*Article* **Clinical and Genetic Characteristics of Patients with Mild Hyperphenylalaninemia Identified by Newborn Screening Program in Japan**

**Shino Odagiri 1, Daijiro Kabata 2, Shogo Tomita 2, Satoshi Kudo 1, Tomoko Sakaguchi 1, Noriko Nakano 1, Kouji Yamamoto 3, Haruo Shintaku <sup>4</sup> and Takashi Hamazaki 1,\***


**Abstract:** Phenylketonuria (PKU) and hyperphenylalaninemia (HPA), both identified in newborn screening, are attributable to variants in *PAH*. Reportedly, the p.R53H(c.158G>A) variant is common in patients with HPA in East Asia. Here, we aimed to define the association between p.R53H and HPA phenotype, and study the long-term outcome of patients with HPA carrying p.R53H. We retrospectively reviewed the genotype in 370 patients detected by newborn screening, and identified the phenotype in 280 (117, HPA; 163, PKU). p.R413P(c.1238G>C) was the most frequently found (*n* = 117, 31.6%) variant, followed by *p.R53H* (*n* = 89, 24.1%). The odds ratio for heterozygous p.R53H to cause HPA was 48.3 (95% CI 19.410–120.004). Furthermore, we assessed the non-linear association between the phenylalanine (Phe) value and elapsed time using the follow-up data of the blood Phe levels of 73 patients with HPA carrying p.R53H. The predicted levels peaked at 161.9 μmol (95% CI 152.088–172.343) at 50–60 months of age and did not exceed 360 μmol/L during the 210-month long observation period. The findings suggest that patients with HPA, carrying p.R53H, do not need frequent Phe monitoring as against those with PKU. Our study provides convincing evidence to determine clinical management of patients detected through newborn screening in Japan.

**Keywords:** phenylketonuria; hyperphenylalaninemia; phenylalanine hydroxylase; genetic analysis; neonatal screening; genotype–phenotype correlation

#### **1. Introduction**

Phenylketonuria (PKU) and hyperphenylalaninemia (HPA) are autosomal recessive disorders characterized by the deficiency of hepatic phenylalanine hydroxylase (PAH) [1]. This enzyme is encoded by *PAH* located on chromosome 12q, and comprising 13 exons and 12 introns. The genotype–phenotype correlations in PKU have been demonstrated using predicted PAH activity, which is the average in vitro residual PAH activity of two alleles [2].

The severity of the disorder is diverse, ranging from HPA to classical PKU that is characterized by high blood phenylalanine (Phe) levels [3]. The European and US guidelines for PKU recommend studying its molecular genetics and accumulating the data about correlations between the genotype and clinical phenotype [4,5]. Furthermore, more than 1100 variants of *PAH* have been recorded in the locus-specific database PAHvdb (http://www.biopku.org/pah/, accessed on 1 March 2021). However, these data are

**Citation:** Odagiri, S.; Kabata, D.; Tomita, S.; Kudo, S.; Sakaguchi, T.; Nakano, N.; Yamamoto, K.; Shintaku, H.; Hamazaki, T. Clinical and Genetic Characteristics of Patients with Mild Hyperphenylalaninemia Identified by Newborn Screening Program in Japan. *Int. J. Neonatal Screen.* **2021**, *7*, 17. https://doi.org/10.3390/ijns7010017

Received: 23 February 2021 Accepted: 15 March 2021 Published: 18 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

mainly gathered from European patients. The frequency of *PAH* variants in Japanese and other East Asian populations is different from that in Europeans. Therefore, recording data on the genotype–phenotype correlation in Japanese patients with PKU is required.

In Japan, a nationwide newborn screening (NBS) for PKU was launched in 1977. The current screening cut-off level for Phe on dried blood spots is set >120 μmol/L. When patients with PKU and related disorders are identified via NBS, their clinical phenotypes are classified as follows: (1) tetrahydrobiopterin (BH4) deficiency is ruled out by pteridine analysis and measurement of dihydropteridine reductase activity, (2) PAH deficiency is classified by pretreatment-Phe levels, and (3) Patients with <60 μmol/L Phe level are diagnosed with HPA while those with >360 μmol/L are further classified as BH4-responsive mild PKU or classical PKU based on BH4 loading test.

According to the integrative Japanese Genome Variation Database (iJGVD, https://ijgvd. megabank.tohoku.ac.jp, accessed on 18 February 2021), the allele frequency of p.R53H in the general Japanese population was reported to be as high as 5% [6]. Several studies have reported that p.R53H is associated with HPA phenotype [3,4,7–9].On the contrary, a few reports exist that correlates PKU with p.R53H [9,10].

In this study, we retrospectively reviewed the genotype–phenotype correlation of 370 patients who were diagnosed with PKU or HPA through NBS in Japan. Consequently, we found that the p.R53H genotype was associated with HPA phenotype, but not with that of PKU. Further, 73 of the patients with HPA carrying the p.R53H variant displayed stable blood Phe levels without needs of permanent treatments during the long-term follow-up. Our findings are especially useful for determining clinical management of such patients in Japan and other East Asian countries.

#### **2. Materials and Methods**

#### *2.1. Study Design and Participants*

This study was approved by the Institutional Review Board of Osaka City University Graduate School of Medicine (Osaka, Japan) (#3687). Written informed consent for the genetic analyses was obtained from all patients or their parents/guardians. The study included 370 Japanese patients analyzed for *PAH* at our facility during January 1998–March 2017, and for secondary examinations of HPA by NBS. We retrospectively reviewed the medical records of the patients to analyze their phenotype and genotype characteristics. The criterion for PKU phenotype was blood Phe level >600 μmol/L, and that for HPA phenotype was 120–600 μmol/L without Phe-restriction diet. Furthermore, we analyzed the genotype–phenotype correlation with *p.R53H* and the follow-up data of blood Phe levels of patients with HPA carrying *p.R53H* from the retrieved medical records.

#### *2.2. Statistical Analysis*

To evaluate whether the p.R53H variant is the cause underlying the HPA phenotype, we performed multivariate logistic regression analysis in the PKU (including HPA) pediatric patients who underwent genetic analysis using the variable indicating presence or absence of the HPA phenotype as the outcome and the number of the p.R53H variants as the explanatory variable. This regression model was adjusted the following covariates; the presence of the p.R413P, p.R241C(c.721C>T), p.R111\*(c.721C>T), IVS4-1G>A(c.442-1G>A), p.R243Q(c.728G>A), p.T278I(c.833C>T), p.R252W(c.754C>T), p.Ex6-96A>G(c.611A>G) and variants in the *PAH* gene and sex. The multiple imputation was conducted to impute the missing values for all variables.

Furthermore, in order to assess the change of Phe level over time, we performed nonlinear regression analysis with the Huber–White robust sandwich estimator of variance– covariance matrix. The robust estimator considers dependence in repeated measures within a single patient. Additionally, non-linear restricted-cubic-spline was conducted to assess the non-linear association between Phe level and the elapsed time. In this mode, Phe levels were used with natural-log transformation to satisfy the assumption of normality of the error distribution.

All statistical tests were performed two-sided with 5% significance level. The analyses were conducted using R (https://www.r-project.org/foundation/, accessed on 21 December 2020) (https://cran.r-project.org/, accessed on 21 December 2020).

#### **3. Results**

We retrospectively examined the genotype and clinical phenotypic characteristics of the 370 subjects with PAH deficiency and blood Phe levels of >120 μmol/L detected through NBS. Frequency of variants in *PAH* observed are summarized in Figure 1, and all PAH variations (*n* = 370) found in the study cohort are shown in Table S1. Briefly, p.R413P was most frequently found (*n* = 117, 31.6%), followed by p.R53H (*n* = 89, 24.1%), p.R241C (*n* = 53, 14.3%), p.R111\* (*n* = 47, 12.7%), and IVS4-1G>A (*n* = 44, 11.9%). Five patients had large deletions involving exons 5 and 6. In all, no variant was identified in three patients while one variant was identified in 27; two in 238; three in 11; and four in two. Phenotypes were identified in 280/370 patients, wherein HPA was identified in 117 and PKU in 163.

**Figure 1.** Frequency of *PAH* variants in 370 patients with elevated Phe levels detected by mass screening. Percentage of patients with mutations in *PAH* was detected by direct sequencing or multiplex ligation-dependent probe amplification (MLPA) in all the participants enrolled in this study. Numbers at the top of the vertical bars represent the actual number of patients with each mutation. Infrequent mutations have been omitted. Phe; phenylalanine, *PAH*; gene encoding phenylalanine hydroxylase, \* stop codon.

We, next, compared the frequency of the p.R53H variant between patients with HPA and PKU (Figure 2A). Consequently, we found that the frequency of p.R53H in those with HPA phenotype was 61.0% (74/117). From these, three were homozygous for p.R53H while 71 were heterozygous. The frequency of p.R53H in the patients with PKU phenotype was 6.7% (11/163). Of them, one was homozygous for p.R53H while 10 were heterozygous. In order to validate that the p.R53H variant is associated with HPA phenotype rather than that of PKU, we performed multivariate logistic regression analysis. Accordingly, we adjusted the influence of the other alleles frequently found in this population. The odds ratio for patients with heterozygous p.R53H to develop HPA phenotype was 48.3 (95% confidence interval [CI] 19.410–120.004) (Figure 2B). The odds ratio for patients with homozygous p.R53H to develop HPA phenotype was 26.2 (95% CI 2.054–334.832). This result demonstrated that patients with homozygous or heterozygous p.R53H variant are more likely to manifest HPA rather than PKU.

**Figure 2.** p.R53H variant is observed at a higher frequency in patients with HPA phenotype than in patients with PKU phenotype. (**A**) Number of patients with R53H variant in PAH (p.R53H) in each of the 280 patients with a known phenotype of either HPA or PKU. (**B**) Odds ratios for patients with no, heterozygous, and homozygous p.R53H variant to develop HPA phenotype. HPA, hyperphenylalaninemia; PKU, Phenylketonuria.

