Dhallan R, Au W, Mattagajasingh S, Emche S, Bayliss P, Damewood M, Cronin M, Chou V, Mohr M. Methods to Increase the Percentage of Free Fetal DNA Recovered From the Maternal Circulation. JAMA. 2004;291(9):1114–1119. doi:10.1001/jama.291.9.1114
Author Affiliations: Ravgen Inc, Columbia, Md (Drs Dhallan, Au, Mattagajasingh, Emche, and Cronin and Mss Chou and Mohr); Department of Obstetrics and Gynecology, York Hospital, York, Pa (Drs Bayliss and Damewood). Dr Bayliss currently is affiliated with the Department of Maternal Fetal Medicine, Lancaster General Women & Babies Hospital, Lancaster, Pa.
Context Noninvasive prenatal diagnostic tests using free fetal DNA provide an
alternative to invasive tests and their attendant risks; however, free fetal
DNA exists in the maternal circulation at low percentages, which has hindered
development of noninvasive tests.
Objective To test the hypothesis that using formaldehyde to reduce cell lysis
could increase the relative percentage of free fetal DNA in samples of maternal
Design, Setting, and Patients The first phase of the study was conducted from January through February
2002 at a single US clinical site; 2 samples of blood were collected from
each of 10 pregnant women, and the percentage of free fetal DNA in formaldehyde-treated
and untreated samples was determined. The second phase of the study was conducted
from March 2002 through May 2003, and measured the percentage of free fetal
DNA in 69 formaldehyde-treated samples of maternal blood obtained from a network
of 27 US clinical sites in 16 states.
Main Outcome Measure Percentage of free fetal DNA in samples of maternal blood.
Results In the first phase of the study, the mean percentage of free fetal DNA
in the untreated samples was 7.7% (range, 0.32%-40%), while the mean percentage
of free fetal DNA in the formaldehyde-treated samples was 20.2% (range, 1.6%-40%)
(P = .02 for difference). In the second phase, a
median of 25% (range, 3.1% to >50%) free fetal DNA was obtained for the 69
formaldehyde-treated maternal blood samples. Approximately 59% of the samples
in this study had 25% or greater fetal DNA, and only 16% of the samples had
less than 10% fetal DNA. In addition, 27.5% of the samples in this study had
50% or greater fetal DNA.
Conclusion Addition of formaldehyde to maternal blood samples, coupled with careful
processing protocols, increases the relative percentage of free fetal DNA,
providing a foundation for development of noninvasive prenatal diagnostic
tests to distinguish fetal DNA from maternal DNA in the maternal circulation.
Prenatal diagnosis is useful for managing a pregnancy with an identified
fetal abnormality and may allow for planning and coordinating care during
delivery and the neonatal period.1 A variety
of prenatal diagnostic tests are available but have limitations. Noninvasive
tests such as maternal serum marker testing and ultrasound can be used to
screen for the presence of chromosomal abnormalities but are not definitive.2- 5 On the
other hand, invasive diagnostic tests (eg, amniocentesis, chorionic villus
sampling, percutaneous umbilical blood sampling) for fetal chromosomal abnormalities
are highly reliable, but the procedure used for each test carries a risk for
loss of pregnancy.6,7 Many patients
who are candidates for these tests decline them because of the risk of pregnancy
An alternative to existing methods for prenatal diagnosis is to use
fetal cells and fetal DNA that exist in the maternal circulation.8- 15 Circulating
fetal DNA has been used to determine the sex of the fetus through detection
of sequences present on the Y chromosome.13 In
addition, several studies have attempted to use free fetal DNA to screen for
chromosomal abnormalities in the fetus.16- 21 However,
the use of free fetal DNA for detecting chromosomal abnormalities has been
limited by the seemingly low percentage of free fetal DNA in the maternal
circulation. Lo et al13 reported a mean of
3.4% free fetal DNA in maternal plasma in the late first to the mid second
trimester and a mean of 6.2% free fetal DNA in the late third trimester. Any
method that can increase the relative percentage of free fetal DNA in the
sample would make it easier to distinguish fetal DNA from maternal DNA. Noninvasive
prenatal diagnostic tests that are DNA-based would benefit from higher percentages
of free fetal DNA in the samples.
We hypothesized that inhibiting cell lysis during sample collection,
shipping, handling, and processing would permit the recovery of a larger percentage
of free fetal DNA. By decreasing the amount of maternal cell lysis, and thus
the amount of free maternal DNA, the relative percentage of free fetal DNA
likely can be increased.
