Ultrasonographic Fetal Nuchal Translucency Measurements and Cytogenetic Outcomes

Key Points Question Is there an association between nuchal translucency measurements less than 3.5 mm and chromosomal anomalies? Findings In this population-based cohort study including 414 268 singleton pregnancies in Ontario, Canada, a significantly increased risk of chromosomal anomalies was associated with each increasing level of nuchal translucency measurement, compared with a reference group of pregnancies with nuchal translucencies less than 2.0 mm. Meaning The findings of this cohort study suggest that pregnancies with nuchal translucency measurements greater than 2.0 mm are at increased risk of chromosomal anomalies, indicating that the widely used threshold of 3.5 mm may need to be reexamined.


Introduction
Since the 1990s, ultrasonographic measurement of nuchal translucency-a collection of fluid behind the fetal neck 1 -has been included in prenatal genetic screening offered to pregnant individuals to identify trisomies 21 and 18, typically as a part of multiple-marker screening. 2,3In addition to these aneuploidies, increased fetal nuchal translucency is associated with other chromosomal anomalies, single gene conditions, and structural defects. 4[7][8][9] These follow-up investigations may include prenatal cell-free DNA (cfDNA) screening and/or confirmatory diagnostic testing through cytogenetic analysis.
0][21][22][23] A small number of these studies 19,20,22 have provided preliminary evidence that pregnancies with nuchal translucency measurements that are elevated but still lower than 3.5 mm could also be at increased risk of clinically significant chromosomal anomalies.These findings require confirmation using robust methodological approaches in large, unselected samples with comprehensive follow-up to adequately assess the risk of chromosomal anomalies across the entire range of nuchal translucency measurements.In this study, we aimed to evaluate the association between all levels of nuchal translucency measurements and cytogenetic outcomes among pregnancies in Ontario, Canada, identified through a population-based provincial registry.

Methods
This study received approval from the research ethics boards of the Children's Hospital of Eastern Ontario and the University of Ottawa.The requirement of informed consent was waived owing to the use of deidentified patient data.All cell counts of less than 6 were suppressed to comply with the privacy requirements of the registry.The study followed the Reporting of Studies Conducted Using Observational Routinely-Collected Health Data (RECORD) reporting guideline.

Data Source
Better Outcomes Registry & Network (BORN) Ontario is a prescribed perinatal registry that collects data directly from all multiple-marker screening, cfDNA screening, and cytogenetics laboratories in Ontario. 24Pregnancy and birth outcomes are also captured, including information on the clinical birth examination and linkage to discharge data from all hospitals through the discharge abstract database of the Canadian Institute for Health Information (CIHI).This enabled comprehensive ascertainment of pregnancy outcomes.Additional details on data sources are provided in eTable 1 in Supplement 1.

JAMA Network Open | Genetics and Genomics
Ultrasonographic Fetal Nuchal Translucency Measurements and Cytogenetic Outcomes

Setting and Study Population
Ontario offers a publicly funded screening program in which all pregnant individuals have access to multiple-marker screening in the first trimester, most often including a nuchal translucency measurement. 25Cell-free DNA screening or cytogenetic testing is offered if the screen result is positive or as a first-tier screen if specific eligibility criteria are met. 26Individuals may also self-pay for cfDNA screening. 26chal translucencies are measured at crown-rump lengths of 45 to 84 mm by sonographers registered in Ontario's Nuchal Translucency Quality Assurance program.27 This study included all singleton pregnancies in Ontario with a valid multiple-marker screening test including a nuchal translucency and with an estimated date of delivery (EDD) from September 1, 2016, to March 31,   2021 (eFigure 1 in Supplement 1).While ethnicity is recorded in the BORN database, we did not incorporate this information into our analysis as it was not expected to influence the associations studied.

Study Exposure
Nuchal translucency measurements for all pregnancies were identified from multiple-marker screening results.The reference group was defined as pregnancies with a measurement less than 2.0 mm, compared with pregnancies with the following categories of nuchal translucency measurements: 2.0 to less than 2.5 mm, 2.5 to less than 3.0 mm, 3.0 to less than 3.5 mm, 3.5 to less than 5.0 mm, 5.0 to less than 6.5 mm, and 6.5 mm or greater.

