Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 1
Journal of Analytical Toxicology, Vol. 34, November/December 2010
Development and Validation of Analytical Methods for
Ultra-Trace Beryllium in Biological Matrices
Vincent Paquette1,2, Pierre Larivière2,*, Diane Cormier2, Ginette Truchon2, Joseph Zayed3, and Huu Van Tra1
1Chemistry
Department, Université du Québec à Montréal (UQÀM), Montréal, Québec, Canada H3C 3P8; 2Institut de recherche
Robert-Sauvé en santé et en sécurité du travail (IRSST), Montréal, Québec, Canada H3A 3C2; and 3Environmental and
Occupational Health Department, Université de Montréal, Montréal, Québec, Canada H3C 3J7
Abstract
Beryllium (Be) is still not well understood from a toxicological
point of view, and studies that involve the determination of
different Be compounds species in tissues need to be conducted. In
this paper we describe the development and validation of reliable
methods for the detection of ultra-trace levels of Be in various
biological matrices. Blood and tissues (liver, lung, spleen, and
kidney) were used in this study. The samples were digested with a
mixture of nitric and perchloric acids for Be and BeAl and an
addition of sulfuric acid was made for BeO. The solutions were
analyzed by inductively coupled plasma mass spectrometry with
6Li as internal standard. The detection limits are in the order of
0.02 ng/g for tissue and 0.03 ng/mL for blood, and were compared
to existing reference methods. To our knowledge, this is the first
study that assesses dissolution of the different Be compounds in
biological matrices, while also undergoing a rigorous optimization
and complete validation. This method has proven that it is reliable,
among the most sensitive available in the literature, and that it can
be used in trace toxicological studies for Be.
Introduction
Beryllium (Be) is a shiny gray metal of the alkaline earth
family. It has an atomic number of 4, a molecular weight of 9.01
g/mol, and an oxidation state of +2. It does not exist in the pure
state in nature. In industry, Be is used mainly in the form of an
oxide (15%), in its metallic form (10%), and in the form of
alloy (75%) (1).
Be is known to increase the hardness of alloys, their resistance to corrosion, and their electrical and thermal conductivity. Be is six times more rigid than steel, and three times
lighter than aluminum. It is resistant to heat deformation and
pressure, has a high heat absorption capacity, and is trans-
* Author to whom correspondence should be addressed: Pierre Larivière, Institut de recherche
Robert-Sauvé en santé et en sécurité du travail (IRSST), 505 de Maisonneuve Ouest, Montréal,
QC, Canada H3A 3C2. Email: [email protected].
562
parent to microwaves and X-rays (1,2). These unique properties
make Be suitable for extensive use in high technology industries
such as aerospace, electronics, nuclear, telecommunications,
and computers, as well as dentistry and ceramics (1).
Exposure to Be can cause sensitization (BeS) in a small percentage of the exposed population. Chronic beryllium disease
(CBD) may develop after subsequent exposure of sensitized individuals. CBD is a progressive granulomatous lung disease
that is characterized by mononuclear cell infiltration and granulomatous inflammation in the lungs (1,3–6). Be toxicity may
depend upon particle properties including the number, concentration, or size of the particles, surface area, and chemical
form (1,7–15).
In order to study the distribution of Be in a target organ, and
depending on the size and nature of Be particles, analytical
methods must be developed. They must take into account the
availability and specificity of the analysis of Be and they must
efficiently dissolve the Be compounds found in biological
matrices.
The main techniques used for the analysis of metals in biological matrices are flame atomic absorption spectroscopy (FAAS), electrothermal atomic absorption spectroscopy (ET-AAS),
inductively coupled plasma optical emission spectroscopy (ICPOES), and inductively coupled plasma mass spectrometry (ICPMS). Of all these techniques, ICP-MS is the one that offers the
best analytical characteristics. In fact, ICP-MS is now widely
used in clinical biology because of its exceptional qualities, including high sensitivity, ability to analyze multiple isotopes of
the same element, high speed of analysis, and its high dynamic
linear range (up to eight orders of magnitude) (16,17).
As yet, no method available in the literature for the biological
matrices that are used (blood, kidney, liver, lungs, and spleen)
has undergone a process of full and rigorous method validation
for different Be species analysis. Obviously, it is quite clear that
the development and validation of reliable analytical methods
for Be detection at trace levels in biological matrices are crucial.
This is the main objective of this study. It is important to note
that urine is a viable biological matrix to assess exposure to Be
(18). Adequate methods are already published (18–21), so this
matrix will not be discussed in this paper.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 2
Journal of Analytical Toxicology, Vol. 34, November/December 2010
Materials and Methods
Reagents
Fine powders of metallic beryllium (99.10%), beryllium
oxide UOX grade (99%+), and beryllium-aluminum alloy
(62.3% Be/37.18% Al) used for the solubilization essay came
from Brush Wellman (Elmore, OH). Claritas-grade beryllium
standard solution (1000 mg/L) used for calibration came from
Spex Certiprep (Metuchen, NJ), and those solutions used for
quality control (10 mg/L) came from SCP Plasma Cal (Baie
d’Urfé, QC). CRM “Seronorm™ Trace elements whole blood L2” came from Accurate Chemical & Scientific (0.59 ± 0.5 μg/L
of Be, Westbury, NY), and interlaboratory materials
(QMEQAS07B-03 at 0.950 ± 0.159 μg/L of Be, QMEQAS06B-02
at 1.74 ± 0.29 μg/L of Be, and QMEQAS06B-08 at 4.43 ± 0.44
μg/L of Be) came from the Institut national de santé publique
du Québec (INSPQ, Québec, QC, Canada).
