Association Between the Gut Microbiota and Blood Pressure in a Population Cohort of 6953 Individuals

Abstract

Background Several small-scale animal studies have suggested that gut microbiota and blood pressure (BP) are linked. However, results from human studies remain scarce and conflicting. We wanted to elucidate the multivariable-adjusted association between gut metagenome and BP in a large, representative, well-phenotyped population sample. We performed a focused analysis to examine the previously reported inverse associations between sodium intake and Lactobacillus abundance and between Lactobacillus abundance and BP. Methods and Results We studied a population sample of 6953 Finns aged 25 to 74 years (mean age, 49.2±12.9 years; 54.9% women). The participants underwent a health examination, which included BP measurement, stool collection, and 24-hour urine sampling (N=829). Gut microbiota was analyzed using shallow shotgun metagenome sequencing. In age- and sex-adjusted models, the α (within-sample) and β (between-sample) diversities of taxonomic composition were strongly related to BP indexes (P<0.001 for most). In multivariable-adjusted models, β diversity was only associated with diastolic BP (P=0.032). However, we observed significant, mainly positive, associations between BP indexes and 45 microbial genera (P<0.05), of which 27 belong to the phylum Firmicutes. Interestingly, we found mostly negative associations between 19 distinct Lactobacillus species and BP indexes (P<0.05). Of these, greater abundance of the known probiotic Lactobacillus paracasei was associated with lower mean arterial pressure and lower dietary sodium intake (P<0.001 for both). Conclusions Although the associations between overall gut taxonomic composition and BP are weak, individuals with hypertension demonstrate changes in several genera. We demonstrate strong negative associations of certain Lactobacillus species with sodium intake and BP, highlighting the need for experimental studies.

Keywords: Lactobacillus; blood pressure; gastrointestinal microbiota; hypertension; salt intake.

Conflict of interest statement

Dr Salomaa has received honoraria from Novo Nordisk and Sanofi for consultations. He also has ongoing research collaboration with Bayer Ltd (all unrelated to the present study). The remaining authors have no disclosures to report.

Figures

Figure 1. Associations between blood pressure (BP) variables and microbial diversity.

The blue heat maps on the left express the change in BP variables per 1‐SD increase in α diversity (Shannon index); log odds are reported for hypertension. The bar plots on the right represent the proportion of variability (R2) in β diversity (Bray‐Curtis distance) explained by BP indexes. Multivariable‐adjusted model is adjusted for age, sex, body mass index, smoking, exercise, diuretics, β blockers, calcium channel blockers, and renin‐angiotensin system blockers. Significant results are marked with asterisk (P<0.05). α indicates effect size for α diversity; and MAP, mean arterial pressure.

Figure 2. Principal coordinate analysis (PCoA) for species‐level bacterial abundances (Bray‐Curtis distance).

Figure 3. Associations for common microbial genera and Lactobacillus species with blood pressure (BP) indexes.

We observed 45 distinct microbial genera and 19 Lactobacillus species that were significantly associated with BP indexes using DESeq2 (P<0.05 for all). The heat map expresses the fold change associated with BP indexes in base 2 logarithm ratios of microbial abundances. For hypertension, the range signifies a change of microbial abundance from 0.5 (blue) to 2 (red) times the bacterial abundance in normotensive participants. For continuous variables, the fold change is expressed per 1‐SD change in BP variable. The models are adjusted for age, sex, body mass index, smoking, exercise, diuretics, β blockers, calcium channel blockers, and renin‐angiotensin system blockers. Association with bacterial plasmid is denoted using asterisk. MAP indicates mean arterial pressure.

Figure 4. Lactobacillus abundance in groups by 24‐hour urinary sodium excretion.

We observed significant association between 24‐hour urinary sodium excretion and two Lactobacillus species with false discovery rate–corrected P<0.05 while the genus‐level association remained insignificant. For Lactobacillus paracasei (log2 fold change, −0.018±0.002; P<0.001), the 24‐hour urinary sodium excretion levels were as follows: quartile (Q) 1, 205.4±68.5 mmol; Q2, 143.0±43.8 mmol; Q3, 123.9±41.1 mmol; and Q4, 96.6±33.7 mmol. For Lactobacillus salivarius (log2 fold change, 0.007±0.002; P=0.004), the 24‐hour urinary sodium excretion levels were as follows: Q1, 132.7±60.3 mmol; Q2, 142.3±63.3 mmol; Q3, 144.0±58.3 mmol; and Q4, 150.1±68.6 mmol. We visualize the associations using quartiles of DESeq2‐fitted abundances against 24‐hour urinary sodium.