We further examined the status of p.R53H in the 11 patients with PKU. Consequently, we found that one patient had four variants, homozygous p.R53H and homozygous p.R158W; seven patients had three variants, heterozygous p.R53H and two other variants, p.R413P/p.R252W, p.R111\*/p.R252W, p.R111\*/IVS9+1G>A(c.969+1G>A), IVS4- 1G>A/p.R252W, p.R252W/p.EX6-96A>G, p.R252W/p.V399V(c.1197A>T), and p.P407S (c.1219C>T)/p.R158W(c.472C>T). Therefore, the presence of these variants other than *p.R53H* explains the manifestation of PKU in eight of the 11 identified patients [11–18]. The remaining three had only one heterozygous variant (p.R413P, p.R111\*, and p.R252W) other than p.R53H.

To determine which of the genetic variants were responsible for HPA phenotype, we analyzed them in the 117 patients with HPA. We categorized the patients into three groups based on their phenotypes predicted from the previously reported genotype–phenotype correlations (Table 1). The left column lists 74 patients with HPA who carried the p.R53H variant. In these patients, genetic variants that are known to cause PKU, such as p.R413P and p.R243Q, were recurrently found. On the contrary, only three of 74 patients had the p.R241C variant, which was associated with BH4-responsive PKU. Patients homozygous for p.R53H were not found in our study. The middle column shows cases carrying variants associated with HPA except for p.R53H. The right column shows remaining cases carrying variants with predicted PKU or unknown phenotype. In this group, p.R241C variant was found in 18 patients. Three patients with homozygous p.R241C manifested the HPA phenotype, while only one patient with same the genotype manifested PKU phenotype (data not shown).


**Table 1.** PAH variants identified in 117 patients with HPA phenotype.

There are no studies that report the long-term follow-up data of patients with HPA carrying p.R53H. We, thus, investigated the detailed clinicopathological characteristics of these patients. In the 74 patients with HPA carrying p.R53H, we excluded one patient carrying two pathogenic variants other than p.R53H. We analyzed the blood Phe levels in 73 patients (Figure 3). These patients included 31 males, 40 females, and two patients for whom gender data were not available. The observation period was of 0–210 months (median: 33 months). The mean of their blood Phe levels was 150 ± 30 μmol/L, and the maximum level found was 340 μmol/L. No patients had impaired mental and physical development. Furthermore, using the blood Phe levels at each visit of these patients and 635 data counts, we assessed the non-linear association between the Phe level and the elapsed time. The predicted Phe level at 0 month was 136.470 μmol/L (95% CI 131.491–141.638), but the Phe level peaked at 161.899 μmol/L (95% CI 152.088–172.343) at 50–60 months. Thereafter, the predicted Phe level gradually decreased to 106.246 μmol/L (95% CI 80.829–139.656) at 200 months of age. The predicted Phe levels did not exceed 360 μmol/L throughout the observation period.

**Figure 3.** Predicted Phe levels until 200-month-olds are less than 360 μmol/L in patients with HPA carrying p.R53H.

Changes in the predicted Phe levels in the patients. Solid line represents mean value; gray area represents 95% confidence interval. HPA, hyperphenylalaninemia; Phe; phenylalanine.

#### **4. Discussion**

We examined the genotypes of 370 patients with HPA or PKU identified by NBS in Japan. Consequently, we found that the p.R53H variant was recurrent in patients with HPA. Our study rigorously demonstrated that carrying compound heterozygous p.R53H and classical PKU-associated variants can predict HPA phenotype. Patients with PKU, carrying p.R53H, were likely to have two other pathogenic variants. Consequently, we found that the levels were maintained lower than 360 μmol/L until adolescence in the absence of Phe-restriction dietary treatment. Our findings provide convincing evidence that can help plan clinical management of patients detected through NBS in Japan.

As shown in Figure 1, the most common variants identified in this study were p.R413P, p.R53H, and p.R241C. A few East Asian countries, namely Japan, China, Taiwan, and Korea, share a common spectrum of *PAH* variants [11,12,17,19,20]. The variants—p.R243Q, p.EX6-96A>G, p.R241C, and p.R413P—have been frequently detected in the these countries. Further, five variants—p.V399V, p.R111\*, p.Y356X(c.1068C>A), IVS4-1G>A, and p.T278I—were shared with each other [12,17]. Allele frequency of p.R53H in the general population is 1.6% in the whole of East Asia, and 2.57% and 4.7% in Korea and Japan, respectively [21,22]. In case of patients detected through NBS, the allele frequency of p.R53H is 12.6% (93/740) in our study, and 1.27% and 2.11% in Korea and Taiwan, respectively [12,17]. Our study predicted the average Phe level for screening positive newborns carrying compound heterozygous p.R53H and other pathogenic variants to be 136.470 μmol/L (95% CI 131.491–141.638). In case of Korea and Taiwan, blood Phe levels of 240–599 μmol/L are used to define HPA; possibly, this is the reason why the p.R53H variant is less frequently detected in their affected populations [12,17]. The frequency of variants and genotypes in Japan is different from that recently published by Hillert et al. [23]. We considered that these differences were made because their study was based on information from only 55 patients.

In this study, 6.7% patients with PKU carried p.R53H (allele frequency, approximately 4%) that is close to 4.7% in the Japanese general population, as mentioned above [21]. This fact also suggested that the p.R53H variant is not directly associated with PKU phenotype. In order to explain the pathogenesis of PKU, these 11 patients must have two other pathogenic variants in addition to p.R53H. However, we found that three of them

carried only one other pathogenic variant. In these three patients we further performed multiple ligation-dependent probe amplification (MLPA) to exclude the possibility of large deletions. It should be noted that some genetic variants could not be identified even with the available methods. In previous studies, the detection rates of pathogenic variants among biochemically confirmed PKU cases were 96% (318/330) in China, 98% (139/142) in Taiwan, and 86.7% (137/158) in Korea [8,12,13]. Therefore, it is reasonable to assume that one other variant has not yet been identified in those patients who carry only one pathogenic variant other than p.R53H.

In an analysis of 512 people in the Japanese general population, 48 people (allele frequency, 4.7%) were heterozygous for p.R53H [21]. These heterozygotes displayed an average 19% increase in plasma Phe levels compared to wildtype homozygotes. Interestingly, Choi in Korea reported that the father of a patient with HPA was homozygous for p.R53H but did not manifest HPA [22]. Consistent with these reports, in our study, patients homozygous for p.R53H were not detected by NBS. Our study revealed that patients with HPA and the p.R53H variant frequently carried compound heterozygous classical PKU-associated variant, such as p.R413P, p.R243Q, and p.T278I (Table 1). They were less likely to have the compound heterozygous mild PKU-associated variant, such as p.R241C. A study using COS cells has reported the residual enzyme activities of p.R413P, p.R53H, and p.R241C as 1%, 79%, and 49%, respectively [12]. Collectively, it is reasonable to assume that the combination of variants with high residual enzyme activity will not exceed 120 μmol/L Phe levels to be detected through NBS.

With respect to the cut-off value of Phe levels set for NBS, there are contradictory views; some argue that this cut-off value should be set higher because the current value detects people as PKU-affected even though they do not require treatment. However, patients with BH4 deficiency present around 120 μmol/L Phe levels [24]. Therefore, we consider that the current cut-off value should not be changed such that patients with BH4 deficiency who sometimes maintain low Phe levels are not missed during the screening process. Furthermore, DNAJC12 deficiency that manifests HPA phenotype but not *PAH* variant or BH4 deficiency has been reported earlier [25]. Further, patients with DNAJC12 deficiency who were diagnosed early and accordingly treated showed normal development, while permanent neurological damage was observed with delayed diagnosis. Since p.R53H is present at a high frequency in the Japanese general population [21], it is possible that patients with DNAJC12 deficiency carry this variant. When patients with p.R53H have persistently high Phe levels and exhibit neurological symptoms, such as developmental delay and behavioral disorder, DNAJC12 deficiency should be considered for differential diagnosis.

Since there are no guidelines for HPA in Japan, patients with this condition are also monitored according to the guidelines for PKU. The European guidelines for PKU recommend frequent measurements of blood Phe levels; weekly up to the age of 12 months, followed by fortnightly till 12 years, and monthly for >12 years [4]. The current Japanese follow-up guidelines of PKU recommend that patients undergo measurement of blood Phe levels every four weeks until elementary school entrance [26]. We found that the predicted Phe levels in patients with HPA carrying p.R53H never exceeded 360 μmol/L without Phe-restricted diet until 200 month of age (Figure 2). These facts suggest that patients with HPA do not require frequent follow-ups. Further, we recommend that guidelines be developed on a priority basis for patients with HPA. In patients with PKU, first trimester of pregnancy should be monitored since the Phe levels tend to elevate during this period [4]. According to the European and US guidelines, when the untreated blood Phe level of women of childbearing-age with PKU is 120–360 μmol/L, treatment is unnecessary [4,5]. Extrapolating our study findings, we expect the blood Phe levels of pregnant women with HPA carrying p.R53H to not exceed 360 μmol/L during pregnancy. Further, our findings suggest that Phe-restricted diet and frequent follow-ups are not necessary for such patients even during pregnancy.

There are several limitations associated with the present study. Owing to its retrospective nature, the protocol for genetic testing was not consistent from subject to subject. Only

a subset of patients who were found to have one or fewer variants by direct sequencing were subjected to MLPA. In addition, for cases in which two or more pathogenic variants were identified by direct sequencing, no further analyses were performed, and the presence of large deletions may have been missed. Gülin Evinç et al. in Turkey report that children with untreated Phe levels of 240–360 μmol/L are at a higher risk for cognitive and behavioral impairment as compared to healthy children [27]. To evaluate neurodevelopmental outcome, appropriate developmental testing should have been performed in this study.

