1. Clin Chem Lab Med 2015; aop
*Corresponding author: Matthew T. Wittbrodt, School of Applied
Physiology, Georgia Institute of Technology, Atlanta, GA 30332, USA,
Phone: +1 231 590-8940, Fax: +1 404 8940-9982,
E-mail: mwittbrodt3@gatech.edu
Sofia Espinoza: School of Industrial and Systems Engineering,
Georgia Institute of Technology, Atlanta, GA, USA
Mindy L. Millard-Stafford: School of Applied Physiology, Georgia
Institute of Technology, Atlanta, GA, USA
Matthew T. Wittbrodt*, Sofia Espinoza and Mindy L. Millard-Stafford
Biological variation of plasma osmolality obtained
with capillary versus venous blood
DOI 10.1515/cclm-2014-1006
Received October 13, 2014; accepted January 1, 2015
Abstract
Background: Plasma osmolality (POsm) is a gold stand-
ard to assess hydration status but requires venipunc-
ture. POsm obtained by lancing a digit, a source of
capillary puncture blood (CAP), has not been validated.
This study compared POsm from CAP versus venous
blood (VEN) and validated its sensitivity to detect
dehydration.
Methods: Healthy young adults (Study A: n = 20 men,
22 women; Study B: n = 23 men, 23 women) participated.
In Study A, CAP and VEN were compared under controlled
euhydration meeting dietary reference intakes for water
(3.7 L men, 2.7 L women). In Study B, CAP was assessed for
sensitivity to detect dehydration with receiver operating
characteristic analysis over two 24 h periods: euhydration
for 24 h followed by water restriction over 24 h. POsm was
measured using freezing point depression.
Results: For all subjects, CAP POsm (283.0±3.9 mOsm/kg)
wasnotsignificantlydifferent(p = 0.07)fromVEN(284.2±3.5)
during euhydration and met analytical goals for individual-
ity and heterogeneity. When outliers (n = 3) were eliminated,
mean difference was –1.6 (±3.2) lower (p < 0.01) with CAP.
Fluid restriction increased (p < 0.001) CAP POsm (284.0±4.4
to 292.8±5.2 mOsm/kg), achieving excellent accuracy (0.92)
and sensitivity (89.1%) to predict mild dehydration (2% body
mass loss).
Conclusions: POsm via CAP exhibited similar coeffi-
cients of variation and analytical goals compared to
VEN combined with excellent accuracy and sensitiv-
ity to detect dehydration. Although CAP values were
approximately 2 mOsm/kg lower than VEN, CAP appears
an adequate substitute for tracking changes in non-clin-
ical settings.
Keywords: capillary puncture; hydration; plasma
osmolality; venipuncture; water intake.
Introduction
Water balance is essential for human health and perfor-
mance [1]. Moreover, body water deficits are a common
daily occurrence in clinical, occupational, and sports
settings [2]. Plasma osmolality (POsm) is a sensitive and
valid clinical measure of hypertonic-hypovolemia [3–5],
serving as the single best biomarker for assessing hydra-
tion status at a static point in time [3]. However, POsm
is often assessed with venipuncture from an antecubital
vein and thus requires a trained phlebotomist which
limits its use to primarily clinical settings. Since
hydration status is a measure of interest by a variety
of sports medicine practitioners, additional methods
to enhance the identification of dehydration would be
advantageous.
Capillary puncture blood (CAP) obtained by lancing
a digit is commonly used for other analytical tests (i.e.,
blood glucose, hemoglobin, arterial blood gases, lactate)
with the practical benefits of lower cost, less techni-
cian skill requirements, less invasive for subjects, and
increased capacity for repeated sampling. CAP speci-
mens are a mixture of arterial, venous, and capillary
blood in addition to interstitial and intracellular fluid.
