A longitudinal study of iodine excretion in normal pregnancy determined by inductively coupled plasma mass spectrometry Carl J. Wardley,a Alan Cox,b Cameron McCleodb and Brian W. Morris*a aDepartment of Clinical Chemistry, Northern General Hospital, Herries Road, ShefÆeld, UK S5 7AU. E-mail: brian@clinchem-ngh.demon.co.uk bCentre for Analytical Sciences, University of ShefÆeld, UK Received 28th July 1999, Accepted 6th September 1999 We report here on a method for the determination of urine iodine by ICP-MS.The method proved to be fast, reliable and precise (within batch CVs v2.5%) and to have a low limit of detection (0.000 38 mmol L21). Iodine was determined in a total of 379 urine samples from 86 healthy pregnant women who gave samples at intervals of 4 weeks from 16 weeks pregnancy to 10 weeks post-partum. Fifty-Æve control urine samples were analysed from age-matched non-pregnant females. Iodine excretion (mmol of iodine per mol of creatinine) increased signiÆcantly from 28±40 weeks of pregnancy (pv0.05), returning to non-pregnant control levels by 10 weeks post-partum. This study conÆrms the ability of our ICP-MS method to analyse large numbers of patient samples with the speed and performance acceptable for a routine assay. Iodine deÆciency has been a world-wide health problem for centuries1 manifesting itself, in its most severe form, as endemic cretinism.Globally, it is estimated that one billion people are at risk from iodine deÆciency today with developing countries being the primary areas of concern.Europe, however, is not exempt and a number of countries, e.g., Denmark, The Netherlands and Northern Italy, are amongst those with a low natural iodine supply.2 Iodine deÆciency has been described as the world's most common and preventable endocrine disease and dietary iodine requirements have been studied in many population groups with urine levels in pregnancy of particular interest.3±5 In healthy pregnant women, the regulation of thyroid function depends upon a number of factors that act independently to increase thyroid hormone requirements.(i) A marked increase in the binding capacity of serum due to high circulating levels of thyroid binding globulin (TBG). (ii) Direct stimulation of the thyroid gland by human chorionic gonadotrophin (hCG) acting as a thyrotrophic hormone. (iii) Increased placental de-iodinating activity which may contribute to thyroid hormone metabolism.The combined result of these events is a physiological adaptation of the maternal thyroid gland to pregnancy so long as the availability of iodine remains sufÆcient. Reduced maternal thyroid concentrations during pregnancy can inØuence foetal neurological development but the effects of iodine deÆciency disorders (IDD) occur at all stages of human development, from the foetus through to adulthood. Throughout gestation, a reduced supply of iodine is associated with chronic stimulation of the thyroid gland and increased renal loss of iodine has been suggested as the cause of thyroid enlargement, the so-called `pregnancy goitre', although there are conØicting reports in the available literature.6 Most ingested iodine eventually appears in the urine and urinary iodine concentration is therefore a good marker for iodine deÆciency.The analysis of urinary iodine has been based, historically, on simple colorimetric procedures allowing reasonably rapid determination of small numbers of samples.7 Our aim in the current study was to investigate the basic physiology of iodine excretion during normal pregnancy, using a highly sensitive ICP-MS technique that required minimal sample preparation and manipulation.Experimental Samples Some 86 healthy pregnant women gave a total of 379 urine samples at intervals of four weeks from 16 weeks gestation to term and up to 10 weeks post-partum. Fifty-Æve age-matched females supplied one random urine sample to provide a nonpregnant reference interval for our study.All urine samples were collected into `iodine-free' sterile plastic universal containers (Medical Wire and Equipment Co., Ltd, Corsham, Wiltshire, UK) and stored at 220 �C prior to analysis. Approval for our study was obtained from the North ShefÆeld Hospital Ethics Committee and informed consent was obtained from each volunteer. Iodine analysis Iodine in urine was determined using a Hewlett-Packard HP4500 ICP mass spectrometer (Palo Alto, CA, USA) Ætted with a Babington V-groove nebuliser and glass spray chamber (Scott double pass), together with a CETAC (Omaha, NE, USA) ASX-500 autosampler for ease of sampling (see Table 1 for instrument settings). Potassium iodate (Sigma Aldrich, Poole, Dorset, UK) was used for the preparation of an aqueous stock standard solution (1000 mg L21), subsequently diluted to give a working stock standard of 100 mg L21.We utilised the method of additions calibration for our analysis using matrix-matched working standards covering the calibration range 5±100 mg L21 (5, 10, 15, 20, 50 and 100 mg L21) together with a urine blank solution.Polypropylene tubes were used throughout (Sarstedt, Leicester, UK). Samples were presented to the ICP via the CETAC ASX-500 autosampler and initially aspirated for 40 s at a high pump Table 1 Main ICP operating parameters (HP4500) Forward power 1300 W Plasma gas Øow 16 L min21 Auxiliary gas Øow 1.0 L min21 Nebuliser gas Øow 1.25 L min21 Sampling depth 7.5 mm J.Anal. At. Spectrom., 1999, 14, 1709±1710 1709 This Journal is # The Royal Society of Chemistry 1999speed (0.30 rps y1 mL min21). The system was stabilised for 30 s (pump speed 0.15 rps y0.5 mL min21) prior to analysis. Data acquisition (pump speed 0.15 rps) was based on measurement in the `spectrum' mode (3 points per mass at 1 s per point for 127I) and 5 replicate measurements were used for each solution (total integration time 15 s).A high pump speed of 0.3 rps for 60 s was used for washout prior to the introduction of the next sample. Quality control and patient samples were diluted 1 : 20 with distilled water and values calculated from the calibration curve. The Ænal urine results were expressed as mmol of iodine per mol of creatinine to correct for urine Øow. Creatinine was measured in a Vitros E250 analyser (Ortho Clinical Diagnostics, Rochester, NY, USA) using dry slide technology.Results and discussion Method performance and quality assurance The Hewlett-Packard 4500 ICP-MS is suitable as a fast screening analytical method for the determination of urinary iodine. Our method gave recoveries, at three levels of iodine (2.36, 3.15 and 3.94 mmol L21) of between 101±104%, a detection limit of 0.000 38 mmol L21 and precision (assessed at three levels) ranged from 1.3±2.3% intra-batch to 6.3±12.5% inter-batch (Table 2).Patients There were no signiÆcant differences in urine creatinine excretion between our pregnant and control groups. Log transformed data showed that during the Ærst 24 weeks of pregnancy, iodine excretion did not change signiÆcantly over that seen in our control population (mean value 85.1 mmol iodine per mol creatinine, range 30±245). SigniÆcant increases were seen, however, commencing at 28 weeks (pv0.05), reaching peak values at 40 weeks (mean 120.2 mmol of iodine per mol of creatinine, range 54±268, pv0.05).Iodine levels returned to control levels by 10 weeks post delivery (Fig. 1). The availability of iodine for the maternal thyroid during pregnancy results from a combination of speciÆc factors and is critically reduced by nutritional deÆciency. Although nutritional deÆciency is unlikely to be a factor in our patient group, the conÆrmation of a signiÆcant increase in iodine excretion during the second and third trimesters might suggest a negative iodine balance in the absence of increased intake.Increased iodine excretion might contribute to the increase in thyroid size reported in pregnancy and although (as we have shown) iodine excretion returns to normal post-partum, is interesting to speculate that thyroid changes in pregnancy might pre-dispose susceptible individuals to thyroid disease in later life. The analysis of iodine using the Hewlett-Packard 4500 ICPMS is a fast screening method for urine estimation. The method is reliable and precise, with low limits of detection and high levels of recovery.The availability of this method offers the opportunity for further work in this area, with particular reference to iodine excretion in patients with renal failure, patients using iodine-containing medication, e.g., amiodarone, and in the differential diagnosis of silent thyroiditis and Graves' disease. References 1 C. A. Furnee, F. van der Haan, C. E. West and J. G. A. Hautvast, Am.J. Clin. Nutr., 1994, 59, 1415. 2 J. Brug, M. R. H. Lowik, M. Wedel and J. Odink, Eur. J. Clin. Nutr., 1992, 46, 671. 3 E. J. Silva and S. Silva, J. Clin. Endocrinol. Metab., 1981, 4, 671. 4 P. Caron, M. Hoff, S. Bazzi, A. Dufor, G. Faure, I. Ghandour, P. Lauzu, Y. Lucas, D. Maraval, F. Mignot, P. Ressigeac, F. Vertongen and V. Grange, Thyroid, 1997, 5, 749. 5 K. M. Pederson, P. Laurberg, E. Iverson, P. R. Knudsen, H. E. Gregersen, O. S. Rasmussen, K. R. Larsen, G. M. Eriksen and P. L. Johannesen, J. Clin. Endocrinol. Metab., 1993, 77, 1078. 6 P. P. A. Smyth, M. T. Hetherton, D. F. Smith, M. Radcliff and C. O'Herlihy, J. Clin. Endocrinol. Metab., 1993, 82, 2840. 7 S. Pino, S. L. Fang and L. E. Braverman, Clin. Chem., 1996, 42, 239. Paper 9/06138J Table 2 Intra- and inter-batch quality control data for iodine estimation using the HP4500 ICP-MS Intra-batch Inter-batch Quality control (n~20) Quality control (n~10) Level/mmol L21 0.76 2.22 3.29 0.76 2.22 3.29 Mean 0.72 2.16 3.03 0.72 2.22 3.21 s 0.01 0.05 0.04 0.09 0.14 0.27 RSD (%) 1.4 2.3 1.3 12.5 6.3 8.4 Fig. 1 Iodine excretion in normal pregnancy: mmol of iodine per mol of creatinine. (Values are means of log transformed data.) *~pv0.05. 1710 J. Anal. At. Spectrom., 1999, 14, 1709±1710