Abstract
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 Sandell–Kolthoff reaction, 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 microplate and the Sandell–Kolthoff reaction was performed 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 μmol/L (14 μg/L iodine). The mean intra- and interassay CVs for samples with iodine concentrations of 0.30–3.15 μmol/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; Sy|x = 0.10) and with the inductively coupled plasma mass spectrometry method (n = 61; r = 0.979; y = 0.962x + 0.03; Sy|x = 0.20). The agreement was confirmed by difference plots. The distributions 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.
The urinary iodine concentration is a good biochemical marker for the control of iodine deficiency disorder (IDDs)1 (1). Most of the popular methods for urinary iodine determination are based on the Sandell–Kolthoff reaction (2). However, substances interfering with the Sandell–Kolthoff 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 Deficiency Disorders (ICCIDD) has compiled seven methods, based on the Sandell–Kolthoff 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 Sandell–Kolthoff 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 recently as a nonhazardous, nonexplosive, and easy-to-use method (4). The persulfate digestion makes possible a comparatively nonhazardous (no chlorine gas) measurement. However, this method is still not completely suitable for testing because it is time-consuming and produces 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 volumetrically equal and to avoid contamination throughout 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 (Corning 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⇓ ).
Sealing cassette.
The cassette is used for digestion after a microplate is placed inside.
The colorimetric measurements were performed in a microplate reader (IMMUNO-MINI; Nalge Nunc International).
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 Erlenmeyer 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 suspension was filtered with a glass filter (5–10 μm 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 hydroxide 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).
Tetraammonium 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 μmol/L (5–600 μg/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 μL each) were pipetted into the wells of a polypropylene (PP) plate, followed by the addition of 100 μL of ammonium persulfate 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-μL 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 μL) was added to the wells and mixed; 50 μL 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.
Inductively coupled plasma mass spectrometry (ICP/MS) method.
An SPQ 8000A1 ICP/MS analyzer (Seiko Instruments) was used. The procedure was carried out according 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 μL) were added to 16 × 160 mm test tubes, and 750 μL 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 ammonium 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 μL) 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 sucking 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 μmol/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 μmol/L). Optimum digestion was defined as having sufficient recoveries of iodine added to various urine samples (final added iodate concentration, 0.79 μmol/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 conversion 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 μmol/L (25, 50, 100, 200, 300 and 400 μg/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 detection 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 experiment was performed on 30 different days.
Recovery.
The recovery of iodine was estimated by assaying in triplicate 12 different urine samples supplemented with potassium iodate solution, and comparing the results with those of water-added urine samples. The iodate-added urine was prepared by adding a given volume (1/10 volume of the urine sample) of potassium iodate solution (3.94 μmol/L) to the urine samples. For the water-added urine, water was added instead of potassium iodate solution. The percentage of recovery was calculated by the following equation: 
Effect of interfering substances.
The effects of three interfering substances (potassium thiocyanate, l-ascorbic acid, and ferrous ammonium sulfate) on the assay were assessed 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 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 process. The recovery of iodine from each iodocompound was estimated by comparing the measured iodine concentration with the calculated concentration.
Linearity.
Pooled urine samples containing low, medium, and high concentrations of iodine were serially diluted with water. The testing was carried out in triplicate.
Comparison with other methods.
Pearson and Spearman correlations were applied to the results. Comparison of the APDM method and the ICP/MS method was performed using a total of 61 urine samples from Mongolia and Pakistan with iodine concentrations of 0.079–3.15 μmol/L. Comparison of the APDM method with the conventional chloric acid digestion method was performed using 70 urine samples from Nepal.
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 concentration of the control plate was 1.17 ± 0.03 μmol/L (range, 1.13–1.28 μmol/L), and that of test plate was 1.21 ± 0.05 μmol/L (range, 1.11–1.34 μmol/L). The recovery of iodine added to urine as potassium iodide (100 μmol/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 ) ammonium persulfate (final concentration), an oven temperature 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 submaximal conditions, the recovery was 0–100%, depending on the urine samples.
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.
assay evaluation
Calibration curve.
The absorbance on a log scale was linear for iodine concentrations between 0 and 3.15 μmol/L. The correlation coefficient for the linearity was >0.998, using six calibrators. An example of a calibration curve is shown in Fig. 3⇓ .
Calibration curve.
Detection limit.
The theoretical detection limit of urinary iodine, defined as 2 SD from the zero calibrator, was 0.11 μmol/L (range, 0.07–1.4 μmol/L), based on five analyses.
Precision.
At medium and high concentrations of urinary iodine, the intraassay CVs were 1.7–2.0%, and the interassay CVs were 4.4–4.5% (Table 1⇓ ). At a concentration of 0.30 μmol/L (Table 1⇓ , low 2), the interassay CV was ∼10%, whereas at 0.15 μmol/L (Table 1⇓ , low 1), the CV was 20%. The working detection limit (tentatively defined as the lowest concentration measured with an interassay CV <10% ) was estimated as ∼0.3 μmol/L.
Precision.
Recovery.
The recoveries of iodine added to urine samples at different concentrations were 89–109% (Table 2⇓ ).
Recovery of iodine added to urine as iodate.
Effect of interfering substances.
In the assay without digestion, the addition of potassium thiocyanate (≥0.1 mmol/L), ascorbic acid (≥4 mmol/L), or ferrous ammonium sulfate (≥16 mmol/L) significantly increased the estimated concentration of iodine in the iodate solution. On the other hand, in the assay with digestion, the addition of up to 16 mmol/L (final concentrations) of these interfering compounds did not affect the results for the urine sample or the iodate solution (Table 3⇓ ).
Effects of potassium thiocyanate, ascorbic acid, and ferrous ammonium sulfate on urinary iodine measurement (n = 3).
