The risks of pregnancy loss after chorionic villus sampling (CVS) and amniocentesis have not been well studied in twin pregnancies. Most current guidelines on prenatal diagnostic techniques do not address this issue. The aim of this review was to assess the available evidence on the risks after CVS and genetic amniocentesis when performed in twin pregnancies. The focus was pregnancy loss after the procedure and evaluation according to different sampling techniques and chorionicity.
The MEDLINE database (1990–2011) was searched for articles that reported pregnancy outcomes after CVS and genetic (midtrimester) amniocentesis. Data were collected regardless of the definition of pregnancy loss. Pregnancy loss (both fetuses) and fetal loss (1 of the 2 fetuses) were distinguished in this assessment. Data included gestation at the time of procedure and chorionicity; transabdominal (TA) versus transcervical (TC) approach, type of needle or catheter used, and sampling error/cross-contamination for CVS; and single- versus double-uterine entry, sampling error/cross-contamination, and use of dye for amniocentesis.
Four of 9 CVS studies reported overall pregnancy loss with pooled incidence of 3.84% (95% confidence interval [CI], 2.48%–5.47%). Estimated pregnancy loss rates before 20 or 28 weeks were 2.75% (95% CI, 1.28%–4.75%) and 3.44% (95% CI, 1.67%–5.81%), respectively. The pooled rate of fetal loss was 5.48% (CI, 4.06%–7.13%). Two studies included a control group; one found an overall fetal loss rate of 2.72% (6/220) in study patients and 3.14% (11/350) for control subjects, and the other 7.35% (15/204) for the study group and 8.41% (17/202) for control subjects. Three studies commented on chorionicity, with 4.46% (18/403) of twins being monochorionic (MC). Of 6 studies reporting on TA and TC approaches, the total fetal loss after the TA approach was 7.09% (10/141) compared with 3.94% (5/127) after the TC approach (RR, 2.08; 95% CI, 0.73–5.91). For a single-needle or double-needle system, the rates of overall pregnancy loss before 24 weeks’ gestation were 4.5% (2/44) and 2.8% (3/104), respectively. Two studies that used a double-needle system described overall fetal loss rates of 12.5% (6/48) and 4.3% (8/186). The rates of fetal losses before 24 weeks were 4.17% (1/24) for single-uterine entry versus 3.75% (3/80) for double-uterine entry. The combined proportion for sampling error and cross-contamination was about 1% (95% CI, 0.49%–1.7%).
Eighteen studies fulfilled selection criteria for amniocentesis. Four reported overall pregnancy loss with the pooled proportion estimated at 3.07% (95% CI, 1.83%–4.61%). Estimated pregnancy and fetal losses before 24 weeks’ gestation were 2.54% (95% CI, 1.43%–3.96%) and 1.87% (95% CI, 0.59%–3.85%), respectively. In 4 of 6 case-control studies comparing pregnancy loss before 24 weeks, the miscarriage rate after amniocentesis was 2.59% compared with 1.53% in control subjects (RR, 1.81; 95% CI, 1.02–3.19). Six studies reported on chorionicity, with 11.16% (135/1210) MC twin pairs. In 2 studies, 4 spontaneous fetal losses occurred before 24 weeks’ gestation in dichorionic twins (4/86), but none in MC twins (0/45). Only 1 study described total fetal loss based on the number of uterine entries, 2.7% (1/37) for single entry versus 4.76% (3/63) for double entry (RR, 0.57; 95% CI, 0.06–5.30). Most investigators did not use any dye, and no study reported pregnancy outcomes based on using dye. Most of the amniocentesis investigations did not report on cross-contamination or sampling errors.
In 2 studies comparing the outcomes between CVS and amniocentesis, pregnancy loss before 28 weeks had a combined loss rate of 2.98% for CVS and of 1.53% for amniocentesis (RR, 1.11; 95% CI, 0.22–5.57). A third study reported pregnancy loss before 24 weeks with a loss rate of 4% (4/100) in CVS and 2.88% (3/104) in amniocentesis (RR, 1.39; 95% CI, 0.23–9.4).
All current evidence is from nonrandomized cohort studies. For counseling of patients carrying twins, it would be helpful to estimate the risk to pregnancy from an invasive procedure. However, without randomized controlled trials with adequate sample size, the excess risk after invasive procedures in twins cannot be estimated accurately. Available data show similar overall pregnancy loss rates for both amniocentesis and CVS with an excess risk of about 1% above the background risk.