Based on study findings, we propose the following measures when p.R53H is detected by genotype search about HPA of NBS in Japan. When identifying patients with PKU carrying p.R53H, presence of two other pathogenic variants should be looked for. Whereas, when patients with HPA carrying p.R53H are identified, *p.R53H* should be considered underlying the phenotype. This study revealed that blood Phe levels in patients with HPA carrying p.R53H were continuously below 360 μmol/L without diet therapy. However, we should not determine the follow-up frequency only by genotype. We suggest monthly follow-up at least in their infancy according to their diet change and increase in protein intake. The patients with persistently low Phe levels during infancy may not require the frequent hospital visits and the frequent measurements of blood Phe levels. In summary, we could clarify genotype–phenotype correlations and long-term follow-up data on blood Phe levels in patients with p.R53H and one genotype associated with PKU in the Japanese population. We therefore consider that the findings of our study hold implications in strategically planning appropriate follow-ups for patients with HPA. Future studies should investigate patient-oriented outcomes in the clinical management.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2409-515 X/7/1/17/s1, Table S1: PAH variations (*n* = 370) found in the study cohort of PKU patients.

**Author Contributions:** Conceptualization and methodology, all authors; writing—original draft preparation, S.O., H.S., and T.H.; statistical analyses, D.K., S.T., and K.Y.; genetic analyses, S.K., T.S., and N.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported in part by Health and Labor Sciences Research Grant 20316977 (to T.H.), and AMED JP16ek0109050, JP19ek0109276, JP20ek0109482 (to T.H.).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Osaka City University Graduate School of Medicine (protocol code 3687, 27/02/2017).

**Informed Consent Statement:** Informed consent was obtained from all subjects or their parents/guardians involved in the study.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author, T.H., upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Review* **Newborn Screening for Congenital Hypothyroidism in Japan**

**Kanshi Minamitani**

Department of Pediatrics, Teikyo University Chiba Medical Center, Chiba 299-0111, Japan; kminami@med.teikyo-u.ac.jp; Tel.: +81-436-62-1211

**Abstract:** Congenital hypothyroidism (CH) is the most common preventable cause of intellectual impairment or failure to thrive by early identification and treatment. In Japan, newborn screening programs for CH were introduced in 1979, and the clinical guidelines for newborn screening of CH were developed in 1998, revised in 2014, and are currently undergoing further revision. Newborn screening strategies are designed to detect the elevated levels of thyroid stimulating hormone (TSH) in most areas of Japan, although TSH and free thyroxine (FT4) are often measured simultaneously in some areas. Since 1987, in order not to observe the delayed rise in TSH, additional rescreening of premature neonates and low birth weight infants (<2000 g) at four weeks of life or when their body weight reaches 2500 g has been recommended, despite a normal initial newborn screening. Recently, the actual incidence of CH has doubled to approximately 1:2500 in Japan as in other countries. This increasing incidence is speculated to be mainly due to an increase in the number of mildly affected patients detected by the generalized lowering of TSH screening cutoffs and an increase in the number of preterm or low birth weight neonates at a higher risk of having CH than term infants.

**Keywords:** newborn screening; lowering of thyroid stimulating hormone screening cutoffs; thyroid dysgenesis; thyroid dyshormonogenesis; transient congenital hypothyroidism; permanent congenital hypothyroidism; delayed rise in TSH; low birth weight

**1. Introduction**

Since thyroid hormone is indispensable for myelin sheath formation during the fetal, neonatal, and infant periods, dysfunction of thyroid hormone during these periods causes irreversible intelligence impairment. Furthermore, thyroid hormone stimulates growth hormone secretion, insulin-like growth factor 1 production, and bone maturation. Therefore, insufficient thyroid hormone activity can result in failure to thrive and early osteoporosis in adulthood. Primary congenital hypothyroidism (CH) is the most common congenital endocrine disorder caused mainly by thyroid dysgenesis or thyroid dyshormonogenesis. CH can be prevented by early detection and optimal treatment, and newborn screening programs for CH have been introduced in many countries worldwide. In Japan, newborn screening programs for CH started in 1979 and have markedly improved neurologic and health outcomes [1–3]. The present review provides an update on newborn screening programs for CH as well as the treatment and long-term outcomes of CH in Japan.

#### **2. Newborn Screening Programs for CH**

Before the development of newborn screening programs for CH, primary CH was mainly diagnosed from clinical symptoms based on a 12-item checklist (persistent jaundice, constipation, umbilical hernia, poor weight gain, xerosis cutis, sluggishness, macroglossia, hoarseness, coldness of limbs, edema, dilation of posterior fontanel, and goiter) [4]. However, because these symptoms are nonspecific in the neonatal period, they were often diagnosed late or overlooked. Newborn screening for CH through an enzyme immunoassay-based thyroid stimulating hormone (TSH) measurement on a filter paper blood spot sample was introduced as a nationwide screening program in 1979, and this method was upgraded to enzyme-linked immunosorbent assay in 1987 [2,5,6]. At present,

**Citation:** Minamitani, K. Newborn Screening for Congenital Hypothyroidism in Japan. *Int. J. Neonatal Screen.* **2021**, *7*, 34. https:// doi.org/10.3390/ijns7030034

Academic Editor: Ralph Fingerhut

Received: 27 May 2021 Accepted: 23 June 2021 Published: 28 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

patients with CH are treated according to the guidelines of mass screening for CH by the Japanese Society for Pediatric Endocrinology and the Japanese Society for Mass Screening, which were developed in 1998 [7] and revised in 2014 [8].

An initial TSH-based screening is performed using a filter paper blood spot sample collected on days 5–7 postpartum. Neonates with a TSH level of 15–30 mIU/L in whole blood on the filter paper blood spot sample are immediately referred to a regional medical facility for closer clinical examination. Neonates with a TSH level of 10–15 mIU/L are retested for TSH using the filter paper blood spot sample. Neonates with a TSH level >10 mIU/L in the retested sample are usually subjected to close examination [8].

Currently, the age of the first visit for close examination of patients ranges from 15.8 to 18 days, with an average of 17.3 days [3].

During close examination in a medical facility, the family history of thyroid disease and mother's history of iodine overload and medication are noted. In Japan, where iodine is abundant, dietary iodine insufficiency is rarely seen. A physical examination is performed mainly based on the abovementioned 12 items of the checklist. Serum TSH, free thyroxine (FT4), free triiodothyronine (FT3), and thyroglobulin levels are measured. The distal femoral epiphyseal ossification center (DFEC) is examined using X-ray, and the thyroid gland is identified using ultrasonography. Thyroid scintigraphy is reliable for the definitive diagnosis of thyroid dysgenesis. However, it is generally not performed in the neonatal period in Japan, probably because it is the only atomic-bombed country in the world.

Treatment is immediately initiated under following conditions: if a case has clinical symptoms, if the appearance of the DFEC is delayed, if the thyroid gland cannot be identified by ultrasonography, or if goiter is found. It is recommended to start treatment if the serum TSH level is ≥30 mIU/L or 15–30 mIU/L and the FT4 level is ≤15 pmol/L.

If no clinical symptoms are found, the serum FT4 level is within the normal range, and the TSH level is <15 mIU/L, a thyroid function test should be performed again. If the serum TSH level is >10 mIU/L at 3–4 weeks after birth, treatment initiation should be considered. It has been suggested that infants with a TSH level of ≥10 mIU/L at <6 months after birth and ≥5 mIU/L at 12 months after birth should be followed up carefully and treated.

#### **3. Incidence of CH in Japan**

Prior to the introduction of newborn screening for CH, the incidence of primary CH was 1:7400 [1]. However, once screening was started, the incidence increased to 1:3000 to 4000 since the 1990s and then to 1:2000 to 2500 since the 2000s [5]. The possible reasons for this increase include an increase in the number of mildly affected patients detected by the generalized lowering of TSH screening cutoffs and an increase in the number of preterm or low birth weight neonates at a higher risk of having CH than term infants, as well as epigenetic factors and changes in iodine intake and dietary habits [9].

The percentage of regions with the positive criterion of a TSH level of ≤30 mIU/L in whole blood on the filter paper blood spot sample doubled from 43.1% in 1993 to 89.4% in 2008 [10]. Therefore, the identification of an additional mild form of CH with gland-in-situ is thought to be responsible for the increase in CH incidence.

Preterm or low birth weight neonates are at a higher risk of having CH than term infants. In Japan, recent dramatic advances in neonatal care have led to an increase in the percentage of low birth weight neonates in Japan, i.e., from 5.2% in 1975 to 9.4% in 2017 [11]. However, there are no reports on the actual incidence of CH in low birth weight neonates in Japan.

#### **4. Epidemiology of CH in Japan**

Iodine deficiency is rare in Japan, which is originally an iodine-sufficient area. According to data from 1989, CH was caused by thyroid dysgenesis in 84% of cases (ectopic thyroid gland in 60% and athyreosis/hypoplasia of the thyroid gland in 24%) and by intrinsic defects of thyroid hormone synthesis (dyshormonogenesis) in the remaining 16% cases [12]. However, several recent studies using lower cutoff points for TSH levels have

reported an increased diagnosis of cases with gland-in-situ. In a 2008 Japanese study, 54% of primary CH cases were caused by thyroid dysgenesis (ectopic thyroid gland in 37% and athyreosis/hypoplasia of the thyroid gland in 17%) and the remaining 46% of CH cases occurred due to dyshormonogenesis [13].

A comprehensive genetic analysis identifies genetic abnormalities in 20% of Japanese patients [14–18]. Mutations in the DUOX2 gene are particularly common, identified in approximately 20% cases of dyshormonogenesis. In contrast, in thyroid dysgenesis, genetic mutations can only be identified in 5–10% of patients.

