Factors affecting low fetal fraction in fetal screening with cell-free DNA in pregnant women: a systematic review and meta-analysis

Cell-Free DNA (cfDNA) is a non-invasive perinatal test (NIPT) used to assess fetal anomalies such as aneuploidies. cfDNA is assessed using maternal blood sampling [1]. Cell-free DNA originates from fetal trophoblasts [2], and its fragment size is smaller than that of maternal DNA [1]. Approximately 11 to 13.4% of Cell-Free DNA in maternal blood is of embryonic origin [3], which appears in the mother’s blood within the 5 to 7 weeks of pregnancy. The amount of cfDNA increases with gestational age and tends to decrease after delivery. It would be cleared within two hours from the mother’s blood after delivery [4]. Compared to standard invasive screening techniques such as Chorionic Villus Sampling (CVS) and amniocentesis, Cell-Free DNA testing is non-invasive, easy to perform, and has no risk of miscarriage [5]. It is recommended to perform Cell-Free DNA in cases of age over 35, positive aneuploidy screening in the presence of increased nuchal translucency, abnormal ultrasound findings, and positive personal or familial history of aneuploidy [6]. cfDNA is a mixture of maternal and fetal cfDNA, and the ability to detect fetal chromosomal aneuploidies is directly related to the fetal to total DNA fraction of a sample. This ratio is the fetal fraction (FF) [7]. If the FF is too small, any abnormalities in the fetal cfDNA will be masked by the overwhelming proportion of euploid maternal cfDNA, thereby making their detection impractical [8]. Recent studies have consistently shown that the average FF is around 10–15% but can range up to 30% or more. The minimum FF is considered 4%, and the test result below 4% is uncertain [9]. Overestimated FF would lead to false-negative results, while underestimated FF may cause the rejection of suitable samples. According to studies, several factors such as body mass index (BMI), gestational age, twin pregnancy, and pregnancy biomarkers in maternal serum affect the amount of Cell-Free DNA and interrupt the interpretation of FF. Pregnancy screening tests impose huge costs on the people and the government, and Incorrect interpretation of results may lead to potentially inappropriate medical decisions [10]. Given the importance of cfDNA tests in prenatal screening and ambiguity around factors affecting LFF, we decided to conduct a systematic and meta-analysis study on studies retrieved through a rigorous search and selection process to detect the factors causing LFF that lead to false results.

The review was conducted according to the PRISMA statement [11]. In addition, a protocol was designed before initiating the search.

Begg’s funnel plots and Egger’s test were applied to assess the potential publication bias (Fig. 3).

No publication bias was detected for any of the conducted meta-analyses. Since only two studies were available for the maternal weight variable according to the results, a publication bias assessment was impossible (maternal age P-value = 0.124, gestational age P-value = 0.447, BMI P-value = 0.903).

cfDNA testing is one of the NIPT tests used to assess aneuploidies in the fetus and is widely used for routine invasive tests [1]. The mother’s bloodstream cDNA consists of maternal and embryonic types. For the test results to be conclusive, there should be a certain minimum amount of embryonic cfDNA. This value might vary based on the technique and kit used in different laboratories; however, a 4% threshold of the FF is considered sufficient [9]. This study is the first meta-analysis that evaluates factors affecting the reduction of FF ratios to less than 4% in the mother’s blood.

Factors that could alter the FF ratio are generally divided into three categories: maternal, fetal-placental, and experimental [23].

Maternal factors mentioned in the literature include maternal age, maternal weight, gestational age, race [14, 15], in vitro fertilization (IVF) [20], consumption of certain drugs [12, 19], and maternal diseases, especially autoimmune diseases [24]. Some studies have reported free β-subunit of human chorionic gonadotropin (free β-hCG) and serum pregnancy-associated plasma protein (PAPP-A) levels to be positively correlated with FF [25, 26].

Several fetal-placental factors other than gestational age have also been reported to alter FF. these include crown-rump length [18, 25], gender [26], twin pregnancies [27, 28], preterm birth [29, 30], and aneuploidies [31, 32]. Experimental causes related to lab procedures include cfDNA fragment size, cfDNA concentration, library concentration, and uniquely mapped reads [33].

We identified four potential factors with enough data for meta-analysis (maternal age, gestational age, maternal weight, and BMI) that could cause LFF. The meta-analysis identified an association between gestational age and LFF; however, there was no significant relationship between the other three named factors and LFF. When the substantial heterogeneity in 95% prediction intervals was accounted for, the results indicated that the association between LFF and maternal age, maternal weight, and BMI became insignificant. These results do not necessarily indicate that there is no impact of named factors on LFF; however, the results do indicate that there is still substantial uncertainty about the significance of the association.