Optima-grade nitric acid (70% v/v, Fisher Scientific, Ottawa,
ON, Canada) and Optima-grade sulfuric acid (95%, p/p, Fisher
Scientific) were used without any other distillation. Trace metal
analysis grade perchloric acid (70%, p/p) used for digestion
came from J.T. Baker (Phillipsburg, NJ). Sigma Ultra-grade
Triton X-100 used for ICP-MS solution came from SigmaAldrich (St. Louis, MO). ACS-grade sodium chloride used for
the physiological solutions came from Fisher Scientific.
Spectro-grade isopropanol (99.5%, v/v) was used to make a
liquid suspension of BeO and came from Fisher Scientific.
Liquid argon with purity greater than 99.99% was used for the
ICP-MS (Les gaz spéciaux MEGS, St-Laurent, QC, Canada).
Water used in the different stages of the method was purified
with a Milli-Q Element A-10 system (Millipore, Billerica, MA).
Lithium-6 (10 mg/L) used for internal standardization came
Table I. Analytical and Instrumental Parameters for
ICP-MS
Parameter
ICP R.F. Power (W)
Sample uptake rate (mL/min)
Gas flow rates (L/min)
Plasma
Auxiliary
Nebulizer
Ion sampling depth (mm)
Axial field voltage (V)
Quantitative mode
Lens optimization
Scan mode
Isotopes
Mass scanned (AMU)
Dwell time per AMU (ms)
Number of sweeps/reading
Number of replicates
Detector mode
Value
1100 or 1300*
1.0
15
1.2
0.80–0.95
5.5
225
Maximum intensity for 9Be
Peak hopping
9Be and 6Li
9.012 for Be and 6.015 for Li
500 for Be and 150 for Li
60
3
Dual
* 1100 W for Be and BeAl method; 1300 W for BeO method.
from Spex Certiprep (Metuchen, NJ), as did the other internal
standard tested [internal standard mix 6Li (enriched 95%), Sc,
Ge, Y, In, Tb, and Bi (10 mg/L)]. Optima-grade hydrochloric
acid (35%, w/w, Fisher Scientific) was used for the dissolution
test.
Biological samples used
The blood used for the development and validation of the
study came from volunteers not exposed to Be, and the kidneys
and livers came from pork purchased from a local distributor.
The spleens and lungs came from male Sprague-Dawley rats
weighing between 150 and 200 g (Charles River Canada, SaintConstant, QC, Canada).
Instrumentation
A Perkin Elmer Elan DRC-II (PerkinElmer Life and Analytical Sciences, Waltham, MA) was used for sample analysis. The
ICP-MS was equipped with a Meinhard™ quartz nebulizer, a
quartz cyclonic spray chamber, a tapered quartz injector with
ball joint and a standard quartz torch, Platinum sampler (orifice diameter of 1.1 mm) and skimmer (orifice diameter of 0.9
mm) cones were used to sample ions. The MS analyzer was a
quadruple. The dynamic reaction cell was not used in this
study. The ICP-MS settings are presented in Table I. The instrument optimization steps followed Perkin Elmer guidelines
(22).
Analytical method
Sample preparation. The organic tissues frozen at –70°C
(blood, kidney, liver, lungs, and spleen) were thawed one day
before digestion at 4°C. The organs were rinsed with a 0.9%
(w/v) sodium chloride physiological solution. The volume of
blood used was 1 mL.
Matrix digestion and powder dissolution. Preliminary tests,
using different reagent mixes of HNO3, HClO4, and H2O2 at
temperatures ranging from 150 to 175°C for 3 to 5 h, showed
that the use of perchloric acid in nitric acid reagent solution allows efficient and fast digestion. The digestion parameters
(temperature, time, reagent ratio) were optimized in this regard.
The method was also verified for powder dissolution in the
presence of organic matrix (kidney). Initial methods were used
on the three different species (Be, BeAl, and BeO) on a mass
equivalent to approximately 100 μg of Be. Different acid mixes
were tested on the BeO species (6 mL of HNO3/HClO4 (3:1); 3
treatments with 6 mL of HNO 3 /HClO 4 (3:1), 6 mL of
HNO3/HClO4 (3:1) followed by 3 mL of HCl added after digestion and HNO3/HClO4 (3:1) with 0.5 or 1 mL of H2SO4 added at
the beginning of digestion). The dissolution method using
H2SO4 was retained.
Laboratory materials were decontaminated in trays containing a solution of 20% (v/v) HNO3 for a period of 12 h. The
beakers containing the samples were placed under a perchloric
acid hood.
The following method was used for the digestion of Be and
BeAl species. A volume of 6 mL of a 3:1 solution of concentrated HNO3/HClO4 is added to each beaker. The beakers are
heated with constant mechanical agitation on a heating plate
563
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 3
Journal of Analytical Toxicology, Vol. 34, November/December 2010
adjusted to a surface temperature between 130°C and 140°C.
When foaming is visible in the solutions, the beakers are removed from the plate for a period of 30 min. After that time,
they are returned to the plate and the digestion continues at a
plate temperature between 170°C and 180°C with constant
mechanical agitation. If the solution in the beakers is not clear
when approximately 1 mL of acid remains, 3 mL of concentrated HNO3 is added to the beakers. This addition is repeated
until a clear solution is obtained. When a clear solution is obtained the sample is then evaporated to dryness. The residue is
put back in solution with 1% HNO3 (v/v) and transferred to a
disposable polyethylene test tube and diluted to 10 mL with 1%
HNO3 (v/v). If necessary, the solutions are filtered with a syringe filter (hydrophilic PVDF 0.45 mm) in another disposable
test tube prior to analysis by ICP-MS.