Blood samples were collected from women carrying a male or a female
fetus; however, the majority of samples (81 of 85 [95.3%]) analyzed were obtained
from women carrying a male fetus. The Y chromosome is the accepted marker
for quantitating percentages of fetal DNA. Each clinical site received institutional
review board approval for participation in this research protocol. All women
were aged 18 years or older, had a singleton pregnancy, and provided written
informed consent prior to enrollment.
First Phase. The first phase of this study
was conducted from January through February 2002, and recruited women pregnant
with a male fetus (identified by ultrasound). Blood samples were collected
from the women at a single clinical site prior to an amniocentesis procedure.
Two tubes of blood (9 mL in each tube) were collected from each of 10 women.
One tube was treated with 0.225 mL of a 10% neutral buffered solution containing
formaldehyde (4% weight per volume) (Sigma, St Louis, Mo), a chemical that
stabilizes cell membranes and impedes cell lysis. The other tube was left
untreated. The tubes were assigned a numerical code and hand-delivered to
our facility for analysis. Laboratory personnel were blinded as to which specimens
Second Phase. For the second phase of the study,
conducted from March 2002 through May 2003, a network of 27 clinical sites,
operating in 16 states in the United States, was established to collect formaldehyde-treated
blood samples from pregnant women prior to amniocentesis or chorionic villus
sampling. Seventy-one samples from women carrying a male fetus and 4 samples
from women carrying a female fetus were analyzed in this phase of the study.
Fetal sex was confirmed by amniocentesis or chorionic villus sampling report.
Each sample was coded so that laboratory personnel did not know whether it
was obtained from a woman carrying a male or a female fetus.
All samples collected in the second phase of the study were treated
with formaldehyde. The clinical sites were provided with a kit used for the
venipuncture procedure, which included 21-gauge needles, 9-mL EDTA blood collection
tubes, a syringe for each tube containing 0.225 mL of a 10% neutral buffered
solution containing formaldehyde (4% weight per volume), an ice pack, and
a shipping container. The clinical sites were instructed to add the formaldehyde
to the tubes and gently invert them immediately after blood was drawn. The
specimens were shipped by commercial carrier for overnight delivery to our
The protocols for isolation of plasma were optimized to reduce cell
lysis. Tubes were centrifuged at 200g for 10 minutes
with the brake and acceleration powers set to zero. Tubes then were centrifuged
at 1600g for 10 minutes with the brake and acceleration
powers set to zero. The supernatant (ie, the plasma) of each sample was transferred
to a new tube and spun at 1600g for 10 minutes with
the brake and acceleration powers set to zero. The plasma was transferred
carefully to a new tube and stored at −80°C. Approximately 0.5 mL
of supernatant was left in the tube to ensure that the buffy coat was not
DNA was isolated from plasma samples using the QIAamp DNA Blood Midi
Kit (Qiagen, Valencia, Calif) for purification of DNA from blood cells, according
to the manufacturer's instructions. DNA was eluted in 100 µL of distilled
Two sets of primers were used: 1 set amplified the sex-determining region
Y gene (SRY), which is located on the Y chromosome
and is thus representative of fetal DNA, and the other set amplified the cystic
fibrosis gene (CYS), which is present on both maternal
template DNA and fetal template DNA. Unique regions of the SRY gene and the CYS gene were identified
by sequence searches using the Blast program available from the National Center
for Biotechnology Information (http://www.ncbi.nlm.nih.gov).
The following primers were designed to amplify the SRY gene: upstream primer: 5′ TGGCGATTAAGTCAAATTCGC 3′;
downstream primer: 5′ CCCCCTAGTACCCTGACAATGTATT 3′. The following
primers were designed to amplify the CYS gene: upstream
primer: 5′ CTGTTCTGTGATATTATGTGTGGT 3′; downstream primer: 5′
The SRY gene and the CYS gene were amplified from plasma free DNA by polymerase chain reaction
(PCR) using the HotStarTaq Master Mix kit (Qiagen). Each PCR reaction used
8 µL of template DNA (diluted or undiluted), 1 µL of each primer
(5 µM), and 10 µL of HotStarTaq mix. The following PCR conditions
were used: (1) 95°C for 15 minutes, (2) 94°C for 30 seconds, (3) 54°C
for 15 seconds, (4) 72°C for 30 seconds, (5) repeat steps 2 through 4
for 45 cycles, and (6) 72°C for 10 minutes.
The percentage of free fetal DNA in the maternal plasma sample was determined
by PCR using serially diluted plasma DNA, which accurately quantifies the
number of genomes that harbor the amplified gene. For example, if the blood
sample contains 100% male fetal DNA, and 1:2 serial dilutions are performed,
then on average the SRY signal will disappear 1 dilution
before the CYS signal, since there is 1 copy of the SRY gene and 2 copies of the CYS gene.
