Aerobic training gave the most consistent drop in chronic inflammation markers
In short
Pooling 54 randomised controlled trials in 2,836 adults aged 60 and over, a Bayesian network meta-analysis compared aerobic, resistance, combined, mind-body and multi-component exercise. Aerobic exercise showed the most consistent signal for reducing CRP (SUCRA 0.89, SMD −0.83, 95% CrI −1.42 to −0.24). Mind-body exercise ranked highly for IL-6, but removing a single extreme-effect study collapsed that ranking (SUCRA 0.845 → 0.415; probability of being best 0.291 → 0.035). Dose-response was non-linear with a probable moderate-dose benefit region, though the uncertainty is too wide to treat as a fixed clinical threshold.
Chronic low-grade inflammation becomes common with age and tracks with multisystem decline. That exercise moves the markers was already known; which kind, and how much, was scattered across studies. This analysis pulled 54 trials into one network and ranked them.
What ranked first for CRP?
Aerobic exercise. Its ranking statistic was the highest at SUCRA 0.89, with an effect estimate of SMD −0.83 (95% CrI −1.42 to −0.24) — an interval that does not cross zero, so the direction is clear. Resistance, combined and multi-component training were also associated with improvement, but the most consistent CRP signal belonged to aerobic work.
Why is the IL-6 result hard to trust?
This is the most honest passage in the paper. In the primary analysis, mind-body exercise ranked highly for IL-6. Influence analysis then identified one extreme-effect study; excluding it dropped SUCRA from 0.845 to 0.415 and the probability of being best from 0.291 to 0.035, with the effect estimate shifting toward the null. The authors state directly that IL-6 rankings, especially the mind-body advantage, should be treated as exploratory rather than definitive.
Did it answer "how much"?
Not fully. Modelling weekly metabolic equivalent minutes (MET-min·week) as a continuous exposure with a quadratic function produced a non-linear dose-response pattern suggesting a probable moderate-dose benefit region. But posterior divergence between model specifications was substantial and credible intervals wide, so the authors explicitly warn against reading fixed clinical thresholds into it. Neither "more is always better" nor "exactly this much" is supported here.
What is left for someone training the big three
Nothing here diminishes resistance training — it says only that on this particular outcome, aerobic work led. A strength score measures strength; inflammatory markers appear nowhere in it. Practically, adding aerobic work alongside a big-three program is the choice that matches this result. On picking an intensity, see zone 2 cardio for lifters; on how modalities split across vascular markers, see lifting and artery health.
Every trial included studied adults aged 60 and over. This analysis offers no evidence that the same ranking holds in young trained individuals.
Frequently asked questions
Which exercise lowers inflammation most effectively?
In this network meta-analysis, aerobic exercise showed the most consistent signal for reducing CRP (SUCRA 0.89, SMD −0.83, 95% CrI −1.42 to −0.24). Resistance and combined training were also associated with improvement, but aerobic work led on consistency.
Is the finding that yoga or tai chi lowers IL-6 reliable?
Treat it cautiously. Mind-body exercise ranked highly for IL-6 in the primary analysis, but excluding one extreme-effect study dropped its SUCRA from 0.845 to 0.415 and its probability of being best from 0.291 to 0.035. The authors label it exploratory.
Is there a specific exercise dose for lowering inflammation?
No fixed dose was established. Modelling weekly metabolic equivalent minutes showed a non-linear dose-response with a probable moderate-dose benefit region, but the authors warn that the uncertainty is too large to treat those regions as clinical thresholds.
Do these findings apply to younger people?
There is no evidence here that they do. All 54 trials studied 2,836 participants aged 60 and over; inflammatory markers in young trained individuals were not part of this analysis.
Source: PubMed