Simultaneous determination of iodide and creatinine in human urine by flow analysis with an on-line sample treatment column

Abstract

This work presents the first flow system for direct analysis of iodide and creatinine suitable for screening of human urine samples. The system had a mini-column packed with strong anion exchange resin for on-line extraction of iodide. After injection of a sample on the column the unretained urine sample was analyzed for creatinine in one section of the flow system using the Jaffe's reaction with spectrometric detection at 520 nm. Iodide was eluted off with 1.42 mL 5 M NaNO₃. A 150 μL fraction of the eluate was analyzed in another section of the same flow system for iodide using the kinetic–spectrometric Sandell–Kolthoff reaction. At the optimum condition, the sample throughput was 12 samples per h. The linear working range covered the normal levels of iodide and creatinine in human urine: 0–200 μg I L⁻¹ and 50–1200 mg creatinine L⁻¹, respectively. Recoveries tested in 10 samples were 87–104% for iodide and 89–104% for creatinine. Bland–Altman plots (n = 50) showed that the scatter of the differences between values obtained by this method and those of reference methods, for both iodide and creatinine, was within mean ± 2SD.

---

1. Introduction

Iodine Deficiency Disorder (IDD) is a term for various adverse health effects resulting from inadequate iodine intake leading to insufficient production of thyroid hormones.¹ The consequence of iodine deficiency during pregnancy is well known for causing mental retardation in children.² Iodine deficiency can occur at any age and it is estimated that more than 2 billion individuals worldwide have an insufficient intake of iodine.³ In order to ensure sufficient iodine intake many countries have implemented salt iodization and iodine supplementation.⁴ However, these strategies need efficient monitoring to avoid excessive intake of iodine as well as evaluation of the current IDD status of the population.

Approximately 90% of the amount of ingested iodine is eventually excreted via urine.⁵ Therefore urinary iodine (UI) has been commonly used to monitor the intake of iodine and IDD status of the population. Since the iodide ion is the most predominant form of iodine in urine,^6,7 quantitation of iodide provides information on total iodine excreted in urine. The measured level of iodide is used for assessing the ‘iodine intake’ and ‘IDD status’ according to WHO/UNICEF/ICCIDD guidelines. For school-aged children, the epidemiological criteria for iodine status are: <20 μg I L⁻¹ (insufficient intake/severe IDD); 20–49 μg I L⁻¹ (insufficient intake/moderate IDD); 50–99 μg I L⁻¹ (insufficient intake/mild IDD); 100–199 μg I L⁻¹ (adequate intake/no IDD); 200–299 μg I L⁻¹ (more than adequate intake/risk of hyperthyroidism) and >300 μg I L⁻¹ (excessive intake/risk of adverse health consequences).^1,8

The most common method for determination of low levels of iodide in urine has been the kinetics-based Sandell–Kolthoff method.¹ However there have also been other methods presented in the literature for measurement of iodide in human urine, such as electrospray ionization tandem mass spectrometry,⁹ ion-pairing reversed-phase high performance liquid chromatography with amperometric detection,¹⁰ single-drop microextraction coupled with gas chromatography¹¹ and fluorescence enhancement, by iodide, of complex formation between carbon dots and mercury ion.¹² Other methods have been recently reviewed by Shelor and Dasgupta.¹³

The classical kinetic method based on the Sandell–Kolthoff reaction¹⁴ is still a method of choice due to its high sensitivity and cost-effectiveness. The method utilizes the catalytic activity of iodide in the redox reaction of Ce(IV) and As(III) in acidic medium and requires only a common spectrometer. Analysis can be carried out in various formats, including batch,^15,16 flow-based^17–20 and microtitre plate.^21–24 All these methods require sample treatment, such as ashing,¹⁵ digestion with chloric acid^{16,19–21,23} or persulfate^{17,18,21,22,24} to destroy organic interferences prior to the kinetic measurement. This step is labor intensive, especially when handling large number of samples. A previous method based on on-line dialysis was subject to positive error.^25,26 Since then there have been various flow methods for on-line total digestion with subsequent spectrometric detection, such as UV-persulfate digestion²⁷ or acid digestion with KMnO₄–K₂Cr₂O₇–H₂SO₄.²⁸

We present a new flow method with an on-line anionic exchange treatment for determination of iodide in urine. The strong anionic exchange resin (SAX) was packed in a glass column incorporated into the flow system to extract iodide from other urine matrices. Iodide was then eluted with sodium nitrate for analysis in a section of the flow system, where the Sandell–Kolthoff reaction was employed. We also incorporated another flow section for simultaneous measurement of creatinine in the unretained urine sample from the SAX column. This employed the Jaffe's reaction.²⁹ The system provided not only the iodide level but also the iodide/creatinine ratio (UI/UCr). This ratio of a spot urine sample is suitable for large scale screening of the iodine status of individuals, as compared to urine samples collected over a 24 h period.³⁰

2. Experimental

2.1 Reagents and samples

All chemicals used in this work were of analytical reagent grade. Deionized distilled water, obtained from the Branstead EASYpure II (USA) unit, was employed for preparation of all aqueous solutions.

