Vapor Pressure and Freezing Point Osmolality Measurements
Applied to a Volatile Screen
EDWIN J. DRAVIAM, PH.D., EDWARD M. CUSTER, PH.D., AND IRWIN SCHOEN, M.D.
This is a report of a rapid and precise screening procedure,
developed for the determination of ethanol in serum using
osmolality measurements. The osmolality of the patient is
determined by freezing point method (freezing point osmometry)
and dew point (water vapor pressure osmometry) method. The
difference between freezing point osmolality and vapor pressure
osmolality (A osm) is due to the presence of volatiles in the
serum, because the volatiles are not measured by vapor pressure
osmometry. The amount of ethanol (mg/dL) in serum is
estimated by multiplying A osm by a factor of 4.2. As. a
comparison method, ethanol also is measured by a spectrophotometry alcohol dehydrogenase method. In addition, a significant difference between an osmometric alcohol assayed value
and enzymatic spectrophotometric measurement indicates the
presence of volatiles, other than ethanol. In addition to ethanol
there is a linear relationship between osmolality and isopropanol
or methanol when added in vitro to serum. (Key words:
Osmolality; Volatiles; Ethanol; Isopropanol; Methanol) Am J
Clin Pathol 1984; 82: 706-709
THE ELEVATION of serum osmolality due to the
presence of ethanol has been reported.7-8 The measurement of freezing point depression is directly proportional
to osmolality. However, ethanol and other water-soluble
volatiles (volatile alcohols, ketones, or aldehydes) are
not detected when one uses a vapor pressure osmometer
in which the measurement is based on dew point
depression. When one uses both types of determinations,
the difference between freezing point and vapor pressure
osmolality is an indication of the presence of watersoluble volatiles. The ingestion of ethanol has been
found to be the most common cause of increased serum
osmolality in healthy subjects.8 The elevation of serum
osmolality due to the presence of other volatiles and
nonvolatiles such as isopropanol, methanol, ethylene
glycol, and propylene glycol has been reported. 1 ' 5 ' 61012
Using both freezing point depression and dew point
measurements in our laboratory, we have established
linear relationships between ethanol, isopropanol, and
methanol and A osmolality.
The osmometric ethanol assay procedure presented
in this article may be used as an effective screening
procedure to determine whether a gas chromatographic
Received November 18, 1983; received revised manuscript and
accepted for publication June 1, 1984.
Address reprint requests to Dr. Schoen: Clinical Laboratory, Department of Pathology, University of Texas Medical Branch, Galveston,
Texas 77550.
706
Department of Pathology, Clinical Chemistry Division, The
University of Texas Medical Branch, Galveston, Texas
analysis for volatiles is necessary. This procedure is
especially suitable for small and medium-sized hospitals,
where osmometers are used routinely but riot gas chrorhatography (GC). Additional equipment usually is not
required, because many hospitals have both freezing
point and vapor pressure osmometers. The most recent
osmometers (Wescor Model 5100B®, List $3,195.00,
Wescor, Inc., Logan, UT 84321 and \i Osmette® Model
5004, List $3,495.00, Precision Systems, Inc., Sudsbury,
MA 01776) are compact in size and require as little as
0.7 and 0.5 square foot of bench space, respectively.
The procedure is simpie to perform on a 24-hour basis
and does not require special skills.
Materials and Methods
Apparatus
For the freezing point osmometry, an Advanced
Cryomatic Osmometer® (Model 3CII, Advanced Instruments, Needham Heights, MA 02194) and a n Osmette
Osmometer (Model 5004) were used. A Wescor Model
5100B Vapor Pressure Osmometer was used for water
vapor pressure osmometry. Enzymatic ethanol determinations were carried out on DuPont aca® (DuPont
Company, Wilmington, DE 19898).
Standards
Calibration standards and assayed run controls were
obtained from Advanced Instruments, Inc., and from
Wescor, Inc. Working standards containing 79, 158,
237, and 316 mg of absolute ethanol per deciliter were
made in pooled serum and in 0.9% NaCl solution.
Isopropanol standards containing 78, 156, 234, and 302
mg/dL in 0.9% NaCl were prepared. Methanol standards
containing 16, 32, 79, and 157 mg/dL also were made
in 0.9% NaCl.
Procedures
The Advanced Cryomatic Osmometer was calibrated
with the use of 100 and 900 mOsm/kg standards. The
Vol. 82 • No. 6
OSMOLALITY MEASUREMENTS APPLIED TO VOLATILE SCREEN
Wescor Vapor Pressure Osmometer was calibrated using
300 and 1,000 mOsm/kg standards. The 300 mOsm
standard was prepared in our laboratory by dissolving
9.429 g NaCl in 1 L distilled deionized water." Ortho®
Normal and Abnormal Unassayed Serum Controls (Ortho Diagnostics, Inc., Raritan, NJ 08869) were used as
daily run controls. Clinitrol 290® (Advanced Instruments,
Inc.) was employed as an assayed accuracy control on
both instruments.
