|Values are valid only on day of printing.|
Published: July 2008Print Record of Viewing
Dr. Matern discusses the use of second-tier testing to reduce the number of false-positive results in newborn screening. This is accomplished using mass spectrometry technology. Also discussed: the unique partnership between Mayo Clinic and the Minnesota Department of Health is credited with reducing the number of false-positive test results, the cost of unnecessary follow-up testing, and parental anxiety for Minnesota parents.
Presenter: Dr. Dietrich Matern
Welcome to Mayo Medical Laboratories' Hot Topics. These presentations provide short discussions of current topics and may be helpful to you in your practice.
Our presenter for this program is Dr. Dietrich Matern, from the Division of Laboratory Genetics at Mayo Clinic. Dr. Matern discusses the use of second-tier testing to reduce the number of false-positive results in newborn screening. This is accomplished using mass spectrometry technology. Dr. Matern also discusses the unique partnership between Mayo Clinic and the Minnesota Department of Health. This partnership is credited with reducing the number of false-positive test results, the cost of unnecessary follow-up testing, and parental anxiety for Minnesota parents.
Let’s begin with a short introduction to newborn screening. Newborn screening was pioneered by Dr. Robert Guthrie in the early 1960s to identify patients with phenylketonuria, better known as PKU, which is an inborn error of amino acid metabolism that causes mental retardation unless treatment with a phenyl-alanine restricted diet is implemented before symptoms occur. Newborn screening was implemented as a public health program for PKU and subsequently other conditions for which early intervention can avoid mortality, morbidity, and disabilities.
Newborn screening comprises a program where, for the most part, blood is collected on a filter paper on the second day of life. The dried blood spots are sent to the screening laboratory where they are analyzed for the biomarkers of the disease included in the screening panel, and short- and long-term follow-up of patients found to be affected.
Since the 1960s, newborn screening evolved by the inclusion of additional conditions to screening panels following a model, where one condition is identified by one test that measures one analyte to which one cutoff is applied.
However, in the 1990s, tandem mass spectrometry, or MS/MS, was applied to newborn screening allowing for improved efficiency and effectiveness, but also representing a more complex technology that allows for multiplexing where in a single dried blood spot punch, multiple conditions can be identified in a single assay that measures more than 60 amino acids and acylcarnitines with analyte-specific cutoffs. In the US, newborn screening is administered under the jurisdiction of each state which determines which conditions each baby should be screened for. Because of this and the slow implementation of tandem mass spectrometry, a significant discrepancy developed in the number of conditions screened for in each state.
To achieve a more uniform screening menu, the American College of Medical Genetics (ACMG) in 2006, recommended that each baby regardless of birthplace, should be screened for at least 29 conditions. Twenty of these core conditions can be identified in a single dried blood spot punch in tandem mass spectrometry. These include inborn errors of amino acid metabolism, such as PKU, inborn errors of fatty acid oxidation and inborn errors of organic acid metabolism. Hemoglobin electrophoresis allows detection o f sickle cell disease and other hemoglobinopathies. Biotinidase deficiency, cystic fibrosis, congenital adrenal hyperplasia, congenital hypothyroidism, galactosemia, and hearing loss each are screened for using specific tests. Hearing loss is actually determined at the baby’s bedside by an audiology exam, so hearing loss is a condition that does not make use of the dried blood spot specimen.
In addition to these 29 core conditions, the ACMG identified 25 so-called secondary targets. These include conditions that are included they are part of the differential diagnosis of the core conditions. Many of these are also clinically significant but no efficacious treatment is yet available and others are of uncertain clinical significance. Twenty-two of the secondary targets are identified by tandem mass spectrometry.
As you now understand, the impact of tandem mass spectrometry on newborn screening was significant because it allows for the detection of more than 40 conditions in a single blood spot analysis looking at more than 100 analytes and analyte ratios. The generated metabolite profiles must be interpreted as a whole because an abnormal concentration of a single parameter in the profile is not unusual but most often not associated with disease.
Reporting an abnormal analyte without interpretation of the complete profile will cause a large number of false-positive results. As our colleagues at the Boston Children’s Hospital have shown in 2003, false-positive results cause significant stress and even serious parent-child dysfunction in some families. In addition, unnecessary follow-up wastes time and money of all involved and can desensitize physicians who might assume that most screening results are wrong anyway.
