Elimination of iodine deficiency disorders (IDD)⁴ is a global public health priority (Maberly et al. 1994, Ramalingaswami 1992, WHO 1991). In 1990, seventy-one heads of state and senior policy-makers from 80 other countries attended "The World Summit for Children" and endorsed "The World Declaration and 1990-2000 Plan of Action on the Survival, Protection and Development of Children" (UNICEF 1990). This "Plan of Action" includes the virtual elimination of iodine deficiency. Universal access to iodized salt is the recommended long-term intervention strategy to eliminate IDD, and many countries set this as a 1995 goal (UNICEF-WHO 1994).

A common misperception is that IDD primarily affects only remote rural populations. This belief may have developed because goiter, the most common visible evidence of iodine deficiency, is usually most prevalent in rural populations. While goiter may be the most common visible evidence of IDD, this clinical sign is just the "tip of the iceberg" of the consequences of IDD, which include lower intelligence quotient (IQ), increased fetal, infant and child mortality, poorer growth and birth defects (Boyages et al. 1989, Hetzel 1994). Mild-to-moderate intellectual impairment is an important consequence of iodine deficiency. A meta-analysis of the effects of IDD on mental development found an average IQ deficit of 13.5 points among iodine-deficient populations (Bleichrodt and Born 1994). The perception of IDD as a problem of rural areas may result in decision makers not being aware of IDD as a major impediment to national development. This perception has also led to attempts to target interventions, such as iodized salt and iodized oil, only to areas thought to be affected. The use of iodized oil may be useful for the short term but is difficult to sustain and can detract from promoting the long-term solution, which in the majority of populations is through the iodization of all salt for human and animal consumption. The targetting of iodized salt to specific areas has been unsuccessful in many regions. Information is required to determine the presence and extent of IDD within the country, including urban areas, in order to garner national support and resources for the elimination of IDD through the universal iodization of salt for human and animal consumption.

This study was conducted to evaluate newborn thyroid stimulating hormone (TSH) testing as a method to assess IDD in urban populations. Sampling newborns in urban hospitals may identify the existence of an IDD problem. Cities in Kyrgyzstan, Malaysia, Pakistan, and the Philippines were studied, countries with national goiter prevalence estimates of 20, 20, 32 and 15%, respectively (WHO/UNICEF/ICCIDD 1994). The proportion of salt adequately iodized in late 1994 was unknown in Kyrgyzstan and Malalysia, and estimated to be 11% in Pakistan and 17% in the Philippines (UNICEF 1994).

SUBJECTS AND METHODS

Table 1. Thyroid stimulating hormone (TSH) results from newborn cord blood in selected cities [View Table]

A commercially available neonatal blood spot method (Spectra-Screen TSH, IEM Diagnostics, Santee, CA) was used to measure TSH. This microplate enzyme-linked immunoassay (ELISA) is based on commonly used TSH-specific antibody sandwich principles (Miyai et al. 1981, Tseng et al. 1985). It was utilized because of the high assay sensitivity (about 2 mU/L) and clear discrimination at low TSH levels, which has been shown to be an important factor in the proposed use of blood spot TSH as a public health tool for monitoring the severity of iodine deficiency in populations (Waite et al. 1986). There are a number of advantages in using dried blood spot specimens compared with venous blood, especially in areas where a lack of adequate storage and transportation facilities exists. Previous studies have shown that TSH in a dried spot matrix remains intact and relatively stable, even when exposed to high humidity and heat (37°C) for up to 30 d (Bourdoux et al. 1990, Waite et al. 1987). Other advantages of blood spot specimens include acceptability and ease of sample collection compared with venous specimens.

Blood spots from Malaysia, Pakistan and Kyrgyzstan were analyzed by the Program Against Micronutrient Malnutrition (PAMM) laboratory located at the Centers for Disease Control and Prevention (CDC), Atlanta, GA. Blood spot specimens from the Philippines were analyzed by the Philippines Department of Public Health, which participates in an external blood spot TSH quality control (QC) program. Internal QC results at the PAMM laboratory include four control specimens of different concentrations run in over 68 assays with the following means, standard deviations, and coefficients of variation: level 1, 1.4 ± 0.6, 43%; level 2, 5.4 ± 0.7, 12.9%; level 3, 14.9 ± 1.65, 11.1%; and level 4, 45.5 ± 5.5, 12.1%. Both the PAMM and Philippines laboratories participate in an international external QC program run by the CDC. Each laboratory receives an identical bi-annual shipment of three levels of TSH blood spot material, and the results during the study period were comparable (information available upon request from the authors).