As for BeO species, a volume of 6 mL of a 3:1 solution of concentrated HNO3/HClO4 and a volume of 0.5 mL of concentrated H2SO4 is added to each beaker. The beakers are heated
with constant mechanical agitation on a heating plate adjusted
to a surface temperature between 135°C and 145°C. When
foaming is visible in the solutions, the beakers are removed
from the plate for a period of 30 min. After that time, they are
returned to the plate and the digestion continues at a plate
temperature between 195°C and 205°C with constant me-
chanical agitation. Contrary to the Be and BeAl method, the
sample is not evaporated to dryness (around 140 μL of reagent
should remain) and is diluted with water to a final volume of
14 mL. The ICP-MS nebulizer and injector plasma torch must
be cleaned with methanol between analysis sequences (around
50 samples) to ensure the quality of the results.
Choice of internal standard. In order to verify which internal standard is the most appropriate for this work, commonly used internal standards (6Li, Sc, and Ge) for low mass
were tested. These elements are used because almost no significant chemical interference is known for them (23). 6Li was
chosen because its mass is close to that of Be (m/z = 6 and m/z
= 9, respectively) (24). 72Ge was also investigated because its
first ionization potential is close to that of Be (eVs of 7.90 and
9.32, respectively) (24). 43Sc is the internal standard of choice
in a majority of laboratories analyzing Be (25).
Internal standards were added online at 4:1 (sample/internal
standard) at a concentration of 100 μg/L. Samples used for
the optimization were 1 μg/L Be solutions with increasing
concentration of the matrix (H2SO4) from 0 to 12% (v/v) with
five replicates for each H2SO4 concentration.
Calibration. The ICP-MS was calibrated using aqueous standard solutions prepared from the stock solutions by subsequent dilution in the range of 0.05–2.00 µg/L (8 calibration
points) of Be in 1% (v/v) HNO3. The minimum linearity required for the calibration curve is a R2 coefficient of 0.990. The
Table II. Verification of CRM Recovery With or Without Correction by Formula
solution for internal standardization con(Eq 1) at Different Moments in the Analytical Sequence*
sists of 100 μg/L of 6Li, 1% (v/v) HNO3
Analytical
Cb
Cc
Expected Value
and 0.004% (v/v) Triton X-100 for the Be
Sequence
(ng/mL)
(ng/mL)
(ng/mL)†
and BeAl method. For the BeO method, it
consists of 100 μg/L of 6Li, 5% (v/v)
Beginning
5.2 ± 0.3 (n = 24)‡,§
5.5 ± 0.2 (n = 24)‡,#
H2SO4 and 0.004% (v/v) Triton X-100.
5.4–6.4
BeO quantification method. Sulfuric
§,
,††
#,
,††
End
4.2 ± 0.8 (n = 23) **
5.5 ± 0.5 (n = 23) **
acid generates important matrix effects
* Where Cc = Corrected concentration of Be by equation 1 for the CRM value (ng/mL) and Cb = Concentration of
in ICP-MS (26). These effects are cumuBe in the CRM solution before correction (ng/mL).
lative and become greater as the analysis
† Analytical value given by the supplier.
‡ Statistically different, confirmed with t test (p value = 0.00).
sequence advances. Some preliminary
§ Statistically different, confirmed with t test (p value = 0.00).
tests have indicated (not shown here) that
# Not statistically different, confirmed with t test (p value = 1.00).
** An instrumental error explains the fact that a CRM value is missing at the end of one analysis.
the standard addition method in the con†† Statistically different, confirmed with t test (p value = 0.00).
centration range studied causes too much
variability. Spike correction with internal
standardization correction is used when
Table III. Powder Dissolution Tests for Be, BeAl, and BeO
sulfuric acid is used. In this method, 50
μL of Be standard is added to 4.95 mL of
Average Mass of Be
Be Recovery
each sample to correct the value obtained
Particle
Protocols
(µg)
(%)
by the calibration curve. The final concentration of the addition is 0.5 ng/mL.
Be
6 mL HNO3/HClO4 (3:1)*
198
103 ± 2% (n = 11)
This method has been verified with Certified Reference Materials (CRM) (see
BeAl
6 mL HNO3/HClO4 (3:1)*
143
100 ± 4% (n = 11)
Table II), and the sample concentration
BeO
6 mL HNO3/HClO4 (3:1)
90
75 ± 9% (n = 6)
can be calculated using equation 1.
3 treatments with 6 mL HNO3/HClO4 (3:1)
6 mL of HNO3/HClO4 and 3 mL of HCl
6 mL of HNO3/HClO4 and 1 mL of H2SO4
6 mL of HNO3/HClO4 and 1 mL of H2SO4*
* With biological matrix (kidney).
564
96
96
96
94
81 ± 4% (n = 12)
77 ± 3% (n = 12)
97 ± 3% (n = 5)
99 ± 2% (n = 12)
Cc = [Cb /(R/100)]
[Eq. 1]
where Cc = corrected concentration of
the analyte solution (ng/mL); Cb = concentration of the analyte solution before
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 4
Journal of Analytical Toxicology, Vol. 34, November/December 2010
correction (ng/mL); and R = spike recovery (%).
Method validation
Precision measurement and limits of the method. The
method detection limit (MDL) was calculated as 3 times the
signal-to-noise ratio obtained from 10 replicates under the
given analytical conditions (27,28). A ratio of 10 was used to
calculate the method quantification limit (MQL). MDLs are
evaluated with each organ and blood sample at a low concentration level (ranging from 6 to 20 ng/L), which must be between 4 to 10 times the obtained MDL. The instrumental limit
of detection is determined with no organic matrix using the
same method, with a Be concentration of 2.5 ng/L for the Be
and BeAl method, and a concentration of 6.0 ng/L the for BeO
method (27,28).
