Journal of Clinical Endocrinology & Metabolism
, doi:10.1210/jc.2006-2156
The Journal of Clinical Endocrinology & Metabolism Vol. 92, No. 3 1019-1022
Copyright © 2007 by The Endocrine Society
Urine Iodine Measurements, Creatinine Adjustment, and Thyroid Deficiency in an Adult United States Population
James E. Haddow,
Monica R. McClain,
Glenn E. Palomaki and
Joseph G. Hollowell
Women and Infants Hospital (J.E.H., M.R.M., G.E.P.), Providence, Rhode Island 02903; and University of Kansas Medical Center (J.G.H.), Kansas City, Kansas 66160
Address all correspondence and requests for reprints to: James E. Haddow, M.D., Division of Medical Screening, 70 Elm Street, 2nd Floor, Providence, Rhode Island 02903. E-mail: jhaddow{at}ipmms.org.
 |
Abstract
|
|---|
Purpose: The purpose of the study was to explore the relationships between urine iodine, creatinine, serum TSH, and total T4 in a diverse population of U.S. adults with the aim of determining whether low urine iodine is associated with thyroid deficiency.
Methods: Using the Health and Nutrition Surveys III data set, we examined median TSH and total T4 values according to deciles of urine iodine (with and without creatinine correction). Stepwise regression analysis was used to further explore these relationships in the context of possible confounding variables. Exclusion criteria included age less than 21 yr, pregnancy, and the presence of thyroid antibodies.
Results: Among the 5963 men and 5722 women, median urine iodine concentrations do not vary with increasing age, whereas median creatinine levels decrease (P < 0.001). Urine iodine and creatinine concentrations are lower among women (P < 0.001). TSH increases with age (P < 0.001), whereas total T4 decreases (P < 0.001). Neither TSH nor total T4 median values are associated with urine iodine. If the urine iodine to creatinine ratio is used instead, an extensive redistribution of study subjects occurs that results in an apparent positive relationship between this ratio and TSH measurements. A multivariate regression analysis that accounts for age, body mass index, race, creatinine, iodine, estrogen use, smoking, and gender reveals only a weak association between levels of urine iodine and markers of thyroid function.
Conclusions: Our analyses indicate that the U.S. nonpregnant adult population is iodine sufficient.
|
Introduction
|
|---|
ARECENT ANALYSIS involving children and adults from 19 countries supports World Health Organization (WHO) criteria for assessing iodine status within any given population, using single urine iodine measurements (1). A population’s median urine iodine value is optimal between 100 and 200 µg/liter, with no more than 20% of individual values less than 50 µg/liter (2). The iodine distribution’s lower tail takes into account varying between-day intake (3). Data from three U.S. National Health and Nutrition Surveys [NHANES I (1971–1974), NHANES III (1988–1994), and NHANES 2001–2002] (4, 5, 6) all satisfy or exceed WHO guidelines for iodine sufficiency, but a fall in median urine iodine from 320 µg/liter (NHANES I) to 145 µg/liter (NHANES III) has led to concerns about decreasing iodine availability in the U.S. diet (7). However, NHANES 2001–2002, the most recent survey, documents a slightly increased median value (168 µg/liter). According to WHO criteria, a population is mildly iodine deficient when its median urine iodine is between 50 and 99 µg/liter; moderate deficiency is a population median between 20 and 49 µg/liter; and less than 20 µg/liter is severe deficiency (2). This has stimulated speculation that an individual whose urine iodine measurement is in this low range might display clinical or biochemical features of hypothyroidism, even within a population defined as iodine sufficient by its median urine iodine value and distribution (8, 9).
The large population sample in NHANES III offers an unusual opportunity to examine how iodine nutrition influences serum TSH and total T4 concentrations in the United States. The impact of using creatinine to adjust for concentration is also important to consider when interpreting urine iodine measurements. This commonly performed adjustment changes the distribution of individual measurements within the population and might either clarify or confound a possible association between urine iodine and thyroid deficiency. The present study further explored the complex interrelationships among urine iodine, creatinine, serum TSH, and total T4 in NHANES III, in the context of age, sex, and race/ethnicity. The purpose was to extend earlier analyses to determine whether any evidence can be found for iodine deficiency within this cohort (5) and also to examine the impact of creatinine correction.
|
Subjects and Methods
|
|---|
NHANES III (1988–1994) was designed to give normative estimates of health and nutrition status for the U.S. civilian, noninstitutionalized population (5). Documentation of consent is in the original publication along with details regarding serum analysis for TSH, total T4 and thyroid antibodies, and urine analysis for iodine and creatinine. Young children, older adults, African-Americans, and Mexican Americans were oversampled. Examinees fasted 10–16 h before morning examinations or 6 h before afternoon or evening examinations. Fasting urine samples are satisfactory for assessing a population’s iodine status (10).