The results of the TSH testing in these cities were compared with results from two iodine-sufficient areas, New South Wales, Australia, and Alberta, Canada (Nordenberg et al. 1992). Blood spots collected in these iodine-sufficient areas were from routine newborn screening programs using the same methods for measuring TSH. These data are based on samples collected on the fourth day of life. A previous study found a high correlation (r = 0.9, P < 0.001) between cord blood and heel stick samples collected at birth using a sensitive monoclonal TSH assay from filter paper specimens (Ma and Lu 1994). The cord blood TSH and heel stick samples collected after the third day of life using sensitive TSH assays were found to be similar (Ma and Lu 1994, Wu 1991). Earlier polyclonal TSH assays tended to cross-react with human chorionic gonadotropin, follicle stimulating hormone, and luteinizing hormone, and therefore the pattern of TSH during the first few days of life differed by generally starting at a high level, surging for the first 3 d of life, and then stabilizing at a lower level (Fisher and Klein 1981). The monoclonal TSH assay used in this study does not show these cross-reactions. In iodine-sufficient areas, using the laboratory methods described above, the proportion of infants with a TSH >5 mU/L whole blood is generally less than 3% (Nordenberg et al. 1992, WHO/UNICEF/ICCIDD 1994). A recent publication from the World Health Organization (WHO/UNICEF/ICCIDD 1994) provides the following epidemiologic criteria by which to classify populations in terms of the severity of IDD based on the proportion of newborns with a TSH >5mU/L whole blood: mild, 3-19.9%; moderate 20-39.9%; severe, >40%. Confidence intervals for the proportion of newborns with a TSH >5mU/L whole blood were calculated using the exact mid-p method using the Epi Pak software program (Epi Pak, Version 1, Atlanta, GA).

RESULTS

DISCUSSION

If iodine intake decreases sufficiently, thyroid hormone synthesis becomes inadequate to promote normal central nervous system development, and TSH levels become elevated in an attempt to stimulate thyroid hormone synthesis (Burrow and Dussault 1980). Therefore, TSH levels directly reflect the adequacy of thyroid hormone in the brain (DeLange 1989). The relatively recent development of whole-blood spot TSH assays sensitive in the low physiologic range means that mild elevations of TSH are detectable and renders TSH a useful screening test for iodine deficiency in populations (Tseng et al. 1985).

In the results presented in Table 1, cord specimens were collected from births occurring in hospitals in the cities studied. Some of the births may have been to women from rural areas, although in the hospitals studied the majority of births were reported to be from women who lived within the city. Because hospitals were not selected at random, there is a potential bias in the results if there is a difference in the socioeconomic status of those affected by IDD, and if individuals from different socioeconomic groups use different hospitals. We would recommend that hospitals serving the lower socioeconomic status be selected to assure that an IDD problem is not missed.

The purpose of screening these newborns was not to provide a representative sample of newborns but to determine whether an IDD problem exists. If a sample of newborns has elevated TSH values, then the interpretation would be that there is evidence of IDD from the area where the mothers of the newborns live.

Urine samples obtained from mothers of the infants tested for TSH from Kuching were consistent with the TSH results in identifying an IDD problem. The median urinary iodine level for the women was 0.26 µmol/L (3.3 µg/dL; n = 194), which according to WHO criteria would indicate a moderate IDD problem (WHO/UNICEF/ICCIDD 1994). Urine, thyroid ultrasound, and palpation results collected on school children in four rural areas at the same time that cord bloods were collected in Kyrgyzstan were consistent in identifying a moderate-to-severe IDD problem (overall prevalence of goiter = 49%; prevalence of thyroids >97th percentile = 39%; median urinary iodine = 0.24 µmol/L or 3.0 µg/dL). In Manila, the median urinary iodine among young school children in early 1993 was 0.32 µmol/L (4.0 µg/dL), which indicates a moderate IDD problem even though the prevalence of goiter was only 2%.

Other factors can effect TSH levels in newborns. In Pakistan, information was collected on gestational age and birth weight. The prevalence of elevated TSH (>5 mU/L) was 84% among low birth weight infants (<2500 g) compared with 72% among infants of normal birth weight (>2500 g). The prevalence of elevated TSH among preterm infants (<37 wk) was 72%, similar to the prevalence of 73% for full-term infants (37 wk). In Manila, information was collected on maternal age, parity, marital status, income, presence of goiter in the mother, sex of child, gestational age, birth weight, birth length, APGAR scores (1 and 5 min), and presence of congenital malformations. Groups with elevated TSH levels were firstborn infants (prevalence of 39% compared with 27% among non-firstborns), males (36% compared with females, 28%), those with a 1-min APGAR <8 (46% compared to those with APGAR score > 8 of 31%), and those with a 5-min APGAR < 9 (49% compared with those with an APGAR >9 of 31%). While other factors may account for some variability of a high prevalence of elevated TSH values in certain areas, the primary determinant is most likely the availability of iodine in the diet of the mothers.

The results of this study estimate that iodine deficiency in these urban populations may be more extensive than previously recognized. TSH has been used in newborns for identifying the existence and magnitude of IDD in other urban areas, including China (Rushworth et al. 1989) and Poland (Nordenberg et al. 1994a and 1994b) and has been consistent with other indicators of IDD in these areas. These results along with other information suggest the importance of directing IDD intervention efforts to both urban and rural populations. Collection of cord blood samples on filter paper from a small number of newborns for TSH analysis may provide a method to rapidly determine the existence of an IDD problem.