1. Introduction
The Great East Japan Earthquake, which occurred on 11 March 2011, along with the subsequent tsunami and nuclear accident at the Fukushima Daiichi Nuclear Power Plant, was the most devastating event in recent Japanese history. In the Fukushima Prefecture, thousands of deaths were reported due to the tsunami, with the devastation affecting many people living in the coastal areas of the Soso and Iwaki districts. After the power plant accident, residents living in coastal areas were forced to evacuate suddenly. The Aizu district, located in a mountainous region in Fukushima Prefecture and far from the power plant, experienced less damage than the coastal regions (
Figure 1).
After the disaster, the Fukushima Health Management Survey (FHMS), a population-based study that included geographical and birth information used to evaluate pregnancy outcomes, was launched by Fukushima Prefecture to provide valuable data on the investigation of the health effect of low radiation dose and disaster-related stress in 2011 [
1]. It is well established that disasters can significantly impact perinatal outcomes. Numerous studies have reported associations between such disasters and various aspects of perinatal health [
2]. Even without direct exposure to the disaster, the surrounding circumstances can increase the risk of adverse pregnancy outcomes, including more unplanned pregnancies and sexually transmitted infections [
2]. Specifically, regarding the Great East Japan Earthquake, some studies have reported its impact on perinatal outcomes in 2011 [
3,
4,
5,
6,
7]. It has been shown that this disaster impacted both immediate outcomes and longer-term postnatal factors, such as the incidence of postnatal depression and the state of breastfeeding nutrition. Only a few studies have examined chronological trends in the occurrence of perinatal outcomes after the disaster. The chronological trends in pregnancy outcomes after the Great East Japan Earthquake is of worldwide interest; currently, the FHMS maintains the data from the investigation of the effect of this disaster on pregnancy and infant care.
This study aimed to examine the 8-year chronological trends in perinatal outcomes after the Great East Japan Earthquake in Fukushima Prefecture by district, using data from the FHMS.
3. Results
The survey questionnaire was mailed to 115,155 women who had experienced pregnancy during the study period. In total, 57,537 women (response rate 50.0%) responded to the questionnaire. Women who had given birth before 11 March 2011 (
n = 459), who received maternal and child health handbooks outside Fukushima Prefecture (
n = 176), for whom sufficient data were unavailable (
n = 128), who were pregnant at the time of the survey (
n = 165), who had had an abortion (
n = 427), and those with triplets (
n = 3) were excluded from the study. After applying these exclusion criteria, 57,375 women were finally included in the analyses. The number of cases in each of the eight years from 2011 to 2018 inclusive was 9299, 7085, 7152, 7024, 6913, 7191, 6332, and 6541, respectively (
Figure 2).
Table 1 shows the chronological changes in response rates by district. The Mantel–Haenzel test showed a significantly decreasing trend in the response rate from 58.2% in 2011 to 51.4% in 2018 (
p < 0.001). This was observed in all districts except Kennan (
p = 0.652). The Chi-square test showed differences in response rates among districts. During the study period, Kenpoku had the highest response rate, except in 2011, when the area with the highest response rate was Soso (65.6%), the location of the coastal area and the Fukushima Daiichi power plant.
Table 2 summarizes the characteristics of the respondents based on the year of delivery. Although maternal age, gestational age at delivery, and ratio of male newborns showed significant differences during the study period (
p < 0.001,
p < 0.001, and
p = 0.046, respectively), there was no significant difference in mean neonatal weight (
p = 0.912). Additionally, there were no significant differences in the rate of single pregnancy (
p = 0.368). The conception rate after sterility treatment and cesarean delivery tended to increase (
p < 0.001 and
p = 0.005, respectively).
Table 3 shows the chronological change in the occurrence of preterm birth by district. The Chi-square test showed no significant difference in the occurrence of preterm birth among the districts, except in 2012 (
p = 0.020). The overall rate of preterm birth was 4.6% in 2011 and increased to 5.2% by 2018. However, there was no significant trend in the occurrence of preterm birth between 2011 and 2018 in all areas (
p = 0.197).
