RESUME La faisabilité opérationnelle d’un programme de dépistage de l’hyperthyroïdie congénitale a été évaluée. Des échantillons de sang du cordon ombilical ont été recueillis dans sept hôpitaux de Téhéran et dans le réseau de santé du District de Damavand. Les nouveau-nés ayant un taux d’hormone de libération de la thyréostimuline (TSH) dans le cordon supérieur ou égal à 20 mU/L ont été rappelés et un traitement par lévothyroxine (L-T4) a été mis en route immédiatement après le diagnostic d’hyperthyroïdie congénitale. Sur les 20 107 échantillons acceptables, 22 nouveau-nés avaient une hyperthyroïdie congénitale (1 pour 914 naissances). Le taux de rappel était de 1,3 %. La couverture du dépistage s’élevait à 90 % des naissances vivantes. Sur tous les échantillons de cordon, 0,2 % seulement étaient inacceptables du fait soit du retard dans le transport soit du mauvais recueil des échantillons. Les âges médians au moment du diagnostic et de la mise en route du traitement étaient de 12 et 8 jours respectivement. Le dépistage de l’hyperthyroïdie congénitale est possible, et un programme national de dépistage est nécessaire.

Congenital hypothyroidism (CH) occurs in 1:3000–4000 births [1] and represents one of the most common preventable causes of mental retardation [2]. Delayed diagnosis and treatment may result in poor neurointellectual outcome [3]. Screening for CH was first introduced in Canada in 1974 [4,5]. It has since become routine in all industrialized and some developing countries [6]. Primary tests measuring thyroid-stimulating hormone (TSH) or serum thyroxine (T4) on dried blood spots have been extensively employed for detection of neonatal hypothyroidism. Primary TSH screening is presently performed in all European countries (except the Netherlands) and in Japan, Australia and parts of the United States of America and Canada. In the Netherlands, primary T4 tests are used more frequently [7].

In 1987, the Islamic Republic of Iran was an iodine deficient area and Azizi et al. introduced the first CH screening in Teheran using cord blood TSH measurement. Iodine deficiency induces hyperthyrotropinaemia and subsequently has a high recall rate of neonates. Because the recall rate of the first screening programme was unusually high (i.e. 5%), the programme was temporarily discontinued [8]. In 1990, Karamizadeh and Amirhakimi conducted a study of 4300 cord blood samples in Fars Province using primary T4 supplemented by TSH back-up tests on specimens lower than the cut-off level. The incidence rate of CH was one in 1433 births [9].

In this country, most deliveries occur in hospitals or birth centres, which belong to primary health care (PHC) or non-PHC systems.

We determined CH incidence rate and assessed the operational feasibility of the programme in the Damavand District health network (PHC) and in Teheran hospitals (non-PHC), both of which offer models for a national screening programme. Proper sampling, timely transportation of samples, maximum neonate coverage, follow-up and treatment of CH neonates were also analysed.

For 40 months between 20 February 1998 and 21 June 2001, cord blood samples were collected in seven hospitals in Teheran. For an additional 12 months therein, a general hospital and a rural birth centre in the health network of Damavand District were also used for sample collection. The Damavand hospital and birth centre and the Teheran hospitals offered models for PHC and non-PHC systems respectively. Damavand hospital and birth centre were 76 and 96 km respectively from the Endocrine Research Centre (ERC). The distances of the Teheran hospitals from the ERC were between 0 km and 25 km.

Figure 1 Flow-chart of the Endocrine Research Centre (ERC) screening programme for congenital hypothyroidism (CH) with diagnoses shown in bold

A total of 20 143 cord blood samples, 18 978 (94.2%) from Teheran and 1165 (5.8%) from Damavand, were collected and sent to the ERC central laboratory. Of these, 36 (0.18%) samples were unacceptable, 28 samples from Teheran and 8 samples from Damavand. Of these 36 samples, 17 were unacceptable because of improper sampling and 19 were unacceptable because of delayed transportation. Of the improperly collected specimens, 15 were from Teheran hospitals and 2 were from the Damavand health network. Of the delayed transportation specimens, 13 were from the Teheran hospitals and 6 were from the Damavand health network. The mean interval and standard deviation between specimen collection in the Damavand health network (PHC system) and receipt and testing by the ERC central laboratory was 2.7 ± 1.5 days with a median of 2 days. These data were not available for Teheran hospitals.

During the course of the programme, 20 107 neonates, i.e. acceptable samples, were screened. Of the total acceptable samples, 18 950 (94.2%) and 1157 (5.8%) were from Teheran and Damavand respectively. Coverage increased progressively during the first year until it reached 90% of total births and then it remained constant throughout the programme. Of 1169 live births in Damavand, 1157 acceptable cord samples were sent to the ERC, which meant there was almost complete coverage (98.9%) in the PHC system. In the Teheran hospitals, coverage was 88.1% (16 183 of 18 362 live births).

Of 56 telephone recalls, 52 parents responded and brought the infant to the ERC for confirmatory tests. The mean interval between birth and confirmation of diagnosis for recalled neonates was 13 ± 6 days with a median age of 12 days and a range of 5–25 days. The mean interval between birth and commencement of treatment in CH neonates was 10 ± 5 days with a median age of 8 days and a range of 4–46 days. L-T4 therapy was started in four CH neonates at 15, 19, 23, and 46 days of life.