Stock solution of iodide (1000 mg I L⁻¹) was prepared by dissolving 0.1303 g of KI powder (Merck, Germany) in 100.0 mL of water. Creatinine stock solution (2000 mg creatinine L⁻¹) was prepared by dissolving an accurate weight of approximately 0.2 g of anhydrous creatinine (Sigma, USA) in 100.0 mL of water. Working standard solutions for calibration were solutions of iodide and creatinine, serially diluted from the stock solutions.

Preparation of the reagents, used in the determination of iodide by the Sandell and Kolthoff reaction, was carried out similar to that reported in a previous work.¹⁹ The arsenious solution was prepared by first dissolving 10.0 g of As₂O₃ (Ajax, New Zealand) and 47.0 g of NaCl (Merck, Germany) in 500 mL of water. The mixture was heated until a clear solution was obtained. After cooling to room temperature, 27.8 mL of concentrated H₂SO₄ was added to the solution which was then made up with water to 1 L in a volumetric flask to give 0.10 M arsenious acid solution. The cerium(IV) solution (0.008 M) was prepared by dissolving 5.0 g of (NH₄)₄Ce(SO₄)₄·2H₂O (Merck, Germany) in a 1 L volumetric flask with 1.75 M H₂SO₄. The solution was allowed to stand overnight before use.

The alkaline picrate solution (0.03 M), used for the Jaffe's method, was freshly prepared by dissolving approximately 1.7 g of picric acid powder (Merck, Germany) in 250 mL of 0.4 M NaOH. The 5 M NaNO₃ eluting solution was prepared by dissolving 215 g of NaNO₃ (Ajax, New Zealand) in 500 mL of water. This solution was filtered through a 0.45 μm cellulose acetate membrane (Sartorius, Germany) before use.

For the validation study, 50 samples of spot urine from healthy volunteers were used. The urine samples were stored at 4 °C until analyzed. Prior to analysis, 5.00 mL of the urine sample was diluted five-fold with water.

2.2 The flow manifold

The optimized flow system for simultaneous determination of iodide and creatinine is shown in Fig. 1. PTFE tubing (1.0 mm i.d., VICI, Switzerland) was employed in the assembly of the flow system. An autosampler (Perkin Elmer AS90, USA) was used for sample loading. The flow system consisted of three peristaltic pumps, P₁, P₂ and P₃ (Ismatec/ISM827, Switzerland) and two 6-port injection valves (Upchurch, USA), IV₁ and IV₂, for loading and injection of the sample and the eluted zone of iodide, respectively. A 3-way switching valve, SV₁ (Upchurch, USA), was used for selecting the flow of the sample or water into the system. A 4-way switching valve, SV₂ (Upchurch, USA), was utilized for switching between the flow of the sample/water and the eluent, E, into the sample treatment column (SAX in Fig. 1). In ‘section b’ of the flow manifold, a relatively larger size tubing (3.17 mm i.d., 8 cm long), denoted ‘dilution tube’, was placed before the mixing coil MC to increase dilution of the urine sample in order to give a suitable signal amplitude in the analysis of creatinine.

Fig. 1 A schematic drawing of the optimized flow system with an on-line SAX column for simultaneous and direct analysis of urinary iodide and creatinine. The system is composed of three sections: ‘section a’ for sample treatment, ‘section b’ for creatinine analysis and ‘section c’ for iodide anlaysis. R₁: 0.03 M alkaline pricrate; R₂: 0.1 M As₂O₃ with 0.8 M NaCl in 0.5 M H₂SO₄; R₃: 0.008 M (NH₄)₄Ce(SO₄)₄ in 1.75 M H₂SO₄; E: 5 M NaNO₃; P₁, P₂ and P₃: peristaltic pump 1, 2 and 3; IV₁, IV₂: injection valve 1 and 2; L₁, L₂: loop 1 (50 μL) and loop 2 (150 μL); SV₁, SV₂: switching valve 1 and 2; RC₁, RC₂: reaction coil 1 and 2 (1.0 mm i.d., 200 cm long); MC: mixing coil (1.0 mm i.d., 50 cm long); dilution tube: 3.17 mm i.d., 8 cm long; SAX minicolumn: glass tube (2.2 mm i.d., 25 mm long) packed with 30 mg of strong anion exchange resin; WB: water bath; D₁ and D₂: LED detector 1 (520 nm) and 2 (420 nm).

