Breathwork: Effects & Evidence — what’s supported and what isn’t

Breathwork: Effects & Evidence — what’s supported and what isn’t

Breathwork (breathing exercises) includes many very different techniques: from calm, “diaphragmatic” breathing to variations with breath-holds (retention) or intentionally higher ventilation. This variety makes it scientifically hard to assess “breathwork” as a single entity. In studies, however, recurring effects show up—especially on stress, mental health, and physiological arousal.

Why breathing exercises might work (and what’s proven)

Short answer: Breathwork could influence stress and arousal via breathing patterns and the regulation of CO₂/O₂ dynamics, and through autonomic regulation. But the mechanisms are not clearly established in the same way for every technique—the evidence points more toward specific protocols rather than “any breathing exercise.”

Breathing exercises plausibly act at several points:

1. Breathing patterns and mechanics: Frequency, rhythm, depth (e.g., “diaphragmatic”), and the ratio of inhalation to exhalation can change how the autonomic nervous system is regulated.
2. CO₂/O₂ dynamics: Certain protocols change CO₂ availability in the respiratory control center. This may be experienced subjectively as “calming” or “wakefulness,” and it can coincide with physiological changes in arousal.
3. Attention & interoception: Controlling breath is also an attention task and a form of body awareness. This can modulate stress-relevant experiences independently of any purely “chemical” effects.

Important: the strength of the scientific claim depends heavily on which technique, how long, with what intensity, and under which study design it was tested. That’s exactly why it’s risky to treat “breathwork” as a blanket term: a protocol that uses retention is not the same as short practice without breath-holds, and “diaphragmatic breathing” is not automatically equivalent to “calming.”

Regarding the evidence base, the most robust signal area right now is: stress/mental health. A meta-analysis of randomized controlled trials summarizes the effects of breathwork on stress and mental health across several RCTs (Fincham et al., 2023, PMID 36624160). That’s a strong argument that a real benefit may exist—while still being an aggregation across techniques and outcomes.

For mood and physiological arousal, there is additional RCT evidence on short, structured breathing practices (Balban et al., 2023, PMID 36630953). Although it is labeled in the content as “breathwork,” transferring the results to arbitrary breathing techniques is not automatically justified: in RCTs the protocol is usually clearly defined, while in real-world use duration, intensity, and coaching often vary.

Lifestyle levers first: sleep, movement, light — before you “stack” breathing techniques

Short answer: If your goal is stress reduction, sleep, regular exercise, and daylight exposure are usually better data-supported levers than breathwork. Breathing exercises can complement these, but the evidence does not establish a general replacement for baseline measures—and sleep loss or fatigue can distort effects.

In practice, much of the “breathing performance” often works indirectly: if your system is already in a stress mode (poor sleep, little daylight, low activity), your physiological and psychological response patterns may differ from those in a “stable” baseline routine. Methodologically, it makes sense not to treat breathwork as a substitute for baseline levers, but as an add-on.

Sleep is especially important because many people use breathing exercises during periods when their stress level is high—and at exactly those times, sleep status can overshadow the effects. If you, for example, sleep restlessly or go to bed late, arousal can be higher during the day and mood may shift accordingly. This fits the broader evidence logic from the sleep regulation field: the better sleep architecture and sleep quality are stabilized, the “cleaner” it becomes to evaluate the effects of individual interventions. If you want to go deeper: Sleep cycles: Effects & Evidence — what’s supported and what isn’t can provide relevant context.

Movement and light also influence stress pathways, circadian rhythm, and mood via established routes. Breathing exercises are most likely useful when you:

• have a consistent baseline routine (bed/wake times as similar as possible),
• get light during the day,
• incorporate movement into your daily life,
• and use breathwork on days when you don’t feel completely “out of sync” with your routine.

Another practical point: many RCTs test short, structured breathing sessions or clearly defined meetings. This supports “start small” rather than stacking long sessions. When the practice varies widely (duration/intensity/retention), it becomes harder to separate real effects from expectation or day-to-day fluctuations. The available evidence more strongly supports structured, short-term approaches for mood/arousal than a universal “more is better.”

So if you use breathwork as a supplement, prioritize the levers that reduce overall variability in your system first. Then breathing exercises are more likely to be an additional effect—rather than an attempt to compensate for missing basics.

What’s well supported: stress, mood, and mental health

Short answer: The best evidence supports breathwork mainly for stress, mental health, mood, and physiological arousal. Meta-analyses and RCTs show effects in controlled designs; at the same time, that does not mean every breathing technique and every intensity works to the same extent for every person.

