Psychiatry · Medical Literature
Breath-hold time and anxiety-related vulnerability: A systematic review and meta-analysis.
Abstract
Altered interoceptive processing and reduced tolerance to aversive bodily sensations are core features of anxiety disorders. Maximal voluntary breath-holding time (mvBHT) has been proposed as a simple, objective behavioral index of interoceptive tolerance. This PRISMA-guided systematic review and meta-analysis evaluated mvBHT as a performance-based marker of interoceptive tolerance in anxiety by examining whether this measure differentiates adults with anxiety disorders (AD) from healthy controls (HC), quantifying its associations with trait and state anxiety, and examining study-level moderators (sex, age, and breath-hold onset). 22 studies (N = 1263 adults) were included. Random-effects models showed that AD exhibited shorter mvBHT than HC, an effect driven by a significant medium-sized reduction in panic disorder (PD) with low heterogeneity, whereas non-PD anxiety disorders yielded a smaller, nonsignificant effect. Multilevel meta-analytic correlations indicated marginal, non-significant associations between mvBHT and trait anxiety once within-study dependence was accounted for, whereas associations with state anxiety and momentary distress assessed during or immediately after respiratory challenge were significant and consistently stronger. Meta-regressions in HC samples indicated that a higher proportion of females predicted shorter mean mvBHT, and that protocols initiating the breath-hold after inhalation yielded longer mvBHT than those initiating after exhalation. In PD samples, sex composition did not predict mvBHT, whereas breath-hold onset remained the primary study-level predictor. Overall, mvBHT reliably distinguishes clinical from non-clinical groups and tracks state anxiety and distress; trait correlations are marginal and non-significant once within-study dependence is properly modelled, supporting its utility as a performance-based behavioral marker of respiratory interoceptive tolerance.
Clinical relevance
Does maximal voluntary breath-holding time (mvBHT) reliably differentiate adults with anxiety disorders from healthy controls, and how strongly does it correlate with trait anxiety, state anxiety, and momentary distress? __SNAPSHOT__ {"studyDesign":"PRISMA-guided systematic review and meta-analysis (random-effects models; multilevel meta-analytic correlations; meta-regression)","setting":"Aggregate of studies identified through systematic search; individual study settings not reported in abstract","sampleSize":"22 studies; N = 1,263 adults","population":"Adults with DSM-defined anxiety disorders (including panic disorder and non-PD anxiety disorders) compared with healthy controls; studies also included non-clinical samples with continuous trait/state anxiety measures","intervention":"Maximal voluntary breath-holding time (mvBHT) administered as a performance-based behavioral task; breath-hold onset varied across studies (post-inhalation vs. post-exhalation)","comparator":"Healthy controls (for group-difference analyses); continuous anxiety measures (for correlational analyses)","primaryOutcome":"Group difference in mvBHT between anxiety disorder and healthy control samples; meta-analytic correlations of mvBHT with trait anxiety, state anxiety, and momentary distress","followUp":"Not reported (cross-sectional measurements within included studies)","fundingSource":"No external funding (PROSPERO-registered protocol: CRD420241047779)"}
Key findings
Across 22 studies (N = 1,263 adults), random-effects models demonstrated that adults with anxiety disorders exhibited shorter mvBHT than healthy controls; however, this overall effect was primarily driven by a significant, medium-sized reduction in panic disorder (PD) specifically, with low heterogeneity across PD studies. Non-PD anxiety disorders yielded a smaller, statistically nonsignificant effect. Exact group-difference effect sizes (Cohen's d or Hedges' g), confidence intervals, and p-values for the overall comparison and the PD-specific subgroup are not reported in the available abstract. Multilevel meta-analytic correlations — which corrected for within-study dependence — showed marginal, non-significant associations between mvBHT and trait anxiety. In contrast, associations between mvBHT and state anxiety and momentary distress assessed during or immediately after the breath-hold challenge were significant and consistently stronger (specific r values, CIs, and p-values not reported in the abstract). Meta-regression in healthy control samples revealed that a higher proportion of female participants predicted shorter mean mvBHT, and that protocols initiating breath-hold after inhalation yielded longer mvBHT than those initiating after exhalation. In PD samples, sex composition did not predict mvBHT, whereas breath-hold onset remained the primary study-level predictor. NNT/NNH are not applicable to this meta-analytic design. No clinically important adverse events were reported (not a treatment trial).
