Abstract
Congenital adrenal hyperplasia (CAH) occurs due to enzyme defects in adrenal steroidogenesis. The 21-hydroxylase deficiency accounts for 90–95% of cases, triggering accumulation of 17-hydroxyprogesterone (17-OHP). Early diagnosis through neonatal screening allows adequate treatment and reduced mortality. The purpose of the study was to determine 17-OHP cutoffs for the diagnosis of CAH in a public newborn screening program in Southern Brazil. A retrospective, descriptive, cross-sectional study was conducted to analyze 17-OHP levels in dried blood samples collected on filter paper of 317,745 newborns screened at a public newborn screening center from May 2014 to April 2017. Neonatal 17-OHP was measured in DBS samples using a time-resolved fluoroimmunoassay (GSP® kit 3305-0010; PerkinElmer). Different cutoffs were determined and stratified by birth weight. The incidence of CAH was 1:15,887 live births in the state of Rio Grande do Sul, with 20 cases of classical CAH diagnosed during the study period. Most newborns (80.73%) were white, and the prematurity rate was 9.8% in the study population. The combination of different percentiles, 98.5th for birth weight 2001–2500 g and 99.8th for the other birth weight groups, decreased false-positive results and increased specificity compared with current reference values to identify classical CAH cases. The local 17-OHP cutoffs determined were higher than those currently used by this screening program for all birth weight groups. The calculation of reference values from local population data and the combination of percentiles proved to be a valuable tool for proper diagnosis of CAH and reduction in the number of false positives.
Introduction
Congenital adrenal hyperplasia (CAH) is an autosomal recessive disorder characterized by inadequate cortisol secretion (with or without insufficient aldosterone production) and androgen excess, caused by deficiency in one of the enzymes responsible for cholesterol degradation (1, 2). CAH is related in 90–95% of cases to mutations in the CYP21A2 gene, which modifies the activity of 21α-hydroxylase (21α-OH) (1, 3, 4), leading to a decrease in the conversion of progesterone to 11-deoxycorticosterone and 17-hydroxyprogesterone (17-OHP) to 11-deoxycortisol, decreasing mineralocorticoid and glucocorticoid levels and increasing progesterone and 17-OHP levels (5, 6).
Clinically, CAH is divided into classical and nonclassical forms. Historically, the classical was often subdivided into salt-wasting CAH, whose enzymatic activity is null, and simple-virilizing CAH, whose enzymatic activity exists but is extremely reduced (3) and now these are recognized as part of an overlapping disease spectrum because many compound heterozygous patients carry more than one mutation on either or both CYP21A2 alleles (7). In salt-wasting CAH, there is no cortisol or aldosterone production with a consequent increase in steroid precursors, which may cause genital virilization (3, 8). In female newborns presenting with ambiguous genitalia, the degree of virilization is classified according to the Prader scale (3). In male patients, genital abnormalities are not so easily detectable (9), hindering the possibility of a clinical diagnosis, which makes neonatal screening crucial in these cases. In about 75% of cases, a salt-wasting crisis will occur in the first weeks of life, which may lead to shock and death without adequate treatment (3, 10). In simple-virilizing CAH, cortisol, and aldosterone are released in small amounts, but when present in insufficient amounts, they induce the same physiological adaptations seen in salt-wasting CAH (3, 11). There is also genital virilization in female newborns and early puberty for both sexes (9). In nonclassical CAH, the enzyme activity is higher, 20–50%, and symptoms appear late and less severe (7).
CAH is a disease with high incidence rates. In Brazil, the average incidence of classical CAH ranges from 1:7500 to 1:18,000 live births (12). However, the incidence varies from state to state (13, 14, 15, 16).
Screening for CAH aims to identify newborns most likely to be affected by the more severe forms of the disease. Both salt-wasting and simple-virilizing CAH cause severe biochemical changes in a short period of time that may lead to death. Therefore, an adequate screening flowchart is necessary. Because 17-OHP is a direct substrate of 21α-OH, an elevated blood level of 17-OHP indicates impaired enzyme activity (11). Local public screening programs for CAH use a technique based on immunoassays performed on a drop of blood collected on filter paper. Brazil officially introduced the screening for CAH in 2013 for all states of the federation, although some states had already implemented CAH screening as isolated programs based on specific state public policies. In May 2014, Rio Grande do Sul, the southernmost state of Brazil, included CAH in its newborn screening program. In 2015, the public health system had already achieved an estimated 80% coverage of the state.
The reference values to determine 17-OHP using the immunoassay technique are often based on gestational age (GA), age at first sample collection, and/or birth weight (BW) (17, 18, 19). The Brazilian Ministry of Health recommends that 17-OHP be stratified by BW (12, 20). To reduce false-positive results, current reference values for CAH have been divided into four BW groups based on a pilot study conducted in the state of São Paulo in 2011 (12, 20, 21). However, it is recommended that each screening program standardize their own reference 17-OHP values for CAH (22) due to the country’s highly mixed-race population, as the population characteristics can be quite variable depending on the region of the country.
