Thyroid hormone levels and incident chronic kidney disease in euthyroid individuals: the Kangbuk Samsung Health Study

Transcription

1 International Journal of Epidemiology, 2014, doi: /ije/dyu126 Advance Access Publication Date: 10 July 2014 Original article Original article Thyroid hormone levels and incident chronic kidney disease in euthyroid individuals: the Kangbuk Samsung Health Study Yiyi Zhang, 1 Yoosoo Chang, 1,2,3 Seungho Ryu, 1,2,3 Juhee Cho, 1,3,4 Won-Young Lee, 5 Eun-Jung Rhee, 5 Min-Jung Kwon, 6 Roberto Pastor-Barriuso, 7 Sanjay Rampal, 1,8 Won Kon Han, 9 Hocheol Shin 10 * and Eliseo Guallar 1 * 1 Department of Epidemiology and Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD, USA, 2 Department of Occupational Medicine, 3 Center for Cohort Studies, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, South Korea, 4 Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University, Seoul, South Korea, 5 Department of Endocrinology and Metabolism, 6 Department of Laboratory Medicine, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, South Korea, 7 National Center for Epidemiology, Carlos III Institute of Health and Consortium for Biomedical Research in Epidemiology and Public Health (CIBERESP), Madrid, Spain, 8 Department of Social and Preventive Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia, 9 Department of Surgery and 10 Department of Family Medicine, Kangbuk Samsung Hospital and Sungkyunkwan University School of Medicine, Seoul, South Korea *Corresponding authors. Hocheol Shin, Department of Family Medicine, Kangbuk Samsung Hospital and Sungkyunkwan University School of Medicine. Pyung-dong, Jongro-Gu, , Seoul, South Korea. Tel. þ(82) Fax þ(82) hcfm.shin@samsung.com; Eliseo Guallar, Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins Bloomberg School of Public Health, 2024 E Monument Street, Baltimore, MD 21205, USA. eguallar@jhsph.edu Accepted 3 June 2014 Abstract Background: Overt and subclinical hypothyroidism are associated with higher levels of serum creatinine and with increased risk of chronic kidney disease (CKD). The prospective association between thyroid hormones and kidney function in euthyroid individuals, however, is largely unexplored. Methods: We conducted a prospective cohort study in South Korean men and women who were free of CKD and proteinuria at baseline and had normal thyroid hormone levels and no history of thyroid disease or cancer. At each annual or biennial follow-up visit, thyroid-stimulating hormone (TSH), free triiodothyronine (FT3) and free thyroxin (FT4) levels were measured by radioimmunoassay. The study outcome was incident CKD, defined as an estimated glomerular filtration rate (egfr) <60 ml/min/1.73 m 2 based on the Chronic Kidney Disease Epidemiology Collaboration creatinine equation. VC The Author 2014; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association 1624

2 International Journal of Epidemiology, 2014, Vol. 43, No Results: After a median follow-up of 3.5 years, 1032 participants developed incident CKD. There was a positive association between high-normal levels of TSH and increased risk of incident CKD. In fully-adjusted models including baseline egfr, the hazard ratio comparing the highest vs the lowest quintiles of TSH was 1.26 [95% confidence interval (CI) 1.02 to 1.55; P for linear trend ¼ 0.03]. In spline models, FT3 levels below 3 pg/ml were also associated with increased risk of incident CKD. There was no association between FT4 levels and CKD. Conclusions: In a large cohort of euthyroid men and women, high levels of TSH and low levels of FT3, even within the normal range, were modestly associated with an increased risk of incident CKD. Key words: Thyroid hormones, thyrotropin, thyroxin, triiodothyronine, kidney disease Key Messages Overt and subclinical hypothyroidism are associated with risk of chronic kidney disease (CKD). Some evidence from cross-sectional studies suggests that thyroid hormone levels within the normal range may also be associated with CKD, but no prospective data were available. Our study was the first prospective study to evaluate the association between thyroid hormones and incident CKD among a large cohort of euthyroid men and women. We found that high-normal TSH and low-normal FT3 levels were associated with an increased risk of incident CKD. Our results suggest that low thyroid function, even within the euthyroid range, may affect kidney function. Introduction Overt and subclinical hypothyroidism are associated with higher levels of serum creatinine and with increased risk of chronic kidney disease (CKD) However, most of the evidence has been derived from cross-sectional studies of patients with either clinical thyroid dysfunction or with kidney disease. Prospective study examining the direction of the association between thyroid function and kidney function is lacking, although some clinical data reported improvement of renal function after thyroid replacement therapy. 