Deletion of a non-imprinted 500kb genomic region at chromosome 15q11.2, between breakpoints 1 and 2 of the Prader-Willi/ Angelman locus (BP1-BP2 deletion), has been associated in previous studies with phenotypes including congenital cardiovascular malformations (CVM). Previous studies investigating association between BP1-BP2 deletion and CVM have tended to recruit cases with rarer and more severe CVM phenotypes; the impact of CVM on relatively unselected population cohorts, anticipated to contain chiefly less severe but commoner CHD phenotypes, is relatively unexplored. More precisely defining the impact of BP1-BP2 deletion on CVM risk could be useful to guide genetic counselling, since the deletion is frequently identified in the neurodevelopmental clinic. Using the UK Biobank (UKB) cohort of~500,000 individuals, we identified individuals with CVM and investigated the association with deletions at the BP1-BP2 locus. In addition, we assessed the association of BP1-BP2 deletions with neuropsychiatric diagnoses, cognitive function and academic achievement. Cases of CVM had an increased prevalence of the deletion compared with controls (0.64%; OR = 1.73 [95% CI 1.08-2.75]; p = 0.03), as did those with neuropsychiatric diagnoses (0.68%; OR = 1.84 [95% CI 1.23-2.75]; p = 0.004). We conclude that BP1-BP2 deletion moderately increases the risk of the generally milder, but commoner, CVM phenotypes seen in this unselected population, in addition to its previously demonstrated association in case/control studies ascertained for CVM.

BackgroundDeletion of a non-imprinted 500Kb genomic region at chromosome 15q11.2, between breakpoints 1 and 2 of the Prader-Willi/Angelman locus (BP1-BP2 deletion) has been associated in previous studies with phenotypes including developmental delay, autism, schizophrenia and congenital cardiovascular malformations (CVM). The deletion has a low baseline population prevalence and large-scale data regarding the magnitude of these associations and any milder effects on cognition phenotypes, in populations not selected for disease, are limited. MethodsUsing the UK Biobank (UKB) cohort of ~500,000 individuals, we identified individuals with neuropsychiatric and CVM diagnoses and investigated their association with deletions at the BP1-BP2 locus. In addition we assessed the association of BP1-BP2 deletions with cognitive function and academic achievement in individuals with no previous diagnosis. Results Casesof neurodevelopmental and CVM disease had an increased prevalence of the deletion compared to controls (0.68%; OR=1.84 [95% CI 1.23 -2.75]; p=0.004 and 0.64%; OR=1.73 [95% CI 1.08 -2.75]; p=0.03 respectively). Excluding participants diagnosed with neurodevelopmental or neuropsychiatric disease, deletion carriers had worse scores in four tests of cognitive function, and while 32.8% of UKB participants without BP1-BP2 deletion had a university or college degree as their highest educational qualification, only 22.8% of deletion carriers achieved this (OR 0.57 [95% CI 0.51-0.64]; p=5.60E-22). ConclusionsWe conclude that BP1-BP2 deletion has an appreciable population prevalence with important life-course impacts on undiagnosed carriers. These data are of potential utility in deciding the circumstances under which clinical testing for BP1-BP2 deletion may be helpful. Key messages• Deletions at chromosome 15q11.2 between breakpoints 1 and 2 (BP1-BP2), which encompass four genes, have been associated with developmental delay, autism, schizophrenia and congenital cardiovascular malformations.• Here, we use the largest cohort studied to date: the UK Biobank cohort of ~500,000 individuals, to explore the association of this deletion with neuropsychiatric and cardiovascular phenotypes as well as its effects on cognitive function and academic achievement.• We find an appreciable population prevalence of the deletion with these phenotypes and demonstrate reduced cognitive function and lower academic achievements in those individuals with no prior diagnosis.

Adults with congenital heart disease (CHD) face increased risk of various comorbid diseases. Previous work on lung dysfunction in this population has mainly focused on restrictive lung disease, in patients with severe CHD phenotypes. We examined the association of mild CHD with chronic obstructive pulmonary disease (COPD) in the UK Biobank (UKB). Electronic health records (EHR) were used to identify 3385 CHD cases and 479,765 healthy controls in UKB, before performing a case–control analysis over a 20-year study period for a total of > 9.5 M person-years of follow-up. Our analysis showed that UKB participants with CHD are at substantially greater risk of developing COPD than healthy controls (8.7% vs 3.1% prevalence, unadjusted OR 2.98, 95% CI 2.63, 3.36, P = 1.40e−53). Slightly increased rates of smoking were observed amongst CHD cases, however the association with COPD was shown to be robust to adjustment for smoking and other factors known to modulate COPD risk within a multivariable-adjusted Cox regression framework (fully adjusted HR 2.21, 95% CI 1.97, 2.48, P = 5.5e−41). Care for adults with CHD should aim to mitigate their increased risk of COPD, possibly via increased smoking cessation support.