Currently, the National Center for Child Health and Development (Tokyo, Japan) analyzes genetic mutations in CH-associated genes, including *DUOX2*, *DUOXA2*, *FOXE1*, *GLIS3*, *IGSF1*, *IYD*, *NKX2-1*, *PAX8*, *SECISBP2*, *SLC26A4*, *SLC5A5*, *TG*, *THRA*, *THRB*, *TPO*, *TRH*, *TRHR*, *TSHB*, and *TSHR*, using next-generation sequencing methods.

#### **5. CH in Low Birth Weight Neonates**

Premature and low birth weight neonates may present with hypothyroxinemia without an increase in the TSH level through a variety of mechanisms, including the hypothalamic– pituitary–thyroid axis immaturity, nonthyroidal illness, dopamine administration, highdose steroid therapy, undernutrition, and exchange transfusion [19]. A delayed rise in TSH is a condition in which although the TSH level is below the cutoff point at initial screening, it increases later. A delayed rise in TSH is particularly common in low birth weight infants. A retrospective single-center matched case-control study shows that the percentage of small-for-gestational age infants was significantly higher in the delayed TSH rise group (71%) than in the comparison group (25%) [20]. In order not to overlook this pattern of delayed rise in TSH, since 1987, additional rescreening of premature neonates and low birth weight infants (<2000 g) at four weeks of life, when their body weight reaches 2500 g, or at discharge from the hospital is recommended, despite a normal initial newborn screening [21,22].

More than 50% of low birth weight infants before 30 weeks of gestational age manifest a temporary pattern of low levels of FT4 and normal or low levels of TSH termed 'transient hypothyroxinemia of prematurity' (THOP), due to the immaturity of the hypothalamicpituitary-thyroid axis, iodine deficiency, the withdrawal of maternal placenta FT4 transfer, nonthyroidal illness and exposure to some medications. The more premature the infants are, the more severely the thyroxine is reduced. Many studies have shown that levothyroxine sodium (LT4) has a poor effect on severe hypothyroxinemia [23–25], and the administration of LT4 to premature infants in Japan has been suggested to cause late onset circulatory collapse [26,27]. Infants with THOP should not be treated with LT4.

#### **6. Treatment of CH**

The Japanese Guidelines classify serum FT4 level <5, 5 to <10, and 10 to <15 pmol/L as indicating most severe, severe, and moderate cases, respectively, taking into consideration the consensus guidelines of the European Society for Pediatric Endocrinology [28].

Treatment starts with the administration of 10 μg/kg/day LT4 in powder form once daily before breakfast. In most severe cases, treatment starts with a dose of 12–15 μg/kg/day LT4. Infants with subclinical CH can be treated with 3–5 μg/kg/day LT4 because they often become hyperthyroid when given 10 μg/kg/day LT4 [8].

The target for serum FT4 levels should be >50% of the reference range by age. The target for TSH level should be the reference range by age. A follow-up is required at one, two, and four weeks after the start of LT4 treatment, at one-month intervals until one year of age, and then at 3–4-month intervals until the adult stage.

#### **7. Re-Evaluation**

A re-evaluation or definitive diagnosis should be made for the patients with CH after the age of three years, including the differentiation of transient from persistent CH [8]. After four weeks of LT4 withdrawal, a 123I thyroid scintigram, 123I uptake rate, saliva/serum

iodine ratio, perchlorate discharge test, thyroid function tests (TSH, FT4, FT3, and thyroglobulin), and thyroid ultrasonography are performed to diagnose athyreosis, hypoplasia, ectopic thyroid gland, hormone organification defect, and iodine concentration deficiency. If no abnormalities are detected upon these examinations, the patient is diagnosed with transient hypothyroidism. Infants treated with less than 1.25 μg/kg/day LT4 at three years of age are more likely to have transient CH [29,30]. In addition, infants who do not require an increase in LT4 dose after three years of age are more likely to have transient CH [31].

#### **8. Psychomotor Development**

Prior to newborn screening for CH, only 19.8% of infants with CH received treatment at an age of less than three months. Therefore, even after treatment, 43% of the patients showed mental retardation with IQ levels below 75, 33.3% of the patients showed IQ levels over 90, and two thirds of the patients were mentally retarded, including those on the borderline [32].

In early newborn screening, the recommended initial dose of LT4 was 5–8 μg/kg/day, and the initiation of treatment was often delayed until 4–5 weeks after birth. The patients with CH had an IQ that was lower by 6–20 points in comparison with controls, and the prognosis was particularly poor in severe children with a blood T4 level < 5 μg/dL at their initial visit. The mean IQ in the first nationwide survey in 1991 was 97.5 ± 14.8 (*n* = 81) [33] and that in the second survey in 1994 was 99.9 ± 13.7 (*n* = 151) [34].

Since the late 1990s, infants with CH have been treated with an initial dose of 10–15 μg/kg/day LT4, with treatment starting within two weeks after birth [35]. In the nationwide follow-up survey of CH children in 2003 [36], the DQ/IQ at 1–5 years of age was good, ranging from 104.1 to 107.3. Serious intellectual disability due to CH has almost been eradicated. However, children with severe hypothyroidism, such as athyreosis, during pregnancy presented with significantly lower IQ levels than those with other types of CH [37]. Furthermore, patients with severe CH also have cognitive, behavioral, and attention deficits in adolescence and adulthood [35,36].

#### **9. Growth, Puberty, Body Composition, and Quality of Life**

Prior to newborn screening for CH, while the frequency of children with a high degree of short stature equal to or less than −3 SD decreased from 45% to 11.8% after LT4 treatment, the frequency of children with a short stature equal to or less than −2 SD represented approximately 30% [32].

A report analyzing the height and body weight of 2341 patients with CH (1030 males and 1311 females) detected neither short stature nor obesity, but normal growth and constitution through a registration in the Medical Aid Program for Chronic Pediatric Disease of Specified Categories in 2002 [38]. A follow-up study in 2006 reported that the patients with CH had a nearly normal physique, with a height of 162.9 ± 8.4 cm and body weight of 60.8 ± 14.3 kg for male adults, height of 157.3 ± 5.2 cm and body weight of 52.4 ± 7.4 kg for female adults, and BMI of approximately 21.1 ± 3.0 for both males and females [39]. A report from Kanagawa shows no significant difference in adolescent growth patterns and adult height between patients with CH and healthy individuals, and no significant correlation between adult height and severity of hypothyroidism or the age of starting treatment was observed [40]. Some reports suggest that, even with good control, puberty tends to be earlier in girls with CH, judging from the age at menarche [41,42].

Patients in whom CH was detected shortly after the introduction of newborn screening have already finished compulsory education and reached the age for employment or marriage. The long-term quality of life (QOL) condition of these patients has been reported [39,41]. Regarding the employment status of these patients, full time employees represent 27% of patients, part-time employees represent 10%, unemployed represent 8%, married and unemployed represent 6%, students represent 43%, and others represent 6%, with no difference in employment status compared with the general population of the same generation. Patients with CH show no differences in academic backgrounds for

employees and the unemployed compared with those of the general population, as 15% of patients graduated from university/college, 7% dropped out from university/college, 22% graduated from vocational school or junior college, 41% graduated from high school, 4% dropped out from high school, and 11% graduated from junior high school. In terms of marital status, 8% of the patients with CH are married. In Japan, more than 90% of all households purchase life insurance. Life insurance aims to cover the loss associated with life, accidents, and sickness and also meets various needs, such as savings and postretirement security, but people with underlying conditions are often refused enrollment by life insurance companies. Among patients with CH, 46% have purchased life insurance and 65% of them applied for their insurance without declaring their disease [39].

#### **10. Summary**

Newborn screening for CH markedly improves the long-term intellectual outcome, physical growth, and QOL of patients with CH.

The incidence of CH is increasing every year. It is important to minimize the damage of hypothyroidism and further improve the outcomes by setting appropriate cutoff values, appropriate initial therapeutic doses, and appropriate treatment for mild CH and low birth weight infants.

Given that some patients with CH are anxious about explaining their disease to their spouses and the inheritance of the disease, proper counseling needs to be provided based on genetic diagnosis. In addition, it is becoming apparent that patients have various issues, including their transition from the pediatric to adult clinic, purchase of life insurance, and burden of medical expenses.

**Funding:** This research received no external funding.

**Data Availability Statement:** No new data were created or analyzed in this study. Data sharing is not applicable to this article.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


## *Article* **Re-Evaluation of the Prevalence of Permanent Congenital Hypothyroidism in Niigata, Japan: A Retrospective Study**

**Keisuke Nagasaki 1,\*, Hidetoshi Sato 1, Sunao Sasaki 1, Hiromi Nyuzuki 1, Nao Shibata 1, Kentaro Sawano 1, Shota Hiroshima <sup>1</sup> and Tadashi Asami <sup>2</sup>**


**Abstract:** Although newborn screening (NBS) for congenital hypothyroidism (CH) in Japan started more than 40 years ago, the prevalence of CH remains unclear. Prevalence estimations among NBS-positive CH individuals include those with transient hypothyroidism and transient hyperthyrotropinemia, and re-evaluation with increasing age is necessary to clarify the actual incidence. Thus, we re-evaluated the incidence of permanent CH. Of the 106,114 patients who underwent NBS in the Niigata Prefecture, Japan, between April 2002 and March 2006, 116 were examined further due to high thyroid-stimulating hormone levels (>8 mIU/L) and were included in the study. We retrospectively evaluated their levothyroxine sodium (LT4) replacement therapy status from the first visit to 15 years of age. Of the 116 NBS-positive patients, 105 (91%) were initially examined in our department. Of these, 72 (69%) started LT4 replacement therapy on the first visit. Subsequently, 27 patients continued LT4 replacement until 15 years of age after multiple re-evaluations. The prevalence of permanent CH in the Niigata Prefecture during this period was 1 in 2500–3500 children. Ultimately, 62.5% of patients on LT4 replacement discontinued treatment by 15 years of age. This is the first study to clarify the true prevalence of permanent CH in Japan.