Five studies [3, 12, 14, 16, 17] discussed gestational age’s effect on LFF [3, 14, 16, 18, 22]. The mean gestational age at Burns et al. (2017) [12] study was reported as 14.3 weeks, and at Krishna et al. (2016) [14] was 16.4 weeks, which was irrelevant with LFF. In Nakamura et al. (2020) [16] study, the gestational age was 12.1 weeks, Wang et al. (2013) [3] was 13.9 weeks, and Zhao et al. (2019) [17] was 14.43 weeks, and in these three studies, it was related to a LFF.

Five studies [12‐14, 16] discussed maternal age’s effect on the LFF. The mean maternal age in Burns et al. (2017) [12] study was 37.3 years, and in Miltoft et al. (2020) [15] was 33.3 years. Krishna et al. (2016) [14] reported 22 to 46 years, while Nakamura et al. (2020) [16] reported 34.8 to 43.4 years, All irrelevant to LFF. The mean maternal age of Dabi et al. (2018) [13] was 34 and related to LFF.

Two studies [3, 13] discussed the effect of maternal weight on the LFF [3‐11, 19]. In Dabi et al. (2018) [13], the maternal weight range was (60-90.3 kg), and the mean was 77 kg. The mean maternal weight in Wang et al. (2013) [3] study was 103 kg, which was related to the LFF in these two studies.

Maternal weight meta-analysis heterogeneity might be due to different gestational ages chosen for inclusion criteria within the studies. The gestational age range in Dabi et al. (2018) [13] study was 11.1 to 27.5 weeks, whereas it was 11.1 to 40.43 weeks in Wang et al. (2013) [3] study. Since there is a constant increase in maternal weight as the pregnancy progresses, the difference in the gestational age range might be the clinical cause of heterogeneity. Of all the studies, Wang et al. (2013) [3] had the broadest range of gestational age and the largest sample size (22,384 pregnant women); we speculate that the uneven distribution of gestational age within the selected studies for gestational age meta-analysis might be the cause of heterogeneity.

Six studies [12‐15, 17] discussed the BMI effect on LFF [13, 14, 16, 18, 19, 22]. The mean BMI was 34.6 in Burns et al.  (2017) [12] study and 29 in Dabi et al. (2018) [13] study and 36.5 in Krishna et al. (2016) [14] study, and 19.5 in Miltoft et al. (2020) [15] study and 23.1 at Zhao et al. (2019) [17] study and it was related to the LFF in these five studies. The mean BMI was 20.9 in Nakamura et al. (2020) [16] study, which was irrelevant.

Other implications for LFF in prenatal care have been suggested as well. In a systematic review by Scheffer et al. published in 2021, LFF in cfDNA testing was associated with adverse pregnancy outcomes, particularly pregnancy-related hypertensive disorders, preterm delivery, and impaired fetal growth [34]. Shree et al. concluded that in mothers with BMI < 30 kg/m2, those with hypertensive disorders of pregnancy have lower fetal fraction; however, obesity affected LFF in such a way that it could not be used for predicting hypertension in obese individuals [35].

Further and more extensive studies are needed to investigate possible factors affecting FF thoroughly. In addition, comprehensive strategies can be developed to reduce the chance of encountering LFF by designing tailored and beneficial interventions targeting modifiable factors.

This review had some limitations that are worth mentioning: the relatively small number of articles included in this review, only studies between the years 2000 and 2021 were examined, only the most prevalent factors with potential effects were included in this study, and finally, only English and Persian articles were examined.

Along with its limitations, this study had noteworthy strength points. The topmost advantage of this systematic review was the low risk of subjective data selection. Predefined criteria guided the search process, quality assessment, and data synthesis and two independent reviewers performed those using well-established tools. This study was the first comprehensive systematic review and meta-analysis that evaluated the factors affecting LFF in fetal screening with Cell-Free DNA in pregnant women.

A number of factors were reported to have a potential effect on the amount of FF such as maternal age, maternal BMI, maternal weight, and gestational age. However, after meta-analyses of the mentioned factors, only gestational age significantly affected the amount of FF in the cfDNA tests. Lower gestational age is significantly associated with LFF. Thus, this factor needs to be considered in interpreting the prenatal cfDNA screening tests to make a more accurate interpretation.