The original work of Lo et al1 demonstrated the presence of circulating cell-free DNA in maternal plasma. Subsequent studies explored various settings in which noninvasive testing utilizing cell-free DNA has clinical value.2–4
Few studies have explored cell-free DNA in nonviable gestations. One study investigating missed abortions found increased levels of cell-free DNA in maternal blood in first-trimester abortions relative to viable pregnancies.5 Approximately 15% of clinically recognized pregnancies result in spontaneous abortion.6 Additionally, an estimated 26,000 intrauterine fetal demises occur per year in the United States.7
Intrauterine fetal demise, or stillbirth, defined as fetal death occurring after 20 weeks of gestation or fetal weight of 350 g or greater, has an estimated incidence of 6.2 per 1,000 live births. Cytogenetic abnormalities account for 50% of fetal deaths at less than 20 weeks of gestation and 6–15% at 20 weeks of gestation or greater.8–11 Data indicate that microarray identifies an additional 41.9% of genetic abnormalities in the intrauterine fetal demise.12 Invasive testing provided superior culture rates (85%) over fetal tissue sampling after birth (28%).13
Karyotype is recommended in the workup of stillbirth unless an obvious cause is identified (such as death occurring after maternal trauma, etc).14 Every spontaneous abortion does not require an assessment of aneuploidy status. After two or three spontaneous abortions, the American Society of Reproductive Medicine recommends cytogenetic evaluation of the products of conception.15 Our study was designed to explore the feasibility of cell-free DNA in situations in which aneuploidy testing is indicated.
The objective of this study was to estimate whether cell-free DNA is present in maternal circulation in the setting of intrauterine fetal demise and missed abortion. Secondary objectives included evaluating factors associated with nonresults and clinical applications.
MATERIALS AND METHODS
Because limited data exist in the use of cell-free DNA in missed abortion and fetal demise, sufficient information to perform power analysis and sample size calculations was not available. Therefore, we performed a prospective cohort study enrolling 50 patients from June 2013 to January 2014 at MedStar Washington Hospital Center, Washington, DC. The target of 50 was reached after a review of funding and demographics of our institution. The study received approval from the institutional review board at the MedStar Health Research Institute at Medstar Washington Hospital Center and was registered at ClinicalTrials.gov (#NCT01916928). Women presenting to our institution with a nonviable singleton pregnancy were screened for study participation. For study purposes, nonviable pregnancies were defined as missed abortion (with or without embryo present) at less than 20 weeks of gestation and intrauterine fetal demises were defined as fetal loss after 20 weeks of gestation. We defined the time since fetal demise as days elapsed between the last documented heartbeat and the time of maternal blood draw. When used clinically for prenatal screening, noninvasive prenatal testing requires a gestational age of 10 weeks or greater.16 In this study, all gestational ages were considered for participation.
After ultrasonographic confirmation of a nonviable singleton or anembryonic gestation, women who could be adequately consented in English, were at least 18 years of age, and had products of conception in utero at the time of blood sampling were eligible for study participation. Women with incomplete or completed abortions or ectopic pregnancies were excluded. Additionally, participants known to carry a maternal genetic abnormality were ineligible.
Recruitment locations included the emergency department, perioperative holding area, and the obstetrics and gynecology clinics at MedStar Washington Hospital Center. Patients were identified by daily review of hospital records and by notification from health care providers in these locations. Once patients were identified, the investigators or clinical coordinator discussed the specifics of the study with the patient and offered enrollment. Individuals involved with patient recruitment did not manage the patient's nonviable pregnancy. Of note, the individuals involved with patient recruitment were not involved in the overall care of the patient. After informed consent, blood was collected in two 10 mL Cell-Free DNA BCT tubes before uterine evacuation. Patients were offered the standard workup for fetal demise, missed abortion, or both in accordance with the recommendations set forth by the American College of Obstetricians and Gynecologists: Management of Stillbirth Practice Bulletin Number 102, March 2009, reaffirmed 2012. Additionally, the providing practitioner ordered tests deemed clinically relevant based on the presenting scenario. Genetic testing ordered by the providing physician was billed to the patient or her insurer as appropriate and was not sponsored by Sequenom Laboratories. As such, the patient evaluation was left to the discretion of the managing health care provider. Therefore, participants were not required to have subsequent genetic analysis (prior noninvasive prenatal testing results, karyotype, microarray, or all of these) performed for study participation. The maternal chart was reviewed by the first author and clinical research coordinator to collect data pertinent to analysis. The deidentified information was recorded on a standardized clinical data form. All plasma samples were sent to the processing laboratory (Sequenom Laboratories, San Diego, CA) for cell-free DNA analysis. The processing laboratory was blinded to the prenatal or postmortem diagnostic results.
