Faculty Bibliography

Objective: To determine whether a machine learning model can predict patients at risk for shoulder dystocia (SD) better than estimations of birthweight (BW) alone.
Study Design: This was a retrospective analysis of 17,731 pregnant individuals from 7/2013 to 10/2018. Utilizing a machine learning model, a total of 122 binary and continuous variables were included. Baseline models were built with different sets of variables during 3 time periods: 57 in antepartum period, 96 in stage 1 and 119 in stage 2 of labor. BW was used as a proxy for estimated fetal weight (EFW) because documented assessment of EFW was not available in all cases. Two decision tree-based models, Random Forest and XGBoost, were used as predictive models and performance was evaluated with 5-fold cross validation. Area under the receiver operating characteristic curve (AUROC) and area under precision/recall curve (AUPR) were used as metrics for evaluating model performance. Mean and standard error of performance metrics were calculated.
Result(s): The cohort included 21,232 vaginal deliveries. There were 415 (1.95%) vaginal deliveries complicated by SD that occurred in 406 patients. An AUROC of 0.73 +/- 0.01 (mean +/- standard error) was achieved for RandomForest model and 0.78 +/- 0.01 for XGBoost model for BW. BW was then added as a variable to variable sets from the antepartum period and each labor stage to assess performance change. The RandomForest model predicted patients at risk for SD better than BW alone, but only the AUROC was statistically significant at all stages (p < 0.05). (Figure 1) This finding was not demonstrated in the XGBoost model.
Conclusion(s): Our machine learning model performed better in predicting SD than EFW alone (using BW as a proxy) at each time period evaluated - the antepartum period, 1st stage and 2nd stage of labor. While our results indicate that our model may enhance the prediction of SD, these findings should be validated using a more robust data set that includes documented EFW to account for the margin of error between BW and EFW. [Formula presented] [Formula presented]
Copyright

Objective: Iron deficiency is the most common cause of anemia in pregnancy, which is associated with maternal and neonatal complications. The objective of this study was to identify the frequency and risk factors for isolated iron deficiency (iID) in the 1^st trimester.
Study Design: This was a secondary analysis of a prospective cohort study to identify risk factors for iID of non-anemic pregnant individuals presenting for prenatal care in the 1^st trimester from February 2022 to June 2022 at NYU Langone Health. iID was defined as serum ferritin level of <= 29 ng/mL. Inclusion criteria included pregnant individuals ages 18-60 years with a singleton gestation enrolled in the 1^st trimester (prior to 14 weeks 0 days gestation) and were non-anemic (hemoglobin >=11.0 g/dl). Patients were excluded if they had 1^st trimester anemia or history of blood transfusion 3 months prior to pregnancy. Univariate analyses were followed by multiple logistic regression (OR [95% CI]) with statistical significance defined at p< 0.05.
Result(s): Of 600 patients enrolled in the study, 89 (14.8%) had 1^st trimester iID. Black/African American patients (19.1% vs. 8.2%, p=0.003), those with uterine fibroids (20.2% vs. 9.4%, p=0.003), and those with higher median BMI (25.2 kg/m² (IQR 22.7-29.6) vs. 23.6 kg/m² (IQR 21.4-27.3), p=0.01) were more likely have to 1^st trimester iID. White patients (41.6% vs. 56.6%, p=0.01) and those with a normal BMI (44.9% vs. 57.3%, p=0.04) were less likely to have 1^st trimester iID. After adjusting for confounders in regression models, Black/African American patients had the strongest association with 1^st trimester iID (aOR 2.18 [1.12-4.11], p=0.02), followed by uterine fibroids (aOR 2.02 [1.05-3.72], p=0.03). Overweight or obese BMIs were not identified as risk factors for 1^st trimester iID.
Conclusion(s): Based on our findings, Black/African American pregnant individuals and those with uterine fibroids are at highest risk for 1^st trimester iID. Future studies should investigate perinatal outcomes for pregnant individuals with 1^st trimester iID. [Formula presented]
Copyright

