Population
Apparently structurally normal fetus; conventional testing has not identified an explanation.
Critical appraisal · Journal club
What does a molecular finding tell us when fetal anatomy is reassuring?
MATERNAL–FETAL MEDICINE · GENOMICS · EVIDENCE REVIEW
Open with the clinical tension: sequencing may reveal variation when ultrasound offers no recognizable phenotype. This presentation critically appraises one 2025 meta-analysis using the supplied assessment. It is an educational review, not a new systematic review or an institutional guideline. Suggested duration: 40–50 minutes, including discussion.
Orientation
Should PES be offered after reassuring anatomy and nondiagnostic chromosome testing?
Apparently structurally normal fetus; conventional testing has not identified an explanation.
Whether additional sequence information creates enough value to justify broader testing.
PES means prenatal exome sequencing. CMA means chromosomal microarray analysis. The review allowed normal karyotype or CMA; the actual four cohorts all followed a normal CMA. Do not equate reassuring cfDNA screening with normal diagnostic CMA. Ask learners which outcome would change their recommendation: a molecular result, a confirmed childhood diagnosis, or demonstrated improvement in care.
Orientation
By the end, learners should be able to interpret the estimate and use it in counseling.
Read pooled proportions, confidence intervals, and inheritance patterns.
Assess classification, phenotyping, selection, and follow-up.
Apply the evidence to counseling, FGR, practice, and research.
The presentation is designed for MFM clinicians, genetics teams, residents, and fellows. The clinical practice judgments are the supplied appraisal’s interpretation, not recommendations formally issued by the study investigators or an adopted APA policy.
Executive assessment
Sotiriadis et al. · Ultrasound Obstet Gynecol. 2025;65(5):552–559.
Approximately 1 in 63
95% CI 1.0–2.6%
Approximately 1 in 200
95% CI 0.1–1.5%
Fetuses in four cohorts
32 reported P/LP-positive cases
P/LP means pathogenic or likely pathogenic. Reciprocal frequencies are rounded from the pooled point estimates, not from a screening-benefit calculation. A severe disease-associated variant is not a 0.5% proven risk of severe disease in a future child. Severity, penetrance, expressivity, and age at onset require separate evaluation.
Conceptual framework
The study directly estimates the first step.
How often was an additional P/LP variant reported?
Will this fetus develop clinically important disease?
Does knowing improve decisions, management, or outcomes?
In a phenotype-negative prenatal setting, a molecular finding may establish a diagnosis for some conditions, indicate predisposition for others, or create substantial prognostic uncertainty. We should not assume these possibilities are interchangeable. Clinical utility may include informed reproductive decisions and care planning, not only measurable neonatal treatment effects.
Methods
PROSPERO CRD42024593349 · PRISMA reporting framework.
2018 to September 25, 2024; search repeated before submission. Multiple databases, registers, and reference lists.
1,030 records → seven full-text articles → four included cohorts. Independent screening and data extraction.
Databases included PubMed, Cochrane Library, Google Scholar, Scopus, International HTA, and ClinicalTrials.gov. Inclusion was restricted to English-language reports. Eligible case series required at least 10 fetuses; overlapping samples were checked. These procedures strengthen transparency but cannot repair limitations of the original cohorts.
Methods
All four included studies were retrospective and used exome sequencing after normal CMA.
Apparently normal fetal phenotype; normal karyotype or CMA; prospective or retrospective cohorts; singleton or twin pregnancies.
Four retrospective cohorts. Mostly trio sequencing; Daum used trio analysis in approximately half of cases.
Trio analysis compares fetal and parental sequences and helps identify de novo variants and phase recessive variants. The actual cohorts do not directly establish the same yield after karyotype alone. Clinical selection and the completeness of prenatal assessment differed across studies.
Methods
The comparison is the added result after prior testing, without randomized testing strategies.
Frequency of additional reported P/LP findings among fetuses who underwent PES.
Outcome benefit, causal effects of a testing policy, or the balance of benefit, cost, and psychological burden.
Incremental here refers to findings beyond prior conventional testing. It is not a risk difference from a trial of sequencing versus no sequencing. Avoid interpreting the number needed to test to find a variant as a number needed to test to prevent disease or improve outcomes.
Methods
Random-effects generalized linear mixed model for pooled proportions.