The maternal whole blood collected in Cell-Free DNA BCT tubes was centrifuged (Eppendorf 5810R plus swing-out rotor), warmed (25°C) at 1,600 g for 15 minutes, and the plasma was collected. The BCT plasma was centrifuged a second time (Eppendorf 5810R plus swing-out rotor) at 25°C at 2,500 g for 10 minutes. After the second spin, the plasma was removed from the pellet that formed at the bottom of the tube, distributed into 4-mL bar-coded plasma aliquots, and immediately stored frozen at −70°C or less until DNA extraction. All samples were processed and plasma aliquots were frozen within 72 hours of collection.
Cell-free DNA was extracted from maternal plasma using the QIAamp Circulating Nucleic Acid Kit as previously described.17 Library preparation was performed as previously described.18
Four isomolar sequencing libraries were pooled and sequenced together on the same lane (four-plex) of an Illumina v3 flowcell on an Illumina HiSeq2000 sequencer. Paired end sequencing by synthesis was performed for two×36 cycles followed by seven cycles to read each sample index.
Sequencing data processing, alignment, and summation of sequencing reads present within nonoverlapping 50-kbp segments of the genome were performed as previously described.18 To normalize the 50-kbp raw segment count, gas chromatography bias was removed by using a locally weighted scatterplot smoothing-based correction for a sample-wise correction, similar to the one described in Alkan et al19 but using an additive rather than multiplicative correction. Subsequently, principal component analysis was used to remove higher order artifacts for a population-based correction.20,21 After data normalization, a laboratory-designed decision tree method was implemented to detect copy number variants.22 Although a limited set of trisomies and copy number variants are reported clinically, we chose to report all detected aberrations during this research study.
Fetal fraction was calculated for all samples using a model trained on the localized differences in relative sequencing read coverage between maternal and fetal circulating cell-free DNA.23
Analyses for the primary objective involved comparisons of variables between cell-free DNA result-yielding and non–result-yielding specimens, with nonresults defined as a fetal fraction less than 3.7% (laboratory designation of insufficient fetal material) or greater than 65% (fetal fraction algorithm was not sufficiently validated on fetal fraction levels exceeding this value). These cutoff points are commonly used in clinical practice as a result of algorithm irregularities that occur at excessively low or high fetal fractions. Comparison of the fetal fraction observed in this study to presumed viable pregnancies was performed using fetal fraction information from 140,377 samples submitted to Sequenom Laboratories for noninvasive prenatal testing. For binary variables, Wilson confidence intervals (CIs) were calculated. Given nonnormal distribution of continuous variables, Wilcoxon rank-sum tests were used to compare medians between groups. To compare frequencies of categorical variables, Fisher's exact tests were used. All statistical analyses were conducted using STATA 13.1. Statistical significance was determined by a P value ≤.05 and in logistic regression modeling, a 95% CI that did not cross one.
The plasma of 50 patients carrying a nonviable fetus were analyzed with characteristics summarized in Table 1. Three of the 50 specimens included were recruited postintraamniotic administration of digoxin as a result of fetal anomalies. The mean age and body mass index (calculated as weight (kg)/[height (m)]2) of patients were 31.4 years (standard deviation [SD] 6.1) and 30.3 (SD 9.1), respectively. The mean clinical gestational age at diagnosis of nonviable pregnancy was 16.9 weeks (SD 9.2, range 5.3–38.4 weeks of gestation) with 48% (n=24) of pregnancies falling in the first trimester, 40% (n=20) in the second trimester, and 12% (n=6) in the third trimester. For the purposes of this study, the first trimester included gestations through 13 6/7 weeks of gestation. The second trimester included gestational ages between 14 0/7 and 27 6/7 weeks; the third trimester included all gestations at 28 0/7 weeks and beyond. Gestational age determined by ultrasonography at the time of diagnosis of nonviable pregnancy differed from established clinical gestational age by an average of 1.4 weeks (SD 1.9). Seventeen pregnancies (34%) had ultrasonographic gestational ages less than 8 weeks. The mean ultrasonographic gestational age at diagnosis was 15.5 weeks (SD 10.1, range 6.1–38.4 weeks). Although 24 patients had clinical dating indicating a first-trimester gestation, 27 pregnancies were consistent with first-trimester gestations based on crown–rump length measurement. The majority (n=31 [62%]) of fetuses were female.