**Keywords:** congenital hypothyroidism; newborn screening; Japan; re-evaluations; prevalence

#### **1. Introduction**

In Japan, newborn screening (NBS) for congenital hypothyroidism (CH) was initiated in 1979. The prevalence of CH in Japan was initially estimated at 1 in 7400 newborns prior to the commencement of the NBS [1]. National survey results after starting NBS indicated that the prevalence of CH in Japan was 1 in 1600–2500 children since the 2000s [1,2]. CH detected by NBS has been reported in various countries worldwide, but the incidence varies from 1 in 1000 to 1 in 6000 children [3]. Recently, reports have indicated that the prevalence of CH detected by NBS is increasing due to lower thyroid-stimulating hormone (TSH) cut-off values, racial composition changes, and an increase in the number of preterm or low-birthweight infants [3–5].

NBS-positive individuals present with transient hypothyroidism and transient hyperthyrotropinemia [6], and re-evaluation with increasing age is necessary to differentiate between these conditions and determine the actual incidence of CH. Permanent CH prevalence, excluding transient hypothyroidism, has not been clarified in Japan, which prompted our retrospective re-evaluation. To clarify the true prevalence of permanent CH in Niigata, Japan, we re-evaluated the patients who were NBS-positive for CH to determine the permanent CH prevalence based on the levothyroxine sodium (LT4) replacement status.

**Citation:** Nagasaki, K.; Sato, H.; Sasaki, S.; Nyuzuki, H.; Shibata, N.; Sawano, K.; Hiroshima, S.; Asami, T. Re-Evaluation of the Prevalence of Permanent Congenital Hypothyroidism in Niigata, Japan: A Retrospective Study. *Int. J. Neonatal Screen.* **2021**, *7*, 27. https://doi.org/ 10.3390/ijns7020027

Academic Editor: Bradford L. Therrell

Received: 29 April 2021 Accepted: 25 May 2021 Published: 28 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **2. Materials and Methods**

This was a single-institution retrospective cohort study. We retrospectively reviewed the LT4 replacement therapy status from the first visit after birth to 15 years of age. In this study, patients on LT4 replacement were defined as CH patients.

#### *2.1. NBS Method in the Niigata Prefecture*

Blood samples were collected on filter paper within the first 4 to 7 postnatal days, and the TSH level in the filter paper sample was measured using an enzyme-linked immunosorbent assay (TSH: Enzaplate N-TSH, Bayer Co., Tokyo, Japan). All CH screening tests were centralized at the Niigata Health Laboratory Center.

If the initial TSH level was between 8 and 30 mU/L, a second specimen was evaluated. If the TSH level in the second specimen was also greater than 8 mU/L, a confirmatory test was performed within 30 days of birth at the patient's medical institution. If the initial TSH level was more than 30 mU/L, a confirmatory test was performed within 14 days of birth at the patient's medical institution. Serum-free T4 (FT4), free T3 (FT3), and TSH levels were measured, and the thyroid morphology was evaluated by ultrasonography at the patient's medical institution.

The included patients comprised 116 newborns who tested positive for high TSH levels among the 106,114 newborns who underwent NBS in the Niigata Prefecture, Japan, between April 2002 and March 2006. Patients who had been initially examined at other hospitals were excluded.

#### *2.2. Re-Evaluations at Ages 2–4 Years*

Patients with a eutopic thyroid gland who underwent LT4 replacement therapy and remained euthyroid without LT4 dose increments after 12 months of age were re-evaluated by discontinuing LT4 for 4 weeks and performing thyroid function tests.

#### *2.3. Etiological Diagnosis Determination for CH after 5 Years of Age*

The methods are detailed in previous studies [7]. To summarize, after discontinuing LT4 replacement therapy, several tests (such as thyroid function test, thyrotropin-releasing hormone stimulation, 123I scintigraphy and radioactive iodine uptake (RAIU), saliva-toplasma radioiodine ratio, and perchlorate discharge (if the RAIU was 20% or more)) and thyroid ultrasonography were performed.

#### *2.4. Re-Evaluations at Final Height*

Patients with a eutopic thyroid gland who underwent LT4 replacement therapy and remained euthyroid after achieving their final height were re-evaluated by discontinuing LT4 for 4 weeks to confirm thyroid function.

#### *2.5. Criteria for Beginning and Discontinuing LT4 Replacement Therapy*

The criteria for beginning LT4 were a serum TSH level of 10–15 mU/L or higher at the time of the initial visit, persistent TSH level of ≥10 mU/L after the age of 3–6 months, or a persistent TSH level of ≥5 mU/L after the age of 1 year. The discontinuation criterion was a serum TSH level of <5 mU/L without LT4 replacement therapy, which was restarted if the TSH level remained above the 5–10 mU/L range.

#### *2.6. Primary and Secondary Outcomes*

The primary outcome was the prevalence of permanent CH detected by NBS, and the secondary outcome was the prevalence of transient CH among patients with CH who received LT4 replacement. This study was approved by the Niigata University Ethics Committee. We have published information related to the content of the research on the hospital's homepage. The patients and their parents were informed of their right to refuse access to their medical records for use in the study.

#### **3. Results**

The background characteristics of the subjects are listed in Table 1. Fifteen percent of the NBS-positive infants had a low birthweight. Of the 116 NBS-positive subjects with high TSH levels, 105 (91%) were evaluated at our hospital (Figure 1). Therefore, this study is based on a population base of 106,114 × 91% (i.e., 96,000 newborns). The LT4 replacement status for each age group is shown in Figure 2. Of these, 73 patients (69%) were initiated on LT4 at their first visit, while 32 (31%) were left initially untreated; 10 of the latter had persistent mildly elevated TSH levels and were initiated on LT4 by the age of 1 year. Thus, 73 out of 87 patients (84%) were treated with LT4 for 2 years, excluding those who were transferred or those for whom the follow-up had ended.

**Table 1.** Subject backgrounds.


*n*, number of patients; NBS, newborn screening; \* Data from 105 patients who underwent detailed examinations at our hospital; \*\* Data are shown as median (interquartile range).

**Figure 1.** Enrollment of the study subjects. Between April 2002 and March 2006, a total of 106,114 newborns were screened for CH in Niigata prefecture, and 116 were referred to pediatric endocrinologists. We evaluated 105 subjects (90.5%). Eleven patients did not visit our hospital due to reasons such as hospitalization in the neonatal intensive care unit. Fourteen patients (including 1 patient who died after cardiac surgery while on LT4 replacement) with normal thyroid function at the first visit, transient hypothyroidism, or transient hyperthyrotropinemia did not receive further follow-up. Twelve patients, including 9 on LT4 replacement, were transferred to another hospital. At the age of 15, 79 patients were being followed up.

**Figure 2.** The LT4 replacement status of each age group. M, month; yr, years; (*n*), the number in both parenthesis indicates the number of patients receiving LT4 replacement; \* transient hyperthyrotropinemia due to maternal antithyroid drugs or blocking TSH receptor antibody; \*\* thyroid function was normal from the first visit and the follow-up ended; \*\*\* the patient died after cardiac surgery.

Among patients aged between 2 and 5 years, 55 patients were re-evaluated and 16 discontinued LT4 replacement therapy. Consequently, 57 patients were on LT4 replacement therapy at the age of 5 years; of these, 52 patients were diagnosed etiologically with CH at the age of 5–7 years. LT4 replacement was discontinued in 24 patients, and 33 patients were continued on LT4 after the CH etiological diagnosis.

At the re-evaluation conducted after reaching final height, six patients discontinued LT4 replacement therapy, and at 15 years of age, 27 of the 79 patients (34%) who were followed up were receiving LT4 replacement therapy.

#### *3.1. Permanent CH Prevalence*

In addition to the 27 patients receiving LT4 replacement at 15 years of age, 10 patients were transferred or died while on LT4 replacement therapy. Thus, the number of patients with permanent CH from April 2002 to March 2006 ranged from 27 to 37, and the permanent CH prevalence was 1 in 2500–3500 children.

#### *3.2. Transient CH Prevalence*

Of the 74 patients who received LT4 replacement at 1 year of age, LT4 was discontinued in 46 patients by the age of 15 years, suggesting transient CH. There were 13 NBS-positive infants with an elevated TSH level and a birthweight of less than 2500 g who were examined at our hospital; 11 were initiated on LT4 replacement therapy, 3 were transferred to other hospitals, and 8 discontinued LT4 by the age of 15 years. Thus, the number of patients with transient CH or transient hyperthyrotropinemia ranged from 46 to 56, and the transient CH or transient hyperthyrotropinemia prevalence was 1 in 1700–2100 children.

#### **4. Discussion**

In this study, the re-evaluated prevalence of permanent CH was 1 in 2500–3500 children. Approximately 60% of the patients who received LT4 replacement therapy had transient CH or transient hyperthyrotropinemia and discontinued LT4 replacement therapy.

The CH prevalence reported worldwide is likely to include patients with transient hypothyroidism. In our study, when a patient with CH was defined as a patient on LT4 replacement therapy, the CH prevalence was approximately 1 in 1300 children at the time

of the first visit after birth, 1 in 1200 children at 1 year of age, and 1 in 1500 children at 5 years of age. The prevalence of CH at 1 year of age increased from the first visit due to the inclusion of patients who started LT4 replacement in early infancy without LT4 replacement at the initial diagnosis, i.e., persistent mild hyperthyrotropinemia. Therefore, when discussing CH prevalence, it is difficult to compare without considering the timing (i.e., age) of the incidences.

Reports indicate that the frequency of transient CH has increased, likely because of lower TSH level cut-off values. The incidence of transient CH in North America is approximately 5% to 10% of the NBS-positive children with CH [8]; however, recent reports indicate that 40% to 53% of the NBS-positive children with CH actually have transient CH [9,10]. Only one-third of patients with CH and eutopic thyroid gland needed to continue LT4 replacement after re-evaluation at the age of 3 to 6 [11]. In our study, approximately 60% of the 74 patients on LT4 replacement therapy at the age of 1 year discontinued it by the age of 15 years. Thus, the TSH cut-off value of 8 mU/L does include a certain number of patients with transient hypothyroidism. Even if LT4 replacement cannot be discontinued by 5 years of age, some patients may be able to discontinue LT4 thereafter and should be re-evaluated for transient CH at the appropriate period.

