Several OPRD1 intronic variants were associated with opioid addiction (OD) in a population-specific manner. This follow-up study aims to further characterize the OPRD1 haplotype pattern of the risk variants in different populations and apply in silico analysis to identify potential causal variants. A population-specific haplotype pattern was revealed based on six OPRD1 eQTL SNPs and five common haplotypes were identified in a sample of European ancestry (CEU). A European-specific haplotype ('Hap 3') that includes SNPs previously associated with OD and is tagged by SNP rs2236861 is more common in subjects with OD. It is quite common (10%) in CEU but is absent in the African sample (YRI) and extends upstream of OPRD1. SNP rs2236857 is most probably a non-causal variant in LD with the causal SNP/s in a population-specific manner. The study provides an explanation for the lack of association in African Americans, despite its high frequency in this population. OD samples homozygous for 'Hap 3' were reanalyzed using a denser coverage of the region and revealed at least 25 potentially regulatory SNPs in high LD. Notably, GTEx data indicate that some of the SNPs are eQTLs for the upstream phosphatase and actin regulator 4 (PHACTR4), in the cortex, and others are eQTLs for OPRD1 and the upstream lncRNA ENSG00000270605, in the cerebellum. The study highlights the limitation of single SNP analysis and the sensitivity of association studies of OPRD1 to a genetic background. It proposes a long-range functional connection between OPRD1 and PHACTR4. PHACTR4, a mediator of cytoskeletal dynamics, may contribute to drug addiction by modulating synaptic plasticity.