The fetal fraction increased proportionally to gestational age, similar to viable pregnancies (Fig. 1A). The fetal fraction of the study population as it relates to ultrasound findings is illustrated in Figure 1B. Of the 38 samples with a reportable result, 29 were classified as euploid, eight samples showed full chromosome trisomies, and one sample was reported to contain a microdeletion. Analysis for sex chromosome aneuploidies was also performed, in which none was identified (Appendix 1). Although 50% of the observed fetal aneuploidies were from commonly observed chromosomes (two trisomy 21; one trisomy 13; one trisomy 18), a number of less common aneuploidies were observed including single instances of trisomy 16, trisomy 14, trisomy 10, and a likely mosaic case of trisomy 7. The remaining 12 samples could not be classified because of fetal fraction outside of the reportable range. A cell-free DNA result was obtained after diagnosis of nonviable pregnancy in 38 (76%) of patients.
Twelve of 50 samples (24%) failed to yield a cell free-fetal DNA result. On further examination of the distribution of results, gestational age was a factor in obtaining cell-free DNA within range. A cell-free DNA result was obtained 87.9% (29/33; 95% CI 72.7–95.2) of the time at ultrasonographic gestational ages 8 weeks or greater compared with 52.9% (9/17; 95% CI 31.0–73.8) of the time at ages less than 8 weeks of gestation (P=.012) (Appendix 2). Cell-free DNA yielded a result for 63% (15/24; 95% CI 42.7–78.8) of samples in the first trimester of pregnancy and for 88% (23/26; 95% CI 71.0–96.0) of samples in the second and third trimesters (P=.047). The ability to obtain results also varied according to the difference in clinical compared with ultrasonographic gestational age (median difference=2.1 weeks in the nonresults group compared with 0 week in the results group, P=.03). The nonresults group did not have a statistically different body mass index than those with a result (median 37.1 compared with 27.5, P=.05). Finally, the distribution of ultrasound findings in the first trimester did not differ between groups in the presence of a fetal pole (P=.06). Both third-trimester fetuses in which results were not obtained were excluded as a result of fetal fractions 65% or greater. The average estimated fetal fraction of these two samples was 91%, although these values are likely overestimated as a result of the lack of algorithm training at elevated fetal fraction levels.
Diagnostic cytogenetic analysis (by antenatal chorionic villus sampling, amniocentesis, or postmortem cytogenetics) was available in 36% (n=18) of the study population (Table 2). Two samples were classified as nonresults as a result of a fetal fraction outside of the prespecified range. Although the small sample size of both cell-free DNA results and confirmatory cytogenetic results (n=16) limit the validity of sensitivity and specificity calculations, 87.5% (14/16 samples) had concordant results. The two discrepancies occurred in a monosomy X (euploid by cell-free DNA) and in a 1.97-MB interstitial deletion of X (trisomy 7 by cell-free DNA).
Although cell-free DNA was present in all of the nonviable pregnancies in the study, fetal fractions reached a concentration of greater than 3.7% in more than 75% after an ultrasonographic gestational age of 8 weeks or greater. Nine of the 38 patients in which there was a result from cell-free DNA analysis demonstrated a genetic abnormality. The most common aneuploidies identified included trisomies 13, 18, and 21, which is concordant with conventional cytogenetic analysis of intrauterine fetal demises.24 Of note, the most commonly found aneuploidies in intrauterine fetal demise are monosomy X (23%), trisomy 21 (23%), trisomy 18 (21%), and trisomy 13 (8%).14 This would result in approximately 25% of fetal aneuploidies being missed if testing were to be restricted to the aforementioned abnormalities. Our study demonstrated the ability of cell-free DNA to identify other abnormalities including trisomies 7, 10, 14, and 16. Therefore, as testing continues to expand in the amount of chromosomes reported, this should become less of an issue.
Similar to previous studies, cell-free DNA was isolated more often with advancing gestational age.25 This could potentially provide clinical information regarding the presence of a genetic abnormality after 8 weeks of gestation among missed abortions compared with the 10-week threshold in viable pregnancies because a fetal fraction of at least 3.7% is likely to be obtained at that time. When genetic analysis is desired, cell-free DNA could possibly avoid a dilation and curettage and provide the opportunity for nonsurgical management. Although results will not be immediately available, the average turnaround time of cell-free DNA is 4.5 days in a commercial setting.16
The small sample size chosen for this test of feasibility limits statistical calculations of sensitivity, specificity, positive predictive value, and negative predictive value. However, our data suggest that for standard aneuploidies (trisomy 13, 18, and 21), the test performed similarly to viable pregnancies. The two main discrepant results occurred in the presence of microdeletions and monosomy. The microdeletion in our sample was identified as a small, 1.97-MB interstitial deletion of XP22.33-P22.33, 46 XY. The changes in this region are associated with chondrodysplasia punctata. This result by cell-free DNA was reported as trisomy 7 with the specification that this could possibly be a mosaic case resulting from discordance between the fetal fraction and the magnitude of the signal. Although no confirmatory analyses were performed, this discordant result may highlight a known limitation of cell-free DNA testing, confined placental mosaicism, which might not be detected by direct testing. The processing laboratory suggested that this sample may be a reflection of confined placental mosaicism and thus may not be detected by direct testing of the product of conception.