As shown in Fig. 1, a water bath, WB (Fisher Scientific/Isotemp 205, UK), was used for maintaining the temperatures of reaction coils, RC₁ and RC₂, at 40 °C. A photometer, D₁ (Bangkok High Lab Co. Ltd, Thailand), equipped with a SMB420/525/640-3100-I LED (Epitex, Japan) as the light source and a 10 mm flow-through cell (Helma, Germany), was used in the creatinine analysis (section b, Fig. 1). In the iodide determination, a second photometer D₂ (Bangkok High Lab Co., Ltd, Thailand), equipped with a SMBD520-1100 LED (Epitex, Japan) as the light source and a 10 mm flow-through cell (Helma, Germany), was used for monitoring the absorbance of Ce(IV) (section c, Fig. 1). An in-house software written with Lab VIEW 8.2™ was used for recording the output from the photometers.

2.3 SAX column

The flow system also comprised a glass column (2.2 mm i.d., 25 mm long), packed with strong anion exchange resin (75–150 μm SAX, Alltech, USA), which had a maximum capacity of 900 μmol anion equivalent/600 mg resin. Slurry of the SAX-resin, made by mixing 30 mg of the resin with water, was loaded into the glass column which was stoppered at each end with cotton wool. Before use, the resin was conditioned by flowing a solution of 5 M NaNO₃ for 15 min, followed by washing with water for 5 min.

2.4 Flow procedure

The optimized operating sequence is shown in Fig. 2. The system operated under continuous-flow mode throughout the analysis. The steps in Fig. 2 describe the operations of the four electronic components SV₁, SV₂, IV₁ and IV₂. Fig. 2 shows the steps for triplicate analysis of a sample together with the signal profiles for iodide and creatinine. According to the procedure in Fig. 2, a 5.0 mL urine sample (after 5-fold dilution) or a standard mixture was selected by switching valves SV₁ and SV₂ to flow into the SAX column. Iodide ions were retained by the quaternary amine groups of the resin. Unretained components in the urine, including creatinine, flowed through the loop L₂ (IV₂ at the ‘load’ position) and loop L₁ (IV₁ at the ‘load’ position) to waste line W₂. In the middle of the period of trapping of iodide by the resin, the urine in loop L₁ was injected (IV₁ is at the ‘inject’ position) into ‘section b’ (Fig. 1) for measurement of creatinine. The urine sample from L₁ was mixed with a stream of water at the first confluence point before entering the dilution tube and then into the mixing coil MC. This diluted plug of urine then merged with the stream of the picrate reagent R₁, being mixed inside the reaction coil RC₁, before flowing into the detector, D₁. The signal profile for the first measurement of creatinine is shown in Fig. 2 with a signal height C₁. At the start of detection of creatinine (t = 2.5 min) injection valve IV₁ was set back to the ‘load’ position and switching valve SV₁ was switched, allowing water to rinse the SAX column for 30 seconds. The next procedure was the elution of iodide from the SAX column into sample loop L₂. For this step, SV₂ was switched to allow 5 M NaNO₃ eluent (E in Fig. 1) to flow into the SAX column at a flow rate of 1 mL min⁻¹ for 1.42 min. The eluate, containing iodide, flowed into the sample loop L₂. At 3.42 min, the valve IV₂ was switched to the ‘inject’ position to inject the eluate in loop L₂ into ‘section c’ for analysis of iodide by the Sandell–Kolthoff reaction. This sample zone merged and mixed with arsenious acid R₂ and ceric reagent R₃, respectively. The reaction zone passed through the heated reaction coil RC₂ for acceleration of the kinetics before flowing into the detector D₂. In the Sandell–Kolthoff reaction the iodide ion catalyzes the reduction of Ce(IV) and thus there is a decrease in absorbance when the sample zone is monitored at D₂. Thus for an injection of a sample there would be 2 flow profiles with the peaks for creatinine and iodide appearing almost simultaneously (see Fig. 2). Fig. 2 shows the flow profiles for triple injections of a sample.