5,7 9,11 Moreover, recent studies in euthyroid individuals have demonstrated associations between lownormal thyroid function and several cardiovascular and metabolic risk factors, including higher fasting glucose, increased insulin resistance, a more atherogenic lipid profile and low-grade inflammation, suggesting that variations in thyroid hormone levels within the normal range may also modulate cardiovascular risks However, few studies have examined the association between thyroid function and CKD in euthyroid individuals, and all of them used a cross-sectional design There are no prospective data regarding the association between thyroid hormone levels and incident CKD in euthyroid individuals. The objective of this study was to evaluate the prospective associations of TSH, free triiodothyronine (FT3) and free thyroxin (FT4) with incident CKD among participants with normal thyroid hormone levels in the Kangbuk Samsung Health Study, a large cohort study of apparently healthy South Korean men and women. Materials and methods Study population The Kangbuk Samsung Health Study is a cohort study of South Korean men and women, 18 years of age or older, who underwent a comprehensive annual or biennial health examination at the clinics of the Kangbuk Samsung Hospital Health Screening Center in Seoul and Suwon, South Korea, from 2002 to Over 80% of participants were employees of various companies and local governmental organizations and their spouses. In South Korea, the Industrial Safety and Health Law requires annual or biennial health screening examinations of all employees, offered free of charge. The remaining participants voluntarily purchased screening at the health examination centre. The present analysis included all study participants with at least one follow-up visit between 1 January 2002 and 28 February 2010 (n ¼ ; Figure 1). We excluded participants with missing data at baseline or at any

3 1626 International Journal of Epidemiology, 2014, Vol. 43, No. 5 Figure 1. Flow chart of study participants. follow-up visit on serum creatinine (n ¼ 6), thyroid hormones (n ¼ ) or any other relevant adjustment covariate (n ¼ 2987). We further excluded participants who had any of the following conditions at baseline: CKD (n ¼ 1596), proteinuria (n ¼ 2326), abnormal levels of TSH, FT3 or FT4 (n ¼ 5047), use of thyroid medications (n ¼ 256), history of thyroid disease (n ¼ 1400) or history of cancer (n ¼ 614). Thus, the final sample for this study included South Korean men and women, years of age at baseline, who were free of CKD and had normal levels of thyroid hormones at baseline and throughout follow-up. Among participants included in the analysis, 38.6% had two examinations, 41.5% had three or four examinations and the remaining 19.9% had five or more examinations. The study was approved by the institutional review board of Kangbuk Samsung Hospital. Data collection Baseline and follow-up examinations were conducted at the clinics of the Kangbuk Samsung Hospital Health Screening Center in Seoul and Suwon. At each visit, demographic characteristics, smoking status, alcohol consumption, medical history and medication use were collected through standardized, self-administered questionnaires. Smoking status was categorized into never, former or current smoker, and alcohol consumption into none, moderate (30 g/day in men and 20 g/day in women) or high intake (>30 g/day in men and >20 g/day in women). Height, weight and sitting blood pressure were measured by trained nurses. Body mass index (BMI) was calculated as weight in kilograms divided by height in metres squared. Obesity was defined as a BMI 25 kg/m 2, a cutoff proposed for Asian populations. 21 Hypertension was defined as a systolic blood pressure 140 mmhg, a diastolic blood pressure 90 mmhg, a self-reported history of hypertension or current use of antihypertensive medications. Serum total cholesterol, high-density lipoprotein (HDL) cholesterol, glucose, thyroid hormones and creatinine were measured in fasting blood samples collected after at least 12 h of fasting. Diabetes mellitus was defined as a fasting serum glucose 126 mg/dl, a self-reported history of diabetes or current use of antidiabetic medications. Serum TSH and FT4 levels were measured by radioimmunoassay using a commercial kit (RIA-gnost VR htsh, FT4, Schering-Cis Bio International, Gif-sur-Yvette, France), with lower detection limits of liu/ml and 0.06 ng/dl, respectively. The normal range was liu/ml for TSH and ng/dl for FT4. The intra- and inter-assay coefficients of variation for quality control specimens were % and %, respectively, for TSH, and % and %, respectively, for FT4. Serum FT3 was measured by radioimmunoassay (RIA-mat, Byk-Sangtec Diagnostica, Germany), with a lower detection limit of 0.6 pg/ml and a normal range of pg/ml. The intra- and inter-assay coefficients of

