Clinical Experience with Noninvasive Prenatal Testing in Twin Pregnancy Samples at a Single Center in Germany

Introduction

Since the clinical availability of noninvasive prenatal testing (NIPT) over a decade ago, there has been a vast amount of published data demonstrating the strong performance of NIPT in the detection of a wide range of fetal anomalies including common trisomies,^[1–5] sex chromosomal aneuploidies,^[6,7] and rarer genome-wide fetal anomalies such as rare autosomal aneuploidies and copy number variants.^[8–10] In addition, a large number of professional medical societies are supportive of the use of NIPT for prenatal screening.^[11–16] In Germany, NIPT has been available since 2012 and is currently used in 50 to 75% of pregnancies.^[17] The use of NIPT in Germany is regulated by several professional societies including the German Society of Human Genetics, the German Society for Ultrasound in Medicine (DEGUM), and Fetal Medicine Foundation (FMF) Germany.^[18,19] Since 2022, NIPT has been covered by insurance companies in Germany under certain conditions.^[17]

While it is well accepted that NIPT has improved performance over traditional prenatal screening in singleton pregnancies, fewer publications have focused on the performance of NIPT in twin pregnancies. A systematic review of traditional prenatal screening for trisomy 21 in twin pregnancies using combined nuchal translucency and first trimester serum screening found a pooled sensitivity of 89.3% and pooled specificity of 94.6%.^[20] Prenatal screening by NIPT offers the possibility of a more accurate screening option for these patients. A recent meta-analysis by Judah et al^[21] noted a pooled weighted detection rate of 99.0% and false-positive rate of 0.02% for trisomy 21 in twin pregnancies. The use of NIPT in twin pregnancies is supported by multiple professional societies with some restrictions.^[13–16,22]

We previously reported on the performance of a paired-end sequencing-based NIPT assay (VeriSeq NIPT Solution v2) for common fetal trisomies in a cohort of cases that included both singleton and twin pregnancy samples^[23,24]; results for monosomy X were also provided in these previous studies but for singleton samples only. Here, we focus on the performance of this NIPT assay in a larger cohort of twin pregnancy samples at our laboratory, one of the largest NIPT laboratories in Germany. Results were provided for trisomy 21, trisomy 18, and trisomy 13, with outcome information obtained for a subset of both the high-risk and low-risk NIPT cases.

Materials and Methods

Results

A total of 102,101 pregnancy samples were analyzed using VeriSeq NIPT Solution v2 assay over the study period, of which 1,658 (1.6%) were from twin pregnancies. Two of the twin pregnancy samples failed to give a result upon NIPT analysis, giving a low failure rate of 0.12% (2/1,658). Of the remaining 1,656 cases, 1,625 (98.1%) were found to be low risk and 31 (1.9%) were high risk for presence of a common fetal trisomy following NIPT (►Fig. 1). Demographics for the low-risk, high-risk, and no-results cohorts are provided in ►Table 1. When we compared the low-risk and high-risk cohorts, we found that there was a significant difference for both maternal age (p < 0.001) and body mass index (BMI) (p < 0.0001).

Close

Fig. 1

Study flowchart. FP, false positive; F/u, follow-up; T, trisomy; TP, true positive.

Export to PPT

Overall, there were 20 (1.2%) trisomy 21, 8 (0.5%) trisomy 18, and 3 (0.2%) trisomy 13 high-risk calls in our study cohort; demographics for each of these smaller high-risk cohorts are shown in ►Table 2. When we compared each of these three groups against the low-risk cohort, we found that there was a significant difference between the trisomy 21 cohort and the low-risk cohort for both maternal age (p < 0.001) and BMI (p < 0.05). BMI was also significantly different between the trisomy 18 cohort and the low-risk cohort (p < 0.01). Follow-up information was available for 320 (19.3%) of the twin pregnancy cases in our cohort, including 301 (18.5%) of the 1,625 low-risk cases and 19 (61.3%) of the 31 high-risk cases (►Table 3). All low-risk cases with follow-up (301/301) were determined to be true negatives, giving an estimated negative predictive value (NPV) of 100%. Seventeen (89.5%) of the 19 high-risk cases with follow-up were true positives; there was one false-positive for trisomy 18 and one false-positive for trisomy 13. Overall, for presence of common trisomies in twin pregnancies, we found a sensitivity of > 99.9%, a specificity of 99.3%, and a positive predictive value (PPV) of 89.5% (potential PPV range of 54.8–93.5%). When we looked at performance of the assay for each of the trisomies, we noted high specificities (≥ 99.7%) for all three trisomies, with a sensitivity of > 99.9% for both trisomy 21 and trisomy 18. In addition, our study had a PPV of 100% for trisomy 21 (potential range of 60.0–100%) and 83.3% for trisomy 18 (potential range of 62.5–87.5%) as shown in ►Table 3.

