pH Effects on Measurements of Ionized Calcium and
Ionized Magnesium in Blood
Sihe Wang, PhD; Elizabeth H. McDonnell, MS; Frank A. Sedor, PhD; John G. Toffaletti, PhD
Context.It is well known that the concentration of ion-
ized calcium in blood is affected by the pH of the speci-
men, since hydrogen ions compete with calcium for bind-
ing sites on albumin and other proteins. However, the re-
lationship between pH and ionized magnesium concentra-
tion is not as well characterized.
Objective.To determine the effects of pH on ionized
magnesium concentration over a wide range of pH values
in serum or plasma.
Design.Both ionized calcium and ionized magnesium
concentrations were measured in 3 sets of samples. (1)
Pools of serum or whole blood at different pH values
(7.207.60) achieved by adding a constant volume of acid
or base (diluted solutions of either hydrochloric acid or
sodium hydroxide) plus saline. These pools consisted of 2
serum and 3 heparinized whole blood pools collected from
leftover blood remaining in clinical specimens in the Clin-
ical Chemistry and Blood Gas Laboratories, respectively, at
Duke University Medical Center. (2) Five whole blood
C alcium (Ca) and magnesium (Mg) have many impor-
tant physiologic functions, and measurement of ion-
ized (free) Ca and Mg concentrations in plasma provides
useful information for clinical diagnosis and management.
Approximately 40% of Ca1 and 30% of Mg2 are bound to
proteins in regular human plasma. Changes of pH in the
specimen affect the binding of these ions to plasma pro-
teins, mainly albumin, because hydrogen ions compete
with Ca and Mg for protein binding sites. Elin and col-
leagues3 studied the effect of pH on ionized Mg test re-
sults by analyzing anaerobic serum samples from patients
and reanalyzing those samples after pH was increased by
in vitro loss of carbon dioxide (CO2). Their findings dif-
fered according to the analytic system used: the NOVA
CRT system (NOVA Biomedical, Waltham, Mass) showed
a decreased ionized Mg concentration with increased pH,
Accepted for publication March 19, 2002.
From the Department of Pathology/Clinical Laboratories, Duke Uni-
versity Medical Center, Durham, NC. Dr Wang is currently with the
Department of Medicine, Northwestern University Medical Center,
Chicago, Ill.
Dr Toffaletti receives research support from Roche AVL Scientific. Dr
Sedor receives a grant from Ortho Clinical Diagnostics.
Presented in part at the American Association for Clinical Chemistry/
Canadian Society of Clinical Chemists 2001 Annual Meeting & Clinical
Lab Exposition, Chicago, Ill, August 1, 2001.
Reprints: John G. Toffaletti, PhD, Department of Pathology/Clinical
Laboratories, Duke University Medical Center, Durham, NC 27710 (e-
mail: toffa002@mc.duke.edu).
Arch Pathol Lab MedVol 126, August 2002
specimens obtained from apparently healthy individual do-
nors. (3) Twenty-six whole blood specimens obtained from
individual patients (leftover blood from the Blood Gas Lab-
oratory) in which pH was varied by in vitro loss or gain of
carbon dioxide.
Results.Both ionized calcium and ionized magnesium
concentrations decreased as the pH in the specimen in-
creased, indicating the stronger binding of these ions with
proteins in the more alkaline environment.
Conclusion.We conclude that the rate of change of
ionized magnesium concentration with pH change (0.12
mmol/L per pH unit) is significantly less than that of ion-
ized calcium (0.36 mmol/L per pH unit). Furthermore, our
findings indicate that if adjustment to pH 7.40 is necessary,
the ionized magnesium test results need to be adjusted
when pH is markedly abnormal, as is sometimes done for
ionized calcium.
(Arch Pathol Lab Med. 2002;126:947950)
whereas the AVL system (988-4 Electrolyte Analyzer, AVL
Roche Diagnostics, Roswell, Ga) indicated that ionized Mg
concentration was not affected by an increase in serum
pH of up to even 0.6 units from the baseline value (7.37
7.52). Ising et al4 also studied the influence of pH on ion-
ized Mg test results determined with the Microlyte Mag-
nesium system (Kone Instrument, Finland) in serum, and
showed a constant exponential factor x in the Siggard-
Anderson equation in predicting ionized Mg concentra-
tion (at a given pH) from ionized Mg concentration (at pH
7.40).
Despite these studies, there is a lack of data showing
the changes of ionized Mg concentration in blood over the
most common physiologic pH range (7.207.60) and the
relationship of these changes to albumin/total protein
concentration in the specimen. The aim of this study was
to determine the changes of ionized Mg concentration
with changes of pH in serum or whole blood, and to es-
tablish the relationship between ionized Mg concentration
and pH in routine specimens encountered in clinical lab-
oratories.
MATERIALS AND METHODS
We obtained patient samples from leftover blood destined for
discard after testing in the Blood Gas Laboratory, and none of
our results were linked to the patients in any way. Therefore, our
study was exempt from institutional review board review. All of
the specimens were obtained by random selection.
