Based on the data retrieved from the CBDMN, our study found that in the northern region, especially in the north-rural region, women have a higher risk of NTDs than their counterparts in the southern region. Nevertheless, NTD subtypes have a different geographic distribution at the provincial level.
The results of our study show that geographic disparities exist in total prevalence of NTDs. Northern rural China still has the highest prevalence of NTDs in the world. Similar geographic disparities were also reported in the United Kingdom and the United States
[19, 20]. Although these studies tried to explain why the geographical differences existed, the true reasons remain unclear. A previous study showed that the mutation frequency of methylene tetrahydrofolate reductase (MTHFR) C677T among people living in northern China was higher than that in southern China and a foetus with this mutated gene had a higher risk of NTDs
[21, 22]. Nevertheless, this is inadequate to explain the existence of such a significant difference between regions. Approximately 5% of NTD cases have a family history of disease, and therefore, environmental factors and interaction with genetic factors are also the important causes of NTDs. Although the cause of NTDs in humans is still poorly understood,many studies have shown that environmental factors, including maternal infections, vitamin B deficiency, maternal cigarette smoking, maternal exposure to secondhand smoke, alcohol, caffeine, exposure to chemical fertilizers and pesticides, hyperthermia, maternal diabetes, maternal obesity, taking antipyretic drugs or antibiotics, a high consumption of dried or pickled vegetables, and maternal fumonisin exposure, can increase the risk of NTDs
[2–4, 23–26]. We think the difference in the prevalence of NTDs between the northern and southern regions may be related to the lifestyle of its citizens, hazardous environmental exposure, and the vitamin supplementation from food. For example, in the north, rural areas in particular, coal is the main fuel used for energy processes, which results in substantial indoor air pollution
. This is not the case in the south. Additionally, women in the north perform fewer outdoor activities because of cold weather, and therefore, maternal indoor air pollution exposure in the north is more serious than in the south. Folate deficiency, intake of less green vegetables, and wheat and corn-based food in the winter are also important reasons for a high prevalence of NTDs in the north. A previous study showed that among rural women in the northern region, approximately 50% and 43% were deficient for plasma and red blood cell folate, respectively, compared with 6% and 4%, respectively, for rural women in the southern region
. Because of the cold climate, residents in the north (especially in the north-rural areas) consume less fresh vegetables and more pickled vegetables than in the south. However, vitamin B deficiency and high consumption of pickled vegetables are two important risk factors for NTDs
[4, 27]. A higher possibility of consumption of fumonisin-contaminated corn-based food among the northern women may be another reason for the high prevalence in the north
. In addition, the report ‘Tobacco Controls in 2007 in China’ shows that the proportion of women in the north who are exposed to secondhand smoke was higher than in the south
. This may help explain the disparities in NTD PRs between the northern and southern regions, which are more obvious in the rural areas.
Previous studies simply revealed south–north differences in the total prevalence of NTDs at the regional level. The strength of our study is that it shows, for the first time, that the three subtypes of NTDs have different geographic distributions at the provincial level in China. Not all NTD subtypes are more prevalent in the northern provinces than the southern provinces. The different geographic distributions of NTD subtypes may provide us with some ideas for developing suitable local intervention strategies. It is worthy to note that the most prevalent NTD is anencephaly in the south and spina bifida in the north. The subtype composition of NTDs differs within each country as well. The major NTDs in Iran is anencephaly and spina bifida in the USA, Canada, and New Zealand.
 It indicates that the causes of the three subtypes of NTD differ and that these differences need further exploration.
An urban–rural disparity in the prevalence of NTDs was also shown in this study. The differences in residents’ education level, health awareness, and maternal nutritional status between urban and rural areas, likely contribute to this disparity. A previous study showed that an educational level of primary school or lower was significantly associated with increased risk for an NTD
. It is reported that 90% of illiterate peoples in China reside in the rural areas. A study on folic acid awareness and intake among women in the 6 northern province of China showed that the proportion of women who knew about folic acid and who use folic acid supplements was 73.0% and 10.5%, respectively, in the urban areas, compared with 50.0% and 7.6%, respectively, in the rural areas.
 A large proportion, 40%, of rural women were deficient for red blood cell folate in the north, compared with 20% in the south.
 In addition, exposure to chemical fertilisers and pesticides during pregnancy is also associated with a relatively high prevalence of NTDs in rural areas
Our study has some limitations. (1) It is possible that hospital-based samples may introduce a referral bias because approximately 15% of birth cases are delivered in the hospitals at the county level of lower. If the foetus is suspected to have congenital anomalies during the prenatal examination, his or her mother is usually at a higher-level hospital that has the capabilities to diagnosis it. It can be estimated that a proportion of these women are willing to deliver in the higher-level hospital to obtain better medical services. This may lead to an overestimation of the total prevalence of NTDs. However, although cases with congenital anomalies are diagnosed in hospitals at the county level or higher, there is also a proportion of mothers who are willing to deliver in the hospitals at the county level of lower because of their relatively low-cost medical services. These cases were not included in our surveillance system. Additionally, because the data used in this study includes wide geographical coverage and a large sample size, the estimation of the PR of NTDs is relatively stable. (2) The denominator of the prevalence of NTDs was the total number of births (live births and still births at greater than or equal to 28 weeks of GA) in the same area and time period. Although inclusion of induced and spontaneous foetal deaths at less than 28 weeks of GA would more closely approximate the incidence of NTDs, it is very impractical, as these pregnancy outcomes are often inaccurately counted compared to live births and stillbirths. In addition, thenumber of induced and spontaneous foetal deaths are small in comparison to the number of live births and stillbirths and are unlikely to greatly affect prevalence. (3) In our NTDs database, the minimum gestational age of all NTDs cases is 12 weeks. Almost all NTD cases that are spontaneously aborted during the first trimester cannot be detected because of the limits of detection technology in our birth defects surveillance network, and therefore, the total prevalence of NTDs shown in our study is underestimated from this perspective. Besides, the ability to ascertain NTDs is different among the member hospitals at different-levels (county-level, municipal-level hospitals and provincial-level) in CBDMN. Approximately 85% of deliveries in the county-level hospitals came from rural areas and 90% of deliveries in the municipal or provincial level hospitals from urban areas. Relatively low ability to ascertain NTDs in the county-level hospitals may induce to underestimate the total NTDs prevalence in the rural.