Increasing numbers of low birthweight and preterm infants may also be associated with CH prevalence [12,13]. Small for gestational age, especially, is a high risk for hyperthyrotropinemia [13]. The incidence of low birthweight was higher among infants with high TSH levels on NBS in our study than among the general population. However, all NBS-positive low-birthweight infants were able to discontinue LT4 replacement by the age of 15 years. Because high TSH levels in low-birthweight infants may be transient, these patients should be excluded when determining the prevalence of permanent CH.

Although this was not a nationwide survey, this is the first study to clarify the true prevalence of permanent CH in Japan. The study was limited by the inability to ascertain the LT4 replacement status in patients transferred to other hospitals, and the lack of investigation of the presence of CH patients who were not detected by NBS. We also did not examine factors associated with transient CH in this study. However, this study was strengthened by uniform data, as a single institution managed almost all the CH screenings in the Niigata Prefecture, and transient CH was excluded from long-term follow-up until the age of 15 years.

#### **5. Conclusions**

In our study, 62.5% of the LT4 replacement patients discontinued treatment by 15 years of age. From these results, the prevalence of permanent CH in the Niigata Prefecture during this period was 1 in 2500–3500 children.

**Author Contributions:** Conceptualization, K.N.; methodology, K.N., T.A., H.S., H.N., S.S., N.S., S.H. and K.S.; data curation, K.N.; writing—original draft preparation, K.N.; writing—review and editing, T.A., H.S., H.N., S.S., N.S., S.H. and K.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of Niigata University (protocol code: 2020-0497; date of approval: 29 April 2021).

**Informed Consent Statement:** Informed consent was obtained from the patients in an opt-out method for the publication of this paper.

**Data Availability Statement:** The data presented are available upon request from the corresponding author. The data are not publicly available because of privacy restrictions.

**Acknowledgments:** We would like to thank Keiko Hokari and Naoko Otabe of the Niigata Health Service Center for providing us with data on NBS in the Niigata Prefecture. We would like to thank all the patients who participated in this study.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Thirty-Year Lessons from the Newborn Screening for Congenital Adrenal Hyperplasia (CAH) in Japan**

**Atsumi Tsuji-Hosokawa 1,2 and Kenichi Kashimada 2,\***


**Abstract:** Congenital adrenal hyperplasia (CAH) is an inherited disorder caused by the absence or severely impaired activity of steroidogenic enzymes involved in cortisol biosynthesis. More than 90% of cases result from 21-hydroxylase deficiency (21OHD). To prevent life-threatening adrenal crisis and to help perform appropriate sex assignments for affected female patients, newborn screening (NBS) programs for the classical form of CAH have been introduced in numerous countries. In Japan, the NBS for CAH was introduced in 1989, following the screenings for phenylketonuria and congenital hypothyroidism. In this review, we aim to summarize the experience of the past 30 years of the NBS for CAH in Japan, composed of four parts, 1: screening system in Japan, 2: the clinical outcomes for the patients with CAH, 3: various factors that would impact the NBS system, including timeline, false positive, and LC-MS/MS, 4: Database composition and improvement of the screening program.

**Keywords:** congenital adrenal hyperplasia; 21-hydroxylase deficiency; newborn screening

#### **1. Introduction**

Congenital adrenal hyperplasia (CAH) is an inherited disorder caused by the loss or severely impaired activity of steroidogenic enzymes involved in cortisol biosynthesis (Figure 1A,B) [1,2]. More than 90 percent of cases result from 21-hydroxylase deficiency (21OHD) caused by mutations in *CYP21A2*. The prevalence of 21OHD is estimated to be 1:15,000–16,000 in the USA and Europe [3] and slightly lower in Japan (1:18,000) [4–6]. The clinical spectrum of the disease ranges from the most severe to mild forms, depending upon the degree of enzyme deficiency [2].

The disease is mainly classified into two forms: classical and nonclassical. The classical form is associated with two major problems: life-threatening adrenal crisis in both sexes and virilization of the external genitalia in 46,XX patients. The classical form is further subdivided into two subtypes, the severest, salt wasting (SW) form, and simple virilizing (SV) form. The SW form is associated with cortisol and aldosterone deficiencies, in which neonates are likely to develop life-threatening adrenal crises with severe hyponatremia and hyperkalemia. Virilization of the external genitalia in newborn females and precocious puberty due to overproduction of androgens by the adrenal cortex are the other major clinical manifestations of the SW and SV forms. However, the clinical phenotypes of the SW type and the SV forms may overlap, and attempts to differentiate them based on endocrinological evaluation without genetic analysis are sometimes inconclusive [2].

To prevent a life-threatening adrenal crisis and help perform appropriate sex assignments for affected female patients, newborn screening (NBS) programs for the classical form of CAH have been introduced in numerous countries [7]. In Japan, the NBS for CAH was introduced in 1989, following that for PKU and congenital hypothyroidism [4].

**Citation:** Tsuji-Hosokawa, A.; Kashimada, K. Thirty-Year Lessons from the Newborn Screening for Congenital Adrenal Hyperplasia (CAH) in Japan. *Int. J. Neonatal Screen.* **2021**, *7*, 36. https://doi.org/10.3390/ ijns7030036

Academic Editor: Toshihiro Tajima

Received: 27 May 2021 Accepted: 25 June 2021 Published: 29 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** Steroid synthesis in the adrenal cortex (**A**) and the pathophysiology of 21OHD (**B**). P5: Pregnenolone, 17αOHP5: 17-hydroxypregnenolone, P: Progesterone, 17αOHP: 17-hydroxyprogesterone, DOC: Deoxycorticosterone, 11DOF: 11 deoxycortisol, B: Corticosterone, 18OHB: 18-Hydroxycorticosterone, DHEA: Dehydroepiandrostendione, A4: Androstenedione, T: Testosterone. 17αOHP and other green steroids are included in the panel of LC-MS/MS screening in Japan. Steroids written in blue suggest its synthesis is reduced.

In contrast to the classical form, the nonclassical form has a milder phenotype in which clinical problems are not obvious during the neonatal period or childhood, generally developing during adolescence or adulthood [1,2]. The prevalence of nonclassical form in Japan is estimated much lower than that in western countries [8–10]. Although some of them are screened by the NBS, the screening program is not designed to detect all the newborns with the nonclassical form.

The aim of this review is to summarize the experience of the past 30 years of the NBS for CAH in Japan, comprising four parts: 1, screening system in Japan; 2, clinical outcomes for patients with CAH; 3, factors that would impact the NBS system, including timeline, false positive, and LC-MS/MS; and 4, database composition and improvement of the screening program.

#### **2. Screening System in Japan**

The NBS in Japan was introduced individually into the prefectural administration according to a government notification by the Ministry of Health and Welfare in 1977. The basis of the NBS system, such as the timeline, and the screening panel are identical in all local governments. After informed consent is obtained from a legal guardian, blood samples are collected by a heel prick blotted on a filter paper from neonates at 4–7 days from birth, and the filter paper samples are immediately sent to a laboratory allocated by the prefectural government.

The details of the screening system are different among laboratories, and as a representative, the screening algorithm in Tokyo was shown in Figure 2. In Tokyo, the 1st screening is divided into two procedures. The level of 17-hydroxyprogesterone (17αOHP) is initially determined by enzyme-linked immunosorbent assay (ELISA) without steroid extraction. We select blood samples in the 97th percentile or higher for 17αOHP values for subjecting the second-tier test, which is carried out after steroid extraction [4,5,11,12]. The cutoff criteria for the second-tier test are shown in Table 1.

The NBS has two different cutoff values: for "screening positive" and for "retest". When the 17αOHP level is higher than the screening positive cutoff value, the neonate is directly referred to a pediatric endocrinologist for further endocrinological evaluation. Neonates with 17αOHP levels more than the retest cutoff value are retested. When the 17αOHP levels are higher than the retest cutoff value two–three times, the screening is considered positive (Figure 2) [4,5,11–13].

**Figure 2.** Algorithm of CAH screening in Tokyo.

In some female patients, blood sampling for the screening is performed ahead of schedule due to atypical genitalia, which is one of the major clinical symptoms in female neonates with 21OHD and is frequently recognized at birth.

To reduce the number of false-positive results in preterm newborns, one of the most serious issues in the screening for 21OHD, some laboratories, including that of Tokyo, employ cutoff values based on gestational age and/or birth weight. The cutoff values were determined according to a pilot study of serum 17αOHP levels in full-term and preterm infants. As a representative screening system in Japan, the algorithm and criteria for the screening in Tokyo are shown in Table 1 and Figure 2 [5], respectively.


**Table 1.** Criteria of CAH mass screening in Tokyo.

\*1a Samples collected before the age of 7 days, \*1b Samples collected at the age of 7 days or after, \*<sup>2</sup> 1st test: body weight = birth weight, 2nd test and after: body weight = corrected body weight calculated by the formula as below. Corrected body weight at test (g) = birth weight (g) + (age at test − 7) × 20 (g). \*<sup>3</sup> For infants born small or large for gestational age, either the criteria of gestational age (corrected gestational age) or body weight was applied, whichever was a lower value. Since 2012, criteria according to weight have not been used, and solely gestational age-stratified cutoff has been used. \*<sup>4</sup> recall for the second (or the third) test of the screening. \*5 refer to hospitals for further endocrinological examinations.

For the quality control of the screening, most screening laboratories perform follow-up surveys of the patients who were referred to hospitals. In the surveys, clinical information of the patients, including confirmed diagnosis, is collected from the pediatric endocrinologists at the hospitals [13].

