Positive predictive value for noninvasive prenatal genetic screening

A recent report shares the odds that a noninvasive prenatal genetic screening (NIPGS) result is suggestive of a true positive in a clinical setting. It also makes several key points about the administration of NIPGS and patient choices.

Whitney A. Neufeld-Kasier, Edith Y. Cheng, and Yajuan J. Lieu, all with the University of Washington School of Medicine, Seattle Washington, have published in BMC Medicine what, to their knowledge, “is the largest independent report of clinical experience using [NIPGS] in a tertiary referral center in the United States.” Here’s what they found.

Background

The team reviewed medical records of 632 patients who had NIPGS between March 2012 and December 2013, the first two years the University of Washington Medical Center offered NIPGS. Being seen at a tertiary referral center, each of the women undergoing NIPGS were considered “high risk.” This is consistent with current practice guidelines. “High risk” was based on advanced maternal age, having had a conventional screen indicating increased risk, family history of aneuploidy, or an ultrasound suggesting the possibility of an aneuploidy.

Of the 632 patients, 578 (91%) received an NIPGS result not suggesting an increased chance for aneuploidy. Of these receiving “normal” NIPGS results, 73% did not have confirmation of their NIPGS result due to lack of followup –usually due to patients declining invasive testing and delivering at a different medical center. Therefore, NIPGS sensitivity, specificity, or negative predictive value (NPV) could not be determined. “Thus, the main empases of this report” is on the positive predictive value (PPV) of NIPGS.

Results

There were 53 patients who received an NIPGS report indicating an aneuploidy. Of these:

• 22 (41%) chose prenatal diagnostic testing:
  • 5 of the 22 (22.7%) had diagnostic results discordant with their NIPGS results (meaning NIPGS was a “false positive”);
  • Of the 18 whose NIPGS results were confirmed by prenatal diagnostic testing, 13 chose termination of pregnancy (TOP), giving a termination rate of 72%, consistent with the most current quoted rate.
• 31 (58%) declined prenatal diagnostic testing.
• 23 (43%) had postnatal diagnostic testing:
  • 7 patients opted for TOP after fetal anatomy ultrasound showed anomalies consistent with the condition suggested by NIPGS;
  • 7 experienced intrauterine fetal demise (IUFD);
  • 9 continued to term; 7 had concordant karyotype results and 2 had discordant results (“false NIPGS positives”).
• 8 (15%) declined both prenatal and postnatal diagnostic testing:
  • 2 (25%) had ultrasound anomalies consistent with NIPGS and experienced IUFD;
  • 4 (50%) had newborn clinical exams discordant with NIPGS result (e.g. “false positives”);
  • 2 (25%) had abnormal follow-up nuchal translucency results, “declined all further services, and opted for TOP.”

Based on the concordant and discordant NIPGS results, the study reports the following on NIPGS PPV:

• PPV for all disorders was 77.4%:
  • PPV for Down syndrome: 90.3%;
  • PPV for Trisomy 18: 63.6%;
  • PPV for sex chromosome aneuploidies (SCAs): 55.6%;
  • PPV for Trisomy 13: 50%.
• Concordance and discordance results correlated to maternal age:
  • For abnormal concordant, e.g. NIPGS screen-positive confirmed with diagnostic testing, median maternal age was 39 years old;
  • For abrnomal disconcordant, e.g. false-positive NIPGS results, median maternal age was 34;
  • For normal concordant, e.g. NIPGS screen-negative confirmed, median age was 36.

Researchers’ take-away points:

Positive take-aways:

1. Normal NIPGS “greatly reduce the probability of a fetus affected with a condition included in the screen. … Consequently, our rate of amniocentesis dropped significantly.” This reduction in diagnostic testing is consistent with other medical centers.
2. NIPGS “proved to be an invaluable tool for patients wanting to learn as much as possible about fetal health prior to delivery without incurring the miscarriage risk associated with CVS and amniocentesis.”

“Our experience also confirmed the limitations of [NIPGS]”:

1. NIPGS tests a limited number of conditions. One patient received a “normal” NIPGS result, but had a fetus with trisomy 22, which was not tested by NIPGS screens.
2. “[A]lthough normal discordant [NIPGS] resuls [i.e. “false negatives”] are rare with such a high NPV, they do occur.” Since their report was issued, they had a patient receive a false negative for Down syndrome.
3. “Third, abnormal [NIPGS] results confer a high risk, but not a diagnosis, of a fetal abnormality.”
4. “Apparent sex discordance can have a variety of causes * * * [t]hus, it is imperative that providers make every effort to confirm [NIPGS] results that are abnormal or appear discordant for fetal sex.”
5. “Maternal karyotyping should be offered after abnormal [NIPGS] results to rule out maternal mosaicism, especially when an SCA is suspected.”
6. Following an abnormal NIPGS result, “patients should always be referred to a center offering genetic counseling, high-resolution fetal ultrasound, and diagnostic genetic testing.”

Further discussion

The results from the University of Washington Medical Center further demonstrate how “99% accuracy,” the oft-made claim by NIPGS laboratory marketing literature, should be understood as not suggesting a 99% chance that a mother will have a child with the tested-for condition. In this, the largest known study in clinical practice, a high-risk mother receiving an abnormal NIPGS result was shown to not have a 99% chance, but, at best, a 90% chance if the result was for Down syndrome, and, at worse, just a 50/50 chance if the result was for Trisomy 13.

The results further support the practice guidelines for NIPGS:

• NIPGS has been shown to be reliable as a second-tier screen, i.e. offered once a mother is already considered at high-risk. This is both shown by this study’s cohort being high-risk mothers and by the results showing discordant results correlate with the younger the mother.
• Patients receiving an abnormal NIPGS result should be referred for genetic counseling and confirmation should be had through diagnostic testing, with amniocentesis being more reliable than CVS.

That last bullet is further demonstrated by 9 (16%) of the pregnancies with an abnormal NIPGS being terminated without having a prenatal diagnosis; one had a discordant result and two others were undetermined.

Lastly, given the myriad of causes for a report of SCA, the authors offer a recommendation heretofore not mentioned by professional guidelines: that the mothers, themselves, should be offered karyotyping  to rule out maternal mosaicism if they receive an abnormal NIPGS result. Adding this to the counseling protocol of patients is likely to further provide mothers some unexpected news when they sought prenatal testing just for reassurance.