The percentage of free fetal DNA in the maternal plasma was calculated
using the following formula: percentage of free fetal DNA = (No. of copies
of SRY gene × 2 × 100)/(No. of copies
of CYS gene), where the number of copies of each
gene was determined by observing the highest serial dilution in which the
gene was detected. The formula contains a multiplication factor of 2, which
is used to normalize for the fact that there is only 1 copy of the SRY gene.
In the first phase of the study, 6 serial dilutions (1:5) were performed
for each sample (1:5 to 1:15625). Provided that there is at least 1 copy of
the SRY gene and that the CYS gene
is detected in the sixth serial dilution, the lowest possible value is 0.0128%
free fetal DNA ([1 × 2 × 100]/15625). All other values increase
by multiples of 5 from 0.0128, depending on the number of dilutions that were
positive for the SRY and CYS genes
(eg, 0.064, 0.32, 1.6, 8, and 40).
In the second phase of the study, 10 serial dilutions (1:2) were performed
for each sample (1:2 to 1:1024). Provided that there is at least 1 copy of
the SRY gene and that the CYS gene
is detected in the 10th serial dilution, the lowest possible value is 0.1953%
free fetal DNA ([1 × 2 × 100]/1024). All other values increase
by multiples of 2 from 0.1953.
The number of fetal genomes per milliliter of plasma was calculated
using the following formula: No. of genomes/mL of plasma = (No. of copies
of SRY gene/volume of DNA in reaction [µL])
× (volume of DNA eluted [µL]/total volume of plasma through column
The primary outcome in the first phase of the study was the difference
in the percentage of free fetal DNA between formaldehyde-treated and untreated
samples. The nonparametric Wilcoxon signed rank test, which assumes that there
is information in the magnitude of differences, was used to analyze the data
from the first phase of the study. All analyses were performed using Analyse-it
General & Clinical Laboratory Statistics, version 1.71 (Analyse-it Software
Ltd, Leeds, England); P<.05 was used to determine
The results from the first phase of the study are summarized in Table 1. Analysis of the untreated samples
revealed a mean of 7.7% (range, 0.32%-40%) free fetal DNA. The formaldehyde-treated
samples had a mean of 20.2% (range, 1.6%-40%) free fetal DNA.
Several of the untreated samples (from participants 1, 3, 4, and 10)
contained percentages of free fetal DNA that were substantially below the
median. Even with these samples, the addition of formaldehyde increased the
relative percentage of free fetal DNA. For instance, in untreated samples
from participants 4 and 10, the percentage of free fetal DNA was 0.32%, whereas
in foraldehyde-treated samples collected from the same women the percentage
of free fetal DNA was increased to 8%.
In 3 of the samples (from participants 2, 6, and 7), there was no measurable
effect of formaldehyde on the percentage of free fetal DNA. However, analysis
of the paired samples from the first phase of the study using the Wilcoxon
signed rank test revealed that, overall, the addition of formaldehyde significantly
increased the percentage of free fetal DNA (P = .02;
W = 28 for y = 7).
Because analysis of the samples from the first phase of the study revealed
that the effect of formaldehyde was statistically significant, the second
phase of the study was designed to evaluate the percentage of free fetal DNA
in formaldehyde-treated samples in a larger patient population from multiple
clinical sites. A total of 75 samples from pregnant women were collected in
this phase of the study. Seventy-one samples were collected from women who
carried a male fetus. However, 2 samples were excluded from analysis because
these samples were not received within 24 hours after collection.
Four samples were obtained from women who carried a female fetus. For
each of these samples, a robust signal was observed for the CYS gene; as expected, no signal was detected for the SRY gene, which is specific for the Y chromosome.
Analysis of the 69 formaldehyde-treated samples revealed a median of
25% (range, 3.1% to >50%) free fetal DNA (Table 2). Approximately 16.0% of the samples (11/69) had less than
10% free fetal DNA; approximately 59% of the formaldehyde-treated samples
had 25% or greater free fetal DNA and 27.5% of the samples had 50% or greater
free fetal DNA (4 [5.8% of total] samples had 3.1% free fetal DNA; 7 [10.1%]
had 6.2%; 17 [24.6%] had 12.5; 22 [31.9%] had 25%; 8 [11.6%] had 50%; and
11 [15.9%] had >50%). Although many of the formaldehyde-treated samples contained
high percentages of free fetal DNA, there was variability in the percentages
Analysis of the formaldehyde-treated samples also revealed a mean of
66.1 fetal genomes/mL of plasma, with a range of 3.0 fetal genomes/mL to 533
fetal genomes/mL (Table 2). Some
of the samples (eg, sample 12) had a limited number of fetal genomes but a
high percentage of fetal DNA. Conversely, some samples had an ample number
of fetal genomes but a lower percentage of fetal DNA. For instance, sample
40 had 112.5 fetal genomes/mL but the percentage of fetal DNA was 6.2% (Table 2).