This study focuses on 11,685 adult men and nonpregnant women age 21 yr and older whose thyroid peroxidase and thyroglobulin antibody measurements were negative and for whom TSH, total T4, iodine, and creatinine values were all available. Children and adolescents were excluded to avoid confounding effects when adjusting iodine for creatinine. Individuals with thyroid antibodies were excluded to avoid confounding effects on TSH. Other variables available for the analysis were gender, race, age, smoking status (serum cotinine 
10 ng/ml = smoker), estrogen use, and body mass index (BMI).
Logarithmic transformation was necessary for urine iodine, creatinine, and serum TSH and T4. Tests for trends and comparisons of medians were performed using PROC GLM and PROC NPAR1WAY, respectively, in SAS (SAS Institute, Cary, NC). In addition to standard univariate analyses, we used stepwise regression to further explore these relationships in the context of possible covariates. The main effects in the regression model were urine iodine, creatinine, age, BMI, race/ethnicity, gender, smoking status, and estrogen use. The initial model included all main effects (linear and squared values) along with all two-way interactions (11). Interaction terms and the squared effects were removed from the model in a stepwise fashion, if found to be nonsignificant (P > 0.05). SAS (version 9.1) was used to perform statistical analyses. We compared results with those from SUDAAN (12). There were no important differences, and the results in this report are from the SAS analyses.
|
Results
|
|---|
Table 1
shows median serum total T4 and TSH, urine iodine, creatinine, and iodine to creatinine ratio (I/Cr) values for the 5963 men and 5722 women, stratified by decade of age, and separately by race/ethnicity. The serum total T4 medians are higher among women than men (P < 0.0001). Total T4 decreases with age (P < 0.0001), whereas TSH increases (P < 0.0001) (5). The median urine iodine concentration follows a U-shaped distribution with age, whereas creatinine shows a steady decrease (P < 0.001). The creatinine association with sex and age is known (11). In the respective age groups, median iodine and creatinine concentrations are both significantly lower among women (P < 0.002 and P < 0.0001, respectively). The median I/Cr, however, is higher among women (P < 0.0001) and increases steadily with age in both genders. Median T4 and TSH values are lower among African-Americans (P < 0.0001 and P < 0.001). The median I/Cr is also lower among African-Americans (P < 0.0001), mainly due to higher creatinine levels.
View this table:
[in this window]
[in a new window]
|
TABLE 1. Median values of total T4, TSH, urine iodine (I), creatinine (Cr), and I/Cr ratio for 5963 men and 5722 nonpregnant women aged 21 yr and older in NHANES III, stratified by age and race/ethnicity
|
|
Table 2 shows median T4 and TSH values, stratified according to decile of urine iodine concentration, before and after creatinine adjustment. Unadjusted iodine levels vary by nearly 17-fold between lowest and highest deciles. Before adjustment, median ages by iodine decile range between 41 and 45 yr, with no linear trend (P = 0.07). Neither T4 nor TSH shows a trend by iodine level. The largest proportion of TSH values above 6 mU/liter (1.8%) occurs at the highest decile of urine iodine, but there is no significant iodine-related trend. This absent relationship between urine iodine and thyroid measurements is also true for the first through ninth centiles and the 91st through 100th centiles (data not shown). For example, among the 118 adult individuals with lowest iodine levels (<16 µg/liter), the median TSH value is 1.6 mU/liter (nearly equivalent to the overall median level of 1.4 mU/liter). None of the 118 TSH values are above 6.0 mU/liter.
