Review Efficacy of CBT and acceptance- and mindfulness-based treatments in adults with bodily distress – a network-meta-analysis 2025 Maas Genannt Bermpohl+

Abstract

Introduction: Many physical complaints cause long-term bodily distress. Meta-analyses show that cognitive behavioral therapy (CBT) and acceptance- and mindfulness-based treatments (AMBT) reduce somatic symptom severity, but evidence on differential efficacy is limited.

Objective: This study evaluates the efficacy of CBT and AMBT for bodily distress (e.g., somatoform disorders, functional somatic syndromes, and related disorders).

Methods: A network meta-analysis of randomized controlled trials on adults with bodily distress compared CBT and AMBT either directly or with non-specific control groups. Cohen’s d based on between-group effect sizes was aggregated using a random effects model. Primary outcome was somatic symptom severity; secondary outcomes included depression, anxiety, and perceived health status.

Results: Based on 74 studies (N = 8,277), CBT (d = -0.50, 95%CI -0.70 to -0.29; between-group effect sizes vs. wait-list) and AMBT (d = -0.55, 95%CI -1.06 to -0.23; between-group effect sizes vs. wait-list) were equally effective in reducing somatic symptoms at post-treatment. AMBT were more effective than CBT in reducing depression (d = -0.31, -0.58 to -0.04; between-group effect sizes) and anxiety (d = -0.42, -0.73 to -0.11; between-group effect sizes) post-treatment. At long-term follow-up, effects were partly maintained; AMBT remained more effective than CBT for anxiety, with no differential effects for other outcomes.

Conclusions: Both treatments showed benefits compared to various controls. Evidence suggests potential differential treatment effects, indicating some patient groups may benefit more from AMBT. Clinicians should view CBT as foundational but remain open to variations, especially for comorbid pathology.

Open access

 

"Eligibility criteria

For a study to be included, the participants needed to be adults (≥ 18 years old). Moreover, the participants need to fulfil the criteria for one of the conditions considered bodily distress, which are characterized by persistent physical symptoms sharing transdiagnostic mechanisms [4]: e.g., somatic symptom disorder, somatoform disorder, medically unexplained symptoms, chronic fatigue syndrome, fibromyalgia/chronic widespread pain, irritable bowel syndrome or an according research diagnosis (see Supplement S1.1 for a complete list of included syndromes and diagnoses)."

 

See also Efficacy and Acceptance of Cognitive Behavioral Therapy in Adults with Chronic Fatigue Syndrome: A Meta-analysis, 2024, Maas genannt Bermpohl et al.

 

No mention of bias in the abstract…

RoB = Risk of bias

The RoB evaluation revealed that most studies had a high RoB, with only a few rated as having “some concerns”
regarding RoB (k = 2 for somatic symptom severity, k = 3 for depression, k = 3 for anxiety, k = 2 for perceived
health status): These studies solely exhibited potential issues in the domain “bias in the selection of reported
results” (see Supplement S3 for details on individual ratings of studies).

(…)

The high prevalence of RoB across most studies suggests a limited reliability of the findings. Studies classified as
having “some concerns” present a moderate risk, indicating that while their results might be less biased than
those with a “high risk”, caution is still warranted when interpreting their outcomes. This assessment underscores
the need for careful consideration of RoB in evaluating the included studies. These findings align with previous
research in the field [10, 18], where conventional psychotherapeutic trials often receive higher RoB ratings due to
a lack of blinding, particularly when subjective outcome measures are used. In these trials, participants, who
serve as raters, cannot be blinded to the treatments when the control condition differs significantly from the
active intervention. Furthermore, less than half of the studies was preregistered, with some preregistered only retrospectively. Hence, selective reporting of outcomes could not be assessed in most cases