NIPT Positive Predictive Value Calculator
NIPT Positive Predictive Value Calculator
A cell-free DNA screen came back high risk. This works out the probability that the fetus is actually affected, from the published detection rate and false positive rate for that specific condition and the prior risk you supply — and shows how steeply that probability falls as the condition gets rarer.
NIPT positive predictive value
Condition + prior → PPVA high-risk cell-free DNA result for trisomy 21 in a woman whose age-related prior at 12 weeks is 1 in 250, with a fetal fraction of 9%
The calculation, in full
- prior
- 1 ÷ the denominator you entered. The probability the fetus is affected before the screen — from maternal age, from first-trimester combined screening, or from a previous affected pregnancy
- sensitivity
- the detection rate: the proportion of affected pregnancies the screen calls high risk. 99.7% for trisomy 21, 97.9% for trisomy 18, 99.0% for trisomy 13, 83.3% for 22q11.2 deletion
- false positive rate
- 1 minus specificity: the proportion of unaffected pregnancies the screen calls high risk. 0.04% for each of the three trisomies and 0.05% for 22q11.2 deletion — small numbers that nevertheless dominate the answer whenever the condition is rare
- what is not in it
- fetal fraction, gestational age, twin pregnancy and the laboratory’s own no-call rate. The published detection rates are conditional on a result having been issued, and there is no sourced function relating sensitivity to fetal fraction, so none is invented here
Worked example
A high-risk cell-free DNA result for trisomy 21 in a woman whose age-related prior at 12 weeks is 1 in 250, with a fetal fraction of 9%
Prior probability = 1 ÷ 250 = 0.004, that is 0.4%
True positives among 1 pregnancy's worth of probability: 0.004 × 0.997 = 0.003988
False positives: (1 − 0.004) × 0.0004 = 0.996 × 0.0004 = 0.0003984
PPV = 0.003988 ÷ (0.003988 + 0.0003984) = 0.003988 ÷ 0.0043864 = 90.9%
So about 9 in 100 women with this result are carrying an unaffected fetus, and that is with the most favourable of the four conditions at a relatively high prior
Change nothing but the condition. Trisomy 13 at the same 1 in 250 prior gives 0.004 × 0.99 ÷ (0.00396 + 0.0003984) = 90.9% as well — because the prior is what dominates, not the small difference in detection rate
Now change the prior instead. Trisomy 13 at a 20-year-old's prior of 1 in 8,000 gives 0.000125 × 0.99 ÷ (0.00012375 + 0.00039995) = 23.6%. Three quarters of those positives are false, with exactly the same test
Positive predictive value by condition and prior — every figure recomputed here
| Condition | Prior 1 in 250 | Prior 1 in 1,000 | Prior 1 in 4,000 | Prior 1 in 8,000 |
|---|---|---|---|---|
| Trisomy 21 (99.7% / 0.04%) | 90.9% | 71.4% | 38.4% | 23.8% |
| Trisomy 18 (97.9% / 0.04%) | 90.8% | 71.0% | 37.9% | 23.4% |
| Trisomy 13 (99.0% / 0.04%) | 90.9% | 71.2% | 38.2% | 23.6% |
| 22q11.2 deletion (83.3% / 0.05%) | 87.0% | 62.5% | 29.4% | 17.2% |
Maternal age-specific prior risk, at 12 weeks
| Maternal age | Trisomy 21 | Trisomy 18 | Trisomy 13 |
|---|---|---|---|
| 20 | 1 in 1,068 | about 1 in 2,500 | about 1 in 8,000 |
| 30 | 1 in 626 | — | — |
| 35 | 1 in 249 | about 1 in 600 | about 1 in 1,800 |
| 40 | 1 in 68 | — | — |
| 45 | 1 in 16 | — | — |
Where the published positive predictive values disagree with this calculation
| Published figure | What this page computes | Why they differ |
|---|---|---|
| 22q11.2 deletion, SMART: PPV 52.6% | 52.2% at the observed prior of 1 in 1,524 | SMART had 10 true positives and 9 false positives, so 10 ÷ 19 = 52.6%. Applying a false positive rate of 0.05% to the 18,277 unaffected pregnancies gives 9.14 false positives instead of 9, and 10 ÷ 19.14 = 52.2%. Rounding a count into a rate costs 0.4 percentage points |
| Trisomy 21, SMART low-risk arm: PPV 85.7% | 77.8% at that arm’s observed prior of 1 in 713 | This one is a real disagreement. SMART’s own false positive rate in that arm was 3 in 12,818, which is 0.023% — a little over half the 0.04% pooled figure this page uses. Substituting SMART’s own rate reproduces 85.7% exactly. So the gap is entirely in which published false positive rate you take, and it is worth eight percentage points of positive predictive value |
Why a very specific test still produces mostly false positives
Cell-free DNA screening for the common trisomies is, by any ordinary standard, an excellent test. It detects 99.7% of trisomy 21 and calls only about 4 in 10,000 unaffected pregnancies high risk. And yet the probability that a high-risk result is correct ranges, across the conditions and priors on this page, from about one in six to about nineteen in twenty. Nothing about the test changes across that range. What changes is how many affected pregnancies there are to find.