The analytical uncertainty (CVa) as determined according to
ISO (29,30) is calculated by taking into account all stages of the
methods. The expanded analytical uncertainty (U) is calculated according to ISO, but by using a coverage factor (k) as
suggested by Bartley (31).
Replicability (in-batch) and repeatability (between batch)
Table IV. Precision Values for the Two Validated Methods
are evaluated on 24 samples covering 4 levels of concentration
(0.1, 0.5, 1.0, and 1.5 μg/L of Be). For repeatability testing, different days of analysis and analysts are used. Equation 2 shows
the formula used for calculating replicability and repeatability
with a confidence level of 95% (27):
[Eq. 2]
where REP = replicability or repeatability (%); tstudent = value
from a student table for a confidence level of 95% with a bilateral probability; RSD = average relative standard deviation
between concentration levels; and n = total number of measures on all concentration levels.
Accuracy measurement. Accuracy is defined as the difference in recovery of the CRM Seronorm target value and the obtained value.
Storage of samples. The stability of the samples is measured using samples prepared and analyzed, according to their
respective methods, on the first day (identified as day 0), and
the same samples are tested again after one of the four following periods: 7, 14, 21, and 28 days.
The samples are stored in glass or
polypropylene tubes at 6°C between periods of analysis.
Method
Matrix
MDL
(pg/mL)
R2
Replicability
(%)
Repeatability
(%)
CVa
(%)
U*
(%)
Be and
BeAl‡
Organ†
Blood
2.0
2.2
0.9999
0.9999
1.0
1.5
2.2
1.7
4.6
3.6
12
9.2
BeO§
Organ
Blood
2.2
2.0
0.999
0.999
3.0
2.5
2.9
2.7
7.5
6.9
19
18
* Expanded uncertainty.
† Average value for the organs (kidney, liver, lung, and spleen). Individual values for each organ are available from
the author.
‡ For a final volume of 10 mL (instrument reading).
§ For a final volume of 14 mL (instrument reading).
Figure 1. Internal standard (IS) signal dependence on the matrix load (H2SO4 concentration).
Results
Matrix digestion and powder
dissolution
The powder dissolution test results
are presented in Table III. The final tests
were in the presence of kidney tissue,
and recoveries of 103 ± 2% for Be and
100 ± 4% for BeAl were obtained. Because the method was ineffective for
completely dissolving BeO particles, different reagents were tested and the addition of H2SO4 was the most effective
technique, with a recovery of 99 ± 2%.
A specific correction method was used
for BeO samples, as shown in equation
1, to correct for the matrix effect. Table
II shows that there was a significant difference between the values obtained at
the beginning and at the end of the analysis sequence when no correction was
used. The correction method eliminated
this problem, as clearly shown in Table
II. It should be noted that the values obtained without correction differed more
significantly from the expected value
than the values obtained with correction. Table II was produced by compiling
the results for 24 analysis batches with 1
measure of CRM at the beginning of
565
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 5
Journal of Analytical Toxicology, Vol. 34, November/December 2010
analysis and 1 measure of CRM after processing a maximum
of 20 samples in triplicate.
Choice of internal standard
Three internal standards were tested (6Li, 43Sc, and 72Ge). By
varying the sulfuric acid content, it is possible to verify whether
the internal standard reacts in the same way as Be. Figure 1
Table V. Accuracy Measurement for the Validated
Methods
Seronorm Whole Blood L-2*
Method
Be and
BeAl
BeO
Be Concentration
Obtained
(ng/mL)
Be analytical
uncertainty
(ng/mL)
Be acceptable
range
(ng/mL)
5.9 ± 0.5†
4.9–6.9‡
5.2 ± 0.2 (n = 74)
5.6 ± 0.4 (n = 36)
* Concentrations provided on the certificate of analysis for “Seronorm Trace
elements whole blood L-2”.
† Expanded Uncertainty (U) for the assigned value with a coverage factor of two
(k = 2).
‡ SERO’s assessment.
shows the results for the internal standards used, and the linearity of the Be/Li signal clearly demonstrates that 6Li showed
the same pattern as 9Be. This fact makes 6Li the internal standard of choice for correcting the 9Be signal.
Method validation
Precision measurement and limits of the method. The evaluation of MDL is an important parameter because concentrations of metals in biological tissues are usually low; methods
with high detection power are therefore needed. The precision
represents the calculation of the replicability and repeatability
of the method. Table IV shows the values for precision measurement.
The MDLs obtained for both methods were all in the same
range, regardless of the biological matrix that was being tested.
The instrumental detection limits were evaluated and were
0.7 pg/mL (1% HNO3 matrix) for the Be and BeAl method,
and 1.2 pg/mL (1% H2SO4 matrix) for the BeO method.
Accuracy measurement. Accuracy of the method was evaluated by the analysis of a commercially available CRM containing Be in a blood matrix (see Table V). The accuracy of the
validated methods was within acceptable certified values. Blood
samples containing Be were also available as proficiency testing
materials for a quality assessment scheme and the results are
reported in Table VI. For all experiments, all sources of contamination (biological matrices, reagents, materials) were verified. Spiked tissues were also used to confirm result.
Storage of samples. Stability was assessed once a week for
four weeks and the results are shown in Figure 2. The stability
was also confirmed with two separate test tubes, polypropylene
and glass (results not shown). As shown in Figure 2, storage of
the samples over a maximum period of 28 days caused no stability problems.
