Multiple Congenital Anomalies after Assisted Reproductive Technology in Japan (between 2004 and 2009)

Multiple congenital anomalies (MCAs) are de�ned as two or more unrelated major structural malformations that cannot be explained by an underlying syndrome or sequence. e percentage and combination patterns of MCAs were analyzed using the European surveillance of congenital anomalies computer algorithms for nationwide data on congenital anomalies aer assisted reproductive technology between 2004 and 2009 in Japan. ere were a total of 177,548 pregnancies and 124,846 live births. About 7 % (96/1,324) were MCAs. Although most maternal/neonatal outcomes between the isolated cases group and the MCAs group were similar, higher early neonatal death rate was observed in the MCAs group than in the isolated cases group (9.8 % versus 3.7 % , resp.). Regarding the major organ system subcategory in ICD-10, the rate of MCAs was around 30 % in “congenital malformations of eye, ear, face, and neck,” “congenital malformations of the respiratory system,” and “congenital malformations of genital organs.” On the other hand, the rate of MCAs was less than 10 % in “congenital malformations of the circulatory system.” e combination patterns of diseases were widely varied. Of them, two or three diseases of the circulatory system, the digestive system, and the musculoskeletal system tended to co-occur in the same individuals.


Introduction
Association among congenital malformations is a concept introduced to designate the nonrandom tendency of some malformations to occur together more oen than expected by chance, without being components of known syndromes [1]. e utility of monitoring for patterns of multiple congenital anomalies (MCAs) has been increasingly recognized [2][3][4][5][6]. e possible reasons for this association are the results of known syndromes, as a consequence of a single primary anomaly, and known or unknown factors causing MCAs. is kind of coexistence is called comorbidity in general, and it is important to clarify the genetic/environmental background of associated malformations. According to Garne et al. [7], MCAs are de�ned as two or more unrelated major structural malformations that cannot be explained by an underlying syndrome or sequence.
Since etiologic heterogeneity may complicate epidemiologic analyses designed to identify risk factors for congenital anomalies, case classi�cation, which uses knowledge of embryologic and pathogenic mechanisms [8], is important to make case groups more homogeneous [9]. Wellesley et al. [10] produced a hierarchical method of classifying congenital anomalies into eight groups for inclusion of the source of data registration in conjunction with, but independent of, International Classi�cation of Disease (ICD) coding. e eight disease categories are chromosome, microdeletion, teratogen, new dominant, familial, syndromes, isolated, and multiple. us, there is fairly agreement in the literature on a gross etiologic classi�cation of congenital anomalies [9][10][11] for surveillance and epidemiologic studies to be used in combination with an anatomic approach based on organ system or an approach based on presumed pathogenesis [7]. e proposed classi�cations contain the following categories: chromosomal syndromes, monogenic syndromes, environmental syndromes, isolated congenital anomalies, and MCAs.
Surveillance of MCAs is considered to be more sensitive for the detection of new teratogens than is surveillance of all or isolated congenital anomalies [7,12]. e European surveillance of congenital anomalies (EUROCAT [13]) is a network of European registers of congenital anomalies carrying out epidemiologic surveillance of congenital anomalies, and it currently covers a population of 1.7 million births per year. Recently, the EUROCAT published a computer algorithm for the classi�cation of major congenital anomaly cases ( [7] see Methods).
e purpose of the present study was to estimate the frequency of MCAs and further examine their characteristics by adapting the EUROCAT computer algorithm and applying it to nationwide data on the congenital anomalies in Japan.