Fig. 2 A graphical representation of the timing sequence of the valves and the corresponding absorbance signals for three replicate injections of a sample for the developed flow system (refer to Fig. 1). Section (a) gives the steps for the sample treatment, section (b) for the creatinine analysis and section (c) for the iodide analysis. At the start of the first injection, the switching valves SV₁ and SV₂ are set to allow the sample to flow through the SAX column to trap the iodide ion. Injection valves IV₁ and IV₂ are in the “load” positions. Unretained urine flows through sample loops L₂ and L₁, respectively, to waste line W₂. At 1.0 min injection valve IV₁ is set to the “inject” position to introduce the urine in the sample loop L₁ into section (b) of the system for analysis of creatinine by reacting with a flow of picric acid (R₁). The profile of the absorbance of the product of the Jaffe's reaction is shown as the signal (D₁). At 2.5 min, the time at which creatinine begins to be observed, switching valve SV₁ is set to allow H₂O to flow through the SAX column for 0.5 min. Also at 2.5 min injection valve IV₁ is set back to the load position. Then at 3.0 min the switching valve SV₂ is set to allow the eluent E (5 M NaNO₃) to elute the trapped iodide off the SAX column for 1.42 min and then set back. The eluted zone flows into the loop L₂. At 3.42 min the injection valve IV₂ is set to the “inject” position to introduce the iodide eluate in the sample loop L₂ into section (c) of the system for analysis of the iodide by the Sandell–Kolthoff kinetic method. The resulting absorbance signal as recorded at detector D₂ is shown as Signal (D₂). At 4.42 min injection valve IV₂ is set back to the “load” position. At 5.0 min switching valve SV₁ is set to allow the sample solution to flow into the SAX column, thus starting the second cycle of the analysis.

3. Results and discussion

3.1 System design and operation

3.1.1 Final design. The flow system in Fig. 1 comprises three sections: section a, ‘sample treatment’; section b, ‘creatinine analysis’; and section c, ‘iodide analysis’. We modified an off-line method for the separation of iodide from urine using SAX as reported by G. E. Abraham et al. in 2006.³¹ In this work, the SAX resin, packed in a glass column, was installed into the flow system in ‘section a’ for on-line clean-up of the urine sample. Unretained creatinine flowed into ‘section b’ for analysis using the Jaffe's reaction. An aliquot of iodide eluate, eluted with 5 M NaNO₃ solution, was delivered to ‘section c’ for kinetic determination of iodide using the Sandell–Kolthoff reaction.

3.1.2 Development of the flow protocol for iodide extraction and measurement. In order to optimize the flow for ‘sample treatment’ and ‘iodide analysis’, ‘section a’ and ‘section c’ were connected without ‘section b’ as shown in Fig. 3a (configuration I) or in Fig. 3b (configuration II), respectively. In configuration I, the SAX column was fitted to the six-port injection valve (IV₂) as the sample loop. In ‘configuration II’, the same SAX column was placed between switching valve SV₂ and a six-port injection valve (IV₂), to which a sample loop L₂ (150 μL) was installed. The 4 operation steps for ‘configuration I’ and ‘configuration II’ are listed in Table 1. Connections of the flow devices are also given in Table 1.

Fig. 3 Comparison of the two flow configurations for sample pretreatment (section a) and iodide analysis (section c). The recording of the absorbances at detector D₂ are for 3 replicate injections of blank sample and a standard solution of 20 μg I L⁻¹. (a) Configuration I: the SAX column is fitted to the injection valve IV₂, replacing the sample loop. (b) Configuration II: a second selection valve SV₂ is added in section a and the SAX column placed between SV₂ and injection valve IV₂, which now has a sample loop L₂ installed. This is the configuration selected for the flow system employed for analysis of urine samples. Using Configuration II the solution in sample loop L₂, when injected into section c (for iodide analysis) always has 5 M NaNO₃ (eluent E) as the solvent. Thus the blank signal is due to the effect of 5 M NaNO₃ on the kinetics of the Sandell–Kolthoff reaction (see Section 3.4).

Table 1 Step operation and device connection in configuration I and configuration II representing ‘section a’: sample treatment and ‘section c’: iodide analysis

Step operation	Configuration I		Configuration II
	Device connection	Duration	Device connection	Duration
1. SAX loading with urine	SV1(std)-IV2(SAX)-W2	2 min and 30 s	SV1(std)-SV2(SAX)-IV2-W2	2 min and 30 s
2. SAX cleanup using water	SV1(H2O)-IV2(SAX)-W2	30 s	SV1(H2O)-SV2(SAX)-IV2-W2	30 s
3. SAX elution using 5 M NaNO3 and iodide analysis in ‘section c’	‘Section c’ (E)-IV2(SAX)-W3	3 min and 15 s	(3.1) ‘Section a’ (E)-SV2(SAX)-IV2-‘section c’(H2O)	25 s
			(3.2) ‘Section a’ (E)-SV2(SAX)-IV2-W2	1 min
4. SAX washing with water	SV1(H2O)-IV2(SAX)-W2	30 s	SV1(H2O)-SV2(SAX)-IV2-W2	30 s
Total operation time/cycle	6 min and 45 s		4 min and 55 s