Discussion
Matrix digestion and powder dissolution
Figure 2. Evaluation of stability of Be and BeAl method (A) and BeO
method (B). Recovery relative to day 0 was evaluated as the mean concentration of Be obtained from seven different samples with error bars representing the standard deviation.
566
To ensure good quality results, ICP-MS cannot tolerate more
than 0.5% (22) total dissolved matter. When faced with this
problem, wet digestion is advantageous because it involves the
destruction of organic matter (34). It is important to obtain
complete digestion as quickly as possible. Incomplete digestion
could result in instrument problems or in a precipitate that
could lead to loss of Be, and thereby affect the quality of results
(34). Moreover, it is known that simple solubilization of a
sample, as in the case of blood, does not permit the analysis of
large quantities of samples, due to rapid clogging of the injector torch of the ICP-MS caused by denatured proteins (35).
HClO4 is recognized as the most powerful oxidizing agent used
in the digestion of organic matter (34). The digestion methods
developed are similar to those used by NIOSH (36) and Verma
et al. (37).
It is also important to ensure the complete solubilization of
Be powders because it is known that Be and its compounds are
poorly biotransformed in the body (1). Given the high solubility
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 6
Journal of Analytical Toxicology, Vol. 34, November/December 2010
of Be and BeAl powder in concentrated acid solutions (38),
tests were carried out directly in the presence of organs, as reported in Table III. It is known that BeO particles formed at
high temperatures (those used in the study) can only be dissolved under specific conditions (38,39). Therefore, as shown
in Table III, H2SO4 is required to ensure complete dissolution
of particles and thereby obtain more accurate values in the unknown samples. However, hydrofluoric acid or dilution ammonium bifluoride could also be used, according to Ashley et
al. (40) and Goldcamp et al. (39), but its use would cause incompatibility problems with the glass components in the ICPMS.
The fact that the BeO method uses H2SO4 is a problem,
since this reagent is not usually used in ICP-MS because of its
high viscosity, high boiling point, and tendency to produce
multiple matrix effects (26). The problems mentioned have an
impact on instrument stability, which leads to a significant and
rapid drop in Be recovery during repeated analysis.
Even after the digestion step, there are still some organic
residues that form an insoluble crust when digestion is taken
to dryness at temperatures above 210°C. Thus, the maximum
digestion temperature is below the boiling temperature of
H2SO4 (339°C), but at this temperature, complete evaporation of the reagent cannot occur. The strategy to leave a trace
of H2SO4 after digestion is also described in NIOSH 8005 (36).
The results in Table II show that the correction method is accurate enough to be used instead of standard addition. The results obtained at the end of the analysis show that reproducibility was achieved inside an analysis
sequence by the correction method
Table VI. Interlaboratory Material Recovery Measurement with the Validated
chosen. The matrix effect was also corMethods
rected because the values obtained with
correction were within the expected range
Interlaboratory
Be Concentration Obtained
Be Target Concentration†
for the CRM.
Materials*
Method
(ng/mL)
(ng/mL)
QMEQAS06B-02
Be and BeAl
BeO
1.66 ± 0.09 (n = 6)
1.79 ± 0.04 (n = 4)
1.74 ± 0.29
QMEQAS07B-03
Be and BeAl
BeO
0.84 ± 0.04 (n = 5)
0.89 ± 0.04 (n = 4)
0.950 ± 0.159
QMEQAS06B-08
Be and BeAl
BeO
4.2 ± 0.1 (n = 6)
4.37 ± 0.06 (n = 4)
4.43 ± 0.44
Choice of internal standard
* Quebec Multielement External Quality Assessment Scheme (QMEQAS): These proficiency testing materials are
made from a pool of human blood collected from non exposed individuals which has been spiked with beryllium
and are provided by Institut National de Santé Publique du Québec (INSPQ).
† Concentrations provided on the certificate of analysis.
The role of the internal standard is to
compensate for physical interference and
instrument instabilities that may affect
the signal of the analyte of interest. To be
effective, an internal standard should behave exactly as the analyte of interest, so
its molecular weight and ionization potential must be as close as possible to
those of the analyte of interest (22,41).
There is currently no consensus in the
Table VII. List of Pertinent Methods for the Analysis for Beryllium in Biological Matrices
Author(s)
Related
Biological Matrices
Preparation
Analysis
Final Volume
(mL)
Internal
Standard
MDL*
Engstrom et al. (47)
Kidney, lungs, liver
Microwave
ICP-SFMS
10
In
6 pg/mL
Kidney, liver,
lungs, spleen
Solid state with 10Be†
AMS
N/A
N/A
0.8 amol of 10Be
Delves (45)
Blood
Solubilization
ICP-MS
10
Li or Sc
10 pg/mL
Heitland and Köster (51)
Blood
Solubilization
ICP-MS
5
Tb
2.4 pg/mL
30 pg/mL
Chiarappa-Zucca et al.
(50)
de Boer et al. (23)
Blood
Solubilization
ICP-MS
10
72Ge
Goullé et al. (52)
Blood
Solubilization
ICP-MS
4
Rh
4.2 pg/mL
Rodushkin et al. (35)
Blood
Microwave
ICP-SFMS
10
In
1.4 pg/mL
Stephan et al. (53)
Blood
Solubilization
GF-AAS
8-fold dilution
N/A
7 pg/mL
2.0 pg/mL for Be and BeAl
2.2 pg/mL for BeO
2.2 pg/mL for Be and BeAl
2.0 pg/mL for BeO
Present study
Kidney, liver,
lungs, spleen
Wet digestion
ICP-MS
10 (Be and BeAl)
14 (BeO)
6Li
Present study
Blood
Wet digestion
ICP-MS
10 (Be and BeAl)
14 (BeO)
6Li
* MDL evaluated from the instrument reading.