Outline of Japanese Birth Defects Data aer Assisted
Reproductive Technology (ART). e method for collecting data is described in detail elsewhere [14,15]. Almost all medical institutions performing ART are registered with the Japan Society of Obstetrics and Gynecology (JSOG), which administers questionnaire surveys for these medical institutions. Some of the survey data are presented in simple annual reports of aggregate, not individual, data. From 2004 to 2009 (the latest), the individual list of all ART pregnancies resulting in congenital anomalies was presented every year in the JSOG annual reports on ART (in Japanese) with an almost 100% response rate. e author used these case report data as initial information for the present existing data analyses. ese case reports were classi�ed according to the computer algorithm developed by the EUROCAT [7], which was based on the International Classi�cation of Disease version 10 (ICD-10) codes. is algorithm gives explicit coding rules for classi�cation and picks out a small subset of cases of potential MCAs. e etiologic classi�cation was translated into a computer algorithm that imposes a hierarchical clas-si�cation according to the ICD codes given for each case [7]. e outline of the computer algorithm is as follows: the �rst category is "chromosomal, " to which a case is allocated if it has a chromosomal ICD code, namely, Q90-Q93 (except Q936) or Q96-Q99. Only cases that are not allocated to "chromosomal" are considered for the next step. e next category is the "genetic syndromes, " "skeletal dysplasia, " and "congenital skin disorder" codes. e next code is the "teratogenic syndrome" code. us, aer 21 steps are performed, any remaining case is allocated to the "potential multiple anomaly" category only if it does not belong to any of the preceding categories. All cases are classi�ed into the following four main categories: chromosomal case, syndromes, isolated single anomalies, and "potential" MCAs. Although the EUROCAT further checked the "potential" multiples to determine whether these cases were "true" MCAs by three clinical geneticists [7], this procedure is beyond the present descriptive epidemiologic study.
According to the EUROCAT de�nition, cases of congenital anomalies for live births and fetal deaths with gestational age ≥20 weeks were selected. In total, 1,263 live births and 61 stillbirths were analyzed in the present study. e number of total ART live births were 124,846 between 2004 and 2009. e number of stillbirths (≥22 weeks, according to the de�nition of the JSOG report) was not reported in the JSOG data. e number of stillbirths deliveries (numbers of mothers) was 747 in the same period.

Statistical Analysis.
First, all cases were classi�ed into the four main categories according to the EUROCAT algorithm. en, demographic characteristics between the isolated cases group and the MCAs group were compared. For the comparison between the isolated cases group and the MCAs group, the -test was performed for maternal age and gestational weeks and the 2 test for method of fertility treatment, blastocyst transfer, plurality, sex, and perinatal outcome. e signi�cance level was set at .05. e proportions of cases classi�ed as isolated anomalies, syndromes (chromosomal, monogenic, and environmental), and MCAs by the ICD-10 main organ system subcategories were analyzed. Finally, typical combination patterns of two ICD-10 subcategories in MCAs were examined. If certain patients had more than two diseases from different ICD-10 subcategories, for example, diseases "A, B, and C", these diseases were broken down as several combinations of two diseases, "A and B, " "A and C, " and "B and C. " Newborn conditions and those of prematurity were not considered MCAs. A higher proportion of patent ductus arteriosus (PDA, Q250) in multiple births is frequently observed [16][17][18] and is attributed to the prematurity or shorter gestational age of multiples [17,19]. Since PDA was also frequently seen among multiple births in the present data [15], the analyses were performed both with and without PDA.