CAH is a disease with a high severity index, and the identification of true-positive cases is both important and highly necessary to reduce the number of false-positive results. An increase in specificity without loss of sensitivity is therefore an important goal for a screening program. Adjusting the reference values for the characteristics of the affected population and using an adequate methodology can contribute to improving diagnostic quality and reducing costs in screening programs. Thus, this study aimed to determine the most accurate 17-OHP cutoffs, stratified by BW, in dried blood samples collected on filter paper, based on the first 3 years of experience of the public screening program for CAH in the state of Rio Grande do Sul.
Methods
Design and population
The study was approved by the Research Ethics Committees of Hospital Materno Infantil Presidente Vargas and Universidade Federal de Ciências da Saúde de Porto Alegre, Brazil, under the approval numbers 2,127,786 and 2,162,936, respectively, and Certificate of Presentation for Ethical Appreciation (CAAE) number 48574015.3.3002.5345.
A retrospective cross-sectional study was conducted based on reports generated from the screening program database including all newborns screened at the Newborn Screening Referral Center of the state of Rio Grande do Sul from May 2014 to April 2017. Dried blood spot (DBS) samples were collected at public health units, general hospitals, and maternity hospitals. All samples analyzed in this study come from the public health system and correspond to approximately 80% of live births in the state. Exclusion criteria were newborns weighing <500 g or >5000 g to minimize weight registration errors and age <2 days or >30 days at sample collection to minimize the influence of false-negative results.
Classical CAH was defined as an elevated 17-OHP level confirmed by retest and/or clinical evaluation, followed by genotyping. False positives were characterized by the absence of genital abnormalities and/or salt wasting, with normal 17-OHP levels on retest.
The data gathered from the screening program database were compiled for later analysis and included outcome, 17-OHP, collection time, age at first sample collection, BW, prematurity (≤36 + 6 weeks of gestation), GA, maternal corticosteroid use, race, sex, blood transfusion, and multiple births.
17-OHP determination by immunoassay
Neonatal 17-OHP was measured in DBS samples using a time-resolved fluoroimmunoassay (GSP® kit 3305-0010; PerkinElmer Life and Analytical Sciences) on the GSP® Instrument (model 2021-0010; PerkinElmer). The assay is based on competitive reaction, and fluorescence is inversely proportional to the 17-OHP concentration in the sample. DBS values were converted to serum-equivalent values (conversion rate: 1 nmol/L = 0.33 ng/mL in whole blood).
Statistical analysis
A database plotted in Statistical Package for the Social Sciences, version 24.0, was used for statistical analysis. Numerical data were analyzed descriptively and expressed as percentages (%) and frequencies (n). BW was divided into four groups for analysis: ≤1500 g, 1501–2000 g, 2001–2500 g, and ≥2501 g. To determine 17-OHP cutoffs according to the four BW groups, the 97.5th, 98.5th, 99.5th, and 99.8th percentiles were calculated, since 17-OHP showed a non-Gaussian distribution according to the Kolmogorov–Smirnov test (P < 0.05) in all groups. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for all BW groups.
Results
According to the reports generated from the screening program database, 322,546 newborns were screened during the study period. A total of 2408 and 2393 newborns were excluded based on the weight and collection time criteria, respectively. Additionally, 350 newborns without 17-OHP results for the initial sampling were excluded, due to lack of weight registration and technically inadequate samples, resulting in a study population of 317,395 newborns.
The median and average time of first sample collection was 5.0 and 5.8 days, respectively. For retest, the median and average time was 21.0 and 25.9 days, respectively.
The descriptive analysis of the study population according to BW group is shown in Table 1. The 17-OHP cutoffs determined for the 97.5th, 98.5th, 99.5th, and 99.8th percentiles according to BW group are presented in Table 2.
Descriptive analysis of the study population according to birth weight.
