Professor of Pediatrics, New York University School of Medicine, New York, New York 10016; and Chief of Pediatric Endocrinology, Schneider Childrens Hospital, New Hyde Park, New York 11040
Address all correspondence and requests for reprints to: Dr. P. W. Speiser, Chief, Division Pediatric Endocrinology, Schneider Childrens Hospital, 269-01 76th Avenue, New Hyde Park, New York 11040. E-mail: pspeiser{at}lij.edu.
In this issue of JCEM, Minutti et al. (1) describe improved positive predictive value of newborn screening for congenital adrenal hyperplasia (CAH) when steroid profiling by liquid chromatograph-tandem mass spectrometry is used as a second-tier test. This disease, present in 1:15,000 to 1:16,000 births, is screened for in most states in the United States and in a number of other countries (2). Considerations in selecting diseases to be screened include population frequency, morbidity and mortality resulting from delayed diagnosis, precision of detection, and efficacy of treatment. Economic costs also weigh into the decision to screen; however, these vary widely depending on the screening protocol, the chosen assay, the selected assay cut-offs, and various efficiencies of scale. Cost-benefit analysis done in the last decade for the Texas program with sample acquisition at birth and at 2 wk of life showed that CAH screening is relatively costly compared with other disease screening programs (3, 4). Despite these data, neonatal CAH screening has been expanding. The rationale includes the fact that at least 50% of cases have not been detected clinically at the time when screening results are reported (5). Earlier detection leads to earlier treatment and avoids severe salt-wasting adrenal crises, thus obviating acute and intensive hospital care associated with even higher costs. Beyond these calculated costs, there are also human costs in failing to identify and treat CAH at an early stage. Until the advent of newborn screening, case reports showed female preponderance, suggesting an as-yet-unknown number of early deaths in affected male infants (6). Infants who survived salt-wasting episodes often suffered permanent neurological damage. Additionally, erroneous gender assignment with attendant long-term psychosexual trauma can be largely avoided through screening. Conversely, mislabeling a healthy child as having a serious inherited disease also carries adverse consequences. Thus, diagnostic precision is of paramount importance.
The most widely used and least costly technique is RIA of blood spot 17-hydroxyprogesterone (17OHP); also available are enzyme-linked immunoassay and fluorescent immunoassay. Among the problems encountered with immunoassays are poor antibody specificity in combination with abundant cross-reacting hormones in newborn blood. Lack of assay standardization and noncomparability of filter paper and serum measurements further hamper interpretation of CAH screening results. Furthermore, stress due to prematurity or critical illness generally increases adrenal cortisol and 17OHP secretion in tandem. If only 17OHP is measured, infants may be incorrectly diagnosed as having CAH. Attempts to improve the predictive value of screening programs have included measuring the ratio of 17OHP to cortisol by immunoassay as a second-tier test (7). Stratifying cut-off values by gestational age (8) and/or birthweight (9) is widely used, but none of these strategies has entirely solved the main problem of large numbers of false-positive test results. It is also distressing that immunoassay screening has, in rare cases, failed to identify newborns affected with classic CAH.
In view of the fact that a high 17OHP measured by immunoassay does not necessarily predict disease, it is crucial to seek alternative or adjunctive screening methods. Genotyping has been proposed as a second-tier test (10, 11). Employing CYP21A2-specific PCR amplification, a panel of probes or primers for 910 mutant alleles can identify about 9095% of affected haplotypes. Unfortunately, however, this means that without sequencing, about 510% of affected alleles could go undetected. There are presently at least 75 known mutant alleles for CYP21A2 (i.e. the active steroid 21-hydroxylase-encoding gene), and the list continues to grow (the Human Gene Mutation Database, http://uwcmml1s.uwcm.ac.uk/uwcm/mg/search/120603.html). Moreover, genotyping, like immunoassay, is not without its own accuracy pitfalls (12). In some PCR-based reports, 1520% of patients had only one allele characterized (13, 14). The highly complex nature of the chromosome 6p21.3 region with gene and pseudogene encoding active and inactive 21-hydroxylase enzymes, respectively, sometimes in multiple copies, can make genotype interpretation quite difficult. Thus, although genotyping is feasible on a limited scale in commercial laboratories and in research settings, it does not appear ideally suitable at this time for mass screening.
Would tandem mass spectrometry be an appropriate choice for a second-tier CAH screening test for perhaps the top 35% of the assay results? It is encouraging that measuring 17OHP, androstenedione, and cortisol by liquid chromatography-tandem mass spectrometry in more than 1200 specimens increased the positive predictive value of CAH screening 9-fold over that for 17OHP immunoassay alone, estimated now at about 0.5%. Minutti et al. (1) set the cut-off point for their assay so as to eliminate all false-negative CAH screening results, maximizing sensitivity over specificity. The results of this study show a positive predictive value of 4.7%, on par with that for other disease screening tests (15).
In addition to diagnosing CAH caused by 21-hydroxylase deficiency, it is at least theoretically possible with tandem mass spectrometry to detect other rarer forms of CAH, such as 11-hydroxylase deficiency. This study measured androstenedione and cortisol in addition to 17OHP. Although these hormones are not specifically diagnostic for 11-hydroxylase deficiency, perturbations would provide preliminary information suggesting the need for further evaluation. Similarly, 3-ß-hydroxysteroid dehydrogenase deficiency could be identified if -5 steroids such as 17-hydroxypregnenolone and dehydroepiandrosterone were measured. One also wonders whether the method of steroid profiling will more efficiently distinguish milder forms of CAH. This is important because nonclassic CAH may not require lifelong medical treatment. Beyond CAH diagnosis, tandem mass spectrometry can be used to simultaneously detect more than 20 inborn errors of metabolism apart from CAH. At least one commercial laboratory in the United States is presently offering this test on a fee-for-service basis to parents of newborns (16). Another advantage of tandem mass spectrometry is its rapid turn-around time, comparable or quicker than current methods at 24 h after sample receipt. Rapid and accurate CAH diagnosis would relieve much of the frustration experienced by health professionals and parents who would be spared dealing with equivocal initial test results.
On the negative side, the cost for this new test is rather high compared with immunoassay, but less than genotyping. In its present configuration, the throughput of a single tandem mass spectrometry machine would be only 25,000 samples annually. One must also realize that this methodology has not yet been tested on a large scale in real-time population screening. Minutti et al. (1) estimate that the throughput could be at least doubled with novel technology. With increased efficiency, costs could be reduced. The authors also make the cogent argument that even if their method is expensive at $26 per assay, there is a high financial and emotional burden associated with all the false-positive results we currently encounter, estimated at more than $800 per infant. Ideally, fewer than 1% of screened subjects should be recalled for subsequent testing (2). Among the infants referred for abnormal newborn screens in 2003 in New York State, fewer than 3% were true positives affected with CAH (Pass, K., personal communication). Even if it eliminates only 90% of false-positive test results and costs $26 per sample, a more accurate second-tier screen would potentially save a great deal in unnecessary healthcare expenditures. Regionalization of this relatively costly second-tier test might make sense until more efficient detection systems are designed.
Even with all of its inefficiencies, CAH newborn screening has been practiced widely for more than three decades and is here to stay. The benefits of screening have been shown to reduce time to diagnosis of affected infants, consequently reducing newborn morbidity and mortality. These are clear aims of performing the screening. We must now strive to better separate the false positives and avoid any false negatives.
Footnotes
Abbreviations: CAH, Congenital adrenal hyperplasia; 17OHP, 17-hydroxyprogesterone.
Received May 24, 2004.
Accepted June 7, 2004.
References