Background: ADHD is associated with multiple adverse outcomes and early identification is important. The present study sets out to identify early markers and developmental characteristics during the first 30 months of life that are associated with ADHD 6 years later. Methods: 9201 participants from the prospective Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort were included. Outcome measures were parent-rated ADHD symptom scores (Strengths and Difficulties Questionnaire, SDQ) and ADHD diagnosis (Development and Wellbeing Assessment, DAWBA) at age 7. Seventeen putative markers were identified from previous literature and included: pre-and peri-natal risk factors, genetic liability (ADHD polygenic risk scores, PRS), early development, temperament scores and regulatory problems. Associations were examined using regression analysis. Results: Univariable regression analysis showed that multiple early life factors were associated with future ADHD outcomes, even after controlling for sex and socioeconomic status. In a multivariable linear regression model; temperament activity scores (B = 0.107, CI = 0.083-0.132), vocabulary delay (B = 0.605, CI = 0.211-0.988), fine motor delay (B = 0.693, CI = 0.360-1.025) and ADHD PRS (B = 0.184, CI = 0.074-0.294) were associated with future symptoms (R 2 = 10.7%). In a multivariable logistic regression model, ADHD PRS (OR = 1.39, CI = 1.10-1.77) and temperament activity scores (OR = 1.09, CI = 1.04-1.16) showed association with ADHD diagnosis. Conclusion:As well as male sex and lower socio-economic status, high temperament activity levels and motor and speech delays in the first 30 months of life, are associated with childhood ADHD. Intriguingly, given that genetic risk scores are known to explain little of the variance of ADHD outcomes, we found that ADHD PRS added useful predictive information. Future research needs to test whether predictive models incorporating aspects of early development and genetic risk scores are useful for predicting ADHD in clinical practice.Anita Thapar and Kate Langley contributed equally to this study.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Pregnant women resident in Avon, UK with expected dates of delivery 1st April 1991 to 31st December 1992 were invited to take part in the study. The initial number of pregnancies enrolled is 14,541 (for these at least one questionnaire has been returned or a "Children in Focus" clinic had been attended by 19/07/99). Of these initial pregnancies, there was a total of 14,676 foetuses, resulting in 14,062 live births and 13,988 children who were alive at 1 year of age. When the oldest children were approximately 7 years of age, an attempt was made to bolster the initial sample with eligible cases who had failed to join the study originally. These individuals were not included in our analyses as we focussed on outcomes prior to age 7 years.The sample is relatively homogeneous in regard to ethnicity (approximately 98% White): all participants are from the Bristol area in the United Kingdom. 1 Please note that the study website contains details of all the data that is available through a fully searchable data dictionary and variable search tool: http://www.bristol.ac.uk/alspac/researchers/our-data/.Further details of the study, measures and sample can be found elsewhere. [1][2][3] Generating polygenic scoresIn total 9912 ALSPAC children were genotyped using the Illumina HumanHap500-quad genotyping array. Individuals were excluded on the basis of gender mismatches; minimal or excessive heterozygosity, disproportionate levels of individual missingness (>3%), insufficient sample replication (IBD <0.8), non-European ancestry (assessed by multidimensional scaling analysis and compared with Hapmap II) and cryptic relatedness (IBD > 0.1). SNPs were excluded based on minor allele frequency (<1%), call rate (<95%) or evidence for violations of Hardy-Weinberg equilibrium (P < 5E-7). Imputation was conducted by the ALSPAC team using Impute V2.2.2 against the 1000 genomes reference panel (Phase 1, Version 3: all polymorphic SNPs excluding singletons), using all 2186 reference haplotypes (including non-Europeans).SNPs were subsequently filtered based on minor allele frequency (<1%) and imputation quality (INFO<0.8). Sensitivity analyses utilised the top 10 principal components (in-line with previously work 4 ) generated using PLINK 5 --pca command (https://gist.github.com/explodecomputer/ab3552ee06b5c0df76139fc587373599). Following quality control and limiting individuals to one child per family, genetic data were available for N=7975.Genome-wide association study (GWAS) were filtered to remove SNPs that were palindromic, insertions/deletions, non-autosomal, INFO score <0.8, missing in N>1 study and duplicates (https://github.com/ricanney).