A sample size of 300 /xL was used for Advanced
Cryomatic Osmometer, and 50 nL was used for /x
Osmette Osmometer. The samples and standards were
transferred into sample cuvettes with an Eppendorf®
pipette (Brinkmann Instruments, Inc., Westbury, NY
11590). In vapor pressure osmometry, an S-fiL sample
was delivered using a Drummond® pipette (Drummond
Scientific Company, Broomall, PA 19008) on a sample
filter paper disc placed in the sample chamber. A 10%
decrease in volume of the sample did not change the
osmolality of a control serum by freezing point and
vapor pressure osmometry, while a 10% increase in
volume resulted in 1% decrease in osmolality. The
difference between freezing point and vapor pressure
osmolality measurements by the two instruments represents delta (A) osmolality. The A osmolality was
determined for ethanol, isopropanol, and methanol standards and plotted against the respective alcohol concentration.
Osmolality was determined on patient serum specimens with significant alcohol levels. Venous blood was
drawn from these patients to completely fill a 10-mL
red-top Vacutainer® (Becton-Dickinson and Company,
Rutherford, NJ 07070) tube with less than 1 cm of
space between the top of the blood and the bottom of
the stopper. The clotted tubes were centrifuged for 5
minutes using a Model IEC Spinette® centrifuge (Damon/
IEC Division, Needham Heights, MA 021894). Serum
promptly was removed from the erythrocytes, and the
ethanol level was measured on Dupont aca, using an
alcohol dehydrogenase (EC 1.1.1.1) method. Two B-D
Microtainer® tubes (Becton-Dickinson) were filled completely with serum, capped, and kept frozen until the
osmolality determination. After thawing the stoppered
tubes at room temperature, the contents were mixed
thoroughly by inversion ten times before sampling for
osmolality determinations.
For osmolality determination, serum is reported to
be stable for 2 hours at room temperature and for
several days in the refrigerator.9 To study the stability
of serum ethanol at room temperature, a control serum
was spiked with ethanol (130 mg/dL) and kept frozen
in aliquots. The aliquots were thawed at room temperature, and the osmolality was determined at various
time intervals (0, 0.5, 1, 1.5, 2, 3, 4, 6 hours). The
707
Table 1. Serum Ethanol A Osmolality Storage Stability
at Room Temperature (23°C)
Time (h)
Freezing Point*
Osmolality (mOsm)
Vapor Pressure*
Osmolality (mOsm)
0
0.5
1
1.5
2
3
4
6
303
303
303
303
303
303
303
303
279
279
276
278
278
281
282
282
* Determinations were carried out in duplicates.
results are shown in Table 1. The SDs of freezing point
and vapor pressure measurements were within the SDs
calculated for the respective methods. The osmolality
was found to be stable at room temperature for at least
six hours for both freezing point and vapor pressure
measurements. It has been reported that there was no
significant change in ethanol levels when blood samples
with anticoagulant were frozen or refrigerated for several
months and analyzed by volatilization of reducing substances in blood.4
Results
Using Ortho Normal Control Serum, the run-to-run
coefficient of variation (CV) for osmolality measurements
was determined as 0.94% (276 ± 2.6 mOsm/kg, mean
± SD, N = 20) for the cryoscopic osmometer and 1.1%
(276 ± 3.1 mOsm/kg, N = 20) for the vapor pressure
osmometer. The instruments were cross-checked with
calibrators of the other instrument. The mean of the
differences between cryoscopic and vapor pressure measurements was 0.3 mOsm/kg (standard error of the
mean = 0.42). The sensitivity of A osmolality measurement near the normal range, calculated as 2 SD of the
difference between duplicate cryoscopic and vapor pressure determinations, was 3.9 mOsm/kg or 16 mg/dL of
ethanol (N = 20).
Two ethanol controls were made by spiking approximately 100 mg/dL and 300 mg/dL of ethanol to serum.
The mean ± SD (N = 20) for the controls as determined
by osmometry were 100 ± 8 mg/dL and 290 ± 8 mg/
dL, respectively. The run-to-run precision as 2 SD for
the A osmolality measurement (N = 20) was 3.8 mOsm/
kg for both the levels.
The serum osmolality for 51 random hospitalized
patients on the cryoscopic osmometer ranged from 262
to 306, with a mean ± SD of 281 ± 9.6 mOsm/kg. The
standard error of the mean (SE*) was 1.3 mOsm/kg.
The serum osmolality on the vapor pressure osmometer
was 280 ± 10.2 mOsm/kg (SE ; = 1.4). Using a Mest,
there was no significant difference between freezing
A.J.C.P. • December 1984
DRAVIAM, CUSTER, AND SCHOEN
708
400 n
A Osmolality (mOsm)/Kg
FIG. 1. Linear regression of A osmolality with standard
solutions of ethanol in serum.