Objective: Reticulocyte hemoglobin (RetHb) is used for early detection of iron deficiency (ID) in the nonpregnant patient population. It provides an indication of iron availability in the bone marrow and is an early marker of iron deficiency (ID) erythropoiesis before anemia is present. Due to the paucity of data regarding RetHb use in pregnancy, the study objective was to establish normal values and trend for RetHb during the 1^st and 2^nd trimester of pregnancy by correlating it with ferritin and Hb.
Study Design: This is a secondary analysis of an observational, prospective cohort study evaluating ID parameters in singleton gestations presenting for prenatal care in the first trimester from 2/2022 to 6/2022. ID was defined as serum ferritin level of <= 29 ng/mL. For this analysis, patients were excluded if they had 1^st trimester anemia (Hb< 11.0 g/dL) or history of blood transfusion 3 months prior to pregnancy. Data were analyzed using student's t-test and linear regression modeling with statistical significance defined at p< 0.05.
Result(s): 209 patients met inclusion criteria. In table 1, demographics of the study cohort are shown. There was a prevalence of 16.3% of ID in the 1^st trimester and 69.9% in the 2^nd trimester. Distribution of RetHb values throughout the first 2 trimesters are demonstrated in figure 1. In the 1^st trimester, the mean RetHb in women with ID was 34.01 pg +/- SD 1.64 compared to 34.68 pg +/- 1.65 in women without iron deficiency (p=0.03). In the second trimester, the mean RetHb in women with ID was 33.02 pg +/- 2.34 compared to 34.20 pg +/- SD 1.94 in women without ID (p< 0.001).
Conclusion(s): A statistically significant physiologic decrease was observed among pregnant individuals both with and without iron deficiency in the first and second trimester of pregnancy. Future studies should evaluate the utility of RetHb use in pregnancy to predict ID and iron deficiency anemia throughout pregnancy. [Formula presented] [Formula presented]
Copyright

Objective: To compare pregnancy latency achieved with oral labetalol versus extended-release nifedipine during expectant management of preterm preeclampsia with severe features (PEC-SF).
Study Design: This is a retrospective cohort study of patients initiated on antihypertensive therapy with oral labetalol or extended-release nifedipine during admission for expectant management of PEC-SF < 34 weeks between 1/2013 and 4/2022. Those on antihypertensive therapy prior to admission or with another indication for delivery < 34 weeks were excluded (monochorionic-monoamniotic twins, higher order multiples, absent or reversed umbilical artery Dopplers). Pregnancy latency (from oral agent initiation to delivery decision) was compared between groups. Secondary outcomes included need for initial agent dose uptitration, addition of second oral agent, acute antihypertensive therapy, and delivery for refractory hypertension. Linear and modified Poisson regression models were used to estimate adjusted mean differences (AMD) with 95% confidence intervals.
Result(s): The cohort included 78 patients (Table 1). Comparing those initiated on labetalol versus extended-release nifedipine (Table 2), there was no difference in latency (6.2 (7.5) vs 5.4 (7.4) days, AMD 1.1 days, 95% CI [-2.1, 4.4]), nor in the proportion of patients achieving 1 week latency (25.0% vs 23.8%, respectively, AMD 2.9%, 95% CI [-16.5, 22.3]). Those initiated on labetalol were less likely to require a second agent (16.7% vs 38.1% for nifedipine, AMD -18.4, 95% CI [-37.3, 0.5]). There were no differences in need for initial agent uptitration, acute antihypertensive therapy, or delivery for refractory hypertension.
Conclusion(s): There was no difference in pregnancy latency among patients with PEC-SF initiated on oral labetalol versus extended-release nifedipine. Patients on labetalol may be less likely to require a second antihypertensive agent, but comparative outcome estimates may be limited by small cohort size. Further investigations with a larger cohort should be performed to evaluate for any relative advantages of the two oral agents. [Formula presented] [Formula presented]
Copyright

