Longitudinal Trajectories and Psychosocial Predictors of Postpartum Sexual Dysfunction from Early Pregnancy to 12 Months Postpartum

Sexual function scores hit a nadir 6–8 weeks postpartum and depression predicted decline better than delivery mode or breastfeeding

Journal: Medicina (Kaunas) | Published: 2026-03-14 | Type: Prospective cohort | PMID: 41901622 Authors: Boarta Aris, Gluhovschi Adrian, Craina Marius Lucian et al. (Victor Babes University of Medicine and Pharmacy, Timișoara, Romania) Funding/COI: Not listed for either

Summary

A Romanian single-center cohort tracked Female Sexual Function Index (FSFI) scores across six time points from early pregnancy through 12 months postpartum in 187 women. Mean FSFI bottomed out at 20.1 ± 4.2 at 6–8 weeks postpartum — well below the dysfunction threshold of 26.55 — before partially recovering to 25.5 ± 4.0 by 6–12 months. The headline finding: depressive symptoms (PHQ-9) and body-image disturbance were stronger predictors of persistent dysfunction than mode of delivery, breastfeeding, or pelvic-floor symptoms.

Claims

• FSFI declined from 27.4 ± 3.9 (first trimester) to a nadir of 20.1 ± 4.2 at 6–8 weeks postpartum (p < 0.001)
• Partial recovery to 25.5 ± 4.0 by 6–12 months, still below the dysfunction cutoff for many women
• Women with dysfunction had higher PHQ-9 scores than those without (8.8 ± 3.3 vs. 6.7 ± 3.5; p < 0.001)
• