Clinical Chemistry 46:4
529 536 (2000)
Simple Microplate Method for Determination of
Urinary Iodine
Toshinori Ohashi,1* Mitsuo Yamaki,1 Chandrakant S. Pandav,2
Madhu G. Karmarkar,2 and Minoru Irie3
Background: Urinary iodine is a good biochemical
marker for control of iodine deficiency disorders. Our
aim was to develop and validate a simple, rapid, and
quantitative method based on the SandellKolthoff re-
action, incorporating both the reaction and the digestion
process into a microplate format.
Methods: Using a specially designed sealing cassette to
prevent loss of vapor and cross-contamination among
wells, ammonium persulfate digestion was performed
in a microplate in an oven at 110 C for 60 min. After the
digestion mixture was transferred to a transparent mi-
croplate and the SandellKolthoff reaction was per-
formed at 25 C for 30 min, urinary iodine was measured
by a microplate reader at 405 nm.
Results: The mean recovery of iodine added to urine
was 98% (range, 89 109%). The theoretical detection
limit, defined as 2 SD from the zero calibrator, was 0.11
mol/L (14 g/L iodine). The mean intra- and interassay
CVs for samples with iodine concentrations of 0.30 3.15
mol/L were <10%. The new method agreed well with
the conventional chloric acid digestion method (n 70;
r 0.991; y 0.944x 0.04; Syx 0.10) and with the
inductively coupled plasma mass spectrometry method
(n 61; r 0.979; y 0.962x 0.03; Syx 0.20). The
agreement was confirmed by difference plots. The dis-
tributions of iodine concentrations for samples from
endemic areas of iodine deficiency diseases showed
similar patterns among the above three methods.
Conclusions: Our new method, incorporating the whole
process into a microplate format, is readily applicable
and allows rapid monitoring of urinary iodine.
2000 American Association for Clinical Chemistry
1
Pharmaceutical Research Laboratory, Hitachi Chemical Co., Ltd., 13-1,
Higashi-cho 4-chome, Hitachi-shi, Ibaraki-ken 317-8555, Japan.
2
International Council for Control of Iodine Deficiency Disorders, New
Delhi 110029, India.
3
International Council for Control of Iodine Deficiency Disorders, Tokyo
112-0001, Japan.
*Author for correspondence. Fax 81-294-24-3602; e-mail tos-
oohashi@hitachi-chem.co.jp.
Received December 13, 1999; accepted January 27, 2000.
Endocrinology and
Metabolism
The urinary iodine concentration is a good biochemical
marker for the control of iodine deficiency disorder
(IDDs)4 (1 ). Most of the popular methods for urinary
iodine determination are based on the SandellKolthoff
reaction (2 ). However, substances interfering with the
SandellKolthoff reaction usually affect the performance
of these methods. Chloric acid digestion is one of the
efficient techniques for removing interfering substances
(3 ). The International Council for Control of Iodine Defi-
ciency Disorders (ICCIDD) has compiled seven methods,
based on the SandellKolthoff reaction (1 ), that are used
in several laboratories around the world. Among these
methods, the chloric acid digestion is the most commonly
used. Although it provides an accurate measurement, the
method also has the following disadvantages (1 ): (a)
production of toxic wastes (5 mL/test) from arsenic
trioxide in the SandellKolthoff reaction; (b) leakage of
gas during sample digestion, requiring a special fume
hood; and (c) difficulty in locating chloric acid from
chemical vendors because of its instability.
On the other hand, an alternative method that uses
ammonium persulfate digestion has been reported re-
cently as a nonhazardous, nonexplosive, and easy-to-use
method (4 ). The persulfate digestion makes possible a
comparatively nonhazardous (no chlorine gas) measure-
ment. However, this method is still not completely suit-
able for testing because it is time-consuming and pro-
duces a nonnegligible amount of toxic waste.
One of our objectives in this study was to seek an
easy-to-use method. We applied a microplate format to all
processes so that we could minimize the amount of toxic
wastes as well as simplify and speed up the procedure.
We tried using a closed system during the digestion
process to keep the reaction mixtures in all wells volu-
metrically equal and to avoid contamination throughout
4
Nonstandard abbreviations: IDD, iodine deficiency disorder; ICCIDD,
International Council for Control of Iodine Deficiency Disorders; APDM,
ammonium persulfate digestion on microplate; PP, polypropylene; and ICP/
MS, inductively coupled plasma mass spectrometry.
529
530 Ohashi et al.: Microplate Method for Urinary Iodine
the process; in other words, to prevent the leakage of
vapor and to achieve an accurate measurement of urinary
iodine.
In this study, we evaluated the analytical performance
of the ammonium persulfate digestion on microplate
(APDM) method.
Materials and Methods
equipment and apparatus
Digestion was performed in a standard oven (ST-450
drying oven; Shibata Scientific Technology) after a
polypropylene SEROCLUSTER 96-well microplate (Corn-
ing Costar Japan) was sealed with a stainless steel cassette
(Sealing Cassette; Hitachi Chemical Techno-plant). The
sealing cassette includes a Teflon (fluorinated ethylene
propylene)-laminated silicon rubber gasket and a handle
to seal the wells of a microplate (Fig. 1).
The colorimetric measurements were performed in a
microplate reader (IMMUNO-MINI; Nalge Nunc Interna-
tional).
chemicals
Potassium iodate for calibrators and analytical grade
arsenic trioxide, potassium chlorate, perchloric acid (700
g/L), ammonium persulfate, tetraammonium cerium (IV)
sulfate dihydrate, sodium chloride, and sulfuric acid were
obtained from Wako Pure Chemical Industries. Glass-
distilled deionized water was used for preparation of
reagent solution and dilution procedures.
solutions
Chloric acid solution (3.3 mol/L). Potassium chlorate (500 g)
was dissolved in 1000 mL of water in a 2000-mL Erlen-
meyer flask with heating for 60 min in a boiling water
bath, after which 375 mL of perchloric acid was added
slowly with constant stirring. The solution was then
stored at 25 C in a freezer overnight. The resulting
Fig. 1. Sealing cassette.
The cassette is used for digestion after a microplate is placed inside.
suspension was filtered with a glass filter (510 m mesh).
The filtrate was stored in a refrigerator (4 C) until use.
Ammonium persulfate solution (1.31 mol/L). Ammonium
persulfate (30 g) was dissolved in water to a final volume
of 100 mL. This solution was prepared fresh just before
use.
Arsenious acid solution (0.05 mol/L). Arsenic trioxide (5 g)
was dissolved in 100 mL of 0.875 mol/L sodium hydrox-
ide solution. Concentrated sulfuric acid (16 mL) was then
added slowly to the solution in an ice bath. After cooling,
12.5 g of sodium chloride was added to the solution, and
the mixture was diluted to 500 mL with cold water and
filtered.
Ceric ammonium sulfate solution (0.019 mol/L). Tetraammo-
nium cerium (IV) sulfate dihydrate (6 g) was dissolved in
1.75 mol/L sulfuric acid and adjusted to a final volume of
500 mL with the same acid solution.
Iodine calibrators. In a 100-mL volumetric flask, 168.6 mg of
potassium iodate was dissolved in water to make a 7.88
mmol/L stock solution (1000 mg/L iodine). The stock
solution was diluted 100- and 10 000-fold, and working
solutions of 0.039 4.73 mol/L (5 600 g/L iodine) were
prepared.
urine samples
Urine samples were collected from the following: (a)
individuals (children and adults) staying in the suburbs of
Ulaanbaatar, Mongolia; (b) patients (children and adults)
attending clinics in Baluchistan, Pakistan; (c) nurses and
other technical staff working in the All India Institute of
Medical Sciences, New Delhi, India; and (d) the general
population in Khopasi, Nepal (suburbs of Kathmandu).
experimental procedure
APDM method. Calibrators and urine samples (50 L each)
were pipetted into the wells of a polypropylene (PP) plate,
followed by the addition of 100 L of ammonium persul-
fate solution (final concentration, 0.87 mol/L). The PP
plate was set in a cassette. The cassette was tightly closed
and was kept for 60 min in an oven adjusted to 110 C.
After digestion, the bottom of the cassette was cooled to
room temperature with tap water to avoid condensation
of vapor on the top of wells and to stop the digestion. The
cassette was opened, and 50-L aliquots of the resulting
digests were transferred to the corresponding wells of a
polystyrene 96-well microtiter plate (MicroWell; Nalge
Nunc International). Arsenious acid solution (100 L) was
added to the wells and mixed; 50 L of ceric ammonium
sulfate solution was then added quickly (within 1 min),
using a multichannel pipette (Finnpipette Varichannel;
Labsystems). The reaction mixture was allowed to sit for
30 min at 25 C, and the absorbance was measured at 405
nm with a microplate reader.
 Clinical Chemistry 46, No. 4, 2000
Inductively coupled plasma mass spectrometry (ICP/MS)
method. An SPQ 8000A1 ICP/MS analyzer (Seiko Instru-
ments) was used. The procedure was carried out accord-
ing to the method of Yoshinaga and Morita (5 ). The
plasma gas was argon at flow of 16 L/min, the auxiliary
gas was argon at a flow of 0.4 L/min, and the nebulizer
gas was argon at a flow of 0.4 L/min. The sample flow
rate was 1 mL/min. The output frequency was 1 kW, 12.2
MHz, and the detector was set at 2 kV.
Conventional chloric acid digestion method in a test tube.
Calibrators and urine samples (250 L) were added to
16 160 mm test tubes, and 750 L of chloric acid
solution was added. The tubes were covered with plastic
caps, placed into the wells (75 mm in depth) of an
aluminum block, and left for 60 min at 110 C in a fume
hood. Because the test tubes were much longer than the
depth of the heating block wells, the vapor refluxed
within the test tubes. After the test tubes were cooled, 3.5
mL of arsenious acid solution was added into each tube
and mixed. Three hundred fifty microliters of ceric am-
monium sulfate solution was added and mixed by vortex-
type mixer; a stopwatch was used to keep a constant
interval. Exactly 20 min after the addition of the ceric
ammonium sulfate solution, the absorbance at 405 nm
was measured with a spectrophotometer.
evaluation of sealing cassette
Airtightness. The airtightness of the sealing cassette was
evaluated by comparing the weight of water in each well
of the PP plate before and after heating with the cassette
in an oven, using the following method: Water (150 L)
was pipetted into one well of the PP plate on a balance,
and the increased weight of the PP plate was determined.
The above step of pipetting and weighing was repeated
sequentially for remaining all wells. The PP plate was
then loaded into the sealing cassette and heated for 60 min
in a 110 C oven. After cooling, the plate was weighed.
The water in one well was completely removed by suck-
ing with a pipette and drying with a cotton swab, and the
plate was reweighed. This step was repeated for all other
wells.
Cross-contamination. Cross-contamination between wells
was evaluated by comparing two 96-well plates: (a) a
control plate in which urine was placed into every second
well (48 wells); and (b) a test plate, in which the same
urine was placed into every second well (48 wells) and KI
urine was placed into the remaining wells (48 wells). KI
urine was prepared by mixing KI solution (0.1 mL of a
1000 mol/L solution) and the urine (0.9 mL). After
digestion with ammonium persulfate, the concentration
of iodine was measured separately.
optimization of apdm method of digestion
Optimization is performed using eight Mongolian urine
samples (iodine concentration, 0.30 1.35 mol/L). Opti-
531
mum digestion was defined as having sufficient recover-
ies of iodine added to various urine samples (final added
iodate concentration, 0.79 mol/L) according to the
method described below.
assay evaluation
Calibration curve and calculation. A calibration curve was
prepared for each plate by plotting the logarithmic con-
version of the means of absorbance at 405 nm (n 2) on
the y axis vs the iodine concentrations [0.20, 0.39, 0.79,
1.57, 2.36, 3.15 mol/L (25, 50, 100, 200, 300 and 400 g/L
iodine)] on the x axis. The urinary iodine concentration
was determined using linear regression. Water was used
for the zero calibrator.
Detection limit. A pooled urine sample at a low iodine
concentration was serially diluted with water. The detec-
tion limit of urinary iodine, defined as 2 SD from the zero
calibrator (replicates of 10), was characterized from five
analyses.
Precision. Pooled urine samples with low, medium, and
high concentrations of iodine were used to determine the
intra- and interassay CVs. In an intraassay experiment,
each urine sample was assayed in eight replicates on the
same plate. Using the same samples, an interassay exper-
iment was performed on 30 different days.
Recovery. The recovery of iodine was estimated by assay-
ing in triplicate 12 different urine samples supplemented
with potassium iodate solution, and comparing the re-
sults with those of water-added urine samples. The io-
date-added urine was prepared by adding a given vol-
ume (1/10 volume of the urine sample) of potassium
iodate solution (3.94 mol/L) to the urine samples. For
the water-added urine, water was added instead of po-
tassium iodate solution. The percentage of recovery was
calculated by the following equation:
Recovery (%)
Iodine concentration in iodate-added
urine Iodine concentration
in water-added urine
100
Iodine concentration of
added iodate solution 0.1
Effect of interfering substances. The effects of three interfer-
ing substances (potassium thiocyanate, l-ascorbic acid,
and ferrous ammonium sulfate) on the assay were as-
sessed in triplicate using the following series: (a) iodate
solution plus interfering substance, without digestion; (b)
iodate solution plus interfering substance, with digestion;
and (c) urine plus interfering substance, with digestion.
The interfering compounds were added to urine or iodate
solution to final concentrations of 0.1 64 mmol/L. The
532 Ohashi et al.: Microplate Method for Urinary Iodine

Fig. 3. Calibration curve.

Results
evaluation of sealing cassette
Airtightness. After the microtiter plate sealed with the
sealing cassette was heated for 60 min in the 110 C oven,
99.4% 1.9% of water was retained.

Cross-contamination. The mean urinary iodine concentra-

Fig. 2. Optimization of the APDM digestion method.
The recovery of added iodate was evaluated with various combinations of
ammonium persulfate concentration, oven temperature, and digestion time. (A),
effect of ammonium persulfate concentration: digestion time, 60 min; oven
temperature, 110 C. (B), effect of oven temperature: ammonium persulfate
concentration, 0.87 mol/L; digestion time, 60 min. (C), effect of digestion time:
ammonium persulfate concentration, 0.87 mol/L; oven temperature, 110 C.
iodine concentrations of urine sample and iodate solution
without interfering substances were measured as controls.
Digestion of iodocompounds. Four organic iodocompounds
(p-iodobenzoic acid, m-iodophenol, 2-iodophenol, and
3-iodotyrosine) were examined with the digestion pro-
cess. The recovery of iodine from each iodocompound
was estimated by comparing the measured iodine concen-
tration with the calculated concentration.
tion of the control plate was 1.17 0.03 mol/L (range,
1.131.28 mol/L), and that of test plate was 1.21 0.05
mol/L (range, 1.111.34 mol/L). The recovery of io-
dine added to urine as potassium iodide (100 mol/L)
was 96.8% 6.7% (range, 87.8 118.6%).
optimization of apdm method of digestion
The recovery of iodine added to eight urine samples was
compared with various combinations of three variables;
i.e., final concentration of ammonium persulfate, oven
temperature, and digestion time. Based on analysis of the
data, the combination of 0.87 mol/L (200 g/L ) ammo-
nium persulfate (final concentration), an oven tempera-
ture of 110 C, and a 60 min-incubation was found to be
optimum and gave the highest recovery with little scatter
(Fig. 2). When digestion was performed under submaxi-
mal conditions, the recovery was 0 100%, depending on
the urine samples.

Linearity. Pooled urine samples containing low, medium,

and high concentrations of iodine were serially diluted Table 1. Precision.
with water. The testing was carried out in triplicate. Mean SD,
Sample mol/L CV, %
Comparison with other methods. Pearson and Spearman Intraassay Low 1 0.13 0.02 15
correlations were applied to the results. Comparison of (8 replicates) Low 2 0.30 0.02 6.7
the APDM method and the ICP/MS method was per- Medium 1.19 0.02 1.7
formed using a total of 61 urine samples from Mongolia High 2.51 0.05 2.0
and Pakistan with iodine concentrations of 0.079 3.15 Interassay (30 Low 1 0.15 0.03 20
different days) Low 2 0.29 0.03 10
mol/L. Comparison of the APDM method with the
Medium 1.14 0.05 4.4
conventional chloric acid digestion method was per-
High 2.42 0.11 4.5
formed using 70 urine samples from Nepal.
 Clinical Chemistry 46, No. 4, 2000 533

Table 2. Recovery of iodine added to urine as iodate.

Iodine concentration in triplicate, mol/L
Iodine recovered,
Urine sample Water-addeda Iodate-addedb mol/L Recovery, %
1 0.58 0.01 0.96 0.05 0.38 96
2 2.03 0.08 2.46 0.06 0.43 109
3 0.33 0.04 0.72 0.02 0.39 99
4 0.46 0.02 0.82 0.04 0.36 91
5 0.85 0.02 1.21 0.18 0.36 91
6 1.80 0.04 2.17 0.08 0.37 94
7 0.50 0.01 0.85 0.02 0.35 89
8 0.99 0.08 1.40 0.14 0.41 104
9 0.21 0.01 0.59 0.05 0.38 96
10 1.86 0.18 2.25 0.07 0.39 99
11 0.61 0.11 1.02 0.04 0.41 104
12 0.63 0.11 1.03 0.10 0.40 102
a
Water-added urine was prepared by adding 1 volume of water to 9 volumes of urine.
b
Iodate-added urine was prepared by adding 1 volume of iodate solution (3.94 mol/L) to 9 volumes urine.

assay evaluation iodine) for these compounds was 101% of the expected
Calibration curve. The absorbance on a log scale was linear value (Table 4).
for iodine concentrations between 0 and 3.15 mol/L. The
correlation coefficient for the linearity was 0.998, using Linearity. The recovery was linear for the medium- and
six calibrators. An example of a calibration curve is shown high-concentration samples with urinary iodine 0.20
in Fig. 3. mol/L. The recovery of urinary iodine was 92106% for
concentrations 0.20 mol/L (Table 5).
Detection limit. The theoretical detection limit of urinary
iodine, defined as 2 SD from the zero calibrator, was 0.11 Comparison with other methods. The APDM method was
mol/L (range, 0.071.4 mol/L), based on five analyses. compared with two other methods by the regression
analysis. The correlation between the APDM and the
Precision. At medium and high concentrations of urinary
iodine, the intraassay CVs were 1.72.0%, and the inter-
Table 3. Effects of potassium thiocyanate, ascorbic acid,
assay CVs were 4.4 4.5% (Table 1). At a concentration of
and ferrous ammonium sulfate on urinary iodine
0.30 mol/L (Table 1, low 2), the interassay CV was
measurement (n 3).
10%, whereas at 0.15 mol/L (Table 1, low 1), the CV Measured iodine, mol/L
was 20%. The working detection limit (tentatively defined
as the lowest concentration measured with an interassay Interfering substances, Iodate Iodate solution
mmol/L Urine solution (without digestion)
CV 10% ) was estimated as 0.3 mol/L.
Control 1.13 1.46 1.36
Potassium thiocyanate
Recovery. The recoveries of iodine added to urine samples
0.1 1.10 1.44 3.15
at different concentrations were 89 109% (Table 2). 1 1.12 1.42 3.15
4 1.16 1.43 3.15
Effect of interfering substances. In the assay without diges- 16 1.18 1.39 3.15
tion, the addition of potassium thiocyanate (0.1 mmol/ 64 1.09 1.34 3.15
L), ascorbic acid (4 mmol/L), or ferrous ammonium Ascorbic acid
sulfate (16 mmol/L) significantly increased the esti- 0.1 1.09 1.33 1.33
mated concentration of iodine in the iodate solution. On 1 1.07 1.31 1.53
the other hand, in the assay with digestion, the addition of 4 1.08 1.39 2.14
up to 16 mmol/L (final concentrations) of these interfer- 16 1.10 1.40 3.15
ing compounds did not affect the results for the urine 64 0.97 1.31 3.15
sample or the iodate solution (Table 3). Ferrous ammonium sulfate
0.1 1.11 1.42 1.35
Digestion of iodocompounds. The recovery of iodine from 1 1.11 1.42 1.35
four organic iodinated compounds (p-iodobenzoic acid, 4 1.12 1.38 1.41
16 1.10 1.34 1.78
m-iodophenol, 2-iodophenol, and 3-iodotyrosine) was as-
64 0.96 1.21 3.15
sayed with digestion. The mean recovery (as inorganic
534 Ohashi et al.: Microplate Method for Urinary Iodine

Table 4. Recovery of iodine from iodocompounds (n 3).

Iodine concentration, mol/L

Iodocompound Expected Measured Recovery, %

p-Iodobenzoic acid 2.02 1.89 0.07 94
m-Iodophenol 2.27 2.37 0.04 104
2-Iodophenol 2.27 2.24 0.05 99
3-Iodo-L-tyrosine 1.63 1.72 0.05 106

ICP/MS was good for measurements of urine samples

with iodine concentrations 3.15 mol/L (n 61; r
0.979; y 0.962x0.03; Syx 0.20). In a subset of the urine
samples with iodine concentrations 0.79 mol/L, which
were classified as iodine deficient, the correlation between
the two methods also was linear (n 27; r 0.978; y
1.124x; Syx 0.05). Comparison with the conventional
chloric acid digestion method showed a higher correlation
than that with ICP/MS, with a small standard error of the
Fig. 4. Difference plots for the two comparisons.
estimate (n 70; r 0.991; y 0.944x0.04; Syx 0.10).
(A), ICP/MS method vs APDM method: d 0.00; SD 0.20; d 1.96 SD
Using the difference plot recommended by Bland and 0.39; d 1.96 SD 0.39 mol/L. (B), conventional method vs APDM
Altman (6 ), we compared APDM with the other methods: method: d 0.01; SD 0.11; d 1.96 SD 0.23; d 1.96 SD 0.21
mol/L.
the conventional chloric acid digestion method and ICP/
MS, respectively. On the abscissa, we plotted mean values
of the two methods compared instead of a reference 4B. The result suggested that the difference plot analysis
method because, to date, none of the three methods has
between the two methods is consistent with the regression
been accepted as the reference method (79 ). Comparison
analysis, showing a wider distribution than that of APDM
between APDM and the conventional method gave a
vs the conventional method.
mean difference (d) of 0.01 mol/L, a SD for the differ-
ences of 0.11 mol/L, and a distribution of d 1.96 SD
evaluation of the method as a public health
0.23 mol/L to d 1.96 SD 0.21 mol/L for urinary
iodine concentrations of 0.16 3.15 mol/L (Fig. 4A). For application
the samples with iodine concentrations 0.79 mol/L as We evaluated APDM as a public health application by
described above in the regression analysis, we obtained a comparing it and the conventional and ICP/MS methods,
much narrower distribution (SD 0.05 mol/L) of dif- using the two sets of urinary samples described above.
ferences between APDM and the conventional method. The proportion of samples below specific cutoff points
The difference plot for APDM vs ICP/MS is shown in Fig. and medians are shown in Table 6. The World Health
Organization and ICCIDD proposed the use of these
cutoffs for interpreting urinary iodine results as follows:
Table 5. Serial dilution of urine samples in water (n 3).
severe deficiency, urinary iodine concentration 0.16
Urinary iodine, mol/L
mol/L; moderate deficiency, urinary iodine concen-
Sample Dilution Measured Expected Recovery, % tration, 0.16 0.39 mol/L; mild deficiency, urinary
1 (Low) 1 0.23 0.02 iodine concentration, 0.39 0.79 mol/L; and normal,
0.8 0.19 0.03 0.18 101 urinary iodine concentration 0.79 mol/L (1, 10 ).
0.6 0.17 0.03 0.14 121 The three methods gave similar distribution patterns.
2 (Medium) 1 1.32 0.01
Particularly, in the case of APDM vs the conventional
0.8 1.01 0.01 1.05 96
method, we obtained good agreement for 70 samples
0.6 0.74 0.00 0.79 93
0.4 0.50 0.02 0.53 94
between the two methods. The median values obtained
0.2 0.24 0.00 0.26 92 with the conventional and APDM methods were 0.67 and
0.1 0.14 0.02 0.13 103 0.68 mol/L, respectively; the median values obtained
3 (High) 1 3.16 0.05 with the ICP/MS and APDM methods were 0.44 and 0.47
0.8 2.68 0.02 2.53 106 mol/L, respectively. The Spearman rank correlation
0.6 2.01 0.04 1.89 106 coefficients for both comparisons were 0.990 (APDM vs
0.4 1.33 0.03 1.26 105 the conventional method) and 0.957 (APDM vs the ICP/
0.2 0.66 0.02 0.63 105 MS), respectively (P 0.0001). These results indicated
0.1 0.30 0.02 0.32 94 acceptable interpretative agreement between two meth-
0.05 0.19 0.03 0.16 120
ods in each comparison.
 Clinical Chemistry 46, No. 4, 2000
Table 6. Statistics of urinary iodine measured with three methods.
Conventional vs APDM (n 70)
Sample by IDD classification
Type Cutoff
Severe 0.16 mol/L
Severe moderate 0.39 mol/L
Severe moderate mild 0.79 mol/L
Median, mol/L