We collected 2 serum pools and 3 heparinized whole blood
pH Effects on Ca and Mg MeasurementWang et al 947
 Effects of pH on Concentrations of Ionized Ca (iCa) and Ionized Mg (iMg) in Serum and Whole Blood (WB)
Mean for Serum Pool Mean for WB Pool Mean for WB Donor
(n 5 2)
pH range 7.127.65
Total protein, g/L 67
Albumin, g/L 33
Protein-bound Ca, mmol/L 2.1
Protein-bound Mg, mmol/L 0.80
DiCa/pH, mmol/L 20.48
DiMg/pH, mmol/L 20.15
DiCa/pH per gram per liter of albumin 20.015
DiMg/pH per gram per liter of albumin 20.0045
DiCa/pH per gram per liter of total protein 20.0072
DiMg/pH per gram per liter of total protein 20.0022
pools from the extra blood remaining in clinical specimens at the
Clinical Chemistry and Blood Gas Laboratories, respectively, at
Duke University Medical Center. We also obtained 5 whole blood
specimens from apparently healthy individual donors. After each
pool or specimen was collected and mixed well, it was divided
into 2-mL aliquots in 3-mL syringes (Becton Dickinson, Ruther-
ford, NJ). The pH in each aliquot was adjusted by adding solution
containing 100 mmol/L of sodium hydroxide or hydrochloric
acid. The total volume added to every 2-mL aliquot was kept at
200 mL by addition of an appropriate volume of saline.
We also obtained 26 individual patient specimens (heparinized
whole blood) from extra blood remaining in the Blood Gas Lab-
oratory. The pH of each specimen was changed by altering the
CO2 content, either by exposing the specimen to air to increase
pH or by introducing CO2 gas (30% in nitrogen, National Spe-
cialty Gases, Durham, NC) to decrease pH.
We measured concentrations of ionized Ca and ionized Mg and
pH with the AVL 988-4 Electrolyte Analyzer (AVL Roche). This
instrument was designated for research use only. Calibration and
quality control samples were run immediately before the speci-
mens were analyzed for each experiment in accordance with the
manufacturers instructions. The analyzer automatically adjusts
the ionized Mg concentration result for the effects of sodium and
ionized Ca concentrations, which affect the response of the ion-
ized Mg-sensing electrode.5 Quality control materials were run
each day to ensure acceptable performance of the instrument. The
concentrations of total protein, albumin, and other analytes in
the specimens were measured with the Vitros 950 analyzer (John-
son & Johnson, Rochester, NY).
RESULTS
To determine the pH effects (targeted range, 7.207.60)
on the measurement of ionized Ca and ionized Mg con-
centrations, we completed 10 experiments using different
specimens (2 serum pools, 3 whole blood pools, and 5
whole blood specimens from individual donors), and the
results are shown in the Table. All of these experiments
showed some degree of pH effects on either ionized Ca
or ionized Mg concentration; concentrations of both ions
decreased with increasing pH of the specimen. The mean
change in ionized Ca concentration per pH unit was 0.50
mmol/L, which was significantly greater than that for ion-
ized Mg concentration (0.16 mmol/L) (P , .001, paired
Student t test, 1-tailed analysis). After normalizing to the
percent changes in concentrations of the ionized forms rel-
ative to the total Ca and Mg concentrations, respectively,
we found that the rate of change of concentration was still
higher for ionized Ca than for ionized Mg (26% vs 18%
per pH unit) (P , .005, paired Student t test, 1-tailed anal-
ysis) (Table).
The albumin and total protein concentrations were
markedly different in the different samples studied (Ta-
948 Arch Pathol Lab MedVol 126, August 2002
Overall Mean
(n 5 3) (n 5 5) (n 5 10)
7.197.61 7.217.56 7.197.59
49 66 61
21 35 31
1.7 2.0 1.9
1.14 0.78 0.89
20.32 20.61 20.50
20.13 20.19 20.16
20.015 20.018 20.016
20.0064 20.0053 20.0053
20.0067 20.0093 20.0082
20.0028 20.0028 20.0027
ble); not unexpectedly, they were highest in the whole
blood samples from healthy donors and lowest in the
pooled whole blood samples from patients in surgery or
critical care. To determine if albumin binding was the
cause of the pH effects on ionized Ca and ionized Mg
concentrations, we analyzed the changes in concentrations
of these ions per pH unit against the different albumin
concentrations of the 10 specimens studied (Figure 1). We
found that (1) the pH effect on ionized Ca concentration
was related to the albumin concentration in the specimen,
with greater changes occurring in the specimens with
higher concentrations of albumin, and (2) the influence of
albumin concentration on the pH dependency of ion con-
centration was less for ionized Mg (0.005 mmol/L per
gram per liter of albumin) than for ionized Ca (0.015
mmol/L per gram per liter of albumin).