Thus, reference values may differ in CAP compared to
venous blood (VEN) for various electrolytes (e.g., sodium
which ultimately influences osmolality). Studies [6, 7]
indicate that both plasma and serum sodium are higher
(by 1.7–1.9 mmol) in VEN compared to corresponding
values obtained by CAP. To our knowledge, previous
research has not compared the validity of employing
CAP for POsm, only whole blood and plasma compari-
sons of osmolality [8]. Theoretically, CAP may be unreli-
able or invalid due to either contamination from wound
debris or dilution from tissue fluids when compressing
the digit to obtain sufficient sample volume. However,
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

2. 2 Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood
the use of free flowing blood samples by heating the
digit should minimize potential problems. Therefore,
the purpose of this study was to examine the biologi-
cal variability of POsm with CAP compared to VEN.
We hypothesized CAP would yield slightly lower mean
POsm values (due to differences observed between
blood sources for sodium) but exhibit similar analyti-
cal goals to VEN and, further, demonstrate acceptable
sensitivity to detect dehydration, in order to serve as an
adequate substitute for VEN.
Materials and methods
Informed written consent was obtained as approved by the Georgia
Tech Institutional Review Board for all participants across two differ-
ent study protocols.
Study A
Forty-two adults (20 men, 22 women) between 18 and 45 years
of age participated by remaining sedentary and consuming con-
trolled, packaged meals for 48 h. No exercise was permitted during
the course of the study (48 h). During the first 24 h, beverages were
consumed ad libitum; however, during the last 24 h, beverages were
provided to ensure that subjects met a total water intake equivalent
to dietary reference values (men: 3.7 L, women: 2.7 L) [1]. Food and
beverage records were obtained to ensure total water intake was met
as prescribed.
Subjects reported to the laboratory at the end of each 24 h
period (16:00 h). After sitting for 10 min with limb immobilized, a
blood sample was obtained by venipuncture into a 4 mL lithium-
heparin tube (BD, Franklin Lakes, NJ, USA). Since pre-conditions
precluded exercise, the peripheral circulation was not vasocon-
stricted nor were any subjects experiencing edema in the limbs.
Next, subjects submerged a digit on the non-dominant hand in hot
water for 5 min before a lancet (1.8 mm Safety Lancet, Fisherbrand,
Houston, TX, USA) was used to obtain a sample drawn into 0.3 mL
lithium-heparin microvettes (Sarstedt, Newton, NC, USA). Finger
stick sampling (avoiding the thumb) was performed according to
procedures previously described [9]. The digit was heated to pro-
mote free flowing capillary blood, preventing extreme tissue com-
pression and potential hemolysis. As a result, there is potential for
the samples to more closely resemble arterialized blood, but reflects
best practices [9]. Microvettes were not always completely filled if
the blood sample could not be obtained without excessive tissue
compression. If insufficient sample was obtained (less than half of
a tube), a different digit was warmed and the process was repeated.
The finger stick sampling took approximately 1 min to fill, on aver-
age, two microvette tubes per subject. Whole blood was immediately
centrifuged for 10 min to separate plasma from cells and processed
within 30 min at room temperature. Samples were not refrigerated or
frozen prior to measuring osmolality.
POsm was analyzed (10 μL samples) using freezing-point depres-
sion (Osmette II, Precision Systems, Natick, MA, USA) by the same
technician. The osmometer was calibrated at the beginning of the
study and daily quality control was performed using 290 mOsm/kg
solution (ClinitrolTM
290 Reference Solution, Norwood, MA, USA) to
ensure accuracy (minimum of three samples within ±2 mOsm/kg).
The order of sampling between CAP and VEN samples was random
during a batch run, but three values from each sample were run con-
secutively prior to switching.
The median of triplicate measures was determined unless the
range of three values exceeded 1% (n = 7 for VEN, n = 10 for CAP) [3, 10].