For 61 patients with positive alcohol levels for the
aca (sensitivity = 5 mg/dL), the difference in osmolality
between cryoscopic and vapor pressure determinations
was compared with the enzymatically assayed ethanol
concentrations. The comparison (slope = 1.01; intercept
= 7.49; Sj,x = 22.1; r = 0.9797) is shown in Figure 2.
All the specimens were negative for acetoacetic acid (<5
mg/dL) and acetone (<50 mg/dL) with Acetest® reagent
tablets (Ames Company, Elkhart, IN 46515). Five specimens with high levels of ethanol (>340 mg/dL) were
analyzed for the ethanol, methanol, isopropanol, and
acetone by gas chromatographic volatile screen. No
other alcohol except ethanol was present. The ethanol
concentrations by GC were, on the average, 9% higher
than the results from osmometry and 13% higher than
the aca results. Because enzymatic determinations using
aca were linear up to only 300 mg/dL, these specimens
were diluted with water. The discrepancy in the aca and
GC results may be due to the error in dilution of the
specimens for analysis by aca. One of the five specimens
was found to contain 30 mg/dL of acetone. This could
be equivalent to 5 mOsm/kg.3
point and vapor pressure measurements (P > 0.50). The
Discussion
A osmolality for these patients ranged from —4 to +6
In order to use osmolality measurements as a screening
mOsm/kg. Three of the patients had a A osmolality of
procedure, the clinically significant concentrations of
5 mOsm/kg and one patient 6 mOsm/kg, which were
the volatile substance need to exceed the sensitivity of
greater than the sensitivity (3.9 mOsm/kg) of the method.
the A osmolality measurements. In addition to ethanol,
This may be due to random error or due to the pressure
isopropanol satisfies the above conditions. Although
of endogenous volatiles.
ethanol increases serum osmolality more than isoproA linear response was indicated when a serum pool
panol on w/w basis, it is not possible to tell by osmolality
spiked with ethanol corresponding to 0, 79, 158, 237,
and 316 mg/dL were plotted against A osmolality (Fig.
1). The concentration of ethanol is obtained by linear
regression as follows.
600
Ethanol (mg/dL) = 4.2 (A osm) - 0.04. (standard error
y = 1.01 x + 7.49
Sv x = 21.13
of estimate, Sy)l = 2.51; correlation coefficient, r
r = 0.9797
>500= 0.9998; N = 5)
/) = 61
«to-» i• Dj
3,7
The slope of 4.2 is consistent with previous reports.
o \CD
With standards made by spiking ethanol to 0.9% NaCl,
F fc 4 0 0 a
concentrations up to 1,000 mg/dL were found to be O<>/ <>/
linear with A osmolality. About 8% decrease in A
ox
300osmolality was observed when the standards were made
*•—
ID
3
o
in saline as opposed to serum. However, this discrepancy
raj
<o 2 0 0 Sr
can be eliminated if the A osmolalities of aqueous
0)
r.
standards are corrected for volume of water (91%) in
the serum.14
100Similar concentrations of isopropanol and methanol
in saline were found to be linear with measured osmo100
200
300
400
500 600
lality.
Ethanol from Enzymatic
Isopropanol (mg/dL) = 5.0 (A osm) - 6.2 (Syx = 6.37;
Determinations (mg/dl)
r = 0.9989; N = 5).
Methanol (mg/dL) = 3.2 (A osm) + 1.6 (Sy.x = 1.50;
FIG. 2. Linear regression of enzymatically determined ethanol and
ethanol determined from A osmolality measurements.
r = 0.9998; N = 5).
—
men
rom
•
i_
JJS
o
Vol. 82 • No. 6
OSMOLALITY MEASUREMENTS APPLIED TO VOLATILE SCREEN
measurement when a mixture of both alcohols is present.
In the emergency room, most of the cases involve
intoxication with high blood levels of ethanol. Very few
cases of isopropanol intoxication are encountered. Because isopropanol is readily metabolized to acetone, the
patients with isopropanol intoxication have significant
levels of acetone in blood. The sensitivity of acetone by
osmometry is 21 mg/dL (3.9 mOsm/kg). Elevation of A
osmolality and a positive acetone may be a practical
indicator of the presence of excess isopropanol or acetone
(e.g., diabetic coma). In this case a quantitative volatile
screen by GC may be necessary to positively identify
the alcohols present.
