We collected information regarding maternal, obstetric, and infant characteristics of successful pregnancies achieved through IVM between years 1999 and 2010 through structured questionnaires sent to 31 assisted reproduction centers located in 22 areas/countries around the world [25]. Specific questions were asked about the presence or absence of any birth defects in the newborns and the description of the defects. The questionnaire drew on data collected at the time of birth, and therefore would include stillbirths but not pregnancy terminations.

Reported birth defects were categorized according to ICD-10, version 2007 [29]. A major birth defect was defined as one that generally causes functional impairment or requires surgical correction [30]. We compared the prevalence of MBDs in IVM infants with two different control groups: (i) infants conceived by IVF and (ii) the general population. First, we conducted a systematic literature review of MEDLINE, EMBASE, and OVID MEDLINE(R) In-Process and Other Non-Indexed Citations databases to identify all studies reporting prevalence of birth defects in IVF infants. Studies with overlapping data were excluded. In case of partially overlapping studies, the study with the larger sample size was included. Studies that did not use a specific definition for MBD, with a sample size of less than 100 IVF children, and studies in which the children were examined after the age of 30 months were also excluded. Annual report of the Centre of the International Clearinghouse for Birth Defects Surveillance and Research (ICBDSR) was used to estimate the prevalence of MBDs in the general population across different countries [31]. ICBDSR is an international organization that affiliated the World Health Organization [32]. ICBDSR conducts worldwide surveillance and research into the occurrence and possible causes of birth defects. ICBDSR operates an international program for regular exchange among its members of information on birth defects in populations covered by the member’s surveillance and research programs. The 2009 ICBDSR annual report, used for this study, included data from 38 member programs [31]. Data collected on selected major birth defects, including trisomies 13, 18, and 21, are presented separately for live births, stillbirths, and pregnancy terminations. The number of major birth defects observed in live births and stillbirths were extracted and combined for each member program. Pregnancy terminations were excluded because we did not have available data for pregnancy terminations for IVM pregnancies.

The heterogeneity of available data on the prevalence of birth defects in the selected control groups, e.g., differences in population characteristics and varying definitions of birth defects , prevented converting the data into a single metric. Therefore, we calculated 95% confidence intervals around the reported crude point estimates in the selected studies and registries, as well as around the prevalence observed in IVM infants. The results were combined in relevant subgroups and sorted in increasing order of point estimates to provide an overview of how the prevalence of MBDs in IVM infants compared with the controls. Of the 1187 IVM infants, a total of 18 were diagnosed with a major birth defect, giving a prevalence of 1.27% (95% CI: 0.81–2.0%).

Twenty-seven manuscripts were considered eligible for our review. Four of the included studies were prospective [30, 33–35], one had a cross-sectional design [36], and the majority was retrospective. However, population-based registries which prospectively collected data were used to identify IVF children and those with birth defects in eight [37–44] out of 22 retrospective studies. Sample size ranged from 150 to 9175 IVF infants. In total, 54,678 IVF children were included. The prevalence of MBDs observed in IVF infants ranged between 0.37% (95% CI: 0.06–0.021%) [45] and 9.1% (95% CI: 7.9–10.4%) [46]. The observed prevalence of MBDs in IVF infants was higher than the observed prevalence in the IVM cohort in 24 studies, and lower in only four studies [35, 41, 45, 47]. However, the 95% confidence intervals of the observed prevalence of MBD in these four studies and in the IVM cohort were overlapping (Fig. 26.1).

The total number of births, including live and stillbirths, reported by the ICBDSR member programs was 2,637,638. MBDs were reported for 32,572 children, yielding an overall prevalence of 1.24% (95% CI: 1.22–1.24%). Prevalence of MBD per birth ranged between 0.2% (95% CI: 0.18–0.26%) and 2% (95% CI: 1.9–2.1%) in the individual registries reporting to ICBDSR. Of the registries collecting information until 2 years of age or less, 15 registries reported a lower prevalence of MBDs than that observed in IVM infants, while 16 reported a higher prevalence (Fig. 26.2). The prevalence of MBDs observed in IVM infants was not higher than the overall prevalence from the registries.

The overall MBDs prevalence of 1.27% (95% CI: 0.81–2.0%) per IVM infant born in the largest population of IVM infants to date appears to not exceed the prevalence of MBDs in either IVF children or the general population. In our view, this represents the most reliable estimate of the prevalence of MBDs in IVM infants available so far.

The lack of information on pregnancy terminations, especially on terminations due to congenital anomalies, can have caused underestimation of the overall prevalence of MBD in IVM pregnancies. However, the figures reported for both control groups also excluded information on pregnancy terminations. While none of the IVF studies included in the first control group presented data on pregnancy terminations, we excluded pregnancy terminations available in the ICBDSR report. Hence, even if the overall prevalence of MBD in IVM pregnancies can be underestimated to some extent, the comparisons with IVF children and population-based registries are unbiased in this regard. The data on IVM infants were mostly collected at birth or shortly after delivery, and this can have led to the underestimation of the true prevalence. However, MBDs are mostly evident at birth, and about two-thirds are diagnosed in the first week of life [38]. In order to prevent a bias in favor of the IVM cohort, we excluded studies in which IVF children were examined for MBDs after the age of 30 months. Similarly, when ICBDSR member registries were categorized according to the time of assessment for MBD, the IVM cohort still remained within the range of the countries where the children were examined no later than the first week of life (Fig. 26.2). Therefore, we believe the comparisons to be unbiased in this regard as well.

The population-based registries contributing to the ICBDSR data did not exclude children conceived after IVF. However, IVF children constitute a small fraction of the total births. Even if their risk of having a malformation at birth is 30–40% higher, as suggested by recent meta-analyses [8, 48], it is unlikely that their presence would distort population-based data to a substantial extent.

Although our cohort represents the largest available IVM population, the numbers are small, especially in order to comment on any specific malformations. For instance, hypospadias was the only MBD observed in more than one IVM infant, resulting in a prevalence of 2.1 per 1000 births. The overall prevalence of hypospadias was 1.7 per 1000 births in the ICBDSR report [31]. Schieve et al. [49] calculated that more than 3000 ART live births were required to demonstrate, as statistically significant (at 5% level), a threefold increase in the prevalence of hypospadias. More than 50,000 ART live births would be required to demonstrate a 1.5-fold increase [49].

Four other studies have reported on the prevalence of birth defects in IVM infants [18, 22–24]. Cha et al. [18] reported on 38 women with PCOS who conceived after IVM and had obstetric follow up at the authors’ institution. Following elective reduction in triplets, spontaneous miscarriages (including one fetus with omphalocele and 45X/46XY mosaicism), and one medical abortion for hydrops fetalis (which is not considered a congenital anomaly), there were 20 singleton and 4 twin live births. One out of 28 children born had a cleft palate, resulting in a prevalence of 3.6 per 100 live births. Fifty-five IVM children delivered at the McGill University Health Centre were already included in the data presented above [22]. Foix-L’helias et al. [23] reported on 35 IVM children evaluated at one year of age in the context of a prospective study and reported that prevalence of congenital anomalies was similar to that in IVF children. Fadini et al. [24] did not observe any major congenital anomalies in their group of 196 IVM infants, but observed several minor defects, albeit at a similar rate in ICSI infants.