Spearman rank correlation (r)
ICP/MS vs APDM (n 61)
Sample by IDD classification
Type Cutoff
Severe 0.16 mol/L
Severe moderate 0.39 mol/L
Severe moderate mild 0.79 mol/L
Median, mol/L
Spearman rank correlation (r)
Discussion
The SandellKolthoff reaction with chloric acid digestion
has been used extensively as a simple and sensitive
method for the estimation of iodine in urine. Recently, a
nonhazardous persulfate digestion has been reported (4 ).
However, this method also is time-consuming, and the
amount of toxic waste generated is not negligible. To
speed the procedure and to minimize the amount of toxic
waste, we developed a new method that uses the micro-
plate format for all processes, including the digestion
process as well as the SandellKolthoff reaction. For the
digestion process, a PP microplate, which is heat-resistant
and chemically inert, was found to be most suitable. To
prevent leakage of vapor and cross-contamination be-
tween the contents of the wells of the microplate, a sealing
cassette was specially designed (Fig. 1). Using this cassette
(commercially available at approximately $1500), we ob-
served no substantial loss of sample volume and no
cross-contamination between wells.
In our study, the recovery of iodine added to urine
showed very large variation when digestion was insuffi-
cient. We tried raising the temperature and/or prolonging
the digestion time, and found the optimum conditions:
60-min digestion in 110 C standard oven in our system.
Although it is natural that the recovery is less with
insufficient digestion, a long digestion time also de-
creased recovery. We surmise that some substances pro-
duced from urine during digestion may have been re-
sponsible for the decrease in recovery. This may depend
mainly on the characteristics of urinary samples. Further
study will be required to make this clear. Regarding
interference, ascorbic acid, potassium thiocyanate, and
ferrous ammonium sulfate in concentrations up to 16
mmol/L did not affect the SandellKolthoff reaction after
digestion in our study, which is consistent with a previous
report by Pino et al. (4 ).
The cutoff points proposed by WHO/UNICEF/IC-
535
Distribution, %
Conventional APDM
1 0
24 23
56 57
0.67 0.68
0.990 (P 0.0001)
Distribution, %
ICP/MS APDM
16 23
48 43
59 56
0.44 0.47
0.957 (P 0.0001)
CIDD for classifying iodine deficiency are based on me-
dian urinary iodine concentrations. As an indicator of
iodine deficiency elimination, they proposed that the
median iodine concentration should be 0.79 mol/L, i.e.,
50% of samples should be above 0.79 mol/L, and not
more than 20% of samples should be below 0.39 mol/L.
The intra- and interassay CVs for samples with iodine
concentrations of 0.30 3.15 mol/L were 10% for the
APDM method. The APDM method has sufficient preci-
sion to assess the indicator of iodine deficiency elimina-
tion (10 ). On the other hand, because the interassay CV
was 20% for samples at 0.16 mol/L, in the case of
monitoring of urinary iodine for severe iodine deficiency
it is necessary to take into consideration the measurement
error.
The APDM method was compared with the conven-
tional chloric digestion method and the ICP/MS method,
which is said to be the most sensitive method for urinary
iodine detection (5, 7, 8, 11, 12 ). There were good correla-
tions between the APDM method and other two methods.
In the difference plot, trends or shifts between these
methods were not observed.
In addition, we compared the distributions and medi-
ans of urinary iodine concentrations measured by three
methods (Table 6). The Spearman rank correlation coeffi-
cients for the APDM method vs the conventional and
ICP/MS methods were 0.99 and 0.95, respectively. The
three methods gave similar distributions. In particular, in
comparison of the APDM method and the conventional
chloric digestion method, good agreement was obtained
for 70 samples between the two methods.
The only equipment required for the APDM method is
an automated microplate reader, which is widely used in
assays for thyroid-stimulating hormone (one of the bio-
chemical indicators of IDD). Because the SandellKolthoff
reaction is a kinetic reaction, ideally the interval between
