Cystic Fibrosis Screen-Positive Neonates with One Pathogenic
Variant Still Warrant Sweat Testing
To the Editor,
Many challenges remain as states seek to improve newborn screening (NBS)
for cystic fibrosis (CF), especially to achieve the goals of equity and
timeliness. Controversies abound with regard to procedures (NBS
algorithms such as the size of panels to identify variants in the cystic
fibrosis transmembrane conductance regulator or CFTR gene),
practices (operations), and policies (e.g., selecting babies for
follow-up diagnostic evaluations). Our recent article1entitled “Refinement of newborn screening for cystic fibrosis with next
generation sequencing” described the evolution of the NBS algorithm in
Wisconsin to maximize detection of infants with CF through an expandedCFTR panel achieved with next generation screening methodology
and updateable reporting software. In brief, we utilized the CFTR2
database to screen for CF-causing variants in infants who had an
elevated immunoreactive trypsinogen level (with a floating cut-off at
the 96th percentile). The number of CF-causing
variants is increasing over time as we learn more about the disease
liability of different CFTR variants. The CFTR2 project is an
international effort to characterize variants as CF-causing
(pathogenic), non-CF-causing, variants of varying clinical consequences
(VVCCs) and variants of unknown significance. Utilizing the CFTR2
database was a cornerstone of our refinement of NBS.
In phase 2 of our NBS studies, we utilized variant lists from CFTR2 that
were published in 2013 and 2016. In phase 3, we utilized four variant
lists published in CFTR2, with the most recent being
CFTR2_24September2021.xlsx (variant lists from
https://www.cftr2.org/mutations_history). This variant list has
382 variants that are CF-causing. When we published our
article,1 there were a total of four infants diagnosed
with CF in phase 2 plus phase 3 who had novel variants not in CFTR2.
(The original published article stated that there were six infants who
had novel variants that were not in CFTR2. Due to typographical errors
in the legacy name of two variants, the correct number of infants with
novel variants is four.)
Since then, there are two more recent variant lists in CFTR2. The
variant list CFTR2_29April2022.xlsx has 401 CF-causing variants and
CFTR2_7April2023.xlsx includes 719. If we had used this most recent
variant list, we would still have identified the same four infants with
CF who each have one detected variant and a novel variant not in CFTR2.
To avoid or at least minimize false negatives and strive for equity, we
have always recommended a sweat test for all infants born in Wisconsin
who have one pathogenic variant, so these infants were diagnosed at an
early age.
McGarry et al2 utilized the CF Foundation Patient
Registry to analyze the detection of 1 or 2 variants using variant
panels of different size (ranging from F508del only to the 401
pathogenic variants in CFTR2 from CFTR2_29April2022.xlsx). When thisCFTR 2 database was utilized, detection of one variant was lowest
in people who were Asian, non-Hispanic (77.4%), and detection of two
variants was particularly low in Asian, non-Hispanic (53.9%), Hispanic
(73.7%) and mixed races, non-Hispanic (76.0%). With CFTR2 now being
updated to 719 pathogenic variants (CFTR2_7April2023.xlsx), we asked
Alex Ebert and Runyu Wu to reanalyze the data in the McGarry article
with this most recent updated variant list. Tables 1 and 2 include the
last two lines from the data in the McGarry et al2publication (CFTR2_22April2022.xlsx which has 401 CF-causing variants),
and an additional two lines in each table utilizing the 719 CF-causing
variants in CFTR2_7April2023.xlsx alone, and 719 CF-causing variants
plus VVCCs. Of note is that there is only a minimal increase in the
detection of 1 or 2 variants in all races/ethnicities with this most
recent larger variant panel. Specifically, detection of one variant in
Asian, non-Hispanic people with CF only increased from 77.4% to 80.6%,
and detection of two variants also increased minimally (Asian,
non-Hispanic: 53.9% to 56.2%; Hispanic: 73.7% to 75.9%; and mixed
races, non-Hispanic: 76% to 78.4%).
We concluded our article1 with a policy assertion
“that all infants with one variant should have sweat testing
performed.” Even though CFTR2 has added 318 more pathogenic variants,
if we had not performed a sweat test in all 1-variant infants, we would
have missed diagnosing CF in 5% of the infants in Wisconsin in phase 2
plus phase 3 of our improved NBS program. Although this proportion of
missed cases is better than the 11% during 2010-2018
nationally,2 preventable false negative results should
obviously be avoided. These data are even more compelling in looking at
the updated analysis from the McGarry et al2publication, particularly in people of color or minority groups.