Case reports have illustrated false-positive results among fetal deletions and duplications. Although the fetuses in this case report ultimately demonstrated a genetic abnormality, the final diagnosis was a deletion or duplication rather than a trisomy.26 Finally, several studies have shown decreased performance with regard to abnormalities of sex chromosomes; therefore, the result of a normal female by cell-free DNA in the presence of Turner syndrome is not unanticipated.27
After an ultrasonographic gestational age of 8 weeks among nonviable pregnancies in which trisomies 13, 18, or 21 are suspected, cell-free DNA is likely to identify these abnormalities. This option may be attractive to clinicians and patients alike because it avoids postmortem tissue sampling or amniocentesis. Although guidelines have been established for the management and evaluation of intrauterine fetal demise, insurance companies vary in their coverage of the evaluation to include the cost of the autopsy and maternal laboratory tests.14 Coverage will certainly vary in the setting of a nonviable pregnancy. Like in current practice, the cost of this testing may be prohibitive to some patients.
In those patients in whom recommended testing is not feasible, either as a result of availability or patient preference, cell-free DNA could be useful in identifying common aneuploidies, especially after 8 weeks of gestation when the products of conception are still in utero. Care must be taken in women with presenting with their first loss under 14 weeks of gestation. As many as 50% of spontaneously aborted conceptus will have an identifiable aneuploidy; therefore, although cell-free DNA may identify this abnormality, the clinical utility for future pregnancies will likely be minimal.
Consistent with the observation that complete clearance of cell-free DNA from the maternal circulation occurs after delivery of the placenta, our results indicate that cell-free DNA remains in maternal plasma when the gestation remains in situ, possibly as a result of continued placental apoptosis after fetal demise.28 Our study is limited by the number of patients who had diagnostic cytogenetics testing available for comparison with the postdemise cell-free DNA results. Given that our protocol allowed for standard evaluation at the health care provider's discretion, not every specimen had cytogenetic, microarray results, or both for comparison. In our institution, cytogenetic evaluation of first-trimester missed abortion is commonly not performed.
Because intrauterine fetal demise is fortunately a rare occurrence, research among this population is difficult. In accordance with recent literature and published guidelines, amniocentesis or postmortem karyotype with microarray is the first line for the evaluation of genetic abnormalities among intrauterine fetal demises11,12; however, when the aforementioned items are not feasible, the use of cell-free DNA can provide useful information and is a reasonable complementary test.
1. Lo YM, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, et al.. Presence of fetal DNA in maternal plasma and serum. Lancet 1997;350:485–7.
2. Hooks J, Wolfberg AJ, Wang ET, Struble CA, Zahn J, Juneau K, et al.. Non-invasive risk assessment of fetal sex chromosome aneuploidy through directed analysis and incorporation of fetal fraction. Prenat Diagn 2014;34:496–9.
3. Norton ME, Brar H, Weiss J, Karimi A, Laurent LC, Caughey AB, et al.. Non-Invasive Chromosomal Evaluation (NICE) Study: results of a multicenter prospective cohort study for detection of fetal trisomy 21 and trisomy 18. Am J Obstet Gynecol 2012;207:137.e1–8.
4. Soothill P, Finning K, Latham T, Wreford-Bush T, Ford J, Daniels G. Use of cffDNA to avoid administration of anti-D to pregnant women when the fetus is RhD-negative: implementation in the NHS. BJOG 2014 Aug 21 Epub ahead of print.
5. Lim JH, Kim MH, Han YJ, Lee da E, Park SY, Han JY, et al.. Cell-free fetal DNA and cell-free total DNA levels in spontaneous abortion with fetal chromosomal aneuploidy. PLoS One 2013;8:e56787.
6. Wilcox AJ, Weinberg CR, O'Connor JF, Baird DD, Schlatterer JP, Canfield RE, et al.. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189–94.
