Prevalence of birth defects in an Arctic Russian setting from 1973 to 2011: a register-based study

The study population included all newborns delivered in Monchegorsk and registered in either the Kola Birth Register (KBR) or the Murmansk County Birth Register (MCBR) for the years 1973–2011.

The KBR was established in 1998 by the retrospective collection of information about all births from week 28 of gestation on that occurred in Monchegorsk as of March 1973. Registration was continued prospectively until 2005. In addition, 102 newborns with age 13–27 weeks of gestation were also registered, of which 82 were stillborn or died during the first week of life. Details of the construction and description of the register’s suitability for epidemiological investigations have been published [8, 9, 16]. A total number of 26 841 pregnancy outcomes were registered. The implementation and use of the KBR demonstrated the need for a medical birth registry for the entire Murmansk region (also referred to as the Kola Peninsula).

The MCBR was established in 2005, and the prospective registration of pregnancy outcomes from 22 weeks of gestation began on the 1^st of January 2006 [17]. A total of 3750 births in Monchegorsk were registered during 2006–2011.

Data from the two registers were merged into one database using the same fields: maternal date of birth; child’s birth date; and status of child (LB or SB); BD diagnosis and its ICD-10 code. Newborns were excluded from the study if diagnoses at birth had been missed, or its interpretation had been uncertain (i.e., not having an ICD-10 code). We also excluded newborns with missing data about their status at birth. The total number of excluded newborns was 144 (0.5%), and thus 30448 newborns were included in the current analysis.

The registration of BD in the two registers included information about the existence of BD and its diagnosis; the latter was based on primary medical documentation (up to five fields of diagnosis), and were coded according to ICD-10. The diagnosis was usually made by a pediatrician or neonatologist after birth and during the maternity house stay. Use of pertinent diagnostic tools such as ultrasound examination (US) or echocardiography supplemented routine examinations. In some cases, prenatal US evidence for suspected BD was confirmed by examination after birth. In case of abortion or fetal death, the diagnosis was based on autopsy results.

The prevalence, structure of BD, and distribution of the different forms were estimated based on the two- and three-level of International Classification of Disease, the Tenth Edition (ICD-10), Chapter XVII, Q codes. The newborns with more than one BD diagnosis were included in the analysis only as newborns with multiple defects as their diagnoses were not coded as Q89.7 (“multiple congenital malformations, not elsewhere classified”). These newborns were not included for any specific defects they had.

The total prevalence and prevalence by BD groups were calculated using the number of newborns (LB plus SB) with BD as the numerator, with the total number of newborns included as the denominator. To make the prevalence data more comparable with those of EUROCAT, we calculated the total BD prevalence and that of all BD malformation groups after excluding all minor anomalies according to the EUROCAT Guide 1.4, Chapter 3.2: “Minor malformation for exclusion” from the total number of BD [18]. The prevalence of malformations for which registration was mandatory in the RF was calculated using the sum of newborns with the BD items listed in Table 1 as the numerator.

Descriptive statistics were calculated using SPSS 21.0. The rates are presented per 1000 newborns (LB plus SB), with 95% confidence intervals (95% CI). A time-trend calculation was done for four 10-years periods (the last one included nine years) using chi-squared test for trend.

The Committee for Research Ethics at the Northern State Medical University (Arkhangelsk, Russia) and REK Regional Committee for Health and Research Ethics, Northern Norway (Tromsø, Norway) approved the current study.

Although the observed prevalence of all BD for the study period was higher than reported in other European registers [1, 20], improved agreement with the latest EUROCAT data was evident after minor anomalies were excluded. However, that for the last decade was higher than in Europe, even after adjustments using the EUROCAT guidelines.

A comparison of our results with other available Russian monitoring data shows that in Monchegorsk the total prevalence was also higher than the mean Russian value for the 2003–2011 period. By contrast, the prevalences for BD that require mandatory reporting were approximately equal [13]. However, use of official monitoring data in the assessment of BD prevalences likely involve less rigorous data collection than that exercised for birth registers, and perhaps is also prone to under-reporting less severe forms of defects. Our prevalence estimates of genital malformations and defects of musculoskeletal system were lower than those reported for Monchegorsk by Vaktskjold et al. [8, 9]. Our neural tube and anterior abdominal wall defect prevalences were also lower than those reported for Arkhangelsk by Petrova and Vaktskjold as incidences (abortions after 12 weeks of gestation were included) [10, 11].

