No, the Miscarriage Risk from Amniocentesis Is Not 1 in 200 — That Number Is Decades Out of Date

The claim that invasive prenatal testing carries a miscarriage risk of about 1 in 200 is partially false. That figure was a reasonable estimate for its time, but it is now significantly out of date for amniocentesis — the most common of these procedures — and continuing to cite it can cause real harm by discouraging people from testing they may genuinely benefit from.

The 1-in-200 figure traces back to a landmark 1978 study by the U.S. National Institute of Child Health and Human Development. For decades it was taught in medical schools and printed in textbooks, and it stuck. But medicine has changed. Ultrasound guidance is now standard, operators are more experienced, and needles are finer. The result is a dramatically safer procedure.

The American College of Obstetricians and Gynecologists updated its official guidance in 2020 to reflect a procedure-related miscarriage risk for amniocentesis of roughly 0.1 to 0.3 percent — about 1 in 300 to 1 in 900. A large 2019 meta-analysis by Tabor and colleagues, covering tens of thousands of procedures, landed in the same range. A major Danish cohort study published in 2016 by Wulff and colleagues found a risk of just 0.18 percent, or about 1 in 556.

To be fair to the claim, CVS — chorionic villus sampling, a different procedure done earlier in pregnancy — does carry a somewhat higher risk, estimated by ACOG at around 0.5 to 1.0 percent, which is closer to the old 1-in-200 figure. And a 2007 systematic review by Mujezinovic and Alfirevic found combined historical rates for both procedures that were higher than modern estimates. So the 1-in-200 number is not invented — it just no longer reflects what happens in a well-equipped clinic today, and it was never a precise fit for amniocentesis specifically.

This matters because risk figures directly shape decisions. A person told their chance of miscarriage is 1 in 200 may decline testing that could give them critical information about their pregnancy. When the actual risk is closer to 1 in 500 or 1 in 900, that decision changes. Outdated statistics, even well-intentioned ones, can do quiet damage.

This kind of misinformation spreads not through bad faith but through institutional inertia. A number gets embedded in training materials and patient leaflets, and it circulates long after the science has moved on. If you are discussing these procedures with a provider, it is worth asking what their center's specific complication rate is — experienced, high-volume operators tend to have the lowest risks of all.