Digestion of iodocompounds.
The recovery of iodine from four organic iodinated compounds (p-iodobenzoic acid, m-iodophenol, 2-iodophenol, and 3-iodotyrosine) was assayed with digestion. The mean recovery (as inorganic iodine) for these compounds was 101% of the expected value (Table 4⇓ ).
Recovery of iodine from iodocompounds (n = 3).
Linearity.
The recovery was linear for the medium- and high-concentration samples with urinary iodine >0.20 μmol/L. The recovery of urinary iodine was 92–106% for concentrations >0.20 μmol/L (Table 5⇓ ).
Serial dilution of urine samples in water (n = 3).
Comparison with other methods.
The APDM method was compared with two other methods by the regression analysis. The correlation between the APDM and the ICP/MS was good for measurements of urine samples with iodine concentrations <3.15 μmol/L (n = 61; r = 0.979; y = 0.962x+0.03; Sy|x = 0.20). In a subset of the urine samples with iodine concentrations <0.79 μmol/L, which were classified as iodine deficient, the correlation between the two methods also was linear (n = 27; r = 0.978; y = 1.124x; Sy|x = 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 estimate (n = 70; r = 0.991; y = 0.944x+0.04; Sy|x = 0.10). Using the difference plot recommended by Bland and Altman (6), we compared APDM with the other methods: 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 method because, to date, none of the three methods has been accepted as the reference method (7)(8)(9). Comparison between APDM and the conventional method gave a mean difference (d) of 0.01 μmol/L, a SD for the differences of 0.11 μmol/L, and a distribution of d + 1.96 SD = 0.23 μmol/L to d − 1.96 SD = −0.21 μmol/L for urinary iodine concentrations of 0.16–3.15 μmol/L (Fig. 4⇓ A). For the samples with iodine concentrations <0.79 μmol/L as described above in the regression analysis, we obtained a much narrower distribution (SD = 0.05 μmol/L) of differences between APDM and the conventional method. The difference plot for APDM vs ICP/MS is shown in Fig. 4B⇓ . The result suggested that the difference plot analysis between the two methods is consistent with the regression analysis, showing a wider distribution than that of APDM vs the conventional method.
Difference plots for the two comparisons.
(A), ICP/MS method vs APDM method: d = 0.00; SD = 0.20; d + 1.96 SD = 0.39; d − 1.96 SD = −0.39 μmol/L. (B), conventional method vs APDM method: d = 0.01; SD = 0.11; d + 1.96 SD = 0.23; d − 1.96 SD = −0.21 μmol/L.
evaluation of the method as a public health application
We evaluated APDM as a public health application by comparing it and the conventional and ICP/MS methods, using the two sets of urinary samples described above. The proportion of samples below specific cutoff points 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: “severe deficiency”, urinary iodine concentration <0.16 μmol/L; “moderate deficiency”, urinary iodine concentration, 0.16–0.39 μmol/L; “mild deficiency”, urinary iodine concentration, 0.39–0.79 μmol/L; and “normal”, urinary iodine concentration >0.79 μmol/L (1)(10).
Statistics of urinary iodine measured with three methods.
The three methods gave similar distribution patterns. Particularly, in the case of APDM vs the conventional method, we obtained good agreement for 70 samples between the two methods. The median values obtained with the conventional and APDM methods were 0.67 and 0.68 μmol/L, respectively; the median values obtained with the ICP/MS and APDM methods were 0.44 and 0.47 μmol/L, respectively. The Spearman rank correlation coefficients for both comparisons were 0.990 (APDM vs the conventional method) and 0.957 (APDM vs the ICP/MS), respectively (P <0.0001). These results indicated acceptable interpretative agreement between two methods in each comparison.
Discussion
The Sandell–Kolthoff 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 microplate format for all processes, including the digestion process as well as the Sandell–Kolthoff 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 between 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 observed 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 insufficient. 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 decreased recovery. We surmise that some substances produced from urine during digestion may have been responsible 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 Sandell–Kolthoff 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/ICCIDD for classifying iodine deficiency are based on median urinary iodine concentrations. As an indicator of iodine deficiency “elimination”, they proposed that the median iodine concentration should be 0.79 μmol/L, i.e., 50% of samples should be above 0.79 μmol/L, and not more than 20% of samples should be below 0.39 μmol/L.
The intra- and interassay CVs for samples with iodine concentrations of 0.30–3.15 μmol/L were ≤10% for the APDM method. The APDM method has sufficient precision to assess the indicator of iodine deficiency “elimination” (10). On the other hand, because the interassay CV was 20% for samples at ∼0.16 μmol/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 conventional 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 correlations 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 medians of urinary iodine concentrations measured by three methods (Table 6⇑ ). The Spearman rank correlation coefficients 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 biochemical indicators of IDD). Because the Sandell–Kolthoff reaction is a kinetic reaction, ideally the interval between addition of the cerium solution to a well and reading by 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 applied 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 μmol/L) and precision with a dynamic range of 0.30–3.15 μmol/L (CV ≤10%), and also reduces assay time to ∼2 h for 80 samples per microplate. The microplate digestion method is almost identical to the conventional method in classifying iodine deficiency or sufficiency. 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 detection. 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.
Acknowledgments
We thank Dr. Sangsom Sinawat (Nutrition Division, Ministry 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, Mongolia) for assessing our method. We also thank Drs. Tomoyuki Igari and Tsune Kuroishi (Himalayan Green Club, Japan) for collecting and sending urine samples.
Footnotes
↵1 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.
1 Water-added urine was prepared by adding 1 volume of water to 9 volumes of urine.
2 Iodate-added urine was prepared by adding 1 volume of iodate solution (3.94 μmol/L) to 9 volumes urine.
- © 2000 The American Association for Clinical Chemistry