† This method is not applicable here because there was no AMS (accelerator mass spectrometry) available for this project.
567
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 7
Journal of Analytical Toxicology, Vol. 34, November/December 2010
literature for the choice of internal standard that is most effective for the analysis of Be by ICP-MS (25).
Figure 1 indicates that only the Be signal corrected with 6Li
did not vary as a function of the load of the matrix. This
demonstrates stability for this internal standard, a quality that
was sought by this experiment. Thus, 6Li is the internal standard of choice for the methods. The phenomenon responsible
for the non-linearity observed with 43Sc and 72Ge is the “space
charge effect”, where low mass ions are pushed off the ion
beam by higher mass ions (which have a higher kinetic energy)
before their entry into the lens (area under high vacuum).
The fraction of small ions in the beam is decreased, and thus
the apparent signal for these ions will also be decreased (42,43).
As 43Sc and 72Ge are heavier, this phenomenon does not affect
their signal in the same manner as with 6Li and Be (22,44).
The isotopic abundance of Li in the natural state is 7.5% for
6Li and 92.5% for 7Li (24). It is mathematically possible to
correct for the interference due to possible contamination by
Li in the samples. This correction may be useful if the method
is to be applied to human blood, because the possible use of antidepressants containing Li could alter the results (16). Also,
contamination can come from the anti-coagulant used in the
case of blood collection, if it contains lithium heparin (45).
Method validation
Precision measurement and limits of the method. The MDL
was an important parameter to estimate in the study. Table VII
shows the relevant methods identified in the literature with
their respective MDL. The method used here was more rigorous than simply using a blank as in many other studies (28).
As reported by Miller and Miller (46), it is difficult to compare
the MDL between methods, because the calculation method
differs from one author to another, but despite this fact, the
MDLs obtained were mostly better (see Table VII) than those
identified in the literature, although better detection limits are
usually observed with ICP-SFMS (47). This is because ICPSFMS has a very low background signal and better ion transmission than ICP-MS with a quadruple analyzer (48). For tissues, the lowest detectable concentrations in the present study
were slightly better than those of Engström’s team (47) (2.0
pg/mL for the Be or BeAl method, and 2.2 pg/mL for the BeO
method versus 6.0 pg/mL). For blood, the lowest detectable
concentrations for both methods were comparable to those
obtained by Rodushkin’s team (35), which is the lowest MDL
found (2.2 pg/mL for the Be or BeAl method, and 2.0 pg/mL for
the BeO method versus 1.4 pg/mL).
The values obtained for the replicability and repeatability of
each method with both types of matrixes were consistent with
the required value of 10% (27,28), showing low variability in
the same and between-sample batches.
Accuracy measurement. As reported by other authors, CRMs
for Be in biological organs need to be developed (38,49). Blood
CRMs containing Be were used in the calculation of accuracy.
However, Be in blood CRMs is spiked in soluble form, so this
material does not determine the solubility of Be particles. This
is one of the reasons why it is important to assess the solubility
of each species by other means.
If we compare the present method with the methods found
568
in the literature, it can be understood that one of the advantages of our method is the rigorous assessment of powder dissolution in the reagent mix. Very low MDLs were reported and
the following reasons mainly explain this fact: rigorous digestion of the organic matrix (unlike simple dilution), which significantly reduces the matrix effects; a longer integration time
because the study focuses only on one element, unlike several
studies that analyze several elements at once; and an internal
standard (6Li) selected specifically for the analysis of Be and optimized for the matrix.
Conclusions
Reliable methods for the detection of ultra-trace levels of Be
in various biological matrices have been developed and validated with methodological detection limits among the best in
the literature. The methods developed are sensitive enough to
be applied to future toxicological study.
Because validation has been done directly with human blood,
these methods can be applied in epidemiological or toxicological studies for the analysis of Be in human blood. The methods
also allow the determination of Be in the lungs of deceased
workers in order to help distinguish between chronic beryllium
disease or pulmonary sarcoidosis (37).
Acknowledgments
The authors acknowledge the financial support of the Institut de recherche Robert-Sauvé en santé et en sécurité du travail from a research grant. They also thank Brigitte Blanchette
(IRSST) and Michel Marion (UQÀM) for their technical support.
References
1. World Health Organization (WHO). Beryllium and Beryllium
Compounds, Concise International Chemical Assessment Document 32, Genève, Switzerland, 2001, pp 6–7, 21–31
2. Institut National de Santé Publique du Québec (INSPQ). Le Test
Sanguin de Prolifération Lymphocytaire au Béryllium (BeLPT),
Santécom, INSPQ-2004-018, Canada, 2004, pp 5–10.
3. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for Beryllium, U.S. Department of Health and
Human Services, Rockville, MD, 2002, ch. 3.
4. C. Saltini, K. Winestock, M. Kriby, P. Pinkston, and R.G. Crystal.
Maintenance of alveolitis in patients with chronic beryllium disease by beryllium-specific helper T cells. N. Engl. J. Med. 320:
1103–1109 (1989).
5. L.S. Newman, M.M. Mroz, R. Balkissoon and L.A. Maier. Beryllium sensitisation progresses to chronic beryllium disease. Am. J.
Respir. Crit. Care Med. 171: 54–60 (2005).
6. M.D. Rossman. Chronic beryllium disease: a hypersensitivity disorder. Appl. Occup. Environ. Hyg. 16: 615–618 (2001).
7. J.W. Martyny, M.D. Hoover, M.M. Mroz, K. Ellis, L.A. Maier,
K.L. Sheff, and L.S. Newman. Aerosols generated during beryllium
machining. J. Occup. Environ. Med. 42: 8–18 (2000).