4 International Journal of Epidemiology, 2014, Vol. 43, No variation for FT3 were % and %, respectively. Euthyroid status was defined as levels of TSH, FT3 and FT4 within their corresponding normal ranges, no self-reported history of thyroid disease and no current use of thyroid medications. 22 Serum creatinine was measured with the kinetic alkaline picrate (Jaffe) method. The within-batch and total coefficients of variation for creatinine determinations were % for the duration of the study. We calculated egfr using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. 23 The CKD-EPI equation has less bias and greater accuracy than the Modification of Diet in Renal Disease Study equation, especially at higher GFR levels. 23 CKD was defined as egfr <60 ml/min/1.73 m 2. Urine protein was semi-quantitatively determined by urine dipsticks (URiSCAN Urine strip, YD Diagnostics, Yong-In, Korea) performed on fresh, midstream, morning urine samples. Urine protein was reported in six grades: absent, trace, 1þ, 2þ, 3þ and 4þ (corresponding to protein levels of undetectable, 10, 30, 100, 300 and 1000 mg/ dl, respectively). Proteinuria was defined as grade 1þ. Statistical analysis The study endpoint was the development of incident CKD. Participants were followed from the baseline visit to the visit of CKD diagnosis or to the last available visit. Followup was censored when participants failed to maintain a euthyroid state, started taking thyroid medications or developed thyroid disease or cancer. Since incident CKD occurred at an unknown time point between the visit of CKD diagnosis and the previous visit, we used two alternative analytical approaches to account for this type of interval censoring. First, we fitted flexible parametric proportional hazards models that extend standard Weibull models by using natural cubic splines to estimate the baseline log cumulative hazards and allow for interval censored events. 24 These models were used to estimate the adjusted hazard ratios for incident CKD associated with thyroid hormone levels and covariates measured at baseline. Second, we performed pooled logistic regression analyses with one observation per person-visit to relate timevarying covariates measured at each visit to the development of CKD in the next between-visit interval. Pooled logistic models closely approximate time-dependent Cox proportional hazards models when the risk is low in each time interval, 25 but address the issue of interval censoring of events. The association between thyroid hormones and CKD was analysed first by categorizing thyroid hormone levels into quintiles based on cutpoints from their distributions. Tests for linear trend across quintiles were conducted by including an ordinal variable with the median thyroid hormone level of each quintile in regression models. For more detailed dose-response analyses, we modelled thyroid hormones using restricted quadratic splines with knots at the 5th, 50th and 95th percentiles of their sample distributions to provide a smooth yet flexible description of the relationship between thyroid hormones and CKD. To adjust for confounding, we used three models with increasing degrees of adjustment. The first model adjusted for age (continuous), sex and study centre (Seoul or Suwon). The second model further adjusted for smoking status (never, former, current), alcohol consumption (none, moderate, high), body mass index (continuous), total cholesterol (continuous), HDL cholesterol (continuous), hypertension and diabetes. The third model further included baseline egfr (continuous). In addition, we repeated all analyses without censoring participants when they failed to main a euthyroid state, and we performed sensitivity analyses simultaneously including TSH, FT3 and FT4 in the same models, with similar results (data not shown). Finally, we performed stratified analyses in prespecified subgroups defined by age (<50, 50 years), sex (male, female), body mass index (<25, 25 kg/m 2 ), hypertension (yes, no) and diabetes (yes, no), and found similar findings across subgroups (data not shown). All analyses were performed using STATA version 12 (StataCorp LP, College Station, TX, USA). Results The average