results in the context of the
c) the use of those results in subsequent decisions affecting the
clinical approach (14).
Laboratory specialists are responsible not only for adopting the
most reliable techniques or methods, but also for selecting the most
appropriate test for the specific clinical situation. They should,
moreover, alert clinicians to the sensitivity, specificity, and predic-
tive values of various tests. The laboratory should provide an inter-
pretation of not only the results of the more complicated tests it
performs, but it should also be able to use informatics to transform
the data it produces into clinically useful information.
Audit and medical needs
Audit in laboratory medicine cannot be merely considered a review
of analytical results made using either internal quality control pro-
cedures or external quality assessment schemes. All aspects of the
service, including the consulting of the laboratory must be audited.
Clinicians should be involved in certain aspects of laboratory audit
and joint clinical and laboratory guidelines and protocols on the
use of laboratory investigations should be issued. The concept of
"quality" as defined by the International Organization for Stan-