Risk of miscarriage with amniocentesis

Despite the inadvisability of performing early amniocentesis (EA) before 15 gestational weeks due to a high rate of miscarriage, a retrospective cohort study has found no significant difference in the procedure-related risk of miscarriage between EA, at around 14 weeks gestation, and mid-trimester amniocentesis (MA).

The authors of the study in the journal Diagnostics noted that an advantage of EA over MA is that it shortens the time interval between screening via a noninvasive prenatal test (NIPT) and amniocentesis.

Principal investigator Philip Loquet, MD, a maternal-fetal medicine specialist at St. Augustinus Hospital in Antwerp and Antwerp University Hospital, Belgium, said he initiated the study to confirm similar study results presented 30 years ago at the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG).

“Most of the recent research was conducted by the first 2 authors of the current study (Kelly Steinfort, MD, and Ellen Van Houtven, MD, both of Antwerp University Hospital), for their doctorate thesis,” Loquet told Contemporary OB/GYN®. “It is a massive task that I never had the time to complete by myself. Still, 30 years ago, I proved that when you have a karyotype before 15 gestational weeks, 90% of patients do not feel the baby move and termination of pregnancy was more acceptable at that time of gestation.”

Now with the advent of NIPT, “the problem has become even more acute,” Loquet said. “NIPT is done at 12 weeks, so the result is at 13 weeks. Then you need to inform the patient to wait 3 more weeks before the result can be confirmed. This is agony for the patients.”

The current study compared the procedure-related risk of miscarriage between MA, at 15 to 17weeks of gestational age, and EA. Procedure-related fetal loss was defined as spontaneous abortion occurring within 4 weeks of the procedure. Multiple gestations, amniocenteses performed after 17 or before 14 weeks, and indications other than prenatal genetic diagnoses and procedures were excluded, as were procedures performed by less experienced ob-gyns.

MA and EA were conducted at Antwerp University Hospital or St. Augustinus Hospital from January 2007 to November 2018

Complete outcome data were available for 73.1% of the 1,515 women: 69.9% in the MA group and 83.2% in the EA group.

No significant difference was found in the procedure-related risk of miscarriage between the two time periods for amniocentesis, EA 0.82% vs. MA 0.36% (P =.646), for a difference of 0.46%.

The 2 cases of procedure-related miscarriages in the EA group were likely caused by preterm premature rupture of membranes (PPROM), whereas in 3 of the 4 cases in the MA group, amniocentesis became complicated by an intrauterine death a few days after the procedure. The fourth MA case became problematic due to chorioamnionitis, which resulted in premature labor.

“I have been doing amniocentesis at 14 weeks for 30 years with minimal complications, and this study seems to confirm that,” said Loquet, who also is a prenatal diagnosis consultant at Antwerp University Hospital.

Loquet emphasized, though, that 14 weeks amniocentesis is different than EA referred to in the literature. “Patients at 12 and 13 weeks were included in those studies,” he said. “For that gestational age, chorionic villus sampling (CVS) is much preferred. But for NIPT, confirmation by amniocentesis is required.”

The only challenge with EA can be the “tenting” of membranes, according to Loquet, thus leading to perhaps a higher risk of PPROM. “To avoid tenting, I make a fast entry of the needle in the amniotic cavity,” he said. “In our study, we did not see a significantly increased PPROM rate. However, skill is required, so I would advocate 14 weeks amniocentesis only in experienced hands.”


Loquet reports no relevant financial disclosures.

Steinfort K, Van Houtven E, Jacquemyn Y, et al. Difference in procedure-related risk of miscarriage between early and mid-trimester amniocentesis: a retrospective cohort study. Diagnostics. 2021 Jun 16;11(6):1098. doi:10.3390/diagnostics11061098