8. K. Kreiss, M.M. Mroz, B. Zhen, J.W. Martyny, and L.S. Newman.
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 8
Journal of Analytical Toxicology, Vol. 34, November/December 2010
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Epidemiology of beryllium sensitization and disease in nuclear
workers. Am. Rev. Respir. Dis. 148: 985–991 (1993).
R.S. Ratney. Is beryllium disease a fossil?—not yet. Int. Arch.
Occup. Environ. Health. 74: 159–161 (2001).
M.S. Kent, T.G. Robins, and A.K. Madl. Is total mass or mass of
alveolar-deposited airborne particles of beryllium a better predictor of the prevalence of disease? A preliminary study of a
beryllium processing facility. Appl. Occup. Environ. Hyg. 16:
539–558 (2001).
P.K. Henneberger, D. Cumro, D.D. Deubner, M.S. Kent, M. McCawley, and K. Kreiss. Beryllium sensitization and disease among
long-term and short-term workers in a beryllium ceramics plant.
Int. Arch. Occup. Environ. Health 74: 167–176 (2001).
M.E. Kolanz. Introduction to beryllium: uses, regulatory history,
and disease. Appl. Occup. Environ. Hyg. 16: 559–567 (2001).
M.A. McCawley, M.S. Kent, and M.T. Berakis. Ultrafine beryllium
number concentration as a possible metric for chronic beryllium
disease risk. Appl. Occup. Environ. Hyg. 16: 631–638 (2001).
D.J. Paustenbach, A.K. Madl, and J.F. Greene. Identifying an appropriate occupational exposure limit (OEL) for beryllium: data
gaps and current research initiatives. Appl. Occup. Environ. Hyg.
16: 527–538 (2001).
G.L. Finch, J.A. Mewhinney, A.F. Edison, M.D. Hoover, and
S.J. Rothenberg. In vitro dissolution characteristics of beryllium
oxide and beryllium metal aerosols. J. Aerosol. Sci. 19: 333–342
(1988).
C. Moesch. Utilisation de l’ICP-MS en biologie clinique. Ann. Toxicol. Anal. 19: 11–21 (2007).
C.A. Burtis, E.R. Ashwood, and D.E. Bruns. TIETZ Fundamentals
of Clinical Chemistry. Saunders-Elsevier, Philadelphia, PA, 2007,
p 131.
P. Apostoli and K.H. Schaller. Urinary beryllium—a suitable tool
for assessing occupational and environmental beryllium exposure? Int. Arch. Occup. Enron. Health 74: 162–166 (2001).
C. Minoia, E. Sabbioni, P. Apostoli, R. Pietra, L. Pozzoli, M. Gallorini, G. Nicolaou, L. Alessio, and E. Capodaglio. Trace element
reference values in tissues from inhabitants of the European community. I. A study of 46 elements in urine, blood and serum of
Italian subjects. Sci. Total Environ. 95: 89–105 (1990).
P. Schramel, I. Wendler, and J. Angerer. The determination of
metals (antimony, bismuth, lead, cadmium, mercury, palladium,
platinum, tellurium, thallium, tin and tungsten) in urine samples
by inductively coupled plasma-mass spectrometry. Int. Arch.
Occup. Environ. Health 69: 219–223 (1997).
J. Angerer and K.H. Schaller. Beryllium, Lithium, Vanadium and
Tungsten in Urine by ICP Emission Spectrometry. Analyses of
Hazardous Substances in Biological Materials, Vol. 5. WileyVCH, Weinheim, Germany, 1997, pp 51–77.
Perkin Elmer. Inductively Coupled Plasma Mass Spectrometry
with ELAN Software, NO220008 Rev. B, ch. 6
J.L.M. de Boer, R. Ritsema, S. Piso, H. van Staden, and W. van den
Beld. Practical and quality-control aspects of multi-element analysis with quadrupole ICP-MS with special attention to urine and
whole blood. Anal. Bioanal. Chem. 379: 872–880 (2004).
A. Montaser and D.W. Golightly. Inductively Coupled Plasma in
Analytical Atomic Spectrometry. VCH, New York, NY, 1987.
M.J. Brisson, A.A. Ekechukwu, K. Ashley, and S.D. John. Opportunities for standardization of beryllium sampling and analysis.
J. ASTM Int. 3: 3–14 (2006).
J. Takahashi and K. Youno. Analysis of Impurities in Semiconductor Grade Sulfuric Acid Using the Agilent 7500cs ICP-MS. Agilent Technologies, 2003, pp 1–6. www.chem.agilent.com/Library/applications/5988-9190EN.pdf (accessed September 2010).
Institut de Recherche Robert-Sauvé en Santé et en Sécurité du Travail (IRSST). Développement, Mise au Point et Validation d’une
Méthode Analytique, I-G-020, Montreal, QC, Canada, 2006, pp
1–15
Centre d’Expertise en Analyse Environnementale du Québec
(CEAEQ). Protocole pour la Validation d’une Méthode d’Analyse
en Chimie, DR-12-VMC, Montreal, QC, Canada, 2002, pp 1–27.
29. International Organization for Standardization (ISO). Guide to the
Expression of Uncertainty in Measurement (GUM), ISO, Genève,
Switzerland, 1992, pp 1–91.
30. E.R. Kennedy, T.J. Fischbach, R. Song, P.M. Eller, and
S.A. Shulman. Guidelines for Air Sampling and Analytical Method
Development and Evaluation, NIOSH Technical Report, 1995.
http://www.cdc.gov/niosh/docs/95-117/pdfs/95-117.pdf (accessed
September 2010).