#### **3. Clinical Outcomes of the Newborn Screening for CAH in Japan**

#### *3.1. The Effects of the Screening*

The clinical profiles of 21OHD before the introduction of the screening differ remarkably from the current profiles [14]. Before the introduction of the screening, Suwa S et al. conducted a nationwide survey and reported the clinical profiles of 21OHD in Japan. According to the survey, the estimated prevalence of 21OHD was 1/43,764, and the average age in days when the patients firstly visited the hospitals was 1102. In the SW type, the average age of the first hospital visit was 55 days (male: 63 days (range, 1 days to 3 years), female: 47 days (range: 0 days to 3.9 years)), and in the SV form, the average age at first visit was 6.4 years (male: 5.9 years (range: 14 days to 34 years), female: 6.5 years (range: 0 days to 44 years)) (Table 2) [15]. The ratio of male to female was 1:1.5, and the number of male patients was significantly lower than that of females, implying that a substantial number of male patients were missed, i.e., the SV form remained undiagnosed or the fatal cases with the SW form in the neonatal-infantile period. Consistently, the survey revealed that the mortality rate was 10.6% in neonates with the SW form, which is consistent with the reports from other countries [15]. In 46,XX cases, 12.9% were firstly assigned as male because of atypical genitalia and corrected to female sex after the diagnosis of 21OHD [15].

**Table 2.** Age at diagnosis before and after implementation of the screening.


Numbers in parentheses indicate numbers of the subjects, \*, \*\*: according to the data reported by Suwa et al., 1994 [15] and Gau et al., 2020 [13], respectively.

> After the introduction of the screening, the clinical outcomes of 21OHD during the neonatal/infantile period were remarkably improved. The average ages at the first visit

were 8.2 and 7.6 days (male: 9.2 days, female: 6.0 days) in Sapporo and Tokyo, respectively (Table 2) [6,13]. To date, no fatal cases have been identified.

Although the follow-up surveys and the screening systems are not designed for detecting false-negative cases, based on a survey for the literature and the annual reports from NBS programs, no false-negative cases have been reported since the introduction of the screening [5,6,13]. We presume that the sex of all 46,XX cases was correctly assigned.

#### *3.2. The Progression of Salt Wasting during the First Two Weeks of Life*

Adrenal crisis is a life-threatening medical emergency, and eradicating the lethal cases of 21OHD is one of the major goals of newborn screening [16]. Although the fact that there were no reported fatal cases suggests the primary goal of the screening has been accomplished, 37.4% of 21OHD neonates already developed severe salt wasting, which is defined by Na < 130 mEq/L,K>7 mEq/L, on arrival at medical hospitals in Tokyo screening [13]. Furthermore, some of the 21OHD neonates exhibited life-threatening salt wasting, such as more than 10 mEq/L of serum K [13].

Severe adrenal crisis during the neonatal to early infantile period would cause neurological comorbidities. According to the nationwide survey before the introduction of NBS in Japan, a substantial number of the 21OHD patients revealed to have neurological comorbidities including intellectual disability and epilepsy. The prevalence associated with the SW form was higher, 18.5%, than with the SV form, 9.4%, suggesting that delayed diagnosis of adrenal crisis causes intellectual disability [17]. Consistently, in the retrospective study from the U.K., where NBS for 21OHD is not introduced, more than 20% of the SW-type 21OHD patients developed learning difficulties [14]. Those suggest that just eradication of lethal cases would not be sufficient for the goal of the 21OHD screening, and avoiding severe adrenal crisis should be considered.

Retrospective analysis of the follow-up survey of the NBS in Tokyo revealed that, in classical 21OHD patients, the serum Na and K levels linearly deteriorated with age in days, and the age when the regression lines reached Na < 130 mEq/L, K > 7 mEq/L approximately coincided at 11.1 and 12.3 days, respectively [13] (Figure 3). The risk of developing severe salt wasting increases during the second week of life without a threshold, and, therefore, an early intervention, ideally during the first week of life, is desirable [13,18,19].

**Figure 3.** Clinical features of serum sodium (Na) and potassium (K) levels in 21OHD neonates. Retrospective analysis of the NBS in Tokyo revealed that, in classical 21OHD patients, the serum Na and K levels linearly deteriorated with age in days, and the age when the regression lines reached Na < 130 mEq/L, K > 7 mEq/L approximately coincided at 11.1 and 12.3 days, respectively. (Modified from Gau et al., 2020 [13]).

#### *3.3. Triage of the Neonates with Salt Wasting by Body Weight Change*

The follow-up survey in Tokyo revealed that from the second week of life, changes in body weight provide a useful index in the evaluation of neonates with positive CAH screening results [13]. Neonates with decreasing body weight from the birth weight are likely to have classical 21OHD, and neonates with increasing body weight after birth are more likely to be false positives [13]. Furthermore, even in cases of 21OHD, the possibility of developing severe salt wasting, such as hyponatremia (<130 mEq/L) or hyperkalemia (>7 mEq/L), is extremely low without loss of body weight (Figure 4). Contrary to body weight change, the relevance of predicting severe salt wasting based on the 17αOHP level is extremely low because the 17αOHP level is not associated with Na or K levels [13].

**Figure 4.** Body weight change from birth is an excellent predictor of 21OHD and the risk of severe salt wasting. Body weight data of 21OHD patients at 7–14 days after birth were collected, and the change in body weight from birth weight was examined. None of the 21OHD patients with severe salt wasting (Na < 130 meq/L or K > 7.0 mEq/L) exhibited increased body weight (**A**,**B**).

Although the findings of body weight change in patients cannot be the direct criteria for the CAH screening protocol, they may assist in some individual cases, e.g., for triaging a neonate with a positive result who is living in a region with limited access to a pediatric endocrinologist or in which there is no CAH screening.

#### **4. Potential Issues of Testing Practices in the Newborn Screening for CAH in Japan**

#### *4.1. The Timeline of the Newborn Screening for 21OHD*

The timeline of the NBSs is becoming earlier worldwide because newly added inborn errors to the screening panel require early intervention immediately after birth. In the U.S., SIMD (Society for Inherited Metabolic Disorders) defines the critical condition as a condition in which serious symptoms may present acutely in the first weeks of life with a short pre-symptomatic window and require immediate treatment to mitigate morbidity and mortality [20]. More than 10 inborn errors of organic acid disorders and fatty acid oxidation disorders are involved in the list, and the SIMD recommends considering the list as an important starting point for discussion between clinicians and laboratories [20]. Accordingly, blood samples for screening are collected 48 h after birth in the U.S., and the recommended age in days when the first results are obtained should be seven [21]. Indeed, in 2018, 64% of the first results were available within 5 days after birth (Table 3) [22]. The situation is similar in the EU, and, in most countries, blood sampling starts 72 h after birth (Table 3) [23].

In the NBS for 21OHD, several factors should be considered in terms of timing for the blood sampling. Especially, given the rate of 37.4% neonates with severe salt wasting in Japan, earlier sampling can be discussed for the prevention of life-threatening salt wasting. However, an increase in serum 17αOHP level has been observed in unaffected neonates during the first 1–2 days of life, and there is evidence of false negatives associated with

the early collection of specimens in the U.S. [24]. Further, the timeline is determined by various factors of other diseases in the screening panels, which are different among countries (Table 3). For optimizing the timeline of the screening, we need careful discussion continuously.

**Table 3.** Summary of newborn screening in European countries, Oceania, and the U.S. (modified table from reference [25]) and the following website (https://www.hrsa.gov/advisory-committees/heritable-disorders/newbornscreening-timeliness.html, https://newbornscreening.hrsa.gov/your-state#w, https://www.newsteps.org/resources/ data-visualizations/newborn-screening-status-all-disorders, visited date, "23 April 2021").


\*<sup>1</sup> European countries whose population is approximately more than 5 million were listed. \*2 x (in screening panel section) = in screening panel, P = pilot/regional screening. \*<sup>3</sup> In the United States section, the number of states that include the disease in the screening panel is listed. In the AAD, OA, and FAOD section, states that implemented more than six of the metabolic disorders were counted. The interval between birth, sampling, and analysis of U.S. is recommended timeline. \*<sup>4</sup> Cord blood is used for some of the screening.

#### *4.2. High Rate of False Positive*

For the 17αOHP measurement, immunoassays have been used because of their sensitivity, cost, and simplicity. However, immunoassays lead to high rates of false positives, seriously affecting the screening efficiency [5,18,26].

One of the major reasons is the cross-reactivity with steroids, such as 17-hydroxy pregnenolone sulfate and 15β-hydroxylated compounds, which is high in preterm infants, and the ratio of false positives is extremely high in preterm infants. To minimize false positives, cutoff points stratified by gestational age and/or birth weight have been used in some screening systems. Although the stratified cutoff improves positive predictive value (PPV) to some extent, its efficiency is limited [5,27–30]. In the Tokyo system, gestational age and birth weight cutoff points have been used since the introduction of the NBS (Table 1). While the average PPV in Japan was reported as 6.6%, the Tokyo screening program achieved 25.8% (Figure 5). On the other hand, the PPV in preterm infants with a gestational age of ≤37 weeks was only 2% [5].

**Figure 5.** PPV (positive predicted value) of CAH screening in Tokyo according to the birth weights and the gestational ages of newborns judged as positive. (Modified from Tsuji et al., 2015 [5]).

Another cause for the high false-positive rate is the nature of 17αOHP itself. Historically, 17αOHP was originally considered as the pathogenic androgen in cases of 21OHD rather than as a diagnostic marker, and it has several shortcomings as a diagnostic for 21OHD [31]. The level of 17αOHP is high in cord blood during the first 1–2 days of life, and stress from other illnesses may result in the 17αOHP remaining high in unaffected neonates. Furthermore, in other forms of CAH, including 11-hydroxylase deficiency (11OHD), 3βhydroxysteroid dehydrogenase deficiency (3βHSDD), and P450 oxidoreductase deficiency (PORD), 17αOHP may be elevated to almost the same level as that of 21OHD [32]. For further improving PPV in 21OHD screening, measuring other biomarkers with high specificity for 21OHD would be required.

