Results from 28 years of newborn screening for congenital adrenal hyperplasia in Sapporo City

Newborn mass screening (MS) for congenital adrenal hyperplasia (CAH) was started in 1989 at nationwide of Japan. To date, the goals for prevention of life-threatening salt-wasting crises and prevention of incorrect sex assignment in virilized females are almost achieved. However in Sapporo city, which is located in the northern part of Japan, a pilot study was started in 1982 after the assessment of its feasibility and efficiency. 
 
Here we report the results of MS from 1982 to 2010 in Sapporo city. 
 
In 28 years, 17-hydroxyprogesterone (17-OHP) was determined in MS samples in 498,147 newborns and a total of 2,466 screened newborns had abnormal 17-OHP. Among them, 26 patients were diagnosed with 21-hydroxylase deficiency (21-OHD), which corresponds to a prevalence of 1:19,160 live birth in Sapporo city. It is almost identical to the worldwide incidence of CAH. 
 
Among 26 patients, 20 patients were classified into salt-wasting forms, 5 patients were classified into simple virilizing forms and one was classified into nonclassic forms. 19 patients were female and 7 patients were male. Two familial cases were detected. The mean examination day of 25 patients (data of one patient was not available) was 6.4 days after birth (from 0 to 18 days). Among 18 female patients, 11 patients were clinically suspected because of virilization prior to MS result becoming available and were referred to neonatal center. 4 female patients were referred after MS screening. One female patient was assigned to male after birth; however the patient was diagnosed as having 21-OHD by MS and the sex was changed after the diagnosis. One female patient was a familial case and during pregnancy prenatal treatment was done and her genitalia was normal female. Another one patient having normal female genitalia and no symptom, was diagnosed as nonclassic forms by MS. By contrast to female patients, all 7 male patients were detected by MS. 
 
Another type of CAH was not detected in our study. 
 
In conclusion, MS for CAH in Sapporo city can be considered reliable.


Introduction
Congenital adrenal hyperplasia (CAH) is characterized by impaired biosynthesis of cortisol and aldosterone and increased secretion of 17-hydroxyprogesterone (17OHP) and androgens (1,2). The most common form is 21-hydroxylase deficiency (21-OHD) (1,2), which affects about 1 child in 18, 000 and results in symptoms that vary with the severity of the enzymatic defect. Classic forms include the salt-wasting (SW) form, for which there is a high risk of life-threatening adrenal insufficiency during the first month of life, and the simple virilizing (SV) form. In both cases, female neonates present with markedly virilized external genitalia. Nonclassic (NC) form can manifest with hyperandrogenism later in life and do not warrant early detection through neonatal screening (3,4). The severity of the clinical presentation depends on the degree of 21-OHD, which is usually caused by deletion or mutations of CYP21A2 (1,2).
The primary goals of newborn mass screening (MS) for CAH are the prevention of life-threatening salt-wasting crisis in the most severe forms of CAH and the prevention of permanent negative effects of androgen overproduction in the SV form (2,4,5). MS for CAH from dried blood spots (DBS) has been implemented in several countries (3)(4)(5)(6)(7)(8)(9)(10)(11). An increased 17-OHP concentration in heel-prick blood is used to indicate patients at risk of having CAH. However, compared with the detection of other diseases by MS, CAH screening is associated with a high frequency for the requirement of a second sample (recall), because preterm infants have elevated 17-OHP levels due to several stresses (5,7,(11)(12)(13). In order to increase the efficiency of CAH MS, body weight-or gestational age-adjusted cutoff levels for 17-OHP have been used (11)(12)(13). However, the false positive rate (FPR) has not been significantly improved by these methods (14).
Most recently, Sarafoglou et al. (14) highlighted the methodology of comparing the first tier with repeat screening and second tier screening using liquid chromatography-tandem mass spectrometry (LC-MS/MS) for CAH.
While Japan initiated national newborn MS for CAH in 1989 (4, 15), a pilot newborn MS for CAH study was started in Sapporo in 1982 after the assessment of its feasibility and efficiency (16,17). The aim of our study was to summarize the results of the past 28 yr of newborn MS for CAH in Sapporo and analyze the efficiency of high-performance liquid chromatography (HPLC) as a second tier test.

Screening population
From 1982 through March 2010, the entire neonatal population of 498,147 newborns was screened for CAH in Sapporo. Blood was usually collected from 4-to 6-d old infants on filter paper and sent by mail to the Sapporo City Institute of Health. Blood samples of neonates suspected to have 21-OHD were collected from day 0 to 3 d after birth.
Abnormal screening results for CAH were directly reported by Sapporo City Institute of Health to coordinating physicians at Hokkaido University Hospital or NTT East Japan Hospital. The coordinating physician traced the individual in question and decided whether the child should have a second heel puncture or should be referred to a pediatric endocrinologist immediately.