---

The design of ‘configuration I’ used ‘section a’ for introducing urine into the SAX column to retain iodide (step 1 in Table 1), followed by rinsing with water (step 2 in Table 1). Iodide was eluted from the column with 5 M NaNO₃ (carrier E) by the six-port injection valve IV₂ (step 3 in Table 1). The eluate directly merged and mixed with the reagents R₂ and R₃ in ‘section c’. It was observed that using the configuration there was not only the signal of iodide (i.e. I₁, I₂ or I₃), but also a large positive peak in front of the negative signal for iodide. This positive peak was also obtained for a blank sample. Water from the rinsing step that remained in the SAX column and in the connecting tubes was the cause of the positive peak. Since the carrier solution in ‘section c’ is 5 M NaNO₃ this water plug reduced the ionic strength of the carrier solution leading to a lower reaction rate and an increase in the background absorbance (vide infra). Thus ‘configuration I’ was not selected.

In ‘configuration II’ (Fig. 3b) a switching valve SV₂ was installed in ‘section a’ and the SAX column connected between this valve and the six-port injection valve IV₂. A sample loop L₂ (150 μL) was installed on the injection valve IV₂. The carrier line of ‘section c’ was changed from NaNO₃ (eluent E) to water (Fig. 3b). The NaNO₃ line now formed a part of ‘section a’ and its flow line was connected to the valve SV₂. The elution of iodide from SAX now constituted ‘section a’.

‘Configuration II’ (Fig. 3b) was operated according to steps 1 to 4 of Table 1. SV₁ was used in conjunction with SV₂ to flow urine, water or eluent (E) through the SAX column. In step 3.1 of Table 1, the volume of the total eluate zone that was in the sample loop L₂ was delivered to ‘section c’ on switching injection valve IV₂. The rest of the eluate was vented to waste line W₂ in step 3.2 of Table 1. The signal from the catalytic effect of iodide was now observed (Fig. 3b). The signal of the blank sample resulted from the increase of the uncatalyzed reaction rate due to the NaNO₃ in the eluent (E). This was shown by a separate experiment using only ‘section c’. The detector D₂ showed a baseline absorbance of 0.9 au, when water was used as the carrier. The baseline absorbance decreased to about 0.45 au when 5 M NaNO₃ was used as the carrier. This showed that increase in ionic strength led to an increase in the rate of the Sandell–Kolthoff reaction.

In the final system the waste line W₂ of ‘section c’ was connected to the six-port injection valve IV₁, which had a 50 μL sample loop L₁ (Fig. 1). Urine that passed through the resin column flowed into this sample loop L₁ (50 μL). The urine in L₁ was then injected into the carrier stream of water for further dilution and reaction with alkaline picrate (R₁) in ‘section b’.

3.2 Selection of the iodide eluate zone for kinetic analysis in ‘section c’

In the final manifold in Fig. 1, only the NaNO₃ eluate (containing iodide ions) in the 150 μL sample loop of injection valve IV₂ was sent to ‘section c’ for kinetic analysis. We therefore need to find the suitable time at which to inject the 150 μL eluate so that the eluate in the sample loop contained the highest concentration of iodide. Fig. 4 shows the signals for a blank (water) sample and iodide standard (100 μg I L⁻¹) in the NaNO₃ eluate at various injection times. The signals for the blank sample increased, reaching a constant value for injection times >30 s. As discussed previously, NaNO₃ increases the kinetics of the uncatalyzed reaction due to the increase in ionic strength. The head zone of the eluate will be diluted by the water remaining in the column, but the later eluting solution (eluting after 30 s) will be 5 M NaNO₃. The signals for the iodide sample also steadily increased with injection times, but reaching a maximum between 25 and 30 s. For longer injection times the signal gradually decreased to the level found for the blank sample (injection times >85 s). This indicated that there was no iodide remaining in the eluate, only 5 M NaNO₃. Thus injection of the eluate in the injection loop at 25 s was the most suitable injection time. The results also showed that a volume of 1420 μL of 5 M NaNO₃ was sufficient to elute all the sample iodide trapped on the column.

Fig. 4 Signals obtained for a blank (water) sample and 100 μg I L⁻¹ iodide (analyte) eluted from the on-line ion-exchange column with 5 M NaNO₃ (eluent). The signals were obtained from solution in the 150 μL sample loop (of injection valve IV₂) injected into ‘section c’ at various injection times.

3.3 Volume and flow rate of the samples

In this work urine samples were diluted 5-fold with water prior to analysis. Therefore a standard iodide solution of 20 μg I L⁻¹ was chosen as representative of the urine at the median level of iodine (100 μg I L⁻¹, the cut off level between mild IDD and normal). The amount of resin packed in the SAX column was capable of trapping 45 μmol of iodide ion. Volume, from 1 to 5 mL, of this solution was loaded on the SAX resin at a flow rate of 1 mL min⁻¹ (1.6 × 10⁻⁴ to 7.9 × 10⁻⁴ μmol I⁻). As expected, the iodide signal recorded at detector D₂ increased linearly with the loading volume (r² = 0.994, data not shown) indicating that the system had high efficiency in both trapping and elution of iodide ion. In this work, 5 mL was selected as the sample volume.