age of study participants at baseline was 38.0 years [standard deviation (SD) 7.4 years], and 64.8% of them were men (Table 1). The average thyroid hormone levels at baseline were 2.1 llu/ml for TSH, 3.2 pg/ml for FT3 and 1.3 ng/dl for FT4, with Spearman correlation coefficients of 0.08 between TSH and FT3, 0.12 between TSH and FT4 and 0.29 between FT3 and FT4. Participants with higher levels of TSH at baseline were more likely to be women, non-smokers, hypertensive and non-diabetic, and to have higher body mass index and total cholesterol and lower HDL cholesterol and egfr. Participants with higher baseline levels of FT3 or FT4 were more likely to be younger, men, current smokers, drinkers and hypertensive (Table 2). During person-years of follow-up (median follow-up of 3.5 years), 1032 participants developed incident CKD (all with egfr in the range of ml/min/ 1.73 m 2 ).There was a positive association between TSH and the risk of incident CKD. Using baseline TSH levels, the hazard ratio for CKD comparing the highest vs the lowest quintiles of TSH was 1.59 (95% CI 1.29 to 1.95;

5 1628 International Journal of Epidemiology, 2014, Vol. 43, No. 5 P for linear trend <0.001) after adjusting for age, sex, study centre, smoking, alcohol, body mass index, total and HDL cholesterol, hypertension and diabetes (Table 3). The corresponding hazard ratio using time-varying TSH levels Table 1. Baseline characteristics of study participants Characteristic n ¼ Age (years) 38.0 (7.4) Sex Male (64.8) Female (35.2) Study centre Seoul (60.2) Suwon (39.8) Smoking Never (40.2) Former (30.7) Current (29.1) Alcohol None (35.9) Moderate (56.5) High 7933 (7.6) Body mass index (kg/m2) 23.5 (3.0) Total cholesterol (mg/dl) (34.6) HDL cholesterol (mg/dl) 54.8 (12.1) Hypertension (15.2) Diabetes 2768 (2.7) TSH (llu/ml) 2.1 (1.0) FT3 (pg/ml) 3.2 (0.2) FT4 (ng/dl) 1.3 (0.2) egfr (ml/min/1.73 m2) 83.9 (10.5) Values are means (SD) or numbers (%). HDL, high-density lipoprotein; TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxin; egfr; estimated glomerular filtration rate. was 1.61 (95% CI 1.31 to 1.97; P for linear trend<0.001; Table 3). Further adjustment for baseline egfr attenuated the association in the model using baseline TSH, but the association persisted in the model using time-varying TSH. In egfr-adjusted models, the hazard ratios for CKD comparing the highest vs the lowest quintiles of TSH were 1.16 (95% CI 0.94 to 1.43; P for linear trend ¼ 0.18) and 1.26 (95% CI 1.02 to 1.55; P for linear trend ¼ 0.03) using baseline and time-varying TSH levels, respectively (Table 3). In spline regression analyses, the association between time-varying TSH levels and CKD was progressive until plateauing at around 3.0 llu/ml (corresponding to the 83th percentile of the TSH distribution; P for nonlinear spline terms ¼ 0.08; Figure 2). The egfr-adjusted hazard ratios for CKD comparing the highest vs the lowest quintiles of FT3 were 0.95 (95% CI 0.79 to 1.15; P for linear trend ¼ 0.66) and 0.89 (95% CI 0.73 to 1.10; P for linear trend ¼ 0.55) in models using baseline and time-varying FT3 levels, respectively (Table 3). In spline regression analyses, however, FT3 levels below 3 pg/ml (corresponding to the 15th percentile of the FT3 distribution) were associated with an increased risk of CKD (P for non-linear spline terms ¼ 0.03; Figure 2). There was no clear association between FT4 and CKD. The egfr-adjusted hazard ratios for CKD comparing the highest vs the lowest quintiles of FT4 were 0.96 (95% CI 0.79 to 1.16; P for linear trend ¼ 0.30) and 1.03 (95% CI 0.84 to 1.25; P for linear trend ¼ 0.87) in models using baseline and time-varying FT4, respectively (Table 3). Spline regression models were also consistent with a null association between FT4 and CKD risk (Figure 2). Table 2. Age-, sex-, and centre-adjusted baseline characteristics of study participants by quintiles of thyroid hormones (n ¼ ) Characteristic TSH (llu/ml) FT3 (pg/ml) FT4 (ng/dl) Quintile 1 ( ) Quintile 5 ( ) P for trend Quintile 1 ( ) Quintile 5 ( ) P for trend Quintile 1 ( ) Quintile 5 ( ) P for trend Participants (n) Age (years) < <0.001 Male < < <0.001 Seoul < < Current smoker < < <0.001 High alcohol iintakeconsumption < <0.001 Body mass index (kg/m 2 ) < < <0.001 Total cholesterol (mg/dl) < < HDL cholesterol (mg/dl) < < <0.001 Hypertension < < <0.001 Diabetes egfr (ml/min/1.73 m 2 ) < < <0.001 Values are means or percentages. P-value for linear trend using an ordinal variable with the median baseline thyroid hormone level in each quintile. TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxin; HDL, high-density lipoprotein; egfr; estimated glomerular filtration rate.

