Maternal Plasma DNA Testing

Materials and Methods

On November 14, 2011, the Prenatal Diagnosis Center (PDC) at Women & Infants Hospital of Rhode Island began offering ccfDNA testing as an option to all women presenting for genetic counseling due to increased risk of Down syndrome (or trisomy 18/trisomy 13). The PDC is the primary referral site serving high-risk obstetrical patients in Rhode Island and, therefore, the women counseled are representative of the statewide population. Before the initiation of the study, the seven certified genetic counselors on the staff met with the Clinical Site Director and Director of Perinatal Genetics to standardize their counseling approach to the ccfDNA test offered to high-risk patients. The standard approach in these sessions was for the counselor to discuss all testing relevant to aneuploidy, including serum screening, ultrasound, CVS and/or amniocentesis, and karyotyping, as well as ccfDNA, as appropriate to the situation. Written materials for ccfDNA testing in the high-risk setting were not yet available. Other testing (i.e., for single gene disorders or microdeletion/duplication syndromes via microarray) was also discussed, if clinically indicated (e.g., certain ultrasound findings).

Patients were told that ccfDNA has a higher detection rate for trisomy 21 and trisomy 18, with a lower FPR than conventional serum/ultrasound screening tests. Specific numbers quoted for the detection rate of ccfDNA testing for the less common chromosomal abnormalities were near 98% for Down syndrome, up to 100% for trisomy 18 (among samples with a reported result), and 92% for trisomy 13 (Palomaki et al., 2011, 2012). At the time these patients were being counseled, these were the only externally validated data available. Patients were also told that false-positive results could occur and that only an invasive procedure followed by karyotyping would be definitive. They were cautioned that ccfDNA does not screen for open neural tube defects and, therefore, second trimester maternal serum AFP testing was discussed. Women who chose ccfDNA were given the option of an ultrasound measurement of the nuchal translucency, if the gestational age was under 14-weeks gestation to rule out cystic hygroma or other identifiable structural anomalies.

The Women & Infants Institutional Review Board approved a retrospective medical record review for the cohort of women attending the PDC during the 10-week period between November 14, 2011, and January 20, 2012, and later extended the review period to include a second 10-week period (April 30 to July 6, 2012). Eligible women were at least 18 years old and candidates for prenatal diagnosis due to increased risk of aneuploidy. If the patients were aged 35 or older at delivery, they also had the option of a level II ultrasound and/or serum-based screening tests. In all women, therefore, ccfDNA would be a secondary screening test.

The record review did not depend on the testing options chosen. A standardized datasheet was designed and used to capture information from existing records, including the maternal age at delivery (completed year), gestational age, indication for referral, aneuploidy risk, familiarity with ccfDNA testing before counseling, testing decision-making and reasoning, and type of insurance coverage. Standard chart notes were used to document the reasoning behind patients' test decisions. Patients' reasoning for their test decisions were determined by a standard genetic counseling technique. If a patient provided a reason for declining ccfDNA testing, the genetic counselors did not pursue this issue further. Insurance coverage was used to estimate patient out-of-pocket expenses. Since only one laboratory was offering ccfDNA testing at this time, patients were not influenced by competitors' marketing/advertising. The only alternative sources of information about ccfDNA testing were media reports or personnel at the referring physician's office.

Categorical variables were compared using Fisher's exact test or the log-likelihood ratio. Continuous variables were compared using the t-test, after appropriate transformations. Significance was two-tailed at the p=0.05 level.

Results

Indications for testing and test acceptance by indication

Table 1 shows the indications for genetic counseling referral. In the first time period, the only indication for diagnostic testing in 69 women (53%) was advanced maternal age (35 years or older), while 25 women (19%) were 35 years or older and also had an abnormal ultrasound or positive serum screening test. Among the remaining women, 22 (17%) had a positive serum screening test and 13 (10%) an abnormal ultrasound finding that increased a patient's risk for aneuploidy.

