Veillonella atypica supplementation reduces fatigue interference and increases voluntary physical activity: A randomized controlled trial with mechanistic validation in mice

Background

The gut microbiome has been implicated in exercise performance and, more recently, in neural pathways that govern exercise motivation. Veillonella atypica —a lactate-utilizing bacterium enriched in elite athletes— enhances forced treadmill performance in mice through lactate-to-propionate metabolism, yet effects on voluntary physical activity, the clinically relevant behavior, remain unknown. We tested whether supplementation with V. atypica reduces fatigue burden and increases voluntary physical activity in humans, with mechanistic validation in mice.

Methods

We conducted a randomized, placebo-controlled, app-mediated trial in healthy adults (8-week protocol: 2-week baseline, 4-week supplementation, 2-week washout; n=151 meeting compliance criteria). Participants received high-dose V. atypica , low-dose V. atypica , or placebo; dose groups were pooled after confirming similar responses. Primary endpoints were Multidimensional Fatigue Inventory (MFI-20) subscales assessed pre- and post-supplementation; secondary endpoints included weekly surveys (fatigue interference, physical activity hours) and daily surveys (sleep quality, energy). In parallel, we profiled voluntary wheel running in mice randomized to PBS, V. atypica , or Lactobacillus johnsonii (peanut butter delivery), with striatal dopamine quantification at endpoint.

Results

In humans (n=146-149 depending on outcome), weekly longitudinal assessments revealed V. atypica -specific benefits not captured by endpoint comparisons. Relative to placebo, V. atypica produced a faster decline in fatigue interference (−0.097 days/week; 95% CI −0.175 to −0.018; p =0.016) and a faster increase in self-reported physical-activity hours (+0.357 h/week; 95% CI +0.031 to +0.684; p =0.032). Model-predicted effects by week 6 corresponded to approximately 0.6 fewer days with fatigue interference and 2 additional activity hours per week if linear trends continued. By contrast, MFI-based fatigue scores improved over time in both arms without between-group differences (treatment×time interactions all p >0.40). Daily measures indicated improved sleep quality during washout in V. atypica versus placebo (difference-in-differences +0.197 Likert units; p =0.0039), with no differences in sleep duration; odds of high-energy days trended higher with V. atypica during washout (OR≈2.36; p =0.07).

In mice, mixed-effects models revealed a significant group×time interaction for daily running distance (F=13.31; p =2.1×10⁻⁴): V. atypica maintained running across 12 weeks (slope not significantly different from zero), whereas PBS and L. johnsonii declined significantly. During the final 5 weeks (days 49–84), baseline-adjusted mean running distance was approximately 2.86 km/day higher in V. atypica versus PBS ( p =2.2×10⁻⁴). Striatal dopamine concentrations were elevated in V. atypica -treated mice relative to both controls (one-way ANOVA p =0.05; V. atypica vs PBS p =0.03, approximately 30% higher).

Conclusions

Across complementary human and murine experiments, V. atypica was associated with reduced day-to-day fatigue interference and increased voluntary physical activity, with convergent evidence of sustained behavioral engagement in mice. Preliminary findings of elevated striatal dopamine are consistent with emerging evidence that gut microbiome regulates exercise motivation through dopaminergic signaling, though mechanistic causality remains to be established. These data extend Veillonella biology beyond performance physiology to motivational domains and support continued investigation of precision microbiome approaches to address barriers to physical activity.