6 International Journal of Epidemiology, 2014, Vol. 43, No Table 3. Hazard ratios (95% confidence intervals) for incident chronic kidney disease by quintiles of thyroid hormones (n ¼ ) Quintiles Baseline thyroid hormones Time varying thyroid hormones No. of events / person years Model 1 a Model 2 b Model 3 c No. of events / person years Model 1 a Model 2 b Model 3 c TSH (llu/ml) Quintile 1 ( ) 150 / Reference Reference Reference 132 / Reference Reference Reference Quintile 2 ( ) 176 / (0.93, 1.47) 1.15 (0.91, 1.44) 1.02 (0.81, 1.28) 163 / (0.94, 1.47) 1.15 (0.92, 1.43) 1.07 (0.86, 1.34) Quintile 3 ( ) 205 / (1.27, 1.97) 1.52 (1.23, 1.89) 1.32 (1.06, 1.64) 215 / (1.13, 1.73) 1.36 (1.09, 1.68) 1.24 (1.00, 1.54) Quintile 4 ( ) 221 / (1.22, 1.87) 1.47 (1.18, 1.82) 1.20 (0.97, 1.49) 231 / (1.17, 1.78) 1.39 (1.12, 1.72) 1.20 (0.96, 1.48) Quintile 5 ( ) 280 / (1.37, 2.08) 1.59 (1.29, 1.95) 1.16 (0.94, 1.43) 291 / (1.39, 2.08) 1.61 (1.31, 1.97) 1.26 (1.02, 1.55) P for trend * <0.001 < <0.001 < FT3 (pg/ml) Quintile 1 ( ) 258 / Reference Reference Reference 278 / Reference Reference Reference Quintile 2 ( ) 225 / (0.77, 1.11) 0.92 (0.77, 1.10) 0.95 (0.79, 1.13) 209 / (0.77, 1.11) 0.91 (0.76, 1.10) 0.92 (0.76, 1.11) Quintile 3 ( ) 190 / (0.73, 1.08) 0.88 (0.72, 1.07) 0.91 (0.75, 1.11) 162 / (0.78, 1.15) 0.92 (0.75, 1.11) 0.96 (0.79, 1.17) Quintile 4 ( ) 179 / (0.80, 1.15) 0.92 (0.77, 1.11) 0.97 (0.81, 1.17) 193 / (0.83, 1.20) 0.95 (0.79, 1.15) 1.01 (0.84, 1.23) Quintile 5 ( ) 180 / (0.80, 1.16) 0.92 (0.76, 1.11) 0.95 (0.79, 1.15) 190 / (0.74, 1.11) 0.85 (0.69, 1.05) 0.89 (0.73, 1.10) P for trend FT4 (ng/dl) Quintile 1 ( ) 256 / Reference Reference Reference 282 / Reference Reference Reference Quintile 2 ( ) 217 / (0.83, 1.18) 1.00 (0.83, 1.19) 1.01 (0.85, 1.21) 213 / (0.82, 1.19) 1.00 (0.83, 1.20) 1.01 (0.84, 1.22) Quintile 3 ( ) 185 / (0.78, 1.13) 0.94 (0.78, 1.14) 0.93 (0.77, 1.12) 180 / (0.76, 1.12) 0.93 (0.76, 1.13) 0.92 (0.75, 1.12) Quintile 4 ( ) 215 / (0.70, 1.04) 0.86 (0.70, 1.04) 0.82 (0.68, 1.00) 161 / (0.73, 1.08) 0.90 (0.74, 1.10) 0.89 (0.73, 1.09) Quintile 5 ( ) 159 / (0.89, 1.30) 1.09 (0.90, 1.32) 0.96 (0.79, 1.16) 196 / (0.88, 1.32) 1.09 (0.90, 1.34) 1.03 (0.84, 1.25) P for trend P value for linear trend using an ordinal variable with the median thyroid hormone level in each quintile. TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxin. a Adjusted for age, sex, and study centre. b Further adjusted for smoking status, alcohol consumption, body mass index, total cholesterol, high-density lipoprotein cholesterol, hypertension, and diabetes. c Further adjusted for baseline estimated glomerular filtration rate.