Coefficients of variation (CV) were determined for analytical (CVa
; 10
samples of 290 mOsm/kg control solution), within-subject (CVi
) or the
variance of repeated sampling on the same subject, and between-sub-
ject (CVb
) variations [11]. An index of individuality (II; [(CVa
2
+CVi
2
)1/2
/
CVb
]) and reference change value (RCV; [21/2
× Z × (CVa
2
+CVi
2
)1/2
] where
Z = 1.96 for unidirectional approach with 95% confidence) were also
calculated [3, 11]. Previously published analytical goals [3, 11–13]
dictate that II should fall between 0.6 (high individuality) and 1.4
(low individuality) [12]. Further, the RCV is only valid if CVi
was not
heterogeneous. To test this, the index of heterogeneity (IH; [CV of:
(SDa
2
+SDi
2
)1/2
]) was divided by the theoretical CV: [2/(n–1)1/2
]. As CAP
and VEN had similar mean determinations per subject (CAP: 3.1, VEN:
3.3), a simplified mean of 3 was used, making the theoretical CV equal
to one. The IH should be less than: {1+2[1/(2n)1/2
]}, or 1.82, when n = 3
[3, 11].
Means for CAP and VEN were analyzed with a two-factor (gen-
der × method) analysis of variance (ANOVA). The mean difference by
sex was also compared with an independent t-test. The relationship
between CAP and VEN was analyzed using a Pearson’s product-
moment correlation coefficient. All analyses were performed with
SPSS version 19 (IBM, Armonk, NY, USA). The level of significance
was established a priori as p ≤ 0.05. A power analysis using G*POWER
version 3.1 (www.gpower.hhu.de) was performed retrospectively to
determine if our sample size was adequate to find differences. All
data are presented as mean±SD.
Study B
A different pool of 46 college students (23 men, 23 women) were
examined over two consecutive 24 h conditions of hydration status:
adequate water intake (euhydrated) followed by fluid restriction
(FR; dehydrated). Additionally, exercise was not permitted beyond
activities of daily living throughout the 48 h period. The participants
restricted their activity to the minimum (i.e., walking to class). For
euhydration, subjects were encouraged to drink additional fluids
above ad libitum food and drink (but refrain from alcohol consump-
tion). After 24 h euhydration, fluid restriction was begun for 24 h. In
consultation with a registered dietitian, the subjects ate a prescribed
diet of foods with low water content (i.e., no fresh fruits, vegetables,
soups, rice) and restricted the ingestion of liquids. For the euhydra-
tion and FR days, subjects arrived at each session without eating or
drinking within at least the previous 3 h. Subjects maintained a food
and beverage log during the 24 h euhydration and fluid restriction.
Total water from food and beverages was calculated using the Food
Processor software version 10.6 (ESHA Research, Salem, OR, USA).
At the end of 24 h over both euhydration and FR days, subjects
arrived at 16:00 h for a CAP sample and body mass measurement.
Nude body mass was measured on a Pennsylvania 50 digital Scale
(Wiggins Scale Co., Atlanta, GA, USA). POsm was analyzed using the
same methods described above.
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

3. Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood 3
CAP and VEN was significantly (p < 0.01) lower (by 1.6±3.2
mOsm/kg) and the correlation between VEN-CAP was
increased to r = 0.51 (p = 0.001). Using this mean differ-
ence and a slightly higher SD (3.7) in a retrospective power
analysis, we found that an n = 43 would result in a large ES
(0.8). Thus, this study was adequately powered to detect
mean differences between blood sources.
The Bland-Altman plot (Figure 2) also identifies sub-
jects by sex. POsm obtained through CAP and VEN did
not differ for men (284.5±4.4 vs. 285.1±3.0 mOsm/kg) or
women (281.3±2.5 vs. 283.3±2.8 mOsm/kg) (p > 0.05); con-
sequently, mean differences in CAP were not significantly
(p = 0.30) lower for women (–2.0±3.2) compared to men
(–0.6±5.2 mOsm/kg). Of note, the three outlying com-
parisons (differences of ±9 mOsm/kg during euhydration)
were predominantly men on the high and low end of POsm
during euhydration, respectively. However, when the data
was averaged across techniques, there was a significant
effect for sex with greater POsm for men (284.8±3.7) com-
pared to women (282.3±3.4 mOsm/kg; p < 0.01). This is
interesting since total water intake (food and beverages)
was controlled for men and women according to dietary
reference intakes [1] and did not different relative to body
mass (44.8±4.1 and 43.8±5.5 mL/kg, respectively).