Study Outcome
Pregnancies with chromosomal anomalies were identified through cytogenetic testing results submitted by all Ontario cytogenetic laboratories to the BORN registry.The primary outcome was defined as any chromosomal anomaly identified on cytogenetic testing, during pregnancy or postnatally, including microarray analysis.As secondary outcomes, we stratified chromosomal anomalies by whether or not the condition is routinely tested through cfDNA screening in Ontario (trisomies 21, 18, and 13 and sex chromosome aneuploidies).
Because only a small number of pregnancies have cytogenetic investigations, we supplemented our outcome data with information from other sources (pregnancies with and without cytogenetic testing are described in eTable 12 in Supplement 1).Specifically, to identify pregnancies without chromosomal anomalies, we first used cytogenetic testing results if performed.If no cytogenetic testing results were available, we used cfDNA screening results, if performed; although a low-risk cfDNA screening result cannot be used clinically to exclude these conditions (trisomies 21, 18, and 13 and sex chromosome aneuploidies), for the purposes of this research, it was considered a reasonable proxy given its negative predictive value of greater than 99.9%. 28Finally, for pregnancies with no cytogenetic testing and no cfDNA screening, we used results from the clinical examination at birth to exclude conditions typically clinically diagnosable at birth, relying on both BORN and CIHI data (eTable 2 in Supplement 1).

Statistical Analysis
Data were analyzed from March 17 to August 14, 2023.The study population was described using means (SDs) for continuous variables and frequencies and proportions for categorical variables.We used multivariable modified Poisson regression models with robust variance estimation and adjustment for gestational age at screening to compare the risk of chromosomal anomalies across pregnancies with varying categories of nuchal translucency measurements with the reference category (<2.0 mm). 29This model also allowed us to account for clustering for individuals with more than 1 pregnancy within the study period. 30Gestational age at screening was identified a priori as a potential confounder through directed acyclic graphs, as nuchal translucency measurements are on a continuum and will change with gestational age (eFigure 2 in Supplement 1).A post hoc analysis with additional adjustment for age of the pregnant individual was also performed.Adjusted risk ratios (ARRs) and risk differences (ARDs) were reported with 95% CIs.
We conducted the following sensitivity analyses to evaluate the potential impact of incomplete or inaccurate ascertainment of the exposure and outcome and to address losses to follow-up.All analyses were performed using SAS, version 9.4 (SAS Institute Inc), and 2-tailed P < .05 was considered statistically significant.

Exposure Measurement Source
When a very high nuchal translucency measurement or cystic hygroma is identified, some pregnant individuals may not complete the multiple-marker screening process and, thus, may not be ascertained by the laboratories.We therefore performed a sensitivity analysis identifying pregnancies with increased nuchal translucency measurements through other sources available within the registry, including data obtained from consultations with genetics or maternal fetal medicine clinics and from documented clinical indications for testing obtained from cytogenetic laboratories (eTable 3 in Supplement 1).

Exposure Definition
Some studies rely on percentiles of the nuchal translucency measurement rather than absolute values.Therefore, we categorized nuchal translucency measurements as less than 90th, 90th to less than 95th, 95th to less than 99th, and 99th percentile or greater 10,21 in an additional sensitivity analysis (eTable 4 in Supplement 1).

Losses to Follow-Up
For some pregnancies, no outcome was recorded.These may reflect pregnancy losses or terminations in the absence of follow-up cfDNA screening or cytogenetic testing or pregnant individuals who had multiple-marker screening in Ontario but subsequently received care outside the province.Because of the unclear nature of the outcome for these pregnancies, we performed a sensitivity analysis in which we randomly classified the losses to follow-up to having twice the prevalence of chromosomal anomalies compared with pregnancies in the same category of nuchal translucency measurement for which an outcome was recorded or half the prevalence (eTable 5 in Supplement 1).A further sensitivity analysis included all pregnancies with varying assumptions of risk of chromosomal anomalies for those lost to follow-up based on the pregnancy outcome (eTable 6 in Supplement 1).

