A Simple Microplate Method with Improved Low Iodine Concentration Sensitivity in Urinary Iodine Measurement

Dear Editor:

Urinary iodine (UI) depends upon very recent dietary iodine intake and is a good marker for assessing iodine nutrition in a population. Among the analytical methods for UI measurement, the Sandell-Kolthoff (S-K) reaction, a redox reaction developed in 1937, remains the most commonly used method (1). The first step is to eliminate (by digestion) the interfering substances in the urine to avoid any interference in the subsequent S-K reaction. Heating urine with strong oxidizing agents such as chloric acid or the less hazardous and nonexplosive ammonium persulfate is the traditional method for urine digestion (2). An alternative to this procedure is the microplate method, which involves incorporating a tube digestion process and the S-K reaction to a microplate format, therefore simplifying the procedure and generating less harmful wastes (3). The digestion is carried out by sealing the microplate in a special stainless steel cassette, which is subsequently baked at 110°C for 60 min in an oven. However, this indirect heating design has the drawbacks of a long incubation time and difficulty in controlling the sample temperature and is not available in many countries. A thermal cycler designed for the polymerase chain reaction (PCR) has the advantage of precise programmable sample temperature control and is a good replacement for a sealing cassette for urine digestion. However, there is a limitation in methods using microplates for urine digestion [i.e. high interassay coefficient of variation (CV) at iodine concentrations <20 μg/L] (3,4). Since 20 μg/L is the cutoff criteria between moderate and severe iodine deficiency proposed by the World Health Organiation, the bias or low sensitivity of the assay might cause classification error, which will affect the iodine policy of a country and raise difficulties in monitoring and assessing the iodine status of a population.

In our laboratory, a thermal cycler is used instead of a sealing cassette for the digestion process of the microplate method, and this method was used to measure the iodine content of more than 10,000 urine samples obtained from The Nutrition and Health Survey in Taiwan (5). The reagent and the methods were the same as the microplate method described previously with some modifications (3). Twenty-five microliters of standard iodine solutions (0, 25, 50, 100, 200, 300, and 400 μg/L) and urine samples were pipetted into a MicroAmp® optical 96-well reaction plate, followed by adding 50 μL/well of 1.35 mol/L ammonium persulfate solution (final concentration 0.9 mol/L). After covering the plate with a reusable MicroAmp® 96-well full plate cover and centrifuging briefly in a desktop microplate centrifuge, a MicroAmp® optical film compression pad was placed on top of the full plate cover to reinforce the compression pressure. Digestion was performed in the GeneAmp® PCR System 9700 Fast Thermal Cycler at 95°C for 30 min and 4°C for 5 min. After digestion, the microplate was centrifuged at 1000 g for 3 min, followed by the S-K reaction step as previously described (3). Briefly, 50 μL aliquots of the resulting digests were transferred to a 96-well reading plate, in which 100 μL of 0.05 mol/L arsenious acid solution had been preloaded. After shaking the plate in the microplate reader, 50 μL of 0.019 mol/L ceric ammonium sulfate solution was added into each well as quickly as possible using a multichannel pipette. The absorbance of the reaction mixture was measured at 405 nm after sitting at room temperature (∼25°C) for 30 min. Samples with concentrations >400 μg/L were diluted with water to fit the calibration curve, while samples <25 μg/L were measured using a modified low UI protocol (LIP). For the LIP, double volumes (50 μL) of the standard iodine solutions (0, 6.25, 12.5, 25, and 50 μg/L) and urine samples were pipetted into the MicroAmp® optical 96-well reaction plate, followed by the addition of a half volume (25 μL/well) of high concentration ammonium persulfate solution (freshly prepared 2.5 mol/L, at a final concentration of 0.83 mol/L). The rest of the digestion and S-K reaction procedures were the same as the standard protocol.

The air tightness and cross-contamination tests showed no significant vapor leakage or cross-contamination between wells during the digestion process. The correlation coefficients for the linearity of the calibration curve of the standard protocol and LIP were >0.999 in more than 99% of the assays. The detection limits (LOD) of the standard protocol and LIP were 6.0 and 1.5 μg/L of iodine, respectively (LOD = 3 × S_0; S₀ was determined by extrapolating the standard deviations of 8 different runs of 5 low iodine concentration samples to zero concentration). The assay precision of the standard protocol and LIP are shown in Table 1. For the standard protocol, the intraassay CV of UI <10 μg/L and interassay CV of UI <20 μg/L were both >10%. However, the intra- and interassay CVs of identical samples measured using the LIP were <10% at concentrations in the range of 10–30 μg/L. The improvement of the intra- and interassay CVs measured by the LIP is sufficient to differentiate moderate and severe iodine deficiency in a population. For the recovery test, the mean iodine recovery of 12 urine samples (>20 μg/L) measured by the standard protocol was 101% (range 96–107%), and mean recovery for 4 samples with concentrations 6–20 μg/L measured by the LIP was 98% (range 96–99%) (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/thy). When urine samples were serially diluted to 5 μg/L, the coefficient of linearity was >0.998 (Supplementary Fig. S1) and better recoveries were noted for most samples of <50 μg/L when measured by the LIP than by the standard protocol (Supplemental Table S2).

The results of 102 external quality assurance samples from the External Quality Assurance Program (EQUIP; Centers for Disease Control and Prevention, Atlanta, GA), measured by inductively coupled plasma mass spectrometry, which is considered to be the “gold standard” for urine iodine measurement (1,3,4), were compared with those of the thermal cycler digestion method described in the current study. The correlation coefficient for the linearity between the two methods was 0.9982 for iodine concentrations in the range of 7.5–490 μg/L. Further comparing the results of the 12 external quality assurance samples of low UI (<25 μg/L) measured by the standard protocol and LIP with inductively coupled plasma mass spectrometry, the correlation coefficients were 0.9079 and 0.9774, respectively.

High interassay CV (≥20%) at iodine concentrations <20 μg/L is a common drawback in microplate digestion (3,4). The thermal cycler digestion method, with the supplementary use of the LIP, provides sufficient accuracy and precision to detect iodine concentrations >10 μg/L, and is therefore sensitive enough to classify and monitor severe iodine deficiency in a population.