The lethal level of methanol (200 mg/dL) is high
enough so that a significant A osmolality will be obtained
with high levels of methanol. Toxic levels of methanol
greater than 400 mg/dL have been reported.5 The sensitivity of osmolality measurements will indicate toxic
levels of methanol as it does for ethanol. Ethylene glycol,
which is toxic (lethal dose: 100 g), and propylene glycol,
which is relatively nontoxic, may attain high blood
levels. Propylene glycol has been demonstrated in high
enough concentrations to be measurable in the sera of
two severely burned patients.1 Additional studies are in
progress concerning the use of osmometry in comparison
with other methods for detecting nonvolatile alcohols
and other volatile aldehydes and ketones of possible
clinical significance.
The A osmolality due to the presence of volatiles
commonly has been determined by finding the difference
between the measured osmolality and calculated osmolality using various formulas taken from the literature
that are based on analyte (Na, K + , etc.) assays.2,313
These literature presentations do not reveal findings as
reliable as those presented in this article. Analytic sensitivity and specificity and correlation coefficients are
not as good as achieved in this study using osmolality
determinations on the vapor pressure osmometer instead
of the calculated osmolality from sodium, glucose, and
urea-nitrogen concentrations.
In this study no significant change in osmolality was
observed by the Wescor instrument when up to 318
mg/dL of ethanol was added to control serum, saline,
or water. Similarly, 157 mg/dL of methanol and 302
mg/dL of isopropanol did not change the osmolality
measured by vapor pressure.
The following procedure is presented for the determination of ethanol in serum. Osmolality measurements
are made with cryomatic and vapor pressure osmometers,
and the A osmolality is multiplied by a factor of 4.2.
The result is reported as follows: Water-soluble volatiles
are present equivalent to XXX mg/dL as ethanol (sen-
709
sitivity: ethanol = 16 mg/dL). If only isopropanol or
methanol was known to be ingested, the respective linear
regression factor could be used for the determination of
concentration, and the result will be reported in terms
of isopropanol or methanol (sensitivity: isopropanol
= 14 mg/dL; methanol = 14 mg/dL).
Water-soluble volatiles equivalent to or greater than
16 mg/dL as ethanol are assayed by a more specific
enzymatic ethanol method. If the enzymatic result is
less than the value obtained from A osmolality by more
than 29 mg/dL (3 SD diff; SD2 diff = SD 2 enz + SD 2
A osm; where enz = enzymatic determination; A osm
= A osmolality measurement), a volatile screen by gas
chromatography is done. If less than 29 mg/dL the
presence of significant amounts of (> 1 /6 of lethal level),
methanol and isopropanol is ruled out.
This volatile screen assay procedure is especially useful
in measuring alcohol in the clinically significant range
of 50 mg/dL to 600 mg/dL.
References
1. Bekeris L, Baker C, Fenton J, Kimball D, Bermes E: Propylene
glycol as a cause of an elevated serum osmolality. Am J Clin
Pathol 1979; 72:633-636
2. Dorwart WV, Chalmers L: Comparison of methods for calculating
serum osmolality from chemical concentrations, and the prognostic value of such calculations. Clin Chem 1975; 21:190-194
3. Glasser L, Sternglanz PD, Combie J, Robinson A: Serum osmolality
and its applicability to drug overdose. Am J Clin Pathol 1973;
60:695-699
4. Glendening BL, Waugh TC: The stability of ordinary blood
alcohol samples held various periods of time under different
conditions. J Forensic Sci 1965: 10:192-200
5. Jacobson D, Bradesen JE, Eide I, Ostborg J: Anion and osmolar
gaps in the diagnosis of methanol and ethylene glycol poisoning.
Acta Med Scand 1982; 212:17-20
6. McCord WM, Switzer PK, Brill HM, Jr: Isopropyl alcohol intoxication. South Med J 1948; 41:639-642
7. Redetski HM, Koerner TA, Hughes JR, Smith AG: Osmometry
in the evaluation of alcohol intoxication. Clin Toxicol 1972;
5:343-363
8. Robinson AG, Loeb JN: Ethanol ingestion Commonest cause of
elevated plasma osmolality? N Engl J Med 1971; 284:12531255
9. Schoen I, Fischer C, Winter S, Barnett R: Patient preparation and
specimen collection and handling including storage stability,
Medical Usefulness Subcommittee of CAP Standards Committee, March 1974
10. Stern EL: Serum osmolality in cases of poisoning (Letter to the
Editor). N Engl J Med 1974; 290:1026
11. Tietz NW: Fundamentals of of clinical chemistry. Second edition.
Philadelphia, WB Saunders, 1976
12. Vasiliades J, Pollock J, Robinson CA: Pitfalls of the alcohol
dehydrogenase procedure for the emergency assay of alcohol:
A case study of isopropanol overdose. Clin Chem 1978; 24:383385
13. Weisberg HF: Osmolality-calculated, "delta," and more formulas.
Clin Chem 1975;21:1182-1185
14. Wolf AV: Aqueous solutions and body fluids. New York: Hoeber
Medical Division, Harper and Row, 1966