The overall project aim is to study children's neurodevelopmental outcomes (including diagnoses of autism, ADHD, and intellectual disability) following exposure to maternal anemia during pregnancy or anemia during the first year of life using national and regional Swedish health-data registers, and to assess children's neurodevelopmental outcomes over the range of maternal hemoglobin levels during pregnancy.

Our overall hypothesis is that early life iron sufficiency may promote appropriate neurodevelopment and reduce risk of neurodevelopmental disorders in children. We will use Swedish register data to define a cohort of maternal-child pairs to follow from gestation through childhood. We will consider exposure to anemia during two key time periods: gestation and the first year of life. Exposure data are recorded in Swedish medical records prospectively relative to the follow-up for the outcomes.

To assess exposure to maternal anemia, we will consider maternal diagnoses of anemia during pregnancy in a national cohort of maternal-child pairs, using ICD codes recorded in pregnancy in the Swedish National Patient Register and the Swedish Medical Birth Register. In a sub-sample of mother-child pairs in Stockholm with data available from antenatal records (i.e., Obstetrix register data), we will also collect data regarding levels of maternal hemoglobin measured longitudinally over the course of pregnancy, and consider this as a continuous variable and divided into categories to reflect the full range of potential exposures from very low to very high. We will consider whether risk of the outcomes associated with exposure to anemia or decreasing hemoglobin levels vary with the time of onset during pregnancy.

To assess exposure to anemia during the first year of life, we will consider infant's diagnoses of anemia during the first year of life in a national cohort of maternal-child pairs, using ICD codes recorded during the year after birth in the Swedish National Patient Register. We will consider whether risk of the outcomes associated with exposure to anemia during vary with the time of onset during the first year of life, and whether anemia during the first year interacts with gestational age at birth.

The main outcome measures will be incidence of children's diagnoses of any of three common neurodevelopmental conditions: autism, attention-deficit/hyperactivity disorder (ADHD), and intellectual disability. Diagnoses will be ascertained using ICD codes recorded in pregnancy in the Swedish National Patient Register. For the sub-sample of mother-child pairs for children born in Stockholm, we will also ascertain outcomes in regional registers that capture further sources of outpatient treatment and care, primarily the VAL database. Secondary analyses will consider the incidence of commonly co-occurring combinations of the disorders (e.g., autism with co-occurring intellectual disability).

Inclusion Criteria:

• children born in Sweden from January 1, 1987 to December 31, 2010 (national cohort)
• children born in Region Stockholm from January 1, 2007 to December 31, 2010 (Stockholm regional cohort)

Exclusion Criteria:

• Adopted children
• Children without records in the Medical Birth Registry (MBR)

We will study the total population of children born in Sweden January 1, 1987 to December 31, 2010 (national cohort) and children born in Region Stockholm from January 1, 2007 to December 31, 2010 (Stockholm regional cohort) with maternal health and birth records available for the index pregnancy. Since records in Medical Birth Register are available for >98% of births in Sweden, we anticipate that we will need to exclude relatively few births. Children will be followed up for neurodevelopmental outcomes using Swedish health registers; the only loss to follow-up that is expected is emigration from Sweden or death of the child.