ORIGINAL ARTICLE

Clinical Practice: Experience with newborn screeningfor congenital hypothyroidism in the Republicof Macedonia—a multiethnic country

Mirjana Kocova & Violeta Anastasovska &

Elena Sukarova-Angelovska & Milica Tanaskoska &

Elizabeta Taseva

Received: 27 May 2014 /Revised: 25 August 2014 /Accepted: 26 August 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract To evaluate the thyroid screening program and toestimate the prevalence of congenital hypothyroidism (CH)among newborns in the Republic of Macedonia, we measuredthyroid-stimulating hormone (TSH) levels in dried blood spotspecimens using the DELFIA fluoroimmunoassay, over aperiod of 12 years. The TSH cutoff level was 10 mU/L blood.A total of 215,077 newborns were screened (94.76 %). Out of254 recalled newborns (a recall rate of 0.15 %), 83 newbornswith CHwere detected, yielding a CH prevalence at screeningof 1/2,591 (female to male ratio, 1.86:1). Of the CH cases, 47/107,754 (56.6 %) neonates were Macedonian, 29/70,330(34.9 %) were Albanian, and 7/15,055 (8.4 %) were Roma.The thyroid gland was undetectable on ultrasound in 43(51.8 %) newborns with CH, thyroid hypoplasia was con-firmed in 8 (9.6 %), while 29 (34.9 %) had a normal thyroidgland. In three newborns (3.6 %), agenesis of one lobe was

confirmed. Therapy with levothyroxin was initiated on aver-age 11.7 days after birth.

Conclusion: The national thyroid newborn screening pro-gram in Macedonia has been successful and effective, provid-ing timely diagnosis and treatment of children with congenitalhypothyroidism.

Keywords Congenital hypothyroidism . Newborn . Thyroidscreening program . Thyroid-stimulating hormone

AbbreviationsCH Congenital hypothyroidismTSH Thyroid-stimulating hormone

Introduction

Congenital hypothyroidism (CH) is the most common con-genital endocrine disorder, affecting 1 in 3,000 to 4,000 new-borns. Although newborn screening for CH has been a routinepractice in developed countries for more than 40 years, devel-oping countries have addressed this concern only in the pastdecade [1]. Thyroid neonatal screening has dramaticallychanged the natural history of CH. It allows for the earlydetection and treatment of CH, thus preventing growth failure,irreversible mental retardation, and a variety of neuropsycho-logical deficits, including the cretinism [2, 3]. The first screen-ing programs were performed in North America [4]. TheRepublic ofMacedonia is a multiethnic country;Macedoniansof Slavic origin are the majority, contributing with 53.2 %;30.3 % are ethnic Albanians; 7 % are Roma; 4.7 % are Turks;and 4.9 % are of other ethnicities. In the Republic of Mace-donia, the neonatal screening program is based on the mea-surement of thyroid-stimulating hormone (TSH) at birth. Itwas introduced as a pilot program in the birth centers of the

Communicated by Beat Steinmann

M. Kocova (*) : E. Sukarova-AngelovskaDepartment of Endocrinology and Genetics, University Clinic forChildren’s Diseases, Vodnjanska 17, Skopje, Macedoniae-mail: mirjanakocova@yahoo.com

E. Sukarova-Angelovskae-mail: esukarova@doctor.com

V. AnastasovskaLaboratory for Neonatal Thyroid Screening, University Clinic forChildren’s Diseases, Skopje, Macedoniae-mail: violeta_anastasovska@yahoo.com

M. Tanaskoska : E. TasevaGenetic Laboratory, University Clinic for Children’s Diseases,Skopje, Macedonia

M. Tanaskoskae-mail: milcka_m@yahoo.com

E. Tasevae-mail: tasevabeti@yahoo.com

Eur J PediatrDOI 10.1007/s00431-014-2413-4

capital city and two larger towns for a period of 5 years.Thereafter, the TSH screening program became mandatorynationwide.

Methods

The screening for CH in the Republic of Macedonia is basedon the measurement of TSH in dried blood spots, 48 h afterbirth. After the regional pilot program from April 2002 toDecember 2006, screening became mandatory nationwide in2007. There is one screening center in the Republic of Mac-edonia, located at the University Children’s Clinic in Skopje,which covers the newborns from the whole country (the totalpopulation being about two million). Approximately 24,000neonates born in 31 public and private birth centers all overthe country are screened every year by measuring blood spotTSH concentrations. Heel prick blood samples were taken bytrained nurses, mostly after 48 h of life (when the baby isdischarged from the birth center) and spotted on Whatman903 filter papers (Guthrie card). They were dried and mailedto the screening laboratory daily or every second day, depend-ing on the number of newborn babies in the birth center.Premature or sick full-term neonates, who usually have aprolonged stay in neonatal intensive care units, were screenedbetween the first and second week of life. Neonatesdischarged from the centers before 48 h of life (upon parents’request) were screened at the moment of discharge. The studywas carried out between April 2002 and December 2013;during this period, a total of 215,077 newborns were screened.