Results
e results of the case classi�cation by the EUROCAT algorithm are shown in Table 1. Two hundred and nine cases with more than one disease were classi�ed into 68 isolated single anomalies, 36 chromosomal cases, 9 syndromes, and 96 potential MCAs. us, �nal classi�cation of the 1,324 cases was as follows: 203 chromosomal cases (15.3%), 45 syndromes (3.4%), 980 (74.0%) isolated single anomalies, and 96 (7.3%) potential MCAs. About half of the cases with more than one disease were reclassi�ed into other categories than potential MCAs. is result was nearly the same as that obtained for the EUROCAT registries, although the present results were "potential" MCAs. About 17% (16/96) of MCAs included more than two anomalies. e perinatal characteristics of the isolated cases group and the MCAs group are shown in Table 2. e frequency of early neonatal death was signi�cantly higher among the MCAs group compared to the isolated cases group (9.8% versus 3.7%, resp.). ere was no difference with regard to method of fertility treatment, blastocyst transfer, maternal age, gestational weeks, plurality, sex, or percentage of stillbirths between the isolated cases group and the MCAs group.
e proportions of cases classi�ed as isolated anomalies, syndromes (chromosomal, monogenic, and environmental), and MCAs are shown according to ICD-10 major organ system categories in Table 3. e rate of MCAs was around 30% in "congenital malformations of eye, ear, face, and neck (Q10-Q18), " "congenital malformations of the respiratory system (Q30-Q34), " and "congenital malformations of genital organs (Q50-Q56). " On the other hand, rate of MCAs was 8% in "congenital malformations of the circulatory system (Q20-Q28). " e patterns of association of ICD-10 subcategory in MCAs are shown in Table 4. Twenty-four cases (25%) had more than one disease within the same ICD-10 subcategory. "Congenital malformations of the circulatory system (Q20-Q28), " "other congenital malformations of the digestive system (Q38-Q45), " and "congenital malformations and deformations of the musculoskeletal system (Q65-Q79)" tended to co-occur. "Cle lip and cle palate (Q35-Q37)" also tended to be associated with other congenital anomalies. Forty-four different patterns of combination at the ICD-10 organ system subcategory level were detected for 96 MCAs. But the diseases co-occurrence at a more detailed disease classi�cation level, namely, the two-digit code aer Q (e.g., Q21 for congenital malformations of cardiac septa), was rarely observed. Eighty-four cases out of 96 MCAs (88%) had different combination patterns at the two-digit code aer Q ISRN Epidemiology 5 T 3: Proportions of isolated anomalies, syndromes (chromosomal, monogenic, and environmental), and MCAs according to the ICD-10 major organ system categories.

ICD-10
Isolated Syndromes MCAs For an explanation of ICD-10 Q codes, see Table 3. Percentage was calculated for the total MCAs ( ).
level. e combination pattern occurring in three or more cases across organ system subcategory were the following: Q17 (other congenital malformations of ear) and Q21 ( 4), Q21 and Q35 (cle palate) ( ), and Q21 and Q37 (cle palate with cle lip) ( ). ere were 7 MCAs cases that included PDA (Q250). If PDA was excluded, three cases transferred to the isolated cases group. us, the percentage of potential MCAs became 7.0 (93/1,324). e other four cases contained more than two anomalies.

