Unevenly distributed: Twins in developing countries

This new information is now published in the scientific journal PLoS One.

Sociologist/economist Jeroen Smits has brought together demographic, socio-economic and health data on millions of households in over one hundred developing countries into one huge database called Database Developing World. The information in this database is being used in various research studies involving inequality. In collaboration with sociologist Christiaan Monden, Smits has used this information to create a twins database for 76 developing countries. The database contains information on approximately 2.5 million births by 1.4 million women. This information was collected between 1987 and 2010 in 150 large household surveys. The purpose of the twins database is to study inequalities between girls and boys in terms of education, infant mortality, health and related factors.

New information

The twins database is therefore a means, not an end. Until now, little was known about twinning rates in developing countries. The information that was available was generally derived from hospital registrations, which are notorious for their selectivity problems: as twin births are associated with more complications, these births tends to be overrepresented among hospital births. Hospital records therefore do not provide a good picture of the national situation.

High twinning rate in Central Africa, low rate in Asia

• On average, 13.6 twins per 1000 births were born in the 76 countries included in this study. The dark band across Central Africa indicates a very high twinning rate in this zone: more than 18 twins per 1000 births. The twinning rate in Asia and Latin America, on the other hand, is very low: often less than 8 per 1000 births and almost always less than 10 per 1000 births.
• The existence of low national twinning rates in Asia was only known for the more developed countries of this region (Japan, Korea, Taiwan and Singapore).
• The low twinning rate in India contradicts the previously held believe that twinning among the predominantly Caucasian Indian population would be at an intermediate level, comparable to that in European countries before the introduction of fertility treatments (9-16 twin births per 1000 births).
• For Africa, the myth is broken that twinning rates in Nigeria are the highest in the world. High national twinning rates are found throughout Central Africa and in several countries twinning incidence is higher than in Nigeria. With 27.9 twins per 1000 births, Benin has the highest national average.
• Little was known about national twinning rates in Latin America. The new data reveal that on the Middle and South American mainland, twinning rates are just as low as those in Asia. The major exception are the Caribbean islands, in particular Haiti (14 twins per 1000 births), where many people from African descent live.
• The data included in the twin database are data on live births. As twins are prone to somewhat higher mortality rates during pregnancy and birth, the actual number of twins could be somewhat higher. Given that Sub-Saharan Africa has the highest infant mortality rate, the differences between this region, Asia and Latin America could therefore be even higher than can be read from the image.

Further research The researchers do not provide an explanation for the vast differences in twin birth rates among the different regions of the developing world. The literature reveals that genetics plays a significant role in twin births and substantial differences have been perceived in the twinning rates of various ethnic groups. In Nigeria, for example, the number of twins born to Yoruba women is much higher than those born to Hausa women. In the United States, more twins are born to women of African descent than to women of European descent. Differences in the age at which women get their children and in the number of pregnancies they experience also play a role. Older women more often get twins than younger women and in a fourth or fifth pregnancy, the chances of getting a twin are higher than in a first or second pregnancy. The new twins database offers many opportunities for further research into the causes of the observed differences.