Loading flow rates of 0.5, 1.0, 1.5, 2.0 and 2.5 mL min⁻¹ were tested with no difference in the observed signals. However, it was found that the flow rate of 2.5 mL min⁻¹ led to back pressure after long use. A flow rate of 2.0 mL min⁻¹ was selected for loading the sample.

3.4 Investigation of the kinetics

For determination of iodide using the kinetics of Ce(IV)–As(III) reaction, it was proposed that the reaction should exhibit pseudo-first order kinetics in Ce(IV) concentration.^20,32 Then the relationship between Ce(IV) concentration and time can be expressed by [Ce(IV)]_t = [Ce(IV)]₀ e^−kt: where [Ce(IV)]_t is the concentration of Ce(IV) at time t; [Ce(IV)]₀ is the initial concentration of Ce(IV); k is the observed pseudo-first order rate constant. This rate constant is equal to (k_uncatalyzed + k_catalyzed[I⁻]), where the two rate constants are for the uncatalyzed and catalyzed processes, respectively. In flow-based analysis the absorbance of Ce(IV) is monitored at a fixed time t after mixing of the reagent(s) and sample. It can be shown that −ln A_t and [I⁻] are linearly related, as shown in eqn (1).

−ln A_t = −ln A_o + t_kuncatalyzed + t_kcatalyzed[I⁻] (1)

where A_o and A_t are the absorbance of Ce(IV) after mixing and at fixed time t, respectively.

The flow manifold of ‘section c’ (Fig. 1) was constructed similar to the manifold presented for the determination of total inorganic iodine in drinking water.³² However, the current manifold has a 40 °C water bath to accelerate the kinetics and improve the sensitivity. For this new flow system, it was necessary to adjust the system to ensure that the Sandell–Kolthoff reaction followed first-order kinetics. A stopped-flow mode^20,32 was employed in the flow system ‘section c’ to investigate the kinetics for five concentrations of iodide (20 to 100 μg I L⁻¹), prepared in 5 M NaNO₃ (to simulate the iodide in the eluate). An exponential decrease in absorbance with time (Fig. S1‡) was observed for all the iodide solutions. Using Sigma plot software³³ it was shown that the exponential fittings had less than 1% fitting errors, comparable to a batch mixing.

3.5 Length of the dilution tube for creatinine analysis

In this work, creatinine was measured using the flow manifold ‘section b’ (Fig. 1). It had been recommended that for the Jaffe's method the urine sample should be diluted approximately 50-fold before reacting with the picrate reagent.³⁴ Thus a large diameter tube (3.17 mm i.d.), the dilution tube, was added to further dilute the urine sample and also to improve the mixing efficiency. Table 2 shows the analysis times and percent recoveries obtained when using various lengths of dilution tube. The experiment was carried out using pooled urine from 5 subjects. It was observed that the recovery decreased to almost 100% as the length of the dilution tube increased, i.e. with increasing dilution. 8 cm tubing was selected for a faster analysis time.

Table 2 Optimization of the length of the dilution tube, carried out using pooled urine (from five samples), and its effect on recovery of the creatinine analysis of ‘section b’

Tube length (cm)	Calculated value		Analysis time (minutes per injection)	%Recovery (n = 3)
	Tube volume (μL)	Dilution factor^a (fold)
a Dilution factor = (volume of dilution tube)/50 μL of L1.
2	160	3.2	3.00	118 ± 2.5
4	320	6.4	3.50	102 ± 4.7
6	470	9.4	4.00	100 ± 6.5
8	630	12.6	4.50	100 ± 0.7
10	790	15.8	5.00	101 ± 0.4

---

3.6 Analytical performance and validation

Using the optimal conditions discussed previously, the flow system in Fig. 1 gave simultaneous measurements of iodide and creatinine in urine. The linear range for iodide was 0–200 μg I L⁻¹ (−ln A_t = (0.0109 ± 0.0001) μg I L⁻¹ + (0.2485 ± 0.0045): r² = 0.999) and for creatinine the linear range was 50–1200 mg creatinine L⁻¹ (ΔA = (0.0007 ± 0.0001) mg creatinine L⁻¹ + (0.0015 ± 0.0045): r² = 0.999) (Fig. S2‡). The limits of detection (LOD, 3σ) were 1 μg I L⁻¹ and 12 mg creatinine L⁻¹ for iodide and creatinine, respectively. The method had a precision of 1.5% RSD (n = 10) (for 20 μg I L⁻¹) and 6.1% RSD (for 50 mg creatinine L⁻¹). Recoveries for 10 urine samples were 87–104% for iodide and 90–104% for creatinine, respectively. The method had a throughput of 12 samples per hour. The SAX column was robust and self-regenerated to the nitrate form during the elution of iodide with the sodium nitrate eluent.