View this table:
[in this window]
[in a new window]
|
TABLE 2. Median values of iodine, total T4, TSH, and age of 11,685 men and nonpregnant women aged 21 and older who are antibody negative (thyroid peroxidase and thyroglobulin) from the NHANES III study, by deciles of urinary iodine
|
|
Table 2
also shows median T4 and TSH values and average age by iodine decile, after dividing the iodine value by creatinine to correct for urine concentration. The extensive redistribution of study subjects among deciles of iodine after creatinine correction is reflected in the steadily rising median age as the ratio of I/Cr increases (from 37 yr at the lowest decile to 57 yr at the highest decile). The upward trends in median TSH levels (P = 0.001) and proportion of individuals with TSH at or above 6 mU/liter (P = 0.0001) likely result from confounding, due to the strong positive association between age and TSH levels (Table 1). TSH median values in the first through ninth and 91st through 100th centiles are consistent with the findings by decile, along with the proportion with values above 6 mU/liter (data not shown). Median levels of T4 do not vary by decile of I/Cr ratio (P = 0.95).
Stepwise regression is a more structured approach to examine the relationship between iodine and TSH and T4 by formally accounting for possible confounding. We performed two regression analyses; first with TSH as the dependent variable and then with T4. The variables could be ranked for impact on TSH and T4 by allowing each variable’s observed values to range from the 10th to 90th centile. All main effects and/or interactions are significant at least at the P < 0.01 level. The impact of variables on TSH (most to least) in men are: age, race, BMI, smoking, creatinine, and iodine. Among women the rank is: age, race, BMI, estrogen use, creatinine, smoking, and iodine. The following example shows the minimal impact of iodine on TSH in this model: the TSH value of a 47-yr-old Caucasian male nonsmoker with a BMI of 27 kg/m2 and urinary iodine value of 500 µg/liter (90th centile) is 0.08 mU/liter lower than a similar male whose urinary iodine is 40 µg/liter (10th centile) (1.49 mU/liter vs. 1.57 mU/liter, respectively). For T4, the ranking for women is estrogen, age, race, creatinine, iodine, smoking, and BMI. The ranking for men is the same (estrogen is removed).
|
Discussion
|
|---|
The present study once again finds no association between low urine iodine and higher TSH and/or lower T4 measurements at the lowest deciles of uncorrected (or corrected) urine iodine. This lack of association persists, even at the first through ninth centiles of urine iodine. Individual cases of iodine deficiency would be most likely to occur in this region, but we found no excess of high TSH or low T4 values. The only hint of iodine-related thyroid deficiency is at the highest iodine decile, in which there is a slight increase in the percent of individuals with elevated TSH measurements, indicating a possible Wolff-Chaikoff effect. Our analysis also underscores the pitfalls associated with dividing urine iodine values by creatinine. The confounding created by such adjustment confuses the interpretation of TSH and/or T4 measurements at the population level.
The I/Cr ratio would be highly effective if all individuals excreted the same amount of creatinine, but such is not the case (13, 14). In the present data set, in which a broad age range is coupled with racial diversity and a balanced gender representation, men excrete more creatinine than women, larger individuals excrete more than smaller, African-Americans excrete more than Caucasians, and young adults of both genders excrete more than older adults. Given all of these sources of variability, creatinine adjustment of urine iodine measurements might be expected to be misleading. In population studies of this type, the best policy might be to analyze urine iodine measurements without correcting for creatinine, accepting that there are inherent limitations on what can be learned. If, however, creatinine is included in the calculations, then a multivariate regression model is necessary (11). Our analyses confirm that, on a population basis, the impact of iodine will be small in comparison with other known covariates.
Authors citing the NHANES III report often incorrectly characterize individuals with iodine concentrations less than 100 µg/liter as having varying degrees of iodine deficiency (9, 15, 16). This overstates what can be learned from single urine iodine measurements because these are useful only as a measure of overall population status. Given the great variability of iodine content of food in the United States (17, 18), it would be unlikely that a given individual would continue to ingest foods with low iodine over an extended period of time. This is confirmed in a 1-yr study of 15 euthyroid men in Denmark (an area of mild to moderate iodine deficiency) that shows that long-term iodine status cannot be determined from a single 24-h urine measurement (3). The present data do not support the contention that iodine deficiency is a problem for nonpregnant adults in the United States at this time. Potentially vulnerable population groups, such as pregnant women, still require monitoring, and continued attention needs to be paid to nutritional sources of iodine (19, 20).
|
Footnotes
|
|---|
Disclosure: The authors have nothing to disclose.
First Published Online January 2, 2007
Abbreviations: BMI, Body mass index; I/Cr, iodine to creatinine ratio; NHANES, National Health and Nutrition Surveys; WHO, World Health Organization.
