Risk for placental abruption following amniocentesis and chorionic villus sampling

Amniocentesis and chorionic villus sampling during the first and second trimesters of pregnancy are widely used methods for prenatal diagnosis. Much interest has focused on the risks of miscarriages and fetal loss as well as fetal morbidity and maternal psychological impact of these invasive procedures (Cederholm and Axelsson, 1997; Cederholm et al., 1999; Cahill et al., 2009). Placental abruption complicates 0.38–1% of pregnancies and is the leading cause of vaginal bleeding in the second half of pregnancy (Nilsen et al., 2008). Approximately 10% of preterm births and even one third of all perinatal deaths are caused by placental abruption (Oyelese et al., 2006). Maternal deaths due to placental abruption are rare, but still, for example, in Finland the maternal mortality ratio (MMR) of placental abruption is seven times higher than the overall MMR (Tikkanen et al., 2009). Few studies have concentrated on the association of amniocentesis or chorionic villus sampling and the obstetric emergencies, such as placental abruption, bleeding and premature rupture of membranes (Working Party of Amniocentesis, 1978; Eriksen et al., 1991), and only a few reports have included non-exposed controls (Cederholm et al., 2003). Some studies (Rhoads et al., 1989) but not all (Tongsong et al., 1998) have shown an association between these procedures and pregnancy complications. In a large study from Sweden including more than 70 000 singleton births (approximately 20 000 women were exposed to amniocentesis and 2000 to chorionic villus sampling among nearly 48 000 control mothers), no association was found between bleeding late in gestation or placental abruption and the use of these procedures (Cederholm et al., 2003). We wanted to study whether amniocentesis and chorionic villus sampling are associated with placental abruption in a large nation-wide register-based case-control setting. The study included all deliveries in Finland from 1 January 1991 to 31 December 2005. We used the Finnish Medical Birth Register (MBR) and the Hospital Discharge Register (HDR), maintained by the National Institute for Health and Welfare (THL). The MBR collects baseline data on maternal care and interventions during pregnancy and delivery, and on the outcomes of the newborns during the first 7 days. The database covers all live births and stillbirths. The HDR collects data on all inpatient episodes in health care facilities with up to three diagnoses, according to the relevant International Classification of Diseases (ICD) codes (ICD 9 1987–1995 and ICD 10 since 1996). Mothers exposed to amniocentesis or chorionic villus sampling and cases with placental abruption were identified by linking these two registers (Gissler and Haukka, 2004). Pearson chi-square test or Fisher's exact test and Mantel-Haenszel test were used for odds ratio (OR) estimates. Statistical analyses were performed by SPSS for Windows version 17.0 or 18.0. Authorization to use health register data was obtained from the register keeping authorities after the study was approved, as required by national data protection legislation. There were 887 439 births (874 004 singleton and 13 435 multiple births), including 3195 stillbirths and 1634 early neonatal deaths during 1 January 1991 to 31 December 2005 in Finland. A total of 3346 (3222 in singleton and 124 in multiple pregnancies) placental abruption cases (0.38%) were reported. The rest of the 884 093 births without placental abruption served as controls. Overall, 31 147 women were exposed to amniocentesis and 8837 women to chorionic villus sampling. The women exposed to amniocentesis were nearly 10 years and the women exposed to chorionic villus sampling 5–10 years older compared with women not exposed to either of these procedures. The age distribution of the cases and controls exposed to these procedures was similar. In the abruption group, there were 165 women (4.9%) exposed to second-trimester amniocentesis and 49 (1.5%) women exposed to first or second-trimester chorionic villus sampling. In the control group, the corresponding figures were 30 982 (3.5%) (OR 1.43 [95% confidence intervals (CI) 1.22–1.67], p < 0.001) and 8788 (1.0%) [OR 1.48 (95% CI 1.12–1.96), p = 0.006]. In the study group of women with singleton pregnancies and placental abruption, 156 women (4.8%) were exposed to amniocentesis and 48 women (1.5%) to chorionic villus sampling. The corresponding figures in the control group were 30 443 (3.5%) [OR 1.40 (95% CI 1.20–1.65), p < 0.001] and 8718 (1.0%) [OR 1.49 (95% CI 1.12–1.99), p = 0.005] (Figure 1). Mantel-Haenszel common odds ratios (OR) with 95% confidence intervals (CI) for placental abruption in women exposed to amniocentesis or chorionic villus sampling both in singleton and multiple pregnancies In the study group of women with multiple pregnancies and placental abruption (n = 124), nine women (7.3%) were exposed to amniocentesis and one woman (0.8%) to chorionic villus sampling. The corresponding figures in the control group were 539 (4.0%) and 70 (0.5%) [OR 1.85 (95% CI 0.94–3.68), p = 0.072 and OR 1.54 (95% CI 0.21–11.16), p = 0.483, respectively] (Figure 1). Our study is the largest register-based case-control study of the risk for placental abruption following amniocentesis or chorionic villus sampling. During the study period, there were nearly 900 000 deliveries in Finland. Overall, 3.5% of the mothers were exposed to amniocentesis and 1.0% to chorionic villus sampling, and the rate of placental abruption was 0.38%. Nearly 5% of the study group of women with singleton pregnancies and placental abruption were exposed to amniocentesis compared with 3.5% of the control women with no placental abruption. The corresponding figures for chorionic villus sampling were 1.5 and 1.0%. Such differences were not seen among women with multiple pregnancies. There are only a few controlled studies of the link between these procedures and placental abruption. The largest population-based cohort study from Sweden showed no association (Cederholm et al., 2003). Other studies have been relatively small and have shown contradictory results (Rhoads et al., 1989; Tongsong et al., 1998). The advanced age of the mother has been linked to placental abruption in some (Kramer et al., 1997) but not all studies (Tikkanen et al., 2006). Not surprisingly, women exposed to these procedures were older than not exposed, but this was the case both among cases and controls; no differences between the groups suggested that this bias had no effect on our findings. We conclude that both first- and second-trimester amniocentesis and chorionic villus sampling performed for fetal karyotyping may increase the risk for placental abruption in singleton, but not in multiple pregnancies. The strengths of our study are long study period (15 years) and large number of patients as well as validated health registers available in Finland. The limitations in this preliminary study include the fact that we did not take into account several other factors possibly increasing the risk for this pregnancy complication. However, more research is needed.