The mechanism is worth stating plainly, because it is the reason the page exists. Imagine 10,000 pregnancies in which the risk of trisomy 13 is 1 in 8,000. Rather more than one affected fetus will be detected — call it 1.2 across 10,000, since 10,000 ÷ 8,000 is 1.25 and the detection rate is 99%. Meanwhile the false positive rate of 0.04% acting on the roughly 9,999 unaffected pregnancies produces about 4 false positives. So the screen generates around five positive results, of which one is real. The test has not performed badly; there simply were not many affected pregnancies for it to find, and a tiny false positive rate applied to a very large denominator beats a near-perfect detection rate applied to a very small one.
This is why sensitivity and specificity are properties of the test and positive predictive value is not. Quote a PPV without saying what prior it assumes and you have said almost nothing. It is also why the correct response to a high-risk result is never to act on it and never to dismiss it, but to confirm it — chorionic villus sampling or amniocentesis with a diagnostic test. That is the one step that converts a probability into a diagnosis, and no screening technology, however good its numbers, replaces it.
Microdeletions make the arithmetic worse in both directions at once. They are rarer than the trisomies, which lowers the prior, and they are detected less reliably from cell-free DNA, which lowers the sensitivity and raises the false positive rate. Only 22q11.2 deletion is offered here, because it is the only one with a prospective, blinded study that confirmed outcomes: 12 confirmed cases among 18,289 pregnancies, a detection rate of 83.3% with a confidence interval running from 51.6% to 97.9%, and a positive predictive value of about a half. Screening for the other microdeletions on commercial panels does not have comparable published performance, and this page will not print numbers that do not exist. That is not caution for its own sake: a made-up sensitivity in a PPV calculation produces a confident-looking figure that is wrong in an unknowable direction.
Two further things a positive predictive value cannot capture. A screen can be genuinely positive in a pregnancy with a chromosomally normal fetus — confined placental mosaicism is the commonest reason, since the cell-free DNA is placental rather than fetal, and a vanishing twin, a maternal copy number variant and, rarely, an occult maternal malignancy do the same. And a low fetal fraction is not a neutral technical failure: it is associated with trisomy 18 and 13 as well as with maternal weight and early gestation, so a no-call sample carries information and should not simply be filed. For the read-count arithmetic behind a fetal fraction see the cell-free DNA fetal fraction calculator; for the general form of this calculation on any test see the sensitivity and specificity calculator and the likelihood ratio calculator.
Frequently asked questions
What is the positive predictive value of NIPT for Down syndrome?
It depends entirely on the prior risk, which is why no single number is correct. Using a detection rate of 99.7% and a false positive rate of 0.04%, a high-risk trisomy 21 result gives a positive predictive value of about 91% at a prior of 1 in 250 (roughly a 35-year-old), about 71% at 1 in 1,000, and about 24% at 1 in 8,000. Sensitivity and specificity belong to the test; positive predictive value belongs to the population you applied it to. Any quoted PPV without a stated prior should be treated as incomplete.
Why do rare conditions give such a low positive predictive value?
Because the false positives come from the unaffected pregnancies, and there are vastly more of them. A false positive rate of 0.04% sounds negligible until it is applied to 9,999 unaffected pregnancies out of 10,000, which produces about four false positives — while a condition affecting 1 in 8,000 supplies only about 1.2 true positives in the same 10,000. The test is not performing worse; the ratio it is working against has changed. This is why a microdeletion screen with a specificity above 99.9% can still be wrong more often than right.
Does a high-risk cell-free DNA result mean my baby has the condition?
No. It means the probability has gone up, sometimes a great deal and sometimes very little, and it always requires confirmation by an invasive diagnostic test — chorionic villus sampling or amniocentesis — before any irreversible decision. Even at a positive predictive value above 90% about one result in ten is false. There are also biological reasons a screen can be correctly positive while the fetus is normal: the cell-free DNA is placental in origin, so confined placental mosaicism produces a true placental finding and a normal baby, and a vanishing twin, a maternal copy number variant or an undiagnosed maternal cancer can each do the same.
Is cell-free DNA screening reliable for microdeletions?