| Sex (% (n)) | Birth weight group | |||
|---|---|---|---|---|
| ≤1500 g | 1501–2000 g | 2001–2500 g | ≥1500 g | |
| Female | 49.5% (1864) | 50.1% (2908) | 53.8% (10,809) | 48.0% (138,206) |
| Male | 47.8% (1800) | 47.7% (2772) | 45.2% (9076) | 51.7% (148,910) |
| NR | 2.7% (102) | 2.2% (126) | 1.0% (195) | 0.3% (977) |
| Race (% (n)) | ||||
| White | 79.0% (2977) | 79.4% (4608) | 79.3% (15,919) | 80.9% (233,022) |
| Black | 3.5% (132) | 2.9% (170) | 4.3% (860) | 4.0% (11,447) |
| Yellow | 0.1% (3) | 0.1% (6) | 0.2% (36) | 0.2% (469) |
| Indigenous | 0.4% (14) | 0.4% (21) | 0.5% (96) | 0.5% (1435) |
| Brown | 12.2% (459) | 12.0% (695) | 10.5% (2109) | 9.0% (26,064) |
| Other | 4.8% (181) | 5.3% (306) | 5.3% (1060) | 5.4% (15,656) |
| Outcome (% (n)) | ||||
| Confirmed CAH | 0% (0) | 0% (0) | 0.03% (6) | 0.0048% (14) |
| Not confirmed CAH | 100% (3766) | 100% (5806) | 99.97% (20,074) | 99.95% (288,079) |
| Corticosteroid use (% (n)) | ||||
| Yes | 57.4% (2162) | 42.8% (2484) | 24.5% (4922) | 8.8% (25,225) |
| No | 37.4% (1410) | 50.9% (2956) | 69.2% (13,896) | 85.0% (244,986) |
| NR | 5.2% (194) | 6.3% (366) | 6.3% (1262) | 6.2% (17,882) |
| Prematurity (% (n)) | ||||
| Yes | 91.4% (3441) | 79.7% (4629) | 40.9% (8220) | 4.4% (12,753) |
| No | 8.1% (306) | 19.8% (1149) | 58.3% (11,714) | 94.9% (273,403) |
| NR | 0.5% (19) | 0.5% (28) | 0.7% (146) | 0.7% (1937) |
| Twin birth (% (n)) | ||||
| Yes | 17.6% (662) | 19.9% (1158) | 11.4% (2291) | 0.8% (2360) |
| No | 82.1% (3092) | 79.7% (4628) | 88.1 % (17,698) | 98.7% (284,366) |
| NR | 0.3% (12) | 0.3% (20) | 0.5% (91) | 0.5% (1367) |
| Blood transfusion (% (n)) | ||||
| Yes | 15.4% (580) | 3.4% (197) | 0.7% (135) | 0.1% (430) |
| No | 83.3% (3136) | 95.2% (5530) | 97.9% (19,653) | 98.6% (283,963) |
| NR | 1.3% (50) | 1.5% (292) | 1.5% (292) | 1.3% (3700) |
NR, not reported.
Neonatal 17-OHP cutoffs for the 97.5th, 98.5th, 99.5th, and 99.8th percentiles according to birth weight as measured in dried blood spot samples.
| Birth weight | Percentiles | |||
|---|---|---|---|---|
| 97.5th (nmol/L) | 98.5th (nmol/L) | 99.5th (nmol/L) | 99.8th (nmol/L) | |
| ≤1500 g (n = 3766) | 393.9 | 487.8 | 691.5 | 927.2 |
| 1501–2000 g (n = 5806) | 159.6 | 182.1 | 228.4 | 301.2 |
| 2001–2500 g (n = 20,080) | 71.2 | 91.5 | 134.8 | 173.9 |
| ≥2501 g (n = 288,093) | 26.3 | 29.6 | 38.1 | 54.8 |
17-OHP, 17-hydroxyprogesterone.
Twenty newborns were diagnosed with classical CAH, with 17-OHP levels ranging from 98.35 to 2218.2 nmol/L (32.50–733.00 ng/mL) for those with BW 2001–2500 g and ≥ 2501 g. Clinical characteristics, laboratorial findings, and genotype are described in Table 3. No confirmed cases of CAH were detected for BW ≤ 1500 g and 1501–2000 g until the time of the present analysis.
Clinical characteristics, laboratorial findings, and genotype of CAH confirmed cases.
| BW category (g) | BW (g) | Sex | Time of first sample (days) | 17-OHP first sample (nmol/L) | Time of second sample (days) | 17-OHP second sample (nmol/L) | Age of diagnosis (days) | Na+ (mEq/L) | K+ (mEq/L) | Genotype | Phenotype |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 3 | 2040 | F | 2 | 1551.4 | 8 | 1578.6 | 7 | 126 | 5.3 | Del 30 kb/IVS2-13A>G | SW |
| 3 | 2270 | F | 6 | 112.1 | 15 | 1072.6 | 15 | 129 | 5.9 | IVS2-13A>G/IVS2-13A>G | SW |
| 3 | 2275 | M | 12 | 324.2 | 25 | 285.1 | 24 | 124 | 4.4 | p.GIn319Ter/p.GIn319Ter | SV |
| 3 | 2340 | M | 5 | 98.5 | 11 | 175.7 | 25 | 131 | 5.9 | p.Ile173Asn/p.Ile173Asn | SV |
| 3 | 2395 | M | 5 | 1421.1 | 20 | 1318.1 | 19 | 118 | 8.9 | IVS2-13A>G//gene conversion | SW |
| 3 | 2490 | F | 16 | 136.0 | 19 | 287.5 | 27 | 132 | 5.6 | IVS2-13A>G/IVS2-13A>G | SV |
| 4 | 2890 | F | 4 | 103.6 | 24 | 403.0 | 21 | 130 | 5.3 | p.Ile173Asn/gene conversion | SV |
| 4 | 2980 | M | 38 | 1396.8 | 51 | 1481.7 | 51 | 109 | 6.5 | Del CYP21A2/p.Arg357Trp | SW |
| 4 | 3160 | F | 4 | 978.7 | 15 | 781.7 | 15 | 133 | 6.1 | p.GIn319Ter, p.Pro454Ser/IVS2-13A/C>G | SW |
| 4 | 3200 | F | 4 | 187.3 | 17 | 633.3 | 17 | 132 | 5.0 | Del CYP21A2/p.Ile173Asn | SV |
| 4 | 3295 | M | 4 | 1157.5 | 42 | NA | 9 | 126 | 6.6 | IVS2-13A>G/IVS2-13A>G | SW |