The most convincing starting point is the meta-analysis on stress and mental health. In a summary of randomized controlled studies, effects of breathwork on stress and mental health were reported across multiple studies (Fincham et al., 2023, PMID 36624160). The advantage of this design: it reduces individual bias because there are comparison groups and the effects are aggregated across studies. That’s precisely why meta-analysis is a good tool to test whether anything exists at all—although results depend on the included protocols.

For specific mood effects, there is RCT evidence on short, structured breathing practices. In the RCT by Balban et al. (Balban et al., 2023, PMID 36630953), improvements in mood and reduced physiological arousal after a short, structured practice are described. The key study logic here is: “short and structured” is likely not random—participants are probably coached, the intervention is clearly defined, and the dose is comparable.

Additionally, there is an RCT with remote “high ventilation breathwork” including retention versus placebo, reporting effects on mental health and well-being (Fincham et al., 2024, PMID 39043650). This is relevant because it doesn’t just capture “slow breathing,” but also a version with stronger ventilation and breath-holds. However, you should not conclude that retention is good or safe in all situations: uniform long-term and safety data are missing in many areas.

What the evidence base also does not provide: proof that every breathwork technique produces the same magnitude of effect for every person. The label “breathwork” covers very different breathing patterns, and even within the evidence, effects can depend strongly on the protocol. Meta-analyses are useful to identify overall signals, but they don’t replace technique-specific interpretation.

Methodologically sound takeaways are therefore:

• If your goal is stress/mental health, orient yourself toward protocols tested in RCTs (not just YouTube/app variants without clear description).
• If you want to try something specific, keep the frame stable (same duration, same instruction, same time of day) and assess outcomes like arousal/tension or mood consistently.

Breathing exercises for conditions: back/asthma topics and gaps

Short answer: For medical conditions, the evidence is mixed. Systematic reviews exist for spinal pain and for breathing techniques in asthma, but results vary depending on the endpoint and the technique. For dysfunctional breathing (adults), the review literature places breath-related approaches in context, and for Buteyko there is methodological critique of a meta-analysis, which complicates interpretation.

When it comes to illnesses, the question “does it work?” becomes automatically more complex: which technique? Which target outcome (pain intensity, lung function, quality of life, symptom scores)? What comparator treatment? And how reliable are the measurement instruments?

For spinal pain, there is a systematic review including a meta-analysis on breathing interventions (Van et al., 2025, PMID 41099211). This suggests breathing interventions aren’t only viewed as relaxation, but may also be a supportive strategy. Still, for spinal pain, study heterogeneity is often high (participant profiles, co-therapies, extent of the breathing intervention), so effects are generally harder to generalize than in clear stress-outcome pathways.

For asthma, a systematic review reports the role of breathing techniques in asthma treatment (Zafar et al., 2024, PMID 39028058). Again, “breathing techniques” are not a single package. Breathing exercises differ in approach (e.g., fast vs. slow, with or without retention, physiologically justified vs. training-based), and asthma endpoints are likewise multifaceted (e.g., symptoms vs. lung function vs. quality of life). Accordingly, evidence varies by technique and endpoint.

For dysfunctional breathing (adults), a systematic review evaluates nonpharmacological interventions and places breathing-related approaches within the broader context (Bondarenko et al., 2025, PMID 40345332). This is particularly important because “dysfunctional breathing” is clinically relevant as a concept, but it doesn’t automatically mean that every breathing exercise produces identical effects. Reviews can show which intervention types have been studied better and where gaps remain.

An additional source of uncertainty involves the Buteyko technique. Zhao et al. (Zhao et al., 2026, PMID 42105647) addresses methodological reflections on the evidence of a meta-analysis for Buteyko technique in asthma management. These kinds of methodological discussions are a warning sign: even if an earlier analysis suggested positive effects, the study design, inclusion criteria, or statistical assumptions can bias conclusions. For you, this means: when techniques are argued strongly via individual meta-analyses, you should be especially critical about study quality and consistency.

In short: research exists for breathing exercises in conditions, but the data are not as consistent as in the stress/mental-health domain. If you consider breathing exercises for asthma or pain, it makes sense to view them as a complementary measure—not a replacement for medical care—and to adapt the technique to your endpoints.

Evidence hierarchy: RCTs, systematic reviews, methodological quality, and limits

Short answer: RCTs and meta-analyses offer the highest level of inference because they include comparison groups and usually provide better control of confounders. Still, conclusions remain limited if protocols vary substantially or if long-term data are missing on safety/effectiveness—especially for intensive variants involving retention.