Outcome
Complete cytogenetic data for microarray testing was only available since January 2018.Therefore, we conducted an analysis restricted to pregnancies with EDD from September 1, 2018, to March 31, 2021 (eTable 7 in Supplement 1).

Time Period
An additional sensitivity analysis excluded pregnancies with an EDD from April 1, 2020, to March 31,   2021.This exclusion accounted for potential effects of the COVID-19 pandemic on prenatal care practices (eTable 8 in Supplement 1).

JAMA Network Open | Genetics and Genomics
mm or greater (Table 1).The mean (SD) maternal age at EDD increased across nuchal translucency categories, from 31.5 (4.7) years for pregnant individuals with nuchal translucency measurements less than 2.0 mm to 33.3 (5.4) years for those with measurements 6.5 mm or greater.We excluded 225 264 pregnancies without a valid multiple-marker screening test including a nuchal translucency measurement, 158 with no screening result report issued, and 3456 where the measurement was performed outside the gestational age range corresponding to the crown-rump length of 45 to 84 mm. Figure 1 and Figure 2 describe the uptake of follow-up investigations (cfDNA screening, cytogenetic testing) and pregnancy outcomes among pregnant individuals with nuchal translucency measurements less than 3.5 mm and 3.5 mm or greater, respectively.Of pregnancies with a nuchal translucency measurement 3.5 mm or greater, 1654 (83.6%) underwent follow-up investigations prenatally, compared with 44 849 (10.9%) of pregnancies with a measurement less than 3.5 mm.
Among those with nuchal translucencies 3.5 mm or greater, 414 (20.9%) had both cfDNA screening and prenatal diagnosis with cytogenetic testing, compared with 1747 (0.4%) among pregnancies with a measurement less than 3.5 mm.Of chromosomal anomalies identified in pregnancies with measurements of 3.5 mm or greater, 496 (72.5%) were identified prenatally, compared with 947 (35.3%) for measurements less than 3.5 mm (eTable 9 in Supplement 1).
Among pregnancies with cytogenetic testing results (n = 15 755), the proportion with chromosomal anomalies increased across nuchal translucency measurement categories, from 1913 pregnancies (16.6%) with nuchal translucency less than 2.0 mm to 256 pregnancies (70.1%) with a measurement 6.5 mm or greater (Table 2).To ascertain potential chromosomal anomalies in pregnancies that did not receive cytogenetic testing, 38 041 (98.2%) received a low-risk cfDNA screening result (Table 2).Next, among pregnancies with no cfDNA screening performed, we identified 331 638 documented live births with no notable clinical findings reported (Table 2).These results were used to estimate the proportion of pregnancies in the study with chromosomal anomalies, increasing from 1913 (0.5%) in pregnancies with nuchal translucency less than 2.0 mm to 256 (52.6%) in pregnancies with nuchal translucency 6.5 mm or greater (Table 2).The proportion of pregnancies excluded from the primary analysis due to an unknown outcome was associated with nuchal translucency category, from 9281 (2.6%) in the group with measurements less than 2.0 mm to 48 (9.9%) in the group with measurements 6.5 mm or greater (Table 2).We therefore conducted a sensitivity analysis randomly classifying the pregnancies lost to follow-up to have twice the prevalence of chromosomal anomalies compared with those for which an outcome was recorded within the given nuchal translucency measurement category, and results were mildly accentuated.The analysis assuming half the prevalence showed mildly attenuated results (eTable 5 in Supplement 1).translucency category of 3.0 to less than 3.5 mm relative to less than 2.0 mm, the ARD was 1.40% (95% CI, 0.77%-2.04%)and the ARR was 4.97 (95% CI, 3.45-7.17).
All additional sensitivity analyses showed comparable findings to the primary analysis, including analyses incorporating pregnancies with nuchal translucency measurements identified by sources other than multiple-marker screening (eTable 3 in Supplement 1); analyses restricted to a timeline with complete capture of microarray data, with an EDD from September Finally, for the pregnancies included in our study, the 99th percentile for nuchal translucency measurement was 2.8 mm, the 95th percentile was 2.2 mm, and the 90th percentile was 2.0 mm.
When defining the exposure by nuchal translucency percentile, pregnancies with a measurement greater than the 99th percentile had a risk of any chromosomal anomaly 34.9 times greater than pregnancies with a measurement less than the 90th percentile (eTable 4 in Supplement 1).