| 4 | 3325 | F | 8 | 1308.9 | 20 | 2299.8 | 21 | 108 | 5.4 | Del CYP21A2/rearrangement | SW |
| 4 | 3370 | M | 4 | 1230.2 | 10 | 1630.1 | 10 | 126 | 8.0 | p.GIn319Ter, del 30kB/ p.GIn319Ter, p.Leu308PhefsTer6 | SW |
| 4 | 3450 | M | 15 | 2221.0 | NA | 354.5 | 15 | 120 | 5.7 | Del CYP21A2/cluster E6 | SW |
| 4 | 3500 | F | 3 | 197.6 | 11 | 948.4 | 11 | 125 | 7.4 | p.Leu308PhefsTer6/IVS2-13A>G | SW |
| 4 | 3580 | F | 4 | 1390.8 | 19 | NA | 18 | 131 | 6.0 | IVS2-13A>G/IVS2-13A>G | SW |
| 4 | 3610 | F | 5 | 836.3 | 11 | 1539.2 | 11 | 129 | 7.6 | p.Arg357Trp/ wildtype | SW |
| 4 | 3640 | M | 8 | 318.2 | 20 | 857.5 | 24 | 133 | 5.9 | IVS2-13A>G/IVS2-13A>G | SW |
| 4 | 3640 | F | 4 | 1099.9 | 9 | NA | 12 | 128 | 5.4 | IVS2-13A>G/IVS2-13A>G | SW |
| 4 | 3730 | F | 4 | 1560.5 | 122 | NA | NA | 127 | 4.3 | p.GIn319Ter/p.GIn319Ter | SW |
BW categories: 1 (<1500 g), 2 (1501–2000 g), 3 (2001–2500 g), 4 (>2501 g).
17-OHP, 17-hydroxyprogesterone; BW, birth weight; F, female; M, male; NA: not available; SV, simple virilizing; SW, salt wasting.
If only the 17-OHP cutoffs established for the 99.5th and 99.8th percentiles were used, BW 2001–2500 g would no longer identify two and three true cases of CAH, respectively, leading to an NPV of 10% and 15% and reducing the sensitivity of the reference values. Based on sensitivity and specificity analyses, different percentiles were combined to identify all true cases of CAH, resulting in the 17-OHP cutoffs presented in Table 4. A comparison between current reference 17-OHP values and the 17-OHP cutoffs determined in the present study is also shown in Table 4, along with their sensitivity, specificity, PPV, and NPV. The numbers in parentheses in Table 4 express the number of false positives, which drops significantly when the new cutoff point is adopted (n = 140 × 6 for the weight range < 1500 g and n = 261 × 11 for the weight range 1501–2000 g).
Current and proposed reference 17-OHP values in the study population.
| Birth weight | Current reference values | Proposed reference values | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 17-OHP (nmol/L) | Sensitivity (% (n)) | Specificity (% (n)) | PPV (% (n)) | NPV (% (n)) | 17-OHP (nmol/L) | Sensitivity (% (n)) | Specificity (% (n)) | PPV (% (n)) | NPV (% (n)) | |
| ≤1500 g | 334.5 | ND | 96.3% (3618) | 3.7% (140) | ND | 927.2 | ND | 99.8% (3752) | 0.2% (6) | ND |
| 1501–2000 g | 130.3 | ND | 95.5% (5532) | 4.5% (261) | ND | 301.2 | ND | 99.8% (5782) | 0.2% (11) | ND |
| 2001–2500 g | 85.4 | 100% (6) | 98.2% (19,683) | 1.8% (367) | 0.0% (0) | 91.5 | 100% (6) | 98.5% (19,757) | 1.5% (293) | 0.0% (0) |
| ≥2501 g | 45.8 | 100% (14) | 99.7% (286,794) | 0.3% (980) | 0.0% (0) | 54.8 | 100% (14) | 99.8% (287,218) | 0.2% (556) | 0.0% (0) |
17-OHP, 17-hydroxyprogesterone; ND, not determined; NPV, negative predictive value; PPV, positive predictive value.
There was a statistically significant difference in specificity and PPV between the 98.5th and 99.5th percentiles currently used for BW 2001–2500 g (P = 0.004). The same difference was observed in specificity and PPV between the 99.8th and 99.5th percentiles for the other BW groups (P ≤ 0.001).
Discussion
The present study determined BW-adjusted 17-OHP cutoffs for CAH screening 36 months after the implementation of a newborn screening program for this disease in the local population. During this period, 317,745 newborns were screened between the second and 30th day of life, which represents 73.3% of live births occurring in the period (23). The median age of newborns at first sample collection was 5.0 days, and the mean age was 5.8 days. Compared with a previous study of the same population over a shorter screening period, there was a decrease in the age at first sample collection, indicating an improvement in the newborn screening program with the possibility of an earlier diagnosis of the disease (24). The age of newborns at first sample collection is in accordance with the guidelines of the Brazilian Ministry of Health, which recommends that the first sample for newborn screening be collected between the third and fifth day of life (12).