Table 1 indicates the CVa
, CVi
, and CVb
, II, and IH for
CAP compared to VEN. Analytical variation (CVa
), CVi
, and
CVb
was similar between CAP and VEN, and both were
similar to previously published reference values [3]. Both
CAP and VEN satisfied the analytical goal of II (between
A logistic regression model was fitted to POsm and the predic-
tive accuracy of the model reported using area under the receiver
operating characteristic curve (AUC). Classification tables from the
logistic regression model predicted sensitivity (% correctly classified
as dehydrated) and specificity (% correctly classified as euhydrated).
The criterion value used to classify dehydration versus euhydra-
tion was determined by providing the highest sensitivity without
sacrificing specificity below 80%. Decisions based on maintaining
specificity and sensitivity above 80% have been utilized in previous
dehydration studies [3].
Results
Study A
During euhydration for all subjects, the difference in POsm
with VEN (284.2±3.5) compared to CAP (283.0±3.9) did not
meet statistical significance (p = 0.07). Figure 1 illustrates
the association between the techniques were significantly
correlated (r = 0.31; p < 0.05). However, the correlation was
not relatively strong, in part due to homogenous values
within a rather narrow range as a result of the controlled
euhydration intervention. A lower CAP value was not
observed consistently since approximately 24% of indi-
vidual CAP values were greater than the corresponding
values obtained with VEN.
Individual differences between the criterion method
(VEN) and CAP are illustrated in the Bland-Altman plot
(Figure 2). Three values fell outside ±1.96 SD. When elimi-
nating these outliers (n = 3), the mean difference between
300
295
290
285
275
0
-0
275
280
285
290
295
300
280
CAPPOsm,mOsm/kg
VEN POsm, mOsm/kg
r=0.313
p<0.05
Figure 1 Correlation between plasma osmolality (POsm) obtained
by capillary puncture blood (CAP) and venous blood (VEN) under
controlled euhydration conditions (n = 42, Study A).
The dashed line denotes the line of identity. Note: closed circles
indicate more than one observation.
15
Men
Women10
5
-5
-15
-0
275
280
290
285
295
-10
0
POsmCAP-POsmVEN,mOsm/kg
(POsmCAP+POsmVEN)/2, mOsm/kg
Figure 2 Bland-Altman plot showing agreement between capillary
puncture blood (CAP) and venous blood (VEN) determinations of
plasma osmolality (POsm). Men (n = 22) are depicted as squares and
women (n = 20) as circles.
Dashed lines indicate mean and ±1.96 SD (95% confidence interval).
Note closed symbols indicate more than one observation.
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

4. 4 Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood
0.6 and 1.4). The RCV was also similar across CAP, VEN,
and reference values. As both CAP and VEN IH were < 1.82,
this RCV was considered valid. CAP for both men and
women also had similar CVi
, CVb
, and RCV, satisfying the
analytical goals for IH and II (Table 1). All analytical goals
were also met when the outliers (n = 3) were eliminated
from the analysis.
Study B
Total water intake was significantly greater (p < 0.05)
during 24 h of euhydration (3.62±1.42 L) compared to
24 h of fluid restriction (0.25±0.18 L), averaging 52.5±19.0
versus 3.9±2.7 mL of water per kg of body mass, respec-
tively. Mean Δ in body mass from euhydration to fluid
restriction for all subjects was –2.1%±0.7%; thus, 24 h
fluid restriction resulted in mild dehydration.
Following 24 h of fluid restriction, POsm increased
significantly (p < 0.001) by 9 mOsm/kg (from 284.0±4.4
to 292.8±5.2 mOsm/kg), as illustrated in the box plot
(Figure 3). Minimum and maximum values ranged from
276 to 292 mOsm/kg during euhydration and 285 to 307
mOsm/kg during fluid restriction. The AUC analysis which
represents the accuracy to detect dehydration from our
prospective fluid restriction study was 0.92, a value con-
sidered excellent for a diagnostic test ( > 0.90). Accuracy
was only slightly lower than previously observed (0.95) to
predict dehydration with POsm using venipuncture [3].