Models binomial outcomes directly and handles sparse events without a continuity correction for zero-event studies.
Only four studies inform the between-study distribution. A pooled mean is not a universal patient-specific probability.
The investigators used metaprop in R and a random-effects GLMM. Do not treat the pooled 1.6% as the simple arithmetic average of cohort percentages or as an exact population prevalence. A confidence interval describes uncertainty around the pooled estimate; it is not a prediction interval for every new clinical population.
Results
Cohort counts are observed data; percentages below are crude cohort yields.
| Cohort | Country | Fetuses | P/LP cases | Crude yield |
|---|---|---|---|---|
| Daum 2023 | Israel | 482 | 4 | 0.83% |
| Gao 2024 | China | 254 | 6 | 2.36% |
| Levy 2024 | Israel | 1,020 | 21 | 2.06% |
| Vaknin 2022 | Israel | 160 | 1 | 0.63% |
| Total | — | 1,916 | 32 | 1.67% |
The meta-analysis reports these denominators and positive counts in Table 1. Levy provides 53.2% of fetuses and 65.6% of positive cases. Those proportions are not GLMM study weights, which the authors did not calculate. Cohort differences include parental selection, reporting policies, and ultrasound definitions. Do not infer geography explains the differences causally.
Results
All estimates are proportions of tested fetuses, not confirmed disease outcomes.
Any P/LP: 1.6% (1.0–2.6%); severe: 0.5% (0.1–1.5%); moderate: 0.5% (0.2–1.5%); de novo moderate/severe: 0.7% (0.4–1.2%). The 0.7% subgroup overlaps the overall and severity categories. These estimates are separately pooled and should not be reconstructed by subtracting or adding point estimates. The chart is newly drawn from published summary values, not a reproduction of the article’s forest plot.
Results
Inheritance counts among the 32 P/LP-positive fetuses.
The separately pooled yields were 1.1% for autosomal dominant findings (95% CI 0.6–2.3%) and 0.3% for autosomal recessive findings (0.1–0.6%). There was only one X-linked recessive case, so its yield could not be robustly pooled. Counts among positives and proportions among all tested fetuses have different denominators.
Results
18 de novo variants among 32 positive cases, all in the dominant group.
0.9%
Pooled yield · 95% CI 0.5–1.7%
0.7%
14 fetuses · 95% CI 0.4–1.2%
Eighteen of 32 positive cases were de novo; eight autosomal dominant variants were inherited, six maternally and two paternally. The 14 moderate/severe de novo cases are a subset, not additional cases. Trio sequencing can recognize these events, but confirming de novo status does not by itself resolve penetrance or clinical severity in this fetus.
Results
The authors applied or revised severity using the Lazarin framework.
| Reported category | Positive cases | Pooled yield |
|---|---|---|
| Mild | 9 | Not pooled: one study |
| Moderate | 10 | 0.5% (95% CI 0.2–1.5%) |
| Severe | 12 | 0.5% (95% CI 0.1–1.5%) |
| Profound | 1 | Not pooled: one case |
These counts sum to 32. Severity classification was a post-hoc subanalysis by clinical experts, with reclassification when needed. A universal prenatal disease-severity standard is lacking. Classification concerns the expected associated disease, not a confirmed individual phenotype. The single profound case should not silently be merged into the published severe estimate.
Critical appraisal
Transparent synthesis is useful even when the underlying evidence is limited.
Registered protocol; systematic search; independent review; explicit inclusion and outcome definitions.
Appropriate sparse-event model; assessment of inheritance and severity; open acknowledgment of missing outcomes.
Avoid dismissing the paper because it contains only four studies. It provides a baseline for a practice already occurring. Its main strengths are transparency, appropriate synthesis, and clear discussion of interpretive limits. At the same time, some severity and inheritance analyses were post hoc and deserve cautious interpretation.
Critical appraisal · 1
Independent reanalysis questioned more than half of P/LP variants in one included cohort.
The authors cite a critique suggesting that over half of the Levy cohort’s P/LP variants could be classified as VUS.
The 1.6% estimate may overstate robust molecular diagnoses or actionable findings. The magnitude is unresolved.
VUS means variant of uncertain significance; the article also uses VOUS. This was a published independent critique, not a definitive reclassification of every included fetus. The original review notes differences in application of ACMG/ClinGen criteria. A P/LP label evaluates disease causation in a defined context; it is not a guarantee of disease, prognosis, or actionability. The investigators proposed reinterpretation after 24–36 months.