Our data show a representative distribution in terms of number of patients in each of the BW groups, which further enhances the results. As expected, there was a higher rate of prematurity, twin birth, blood transfusion, and maternal corticosteroid use in the lower BW groups. In addition, the characteristics of the study population are in agreement with data from the Brazilian Institute of Geography and Statistics for Southern Brazil, where approximately 80% of the total population is white (25). Based on the analysis of 317,745 newborns, the incidence of classical CAH was 1:15,887 live births in the state of Rio Grande do Sul. This is consistent with a previous study analyzing 108,409 newborns and similar to the incidence reported for the neighboring state Santa Catarina of 1:14,967 live births (24, 26), which may be related to the similarity between the two state populations with a predominantly Caucasian background (24). However, this rate seems to differ from that of states located further to the Midwest and Southeast of the country, such as Goiás, São Paulo, and Minas Gerais, with incidence rates of 1:10,325, 1:10,460, and 1:19,939 live births, respectively (21, 22, 27).
A complex issue in neonatal screening for CAH is the high rate of false-positive results. Low BW, prematurity, and GA are variables that may alter 17-OHP results, increasing the rate of false positives (3, 7, 26), which can lead to doubtful conclusions that are likely to increase family stress. A high false-positive rate also increases the costs of the program, since repeat 17-OHP tests, serum confirmatory tests, and outpatient follow-up with medical staff are required to confirm the diagnosis (22, 24). Moreover, although fluoroimmunoassays are simple, fast to perform, and highly sensitive, their use for serum analyses is associated with high false-positive rates and low specificity (28, 29, 30).
In an attempt to reduce the false-positive rate, it is recommended that each laboratory determine the 17-OHP cutoffs according to the characteristics of the population to be screened (22), since miscegenation and geographical factors have a significant impact on population characteristics and, consequently, can influence the outcome and degree of disease involvement (3, 31). Quantification of 17-OHP is mainly based on GA or BW, as this information can reduce false-positive results while distinguishing between more and less severe forms of the disease (18, 21, 26). In the present study, we chose to determine the reference values based only on BW, since the data collected on GA were either not reliable or not systematically available in our setting. In Turkish CAH screening, 17-OHP reference ranges for GA were compatible with the ones stratified for BW. The same fluoroimmunoassay for 17-OHP measurement is used for the first step of screening in Turkey, and similar 99.8% cutoff values for 17-OHP were determined (32). Therefore, our choice to stratify the values by four BW groups is consistent with these findings and with other Brazilian newborn screening strategies and appears to increase PPV without affecting the sensitivity of the assay (12, 16).
The 17-OHP cutoffs proposed in the present study were based on data from the newborn screening program of the state of Rio Grande do Sul and determined from the combination of the 98.5th and 99.8th percentiles. The 98.5th percentile was used to determine 17-OHP cutoffs for BW 2001–2500 g because it yielded higher specificity and PPV, despite being below the percentile currently used and determined in this study for the other BW groups. The percentiles that best fitted the present study population showed a heterogeneous profile, differing from that reported in a study conducted in the Brazilian state of São Paulo, where the 99.8th percentile had the best results for false-positive rate, PPV, and specificity in all BW groups, considering both BW and age at first sample collection (22). These differences between the studies may be explained, at least in part, by the different inclusion and exclusion criteria. In the present study, while not stratifying the groups by collection time (< 36 or > 36 h), we excluded newborns at extreme ranges of weight (< 500 g or > 5000 g) and age at sample collection (< 2 days or > 30 days).
In a study conducted in 2008 in the state of Goiás, the authors stratified reference values by two BW groups (< 2500 g and ≥ 2500 g) and determined that 17-OHP levels lower than 242.4 nmol/L (80 ng/mL) could be considered of low risk for both groups, whereas 17-OHP levels below 121.2 nmol/L (40 ng/mL) and 90.9 nmol/L (30 ng/mL), respectively, were considered normal (32). In a study conducted in 2012 in the state of Minas Gerais, the authors stratified 17-OHP by four BW groups and determined a 17-OHP cutoff of 484.8 nmol/L (160 ng/mL) for newborns with BW <2500 g and 242.4 nmol/L (80 ng/mL) for those with BW ≥ 2500 g (16). In a study conducted in 2014 in the state of Santa Catarina, the reference values were divided into three BW groups (< 1250 g, 1250–2249 g, and ≥2500 g) and the cutoffs used were 169.6 nmol/L (56 ng/mL), 100 nmol/L (33 ng/mL), and 42.4 nmol/L (14 ng/mL), respectively (26). In line with the studies conducted in the states of Goiás, Minas Gerais, and Santa Catarina, the present study also identified significant differences between the reference 17-OHP values recommended by the Brazilian Ministry of Health (12). The differences between the reference values determined in different populations remained even when the same equipment was used for analysis, as demonstrated in the study conducted in Santa Catarina (26).