OPEN IN VIEWER

Table 1. Maternal Plasma ccfDNA Testing by Indication for Referral for the Two Study Cohorts, Stratified by Expected Out-of-Pocket Cost

There was a similar distribution of these four indications in women seen during the second time period (59%, 15%, 13%, and 14%, respectively). Table 1 also shows the percentage of women accepting the offer of ccfDNA testing according to indication for referral. In the first time period, the rates of acceptance ranged from a low of 9% among women aged 35 and older to 23% among those with a positive serum screening test. In the second time period, the acceptance rate among AMA women doubled to 18%, but the rates for other indications remained constant. The overall acceptance rates for the two time periods were 13% and 18%, respectively (p=0.27). There were no statistically significant differences in acceptance by indication (G2=1.77, p=0.62).

Reasons provided by the women for selecting or declining DNA testing

Table 2 compares the women's reported reasons underlying the ccfDNA test decision-making. Among those women choosing testing, the desire to avoid the risk of an invasive procedure was cited by three-quarters of the women in each cohort. The most commonly cited reason for declining testing was that no further testing was desired (33% and 57%, respectively). Many of those women cited that additional testing “wouldn't change anything” about how they managed their pregnancy. The multiple reasons for declining the test were not pursued. The percentage of women choosing an invasive procedure decreased from 21% to 12%, a 43% decrease. For comparison, there were 39 invasive procedures (14 CVS and 25 amniocenteses) performed in the 3 months before cohort 1, but only 25 procedures (13 CVS and 12 amniocenteses) in the 3 months after cohort 2. This represents a 36% reduction in any invasive procedure, but there was a 52% reduction in amniocenteses.

OPEN IN VIEWER

Table 2. Primary Reasons for Choosing or Declining ccfDNA Testing Among High-Risk Women in the Two Cohorts

About one in five women (many aged 35 and older without any other screening) decided to have conventional serum/ultrasound screening before making a decision about ccfDNA or invasive testing (Table 2). All of those results were screen negative and none of those women chose further testing. Approximately 12% of all women declining ccfDNA testing cited cost as their reason. An average of 17% of each cohort did not provide a reason for their decisions (Table 2, last column).

Results of Plasma DNA Testing

A total of 42 women completed ccfDNA testing by the end of the second trimester; all received reports that were negative for trisomy 21, 18, and 13. Interpretations for the sex chromosomes were not available at that time. The median turnaround time in the first group of 13 samples was 9 days (range 6–13 days). During the second time period, turnaround times were slightly longer (median 12 days, range 7–18 days).

One woman who initially declined ccfDNA testing after counseling reversed her decision at 33 weeks of gestation following an ultrasound study that identified duodenal atresia. This was the only case associated with an aneuploidy (Down syndrome) among women in this study. The turnaround time for this sample was 14 days. The diagnosis was confirmed on products of conception (cord and placenta) after a spontaneous fetal demise at 35 weeks of gestation.

Discussion

Other studies have examined ccfDNA test uptake in the high-risk setting, either through preference surveys (Hill et al., 2012; Carroll et al., 2013) or by experience (Chetty et al., 2013; Taylor et al., 2014; Vahanian et al., 2014). These latter three studies (Chetty et al., 2013; Taylor et al., 2014; Vahanian et al., 2014) provide extensive information regarding how factors such as insurance and race influence patient uptake, but they did not survey women on the reasons they declined (Taylor et al., 2014). To our knowledge, ours is the first report to address the underlying reasons patients provide for their decision-making. We documented that most women were unfamiliar with the availability of ccfDNA testing for aneuploidy before their counseling session, both at the time testing first introduced in late 2011 and several months later (4% vs. 14%). In the later time period, three women had already initiated ccfDNA testing at their primary prenatal care office before being referred to the PDC for counseling, an action that was not observed earlier. The offer of ccfDNA testing to the women at both these time periods preceded the publication of the December 2012 committee opinion from the American College of Obstetricians and Gynecologists (ACOG, 2012). That opinion supported the offering of ccfDNA testing in women at high risk of fetal aneuploidy as a secondary screening test. The increase seen in test uptake between the first and second time periods of study likely reflected an increasing awareness that ccfDNA testing can reduce referral for invasive diagnostic procedures associated with discomfort and risk (Hill et al., 2012). Our study supported this notion, with a 43% reduction in invasive procedures between the two cohorts over the 10-week period. It was not possible to assess patient decision-making after a positive ccfDNA test result, as only one such finding occurred.