The developed method was applied to the determination of iodide and creatinine in 50 urine samples collected from volunteers with no known history of IDD. These values were compared with values obtained using the reference methods by means of the Bland–Altman plot³⁵ (see Fig. 5a and b for iodide and creatinine, respectively). The plots showed that all data lay within ±2SD of the mean of their differences, showing that this method was equivalent to the reference methods. Pearson's correlations (Fig. S3‡)³⁵ also confirmed that our method gave results that did not differ significantly from values using the reference methods (r = 0.988 for iodide, r = 0.992 for creatinine).

Fig. 5 Bland–Altman plots for comparing the data obtained from our method for (a) iodide, compared with chloric acid digestion method¹⁹ and (b) creatinine, compared with the standard Jaffe's method.²⁹

As reported in the literature,^36,37 the levels of urinary iodine and creatinine showed diurnal fluctuation over a 24 hour period (Fig. 6). However, the iodide to creatinine ratios (UI/UCr) almost eliminated this diurnal fluctuation (see Fig. 6a and b). Therefore, this method allows the evaluation of the UI/UCr ratio of a spot urine sample which can replace a 24 h urine sample for evaluating the IDD status of a large number of subjects. A recent paper³⁸ also measured the UI/UCr ratio in urine dried on filter paper strips. The extracted sample was oxidized with persulphate prior to analysis by the Sandell–Kolthoff reaction. A separate Jaffe's method was employed for creatinine determination.

Fig. 6 Levels of iodide, creatinine and iodide/creatinine ratio of spot urine samples collected over 24 hours from two subjects, (a) and (b).

4. Conclusion

To our knowledge, this is the first system that provides simultaneous determination of iodide and creatinine in urine, and hence the UI/UCr ratio. Thus survey of the iodine status of a population can be conveniently carried out using spot urine. The urine sample is directly injected into the flow manifold without any sample pre-treatment, except a 5-fold dilution with water. An in-line SAX column in the nitrate form extracts iodide from the urine matrix while the non-retained urine is analyzed for creatinine content using alkaline picrate. A selected fraction of the eluate from the SAX column is analyzed for iodide using the Sandell–Kolthoff reaction. A 5 M NaNO₃ solution is used for eluting iodide from the SAX column as well as for regenerating the SAX resin for the next sample. The SAX column has been employed for more than 200 injections. We believe that this system is a step towards resolving the outstanding problem of a facile sample preparation as commented in the review of Shelor and Dasgupta.¹³

Acknowledgements

The work was financially supported by the research grant (B04/2554) of Mahidol University (DN), S&T Research Grant of the Thailand Toray Science Foundation, 2010 (DN) and the Office of Higher Education Commission and Mahidol University under the National Research Universities Initiative, chaired by Prof. Manat Pohmakotr. We would like to thank the Thailand Research Fund (TRF) for providing the student scholarship of the Royal Golden Jubilee Ph.D. program to JS. We are grateful to the postdoc scholarship of Mahidol University (TM). The authors would also like to acknowledge the Center of Excellence for Innovation in Chemistry (PERCH-CIC), Commission on Higher Education, Ministry of Education for the support of equipment and partial scholarships for students (JS and PI). The authors wish to thank all the volunteers for providing us with urine samples.