Received October 3, 2006.
Accepted December 26, 2006.
|
References
|
|---|
- Delange F, de Benoist B, Burgi H 2002 Determining median urinary iodine concentration that indicates adequate iodine intake at population level. Bull World Health Organ 80:633–636[Medline]
- WHO/UNICEF/ICCIDD 1994 Indicators for assessing iodine deficiency disorders and their control through salt iodization. WHO/Nut 6:36
- Andersen S, Pedersen KM, Pedersen IB, Laurberg P 2001 Variations in urinary iodine excretion and thyroid function. A 1-year study in healthy men. Eur J Endocrinol 144:461–465[Abstract]
- Hollowell JG, Staehling NW, Hannon WH, Flanders DW, Gunter EW, Maberly GF, Braverman LE, Pino S, Miller DT, Garbe PL, DeLozier DM, Jackson RJ 1998 Iodine nutrition in the United States. Trends and public health implications: iodine excretion data from National Health and Nutrition Examination Surveys I and III (1971–1974 and 1988–1994). J Clin Endocrinol Metab 83:3401–3408[Abstract/Free Full Text]
- Hollowell JG, Staehling NW, Flanders WD, Hannon WH, Gunter EW, Spencer CA, Braverman LE 2002 Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab 87:489–499[Abstract/Free Full Text]
- Caldwell KL, Jones R, Hollowell JG 2005 Urinary iodine concentration: United States National Health And Nutrition Examination Survey 2001–2002. Thyroid 15:692–699[CrossRef][Medline]
- Lee K, Bradley R, Dwyer J, Lee SL 1999 Too much versus too little: the implications of current iodine intake in the United States. Nutr Rev 57:177–181[Medline]
- Soldin OP, Tractenberg RE, Pezzullo JC 2005 Do thyroxine and thyroid-stimulating hormone levels reflect urinary iodine concentrations? Ther Drug Monit 27:178–185[CrossRef][Medline]
- Utiger RD 2006 Iodine nutrition—more is better. N Engl J Med 354:2819–2821[Free Full Text]
- Thomson CD, Smith TE, Butler KA, Packer MA 1996 An evaluation of urinary measures of iodine and selenium status. J Trace Elem Med Biol 10:214–222[Medline]
- Barr DB, Wilder LC, Caudill SP, Gonzalez AJ, Needham LL, Pirkle JL 2005 Urinary creatinine concentrations in the U.S. population: implications for urinary biologic monitoring measurements. Environ Health Perspect 113:192–200[Medline]
- Shah B, Barnwell BG, Bieler GS 1997 SUDAAN user’s manual, release 7.5. Research Triangle Park, NC: Research Triangle Institute
- Kesteloot H, Joossens JV 1996 On the determinants of the creatinine clearance: a population study. J Hum Hypertens 10:245–249[Medline]
- Boeniger MF, Lowry LK, Rosenberg J 1993 Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: a review. Am Ind Hyg Assoc J 54:615–627[Medline]
- Delange F 2004 Optimal iodine nutrition during pregnancy, lactation and the neonatal period. In J Endocrinol Metab 2:1–12
- de Benoist B, Andersson M, Egli I, Takkouche B, Allen H 2004 WHO global database on iodine deficiency. Geneva: World Health Organization
- Pearce EN, Pino S, He X, Bazrafshan HR, Lee SL, Braverman LE 2004 Sources of dietary iodine: bread, cows’ milk, and infant formula in the Boston area. J Clin Endocrinol Metab 89:3421–3424[Abstract/Free Full Text]
- Pennington JA, Schoen SA 1996 Contributions of food groups to estimated intakes of nutritional elements: results from the FDA total diet studies, 1982–1991. Int J Vitam Nutr Res 66:342–349[Medline]
- Becker DV, Braverman LE, Delange F, Dunn JT, Franklyn JA, Hollowell JG, Lamm SH, Mitchell ML, Pearce E, Robbins J, Rovet JF 2006 Iodine supplementation for pregnancy and lactation-United States and Canada: recommendations of the American Thyroid Association. Thyroid 16:949–951[CrossRef][Medline]
- Glinoer D 2006 Iodine nutrition requirements during pregnancy. Thyroid 16:947–948[CrossRef][Medline]
Top
Abstract
Introduction