Critical appraisal · 2
No included study confirmed normal phenotype systematically at birth or later pregnancy.
Which fetuses developed clinically significant disease? Which remained asymptomatic? When did symptoms emerge?
Postnatally confirmed phenotype, uncertain variants, and expected age of onset could not be pooled.
No systematic postnatal confirmation does not mean every molecular result was incorrect or every child lacked follow-up of any kind. It means sufficient data were not reported to perform the review’s planned analyses. Developmental, neurobehavioral, and adult-onset conditions may require years of follow-up.
Critical appraisal · 3
Three of four cohorts were conducted in Israel; one was conducted in China.
Reasons varied. Gao’s cohort was strongly enriched for a previous affected child or pregnancy, despite current reassuring assessment.
Routine preconception carrier screening in Israel may reduce observed recessive diagnoses and alter the disease mix.
The population is not equivalent to consecutive unselected patients at routine anatomy screening. A known prior familial disorder was tested before inclusion where relevant, but prior history and parental concerns may still shape selection. Country itself is not the causal factor; screening programs, access, consent, referral patterns, and family selection all matter.
Critical appraisal · 4
Definitions and timing of fetal assessment differed across cohorts.
Subtle findings, evolving abnormalities, family phenotype, and gestational age may affect interpretation.
FGR was not consistently or explicitly excluded across the evidence base, although the largest cohort did exclude IUGR.
The review’s discussion identifies incomplete exclusion of FGR as a limitation. Table 1 explicitly lists IUGR among Levy’s exclusions, so saying that no study excluded FGR would be inaccurate. The evidence does not isolate the contribution of growth abnormalities, and it does not provide an isolated-FGR subgroup estimate.
Critical appraisal · 5
Reporting policies varied; only one study reported VUS.
No pooled VUS rate. The psychological and interpretive burden is not quantified.
Policies differed for fetal versus parental results, childhood versus adult onset, and what counted toward yield.
Gao included adult-onset secondary findings among positive fetal cases, whereas Daum excluded a RET secondary finding from primary diagnoses. The authors note that exchanging selected classifications would leave the headline yield at 1.6%, while changing what that yield means clinically. Discuss consent, future child autonomy, incidental parental findings, and reporting preferences without implying a known rate of harm.
Critical appraisal
The authors used a modified Newcastle–Ottawa Scale for single-arm cohorts.
Three studies scored 6/6 stars; one scored 5/6. All were labeled low risk of bias.
Transportability, variant validity, missing longitudinal outcomes, or the net value of universal testing.
Because these were single-arm cohorts, the comparability section was removed from NOS, reducing the maximum to six stars. The resulting scores are reasonable within that framework. They should not be conflated with high certainty of an individual fetus’s prognosis or a proven clinical benefit of a sequencing program.
Critical appraisal
The review does not establish a net benefit from routine PES in this population.
Targeted neonatal assessment, anticipatory planning, reproductive information, and informed family decisions.
Management benefit, cost-effectiveness, uncertainty, psychological effects, and the consequences of broader testing.
Absence of evidence of benefit is not proof of no benefit. Some condition-specific results may be highly useful. The missing claim is that a routine policy creates net benefit across otherwise reassuring pregnancies. This would require management, outcome, patient-reported, and economic evidence beyond variant counts.
Decision analysis
Illustration using pooled point estimates in a comparable population.
95% CI scaled: 100–260
95% CI scaled: 10–150
95% CI scaled: 40–120
Each number is the published pooled proportion multiplied by 10,000. Scaled CIs describe uncertainty in the expected count from the pooled estimate, not prediction intervals for an actual future cohort. The estimates are not additive. There is no single observed cohort of 10,000 in this paper. The uncertainty range for severe-associated findings is especially wide.
Decision analysis
The same estimate can be valued differently by different families.
A potentially important diagnosis before birth may enable earlier assessment, specialty planning, or reproductive choices.
About 99.5% would have no severe-associated finding, while uncertainty, costs, and outcome benefit remain unresolved.
The 99.5% complement applies only to the severe-associated finding category. It does not mean 99.5% have no genetic condition, no other finding, or guaranteed normal outcomes. Some families value information highly despite uncertain immediate actionability; others prefer to avoid ambiguous prognostic information. Good counseling makes those preferences explicit.
