Background
Universal salt iodization was successfully implemented in Sri Lanka in 1995 to prevent iodine deficiency (1). An increase in serum thyroglobulin antibodies (TgAbs) was found in the years after iodine supplementation was implemented (2), but it is unclear whether thyroid autoimmunity persists in populations exposed to sustained iodine supplementation and whether such populations have an increased prevalence of thyroid dysfunction.

This study (3) was designed to evaluate the prevalence of serum thyroid autoantibodies (TgAbs and thyroid peroxidase antibodies [TPOAbs]) and thyroid dysfunction in Sri Lankan children and adolescents after more than two decades of sustained iodine supplementation.

Methods
This cross-sectional study was undertaken between July and November 2017. By random selection, 15 schools from nine districts were included, from which one class per school with 10 girls and 10 boys was randomly selected. All males and females aged 10 to 18 years of age were eligible, although participants with chronic disorders requiring long-term treatment were excluded.

Anthropometric data and blood and urine samples were collected, from which serum thyrotropin (TSH), thyroglobulin, TgAbs, TPOAbs, free triiodothyronine and free thyroxine were measured using immunoassays. Based on this, reference ranges were derived for all thyroid variables according to guidelines from National Association of Clinical Biochemistry (4). Goiter was defined as ultrasound-based thyroid volumes at >97th percentile for age and body-surface area. Urinary and drinking water iodine concentrations were also measured.

Results
The study consisted of 2507 subjects (mean age, 14 years; IQR, 10–14; 53.1% female) with blood samples from all participants and urine samples from 2473 participants. Thyroid ultrasound scans were available from 882 participants. Local population-specific reference ranges for the total cohort were TSH 0.75 to 5.88 mU/L, free thyroxine 0.85 to 1.80 ng/dl, and triiodothyronine 2.21 to 5.22 pg/ml. Because of the significant differences between manufacturer-recommended and reference population–derived autoantibody cutoffs, a population-derived functional cutoff was used for TPOAbs (28 IU/L) and TgAbs (50 IU/L). Thus, prevalences of TPOAbs and TgAbs were 10.3% and 6.4%, respectively.

Of the TPOAb-positive participants, 66.2% had a low TPOAb concentration, defined as one to three times the upper limit of the reference range (ULRR), but the strongest association with subclinical hypothyroidism was found among those with higher TPOAb concentrations (>4 ULRR). The overall prevalence of subclinical hypothyroidism was 3% using the cohort-specific reference ranges and 16.8% using the manufacturer's reference ranges.

The median urinary iodine concentration was 138.5 µg/L (IQR, 79.4–219.0) and the median serum thyroglobulin 8.3 ng/ml (IQR, 4.1-13.5). The median iodine content in drinking water was 10.1 µg/L (IQR, 3.1–24), as compared with 17.5 ppm (IQR, 11.1–22.8) in commercially available salt samples, although there was regional variability in the latter.

Conclusions
After more than two decades of sustained iodine supplementation in Sri Lanka, the prevalence of subclinical hypothyroidism has remained low despite a high prevalence of TPOAbs. Furthermore, population-derived age-specific serum thyroid function reference ranges are crucial in epidemiologic studies.