#### *4.3. LC-MS/MS Analysis of 17αOHP as a Second-Tier Test and Diagnostic Test for 21OHD*

To improve PPV, an alternative methodology should measure disease-specific markers other than 17αOHP or has high specificity for the target steroids. When used appropriately under highly regulated conditions, liquid chromatography-tandem mass spectrometry (LC-MS/MS) is considered as the gold standard for steroids assays [33–39], and the guideline of the Endocrine Society have recommended to employ LC-MS/MS for measuring 17αOHP of the second tier in neonatal screening since 2018 [16].

In addition to its specificity, the advantage of LC-MS/MS is the capability for the simultaneous assay of multiple steroids [33,34,36]. In Japan, a steroid profile panel from Siemens Healthineers AG (Frankfurt, Germany), "MS2-screening CAH" was developed for the CAH screening. Five steroids were selected for the panel: 17αOHP, 21-deoxycortisol (21- DOF), 11-deoxycortisol (11-DOF), 4-androstenedione (4AD), and cortisol (F). Accordingly, the cutoff criteria of the LC-MS/MS assay are not solely based on 17αOHP, but on 21DOF and the ratios of steroids, such as (17αOHP + 4AD)/F, 11DOF/17αOHP (Table 4) [33]. The combination of highly specific LC-MS/MS and simultaneous assays of five steroids is expected to dramatically improve the efficiency of the screening [33,34,36].

**Table 4.** Cutoff level of 17αOHP and other steroids assayed by LC-MS/MS in Saitama, Sapporo and Tokyo, Japan \*1.


\*<sup>1</sup> The algorithm of the screening is the same as shown in Figure 1. When the retest values are documented twice, the patients are judged as positive. \*<sup>2</sup> The result is judged as positive or retest when all parameters meet the criteria. \*3 The patient with the result that meets criteria A or B is considered as screening positive.

Indeed, the outcomes of LC-MS/MS are excellent. In 2018, the LC-MS/MS assay for 21OHD was employed in 5 of 37 prefectural laboratories in Japan. In immunoassay screening, out of 653 subjects with positive results, there were 38 confirmed cases of 21OHD, resulting in a PPV of 5.8% (38/653). On the other hand, the PPV in LC-MS/MS screening was 40.0% (6/15), indicating that the specificity of LC-MS/MS is remarkable [40]. Accordingly, in 2018, the Ministry of Health, Labor, and Welfare in Japan added LC-MS/MS to the list of recommended methodologies for the second-tier test of 21OHD screening [41].

In addition to improving the efficiency of 21OHD screening, the steroid profile assay by LC-MS/MS may bring further advantage to the screening, that is, assisting definitive diagnosis of 21OHD [34]. Although 21OHD can be diagnosed endocrinologically, the procedures and cutoff criteria are complicated because other rare forms of CAH, such as 11OHD, PORD, and 3βHSDD, should be differentiated from the diagnosis of 21OHD as we described in the previous section [32,42–48]. The nonspecific increase in 17αOHP levels in other forms of CAH has been considered as a potential clinical pitfall.

Currently, reliable methods for differentiating 21OHD from other forms of CAH are an adrenocorticotropic hormone (ACTH) stimulation test [16], urine steroid profile analyses using gas chromatography mass spectrometry [49], and genetic test, which cannot be used as a first-line diagnostic test because the procedure of *CYP21A2* gene analysis is extremely complicated [50–52].

It has been suggested that the ratios of steroids (17αOHP + 4AD)/F, 11-DOF/17αOHP, and 21-DOF may be more specific biomarkers for the diagnosis of 21OHD than that of 17αOHP and are expected to differentiate 21OHD from other types of CAH in which 17αOHP levels are elevated, such as 3βHSDD, 11OHD, and PORD. Although there are few reports of the levels of (17αOHP + 4AD)/F, 11-DOF/17αOHP, or 21-DOF in these forms of CAH, some cases suggest the potential usefulness of the five steroids in the screening panel. 21DOF would not be elevated in 3βHSDD because, in a model of partial 3βHSD deficiency preterm infants, 21DOF is not grossly elevated [34]. In patients with 11-OHD, 21-DOF levels are reported to be normal, and 11-DOF is markedly elevated, presumably increasing the 11-DOF/17αOHP ratio [53]. Urinary steroid profile analyses of PORD suggested that the ratio of pregnanetriolone (Ptl)/tetrahydrocortisone steroids (THEs) and a specific cutoff of 11β-hydroxyandrosterone (11HA) would be useful for differentiating PORD from 21OHD. Ptl, THEs, and 11HA are metabolites of 21DOF, 11DOF + cortisol, and 4AD, respectively, which are included in the LC-MS/MS screening panel [54].

Accordingly, in combination with other clinical symptoms and signs, such as poor body weight gain and high ACTH, 21OHD can be diagnosed based on the results of NBS by LC-MS/MS [31,34]. However, we cannot directly apply the screening criteria to the diagnostic criteria, and for establishing the diagnostic criteria, an accumulation of the cases is required.

#### **5. Database Composition and Improvement of Screening Program**

For better and more efficient management of the CAH, the screening programs need persistent improvement in quality. By examining reliable follow-up studies, the outcomes and the experiences of the screening should be retrospectively evaluated and shared among the screening laboratories [25,55,56]. For short-term outcomes, most laboratories and local governments have introduced follow-up surveys in Japan, using the results for more efficient screening by decreasing false positives and early availability of screening results.

On the other hand, the assessment of long-term outcomes for CAH patients identified by screening is challenging. A nationwide registry system is required to establish efficient long-term follow-up systems. In Japan, the current screening system depends on each local government; thus, the demanding task of organizing a cross-regional collaborative system that involves local governments, local laboratories, and medical institutes is required.

Recent studies have revealed that 21OHD patients have substantial risks for metabolic syndrome in adulthood [57–60]. The metabolic syndrome in 21OHD has been assumed to be due to long-term glucocorticoid therapy [61]. However, other causes, such as fetal environments, have also been suggested [62,63], and the pathophysiology of the condition remains unknown. Further, the quality of life of 21OHD patients in adulthood is largely unknown. Especially in female patients, their gender issues should be clarified in detail [64–66]. As mentioned previously, before the introduction of the screening, 21OHD patients had substantial risks for neurological sequelae, which are presumably caused by the adrenal crisis during the neonatal period [14,17]. Therefore, the introduction of NBS would reduce the risk for neurological sequelae [67], but currently, available data is limited. We should keep in mind that, even after the introduction of the screening, there are a substantial number of 21OHD patients who developed severe salt wasting before the introduction of therapy. Further, a recent study suggested the number of hyponatremic episodes is an independent risk for lower IQ, suggesting that for optimizing the management of 21OHD patients during childhood, preventing episodes of severe adrenal crisis is crucial [68]. Thus, clarifying long-term outcomes will provide valuable information for improving the screening logistics.

Long-term outcomes will also provide valuable insights for evaluating the costeffectiveness of screening. Although several analyses have been performed economically, they were based on short-term outcomes with various analytical models, leading to inconsistent results [69–71]. In Japan, economic analyses based on detailed clinical data have not been performed. To better understand the cost-effectiveness of screening, comprehensive approaches based on long-term outcomes are essential.

Despite not covering 21OHD patients, the introduction of several registry systems for rare congenital diseases has encouraged us. In the EU, some international collaborationbased registry systems for rare congenital diseases, such as the European Registry and Network for Intoxication type Metabolic Diseases (E-IMD) and the European Registry and Network for Homocystinurias and Methylation Defects (E-HOD), have been established [25,72–75]. Further, the Japanese Society for Inherited Metabolic Disease has successfully introduced the registry system, "JaSMIn", for patients with inherited metabolic

disease (https://www.jasmin-mcbank.com/, visited 23 April 2021). The limitation of this registry system is the unknown coverage rate due to voluntary patient registration. However, the registry is designed to cover a broad spectrum of rare inherited metabolic diseases that can be discovered by NBS, and it will provide valuable insights, enabling feedback on newborn screening in the future, including economic aspects.

Ideally, for establishing feedback systems with long-term follow-up surveys, close collaboration among the screening laboratory professionals, pediatricians, primary care providers, and clinical epidemiologists is essential [55,56,76,77]. Currently, to share the outcomes of the screening and updating of technical information, the Japan Society for Neonatal Screening has a collaborative laboratory integrated committee (Gijutsubu-kai). We expect that with this committee, pediatric endocrinologists and local governments would be able to construct large collaboration-based reports and infrastructure. A nationwide registry system in which all results of infants are registered and evaluated periodically would lead to further methodological improvements.

**Author Contributions:** Conceptualization, writing—original draft preparation, A.T.-H. and K.K.; supervision, K.K. Both authors have read and agreed to the published version of the manuscript.

**Funding:** The present study was partly supported by the grant from the Ministry of Health, Labour and Welfare of Japan (grant Nanbyo-Ippan046 to K.K. for research on intractable diseases, Principal Investigator: Tomonobu Hasegawa) (# 201811056A).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy policy.

**Acknowledgments:** We thank Masaru Fukushi for critical reading and valuable suggestions. We also thank Maki Gau (Tokyo Medical and Dental University), Nobuyuki Ishige (Tokyo Health Service Association, Newborn Screening), and Kazuhiro Watanabe (Tokyo Health Service Association, Newborn Screening) for valuable suggestions. We express our deep gratitude to the medical staff who provided the clinical information in the follow-up survey in Tokyo.

**Conflicts of Interest:** Atsumi Tsuji-Hosokawa received a research grant from the Japanese Society for Pediatric Endocrinology and Novo Nordisk Pharma Ltd. Kenichi Kashimada declare no conflict of interest.

#### **Abbreviations**


#### **References**