7 1630 International Journal of Epidemiology, 2014, Vol. 43, No. 5 Figure 2. Hazard ratios for incident chronic kidney disease by thyroid-stimulating hormone (TSH), free triiodothyronine (FT3) and free thyroxin (FT4) levels. Curves represent adjusted hazard ratios (solid lines) and their 95% confidence intervals (dashed lines) based on restricted quadratic splines for time-varying thyroid hormone levels with knots at the 5th, 50th and 95th percentiles of their sample distributions. The reference values (diamonds) were set at the 10th percentile of the thyroid hormone distributions (corresponding to 0.94 llu/ml, 2.97 pg/ml and 1.10 ng/dl for TSH, FT3 and FT4, respectively). Results were obtained from pooled logistic regression models with one observation per person-visit and adjusted for time-varying values of age, sex, study centre, smoking status, alcohol consumption, body mass index, total cholesterol, high-density lipoprotein cholesterol, hypertension and diabetes, as well as for baseline estimated glomerular filtration rate. Bars represent the frequency distribution of thyroid hormone levels. Discussion In this large cohort of euthyroid men and women from South Korea, high-normal levels of TSH and low-normal levels of FT3 were modestly associated with an increased risk of incident CKD. We excluded participants who had subclinical or overt hypothyroidism at baseline, and also censored participants when they developed hypothyroidism during follow-up. As a consequence, our data indicate that even within the euthyroid range, low thyroid function is associated with incident cases of CKD in apparently healthy subjects, independently of multiple demographic and cardiovascular risk factors. Overt and subclinical hypothyroidism were associated with higher levels of serum creatinine and with lower egfr, whereas hyperthyroidism was associated with lower levels of creatinine and higher egfr These abnormalities were reversible with treatment, and renal function improved after regaining euthyroidism. 5,7 9 Moreover, in a study of 309 patients with subclinical hypothyroidism and CKD, those who did not take thyroid hormone replacement therapy had a greater decline in egfr compared with the treatment group. 11 Only a few studies, all of them cross-sectional, have evaluated the association between thyroid hormone levels and the risk of CKD in the euthyroid range An inverse association between TSH levels and egfr was found in studies in Italy (n ¼ 9888), 18 Norway (n ¼ ), 17 Taiwan (n ¼ 8418) 20 and Korea (n ¼ 2284). 19 Only the study conducted in Korea measured total T3 levels, finding that participants with CKD had marginally higher TSH and lower T3 levels compared with participants without CKD. 19 Because of their cross-sectional design, these studies could not evaluate the direction of the association between thyroid hormones and CKD, an important factor for establishing causality. In contrast, our analysis was based on cohort data with repeated measurements of thyroid hormones and serum creatinine over time in a large sample of healthy individuals. In our study, the association between thyroid hormone levels and CKD was attenuated but persisted after adjusting for egfr, thus strengthening the evidence for a causal effect of lower thyroid function on the risk of future CKD in euthyroid individuals. Several possible mechanisms may explain the link between thyroid hormones and renal function. In overt hypothyroidism, cardiac output is decreased and systemic vascular resistance is increased, resulting in decreased renal blood flow, decreased GFR and creatinine clearance and increased serum creatinine. 5,7,9,17,26 Indeed, GFR reductions in patients with hypothyroidism have been established using gold standard approaches such as plasma clearance of 51 Cr-EDTA or 125 I-iothalamate. 6,9 Conversely, cardiac output and circulating volume are increased in hyperthyroidism, resulting in increased creatinine clearance and decreased serum creatinine. 5,10 Patients with hypothyroidism may also have increased release of creatinine from muscles as a result of myopathy or rhabdomyolysis, whereas patients with hyperthyroidism may experience a decrease in serum creatinine due to the decrease in lean body mass. 5,27,28 In addition, thyroid hormones have important effects on kidney tubular transport, 5,29,30 and patients with hypothyroidism may have diminished tubular creatinine secretion leading to higher serum creatinine levels and lower egfr. 5,29,30 These mechanisms can explain a higher CKD risk in patients with hypothyroidism, but their role in euthyroid individuals with low-normal thyroid function is unclear. Experimental

