Causal association between circulating blood cell traits and pulmonary embolism: a mendelian randomization study

Abstract

Background: Pulmonary embolism (PE) is a life-threatening thromboembolic disease for which there is limited evidence for effective prevention and treatment. Our goal was to determine whether genetically predicted circulating blood cell traits could influence the incidence of PE.

Methods: Using single variable Mendelian randomization (SVMR) and multivariate Mendelian randomization (MVMR) analyses, we identified genetic associations between circulating blood cell counts and lymphocyte subsets and PE. GWAS blood cell characterization summary statistics were compiled from the Blood Cell Consortium. The lymphocyte subpopulation counts were extracted from summary GWAS statistics for samples from 3757 individuals that had been analyzed by flow cytometry. GWAS data related to PE were obtained from the FinnGen study.

Results: According to the SVMR and reverse MR, increased levels of circulating white blood cells (odds ratio [OR]: 0.88, 95% confidence interval [CI]: 0.81-0.95, p = 0.0079), lymphocytes (OR: 0.90, 95% CI: 0.84-0.97, p = 0.0115), and neutrophils (OR: 0.88, 95% CI: 0.81-0.96, p = 0.0108) were causally associated with PE susceptibility. MVMR analysis revealed that lower circulating lymphocyte counts (OR: 0.84, 95% CI: 0.75-0.94, p = 0.0139) were an independent predictor of PE. According to further MR results, this association may be primarily related to HLA-DR⁺ natural killer (NK) cells.

Conclusions: Among European populations, there is a causal association between genetically predicted low circulating lymphocyte counts, particularly low HLA-DR⁺ NK cells, and an increased risk of PE. This finding supports observational studies that link peripheral blood cells to PE and provides recommendations for predicting and preventing this condition.

Keywords: HLA-DR+ NK cells; Lymphocyte subsets; Mendelian randomization; Peripheral blood cell; Pulmonary embolism.

Fig. 1

Flowchart of the overall MR design. In assumption 1, exposure is robustly correlated with instrument variables (IVs); In assumption 2, IVs are not affected by confounders; In assumption 3, IVs only affect outcome through exposure. Abbreviation SNPs, single-nucleotide polymorphisms; GWAS, genome-wide association study; NK cell, nature kill cell; IVs, instrumental variables; IVW, inverse variance weighted; BMI, body mass index; CRP, C-reactive protein

Fig. 2

Forest plot for the causal effect of circulating blood cell counts on the risk of pulmonary embolism derived from IVW method. Abbreviation OR, odds ratio; CI, confidence interval; IVW, inverse variance weighted; SVMR, single-variable mendelian randomization; MVMR, multivariate mendelian randomization

Fig. 3

The scatter plots of bidirectional MR analysis results between circulating white blood cell traits and PE. The scatter plots of genetic associations between white blood cell counts (A), lymphocyte cell counts (B) and neutrophil cell counts (C) associated SNPs against the genetic associations of PE. The scatter plots of genetic associations between PE associated SNPs against the genetic associations of white blood cell count (D), lymphocyte cell count (E) and neutrophil cell count (F). The sky-blue, dark-blue, bottle-green, red and light-green line represents the IVW, MR-Egger, weighted median, weighted mode and simple mode effect, respectively. The slope of the line represents the mendelian randomization effect size. Abbreviation PE, pulmonary embolism; MR, mendelian randomization; IVW, inverse variance weighted

Fig. 4

Forest plot for the causal effect of lymphocyte cell counts on the risk of PE derived from IVW with three models. Abbreviation PE, pulmonary embolism; IVW, inverse variance weighted; OR, odds ratio; CI, confidence interval; SVMR, single-variable mendelian randomization; MVMR, multivariate mendelian randomization

Fig. 5

Forest plot for the causal effect of circulating lymphocyte subgroup on the risk of PE in SVMR and MVMR. Abbreviation PE, pulmonary embolism; OR, odds ratio; CI, confidence interval; SVMR, single-variable mendelian randomization; MVMR, multivariate mendelian randomization