Discussion

In our study, we found that the VeriSeq NIPT Solution v2 assay showed strong performance in the detection of common fetal trisomies in twin pregnancy samples, with high sensitivities, specificities, and PPVs observed based on known clinical outcomes. In addition, our study had a very low failure rate of 0.12%. Some previous studies with NIPT in twin pregnancies have shown higher failure rates, with the most recent Position Statement from the International Society for Prenatal Diagnosis^[22] noting that initial failure rates for NIPT in twin pregnancies across 10 different studies ranged from 1.6 to 13.2%, with a median of 3.6%. A recent multicenter study by van Riel et al^[26] found that initial failure rates ranged from 0 to 11.7% among the different genetic centers, which could be reduced to an overall rate of 1.2% after resampling.

As noted earlier, first trimester combined screening in twin pregnancies has been shown to have a relatively low detection rate of 89.3% and a high false-positive rate of 5.4%.^[20] As patients with high-risk screening results are typically counseled to undergo confirmatory diagnostic testing, a high false-positive rate can lead to unnecessary invasive diagnostic procedures which may have a higher risk of loss in twin pregnancies.^[27] As our study has shown, NIPT offers patients with twin pregnancies a more accurate approach to screen for fetal trisomies, in particular for trisomy 21. Based on known clinical outcomes, we observed both a sensitivity and specificity of > 99.9% for trisomy 21. A meta-analysis by Liao et al^[28] also found that NIPT has both high sensitivity and specificity for trisomy 21 screening in twin pregnancies. In addition, our study had a high overall PPV of 89.5%, with a PPV of 100% for trisomy 21 and 83.3% for trisomy 18. This is similar to that observed in a recent study which found that a single-nucleotide polymorphism-based NIPT approach in twin pregnancies gave a PPV of 88.7% for trisomy 21 and 72.7% for trisomy 18.^[29] For patients with twin gestations, it is very important that they receive appropriate counseling prior to prenatal screening and that they are made aware of the limitations of each of the different types of prenatal screening options and also the need for diagnostic confirmation of any detected aneuploidy.

One of the strengths of our study was that it involved a large number of twin samples from a general pregnancy population, allowing us to provide evidence on the performance of the NIPT assay in that patient population. Another strength is that the study was carried out at a single prenatal center, which is one of the largest labs in Germany, where we have over 10 years of experience with NIPT. A limitation of our study was that follow-up was not available for all samples in our cohort, particularly for the low-risk cases. The limited follow-up data on the low-risk cases prevented us from providing a true calculation of the NPV. Based on the 18.5% of these cases that did have clinical outcomes available, we estimated the NPV to be 100%. It is possible that there were some false-negative cases in our cohort that we were not made aware of, although discrepancies between a NIPT result and the birth report are typically reported back to our laboratory immediately. We were, however, able to obtain follow-up on a majority (61.3%) of the high-risk samples. Future steps will include obtaining diagnostic testing outcomes or birth outcomes on all study samples.

Another limitation of our study was the lack of zygosity data for our twin samples. Overall, about one-third of twin pregnancies are monozygotic and two-thirds are dizygotic.^[30] NIPT is more complex in dizygotic pregnancies as the two fetuses have different genotypes but typically only one fetus (if any) will have an aneuploidy.^[31] In addition, although the overall fetal fraction is higher in twin pregnancies, the individual contribution per fetus is lower and each fetus can contribute different amounts of fetal fraction.^[31–35] This can lead to a decreased performance of NIPT in dizygotic pregnancies. A recent study by Kantor et al noted that the PPVs of NIPT for common trisomies were lower in dizygotic compared to monozygotic pregnancies.^[29] As monozygotic twins have the same genotype except in rare cases, and with the higher overall fetal fraction observed in twin pregnancies, the performance of NIPT in monozygotic pregnancies will be at least equivalent to that observed in singleton pregnancies.^[31,34] Therefore, the higher the proportion of monozygotic twins in a twin cohort, the closer the accuracy of the assay will be to that of singletons.

In conclusion, based on the clinical outcomes available at this time, we determined that the VeriSeq NIPT Solution v2 assay allows for screening of fetal chromosomal anomalies in twin pregnancies with a high overall sensitivity, specificity, and PPV, along with a very low failure rate.