To further show and establish a relationship between
ionized Mg concentration and pH in a wide range of rou-
tine specimens encountered in blood gas laboratories, we
measured ionized Mg and Ca concentrations in 26 hepa-
rinized whole blood specimens in which pH changes were
made by altering the CO2 content. As shown in Figure 2,
both ionized Mg and ionized Ca concentrations increased
as pH decreased in the specimens, with a significantly
greater change for ionized Ca concentration (slope, 20.36)
than for ionized Mg concentration (slope, 20.12) (P ,
.001, ANOVA regression test).
COMMENT
We first studied the pH effects on ionized Mg and ion-
ized Ca concentrations by adding diluted hydrochloric
acid or sodium hydroxide solutions (Table). In agreement
with the findings of Elin et al,3 the results for ionized Ca
concentration decreased with increasing pH in the speci-
mens. However, contrary to their findings regarding the
pH effect on ionized Mg concentration as measured with
the AVL analyzer, we found that the ionized Mg concen-
tration also decreased with increasing pH of the specimen.
To avoid changing the ionic strength of the specimen,
which may affect the electrode response to ionized Mg
and/or Ca, we also altered the pH of 26 patient specimens
by varying CO2 content in the specimens. We were able to
confirm that ionized Mg concentration, as well as ionized
Ca concentration, decreased with increasing pH.
Our findings also indicate the stronger binding of Ca
and Mg with proteins, mainly albumin, in a more alkaline
environment. From the results shown in the Table, albu-
min concentration has less of an influence on the pH de-
pendency of ionized Mg concentration than on the pH
pH Effects on Ca and Mg MeasurementWang et al

dependency of ionized Ca concentration (Figure 1), indi-
cating that the binding of Mg with albumin in the speci-
men is less sensitive than that of Ca to pH change.
Elin and colleagues3 showed the effect of pH on ionized
Mg concentrations by analyzing serum samples from pa-
tients after pH had been increased by in vitro loss of CO2.
Results obtained with the NOVA CRT showed a decrease
in ionized Mg concentration with increasing pH, whereas
results obtained with the AVL 988-4 analyzer were unaf-
fected by an increase in serum pH of up to 0.6 pH units
from the baseline value (7.377.52). In contrast, we found
Arch Pathol Lab MedVol 126, August 2002
Figure 1. Effect of pH dependency of ion-
ized calcium (iCa) and ionized magnesium
(iMg) concentrations related to albumin con-
centration for 10 specimens described in the
Table. Slope of DiCa/pH, 0.0153; slope of
DiMg/pH, 0.0037.
Figure 2. Concentrations of ionized mag-
nesium (iMg) (slope, 20.12) and ionized cal-
cium (iCa) (slope, 20.36) in relation to pH
changes induced by gain or loss of CO2 in
the heparinized whole blood of 26 patients.
that ionized Mg concentration measured with the AVL
988-4 analyzer was affected by a change in pH in the
range of 7.20 to 7.60, indicating that the relationship may
be similar for the NOVA CRT and AVL analyzers.
The physiologic significance of pH-dependent binding has
never been established for either ionized Ca or ionized Mg.
When pH is abnormal, knowledge of the relationship be-
tween pH and ionized Ca or ionized Mg concentrations
could predict how pH correction would affect these electro-
lyte concentrations. Although knowledge of the relationship
between pH and either ionized Ca or Mg concentration may
pH Effects on Ca and Mg MeasurementWang et al 949
be helpful when acid-base balance is markedly abnormal, no
clinical studies have been attempted to determine if pH-cor-
rected ionized Ca or Mg concentrations are useful for this
purpose.
In conclusion, both ionized Ca and ionized Mg concen-
trations decrease with increasing pH in serum or whole
blood, although changes for ionized Mg are about one
third as great as those for ionized Ca. These decreases are
related to the change in pH and to albumin and total pro-
tein concentrations. Our findings indicate that if pH is
markedly abnormal, the ionized Mg results may need to
be adjusted as is sometimes done for ionized Ca results.
950 Arch Pathol Lab MedVol 126, August 2002
We thank Mr Ray Dean in the Blood Gas Laboratory for his
technical assistance.
References
1. Toffaletti J, Savory J, Gitelman HJ. Use of gel filtration to examine the dis-
tribution of calcium among serum proteins. Clin Chem. 1977;23:23062310.
2. Elin RJ. Assessment of magnesium status. Clin Chem. 1987;33:19651970.
3. Elin RJ, Hristova EN, Cecco SA, Niemela JE, Rehak NN. Comparison of
precision and effect of pH and calcium on the AVL and NOVA magnesium ion-
selective electrodes. Scand J Clin Lab Invest. 1996;56(suppl 224):203210.
4. Ising H, Bertschart F, Gunther T, Jeremias E, Jeremias A. Measurement of
free magnesium in blood, serum and plasma with ion-selective electrode. Eur J
Clin Chem Clin Biochem. 1995;33:365371.
5. AVL Scientific Corporation. 988-4 Electrolyte Analyzer Operators Manual.
3rd ed. Roswell, Ga: AVL Scientific Corporation; 1995:78.
pH Effects on Ca and Mg MeasurementWang et al