Female Sexual Dysfunction in Type 1 Diabetes: A Systematic Review and Meta-Analysis

About 1 in 3 women with T1DM has sexual dysfunction — twice the rate of healthy controls, and higher than T2DM

Journal: Endocrine | Published: 2026-05-28 | Type: Systematic Review, Meta-Analysis | PMID: 42209931 Authors: Banerjee M, Maisnam I, Mukhopadhyay P, Ghosh S (IPGMER & SSKM Hospital, Kolkata, India; R N Tagore Hospital, Kolkata) Funding/COI: Funding not disclosed. Authors declare no competing interests.

Summary

Female sexual dysfunction (FSD) affects roughly 37% of women with type 1 diabetes — more than double the rate in healthy controls — according to this meta-analysis of 18 studies covering 2,476 women. T1DM carries a higher FSD burden than T2DM, and the authors attribute this to earlier onset and longer disease duration. Continuous subcutaneous insulin infusion (insulin pump therapy) emerged as a protective factor; depression and poor glycemic control were the most consistent risk factors.

Claims

• Pooled FSD prevalence in T1DM: 37% (95% CI: 30–43%; 18 studies, N = 2,476; I² = 89%, random-effects)
• Restricted to studies using the validated FSFI cutoff ≤ 26.5 (n = 12): 34% (95% CI: 26–41%; I² = 89%)
• FSD risk vs. age-matched controls (trim-and-fill adjusted): RR 2.10 (95% CI: 1.30–3.39; p = 0.002; N = 10 studies; I² = 76%)
• T1DM has significantly higher FSD incidence than T2DM (p-interaction < 0.01); diabetes duration explains 70% of this gap (meta-regression R² = 70%)
• Depression and diabetes control indicators were the most consistent individual predictors of FSD
• Continuous subcutaneous insulin infusion (CSII/insulin pump) showed a protective role vs. multiple daily injections

Study Quality

This is a systematic review and meta-analysis drawing on PubMed/MEDLINE, Embase, Scopus, and Web of Science, restricted to cross-sectional and cohort designs that included T1DM participants. The FSFI-restricted subgroup analysis (n = 12) standardizing on the 26.5 cutoff is methodologically sound and the prevalence estimate is consistent with the main analysis. The trim-and-fill adjustment for publication bias is appropriate given the small number of controlled comparisons (10 studies). Meta-regression anchoring the T1DM vs. T2DM gap to disease duration is the most analytically novel piece and holds up given the R² of 70%.

The critical limitation is the I² of 89%, which is severe — nearly all variance across studies is between-study heterogeneity, not sampling noise. This means the "37%" pooled prevalence should be read as a rough central estimate across highly varied populations, measurement tools, and clinical contexts, not a stable true prevalence. The authors use a random-effects model, which is appropriate but does not cure heterogeneity this extreme.

Red Flags

• I² = 89% throughout — this level of heterogeneity makes the pooled estimate unreliable as a precise number; the confidence intervals are wide for a reason
• Only 18 studies with N = 2,476 total; per-study samples appear small
• All source studies are cross-sectional or cohort — causality cannot be established; depression, glycemic control, and FSD likely bidirectionally influence each other
• Funding not disclosed for the meta-analysis itself; individual study funding sources and COI not reported in aggregate
• Publication bias likely exists even after trim-and-fill; smaller studies showing no association may not have been published
• The full-text data provided mixes sources: the Methods, Results, and Discussion sections appear to be from a single pilot case-control study (N = 72, Sant'Andrea Hospital, Italy, 2016–2017 using FSFI-6 with cutoff ≤ 19), not from the meta-analysis itself. This is likely a component study included in the review, suggesting the data extraction pipeline conflated two papers. Readers verifying claims should go to the primary PMID.
• CSII protective effect is based on subgroup analysis within the included studies — not a dedicated comparison; confounding by patient selection (pump users tend to be more engaged in care) is unaddressed

Strengths

• Addresses a genuinely underrepresented complication — FSD in T1DM is largely absent from major diabetes guidelines
• Novel T1DM vs. T2DM direct comparison with meta-regression on duration
• Publication bias formally assessed via trim-and-fill
• Sensitivity analysis restricted to standard FSFI cutoff reduces measurement heterogeneity
• Multi-database search across PubMed, Embase, Scopus, and Web of Science

Verdict

This paper makes a credible case that FSD is a common, underappreciated complication of T1DM — roughly 1 in 3 women affected, twice the rate of healthy controls. The T1DM vs. T2DM comparison and the diabetes duration meta-regression are the most novel and defensible contributions. But the I² of 89% is too high to trust the point estimate, and the CSII finding, while intriguing, is confounded by patient selection. This is a useful paper for establishing that the problem is real and neglected by clinical guidelines. It is not a paper to cite for a precise prevalence figure.