Abstract Transcriptional enhancers regulate gene expression in time and space, commonly engaging in long-range chromosomal contacts with gene promoters. However, the relationship between enhancer activity, enhancer-promoter contacts and gene expression is not fully understood. Here, we leveraged human genetic variation as a “natural perturbation” to dissect this relationship, focusing on distal enhancers containing expression quantitative trait loci (eQTLs) – genetic variants linked to specific gene expression levels. We devised eQTL-Capture Hi-C to profile the chromosomal contacts of these loci globally and at high resolution in primary monocytes isolated from 34 donors, and generated chromatin accessibility and gene expression profiles from the same samples. Extending a Bayesian approach that considers both intra- and inter-individual variation, we detected 19 eQTLs linked to distinct promoter contacts, most of which also associated with enhancer accessibility and activity. Capitalising on these shared effects, we next employed a multi-modality Bayesian strategy, identifying hundreds of variants jointly associated with enhancer activity, connectivity, and gene expression. Many of these variants influenced the predicted binding of the architectural protein CTCF and the core myeloid transcription factors GABPA and SPI1; however, they typically did not perturb the canonical binding motifs of these factors. In contrast, one variant associated with PCK2 promoter contact directly disrupted a CTCF binding motif and impacted the insulation of this promoter from downstream enhancers. Finally, many identified QTLs overlapped with disease susceptibility loci, underscoring the potential role of enhancer-promoter communication in mediating the pathological effects of non-coding genetic variation. Jointly, our findings suggest an inherent functional link between the activity and connectivity of enhancers with relevance for human disease, and highlight the role of genetically-determined chromatin boundaries in gene control.