OBJECTIVE:The ideal time for birth in pregnancies diagnosed with vasa previa remains unclear. We conducted a systematic review aiming to identify the gestational age at delivery that best balances the risks for prematurity with that of pregnancy prolongation in cases with prenatally diagnosed vasa previa. DATA SOURCES/METHODS:Ovid MEDLINE, PubMed, CINAHL, Embase, Scopus, and Web of Science were searched from inception to January 2022. STUDY ELIGIBILITY CRITERIA/METHODS:The intervention analyzed was delivery at various gestational ages in pregnancies prenatally diagnosed with vasa previa. Cohort studies, case series, and case reports were included in the qualitative synthesis. When summary figures could not be obtained directly from the studies for the quantitative synthesis, authors were contacted and asked to provide a breakdown of perinatal outcomes by gestational age at birth. METHODS:Study appraisal was completed using the National Institutes of Health quality assessment tool for the respective study types. Statistical analysis was performed using a random-effects meta-analysis of proportions. RESULTS:The search identified 3435 studies of which 1264 were duplicates. After screening 2171 titles and abstracts, 140 studies proceeded to the full-text screen. A total of 37 studies were included for analysis, 14 of which were included in a quantitative synthesis. Among 490 neonates, there were 2 perinatal deaths (0.4%), both of which were neonatal deaths before 32 weeks' gestation. In general, the rate of neonatal complications decreased steadily from <32 weeks' gestation (4.6% rate of perinatal death, 91.2% respiratory distress, 11.4% 5-minute Apgar score <7, 23.3% neonatal blood transfusion, 100% neonatal intensive care unit admission, and 100% low birthweight) to 36 weeks' gestation (0% perinatal death, 5.3% respiratory distress, 0% 5-minute Apgar score <7, 2.9% neonatal blood transfusion, 29.2% neonatal intensive care unit admission, and 30.9% low birthweight). Complications then increased slightly at 37 weeks' gestation before decreasing again at 38 weeks' gestation. CONCLUSION/CONCLUSIONS:Prolonging pregnancies until 36 weeks' gestation seems to be safe and beneficial in otherwise uncomplicated pregnancies with antenatally diagnosed vasa previa.

BACKGROUND:Cell-free DNA (cfDNA) non-invasive prenatal screening for trisomy (T) 21, 18, and 13 has been rapidly adopted into clinical practice. However, prior studies are limited by lack of follow up genetic testing to confirm outcomes and accurately assess test performance, particularly in women at low-risk for aneuploidy. OBJECTIVE:To compare the performance of cfDNA screening for T21, T18 and T13 between women at low and high-risk for aneuploidy in a large, prospective cohort with genetic confirmation of results. STUDY DESIGN/METHODS:A multicenter prospective observational study at 21 centers in 6 countries. Women who had SNP-based cfDNA screening for T21, T18 and T13 were enrolled. Genetic confirmation was obtained from prenatal or newborn DNA samples. Test performance and test failure (no-call) rates were assessed for the cohort and women with low and high prior risk for aneuploidy were compared. An updated cfDNA algorithm, blinded to pregnancy outcome, was also assessed. RESULTS:20,194 were enrolled at median gestational age of 12.6 weeks (IQR:11.6, 13.9). Genetic outcomes were confirmed in 17,851 (88.4%): 13,043 (73.1%) low-risk and 4,808 (26.9%) high-risk for aneuploidy. Overall, 133 trisomies were diagnosed (100 T21; 18 T18; 15 T13). cfDNA screen positive rate was lower in low- vs. high-risk (0.27% vs. 2.2%, p<0.0001). Sensitivity and specificity were similar between groups. The positive predictive value (PPV) for the low and high-risk groups was 85.7% vs. 97.5%, p=0.058 for T21; 50.0% vs. 81.3%, p=0.283 for T18; and 62.5% vs. 83.3, p=0.58 for T13, respectively. Overall, 602 (3.4%) patients had no-call result after the first draw and 287 (1.61%) after including cases with a second draw. Trisomy rate was higher in the 287 with no-call results than patients with a result on a first draw (2.8% vs. 0.7%, p=0.001). The updated algorithm showed similar sensitivity and specificity to the study algorhitm with a lower no-call rate. CONCLUSIONS:In women at low-risk for aneuploidy, SNP-based cfDNA has high sensitivity and specificity, PPV of 85.7% for T21 and 74.3% for the three common trisomies. Patients who receive a no-call result are at increased risk of aneuploidy and require additional investigation.