Screening protocol of the study

When the TSH value on the Guthrie card was lower than10 mU/L of whole blood, it was considered negative and nofurther action was pursued. Results between 10 and 20 mU/Lwere considered borderline, and a blood specimen on a newGuthrie card was requested by phone. TSH values greater than20 mU/L on the initial Guthrie card were considered positivefor CH and the newborn was referred for biochemical andclinical evaluation immediately.

When venous TSH levels were >20 mU/L, irrespective ofthe T4 levels (as measured by IMMULITE chemiluminescentenzyme immunoassay—SIEMENS), replacement therapywith L-thyroxine of 10–15 μg kg day [5] was initiated. WhenTSH serum levels were between 5 and 9 mU/L (gray zone)and T4 levels were normal, the baby was followed up every15 days. Infants with repeatedly increased serum TSH levels(>5 mU/L) up to 2–3 months of life also received substitutiontherapy with L-thyroxine. Ultrasonography of the thyroidgland was performed during the first month of life.

Method of TSH measurement

A time-resolved fluoroimmunoassay was used to measure theTSH concentration in the dry blood spot specimen using aDELFIA neonatal hTSH (human TSH) kit manufactured byPerkin Elmer, Turku, Finland. The DELFIA neonatal hTSHassay is a solid phase, two-site fluoroimmunometric assaybased on the direct sandwich technique in which two mono-clonal antibodies (derived frommice) are directed against twoseparate antigenic determinants on the hTSH molecule. Cali-brators, controls, and the specimen containing hTSH simulta-neously react with immobilized monoclonal antibodies direct-ed against a specific antigenic site on the βhTSH subunit andeuropium-labeled monoclonal antibodies (directed against adifferent antigenic site located partly on the β subunit andpartly on the α subunit) in an assay buffer. The assay bufferelutes hTSH from the dried blood spots. The complete assayrequires only one incubation step. After washing, an enhance-ment solution dissociates the europium ions from the labeledantibody into the solution where they form highly fluorescentchelates with components of the enhancement solution. Thefluorescence in each well is then measured. The fluorescenceof each sample is proportional to the concentration of hTSH inthe sample. The results are expressed as milliunits per liter ofwhole blood (milliunits per liter of whole blood×2.2=milliunits per liter of serum, assuming a hematocrit of 55 %for the first days of life). The TSH calibrators and controlshave been prepared from human blood with a hematocritvalue of 50–55 % and calibrated against the thyroid-stimulating hormone, human, for immunoassay, Third Inter-national Standard, National Institute for Biological Standardsand Control Code 81/565. Internal and external quality assur-ance controls were included in the TSH assay. Internal con-trols were made simultaneously with samples to assure theday-to-day validity of the results. External controls from“Referenzinstitut fűr Bioanalytik,” Bonn, Germany were in-cluded every 3 months. The analytical sensitivity of theDELFIA Neonatal TSH assays is typically better than 2 mU/L blood. The accuracy of the TSH test was assessed usingboth intra- and inter-assay. Intra-assay coefficients of variationat TSH concentrations of 15, 23.7, and 66.1mU/Lwere 7, 7.8,and 7.5 %, respectively, while the inter-assay coefficients ofvariation at the same TSH concentrations were 8, 9, and7.9 %, respectively.

Statistical analysis

The data were analyzed using SPSS 20 software (Chicago, IL)and the statistical significance was set at p<0.05. Pearson χ2

was used for the comparison of proportions.Mann-WhitneyUtest was used for the comparison of TSHmean values between

Eur J Pediatr

different groups (according to the day of sampling, birthweight, and gestational age at birth). F-ANOVA test was usedfor establishing the interaction between categorical variablesand the Spearman’s coefficient for correlation.

Results

Out of 226,962 newborns, a total of 215,077 have beenscreened, of which 111,065 (51.64 %) males and 104,012(48.36 %) females, with a male to female ratio of 1.068.