Discussion
This population-based cohort study leveraged linked multiple-marker screening, cytogenetic testing, cfDNA screening, and birth registry data capturing pregnancy outcomes and findings from the clinical examination at birth to quantify the association of increased risk of chromosomal anomalies with increasing nuchal translucency measurement.We found a strongly increased risk of  chromosomal anomalies with increased nuchal translucency relative to values less than 2.0 mm, particularly for measurements of 3.0 mm or higher.The findings were consistent through several sensitivity analyses.
To our knowledge, this is the first population-based study assessing the risk of chromosomal anomalies across all levels of nuchal translucency measurements and incorporating information from antenatal as well as postnatal cytogenetic testing, cfDNA screening, pregnancy outcomes, and newborn clinical examinations.21,31 The concern is that pregnant individuals who opted for cytogenetic testing in these studies may have had a higher risk of chromosomal anomalies, as additional findings beyond nuchal translucency measurement may have led them to have prenatal diagnostic testing; excluding pregnancies at lower risk for which outcomes are not available would, therefore, tend to overestimate the risk.This is particularly important when investigating nuchal translucency measurement values that would not independently trigger an offer of a follow-up investigation.As expected, we observed a substantial difference in the proportion of pregnant individuals who elected to have follow-up investigations prenatally in the group with nuchal translucency measurements less than 3.5 mm (10.9%) compared 2.0 to <2.5 0.55 (0.45-0.66)Our finding that nuchal translucency values below 3.5 mm, particularly those from 3.0 to less than 3.5 mm, are associated with chromosomal anomalies relative to values less than 2.0 mm, has important implications for prenatal genetic screening and counselling.Beyond the common aneuploidies (trisomies 21, 13, and 18 and sex chromosome aneuploidies) routinely identified by cfDNA screening in Ontario and included in many screening programs internationally, 32 we report that increased nuchal translucency measurements also yield an increased risk for other chromosome anomalies, although weaker.This has important policy implications given that in some jurisdictions, prenatal cfDNA screening is offered following the identification of an increased nuchal translucency measurement, whereas in others, cfDNA screening is part of first-tier prenatal screening, with or without accompanying nuchal translucency measurement.Screening programs should consider the value of nuchal translucency measurements when making decisions about whether to replace such measurements with cfDNA screening alone and may wish to reexamine the threshold of nuchal translucency at which diagnostic investigations are offered. 10,32Further research, including economic modeling to estimate benefits and costs, is needed to inform the best options for policy and practice. 10,32e results of this study also have implications for the quality assurance of nuchal translucency: the 99th percentile for nuchal translucency measurement was well below the expected 3.5 mm, implying that using a cutoff of 3.5 mm to offer follow-up investigations may not be sufficient.While the increased risk of chromosomal anomalies in pregnancies with nuchal translucencies less than 3.5 mm could be partly due to chronic undermeasurement, it is unlikely that this would be the only factor.Indeed, some jurisdictions have adopted thresholds to offer follow-up investigations much lower than the 99th percentile (eg, 95th percentile in Finland, the Netherlands, Germany, and Switzerland). 33[36][37][38][39]

Limitations
An inherent limitation of this study is that cytogenetic outcomes were not available for all pregnancies, as cytogenetic testing is only offered under specific clinical indications.We therefore included additional information from cfDNA screening and pregnancy outcomes recorded in the birth registry to maximize ascertainment of chromosomal anomalies.Moreover, sensitivity analyses with varying assumptions about outcome ascertainment yielded the same conclusions.
Additionally, some chromosomal anomalies may not have clinically significant features at birth, prompting postnatal cytogenetic investigations, such that an infant could possibly be misclassified as not having a chromosomal anomaly.Because of the difference in ascertainment of pregnancies with nuchal translucencies less than 3.5 mm and 3.5 mm or greater, it is possible that we overestimated the risk of chromosomal anomalies in pregnancies with measurements of 3.5 mm or greater.For this reason, our findings allow us to draw conclusions on the association between nuchal translucency measurements and chromosomal anomalies that have clear features at birth, whereas more careful consideration is needed for chromosomal anomalies for which a clear phenotype is not expected at birth.Additionally, single-gene conditions associated with increased nuchal translucency measurements such as RASopathies are not captured in the registry and were therefore not included. 40Although there is evidence that some factors may influence the choice to have prenatal genetic screening (eg, maternal age, rural residence), 25,41 there is no reason to expect the association between the nuchal translucency measurement and chromosomal anomalies to differ in this population excluded from our study.