Much less so than for the common trisomies, and for most microdeletions the performance is not established. 22q11.2 deletion is the exception with real evidence: the SMART study screened 18,289 pregnancies with confirmed outcomes, found 12 cases, and reported a detection rate of 83.3% with a 95% confidence interval from 51.6% to 97.9%, a false positive rate of 0.05% and a positive predictive value of about 53%. That is roughly a coin toss, and it is the best-evidenced microdeletion on any panel. 1p36 deletion, Wolf-Hirschhorn, Cri-du-chat, Prader-Willi and Angelman are screened commercially without comparable published sensitivity or specificity, and no professional body recommends routine screening for them.
Does fetal fraction change the positive predictive value?
Not in the way people expect, and this page deliberately does not pretend it does. The published detection rates are conditional on a result having been issued at all, and there is no soundly published function relating sensitivity to fetal fraction that could be built into a calculation. What fetal fraction governs is whether a result can be issued: below about 4% most laboratories will not report one. A low or failed sample is not neutral, because low fetal fraction is associated with trisomy 18 and trisomy 13 as well as with higher maternal weight and earlier gestation, so a no-call warrants a repeat or a different approach rather than reassurance.
What prior risk should I use?
For the trisomies, the maternal age-specific risk at the gestation the sample was taken, from the table on this page, unless you have a risk from first-trimester combined screening — which already includes maternal age along with nuchal translucency and the biochemical markers, and is the better prior when it is available. A previous affected pregnancy, an abnormality on ultrasound or a parental balanced translocation all raise the prior above the age-related figure, sometimes greatly, and in those situations the age table understates the answer. For 22q11.2 deletion the prior is not age-related at all: use about 1 in 1,500 to 1 in 4,000.
Related calculators
References
- Gil MM, Accurti V, Santacruz B, Plana MN, Nicolaides KH. Analysis of cell-free DNA in maternal blood in screening for aneuploidies: updated meta-analysis. Ultrasound Obstet Gynecol. 2017;50(3):302–314. doi:10.1002/uog.17484. Pooled weighted detection rate and false positive rate in singleton pregnancies: trisomy 21, 99.7% (95% CI 99.1–99.9) and 0.04% (0.02–0.07); trisomy 18, 97.9% (94.9–99.1) and 0.04% (0.03–0.07); trisomy 13, 99.0% (65.8–100) and 0.04% (0.02–0.07).
- Dar P, et al. Cell-free DNA screening for prenatal detection of 22q11.2 deletion syndrome. Am J Obstet Gynecol. 2022 (the SMART study), article PII S0002-9378(22)00006-0. 18,289 participants with confirmed outcomes and 12 confirmed cases, a prevalence of 1 in 1,524; original algorithm sensitivity 75.0% (95% CI 42.8–94.5), specificity 99.84% (99.77–99.89), PPV 23.7% (11.44–40.24); improved algorithm 10 of 12 cases, 83.3% (51.6–97.9), false positive rate 0.05% against 0.16%, PPV 52.6%.
- Dar P, et al. Cell-free DNA screening for trisomies 21, 18, and 13 in pregnancies at low and high risk for aneuploidy with genetic confirmation. Am J Obstet Gynecol. 2022, article PII S0002-9378(22)00041-2. 17,564 participants; trisomy 21 in the low-risk arm, sensitivity 100% (18/18), specificity 99.98% (12,815/12,818), PPV 85.71% (18/21) — the figure this page reproduces only when SMART’s own false positive rate is substituted for the pooled one.
- Snijders RJM, Sundberg K, et al. Maternal age- and gestation-specific risk for trisomy 21. Ultrasound Obstet Gynecol. 1999;13(3):167–170. The trisomy 21 column of the prior-risk table: at 12 weeks, 1 in 1,068 at age 20, 1 in 626 at 30, 1 in 249 at 35, 1 in 68 at 40 and 1 in 16 at 45.
- Nicolaides KH. Screening for fetal aneuploidies at 11 to 13 weeks. Fetal Medicine Foundation review, fetalmedicine.org. “The estimated risks for fetal trisomies 21, 18 and 13 for a woman aged 20 years at 12 weeks of gestation are about 1 in 1000, 1 in 2500 and 1 in 8000, respectively”; and at 35 years, “about 1 in 250, 1 in 600 and 1 in 1800”.
- American College of Obstetricians and Gynecologists and Society for Maternal-Fetal Medicine. Screening for Fetal Chromosomal Abnormalities. Practice Bulletin No. 226. Obstet Gynecol. 2020. Cell-free DNA is a screening test, a positive result requires diagnostic confirmation, and routine screening for microdeletions is not recommended.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.