8 International Journal of Epidemiology, 2014, Vol. 43, No studies have also demonstrated the expression of TSH receptors in extra-thyroidal tissues including the kidney. 31 It is thus possible that TSH may affect renal function independently of FT4 or FT3. Lastly, recent studies in euthyroid individuals found that low-normal thyroid function were associated with multiple risk factors for cardiovascular disease and CKD, including hyperlipidaemia, higher fasting glucose, increased insulin resistance and low-grade inflammation, supporting a biologically plausible role of low-normal thyroid function as a risk factor for CKD Of note, the association between low-normal thyroid function and CKD in our analysis persisted after adjusting for multiple cardiometabolic risk factors, suggesting that other mechanisms may also be involved in explaining the potential effects of TSH and FT3 on CKD risk. Some limitations of the study should be considered. First, we used the CKD-EPI equation to estimate GFR and define CKD. The CKD-EPI equation was developed mainly in Whites and Blacks, and its accuracy for estimating GFR in other ethnic groups, including Asians, was unclear. Indeed, studies of the performance of the CKD-EPI equation in Asians and other ethnicities have been inconclusive and require additional studies. 32,33 Second, we did not have measurements of thyroid antibodies. Autoimmune thyroid disease may lead to the deposition of thyroglobulin-containing immunocomplexes in glomeruli, causing glomerular injury in autoimmune thyroiditis. 34 However, one study found similar associations between TSH and CKD in hypothyroidism patients with and without thyroid peroxidase antibodies, suggesting that autoimmune thyroid disease was unlikely to explain the observed association between TSH and CKD. 17 Third, our study population consisted of apparently healthy young and middleaged Korean men and women, and therefore results from this analysis may not be generalizable to other races/ ethnicities or to an older population with a higher burden of comorbidities. Previous population-based studies in Norway and Italy, however, have found similar associations between TSH and CKD, suggesting that our findings may extend to other race/ethnicity groups. 17,18 In addition to the cohort design, our study had several other strengths. First, the Kangbuk Samsung Health Study is by far the largest population-based study evaluating the association between thyroid hormones and renal function. The majority of previous studies have focused on patients who had already developed overt or subclinical thyroidism, and thus may not be applicable to the general healthy population. The study population of this analysis included individuals with a euthyroid status at baseline and throughout the follow-up, and findings from this study can extend existing knowledge on thyroid hormones and kidney function to general population samples. Second, we also measured FT3 and FT4, allowing for a detailed characterization of the thyroid status of participants beyond TSH measurements. Third, we used high-quality laboratory procedures with extensive quality control. Fourth, the availability of repeated measurements allowed us to assess the time-dependent association between thyroid hormones and CKD. Finally, we also measured an extensive panel of health-related variables, allowing for detailed adjustment of potential confounders. Even though the association between thyroid hormones and CKD was modest in this study, it may still have important clinical implications. Our finding that variations in thyroid hormones, even within the normal range, were associated with incident CKD, independently of multiple demographic and cardiometabolic risk factors, suggest that individuals with low-normal thyroid function may already be different from truly euthyroid individuals. This is in line with the current argument that the upper limit of normal TSH should be reduced to 2.5 liu/ml (as opposed to the current upper limit of 4.5 to 5.0 liu/ml), at least for the young and middle-aged population. 35,36 In conclusion, we found that high-normal levels of TSH and low-normal levels of FT3 were modestly associated with an increased risk of incident CKD in a large cohort study of euthyroid men and women from South Korea. Our results suggest that variations in TSH and FT3 levels, even within the normal range, may affect kidney function. Further research is needed to confirm these findings in other ethnic groups, and to better understand the underlying mechanisms relating thyroid hormones and CKD. Conflict of interest: None declared. References 1. Chonchol M, Lippi G, Salvagno G, Zoppini G, Muggeo M, Targher G. Prevalence of subclinical hypothyroidism in patients with chronic kidney disease. Clin J Am Soc Nephrol 2008;3: Lo JC, Chertow GM, Go AS, Hsu CY. Increased prevalence of subclinical and clinical hypothyroidism in persons with chronic kidney disease. Kidney Int 2005;67: Saini V, Yadav A, Arora MK, Arora S, Singh R, Bhattacharjee J. Correlation of creatinine with TSH levels in overt hypothyroidism a requirement for monitoring of renal function in hypothyroid patients? Clin Biochem 2012;45: Merla R, Martinez JD, Martinez MA et al. Hypothyroidism and renal function in patients with systolic heart failure. Tex Heart Inst J 2010;37: den Hollander JG, Wulkan RW, Mantel MJ, Berghout A. Correlation between severity of thyroid dysfunction and renal function. Clin Endocrinol (Oxf) 2005;62: Karanikas G, Schutz M, Szabo M et al. Isotopic renal function studies in severe hypothyroidism and after thyroid hormone replacement therapy. Am J Nephrol 2004;24:41 45.