Figure 2 .
Figure 2. Investigations Following an Ultrasonographic Nuchal Translucency Measurement 3.5 mm or Greater Under Current Practice

Table 1 .
Characteristics of Study Population by Nuchal Translucency Measurement Abbreviations: ART, assisted reproductive technology; EDD, estimated date of delivery; IVF, in vitro fertilization; NA, not applicable.a Missing data for more than 10.0% of the pregnancies.

Table 3
further categorizes chromosomal anomalies into a group of conditions routinely screened by cfDNA screening (trisomies 21, 18, and 13 and sex chromosome aneuploidies) and a group of conditions beyond the cfDNA screening options consistently available in Ontario (other autosomal aneuploidies, triploidy, mosaic autosomal and sex chromosome aneuploidies, and copy number variants).The risk of chromosomal anomalies routinely screened by cfDNA screening increased with increasing nuchal translucency measurements: for the nuchal translucency category of 3.0 to less than 3.5 mm relative to less than 2.0 mm, the ARD was 8.62% (95% CI, 7.27%-9.96%)and the ARR was 52.15 (95% CI, 43.98-61.84).The risk also increased but with weaker magnitude for the subgroup with other chromosomal anomalies (detailed in eTable 11 in Supplement 1): for nuchal Figure 1.Investigations Following an Ultrasonographic Nuchal Translucency Measurement Less Than 3.5 mm Under Current Practice

Table 2 .
Chromosomal and Pregnancy Outcomes by Nuchal Translucency Measurement a Includes all other chromosomal anomalies, including other autosomal aneuploidies, mosaic aneuploidies, copy number variants.A detailed list of all chromosomal anomalies is available in eTable 2 in Supplement 1.
b Includes those with choromosomal anomaly identified.c Includes trisomies 21, 18, and 13 and sex chromosome aneuploidies.d g Based on cytogenetic testing results.h Includes pregnancy loss, stillbirth, termination, live birth with clinical findings, or a high-risk cfDNA screening result.

Table 3 .
Chromosomal Anomalies by Nuchal Translucency Measurement a Outcome determined by cytogenetic testing.Not having the outcome is determined by normal cytogenetic testing result if performed, normal cfDNA screen result if condition was tested, and by live birth without clinical findings if condition can be clinically diagnosed.withthe group with measurements of at least 3.5 mm (83.6%), illustrating the importance of including outcomes beyond cytogenetic results from prenatal diagnosis.

Open | Genetics and Genomics eTable 3.
Regression Models Including Pregnancies Identified Through Other Data Sources Beyond Multiple-Marker Screening eTable 4. Regression Model With Varying Definition of Exposure by Nuchal Translucency Percentile eTable 5. Sensitivity Analysis for Losses to Follow-Up eTable 6. Regression Models Including All Pregnancies With Nuchal Translucency Measurements eTable 7. Regression Models Including Only Pregnancies With an Estimated Date of Delivery From September 1, 2018, to March 31, 2021 eTable 8. Regression Models Excluding Pregnancies With an Estimated Date of Delivery From April 1, 2020, to March 31, 2021 eTable 9. Time of Cytogenetic Testing of Chromosomal Anomalies Identified by Nuchal Translucency Measurement eTable 10.Post Hoc Analysis Including Additional Adjustment for Age of the Pregnant Individual at Estimated Date of Delivery eTable 11.Chromosomal and Pregnancy Outcomes by Nuchal Translucency Measurement: Detailed Results eTable 12. Characteristics of Pregnancies With and Without Cytogenetic Testing