Discussion
e main aim of this study was to show descriptive epidemiologic/statistic data on the MCAs using nationwide data and not to refer to the pathogenesis of MCAs. e subjects of this study were all live births/stillbirths aer ART. Concerning the comparison of congenital anomalies between spontaneous and ART, larger studies have suggested that children born aer ART have an increased risk of congenital anomalies compared with children conceived spontaneously [20]. Data from meta-analyses consistently suggest that the overall risk of major birth defects in children born aer ART is about 30 higher than in children conceived spontaneously [21,22]. A nationwide survey in Sweden also showed a slightly increased risk for congenital anomalies aer IVF (in vitro fertilization), even adjusting for possible confounding factors, such as year of birth, maternal age, and parity [23]. To my knowledge, there was no population/registry-based study examining ART and MCAs. In other words, the present study sheds light on the necessity for this kind of study.
e percentage of potential MCAs was around 7 in the present subjects. e present distribution of percentage of the four main categories classi�ed using the EUROCAT algorithm was almost the same as those of European registries (Table 1), although the present percentage of MCAs included potential false-positive cases, not necessarily true cases. If more frequent and improved chromosomal analysis and molecular testing become widely used, the percentage of any chromosomal abnormalities/syndromes would become higher.
Given the complicated nature and severity of MCAs, higher early neonatal death observed in the MCAs group than in the isolated cases group seemed reasonable. e percentage of multiple births among the present subjects was obviously higher than that of the general population due to the effect of ART [24]. In addition, the prevalence of congenital anomalies in multiple births is higher than that of singletons [17,18,[25][26][27]. ere was no difference in the percentage of MCAs between singletons and multiple births in the present study. One possible reason was that multiple births aer ART usually produce dizygotic twins, and their pre-and perinatal outcome, including the prevalence of congenital anomalies, is better than that of monozygotic twins [28]. Unfortunately, the present data have no information on the correct zygosity for the same-sexed twin pairs. erefore, the higher proportion of congenital anomalies in multiple births compared to singletons frequently seen in spontaneous pregnancies or total pregnancies may be diminished in ART pregnancies.
e percentage of MCAs according to ICD subcategory and their combination patterns for several speci�c congenital anomalies were also analyzed. According to the recent systematic population-based study by Miller et al. [29], among 7,984 live-born and stillborn infants and fetuses with congenital heart diseases (CHDs), 71.3% had isolated anomalies, 13.5% had MCAs, 13.1% had syndromes, and 2.0% had laterality defects. e rate of CHDs in other population-based epidemiologic studies ranged from 6.5% to 14.1% [30][31][32][33]. e present percentage (8.1%) of MCAs was well within this range. Contrary to the high prevalence of CHDs, multiple cases occupied a small percentage compared to the percentage of MCAs in other ICD categories, as shown in Table 3.
In a study in a tertiary hospital for the treatment of oral cle patients, Barbosa et al. [34] reported that 21 patients had CHDs among 220 child patients with cle lip and/or cle palate (9.5%). In the present subjects, 6.9% (6/87) of subjects with cle lip and/or cle palate had any CHD (not shown in Table).
Two or three diseases of the circulatory system (Q20-Q28), the digestive system (Q38-Q45), and the musculoskeletal system (Q65-Q79) tended to co-occur in the same individual, as shown in Table 4. According to Miller et al. [29], for cases with any CHDs as a group, the most common noncardiac defects were any skeletal defects (35.0%), followed closely by gastrointestinal (25.2%) and renal defects (23.1%). Some of these associations might be explained by a common pathogenic mechanism, that is, early differentiation and migration of mesodermal cells for cardiac and skeletal disease or disturbances of neural crest cell migration for cardiac and gastrointestinal disease [29].
Most of these combinations, which occurred at the main organ system level, however, disappeared at a more detailed combination of anomaly level in each subcategory, suggesting the complicated causes of MCAs. e wide variation in the proportion and type of MCAs reported has been due mainly to differences in the types of case (whether based on autopsy report, clinical study, or epidemiologic study) and defect classi�cations [29]. Moreover, there is some debate about whether individual defect types, isolated or MCAs, represent etiologically distinct groupings [7]. According to Khoury et al. [12], expanding the de�nition of MCAs from cases with three or more defects to cases with two or more defects not only increases the reporting case load but also includes in the database some infants who have the two defects in question on the basis of chance alone. If combinations of defects must be monitored, then the number of possibilities increases greatly. e number of biologically possible combinations would be lower, and a compromise between splitting and lumping of diseases combination must be achieved [12]. erefore, a decision must be made on a list of defects to monitor together with their combination to minimize the problems mentioned above. e statistical methods of classifying MCAs have progressed [1,11,12,[35][36][37]. Nevertheless, a lower prevalence of each MCA makes sophisticated statistical analysis difficult. A population-based twin study using multivariate genetic analyses based on structural equation modeling is one of the most powerful tools to clarify the genetic/environmental contribution to the comorbidity of the diseases [38], including congenital anomalies. Most registries of congenital anomalies do not seem to have precise information on the twin or multiple births cases, including information on concordance/discordance of pairs, sex combination of pairs, and their zygosity. Record linkage between congenital anomalies, including information on the MCAs, and twin registry would make effective statistical genetic analysis possible.
is study has the following limitations, most of which could be attributed to the dataset and due to the fact that individual information was obtained only from the subjects with congenital anomalies. e �rst and greatest limitation is that the author could not check the reliability of the data directly. is is the essential limitation of existing data analysis. Second, the author could not control for confounding factors that can affect ART and/or congenital anomalies, such as maternal age, parity, smoking, socioeconomic status, medical history, and prenatal care. However, to date, many studies have not necessarily controlled for confounding factors. ird, followup aer birth was limited to the neonatal period at the latest and was incomplete. Some congenital anomalies are not obvious within a few days aer birth.

Conclusions
MCAs comprised about 7% (96/1,324) of cases among multiyear nationwide data on congenital anomalies aer ART in Japan. e combination patterns of diseases were widely varied. Of them, two or three diseases that belong to the circulatory system, the digestive system and the musculoskeletal system tended to co-occur in the same individual. Future research based on the standardizing populations and methods of analysis is needed.