Measurement of 17-OHP from DBS and cutoff levels
The methods of measurement of 17-OHP from DBS are summarized in Table 1. Both the method of the determination of 17-OHP and cutoff levels were altered at different points during the screening period (16,17). As a first tier test, from 1982 to 1985, 17-OHP levels were measured by an in-house ELISA kit after extraction by diethyl ether. From 1986 onwards, 17-OHP levels were measured using a commercially available ELISA kit (Enzaplate-N-17α-OHP ® from 1986 to 2001, Enzaplate-Neo-17α-OHP ® from 2002 to 2010, Siemens Japan, Tokyo, Japan) without prior extraction. Cutoff levels for 17-OHP in the first and second tests are also summarized in Table 1

Clinical and genetic analysis
Where available, clinical data were collected, including birth weight, gestational age, sex, 17-OHP levels at first screening, day of referral to the hospital, day at start of treatment and serum Na + and K + levels at first evaluation. The genotype of CYP21A2 was determined as previously reported (20,21).
Patients were defined as having the SW form when they had a serum sodium concentration <135 mEq/l or required fludrocortisone and NaCl supplementation for normal weight gain during hydrocortisone therapy. Patients were defined as having the SV form when they were treated with hydrocortisone only and had no symptoms of salt wasting. Patients were defined as having the NC form when they showed no symptoms.

Results of screening
During the 28-year study period, a total of 498,147 newborns were screened for CAH. Of the 26 CAH patients, 20 were classified as having the SW form, 4 were classified as having the SV form, and 2 were classified as having the NC form. Therefore, the frequency of the classical type of CAH was 1 in 20,756.
Of these, 26 individuals (weight range 2,340-3,774 g) were diagnosed with 21-OHD. In order to decrease the recall rate, we tried different methods for the second tier test, as shown in Table 1. While we did not employ birth weight and/or gestationally-adjusted cutoff levels for 17-OHP, the gestational week of neonates has been registered since 1995.
The recall rates are summarized in Table  2. From 1995 to 1999, 87,194 screening tests for CAH were performed. In this period, the total recall rate was 0.24%. The recall rates for premature and mature babies were 2.60 and 0.10%, respectively. From 2000 to 2010, 164,798 screening tests for CAH were performed, and 669 neonates required a second DBS after a second tier test, corresponding to a total recall rate of 0.41%. The recall rates for premature and mature babies were 5.38 and 0.06%, respectively.
The numbers of medical evaluations after recall are also summarized in Table 2. From 1995 to 1999, seven babies (0.008%) were subjected to medical evaluation in hospitals were after recall, and none of them were diagnosed as having 21-OHD. From 2000 to 2010, six babies (0.004%) were subjected to medical evaluation in hospitals after recall, and two of full-term infants were diagnosed as having 21-OHD (patient 17 and 24 in Table 3).

Characteristics of 21-OHD patients detected by MS
The characteristics of 21-OHD patients detected by MS are summarized in Table 3. Of 26 CAH patients, 20 were classified as having the SW form, 4 were classified as having the SV form, and 2 were classified as having the NC form. There were 19 female patients and 7 male patients. Two familial cases (patients 1 and 3 and patients 9 and 13) were detected. The mean day at examination was 6.4 d after birth (range=0 to 18 d). Among the 18 female patients, 11 patients were suspected as having CAH prior to the MS result becoming available as a result of detection of virilization and were referred to neonatal centers. Five female patients were referred after MS screening.
One patient (patient 17) had normal female genitalia and no symptoms; however, repeated measurements detected elevated levels of 17-OHP, and an ACTH stimulation test was performed at 2 mo of age. After ACTH stimulation, the serum 17-OHP level rose from 2.6 to 86.6 ng/ml. Based on these results, she was diagnosed as having the NC form (2). She is currently 11 yr old and shows no symptoms of androgen excess; however, her serum 17-OHP level remained high (4-9 ng/ml). In contrast to the female patients, all male CAH patients were diagnosed by MS. In the 21-OHD patients for whom we had data, the average day of initiation of therapy in those patients who had  undergone glucocorticoid therapy was 8.2 d after birth (range=2 to 21 d). Failure to thrive, severe hyponatremia, and hyperkalemia requiring immediate intravenous fluid replacement therapy were observed in two male patients (patient 14 and 22). In 4 patients (2 girls and 2 boys), laboratory signs of beginning salt wasting were found at the time of medical examination (sodium, <135 mEq/l). Of the remaining patients, 13 showed no detectable signs of salt wasting at the first examination; however, during followup, 11 patients demonstrated reduced serum sodium levels and poor weight gain despite appropriate glucocorticoid therapy. Accordingly, supplemental fludrocortisone and NaCl were administered in addition to hydrocortisone, and these patients were classified as having the SW form.
In patient 22, the time of referral to our hospital was 12 d after birth. Due to the time taken to contact the parents of the patient, performance of a full medical examination was delayed.
Regarding the genotype of CYP21A2, the most frequent genetic defects were deletions of CYP21A2 and IVS2-13A/C->G, which is in agreement with other studies (1, 2). Two SV patients had an I172N mutation in at least one allele. This mutation resulted in an enzyme with a small amount of activity. We were unable to determine the genotype of the NC patient (patient 17) using the conventional methods we employed. Patient 24 (NC) showed a very slight elevation of 17-OHP in the MS, which was verified by repeated measurement. As he did not have any symptoms, he was considered to have the NC form of 21-hydroxylase deficiency. Genetic analysis of CYP21A2 showed that this patient was a compound heterozygote for P30L and I172N+R356W. P30L has been reported as a very mild mutation (1,2), and this genotype could explain his mild phenotype.