31. D.L. Bartley. Analytical performance criteria—reconciling traditional accuracy assessment with the ISO guide to the expression
of uncertainty in measurement (ISO/GUM). J. Occup. Environ.
Hyg. 1: D37–D41 (2004)
32. E. de Hoffman and V. Stroobant. Spectrométrie de Masse, 3rd ed.
Dunod, Paris, France, 2005, pp 65–68
33. J. Begerow, M. Turfeld, and L. Dunemann. New horizons in
human biomonitoring of environmentally and occupationally
relevant metals—sector-field ICP-MS versus electrothermal AAS.
J. Anal. At. Spectrom. 15: 347–352 (2000)
34. J. Angerer and K.H. Schaller. Digestion procedures for the determination of metals in biological materials. In Analyses of Hazardous Substances in Biological Materials, Vol. 8. Wiley-VCH,
Weinheim, Germany, 2003, ch. 5-2
35. I. Rodushkin, F. Ödman, R. Olofsson, and M.D. Axelsson. Determination of 60 elements in whole blood by sector field inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom.
15: 937–944 (2000).
36. National Institute for Occupational Safety and Health (NIOSH).
ELEMENTS in Blood or Tissue (8005), NIOSH Manual of Analytical Methods, 4th ed., 1994, pp 1–6. www.cdc.gov/niosh/docs/
2003-154/pdfs/8005.pdf (accessed September 2010).
37. D.K. Verma, A.C. Ritchie, and M.L. Shaw. Measurement of beryllium in lung tissue of a chronic beryllium disease case and cases
with sarcoidosis. Occup. Med. (Lond.) 53: 223–227 (2003)
38. A.B. Stefaniak, M.D. Hoover, G.A. Day, A.A. Ekechukwu,
G.E. Whitney, C.A. Brink, and R.C. Scripsick. Characteristics of
beryllium oxide and beryllium metal powders for use as reference
materials. J. ASTM Int. 2(10): 47–61 (2005).
39. M. J. Goldcamp, D.M. Goldcamp, K. Ashley, J.E. Fernback,
A. Agrawal, M. Millson, D. Marlow, and K. Harrison. Extraction
of beryllium from refractory beryllium oxide with dilute ammonium bifluoride and determination by fluorescence: a multiparameter performance evaluation. J. Occup. Environ. Hyg. 6: 735–
744 (2009).
40. K. Ashley, M.J. Brisson, and S.D. Jahn. Standard methods for
beryllium sampling and analysis: availabilities and needs. J. ASTM
Int. 3: 15–26 (2006).
41. H.E. Taylor. Inductively Coupled Plasma-Mass Spectrometry: Practices and Techniques. Academic Press, San Diego, CA, 2001, pp
104–139.
42. K. Busch. Space charge in mass spectrometry. Spectroscopy 19:
35–38 (2004).
43. I.I. Stewart and J.W. Olesik. Time-resolved measurements with
single droplet introduction to investigate space-charge effects in
plasma mass spectrometry. J. Am. Soc. Mass Spectrom. 10:
159–174 (1999).
44. R. Thomas. Practical Guide to ICP-MS. Marcel Dekker, New York,
NY, 2004, pp 145–146
45. T. Delves. Valid analytical measurements in clinical applications
of ICP-MS. VAM Bull. 20: 16–21 (1999).
46. J.N. Miller and J.C. Miller. Statistics and Chemometrics for Analytical Chemistry, 5th ed. Pearson-Prentice Hall, Essex, England,
2005, pp 121–124.
47. E. Engström, A. Stenberg, S. Senioukh, R. Edelbro, D.C. Baxter,
and I. Rodushkin. Multi-elemental characterization of soft biological tissues by inductively coupled plasma-sector field mass
spectrometry. Anal. Chim. Acta 151: 123–135 (2004)
48. I. Rodushkin, F. Ödman, and S. Branth. Multielement analysis of
whole blood by high resolution inductively coupled plasma mass
spectrometry. Fresenius J. Anal. Chem. 364: 338–346 (1999)
49. R.L. Watters, Jr., M.D. Hoover, G.A. Day, and A.B. Stefaniak. Op-
569
Lariviere.qxd:JATLynneTemplate
10/19/10
3:57 PM
Page 9
Journal of Analytical Toxicology, Vol. 34, November/December 2010
portunities for development of reference materials for beryllium.
J. ASTM Int. 3: 29–46 (2006).
50. M.L. Chiarappa-Zucca, R.C. Finkel, R.E. Martinelli, J.E. McAninch,
D.O. Nelson, and K.W. Turteltaub. Measurement of beryllium in
biological samples by accelerator mass spectrometry: applications
for studying chronic beryllium disease. Chem. Res. Toxicol. 17:
1614–1620 (2004)
51. P. Heitland and H.D. Köster. Biomonitoring of 37 trace elements
in blood samples from inhabitants of northern Germany by ICPMS. J. Trace Elem. Med. Biol. 20: 253–262 (2006)
52. J.-P. Goullé, L. Mahieu, J. Castermant, N. Neveu, L. Bonneau,
G. Lainé, D. Bouige, and C. Lacroix. Metal and metalloid multi-
570
elementary ICP-MS validation in whole blood, plasma, urine and
hair. Forensic Sci. Int. 153: 39–44 (2005)
53. C.H. Stephan, M. Fournier, P. Brousseau, and S. Sauvé. Graphite
furnace atomic absorption spectrometry as a routine method for
the quantification of beryllium in blood and serum. Chem. Cent.
J. 2: 14 (2008).
Manuscript received November 19, 2009;
revision received May 12, 2010.