Table 4 shows the chronological change in the occurrence of LBW by district. There was no significant difference in the occurrence of LBW during the study period. In 2011, a significant difference in the occurrence of LBW was observed between Kempoku (7.6%) and Iwaki districts (10.3%). In 2012, there was also a significant difference with respect to the occurrence of LBW between Kenpoku (7.6%) and Aizu districts (11.3%). The overall rate of LBW was 8.6% in 2011 and rose to 9.1% by 2018; however, no significant increasing trend was observed over time (
p = 0.500).
Table 5 shows the chronological change in the occurrence of anomalies in newborns. The rate of anomalies across all regions was 2.85% in 2011; however, it was 2.23% in 2018. There was no significant trend in the occurrence of newborn anomalies between 2011 and 2018 in all areas (
p = 0.069), and the Kenchu district showed a decrease in the rate of anomalies in newborns (
p = 0.011).
Table 6 delineates the incidence rate of anomalies each year, classified into 11 different types. Excluding the category of “Others,” the most frequently observed anomaly each year was “Cardiac malformation,” which was found at a rate of 1.01% in 2013. There was no significant trend in the occurrence of each neonatal anomaly.
4. Discussion
To our knowledge, this is the first population-based study that examined chronological trends for pregnancy outcomes following the disaster. Although the response rate in 2011 was the highest, especially in the Soso district, where the most disaster-related damage had occurred, the response rate decreased over the years, with more than 50% reported in 2018. We also found an increasing trend for mean maternal age, rate of conception after sterility treatment, and rate of cesarean delivery over the years. Regarding PTB, LBW, and fetal anomalies, there were no distinct changes in the trend of occurrence.
The response rate in this study was approximately 50%, which varied significantly over the years and between districts. While Kenpoku constantly had a higher response rate than the remaining districts, Soso had a higher rate only in 2011. This difference may be related to the concern of pregnant women in the Kenpoku and Soso districts that they were exposed to relatively higher radiation doses. Disasters potentially influence a range of reproductive outcomes [
8]. Previous studies examined the effects of exposure to disasters on pregnancy outcomes, and these exposures were usually from so-called “attacks” such as the World Trade Center Disaster, the bombing attack in Serbia, and the Madrid train bombing; environmental and chemical disasters such as the Bhopal gas release in India, the Three Mile Island accident, and the Chernobyl accident; and natural disasters such as earthquakes, hurricanes, floods, and avalanches [
2]. The Great East Japan Earthquake and Fukushima Daiichi nuclear accident form a complex disaster because they included natural disasters such as the Great Earthquake and tsunami and environmental/industrial disasters such as the nuclear power plant accident.
4.1. Disaster in Fukushima and Preterm Birth or LBW
Findings on the association between environmental/chemical disasters and gestational age or birth weight are conflicting. For instance, Goldman et al. reported that the Love Canal disaster in the USA showed no significant association with gestational age among 227 residents [
9]. Meanwhile, Levi et al. reported that the Chernobyl accident affected gestational age and maternal anxiety among 88 Swedish women who were early in their pregnancies during the disaster [
10]. Inconsistent with our study, the Wenchuan earthquake disaster in China increased the risk of PTB. Tan et al. compared the incidence of PTB between 6638 pregnant women before the Wenchuan earthquake disaster and 6365 pregnant women after the disaster [
11]. The incidence of PTB was 5.6% and 7.4%, respectively, significantly higher after the disaster (
p < 0.01). In Japan, high-risk pregnancies have increased due to advanced maternal age and complicated pregnancies [
12,
13,
14]. The incidences of PTB at <37 gestational weeks (5.7%) and LBW of <2500 g (9.4%) in the 6 areas in the study have almost been stable after the disaster. The evidence on the effect of the Fukushima disaster is also conflicting. Hyashi et al. used the same FMHS data and found that the Great East Japan Earthquake had no significant association with the incidence of PTB at <37 gestational weeks during the first year of the disaster among total Fukushima residents [
15]. Several pregnant women were forced to evacuate during the disaster, resulting in maternal depressive symptoms [
4,
16]. Suzuki et al. reported that pregnant women who changed their perinatal checkup institution due to medical indication were significantly associated with shorter gestational duration (β = −10.6,
p < 0.001) and preterm birth (adjusted odds ratio, 8.5; 95% confidence interval, 5.8–12.5) compared with women who visited only one institution [
17].