In the present study, the ability to predict dehydration
(sensitivity) was 89% and correctly classify euhydration
Table 1 Analytical (CVa
), within-subject (CVi
), and between subject
(CVb
) coefficient of variation, indexes of individuality (II) and hetero-
geneity (IH), and reference change values (RCV) of plasma osmolal-
ity using capillary, venous blood (Study A), and reference values
using venous blood [3].
Capillary blood
(men, women)
Venous blood
(men, women)
Reference
venous
blood [3]
Study A
CVa
, % 0.4 0.4 0.4
CVi
, % 1.2 (1.4, 0.7) 0.9 (1.0, 0.8) 1.3
CVb
, % 1.4 (1.6, 0.9) 1.2 (1.0, 1.4) 1.5
II 0.90 (0.91, 0.90) 0.70 (1.08, 0.64) 0.90
IH 1.45 (1.60, 1.11) 1.35 (1.41, 1.29) 1.35
RCV, % 3.2 (3.6, 2.0) 2.5 (2.7, 2.2) 3.1
RCV, mOsm/kg 9 (10, 6) 7 (8, 6) 9
Sex-specific measures are also included for men (n = 22) and women
(n = 20). Reference venous blood [3] was based on n = 18 subjects (13
men, 8 women).
Plasmaosmolality,mOsm/kg
Truepositive,sensitivity
Euhydration Dehydration
310 1.0
0.8
0.6
0.4
0.2
0
0 0.2 0.4
False positive (1-specificity)
0.6 0.8 1.0
305
300
295
290
285
275
0
280
Figure 3 Box plot of plasma osmolality (POsm) measured with cap-
illary puncture blood during euhydrated and fluid restriction (mild
dehydration) conditions.
Black square indicates mean for each condition and open circles
identify outliers. The box represents the 25th, 50th (median), and
75th percentiles scores, respectively. The bars indicate the inter-
quartile range (95% of data). The insert displays the area under the
receiver operator characteristic (ROC) curve for capillary puncture
blood POsm.
(specificity) was 80% compared to 90% and 100%, respec-
tively, in the previously referenced study [3]. The criterion
value that classified “euhydrated” from “dehydrated” was
288 mOsm/kg.
Discussion
The ability to accurately classify an individual’s hydra-
tion status at a specific point in time is valuable informa-
tion for a variety of settings both clinical and in the field.
Although easily obtainable measures of body mass and/
or urine concentration can track hydration status changes
over time [14, 15], urinary changes may lag POsm during
progressive acute dehydration [16, 17]. There is no univer-
sal agreement on the “gold standard” to predict hypertonic
hypovolemia [18], but POsm appears to provide the best
single point assessment to classify hydration status [3, 5].
To our knowledge, this is the first study to demonstrate
excellent diagnostic accuracy and sensitivity with POsm
using CAP to detect dehydration coupled with favorable
comparisons to VEN-derived POsm based upon analytical
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

5. Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood 5
goals and RCVs. Collectively, these data indicate CAP can
be substituted for VEN when measuring POsm to assess
hydration status.
CAP has been previously validated against VEN for
several other biochemical and hematological markers
[19]. Although CAP is not a homologous source of blood
unlike VEN [6], the observed difference of < 2 mOsm/kg
(0.5%) between CAP and VEN was consistent with our
hypothesis and smaller than other biomarkers, such as
hemoglobin, hematocrit, and neutrophils (2.1%) [19]. In
40 subjects, serum sodium obtained from skin puncture
was significantly lower than venipuncture by 1.9 mmol,
and this difference was deemed “clinically unimportant”
[6].
As POsm was the only biomarker of hydration status
which satisfied analytical goals for IH and II compared to
other body fluids [3], it was important that CAP also ful-
filled these criteria. This was accomplished in the present
study for the total group as well as subdividing by men
and women. Each index provides useful assessment to
further understand the variability of the measure: meeting
the II ensured that POsm via CAP exhibited appropriate
distribution of the CVi
and CVb
, and the IH indicated the
probability of false alarms is not > 5% [11, 12]. The RCV of
9 mOsm/kg was similar to the actual mean change dem-
onstrated with fluid restriction, and is in accordance with
previous RCVs for POsm obtained with VEN [3]. A mean
change of approximately 9 mOsm/kg previously indicated
that dehydration was 95% likely [4].