Practice implications
Practice interpretation for MFM and genetics teams.
| Clinical situation | Approach |
|---|---|
| Normal anatomy; no relevant family history | Do not routinely recommend PES on the strength of this meta-analysis alone. |
| Known familial monogenic disorder | Offer testing appropriate to the familial variant; broad PES may be unnecessary. |
| Structural anomaly; nondiagnostic CMA | Consider PES with genetics consultation when a monogenic disorder is suspected. |
These are the supplied critical appraisal’s practice judgments. They are not trial-derived treatment recommendations. Targeted familial testing may be appropriate even with entirely normal anatomy. For structural anomalies, sequence interpretation is strengthened by a defined phenotype. ISPD guidance supports responsible clinical use in selected pregnancies and does not endorse routine sequencing for all pregnancies.
Practice implications
The article cites approximately 10% additional yield in unselected structural anomalies.
Approximately 10% in early major cohorts after nondiagnostic conventional testing.
Substantially higher yields in selected conditions; the review cites up to 50–68% in some phenotype-specific settings.
These comparator figures are context cited in the reviewed paper, not new pooled estimates produced by this review. They should not be used as a universal yield for every anomalous fetus or as a head-to-head comparison with the normal-anatomy cohorts. The two populations differ in prior probability and interpretability.
Practice implications
Informed choice can coexist with a recommendation against routine use.
Reassess imaging, growth, family history, prior diagnostic tests, and what information the family hopes to obtain.
Discuss trio analysis, turnaround, limitations, result categories, secondary findings, and genetics follow-up.
Ask what a result would change, including pregnancy decisions, neonatal planning, or future reproduction. Clarify the laboratory’s reporting policy and options for reinterpretation. Explain that a negative exome does not exclude all genetic disease. Sequencing methods and coverage can miss some variant classes; CMA and exome answer different questions.
Counseling
Suggested patient-facing language, adapted from the supplied appraisal.
“When the baby’s anatomy appears normal and chromosome testing is reassuring, exome sequencing may find a pathogenic or likely pathogenic genetic change in about 1–2% of cases. A finding does not always tell us whether a child will develop a significant medical condition.”
Use plain language. Pathogenic and likely pathogenic are classifications of a variant’s relationship to disease, not individual outcome guarantees. Some variants have variable penetrance or expression. For a given condition, the evidence may support more or less certainty than this general discussion. Offer written materials and time for genetics counseling.
Counseling
Suggested continuation of the counseling discussion.
“The estimated chance of finding a change linked to severe disease is about 0.5%, or 1 in 200. These estimates come from small studies. Routine testing in all pregnancies with normal anatomy has not been shown to improve outcomes. A prenatal genetic counselor can help you consider what the information could mean for your family.”
Explain that the severe-associated estimate is imprecise (95% CI 0.1–1.5%) and comes from retrospective selected cohorts. The language is a starting point, not a substitute for individualized informed consent. Ask the patient to explain back the difference between a variant and a predictable outcome, and explore preferences for uncertain or secondary findings.
Fetal growth restriction
Absence of a structural malformation does not establish a low-risk genetic context.
Placental dysfunction, constitutional smallness, chromosomal disease, or a monogenic condition.
Inconsistent FGR exclusion complicates the interpretation of a “normal” population.
FGR means fetal growth restriction. The review neither proves that FGR caused the reported molecular yield nor establishes a yield for isolated FGR. The correct response is to examine the phenotype carefully, including growth trajectory, Doppler assessment, structural imaging, gestational age, and family history. Do not transfer a 1.6% general pooled estimate into an FGR-specific counseling statistic.
Fetal growth restriction
A framework for individualized assessment, not a new testing algorithm.
Additional concerning findings or an evolving phenotype may increase suspicion for an underlying genetic condition.
Reassuring anatomy and surveillance create a different prior probability and counseling context.
Use established FGR evaluation and surveillance pathways. Genetic assessment may include diagnostic chromosome testing and, in selected circumstances, molecular testing after specialist consultation. This presentation intentionally does not invent an FGR sequencing threshold from a study that did not analyze the relevant subgroup.
Research agenda
A prospective multicenter design should connect variant detection with lived outcomes.