The 17-OHP cutoffs proposed here are higher than those currently used by the local newborn screening program, which would have a positive impact on the program by reducing false-positive results and improving specificity, especially for BW 2001–2500 g and BW ≥2501 g, considering the main goal of early detection of true cases of classical CAH (7, 12, 13). Güran et al. (33), in a recent study conducted in Turkey, showed that only 11% of newborns would require a repeat 17-OHP test if the cutoff values were increased, resulting in a 75% decrease in false-positive results. In contrast, it is known that the use of lower reference values improves screening sensitivity, but this did not apply to our study as sensitivity at different cutoffs remained unchanged. In addition, the use of a more representative sample, according to the prematurity rate (9.8%) of the study population, may also have influenced the improvement of these rates, in line with previous studies evaluating newborn screening for CAH (16, 21, 24, 26, 34).
The strengths of this study are the large sample size, the collection of a second sample of 17-OHP, and genotyping as a routine investigation, as already detailed in previously published data (13).
The potential limitations of this study are related to technical losses; erroneous sample collection, with clots, hemolysis, or insufficient material; and transport logistics for newborns to visit the health-care center for blood collection in the indicated period, resulting in approximately 1% of the newborns in the total sample not being covered by the program. The lack of data on GA and age at first sample collection for several cases prevented us from extending our analysis beyond the adjustments for BW. Therefore, future studies considering the correlation of BW with GA and/or age at first sample collection might further refine the diagnosis of true cases of CAH. An increased number of cases tested will be needed to determine if the new cutoffs are suitable for birth weight ranges <2000 g.
Conclusion
Based on the analysis of data obtained from 317,745 newborns, it can be concluded that the determination of new 17-OHP cutoffs using the combination of different percentiles (98.5th and 99.8th) was more effective in screening for classical CAH in our study population. Also, the 17-OHP cutoffs determined in this study are higher than those currently used for all BW groups. The combination of the different percentiles decreased the false-positive rate and increased specificity, thus reducing uncertainty about 17-OHP cutoff levels determined in another population (12, 21, 22).
Declaration of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Funding
This work was supported in part by the Brazilian National Institute of Hormones and Women’s Health/Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq INCT 465482/2014-7), INCT/FAPERGS 17/2551-0000519-8, and FAPERGS/MS/CNPq/SESRS 002/2013. The funders had no role in any part of the conduct of the study.
Author contribution statement
Simone M. de Castro: Conceptualization, Methodology, Formal Analysis, Investigation, Visualization, Writing – Original Draft, Review and Editing, Funding Acquisition. Paloma Wiest: Methodology, Formal Analysis, Investigation, Visualization, Writing – Original Draft. Poli Mara Spritzer: Conceptualization, Funding Acquisition, Writing – Review and Editing. Cristiane Kopacek: Conceptualization, Methodology, Funding Acquisition, Supervision, Writing – Review and Editing.
Acknowledgements
We thank Prof. Dr Mario Wagner for his support with statistical analysis, Prof. Dr Tania Bachega for her comments on the preliminary results, MSc Mayara Jorgens Prado and MD Luciana Amorim for their initial assistance in data collection, and Ricardo Smejoff for English review.
References
- 1↑
Janzen N, Peter M, Sander S, Steuerwald U, Terhardt M, Holtkamp U, & Sander J. Newborn screening for congenital adrenal hyperplasia: additional steroid profile using liquid chromatography-tandem mass spectrometry. Journal of Clinical Endocrinology and Metabolism 2007 92 2581–2589. (https://doi.org/10.1210/jc.2006-2890)
- 2↑
Schwarz E, Liu A, Randall H, Haslip C, Keune F, Murray M, Longo N, & Pasquali M. Use of steroid profiling by UPLC-MS/MS as a second tier test in newborn screening for congenital adrenal hyperplasia: the Utah experience. Pediatric Research 2009 66 230–235. (https://doi.org/10.1203/PDR.0b013e3181aa3777)
- 3↑
White PC, & Speiser PW. Congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Endocrine Reviews 2000 21 245–291. (https://doi.org/10.1210/edrv.21.3.0398) Erratum in: Endocrine Reviews 2000 21 245–291. (https://doi.org/10.1210/edrv.21.5.0411)
