Prenatal Fish Oil Supplementation, Maternal COX1 Genotype, and Childhood Atopic Dermatitis

Importance Eicosanoids have a pathophysiological role in atopic dermatitis (AD), but it is unknown whether this is affected by prenatal ω-3 long-chain polyunsaturated fatty acid (n-3 LCPUFA; ie, fish oil) supplementation and genetic variations in the cyclooxygenase-1 (COX1) pathway. Objective To explore the association of n-3 LCPUFA supplementation during pregnancy with risk of childhood AD overall and by maternal COX1 genotype. Design, Setting, and Participants This prespecified secondary analysis of a randomized clinical trial included mother-child pairs from the Danish Copenhagen Prospective Studies on Asthma in Childhood 2010 birth cohort, with prospective follow-up until children were aged 10 years. In the trial, maternal and child COX1 genotypes were determined, and urinary eicosanoids were quantified when the child was 1 year of age. The present study was conducted from January 2019 to December 2021, and data were analyzed from January to September 2023. Intervention A total of 736 pregnant women at 24 weeks’ gestation were randomized 1:1 to 2.4 g of n-3 LCPUFA (fish oil) or placebo (olive oil) per day until 1 week post partum. Main Outcomes and Measures Risk of childhood AD until age 10 years overall and by maternal COX1 genotype. Results At age 10 years, 635 children (91%; 363 [57%] female) completed the clinical follow-up, and these mother-child pairs were included in this study; 321 (51%) were in the intervention group and 314 (49%) in the control group. Pregnancy n-3 LCPUFA supplementation was associated with lower urinary thromboxane A2 metabolites at age 1 year (β, –0.46; 95% CI, –0.80 to –0.13; P = .006), which was also associated with COX1 rs1330344 genotype (β per C allele, 0.47; 95% CI, 0.20-0.73; P = .001). Although neither n-3 LCPUFA supplementation (hazard ratio [HR], 1.00; 95% CI, 0.76-1.33; P = .97) nor maternal COX1 genotype (HR, 0.94; 95% CI, 0.74-1.19; P = .60) was associated with risk of childhood AD until age 10 years, there was evidence of an interaction between these variables ( P < .001 for interaction). Among mothers with the TT genotype, risk of AD was reduced in the n-3 LCPUFA group compared with the placebo group (390 mother-child pairs [61%]; HR, 0.70; 95% CI, 0.50-0.98; P = .04); there was no association for mothers with the CT genotype (209 [33%]; HR, 1.29; 95% CI, 0.79-2.10; P = .31), and risk was increased among offspring of mothers with the CC genotype (37 [6%]; HR, 5.77; 95% CI, 1.63-20.47; P = .007). There was a significant interaction between n-3 LCPUFA supplementation and child COX1 genotype and development of AD ( P = .002 for interaction). Conclusions and Relevance In this secondary analysis of a randomized clinical trial, the association of prenatal n-3 LCPUFA supplementation with risk of childhood AD varied by maternal COX1 genotype. The findings could be used to inform a personalized prevention strategy of providing supplementation only to pregnant individuals with the TT genotype. Trial Registration ClinicalTrials.gov: NCT00798226