7. MacDorman MF, Kirmeyer S. Fetal and perinatal mortality, United States, 2005. Natl Vital Stat Rep 2009;57:1–19.
8. Goddijn M, Leschot NJ. Genetic aspects of miscarriage. Baillieres Best Pract Res Clin Obstet Gynaecol 2000;14:855–65.
9. Nagaishi M, Yamamoto T, Iinuma K, Shimomura K, Berend SA, Knops J. Chromosome abnormalities identified in 347 spontaneous abortions collected in Japan. J Obstet Gynaecol Res 2004;30:237–41.
10. Baena N, Guitart M, Ferreres JC, Gabau E, Corona M, Mellado F, et al.. Fetal and placenta chromosome constitution in 237 pregnancy losses. Ann Genet 2001;44:83–8.
11. Reddy UM, Page GP, Saade GR. The role of DNA microarrays in the evaluation of fetal death. Prenat Diagn 2012;32:371–5.
12. Reddy UM, Page GP, Saade GR, Silver RM, Thorsten VR, Parker CB, et al.. Karyotype versus microarray testing for genetic abnormalities after stillbirth. N Engl J Med 2012;367:2185–93.
13. Korteweg FJ, Bouman K, Erwich JJ, Timmer A, Veeger NJ, Ravisé JM, et al.. Cytogenetic analysis after evaluation of 750 fetal deaths: proposal for diagnostic workup. Obstet Gynecol 2008;111:865–74.
14. Management of stillbirth. ACOG Practice Bulletin No. 102. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009;113:748–61.
15. Practice Committee of the American Society for Reproductive Medicine. Evaluation and treatment of recurrent pregnancy loss: a committee opinion. Fertil Steril 2012;98:1103–11.
16. McCullough RM, Almasri EA, Guan X, Geis JA, Hicks SC, Mazloom AR, et al.. Non-invasive prenatal chromosomal aneuploidy testing—clinical experience: 100,000 clinical samples. PLoS One 2014;9:e109173.
17. Palomaki GE, Kloza EM, Lambert-Messerlian GM, Haddow JE, Neveux LM, Ehrich M, et al.. DNA sequencing of maternal plasma to detect Down syndrome: an international clinical validation study. Genet Med 2011;13:913–20.
18. Jensen TJ, Zwiefelhofer T, Tim RC, Džakula Ž, Kim SK, Mazloom AR, et al.. High-throughput massively parallel sequencing for fetal aneuploidy detection from maternal plasma. PLoS One 2013;8:e57381.
19. Alkan C, Kidd JM, Marques-Bonet T, Aksay G, Antonacci F, Hormozdiari F, et al.. Personalized copy number and segmental duplication maps using next-generation sequencing. Nat Genet 2009;41:1061–7.
20. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet 2006;38:904–9.
21. Leek JT, Storey JD. Capturing heterogeneity in gene expression studies by surrogate variable analysis. PLoS Genet 2007;3:1724–35.
22. Zhao C, Tynan J, Ehrich M, Hannum G, McCullough R, Saldivar JS, et al.. Detection of fetal subchromosomal abnormalities by sequencing circulating cell-free DNA from maternal plasma. Clin Chem 2015 Feb 20 Epub ahead of print.
23. Rolnik DL, O'Gorman N, Fiolna M, van den Boom D, Nicolaides KH, Poon LC. Maternal plasma cell-free DNA in the prediction of pre-eclampsia. Ultrasound Obstet Gynecol 2015;45:106–11.
24. Wapner RJ. Genetics of stillbirth. Clin Obstet Gynecol 2010;53:628–34.
25. Hudecova I, Sahota D, Heung MM, Jin Y, Lee WS, Leung TY, et al.. Maternal plasma fetal DNA fractions in pregnancies with low and high risks for fetal chromosomal aneuploidies. PLoS One 2014;9:e88484.
26. Clark-Ganheart CA, Iqbal SN, Brown DL, Black S, Fries MH. Understanding the limitations of circulating cell free fetal DNA: an example of two unique cases. J Clin Gynecol Obstet 2014;3:38–70.
27. Porreco RP, Garite TJ, Maurel K, Marusiak B; Obstetrix Collaborative Research Network, Ehrich M, et al.. Noninvasive prenatal screening for fetal trisomies 21, 18, 13 and the common sex chromosome aneuploidies from maternal blood using massively parallel genomic sequencing of DNA. Am J Obstet Gynecol 2014;211:365.e1–12.
28. Lo YM, Zhang J, Leung TN, Lau TK, Chang AM, Hjelm NM. Rapid clearance of fetal DNA from maternal plasma. Am J Hum Genet 1999;64:218–24.
Detailed Sample Information
Study Diagram and Results Before and After 8 Weeks of Gestation