The evidence hierarchy is particularly important in breathing research because the interventions are heterogeneous. A randomized design helps ensure differences between groups are more likely due to the intervention than to general lifestyle or expectation effects. Meta-analyses (e.g., on stress and mental health) pool these RCTs to obtain a more stable signal (Fincham et al., 2023, PMID 36624160).

That said, typical limitations are visible:

• Technique and protocol differences: A breathing exercise is not just “breathing.” Even under the same name, duration, breathing depth, frequency, inhalation/exhalation ratio, and whether retention is included can differ. This reduces transferability.
• Outcome measurement: Stress and mood can be measured using different scales (e.g., self-report vs. physiological markers). Even with the same scale, timing and expectation management are crucial.
• Quality of primary studies: Systematic reviews aggregate, but they don’t automatically improve poor primary study quality. The review format increases robustness only if the included RCTs are solid (Zafar et al., 2024, PMID 39028058; Van et al., 2025, PMID 41099211; Bondarenko et al., 2025, PMID 40345332).

Another central point is safety and long-term effect. In many breathing protocols, intensity and retention can increase physiological burden. For general “use as you like it,” data are not equally reliable in every detail. Practical consequence: with intensive variants, caution is sensible—especially if you have pre-existing conditions or unstable states (even if the studies cited here do not automatically cover every contraindication in detail).

What you should infer for interpretation:

• If your goal is stress/mental health, the evidence is relatively more consistent (Fincham et al., 2023, PMID 36624160; Balban et al., 2023, PMID 36630953).
• If your goal is clinical symptom control in asthma or back/spinal pain, the evidence base is better read “by endpoint and by technique,” not as a broad overall promise (Zafar et al., 2024, PMID 39028058; Van et al., 2025, PMID 41099211; Bondarenko et al., 2025, PMID 40345332).
• If a meta-analysis is methodologically criticized (e.g., Buteyko), you should factor in uncertainty (Zhao et al., 2026, PMID 42105647).

Study shortcut: which formats were investigated (and how to interpret the information)

Short answer: The RCT and review evidence refers to specific breathing protocols: short structured exercises, remote versions with high ventilation and retention, and breathing techniques in the contexts of asthma, spinal pain, and dysfunctional breathing. That means you always need technique-specific interpretation for “evidence.”

The following table shows which intervention types are emphasized in the provided study list, what the typical comparison situation is, and what the primary outcome goal is in each case. This helps you assess whether “breathwork” is likely to fit your needs—or whether you’re extrapolating the wrong technique type.

This is how you interpret it practically:

1. If your goal is stress/mental health: Start with formats that were pooled in the meta-analysis for stress/mental health (Fincham et al., 2023, PMID 36624160) and with RCTs testing short structured sessions (Balban et al., 2023, PMID 36630953).
2. If you consider a very intensive technique (e.g., high ventilation + retention): The RCT shows effects in the study design (Fincham et al., 2024, PMID 39043650), but that doesn’t mean you should transfer the same intensity on your own—especially without clear coaching and context.
3. If you use breathing exercises for conditions: Read the reviews by endpoint. Asthma endpoints are not the same as back-pain endpoints (Zafar et al., 2024, PMID 39028058; Van et al., 2025, PMID 41099211). For dysfunctional breathing, the review framing helps you place intervention types into the broader context (Bondarenko et al., 2025, PMID 40345332).
4. If your technique is strongly based on Buteyko “asthma”: Take the methodological critique seriously—the evidence interpretation can be uncertain (Zhao et al., 2026, PMID 42105647).

What you take from this (Bottom Line)

• Breathwork is best supported for stress, mood, and mental health—especially in RCTs and meta-analyses (Fincham et al., 2023, PMID 36624160; Balban et al., 2023, PMID 36630953).
• Think technique-specific: “Breathing exercise” is not a uniform tool. Protocols differ (including retention/high ventilation), so transfer is not automatically correct.
• Lifestyle first: sleep, movement, and daylight often provide stronger, more stable baseline effects; breathwork is most likely a sensible add-on.
• For conditions, the evidence is mixed and endpoint-dependent (asthma, spinal pain, dysfunctional breathing) — reviews help with interpretation, but they do not replace medical treatment (Zafar et al., 2024, PMID 39028058; Van et al., 2025, PMID 41099211; Bondarenko et al., 2025, PMID 40345332).
• Long-term safety/optimal dose is not equally well supported for every variant. Especially with more intensive approaches involving retention, be particularly cautious and stick to coached, studied formats.