Although the difference in mean POsm may, in fact, be
lower with CAP (between 1.2 and 1.6 mOsm/kg or < 0.5%),
the CVi
and CVb
were similar to VEN. The presence of a few
spurious outliers (Figure 2) with marked discrepancies
cannot be directly attributed to any one factor (i.e., pre-
analytical factors or greater variation within these specific
samples). The fact that these outliers were primarily men
is of interest, but is not believed related to any inherent
biological sex difference in POsm. The POsm in women
may fluctuate according to their monthly cycle [20] and
when combined with a similar relative total water intake
may explain the slightly lower values compared to men
observed in the present study; however, this should not
affect the difference across sampling methods per se.
Therefore, we re-examined the mean differences without
these outliers and indeed found CAP versus VEN to yield a
small underestimate, which is consistent with reports on
comparisons with plasma or serum sodium [6, 7].
Another key finding was that POsm using CAP to
predict mild dehydration for women and men resulted
in excellent accuracy (0.92) as a diagnostic test based on
the AUC for the receiver operating characteristic curve.
Sensitivity was also similar (approx. 90%) to reported
values for VEN [3], although yielding a lower specificity
(80%). As a capillary blood source is a mixture of arterial
blood, VEN, and interstitial fluid [6], a similar sensitivity
to VEN indicates the heterogeneous mixture of blood did
not confound tracking POsm changes to detect dehydra-
tion. However, our test protocol yielded a lower predictive
accuracy to assure individuals were euhydrated as illus-
trated by the overlap between POsm values in Figure 3.
This may partly be attributed to the tight defense of POsm
through the kidney’s regulation of water balance [21], and
as such may not always track water conservation efforts
during mild dehydration [18]. It also remains difficult to
designate a specific value for euhydration as reviewed
elsewhere [22]. Reference values for “euhydration” in men
(not involved in daily exercise over 12 days) were 289–291
mOsm/kg based upon morning serum osmolality, repre-
senting the 41–60th percentile in their subjects [1]. The
use of additional biomarkers (i.e., body mass change,
urine specific gravity) combined with POsm may improve
diagnostic accuracy [23]. CAP, therefore, removes many of
the practical limitations of obtaining blood, making more
complete evaluation profiles possible.
Although our findings show a close relationship
between CAP and VEN, we acknowledge other factors
which should be considered in applying this method.
In the current study, the state of peripheral circulation
was standardized through seated rest and limb immobi-
lization for 10 min, followed by warming the digit. There
were no signs of vasoconstriction or edema in the periph-
eral circulation and the incidence of hemolysis in CAP
samples was not different than VEN ( < 3 each overall). In
situations with uncontrolled pre-analytical conditions,
CAP may yield slightly different results. Second, CAP
was analyzed on an osmometer requiring plasma sample
sizes of only 10 μL, providing advantages to allow for
a median based upon 3–5 values. It is unclear whether
“excessive” tissue compression, in order to obtain suf-
ficient blood samples for specific osmometers, repre-
sents a potential trade-off impacting test reliability. To
illustrate this, in a subset of samples (n = 24) collected
under euhydration and analyzed with a 50 μL osmom-
eter (μOsmette, Precision Systems, Natick, MA, USA), the
variability appears similar for CVi
(0.6, 0.7) and CVb
(1.3,
1.6) for CAP and VEN, respectively; however, a slightly
greater mean difference (–2.0±3.1 mOsm/kg, p < 0.01).
Whether even larger sampling sizes would alter the dis-
crepancy is uncertain since Cheuvront et al. [8] observed
whole blood and POsm differed significantly in smaller
20 μL samples (by 10±3 mOsm/kg) but less in 250 μL
samples (3±2 mOsm/kg).