Compared to EUROCAT data (1980–2011) [1, 20], the prevalence of cardiovascular malformations in Monchegorsk was lower; those of musculoskeletal and urinary BD were higher; and of comparable magnitude for other malformations (see Table 3). Prevalences of the severe malformations for which registration in Russia is mandatory were mostly similar, but were higher for Down syndrome, severe cardiovascular malformations and diaphragmatic hernia. A likely reason for this is that EUROCAT includes cases of termination of pregnancy due to fetal anomaly (TOPFA), while in the RF mandatory autopsies of aborted fetuses are performed only after 22 weeks of gestation.

The low rate of cardiovascular defects in Monchegorsk can at least be partly explained by an underestimation of the true number of minor chamber defects because of an absence of symptoms during the first week of life. A prenatal diagnosis in regional districts of severe malformation that potentially could be surgically corrected may have led to a transfer of a delivery to regional (Murmansk) or federal (Moscow) centers. Such births were not registered in either of the registers.

Our data for BD that required mandatory reporting are comparable with those available from the Medical Birth Registry of Norway [21] for the same time frame. However, the following prevalences were higher in Norway: transposition of great vessels (0.21/1000), cleft lip with or without cleft palate (1.36/1000), Down Syndrome (1.34/1000), and lower for congenital hydrocephalus (0.43/1000) [21]. The total prevalence of all defects and deformations in the observation period years was also comparable, although the Norwegian data includes TOPFAs from 2000 and based on this we might expect lower values if they had not been.

The positive trend of the total BD prevalence across the study period could be the result of an increase in the prevalence of congenital malformations of the genital organs and urinary system. However, the interpretation of such dependence is complex and might well reflect changes in the health-care system, birth registration protocols, the socio-economic situation and changes in coding practices (the KBR was established retrospectively and historical codes were re-assigned during the past 40 years to conform with ICD-10). Furthermore, the prevalence of compatible-with-living defects, such as urinary malformations, has increased because most of them can be easily visualized by US-screening [22]. Interestingly, the most severe BD that are often incompatible with life (such as anencephaly) did not exhibit substantial changes in prevalence after the 1970–1980 period. This observation presumably reflects an impact of prenatal diagnostics established in 1994, which included one US examination and more complex prenatal screening procedures after the year 2000. Evidently, these improvements in the prenatal detection led to the reduction of severe malformations and better diagnosis of minor ones at the first week of life.

An increase in the prevalence of urinary system malformations during 2003-2011was no doubt influenced by an increase of two defects, namely Q62.0 (congenital hydronephrosis) and Q63.0 (another malformations of kidney, unspecified). Both forms could be symptomless during the first days of life and thus could only have been diagnosed during the time-frame when US-examinations were conducted. On the other hand, the observed increased occurrence of these forms could also reflect over diagnosis, but this would require a detailed follow-up study to verify.

Increased prevalence of BD over time could also represent a true rise, perhaps due to environmental factors. For example, the prevalence of some group of BD (malformations of genital organs, musculoskeletal abnormalities and multiple BD) rose two- to three-fold between 1973 and 1992, which overlaps increased production at the local nickel refinery complex that reached a maximum between 1982 and 1988. The observed peak in BD prevalence during 1983–1992 (after exclusion of urinary system defects) also coincides. The primary pollutants were sulfur dioxide and particulates containing a suite of inorganic elements including toxic metals and nonmetals [23]. The latter can accumulate in soil and be transferred to the watershed over time. Although these emissions have been systematically reduced since the early 1990s, they remain substantial. Other possible explanations of increasing BD prevalence include higher maternal age and an increase in alcohol abuse and smoking among mothers; specifically, 30.8% of women in Russia of reproductive age were smokers in 2008–2010 [24] compared to 19% in 1990 [25].

Our study is the first to examine all types BD prevalences recorded in two population-based registers in Russia that can be compared with European registers. A possible limitation of our study pertains to the retrospectively established KBR database. Only diagnoses made in the maternity houses were taken into account, which could be a reason for underestimation. Minor anomalies or defects that might be revealed later (e.g., small septal heart defects without heart failure) were not identified. Neither were TOPFAs included in either database, which could mean that the prevalence could be higher than calculated after the year 2000 when prenatal screening was established. On the other hand and in terms of the register data, only a small number of newborns had missing information (0.5%). Difference in the gestational age limits of registration of the newborns in the merged registries and retrospective type of data collection in the KBR could potentially have influenced the results, although this would not explain the total BD prevalence increase seen for 2003–2011. There were 17 newborns under 28 weeks in 2006–2011 registered in the MCBR. Among them, only three additional cases of BD were found and the total number of newborns with BD for that period was 134. Inclusion of these data did not change the BD prevalence significantly. Furthermore, when the 102 newborns with gestational age less than 28 weeks noted in the KBR were included in the analysis, there was also little impact on our findings. One might have expected the prevalence of the most severe BD to have increased as well, but we did not observe this.