Among the screened neonates, 107,754 (50.1 %) were Mac-edonian, 70,330 (32.7 %) Albanian, 15,055 (7 %) Roma,9,679 (4.5 %) Turk, and 12,259 (5.7 %) of other nationalities.The coverage of the screened newborns was 94.76 % average(range, 88.34–97.63 %; Table 1). The distribution of the TSHlevels was not significantly different between the male andfemale newborns, although male newborns tended to havehigher mean TSH levels (Fig. 1). In 83.1 % of the neonates,samples were taken after 48 h of birth and the mean TSH levelwas 1.47±1.30 mU/L. The mean TSH level in the newbornsscreened before 48 h of birth (14.8 %) was 2.13±1.75 mU/L.

Fig. 1 TSH levels in the screenedneonates by gender

Table 1 Neonatal thyroid screening coverage and frequency of congenital hypothyroidism by period

Period Coverage (%) Number of CH cases Prevalence (per live birth) Time of initiation of therapy(average day)

April 2002a 91.93 4/7,225 0.55/1,000 (1:1,806) 10

2003a 88.34 4/9,883 0.4/1,000 (1:2,471) 10

2004a 97.63 2/10,239 0.2/1,000 (1:5,120) 7

2005a 92.90 3/9,821 0.31/1,000 (1:3,274) 11.8

2006a 89.69 5/19,744 0.25/1,000 (1:3,950) 10.2

2007a 96.89 11/21,856 0.5/1,000 (1:1,987) 13.5

2008 95.85 6/21,982 0.27/1,000 (1:3,664) 9.7

2009 95.63 9/22,845 0.39/1,000 (1:2,538) 10.6

2010 96.24 11/23,569 0.47/1,000 (1:2,143) 15.1

2011 96.15 10/21,972 0.46/1,000 (1:2,197) 16.3

2012 97.21 9/23,120 0.39/1,000 (1:2,569) 12.5

2013 97.23 9/22,821 0.39/1,000 (1:2,536) 13.1

Total 94.76 83/215,077 0.386/1,000 (1:2,591) 11.7

a Pilot study

Eur J Pediatr

There was a statistically significant difference between theTSH levels in relation to the timing of blood sampling(p<0.05; Fig. 2). The results from the specimens sampledbefore 48 h of birth were not included in the study. Theunsatisfactory sampling rate during the study period was onaverage 2.1 %. The analysis of all screening samples betweenthe second and fifth day of life showed that 95.6 % of sampleshad values between 0–5 mU/L (0–1 mU/L—27.11 %, 1–2 mU/L—37.60 %, 2–3 mU/L—18.84 %, 3–4 mU/L—8.22 %, and 4–5 mU/L—3.9 %). In 3.13 % newborns, theTSH levels were above 5 mU/L. There was a statisticallysignificant negative correlation between the TSH level andthe gestational age (p<0.05). Linear multivariate regressionanalysis performed in order to test the relation between theTSH level as a dependent variable and the newborn birthweight (p=0.008), newborn gender (p=0.055), timing ofnewborn screening sampling (p=0.001), and gestational age(p=0.002) showed that only the newborn gender was not astatistically significant TSH level predictor variable. Spear-man rank correlation analysis between the TSH level, new-born birth weight, and gestational age showed a significant

positive correlation between the birth weight and the gesta-tional age (r=0.371, p<0.01), and a significant negative cor-relation between the gestational age and the TSH level (r=0.29, p<0.05) and the birth weight and the TSH level (r=0.31,p<0.05).

Overall, 254 neonates (0.15 %) were recalled. Of these,171 (67.3 %) were categorized as healthy neonates and 83(32.7 %) were diagnosed as CH cases. The prevalence of thedisease was 1/2,591 average (Table 1). There was only onebirth center with a high prevalence of congenital hypothyroid-ism (1:1,220) that was stable over the analyzed period ofscreening. Among the newborns with CH, 54 (65.1 %) werefemales and 29 (34.9 %) were males (female to male ratio,1.86:1). The prevalence of CH among male and female neo-nates was 0.26/1,000 (1:3,830) and 0.52/1,000 (1:1,926),respectively; the difference was statistically significant(p<0.05). Of the CH cases, 47 (56.6 %) neonates were Mac-edonian, with a prevalence of 1/2,293; 29 (34.9 %) wereAlbanian, with a prevalence of 1/2,425; and 7 (8.4 %) wereRoma, with a prevalence of 1/2151 (Table 2.). The distributionof TSH levels showed that 48 (57.8 %) of the diagnosed CH

Fig. 2 Variation of the TSH levelbased on the timing of newbornscreening sampling (p<0.05)

Table 2 Total number ofscreened newborns, number ofpositive CH, and prevalence foreach ethnic group of newbornsduring the period 2002–2013