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

6. 6 Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood
Another potential limitation in the present study is
our “lower” criterion value (288 mOsm/kg) relative to the
literature based on VEN [3, 24], although we are unaware
of other comparative studies that have performed either a
diagnostic accuracy of VEN POsm using a resting dehydra-
tion protocol or using CAP. Passive, resting conditions of
dehydration may generate a lower RCV in osmolality than
exercise-induced dehydration [24]. Most free living older
community dwellers (based on National Health and Nutri-
tion Examination Survey data) have POsm between 285
and 295 mOsm/kg when reportedly consuming > 3 L of
fluid a day [25]. Yet, serum osmolality across a wide range
of fluid intakes (up to and exceeding 3 L/day) is report-
edly stable at 279–281 and 276–278 mOsm/kg for men and
women < 50 years of age [1]. In contrast, Bohnen et al. [10]
indicate in healthy, “well-hydrated” individuals, POsm
rarely deviate by > 1%–2% from basal values of approxi-
mately 287 mOsm/kg, which appear congruent with our
values. Third, our POsm were performed on fresh samples
(not refrigerated or frozen), centrifuged without delay, and
were obtained on subjects at 16:00 h, which might result
in different criterion values compared to previous investi-
gations measuring first morning values after an overnight
fast [3]. In our studies, fluids were not consumed within
3 h of testing which is another consideration since the
timing of fluid ingestion may not always correlate with
the dilution in POsm in all individuals [26]. Therefore,
without additional resting studies as confirmation, it may
be premature to utilize our criterion value per se to classify
hydration status.
In summary, we conclude that micro-sampling
determination of POsm obtained through CAP is a valid
alternative to VEN in adult men and women. It provides
advantages under conditions when repeated sampling
is required to assess dehydration with the additional
benefits of less blood loss, reduced need for phlebotomy
training, and typically greater acceptability by subjects.
The accuracy of CAP POsm may potentially permit a
single time point identification of hypertonic hypov-
olemia or aid interpretation of hydration status when
combined with other biomarkers easily obtained in the
field for athletes and military personnel outside of clini-
cal settings.
Acknowledgments: The authors appreciate the editorial
comments of Dr. Michael N. Sawka in developing the man-
uscript and Michael L. Jones and Namrita Kumar in assist-
ing with the data collection. The study was supported, in
part, with funding obtained by The Coca-Cola Company,
Atlanta, GA. There are no conflicts of interest for any of the
authors in regards to this study.
Author contributions: All the authors have accepted
responsibility for the entire content of this submitted
manuscript and approved submission.
Financial support: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
Competing interests: The funding organization(s) played
no role in the study design; in the collection, analysis, and
interpretation of data; in the writing of the report; or in the
decision to submit the report for publication.
References
1. Institute of Medicine. Dietary reference intakes for water,
potassium, sodium, chloride, and sulfate. Washington, DC:
National Academies Press, 2004.
2. Sawka MN, Burke LM, Eichner ER, Maughan RJ, Montain SJ,
Stachenfeld NS. American College of Sports Medicine posi-
tion stand. Exercise and fluid replacement. Med Sci Sport Exer
2007;39:377–90.
3. Cheuvront SN, Ely BR, Kenefick RW, Sawka MN. Biological
variation and diagnostic accuracy of dehydration assessment
markers. Am J Clin Nutr 2010;92:565–73.
4. Cheuvront SN, Fraser CG, Kenefick RW, Ely BR, Sawka MN. Refer-
ence change values for monitoring dehydration. Clin Chem Lab
Med 2011;49:1033–7.
5. Cheuvront SN, Kenefick RW, Charkoudian N, Sawka MN. Physi-
ologic basis for understanding quantitative dehydration assess-
ment. Am J Clin Nutr 2013;97:455–62.
6. Blumenfeld TA, Hertelendy WG, Ford SH. Simultaneously
obtained skin-puncture serum, skin-puncture plasma, and
venous serum compared, and effects of warming the skin before
puncture. Clin Chem 1977;23:1705–10.
7. Loughrey CM, Hanna EV, McDonnell M, Archbold GP. Sodium
measurement: effects of differing sampling and analytical meth-
ods. Ann Clin Biochem 2006;43:488–93.
8. Cheuvront SN, Kenefick RW, Heavens KR, Spitz MG. A compari-
son of whole blood and plasma osmolality and osmolarity. J Clin
Lab Anal 2014;28:368–73.