Discussion
We summarized the results of neonatal MS for CAH in Sapporo. MS detected 26 patients affected by 21-OHD, and no other type of CAH was detected. Among them, 24 patients were diagnosed with the classical type of CAH. Therefore, the frequency of the classical type of CAH was 1 in 20,756, which is almost identical to the worldwide incidence of CAH (4)(5)(6)(7)(8)(9). On the other hand, the incidence of 21-OHD in Japan during the prescreening era was reported as 1 in 43,764 (15). These findings confirm that some 21-OHD patients went undiagnosed during the prescreening era and that screening for CAH has improved the rate of case detection by one-third.
The sex distribution of CAH patients in our study was 19 girls (73%) and 7 boys (27%). While a lower prevalence of MS-detected CAH in males has been reported in another study in Europe (22), our number of patients is currently of insufficient size to discuss sex distribution.
All male patients were detected by screening, as were 5 of 18 female neonates. Eleven of 18 female newborns were identified by genital abnormalities before screening, indicating that aside from the NC cases, 50% of the 21-OHD patients in our study were identified by MS. Other studies have shown that clinical criteria alone are insufficient to detect about 50% of CAH cases in the newborn period (22), a rate comparable with our results.
In our study, the average day of initiation of therapy was 8.2 d after birth (range=2 to 21 d). This is later than the 6.7 d reported by Steigert et al. (10) but earlier than the 17.6 d reported by Suwa et al. in a previous study in Japan (15).
Our screening identified two patients (patients 17 and 24) with the NC form, even though our screening program was not specifically designed to detect this form of CAH. It has been reported that several cases of the NC form showed elevated 17-OHP levels and were identified by neonatal screening (2,3,23,24). Identification of this form would require a lower cutoff value and entail a significant increase in false positive recalls. Therefore, some cases with the NC form are thought to have been missed by our screening.
MSs for CAH are still suboptimal because of low specificity, particularly in premature infants. To reduce the recall rate, we tried several methods of second tier test. In addition to the measurement of 17-OHP in DBS with extraction, from 1986 to 1999, the ratio of 17-OHP/cortisol was used, and from 2000 to 2010, HPLC was used. As a result, the total recall rate increased from 0.24 to 0.41% after the introduction of HPLC. The recall rate increased particularly in preterm infants, from 2.60 to 5.38%. One of the reasons for this might be our adoption of a low cutoff value of 17-OHP as a second tier test (2.5 ng/ml). A low cutoff value increases the recall rate, while a higher cutoff value would result in a loss of sensitivity. It must be kept in mind that the appropriate cutoff level of 17-OHP is still debatable.
If only term neonates are considered, HPLC decreased the recall rate (from 0.10 to 0.06%). Furthermore, the rate of medical evaluation after recall decreased (from 0.008 to 0.004%). Thus, HPLC is considered to be effective for these two points in term neonates. However, HPLC is timeconsuming (one analysis takes 40 to 50 min), and the difficulty of analyzing many samples using this method has limited its adoption in Japan. To avoid unnecessary recalls, several studies in the USA and Europe have reported the development of a liquid chromatographytandem mass spectrometry (LC-MS/MS)-based assay as the second tier test for CAH screening (14,25,26). This method can detect not only levels of 17-OHP but also those of other steroids such as androstenedion and cortisol, resulting in a decrease in recalls for CAH MS. In Japan, LC-MS/MS has been employed for neonatal screening of congenital metabolic diseases (27), and measurement of several steroids by LC-MS/ MS has been reported (28). A pilot study of LC-MS/MS as the second tier test for CAH screening is now ongoing (29). In the future, LC-MS/MS will be used as the second tier test for CAH, and we anticipate a further increase in screening efficiency.
In conclusion, MS for CAH in Sapporo is useful for identification of the SW and SV forms of 21-OHD. However, more efficient methods of MS for CAH should be developed in the future.