- 4↑
de Carvalho DF, Miranda MC, Gomes LG, Madureira G, Marcondes JA, Billerbeck AE, Rodrigues AS, Presti PF, Kuperman H, Damiani D, et al.Molecular CYP21A2 diagnosis in 480 Brazilian patients with congenital adrenal hyperplasia before newborn screening introduction. European Journal of Endocrinology 2016 175 107–116. (https://doi.org/10.1530/EJE-16-0171)
- 5↑
Prado MJ, de Castro SM, Kopacek C, de Mello MP, Rispoli T, Grandi T, da Silva CMD, & Rossetti MLR. Development of CYP21A2 genotyping assay for the diagnosis of congenital adrenal hyperplasia. Molecular Diagnosis and Therapy 2017 21 663–675. (https://doi.org/10.1007/s40291-017-0296-6)
- 6↑
Fontes N, Pereira M, Nascimento M, Oliveira E, Espada FV, & Fonseca M. Hiperplasia congénita da suprarrenal por deficiência de 21-hidroxilase: correlação genótipo-fenótipo. Sociedade Portuguesa de Endocrinologia, Diabetes e Metabolismo 2012 7 8–12. (https://doi.org/10.1016/S1646-3439(1270003-2)
- 7↑
Speiser PW, Arlt W, Auchus RJ, Baskin LS, Conway GS, Merke DP, Meyer-Bahlburg HFL, Miller WL, Murad MH, Oberfield SE, et al.Congenital adrenal hyperplasia due to steroid 21-hydroxylase deficiency: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology and Metabolism 2018 103 4043–4088. (https://doi.org/10.1210/jc.2018-01865) Erratum in: Journal of Clinical Endocrinology and Metabolism 2019 104 39–40. (https://doi.org/10.1210/jc.2018-02371)
- 8↑
Turcu AF, & Auchus RJ. Adrenal steroidogenesis and congenital adrenal hyperplasia. Endocrinology and Metabolism Clinics of North America 2015 44 275–296. (https://doi.org/10.1016/j.ecl.2015.02.002)
- 9↑
Mendes C, Vaz Matos I, Ribeiro L, Oliveira MJ, Cardoso H, & Borges T. Hiperplasia congénita da suprarrenal por deficiência de 21-hidroxílase: correlação genótipo-fenótipo. Acta Médica Portuguesa 2015 28 56–62. (https://doi.org/10.20344/amp.5579)
- 10↑
Kraemer GC. Avaliação dos casos com resultado positivo na Triagem Neonatal para Hiperplasia Adrenal Congênita no Paraná. Curitiba. [Mestrado em Saúde da Criança e do Adolescente]. Masters Dissertation: Universidade Federal do Paraná, 2017.
- 11↑
Nordenström A, Wedell A, Hagenfeldt L, Marcus C, & Larsson A. Neonatal screening for congenital adrenal hyperplasia: 17-hydroxyprogesterone levels and CYP21 genotypes in preterm infants. Pediatrics 2001 108 E68. (https://doi.org/10.1542/peds.108.4.e68)
- 12↑
Araújo JPB, & Goldbeck AS.Triagem Neonatal: Hiperplasia Adrenal Congênita. Ministério da Saúde, Secretaria de Atenção à Saúde, Departamento de Atenção Especializada e Temática. 46p. Editora MS-OS: Brasília, Brazil: Ministério da Saúde, 2015.
- 13↑
Kopacek C, Prado MJ, da Silva CMD, de Castro SM, Beltrão LA, Vargas PR, Grandi T, Rossetti MLR, & Spritzer PM. Clinical and molecular profile of newborns with confirmed or suspicious congenital adrenal hyperplasia detected after a public screening program implementation. Jornal de Pediatria 2019 95 282–290. (https://doi.org/10.1016/j.jped.2018.03.003)
- 14↑
Souza JFG de P, Filho DCC, Araújo LDVCP, & Alvarenga AR. Vista da Incidência e distribuição geográfica de hiperplasia adrenal congênita em Goiás entre 2006 a 2012. RESU 2015 3 31–38.
- 15↑
Nunes AKC, Wachholz RG, Rover MRM, & Souza LC. Prevalência de patologias detectadas pela triagem neonatal em Santa Catarina. Arquivos Brasileiros de Endocrinologia e Metabologia. 2013Jul 2013 57 360–367. (https://doi.org/10.1590/S0004-27302013000500005)
- 16↑
Barra CB, Silva IN, Pezzuti IL, & Januário JN. Neonatal screening for congenital adrenal hyperplasia. Revista da Associação Médica Brasileira 2012 58 459–464.
- 17↑
Lee JE, Moon Y, Lee MH, Jun YH, Oh KI, & Choi JW. Corrected 17-alpha-hydroxyprogesterone values adjusted by a scoring system for screening congenital adrenal hyperplasia in premature infants. Annals of Clinical and Laboratory Science 2008 38 235–240.
- 18↑
van der Kamp HJ, Oudshoorn CG, Elvers BH, van Baarle M, Otten BJ, Wit JM, & Verkerk PH. Cutoff levels of 17-alpha-hydroxyprogesterone in neonatal screening for congenital adrenal hyperplasia should be based on gestational age rather than on birth weight. Journal of Clinical Endocrinology and Metabolism 2005 90 3904–3907. (https://doi.org/10.1210/jc.2004-2136)
- 19↑
Bialk ER, Lasarev MR, & Held PK. Wisconsin's screening algorithm for the identification of newborns with congenital adrenal hyperplasia. International Journal of Neonatal Screening 2019 5 33. (https://doi.org/10.3390/ijns5030033)
- 20↑
Kopacek C. Triagem neonatal da hiperplasia adrenal congênita no SUS. Boletim Científico de Pediatria 2015 4 59–64.