9. McCall RE, Tankersley CM. Phlebotomy essentials, 4th ed. Phila-
delphia: Lippincott Williams & Wilkins, 2007.
10. Bohnen N, Terwel D, Markerink M, Ten Haaf JA, Jolles J.
Pitfalls in the measurement of plasma osmolality pertinent
to research in vasopressin and water metabolism. Clin Chem
1992;38:2278–80.
11. Fraser CG, Harris EK. Generation and application of data on
biological variation in clinical chemistry. Crit Rev Clin Lab Sci
1989;27:409–37.
12. Harris EK. Effects of intra- and interindividual variation on the
appropriate use of normal ranges. Clin Chem 1974;20:1535–42.
13. Fraser CG, Hyltoft Peterson P, Larsen ML. Setting analytical
goals for random analytical error in specific clinical monitoring
situations. Clin Chem 1990;36:1625–8.
14. Armstrong LE, Soto JA, Hacker FT, Jr., Casa DJ, Kavouras SA,
Maresh CM. Urinary indices during dehydration, exercise, and
rehydration. Int J Sport Nutr 1998;8:345–55.
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM

7. Wittbrodt et al.: Variability of plasma osmolality with capillary versus venous blood 7
15. Cheuvront SN, Carter R, 3rd, Montain SJ, Sawka MN. Daily body
mass variability and stability in active men undergoing exercise-
heat stress. Int J Sport Nutr Exerc Metab 2004;14:532–40.
16. Oppliger RA, Magnes SA, Popowski LA, Gisolfi CV. Accuracy of
urine specific gravity and osmolality as indicators of hydration
status. Int J Sport Nutr Exerc Metab 2005;15:236–51.
17. Popowski LA, Oppliger RA, Patrick Lambert G, Johnson RF,
Kim Johnson A, Gisolf CV. Blood and urinary measures of hydra-
tion status during progressive acute dehydration. Med Sci Sport
Exer 2001;33:747–53.
18. Armstrong LE, Maughan RJ, Senay LC, Shirreffs SM. Limitations
to the use of plasma osmolality as a hydration biomarker. Am J
Clin Nutr 2013;98:503–4.
19. Nunes LA, Gandra PG, Alves AA, Kubota LT, Vaz de Macedo D.
Adequacies of skin puncture for evaluating biochemical and
hematological blood parameters in athletes. Clin J Sport Med
2006;16:418–21.
20. Stachenfeld NS, Splenser AE, Calzone WL, Taylor MP, Keefe DL.
Sex differences in osmotic regulation of AVP and renal sodium
handling. J Appl Physiol 2001;91:1893–901.
21. Perrier ET, Armstrong LE, Daudon M, Kavouras S, Lafontan M,
Lang F, et al. From state to process: defining hydration. Obes
Facts 2014;7(Suppl 2):6–12.
22. Armstrong LE. Assessing hydration status: the elusive gold
standard. J Am Coll Nutr 2007;26(5 Suppl):575S–84S.
23. Armstrong LE, Johnson EC, Munoz CX, Le Bellego L, Klein A,
McKenzie AL, et al. Evaluation of Uosm:Posm ratio as a hydra-
tion biomarker in free-living, healthy young women. Eur J Clin
Nutr 2013;67:934–8.
24. Munoz CX, Johnson EC, Demartini JK, Huggins RA, McKenzie AL,
Casa DJ, et al. Assessment of hydration biomarkers including
salivary osmolality during passive and active dehydration. Eur J
Clin Nutr 2013;67:1257–63.
25. Stookey JD. High prevalence of plasma hypertonicity among
community-dwelling older adults: results from NHANES III. J Am
Diet Assoc 2005;105:1231–9.
26. Sollanek KJ, Kenefick RW, Cheuvront SN, Axtell RS.
Potential impact of a 500-mL water bolus and body mass
on plasma osmolality dilution. Eur J Appl Physiol 2011;111:
1999–2004.
Authenticated | mwittbrodt3@gatech.edu author's copy
Download Date | 2/15/15 3:24 AM