Ethnicity of newborns Male Female Total Number of CH cases Prevalence

Macedonian 55,645 52,110 107,755 47 1/2,293

Albanian 36,320 34,010 70,330 29 1/2,425

Roma 7,778 7,277 15,055 7 1/2,151

Turk 4,999 4,679 9,678 / /

Other 6,323 5,936 12,259 / /

Total 111,065 104,012 215,077 83 1/2,591

Eur J Pediatr

neonates had initial TSH values above 50 mU/L, 13 (15.7 %)had TSH values within 20–50 mU/L and 22 (26.5 %) hadborderline initial TSH values (10–20 mU/L). The thyroidgland was undetectable on ultrasound in 43 (51.8 %) new-borns with CH, thyroid hypoplasia was confirmed in 8(9.6 %), while 29 (34.9 %) had a normal thyroid gland. Inthree newborns (3.6 %), agenesis of one lobe was confirmed.The median age of initiating treatment with levothyroxin was11.7 days after birth (Table 1). The efficacy indicators for ournational neonatal thyroid screening program as compared tostandard indicators are shown on Table 3.

Discussion

CH is detected shortly after birth through newborn screening.Newborn screening is considered a critical public health pro-gram since the 1960s and is designed to detect inborn disor-ders that, if left untreated, can lead to early mortality orlifelong morbidity and disability [1, 3]. In the Republic ofMacedonia, it is mandatory to offer CH screening to allnewborns; however, parents have the right to opt out of thescreening, similarly as in other countries. The timing of new-born screening sampling is important since most screenedanalytes are within the normal range and vary considerablywithin the first hours or days of life due to the physiologicalneonatal TSH surge. This elevates the TSH levels and mayresult in a high rate of false positive results of the CH screen-ing [3, 6]. However, the false positive result rate is usuallyhigher in newborn screening programs where T4 is used as aprimary analyte (0.30%) compared to programswhere TSH isused (0.05 %) [7, 8]. This study shows that there was astatistically significant correlation between the TSH leveland the timing of the newborn screening sampling (p<0.05).Also, there was a statistically significant negative correlationbetween the TSH level and the gestational age (p<0.05).Preterm newborns account for the largest percentage of new-borns with false positive CH screening results due to their lowT4 levels resulting from the immature hypothalamic-pituitary-

thyroid axis, the maturity of which depends on the gestationalage [9].

The prevalence of CH among neonates in the Republic ofMacedonia has not been previously determined for a longerperiod of time. This study shows that the estimated prevalenceof CH is 1 in 2,591 (Table 2) live births, which is similar toother European countries [4, 8]. The CH prevalence is influ-enced by gender; it is almost twice as common in femalescompared to males (female to male ratio, 1.86:1) (p<0.05).The cause for the increased risk in females is not well under-stood but is thought to be related to autoimmunity, which isgenerally more common in females [7, 10–14]. CH is higherin Hispanics, Asians, and Native Americans, and lower inAfrican-Americans [10, 15]. The results of our study alsoshow a varying prevalence between the different ethnicitiesin Macedonia, which suggests an involvement of geneticfactors not yet elucidated [12]. In our study, we found a higherprevalence in one of the screened birth centers; this may bedue to the low total number of screened newborns in thatcenter or the large Roma population in that region (17 % inthis birth center versus 7 % in the entire population). Furtheranalysis of the prevalence of CH in this ethnicity is warrantedsince no such data exist thus far.

Once the diagnosis of CH is confirmed, treatment shouldbe started immediately, and then additional optional diagnos-tic studies may be performed to determine the underlyingetiology of CH. In our study, the mean age at treatmentinitiation was 11.7 days, which is in line with other reports[16, 17].

Conclusion

The relatively low birth rate (about 24,000 live births peryear), the short geographical distances, the good communica-tion and networking between birth centers, and the centrali-zation of screening are favorable circumstances for the devel-opment and maintenance of a screening program in the Re-public of Macedonia. Therefore, the national newborn

Table 3 Indicators for neonatal thyroid screening efficacy

Indicator National programachieved (%)

Indicatorstandard (%)a

Age of sampling (% sampling at age <10 days) 98.5 100

Timing of specimen delivery (% specimens received at laboratory at/before five working days after sampling) 84 95

Timing of results (% results available at/before two working days after receipt at laboratory) 88.7 90

Age of newborn at time of recall (% results of first specimens available at/before 20 days postpartum) 92.2 95

Age of initiation of treatment (% positive cases started with treatment at/before 21 days postpartum) 100 95

Coverage 94.75 99.95

a Indicators for program auditing [18]

Eur J Pediatr

screening program for congenital hypothyroidism has beensuccessful and quite effective. Further improvement of cover-age is warranted.