Imaging, growth, family history, prior testing, and consecutive recruitment.
Trio sequencing, consistent classification, reporting rules, and reinterpretation.
Confirm phenotype, penetrance, clinical consequences, and delayed onset.
A larger denominator alone will not solve uncertain variant validity and missing follow-up. A useful prospective registry or comparative study should record the pathway from pretest indication through result disclosure, management decisions, birth, and childhood. Systematically document nonparticipation and attrition to assess selection and follow-up bias.
Research agenda
A sequencing program must be evaluated as a care pathway.
Comparative designs can test whether sequencing alters care or outcomes relative to usual pathways; universal randomization may not be the only feasible design. Record whether a change in management was beneficial, neutral, or harmful, not merely whether a change occurred. Patient-reported outcomes should respect differing values rather than presuming all additional information is either good or bad.
Overall judgment
Qualitative appraisal of the evidence supplied for this review.
| Domain | Assessment |
|---|---|
| Clinical relevance | High |
| Methodological rigor | Good |
| Sample size | Limited |
| Certainty of reported diagnostic yield | Moderate to low |
| Evidence of clinical benefit | Insufficient |
| Immediate impact on routine practice | Limited |
These qualitative domain ratings are the supplied appraisal’s interpretation and not a formal GRADE profile. Diagnostic yield is subject to uncertain variant classification and selected recruitment. Clinical benefit is not disproven; it is insufficiently evaluated. Avoid calling this a practice-changing trial or a screening guideline.
Implementation
Proposed actions for APA and OpenMFM; not an adopted institutional policy.
Incorporate estimates and limitations into prenatal genetics counseling and provider education. This study alone does not support universal PES.
Teach the distinction between diagnostic yield, predictive validity, and clinical utility through this evidence review.
APA refers to the organization named in the supplied brief, consistent with Atlanta Perinatal Associates in existing repository branding. These proposed recommendations should not be displayed as an official policy already approved by APA. Decisions about offering and funding testing require local clinical governance and up-to-date guidance.
Journal club
Hypothetical case for discussion.
A fetus has normal anatomy and nondiagnostic CMA. Trio PES identifies a P/LP variant linked to a variably penetrant dominant condition.
Have we established disease, identified predisposition, or created a category of prenatal uncertainty?
What condition-specific evidence, parental phenotype, laboratory review, and follow-up plan would change the interpretation?
No real patient is represented. Do not ask learners to resolve an unnamed variant without the gene, variant evidence, inheritance details, and phenotype. The teaching goal is recognizing what information is missing. Condition-specific consultation may support a firm diagnosis, a probabilistic risk estimate, or a need for reinterpretation and observation.
Journal club
Use these prompts before translating the findings into a recommendation.
No. The per-10,000 categories overlap and are separately pooled.
No. It is the yield of variants associated with a severe disease category.
No. No isolated-FGR subgroup yield was established.
Use this as a brief knowledge check or ask the audience to answer before revealing the slide. The denominators, categories, and distinction between findings and disease are the highest-value interpretation skills in this review. Each answer should connect to a specific limitation in the methods or source evidence.
Take-home message
Our ability to detect variation exceeds our ability to predict all of its consequences.
A baseline for counseling.
A meaningful but uncertain signal.
Unproven
Universal PES is not established.
Return to the opening question. This paper advances the evidence for prenatal genomic detection but does not establish the case for universal exome sequencing in structurally normal pregnancies. The 0.7% moderate/severe de novo signal deserves further study, especially because carrier screening does not reliably anticipate those events. The appropriate response is better counseling and better outcome evidence.
References
The article’s tables, results, and discussion underpin the quantitative appraisal.
The included cohort counts and reported diagnostic categories were verified against the open-access article. Primary cohort details and critiques are linked for deeper review. This slideshow synthesizes the supplied appraisal and the named sources; it is not a comprehensive update of studies published after the review’s search.
References
Evidence review prepared November 12, 2026; primary review search ended in 2024.
Sources support different claims: PRISMA addresses reporting, ACMG/AMP addresses variant interpretation, Lazarin supplies a severity framework, and ISPD supplies clinical context. Do not infer that any one of these documents endorses routine sequencing in otherwise low-risk pregnancies. The deck reviews the specified 2025 paper rather than claiming to summarize all evidence available in 2026.