Effectiveness of non-pharmacological treatments for sexual difficulties in adults with multiple sclerosis: a systematic review and meta-analysis

Counseling for MS-related sexual problems looks massively effective, but most trials are tiny, unregistered, and from one country

Journal: Sexual Medicine Reviews | Published: 2026-06-30 | Type: Systematic Review, Meta-Analysis | PMID: 42407255 Authors: White E, Adanijo A, Papiernik-Berkhauer P, Norton S, Moss-Morris R, Brown A (Health Psychology Section, Institute of Psychiatry, Psychology and Neuroscience, King's College London) Funding/COI: Funded by King's College London's Institute of Psychiatry, the NIHR Specialist Biomedical Research Centre for Mental Health, Research for Patient Benefit, and South London and Maudsley NHS Foundation Trust. No conflicts of interest disclosed.

Summary

This review pooled 32 trials (1,844 people with multiple sclerosis) testing non-drug treatments, mostly psychosexual counseling, for sexual difficulties. The headline effect size is enormous, but the authors themselves flag that the underlying trials are small, methodologically weak, and drawn overwhelmingly from one demographic in one country, so the size of the effect should not be taken at face value.

Claims

• 32 studies met inclusion criteria (N = 1,844); 23 contributed to the meta-analysis
• Most interventions were psychobehavioral (24 of 32), with 4 exercise/physiotherapy trials, 3 combined-component trials, and 1 other
• Pooled effect was Hedges g = 2.1 (SE = 0.62, P = .001), with I2 = 98.88% heterogeneity
• Removing 2 outlier studies dropped the pooled effect to g = 1.27, with I2 still at 85.15%
• PLISSIT-based structured counseling models for women showed the largest benefits among intervention types
• GRADE-rated certainty of evidence was very low

Study Quality

An I2 above 98% means the included studies are not measuring anything close to the same underlying effect, and a Hedges g of 2.1 is an implausibly large effect size for a behavioral intervention, the kind of number that usually signals small-sample noise, publication bias, or inconsistent outcome measures rather than a true clinical effect. The authors' own sensitivity check, dropping just 2 studies to cut the effect nearly in half while heterogeneity barely moves, confirms the pooled estimate is fragile and driven by a handful of outlier trials.

The review used registered methods (PROSPERO, PRISMA) and applied Cochrane RoB 2 and ROBINS-I appropriately, which is good practice. But the underlying evidence base it's grading is thin: most trials were small, unregistered, and used psychobehavioral designs vulnerable to bias (lack of blinding, self-report outcomes, no control for placebo/attention effects). GRADE rating the certainty as "very low" is the review's own admission that these numbers shouldn't be trusted as a reliable effect estimate yet.

Red Flags

• Extremely high heterogeneity (I2 = 98.88%, dropping only to 85.15% after outlier removal) makes the pooled effect size largely uninterpretable
• Most included studies were small and unregistered
• Sample skewed heavily toward heterosexual, married women from Iran, limiting generalizability to other populations, genders, and relationship structures
• High risk of bias across the included trials
• Very low GRADE certainty of evidence, by the authors' own assessment

Strengths

• Registered protocol (PROSPERO CRD42023399271) and PRISMA-compliant reporting
• Comprehensive search across 8 databases spanning multiple time points through March 2026
• Formal risk-of-bias assessment (RoB 2 and ROBINS-I) and GRADE certainty rating applied transparently
• Sensitivity analysis removing outliers, showing the authors didn't just report the headline number without stress-testing it

Verdict

This is a methodologically careful review of a genuinely weak evidence base, and the authors deserve credit for being upfront about that mismatch rather than burying it. The g = 2.1 pooled effect is not a number to cite as evidence that psychosexual counseling works for MS-related sexual difficulties; it's a number that tells you the field needs larger, registered, more diverse trials before anyone can say anything with confidence. Worth reading for the honest accounting of how little solid evidence exists here, not for the effect size in the abstract.