Acknowledgments We would like to acknowledge significant contri-bution of the International Society for Neonatal Screening (Dr. ToniTorresani, Zűrich) and the International Health Fund (Dr. Thomas Foley,Pittsburgh) to the introduction of the neonatal thyroid screening in theRepublic of Macedonia.

Ethics The research was conducted in accordance with the Declarationof Helsinki ethical guidelines and approved by the institutions where itwas conducted.

Conflict of interest The authors declare no conflicts of interest.

References

1. Abduljabbar MA, Afifi AM (2012) Congenital hypothyroidism. JPediatr Endocrinol Metab 25(1–2):13–29

2. Bongers-Schokking JJ, Koot HM, Wiersma D, Verkerk PH, deMuinck Keizer-Schrama SM (2000) Influence of timing and doseof thyroid hormone replacement on development in infants withcongenital hypothyroidism. J Pediatr 136(292–7):183

3. Burgard P, Rupp K, Linder M, Haege G, Rigter T, Weinreich SS,Loeber JG, Taruscio D, Vittozzi L, Cornel MC, Hoffmann GF (2012)Newborn screening programmes in Europe; arguments and effortsregarding harmonization. Part 2—from screening laboratory resultsto treatment, follow-up and quality assurance. J Inherit Metab Dis35(4):613–625

4. Dubuis JM, Glorieux J, Richer F, Deal CL, Dussault JH, Van Vliet G(1996) Outcome of severe congenital hypothyroidism: closing thedevelopmental gap with early high dose levothyroxine treatment. JClinEndocrinol Metab 81:222–7

5. Foley T, Kaplowitz PB, Kaye CL, Sundararjan S, Varma SK (2006)Update of newborn screening and therapy for congenital hypothy-roidism, American Academy of Pediatrics, Rose SR, Section onEndocrinology and Committee on Genetics, American Thyroid

Association, Brown RS, Public Health Committee, Lawson WilkinsPediatric Endocrine Society. Pediatrics 117:2290–2303

6. Foley TP, Jr (1996) Congenital hypothyroidism. In: Braverman LE,Utiger (eds). Werner’s and Ingbar’s. The Thyroid, Lipincott-Raven,Philadelphia 988–992

7. Hinton CF, Harris KB, Borgfeld L, Drummond-Borg M, EatonR, Lorey F, Therrell BL, Wallace J, Pass 197 KA (2010) Trendsin incidence rates of congenital hypothyroidism related to selectdemographic factors: data from the UnitedStates, California,Massachusetts, New York, and Texas. Pediatrics(Suppl 2):S37-47

8. Kaye CI, Accurso F, La Franchi S, Lane PA, Hope N, PangS, Schaefer GB, Lloyd-Puryear MA (2006) Introduction to thenewborn screening fact sheets. Pediatrics 118(3):1304–1312

9. Kaye CI, Accurso F, La Franchi S, Lane PA, Hope N, Pang S,Schaefer GB, Lloyd-Puryear MA (2006) Newborn screening factsheets. Pediatrics 118(3):e934–963

10. Kirsten D (2000) The thyroid gland: physiology and pathophysiolo-gy. Neonatal network: NN 19(8):11-26

11. Klett M (1997) Epidemiology of congenital hypothyroidism. ExpClin Endocrinol Diabetes 4:19–23

12. Kratzsch J, Pulzer F (2008) Thyroid gland development and defects.Best Pract Res Clin endocrinol metab 22(1):57–75

13. Loeber JG, Burgard P, Cornel MC, Rigter T, Weinreich SS,Rupp K, Hoffmann GF, Vittozzi L (2012) Newborn screeningprogrammes in Europe; arguments and efforts regarding har-monization. Part 1. From blood spot to screening result. JInherit Metab Dis 35(4):603–611

14. Lorey FW, Cunningham GC (1992) Birth prevalence of primarycongenital hypothyroidism by sex and ethnicity. Hum Biol 64(4):531–538

15. Rastogi MV, LaFranchi SH (2010) Congenital hypothyroidism.Orphanet journal of rare disease 5(17):216

16. Rose SR, Brown RS, Foley T, Kaplowitz PB, Kaye CI,Sundararajan S, Varma SK (2006) Update of newborn screen-ing and therapy for congenital hypothyroidism. Pediatrics117(6):2290–2303

17. Waller DK, Anderson JL, Lorey FW, Cunningham GC (2000) Riskfactors for congenital hypothyroidism: an investigation of infant’sbirth weight, ethnicity and gender in California 1990-1998.Teratology 62(1):36–41

18. Weston M. Screenwatch: neonatal screening. National Audit of theNeonatal (Guthrie) Screening Programme. Bandolier, 1995, 14-7

Eur J Pediatr