BACKGROUND:Prenatal screening has historically focused primarily on detection of fetal aneuploidies. Cell-free DNA (cfDNA) now enables noninvasive screening for subchromosomal copy number variants, including 22q11.2 deletion syndrome (22q11.2DS or DiGeorge syndrome), which is the most common microdeletion and a leading cause of congenital heart defects and neurodevelopmental delay. Although smaller studies have demonstrated the feasibility of screening for 22q11.2DS, large cohort studies with postnatal confirmatory testing to assess test performance have not been reported. OBJECTIVE:To assess the performance of SNP-based cfDNA prenatal screening for detection of 22q11.2DS. STUDY DESIGN/METHODS:Patients who had SNP-based cfDNA prenatal screening for 22q11.2DS were prospectively enrolled at 21 centers in 6 countries. Prenatal or newborn DNA samples were requested in all cases for genetic confirmation with chromosomal microarray. The primary outcome was sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of cfDNA for detection of all deletions, including the classical deletion and nested deletions that are ≥500kb, in the 22q11.2 low copy repeat A-D region. Secondary outcomes included the prevalence of 22q11.2DS and performance of an updated cfDNA algorithm that was evaluated blinded to pregnancy outcome. RESULTS:Of 20,887 women enrolled, genetic outcome was available in 18,289 (87.6%). Twelve 22q11.2DS cases were confirmed in the cohort, including five (41.7%) nested deletions, yielding a prevalence of 1:1524. In the total cohort, cfDNA reported 17,976 (98.3%) as low risk for 22q11.2DS and 38 (0.2%) as high-risk; 275 (1.5%) were non-reportable. Overall, 9 of 12 cases of 22q11.2 were detected, yielding a sensitivity of 75.0% (95% CI: 42.8, 94.5); specificity of 99.84% (95% CI: 99.77, 99.89); PPV of 23.7% (95% CI: 11.44, 40.24) and NPV of 99.98% (95% CI: 99.95, 100). None of the cases with a non-reportable result was diagnosed with 22q11.2DS. The updated algorithm detected 10/12 cases (83.3%; 95% CI: 51.6-97.9) with a lower false positive rate (0.05% vs. 0.16%, p<0.001) and a PPV of 52.6% (10/19; 95% CI 28.9-75.6). CONCLUSIONS:Noninvasive cfDNA prenatal screening for 22q11.2DS can detect most affected cases, including smaller nested deletions, with a low false positive rate.

Objective: Non-invasive prenatal screening with cell free DNA (cfDNA) includes the option to screen for microdeletion syndromes but data on test performance are limited. We report on performance of cfDNA for detection of 4 microdeletion syndromes: Cri-Du-Chat (5p-), Prader-Willi syndrome (PWS), Angelman syndrome (AS) and 1p36del syndrome.
Study Design: Secondary analysis of the SMART multicenter prospective study, which assessed cfDNA performance for 22q11.2 deletion. Newborn or fetal samples were requested in all cases for genetic confirmation with chromosomal microarray (CMA). SNP-based cfDNA screening for 4 microdeletion syndromes was performed using an investigational algorithm on patients who requested testing for these syndromes or who agreed to future research; results were compared to blinded CMA confirmation. Differentiation between PWS and AS, caused by a similar deletion/imprinting mechanism, was accomplished by comparing neonatal SNPs on CMA and maternal SNPs, if available from the cfDNA sample, in the affected region. Deletions >500kb in the syndrome critical region were considered positive. Deletions < 500kb or not including the disease-causing genes were classified as variants of uncertain significance (VUS).
Result(s): Overall, 10,971 had both cfDNA and DNA confirmation results. Median gestational age at enrollment was 13.3 weeks (8.9-36.1). CMA confirmed 5 PWS cases (1:2194), one case of PWS/AS and one 5p- (Table). Four 5p- deletions and one 1p36del were classified as VUS. Of the 7 confirmed microdeletion cases, 6 were detected by cfDNA (86.7%), including all PWS/AS cases; the 5p- case was not detected. cfDNA was reported as high-risk in 14 cases (0.12%), with FP=0.07%, a PPV of 62.5% (5/8) for PWS, and 0% (0/6) for the remaining 3 conditions. None of the confirmed cases had increased NT or structural anomalies at the time of the anatomic survey. One PWS case was diagnosed with fetal anomalies at 32wks and 2 PWS cases were diagnosed with abnormalities after birth.
Conclusion(s): cfDNA prenatal screening detected 6/7 microdeletions, including all cases of PWS/AS, with a low false positive rate. [Formula presented]
Copyright