The first screening for CH in the Islamic Republic of Iran by Azizi et al. during 1987–89 was discontinued after two years because of the high recall rate due to iodine deficiency and hyperthyrotropinaemia [8]. The second study in the country was conducted by Karamizadeh and Amirhakimi on 4300 cord blood samples and found an incidence rate of 1:1433 births for CH [9]. Our study was conducted in seven hospitals in Teheran for 40 months and in a general hospital and a rural birth centre in the Damavand District health network for 12 months. The former was a representative model for non-PHC systems and the latter for PHC systems. Of all live births, 88.8% were screened and the incidence rate of CH was 1:914 live births. This was higher than reported in North American and European countries and also in Japan [15]. To determine the incidence rate of permanent CH, a prolonged study is needed to differentiate transient from permanent types of CH.

Preliminary results of our study indicated a recall rate of 1.6%. This was significantly lower than during the iodine deficiency period when the recall rate was 5% (P < 0.001) [16]. This significant decrease in the recall rate was one of the indicators of iodine supplementation in the region and supported continuation of the programme [1]. As the programme progressed, the recall rate settled at the present rate of 1.3%.

The interval between birth and confirmation of diagnosis in recalled neonates was used as an indicator of the quality of job performance by the specimen collectors. These collectors included in addition to workers in the PHC and the non-PHC systems, the Damavand health workers and the ERC central laboratory, tracking system and screening office personnel. The mean birth–confirmation of diagnosis interval was 13 ± 6 days with a median of 12 days and a range of 5–25 days and therefore the performance of the screening programme in both PHC and non-PHC systems was acceptable [17].

Heel prick method, which is usually performed during the first 2–6 days of life, is the preferred method for specimen collection [11,12]. However, cord blood sampling is still used in some regions [18,19]. In the United States of America and Canada, T4 is initially measured and TSH is only measured for those with the lowest 10%–20% of T4 results. In other areas of the world, direct TSH screening has commonly been used. In effect, all neonatal thyroid-screening programmes include screening for elevated serum TSH concentration as the most reliable indicator of primary hypothyroidism [6]. Selection of primary TSH or T4 methods depends greatly on the local laboratory or regional screening programme preferences. The incidence of CH has been shown to be similar with both screening approaches [7]. The threshold value of significant TSH elevation in most screening programmes is 20–25 mU/L. Recent increases in the number of maternal–neonatal discharges from hospitals within 48 hours of delivery have been a burden on screening programmes by increasing recall rates and false positive cases [6]; therefore, some programmes use higher values for the TSH cut-off point [6,20]. Whatever method is used, the goal of the screening programme is to obtain samples from neonates and to provide early treatment of CH while keeping the number of false positive results as low as possible, particularly because of the emotional disturbance to parents caused by a recall [21].

In the Islamic Republic of Iran, the heel prick method could possibly result in large numbers of neonates not getting screened. In addition, mothers and neonates are usually discharged within 24 hours of delivery. Therefore, in the present study, in order to obtain specimens from all neonates, the preferred method was cord blood sampling. This might have resulted in a high recall rate because of the physiologic TSH surge within the first 24 hours of life. Some studies using cord blood samples have used a higher level for the cut-off point [20,22]. Since we planned to determine incidence and to design a model to serve as the nationwide programme, we used one of the lowest cut-off points reported, i.e. 20 mU/L. Initial results of the present study supported the use of higher cut-off values, e.g. 30–35 mU/L as CH was not found in neonates with cord TSH less than 40 mU/L. However, this concept was later abandoned when three hypothyroid neonates with primary TSH values of 23.3 mU/L, 33.2 mU/L and 35 mU/L were detected.

One of the main objectives of a CH screening programme is to establish early treatment for affected neonates. Although starting treatment within 14 days of life is preferred [7,23], this is not always possible. Due to screening and early diagnosis, treatment of affected infants has usually been initiated within 45 days of birth and IQ levels in treated infants measured at 5 to 7 years of age have been normal [6]. Some researchers have reported that the turnaround time from collection of specimen to treatment must be less than 3 months or, ideally, less than 1 month [18]. In our study, the mean interval between birth and starting treatment for hypothyroid neonates was 10 ± 5 days with a median of 8 days and a range of 4–46 days. The median age of start of treatment compared favourably with reports from Europe [24,25]. That the telephone recalls worked so well indicated that parents were reasonably cooperative and reduced the need for home visits considerably (and thus reduced the cost of the programme). This is important if the programme is to be conducted on a national scale. In this regard, public awareness plays a critical role.

A few difficulties and obstacles were observed. The large number of recalled neonates with ‘undetermined diagnosis’ and the few CH cases with delayed treatment will necessitate public awareness as the programme continues. Also, despite regular and careful monitoring of the programme (i.e. personnel, tests and equipment), human error was inevitable. For example, delayed reporting by central laboratory personnel of abnormal venous TSH and T4 values of one CH neonate resulted in delayed treatment for that child until day 46 of life.

Unless the programme can be legislated and supported by the government, initiation and operation of the neonate-screening programme for CH will not result in complete participation of the population at risk. Because of the high incidence rate of CH, a national screening programme for CH is needed in the Islamic Republic of Iran. The elimination of iodine deficiency, the low recall rate and the operational feasibility of the CH screening programme for both the PHC and the non-PHC systems have laid the groundwork for a successful programme. The next step will entail obtaining parliamentary legislation, government support and public awareness via mass media in order to establish a nationwide neonate screening programme for CH.

This research project was supported by the research grant No. 115 NRCI Research Projects from the National Research Council of the Islamic Republic of Iran.

25. Grant DB, Smith I. Survey of neonatal screening for primary hypothyroidism in England, Wales, and Northern Ireland 1982–4. British medical journal, 1988, 296(6633):1355–8.