The association between the environment or a natural disaster and fetal growth is also controversial [
15,
18,
19,
20,
21]. Regarding the Great East Japan Earthquake and Fukushima Daiichi nuclear accident, we have previously reported no evidence that the disaster increased the incidence of small gestational age in the Fukushima Prefecture during the first year of the disaster [
5]. Using an institution-based investigation of the coastal area where the most catastrophic damage occurred, Leppold et al. reported no increased proportions in preterm births or LBW in any year after the disaster (merged post-disaster risk ratio of preterm birth: 0.68, 95% confidence interval: 0.38–1.21 and LBW birth: 0.98, 95% confidence interval: 0.64–1.51) [
22]. In Japan, pregnant women may have better access to relief programs or receive adequate support from their families, society, and government during disasters [
4].
4.2. Congenital Anomalies
The association between disasters and congenital anomalies is a major public concern. Several major environmental or industrial disasters have been related to congenital anomalies. These include the nuclear reactor accidents at Chernobyl, Ukraine, in 1986 and Three Mile Island, Pennsylvania, in 1979. The accident at Chernobyl involved a much larger radiation dose exposure and affected more people than the Three Mile Island or Fukushima incidents. Reviews on the effect of the Chernobyl disaster indicated increased microcephaly and neural tube defects [
23,
24,
25]. However, the incidence of most congenital anomalies did not increase in most European countries [
26,
27,
28]. Previous studies have reported that 2–3% of all newborns have a major congenital abnormality, which is detectable at birth [
29,
30] In Japan, from 2011 to 2016, the incidence of congenital anomalies was 2.43–2.59%, according to a report of the International Clearinghouse for Birth Defects Surveillance and Research Japan Center [
31]. Using a Japanese birth cohort study, which included 12,804 pregnant women in Fukushima Prefecture, Kyozuka et al. reported that the prevalence of major congenital anomalies at delivery between 2011 and 2014 in Fukushima Prefecture was 1.6–3.2%, depending on maternal age [
6]. Using the same data set as this study, Fujimori et al. reported that the occurrence of congenital anomalies in Fukushima Prefecture during the first year after the disaster was 2.72% (238/8672) [
3].
Environmental endocrine disrupters, a group of compounds with potentially adverse health effects, are thought to be associated with cryptorchidism [
32]. Kojima et al. suggested that it is difficult to clarify the prevalence of cryptorchidism due to the complexities of the design settings of epidemiological surveys of this disease. They rejected the hypothesis that cryptorchidism increased anywhere in Japan due to the Fukushima Daiichi Nuclear Power Plant accident [
33]. Hirata et al., using the All Japanese Cardiovascular Surgery Database, reported no increase in the number of patients with congenital heart disease during the period 2010–2013 [
34].
Our study has several strengths. In Japan, few epidemiological studies include pregnant women. Large-scale studies and data supported by the government are considered valuable. In addition, we obtained relatively precise data on gestational ages and birth weights from participants’ maternal and child health handbooks. Nevertheless, this study has some limitations. First, because the response rate was only 50–60% throughout the study period, the incidence of negative outcomes may have been overestimated due to the possible over-representation of women affected the most by the disasters, especially those pregnant between 2011 and 2012. Second, as this study used a self-administered questionnaire, it is assumed that the mothers answered correctly, especially regarding fetal anomalies. Lastly, this survey analyzed each district but did not investigate the relationship with individual radiation exposure doses.