In neonatal screening, the cut-off is a key value or concentration: any result above this level will be rechecked. The definition of the cutoff is influenced by many factors and should ideally be specific to each laboratory. In this article, ZenTech shares with you some tips to guide your choice.
In medicine, the cut-off can be generalized as a threshold that discriminates sick from healthy patients. For newborn screening, it represents a good compromise: avoid false-negatives, limit false-positives.
Manufacturers recommend a cutoff value in their instructions for use but it may not be adapted to your practices... Weight, gestational age, ethnicity, sampling method... The factors impacting the cut-off may be numerous.
This article aims to demystify this threshold and share the feedback from a laboratory in Belgium, the Human Genetic Biochemistry Laboratory of the University Hospital of Liège.
The cut-off: a single role for different approaches
The motto in newborn screening is the same for all diseases: identify sick babies, avoid the recall of healthy children. A threshold of disease risk, the cutoff, allows to differentiate them. Two main approaches exist to define it:
- Based on scientific literature (regional or global)
- Based on statistics such as a percentile or a multiple of the median
These methods can then be adapted – definition of cut-offs by age group, multivariate consideration of a parameter (including gestational age), etc.
In all cases, the Association of Public Health Laboratories (APHL) recommends that all laboratories determine a preliminary cut-off and validate it. This guidance is published and available online [1].
Which factors influence the cut off?
These factors are specific to each disease. A literature search may be useful. Here is a non-exhaustive list illustrated by examples from the literature.
Newborn screening technique
The cutoff for newborn screening for phenylketonuria (PKU) may differ depending on the method used. The Latin American consensus recommends the following values [2]:
- 0 to 2.5 mg/dL for enzymatic tests in colorimetry or fluorescence
- 0 to 1.5 mg/dL for tandem mass spectrometry
- 0 mg/dL for bacterial inhibition tests, the least sensitive technique
Gestational age, birth weight and age at sampling
17-OHP concentrations are higher in premature and low-weight infants, and decrease with increasing age. Establishing multiple cut-offs, sorted by weight, age, and time from birth to collection, improves the efficiency of screening for congenital adrenal hyperplasia (fewer false positives) [3].
Prematurity may also increase the risk of false-negative/positive tests for congenital hypothyroidism (TSH, T4), certain metabolic diseases including phenylketonuria (PKU) and maple syrup disease (MSUD), or cystic fibrosis (PAP) [4 - 6]. This factor should therefore be considered in relation to the cutoff.
Sample type
In some countries, congenital hypothyroidism is screened according to two cut-offs, specific to each type of blood sample [7]:
- 20 mIU/L of TSH for heel blood
- 25 mIU/L of TSH for cord blood
Ethnicity
A 2020 study found an ethnic/race difference for 79.1% of the Recommended Uniform Screening Panel (RUSP) biomarkers [8]. This confirms the need for laboratory-specific cut-offs.
False positives/negatives, what impact on the cut-off?
If the statistical approach is used, thousands of samples are considered (two to three thousand at least). The presence of a false positive or negative will therefore have little influence on the cutoff value. However, if several false negatives are reported, a redefinition of the cut-off must be considered.
A concrete laboratory experience
In the Human Genetic Biochemistry Laboratory (University Hospital of Liège), the prevailing approach is that of percentiles: a minimum of two thousand patients are considered to calculate the cut offs. They are reviewed annually.
In some cases, this statistical calculation includes covariates. For example, adrenal hyperplasia screening (17-OHP) includes gestational age or birth weight.
The definition of two cutoffs is also practiced. The example is illustrated by the TSH measurement. The 99.5th percentile is the “positivity threshold”; the sample will be retested in duplicate. The 99.99th percentile is the “urgent recall threshold”: the paediatrician in charge of the patient is immediately called back.
Want some support for your own cut off determination?
Want to learn more about the Recommended Uniform Screening Panel
Sources:
[1] APHL, Overview of Cutoff Determinations and Risk Assessment Methods used in Dried Blood Spot Newborn Screening – Role of Cutoffs and Other Methods of Data Analysis https://www.aphl.org/programs/newborn_screening/Documents/Overview%20on%20Cutoff%20Determinations%20and%20RIsk%20Assessment%20Methods_final.pdf
[2] Borrajo, G.J.C. (2016) ‘Newborn Screening for Phenylketonuria: Latin American Consensus Guidelines’, Journal of Inborn Errors of Metabolism and Screening, 4.
[3] Olgemöller, B. et al. (2003) ‘Screening for Congenital Adrenal Hyperplasia: Adjustment of 17-Hydroxyprogesterone Cut-off Values to Both Age and Birth Weight Markedly Improves the Predictive value’, The Journal of Clinical Endocrinology & Metabolism, 88(12), 5790–5794.
[4] Wilson, K. et al. (2013) ‘Metabolomics of prematurity: analysis of patterns of amino acids, enzymes, and endocrine markets by categories of gestational age’, Pediatric Research, 75, 367–373.
[5] Vernooij-van Langen, A. M. M. et al. (2012) ‘The influence of sex, gestational age, birth weight, blood transfusion, and timing of the heel prick on the pancreatitis-associated protein concentration in newborn screening for cystic fibrosis’, Journal of Inherited Metabolic Disease, 36(1), 147–154.
[6] Hashemipour, M. et al. (2018) ‘Screening of congenital hypothyroidism in preterm, low birth weight and very low birth weight neonates: A systematic review’. Pediatric Neonatology, 59, 3–14 (2018).
[7] Paul, P. G. et al. (2021) ‘Optimizing Cord Blood Thyroid Stimulating Hormone Cutoff for Screening of Congenital Hypothyroidism—Experience from Screening 164,000 Newborns in a Tertiary Hospital in India’. Indian Journal of Endocrinology and Metabolism, 25(4),348–353.
[8] Peng, G. et al. (2020) ‘Ethnic variability in newborn metabolic screening markers associated with false-positive outcomes’. Journal of Inherited Metabolic Disease, 43(5), 934–943.
Source : Madeleine Boulanger, PhD