Not surprisingly, ccfDNA test acceptance was also influenced by expected out-of-pocket expenses. Unlike an earlier study that found no influence in uptake with insurance status (Taylor et al., 2014), our patients with lower expected expenses were more likely to request and complete ccfDNA testing than patients with higher expenses. In fact, the increase in uptake mainly occurred among those women with the lowest ($235) out-of-pocket expense. The increase in uptake between periods was higher for those already in the group with the lower $235 out-of-pocket expense. When the expense was $475 (or higher), the uptake was lower, but similar even when the expense was $1,933. This indicates a possible threshold in acceptable out-of-pocket expenses. This is especially important, given that 41% of our population is on some form of public assistance. This finding is consistent with a survey demonstrating that women were willing to trade other attributes of screening (e.g., detection rate, turnaround time) to pay a lower price (Carroll et al., 2013). One in 10 women who declined testing indicated that the test cost was at least a factor in their decision-making. One limitation of our study was the inability to reliably obtain certain out-of-pocket expenses for each individual patient, due to lack of information regarding potential assistance from the testing laboratory.

We found evidence for two distinct subgroups of women in our study. The largest group declined ccfDNA testing because they wanted no further testing of the pregnancy at all, for example, “the test result would not change anything” about their pregnancy. The 33% may be an artificially low estimate, as 12% of women did not provide a reason in the first cohort. A second group consisted of women willing to accept the risk associated with current invasive diagnostic testing to have definitive results, while choosing to avoid the delay and expense of a secondary screening test. However, since the percentage of women opting for an invasive procedure declined from 21% to 12% over time, this group may actually be relatively small. The low uptake of ccfDNA testing was consistent with the California group (Taylor et al., 2014), which reported an overall 28% acceptance rate of ccfDNA rate.

The current study was limited mainly to second trimester pregnancies, with the median gestational age of 18 weeks at counseling. Uptake may have been higher where testing offered in the first trimester, as this would provide more time and privacy for decision-making. If women expressed concern, they were counseled about the 8 to 10 day turnaround time and told that diagnostic testing may be a better option. None of the women stated that limited availability of termination was a reason for declining testing. Our retrospective chart review did not allow for structured questioning (or questionnaire) to elicit more detailed information about their decision-making. Such a design would have required patient consent, resulting in a more limited participation. Rather, the counselors documented relevant patient comments related to decision-making in their contemporary notes.

Our results are consistent with others (Vahanian et al., 2014) who found that patients with public insurance were less likely to choose ccfDNA testing than those with private insurance. Our overall reported uptake rate is also consistent with another publication from California that focused on women with abnormal serum/ultrasound screening results (Chetty et al., 2013). They reported the uptake of ccfDNA testing to be 39% compared with 23% in our study. Their rate may be slightly higher due to a more extended and current window for enrollment. The California report also stated that about 14% of Caucasian women declined all testing, similar to the 21% and 12% we reported. We did not collect information on race, but 80% of Rhode Island births are classified as Caucasian.

In summary, our early experience in offering this test to high-risk women shows growing understanding and acceptance by women, with utility extending through late pregnancy (third trimester). If the out-of-pocket cost of testing drops further, ccfDNA testing is likely to become more attractive as both a secondary screening test and possibly as an alternative to primary screening using conventional ultrasound and/or serum testing.