To illustrate the impact of newborn screening performance, let’s review best case-worst case scenarios as they are currently encountered in different parts of the US when it comes to newborn screening with tandem mass spectrometry. There are more than 4 million babies born in the US each year. In the best case, the false-positive rate of a screening program is 0.07% and the positive detection rate is one true positive in 2,249 live births. That would identify 1,780 affected babies per year and 2,800 false-positives. The positive predictive value of tandem mass spectrometry screening would be 39%, and based on a population of 100,000 babies per year, 1.4 babies would have to undergo unnecessary follow-ups per week.
In the worst case scenario, where the false-positive rate is 3.0% and the detection rate is one true positive in 15,000 live births, the positive predictive value would only be 0.2% and 58 babies as opposed to 1.4 in the best case scenario would have to undergo unnecessary follow-up testing each week.
Let’s take a look at the cost of false-positive newborn screening results using another real-life example. Newborn screening for congenital adrenal hyperplasia in Minnesota.
First a few words about newborn screening in Minnesota. Four years ago a public/private partnership was formed between the Minnesota Department of Health and the Mayo Clinic. This partnership combines the strength of the Minnesota Department of Health and the administration of a comprehensive newborn screening program of approximately 75,000 babies born each year and Mayo Clinic’s analytical and interpretive expertise in the performance of complex assays and technologies such as tandem mass spectrometry.
If we look at the conditions that are currently screened for in Minnesota and who tests for which condition, the Minnesota Department of Health screens for biotinidase deficiency, cystic fibrosis, congenital adrenal hyperplasia, congenital hypothyroidism, galactosemia, and hemoglobinopathies. Whereas at the Mayo Clinic, we screen for amino acidopathies, organoacidopathies, and fatty acid oxidation disorders using tandem mass spectrometry. What you also see is that at the Mayo Clinic, we perform the second tier tests for congenital adrenal hyperplasia and several of the tandem mass spectrometry tests.
Let’s get back to the example of congenital adrenal hyperplasia screening in Minnesota. If you look at the numbers for 2007, there are 710 false-positive results, or a false-positive rate of 1%. The cost of the clinical follow-up would have been about $600,000 and this is without a second-tier test.
What is second-tier testing? Second-tier testing is confirmatory testing which is performed when primary screening tests, either by tandem mass spectrometry or another method, yields an equivocal result. Basically, there is overlap of the normal control range and the disease range, so there is poor specificity. The confirmatory test measures the disease specific analyte or an analyte profile.
The normal second-tier test result then overrules primary screening results; the second-tier test utilizes the same specimen so there is no additional patient contact necessary.
The second-tier testing is performed for the following problem analytes. Again, these analytes have concentrations in the normal population that overlap with those in affected patients. There is 17-Hydroxy progesterone, the marker for congenital adrenal hyperplasia, that is particularly problematic among premature infants. There is methionine, which is a non-specific marker for homocystinuria and remethylation defects. There is C3 acylcarnitine, a propionylcarnitine, which is a nonspecific marker for propionic scidemia and methymalonic acidemias. And there are branch-chain amino acids, such as leucine, valine and isoleucine, which are nonspecific markers for maple syrup urine disease and mostly elevated in patients that receive total parental nutrition.
Why do we do second-tier testing? Because it increases the specificity of the testing and the positive predictive value and reduces the false-positive rate, parental anxiety, and follow-up efforts and costs.
Coming back to our example for screening for congenital adrenal hyperplasia in Minnesota, if we look at the 2007 data and look at it with the second-tier test that was actually implemented, with the second-tier test, the false-positive results were not 710, but only 41. The false-positive rate was only 0.06% as opposed to almost 1%. The cost of clinical follow-up was reduced to $38,000; however, one has to pay for the second-tier test. Overall, the total follow-up cost is not $600,000, but $62,000 with a savings of 89%.
What is the impact of second-tier testing on the performance of newborn screening by tandem mass spectrometry at the Mayo Clinic? We have a very low false-positive rate of 0.07%, a high-positive predictive value of 39%, a positive detection rate, where we find one affected case in 2,249 babies screened, there are very few unnecessary follow-up evaluations, basically 1.4 per week in 100,000 births. This, as you will recognize, is the "best case scenario."