- 21↑
Hayashi G, Faure C, Brondi MF, Vallejos C, Soares D, Oliveira E, Brito VN, Mendonca BB, & Bachega TA. Weight-adjusted neonatal 17OH-progesterone cutoff levels improve the efficiency of newborn screening for congenital adrenal hyperplasia. Arquivos Brasileiros de Endocrinologia e Metabologia 2011 55 632–637. (https://doi.org/10.1590/s0004-27302011000800019)
- 22↑
Hayashi GY, Carvalho DF, Miranda de MC, Faure C, Vallejos C, Brito VN, Rodrigues AS, Madureira G, Mendonca BB, & Bachega TASS. Neonatal 17-hydroxyprogesterone levels adjusted according to age at sample collection and birthweight improve the efficacy of congenital adrenal hyperplasia newborn screening. Clinical Endocrinology 2017 86 480–487. (https://doi.org/10.1111/cen.13292)
- 23↑
Kopacek C, Castro SM, Chapper M, Amorim LB, Lüdtke C, & Vargas P. Evolução e funcionamento do Programa Nacional de Triagem Neonatal no Rio Grande do Sul de 2001 a 2015. Boletim Científico de Pediatria 2015 4 70–74.
- 24↑
Kopacek C, de Castro SM, Prado MJ, da Silva CM, Beltrão LA, & Spritzer PM. Neonatal screening for congenital adrenal hyperplasia in Southern Brazil: a population based study with 108,409 infants. BMC Pediatrics 2017 17 22. (https://doi.org/10.1186/s12887-016-0772-x)
- 25↑
IBGE Censo Demográfico. População Residente por Cor Ou Raça, Segundo A Situação do Domicílio, o Sexo e A Idade. Rio de Janeiro, Brazil: Instituto Brasileiro de Geografia e Estatística, 2023. (available at: https://sidra.ibge.gov.br/Tabela/3175#resultado)
- 26↑
Nascimento ML, Cristiano ANB, Campos de T, Ohira M, Cechinel E, Simoni G, Lee JS, Linhares RMM, & Silva PCA. Ten-year evaluation of a Neonatal Screening Program for Congenital Adrenal Hyperplasia. Arquivos Brasileiros de Endocrinologia e Metabologia 2014 58 765–771. (https://doi.org/10.1590/0004-2730000003310)
- 27↑
Silveira EL, dos Santos EP, Bachega TA, van der Linden Nader I, Gross JL, & Elnecave RH. The actual incidence of congenital adrenal hyperplasia in Brazil may not be as high as inferred--an estimate based on a public neonatal screening program in the state of Goiás. Journal of Pediatric Endocrinology and Metabolism 2008 21 455–460. (https://doi.org/10.1515/jpem.2008.21.5.455)
- 28↑
Rossi C, Calton L, Hammond G, Brown HA, Wallace AM, Sacchetta P, & Morris M. Serum steroid profiling for congenital adrenal hyperplasia using liquid chromatography-tandem mass spectrometry. Clinica Chimica Acta 2010 411 222–228. (https://doi.org/10.1016/j.cca.2009.11.007)
- 29↑
Carvalho VM, Nakamura OH, & Vieira JGH. Simultaneous quantitation of seven endogenous C-21 adrenal steroids by liquid chromatography tandem mass spectrometry in human serum. Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences 2008 872 154–161. (https://doi.org/10.1016/J.JCHROMB.2008.07.035)
- 30↑
Travers S, Martinerie L, Bouvattier C, Boileau P, Lombès M, & Pussard E. Multiplexed steroid profiling of gluco- and mineralocorticoids pathways using a liquid chromatography tandem mass spectrometry method. Journal of Steroid Biochemistry and Molecular Biology 2017 165 202–211. (https://doi.org/10.1016/j.jsbmb.2016.06.005)
- 31↑
Pang SY, Wallace MA, Hofman L, Thuline HC, Dorche C, Lyon IC, Dobbins RH, Kling S, Fujieda K, & Suwa S. Worldwide experience in newborn screening for classical congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Pediatrics 1988 81 866–874. (https://doi.org/10.1016/S0022-5347(1742164-1)
- 32↑
Güran T, Tezel B, Çakır M, Akıncı A, Orbak Z, Keskin M, Selver Eklioğlu B, Ozon A, Özbek MN, Karagüzel G, et al.Neonatal screening for congenital adrenal hyperplasia in Turkey: outcomes of extended pilot study in 241,083 infants. Journal of Clinical Research in Pediatric Endocrinology 2020 12 287–294. (https://doi.org/10.4274/jcrpe.galenos.2020.2019.0182) Erratum in: Journal of Clinical Research in Pediatric Endocrinology 2021 13 250. (https://doi.org/10.4274/jcrpe.galenos.2021.e003)
- 33↑
Güran T, Tezel B, Gürbüz F, Selver Eklioğlu B, Hatipoğlu N, Kara C, Şimşek E, Çizmecioğlu FM, Ozon A, Baş F, et al.Neonatal screening for congenital adrenal hyperplasia in Turkey: a pilot study with 38,935 infants. Journal of Clinical Research in Pediatric Endocrinology 2019 11 13–23. (https://doi.org/10.4274/jcrpe.galenos.2018.2018.0117)
- 34↑
Pezzuti IL, Barra CB, Mantovani RM, Januário JN, & Silva IN. A three-year follow-up of congenital adrenal hyperplasia newborn screening. Jornal de Pediatria 2014 90 300–307. (https://doi.org/10.1016/j.jped.2013.09.007)