Requiring that 2-variants be present to proceed to sweat chloride
determination would result in not establishing an early diagnosis
in 43.8% of Asian, non-Hispanic patients, 37.3% of African
American/Black non-Hispanic patients and 14.1% of Hispanic patients.
But continuing to expand CFTR panels is important to strive for
equity. Thus, Wisconsin’s panel now includes 689 pathogenic variants
from the CFTR2 update.
Although there are several arguments of not performing sweat testing in
infants with one variant and arguments against increasing the size ofCFTR variant panels, we do not believe that those arguments are
compelling, particularly in the emerging era of next generation
sequencing. A list of misperceptions and counterarguments related to
policies follow:
- Sweat test labs could be overwhelmed with the number of sweat
tests . Prior to CF NBS, providers typically ordered sweat tests for
children with gastrointestinal or respiratory symptoms. With routine
NBS, providers now order many fewer sweat tests. We have
documented3 that the number of sweat tests decreased
in Wisconsin with the introduction of CF NBS from 1670 to 804
(including 134 follow-up tests from screening). Thus, in the NBS era,
sweat test labs are performing many fewer sweat tests compared to
pre-newborn screening but need to perform enough annually to maintain
proficiency.
- Sweat testing infants will result in unacceptably elevated
quantity not sufficient (QNS) sweat collections. It is true that
there are challenges in obtaining sufficient sweat from young
infants.4 Factors such as low birth weight and
gestational age can have impacts on obtaining an adequate quantity of
sweat. However, the CF Foundation has committed to improved oversight
and performance of sweat testing in CF Centers. The site visit process
includes an in-depth examination of sweat collection and analysis,
with scrutiny of the protocol by the Sweat Test Advisory Committee.
Additionally, individual centers can engage in quality improvement
activities to decrease their sweat test QNS rate.4Even if there is a QNS sweat collection, genetic counseling delivered
at the time of the sweat test is a benefit that most parents
appreciate.
- The Illumina MiSeqDx next generation sequencing assay is only
approved for 139 variants. The FDA approval for the 139 variant
method applies only to blood specimens. We performed a full validation
of using this technology on dried blood spots.5 The
variant calling file can be updated through the reporting software as
needed to add more variants when CFTR2 is updated. Variants that are
not the FDA approved 139 variants can be verified via Sanger
sequencing. We include a statement on our screening results reports to
providers as follows: “These tests have not been cleared or approved
by the FDA. The laboratory is regulated under CLIA as qualified to
perform high-complexity testing. These tests are used for clinical
purposes and should not be regarded as investigational or for
research.”
- The next generation sequencing assay is too time consuming. We
demonstrated that our timeliness benchmarks (6-9 days) were being met
with next generation sequencing.1
- It is too costly to increase the number of CFTR variants. The
cost is no different for 139, 401 or more variants. The variant
calling file can be readily updated with each expansion of CFTR2
variant as we have done. And our current costs of analysis for 689
variants are the same as the cost of analysis for 401 variants.
There is documentation showing that historically marginalized groups
have delayed entry into the health care system and worse nutritional
outcomes.6 As it stands now, early diagnosis of CF via
newborn screening depends on where and when a baby is born. An
African/American or Hispanic or Asian infant born in Wisconsin has the
same likelihood to be diagnosed early via newborn screening. If that
same baby was born in states that perform sweat testing only if 2
variants are identified, a practice introduced in California, there is a
much higher chance that the 1-variant infant will have a delayed
diagnosis. Thus, those protocols are not equitable across all races and
ethnicities. We believe that the data in this letter clearly demonstrate
that even with CFTR2 expansion to 719 variants, striving to achieve
equity of early diagnosis of CF via screening requires states to perform
a sweat test in all infants with a high IRT level and one identifiedCFTR variant. This recommended policy can be debated but sweat
testing overload should not be the argued as the barrier and CF
specialists need to recognize that CFTR2 may never include all of the
very rare, “private” pathogenic variants nor will next generation
sequencing cover the structural variants such as deletions and
duplications.
Acknowledgements. We thank Runyu Wu and Alex Ebert of the CF
Foundation for their expert analysis of the updated CFTR2 panel and
their continued outstanding leadership of the Patient Registry.