References

1. World Health Organization, United Nations Children’s Fund, International Council for the Control of Iodine Deficiency Disorders, in Assessment of iodine deficiency disorders and monitoring their elimination – A guide for programme managers, WHO, Geneva, 3rd edn, 2007.
2. A. Melse-Boonstra and N. Jaiswal, Best Pract. Res., Clin. Endocrinol. Metab., 2010, 24, 29–38.
3. M. B. Zimmermann, Endocr. Rev., 2009, 30, 376–408.
4. M. Andersson, B. D. Benoist and L. Rogers, Best Pract. Res., Clin. Endocrinol. Metab., 2010, 24, 1–11.
5. M. B. Zimmermann, P. L. Josste and C. S. Pandav, Lancet, 2008, 372, 1251–1262.
6. B. Michalke and P. Schramel, Electrophoresis, 1999, 20, 2547–2553.
7. B. Michalke, P. Schramel and H. Witte, Biol. Trace Elem. Res., 2000, 78, 67–79.
8. P. L. Jooste and E. Strydom, Best Pract. Res., Clin. Endocrinol. Metab., 2010, 24, 77–88.
9. K. Minakata, I. Yamagishi, S. Kanno, H. Nozawa, M. Suzuki and O. Suzuki, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2010, 878, 1683–1686.
10. V. T. P. Nguyen, V. Piersoel and T. E. Mahi, Talanta, 2012, 99, 532–537.
11. M. Hu, C. Chen, Y. Jiang and H. Zhu, Chem. Pap., 2013, 67, 1255–1261.
12. F. Du, F. Zeng, Y. Ming and S. Wu, Microchim. Acta, 2013, 180, 453–460.
13. C. P. Shelor and P. K. Dasgupta, Anal. Chim. Acta, 2011, 702, 16–36.
14. E. B. Sandell and I. M. Kolthoff, J. Am. Chem. Soc., 1934, 56, 1426.
15. G. Aumont and J. C. Tressol, Analyst, 1986, 111, 841–843.
16. J. T. Dunn, H. E. Crutchfield, R. Gutekunst and A. D. Dunn, Thyroid, 1993, 3, 119–123.
17. S. Pino, S. L. Fang and L. E. Braverman, Clin. Chem., 1996, 42, 239–243.
18. S. Pino, S. L. Fang and L. E. Braverman, Exp. Clin. Endocrinol. Diabetes, 1998, 106, S22–S27.
19. D. Nacapricha, S. Muangkaew, N. Ratanawimarnwong, J. Shiowatana and K. Grudpan, Analyst, 2001, 126, 121–126.
20. D. Nacapricha, N. Ratanawimarnwong, J. Suwannachoat, P. Wilairat, J. Shiowatana and K. Grudpan, Anal. Sci., 2001, 17, i33–i36.
21. T. Ohashi, M. Yamaki, C. S. Pandav, M. G. Karmarkar and M. Irie, Clin. Chem., 2000, 46, 529–536.
22. H. Hussain and W. N. W. Mohamud, Trop. Biomed., 2006, 23, 109–115.
23. M. Hedayati, M. Khazan, P. Yaghmaee, M. Z. Yeghaneh, L. Behdadfar and M. S. Daneshpour, Clin. Chem. Lab. Med., 2011, 49, 281–284.
24. A. Mina, E. J. Favaloro and J. Koutts, J. Trace Elem. Med. Biol., 2011, 25, 213–217.
25. W. May, D. Wu, C. Eastman, P. Bourdoux and G. Maberly, Clin. Chem., 1990, 36, 865–869.
26. H. C. Ford and L. A. Johnson, Clin. Chem., 1991, 37, 759.
27. K. Tsuda, H. Namba, T. Nomura, N. Yokoyama, S. Yamashita, M. Izumi and S. Nagataki, Clin. Chem., 1995, 41, 581–585.
28. Z. Yaping, Y. Dongxing, C. Jixiang, L. Tianshiu and C. Huiqin, Clin. Chem., 1996, 42, 2021–2027.
29. R. W. Bonsnes and H. H. Taussky, J. Biol. Chem., 1945, 158, 581–591.
30. P. Vejbjerg, N. Knudsen, H. Perrild, P. Laurberg, S. Andersen, L. B. Rasmussen, L. Ovesen and T. Jørgensen, Thyroid, 2009, 19, 1281–1286.
31. G. E. Abraham, R. C. Handal and J. C. Hakala, The Original Internist, 2006, 13, 125–135.
32. N. Choengchan, K. Lukkanakul, N. Ratanawimarnwong, W. Waiyawat, P. Wilairat and D. Nacapricha, Anal. Chim. Acta, 2003, 499, 115–122.
33. Sigma plot®, Programming Guide, SPSS Inc., Chicago, 2001.
34. T. Songjaroen, T. Maturos, A. Sappat, A. Tuantranont and W. Laiwattanapaisal, Anal. Chim. Acta, 2009, 647, 78–83.
35. S. A. Glantz, in Primer of Biostatistics, The McGRAW-HILL, New York, 3rd edn, 2005, ch. 8, pp. 305–310.
36. L. B. Rasmussen, L. Ovesen and E. Christiansen, Eur. J. Clin. Nutr., 1999, 53, 401–407.
37. F. KÖnig, M. Andersson, K. Hotz, I. Aeberli and M. B. Zimmermann, J. Nutr., 2011, 14, 2049–2054.
38. T. T. Zava, S. Kapur and D. T. Zava, Anal. Chim. Acta, 2013, 764, 64–69.

---

Footnotes

† This work is dedicated to the 60^th Birthday of Prof. Dr Kate Grudpan.

‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/c4an01224k

---