Practice implications
This meta-analysis establishes that mvBHT is a reliable behavioral discriminator between panic disorder and healthy controls, with a medium effect size and low heterogeneity — a clinically noteworthy signal in a low-cost, easily administered task. However, the discriminative utility appears specific to PD and does not generalize to non-PD anxiety disorders based on current evidence. The dissociation between mvBHT and trait anxiety (marginal, nonsignificant once within-study dependence is modelled) versus state anxiety and momentary distress (significant, stronger associations) is theoretically coherent with contemporary interoception frameworks, suggesting mvBHT captures in-the-moment interoceptive reactivity rather than stable dispositional anxiety. The authors explicitly caution that mvBHT should not be used in isolation but rather integrated into multimodal assessment batteries combining self-report, physiological, and performance-based measures. Patients with non-PD anxiety diagnoses (e.g., generalized anxiety disorder, social anxiety disorder, specific phobia) were not adequately differentiated from controls in this synthesis, limiting cross-diagnostic generalizability. Procedural heterogeneity in breath-hold onset (inhalation vs. exhalation) represents a major source of measurement variance that must be standardized before clinical adoption. IMPORTANT LIMITATION: Only the abstract was available for this summary; full effect sizes, confidence intervals, publication bias assessments, and risk-of-bias ratings from the complete paper were not reviewable. Findings should not be used to make individual clinical diagnostic decisions in the absence of validated clinical protocols. __LIMITATIONS__ Only 22 studies with a total of 1,263 participants were included — a modest aggregate for a meta-analysis, with unknown statistical power for moderator analyses.,Substantial procedural heterogeneity across included studies (breath-hold onset after inhalation vs. exhalation, variable instruction sets) complicates direct comparison of mvBHT values and limits pooled interpretability.,The PD-specific effect was significant with low heterogeneity, but non-PD anxiety disorders yielded a non-significant effect; it is unclear whether this reflects a true null effect or insufficient power due to smaller non-PD subgroup sizes.,Publication bias was not assessable from the abstract alone; small-study effects or selective reporting in the included literature are possible.,All included studies are cross-sectional, precluding causal inference about whether reduced mvBHT is a marker, mediator, or consequence of anxiety pathology.,Sex composition of samples moderated mvBHT in healthy controls, introducing potential confounding across studies with differing sex ratios.,Trait anxiety showed only marginal, non-significant associations with mvBHT once within-study dependence was properly modelled — raising questions about whether mvBHT adds incremental diagnostic value beyond simpler self-report instruments.,Generalizability is limited to adults; findings cannot be extrapolated to pediatric or geriatric populations.,AI-assisted tools (Rayyan, ChatGPT) were used in study selection and manuscript preparation, which, while disclosed and reviewed by authors, introduces a methodological transparency consideration.,Study received no external funding, which minimizes industry-sponsorship bias, but also means resource limitations may have affected search comprehensiveness. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ mvBHT reliably distinguishes adults with panic disorder from healthy controls with a medium effect size and low heterogeneity, supporting its potential utility as an objective, low-cost behavioral index of interoceptive tolerance specifically in PD — but not across non-PD anxiety disorders. Its associations with state anxiety and momentary distress are significant and stronger than those with trait anxiety, suggesting its primary value is as a state-sensitive performance measure rather than a diagnostic screener. Standardization of breath-hold protocols and replication in larger, diagnostically diverse samples are needed before clinical implementation can be recommended.