We have shown that the relative percentage of free fetal DNA recovered
from maternal blood samples can be increased. Addition of formaldehyde to
maternal blood samples, coupled with careful processing protocols, resulted
in an increase in the percentage of free fetal DNA recovered from the maternal
circulation. This increase in the relative percentage of free fetal DNA likely
resulted from a combination of factors.
First, formaldehyde stabilizes cell membranes, thereby preventing cell
lysis and the release of DNA. Prior to the venipuncture procedure, the amount
of free maternal DNA in the maternal circulation likely is low. However, the
maternal cells may lyse during sample collection, shipping, handling, and
processing. For example, during centrifugation the cells are exposed to gravitational
forces, which may rupture cells. The presence of formaldehyde protects the
cells from lysis.
Second, the addition of formaldehyde may allow a larger recovery of
free fetal DNA by inhibiting enzymes that destroy DNA, such as DNases. For
the samples analyzed in the second phase of the study, a mean of 66.1 fetal
genomes/mL was obtained, which represents a 2.6-fold increase over the mean
reported in the literature (25.4 fetal genomes/mL).13 Inhibition
of enzymes that destroy DNA would permit a larger recovery of DNA (including
free fetal DNA) already present in the sample. Also, the addition of formaldehyde
may stabilize and preserve the structure of DNA, which may increase the amount
of DNA recovered.
Third, in conjunction with the addition of formaldehyde to the maternal
blood samples, sample-processing protocols designed to minimize cell lysis
likely contributed to increases in the percentage of free fetal DNA. A centrifugation
protocol was designed to minimize gravitational forces imposed on the cells.
Samples initially were spun at a low speed, which allowed the majority of
cells to separate from the plasma under minimal forces. In addition, all centrifugation
steps were performed with the acceleration and brake powers set to zero. This
reduced the formation of a vortex during centrifugation, minimizing mixing
of the plasma and the buffy coat, which contains maternal cellular material.
Also, when removing the plasma sample, care was taken to ensure that the buffy
coat was not disturbed.
A larger study would be useful to delineate factors contributing to
the results presented in this study, and to understand why some samples have
a higher percentage of free fetal DNA than others. For instance, in 3 samples
from the first phase of the study, there was no measurable effect of formaldehyde
on the percentage of free fetal DNA. It is possible that formaldehyde was
not added or mixed properly with the treated samples or may have been added
to both treated and untreated samples.
Also, in some samples there may be a greater amount of free maternal
DNA already present in the maternal circulation. While the addition of formaldehyde
will impede cell lysis that occurs during sample collection, shipping, handling,
and processing, it likely will not reduce the concentration of free maternal
DNA already present in the sample. A larger study comparing the percentage
of free fetal DNA in formaldehyde-treated and untreated samples will help
to address these issues.
Furthermore, several controllable factors likely contribute to the variation
in the percentages of free fetal DNA. One such factor is the time interval
between the venipuncture procedure and sample processing. The shorter this
time interval the more likely it is the integrity of the sample will be preserved
and cell lysis kept to a minimum. Similarly, the length of time between the
venipuncture procedure and the addition of formaldehyde is thought to be critical.
If formaldehyde is not added shortly after the venipuncture procedure, cells
may lyse and release DNA. Also, it is important that the formaldehyde is gently
mixed throughout the tube to allow maximum exposure to the reagent.
With an increased percentage of free fetal DNA in the maternal blood
samples, the sequence of fetal DNA can be discerned from maternal DNA using
natural genetic markers, such as single nucleotide polymorphisms. For example,
at certain genomic sites, the maternal genome will be homozygous for allele
A, while the paternal genome is homozygous for allele B, which means the fetal
genome will be heterozygous at this genomic site. Allele B represents a distinct
fetal signal in the maternal blood sample. The detection and quantitation
of fetal DNA, in this case allele B, is more attainable with an increased
percentage of fetal DNA, and can be used to diagnose single-gene disorders
and chromosomal abnormalities.
A ratio for alleles A and B can be quantitated and used to detect chromosomal
disorders. When samples have a high percentage of free fetal DNA, the difference
between the expected ratio of the chromosomes for a healthy fetus and that
for an abnormal fetus is greater, which makes it easier to diagnose chromosomal
abnormalities. Thus, the methods described herein for increasing the percentage
of free fetal DNA provide a solid foundation for the development of a noninvasive
prenatal diagnostic test.