addition of the cerium solution to a well and reading by
536 Ohashi et al.: Microplate Method for Urinary Iodine
the microplate reader would be the same for all wells.
However, if an automated reader (reading time, 20 50 s)
and a multichannel pipette (pipetting time, 60 s) are
used, the time lag between pipetting and reading is at
most 40 s, which is only 2% error for a 30-min reaction.
The number of samples that could be assayed in a day is
300 500 if three sealing cassettes are used. Recently, one
of the co-authors (M.G. Karmarker) has successfully ap-
plied this technique in Nepal to analyze 7740 urine
samples in the IDD survey with a proper internal quality
assessment. The required testing period, which had been
estimated as 3 4 months for the conventional method,
was only 20 days by our method.
In conclusion, our new method has a good detection limit
(0.11 mol/L) and precision with a dynamic range of
0.30 3.15 mol/L (CV 10%), and also reduces assay
time to 2 h for 80 samples per microplate. The micro-
plate digestion method is almost identical to the conven-
tional method in classifying iodine deficiency or suffi-
ciency. The APDM method is advantageous from the
viewpoint of safety, ease of operation, and stability of
ammonium persulfate. The APDM method enables easy,
portable, rapid, and nonhazardous urinary iodine detec-
tion. Although further interlaboratory comparisons are
necessary, we believe that the performance of the APDM
method may be useful for the purpose of urinary iodine
monitoring.
We thank Dr. Sangsom Sinawat (Nutrition Division, Min-
istry of Public Health, Thailand), Dr. Ananda B. Joshi
(Department of Community Medicine, Nepal), Dr.
Masamine Jimba (School & Community Health Project,
HMG/JICA/JMA, Nepal), and Dr. Harumichi Ito and
Chieri Yamada (Maternal and Child Health Project, Mon-
golia) for assessing our method. We also thank Drs.
Tomoyuki Igari and Tsune Kuroishi (Himalayan Green
Club, Japan) for collecting and sending urine samples.
References
1. Dunn JT, Crutchfield HE, Gutekunst R, Dunn AD. Methods for
measuring iodine in urine. The Netherlands: International Council
for Control of Iodine Deficiency Disorders, 1993:716.
2. Sandell EB, Kolthoff IM. Micro determination of iodine by catalytic
method. Microchem Acta 1937;1:9 25.
3. Zak B, Willard HH, Myers GB, Boyle AJ. Chloric acid method for
determination of protein-bound iodine. Anal Chem 1952;24:
1345 8.
4. Pino S, Fang S, Braverman LE. Ammonium persulfate: a safe
alternative oxidizing reagent for measuring urinary iodine. Clin
Chem 1996;42:239 43.
5. Yoshinaga J, Morita M. Determination of iodine in biological fluids
by ICP-MS. Biomed Res Trace Elements 1991;2:143 4.
6. Bland JM, Altman DG. Statistical methods for assessing agree-
ment between two methods of clinical measurement. Lancet
1986;i:30710.
7. May SL, May WA, Bourdoux PP, Pino S, Sullivan KM, Maberly GF.
Validation of a simple, manual, urinary iodine method for estimat-
ing the prevalence of iodine-deficiency disorders, and inter-labora-
tory comparison with other methods. Am J Clin Nutr 1997;65:
14415.
8. Haldimann M, Zimmerli B, Als C, Gerber H. Direct determination of
urinary iodine by inductively coupled plasma mass spectrometry
using isotope dilution with iodine-129. Clin Chem 1998;44:817
24.
9. Peterson PH, Stockl D, Blaabjerg O, Pedersen B, Birkemose E,
Thienpont L, et al. Graphical interpretation of analytical data from
comparison of a field method with a reference method by use of
difference plots. Clin Chem 1997;43:2039 46.
10. World Health Organization. Indicators for assessing iodine defi-
ciency disorders and their control through salt iodination. WHO
publication NUT/94.6. Geneva: WHO, 1994.
11. Poluzzi V, Cavalchi B, Mazzoli A, Alberini G, Lutman A, Coan P, et
al. Comparison of two different inductively coupled plasma mass
spectrometric procedures and high-performance liquid chromatog-
raphy with electrochemical detection in the determination of
iodine in urine. J Anal At Spectrom 1996;11:731 4.
12. Stark HJ, Mattusch J, Wennrich R, Mroczek A. Investigation of the
IC-ICP-MS determination of iodine species with reference to
sample digestion procedures. J Anal Chem 1997;359:371 4.