9 1632 International Journal of Epidemiology, 2014, Vol. 43, No Kreisman SH, Hennessey JV. Consistent reversible elevations of serum creatinine levels in severe hypothyroidism. Arch Intern Med 1999;159: Montenegro J, Gonzalez O, Saracho R, Aguirre R, Martinez I. Changes in renal function in primary hypothyroidism. Am J Kidney Dis 1996;27: Villabona C, Sahun M, Roca M et al. Blood volumes and renal function in overt and subclinical primary hypothyroidism. Am J Med Sci 1999;318: Woodward A, McCann S, Al-Jubouri M. The relationship between estimated glomerular filtration rate and thyroid function: an observational study. Ann Clin Biochem 2008;45(Pt 5): Shin DH, Lee MJ, Kim SJ et al. Preservation of renal function by thyroid hormone replacement therapy in chronic kidney disease patients with subclinical hypothyroidism. J Clin Endocrinol Metab 2012;97: Roos A, Bakker SJ, Links TP, Gans RO, Wolffenbuttel BH. Thyroid function is associated with components of the metabolic syndrome in euthyroid subjects. J Clin Endocrinol Metab 2007;92: Fernandez-Real JM, Lopez-Bermejo A, Castro A, Casamitjana R, Ricart W. Thyroid function is intrinsically linked to insulin sensitivity and endothelium-dependent vasodilation in healthy euthyroid subjects. J Clin Endocrinol Metab 2006;91: Boggio A, Muzio F, Fiscella M, Sommariva D, Branchi A. Is thyroid-stimulating hormone within the normal reference range a risk factor for atherosclerosis in women? Intern Emerg Med 2014;9: Gumieniak O, Perlstein TS, Hopkins PN et al. Thyroid function and blood pressure homeostasis in euthyroid subjects. J Clin Endocrinol Metab 2004;89: Roef GL, Rietzschel ER, Van Daele CM et al. Triiodothyronine and free thyroxine levels are differentially associated with metabolic profile and adiposity-related cardiovascular risk markers in euthyroid middle-aged subjects. Thyroid 2013;23: Asvold BO, Bjoro T, Vatten LJ. Association of thyroid function with estimated glomerular filtration rate in a population-based study: the HUNT study. Eur J Endocrinol 2011;164: Lippi G, Montagnana M, Targher G, Salvagno GL, Guidi GC. Relationship between thyroid status and renal function in a general population of unselected outpatients. Clin Biochem 2008;41: Song SH, Kwak IS, Lee DW, Kang YH, Seong EY, Park JS. The prevalence of low triiodothyronine according to the stage of chronic kidney disease in subjects with a normal thyroid-stimulating hormone. Nephrol Dial Transplant 2009;24: Sun MT, Hsiao FC, Su SC, Pei D, Hung YJ. Thyrotropin as an independent factor of renal function and chronic kidney disease in normoglycemic euthyroid adults. Endocr Res 2012;37: Wen CP, David Cheng TY, Tsai SP et al. Are Asians at greater mortality risks for being overweight than Caucasians? Redefining obesity for Asians. Public Health Nutr 2009;12: Garber JR, Cobin RH, Gharib H et al. Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Thyroid 2012;22: Levey AS, Stevens LA, Schmid CH et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 200;150: Royston P, Parmar MK. Flexible parametric proportionalhazards and proportional-odds models for censored survival data, with application to prognostic modelling and estimation of treatment effects. Stat Med 2002;21: D Agostino RB, Lee ML, Belanger AJ, Cupples LA, Anderson K, Kannel WB. Relation of pooled logistic regression to time dependent Cox regression analysis: the Framingham Heart Study. Stat Med 1990;9: Klein I, Ojamaa K. Thyroid hormone and the cardiovascular system. N Engl J Med 2001;344: Nakahama H, Sakaguchi K, Horita Y et al. Treatment of severe hypothyroidism reduced serum creatinine levels in two chronic renal failure patients. Nephron 2001;88: Norrelund H, Hove KY, Brems-Dalgaard E et al. Muscle mass and function in thyrotoxic patients before and during medical treatment. Clin Endocrinol (Oxf) 1999;51: den Hollander JG, Wulkan RW, Mantel MJ, Berghout A. Is cystatin C a marker of glomerular filtration rate in thyroid dysfunction? Clin Chem 2003;49: Azuma KK, Balkovetz DF, Magyar CE et al. Renal Naþ/Hþ exchanger isoforms and their regulation by thyroid hormone. Am J Physiol 1996;270(2Pt1):C Sellitti DF, Akamizu T, Doi SQ et al. Renal expression of two thyroid-specific genes: thyrotropin receptor and thyroglobulin. Exp Nephrol 200;8: Horio M, Imai E, Yasuda Y, Watanabe T, Matsuo S. Modification of the CKD epidemiology collaboration (CKD- EPI) equation for Japanese: accuracy and use for population estimates. Am J Kidney Dis 2010;56: Stevens LA, Claybon MA, Schmid CH et al. Evaluation of the Chronic Kidney Disease Epidemiology Collaboration equation for estimating the glomerular filtration rate in multiple ethnicities. Kidney Int 2011;79: Iglesias P, Diez JJ. Thyroid dysfunction and kidney disease. Eur J Endocrinol 2009;160: Wartofsky L, Dickey RA. The evidence for a narrower thyrotropin reference range is compelling. J Clin Endocrinol Metab 200;90: Surks MI, Hollowell JG. Age-specific distribution of serum thyrotropin and antithyroid antibodies in the US population: implications for the prevalence of subclinical hypothyroidism. J Clin Endocrinol Metab 2007;92: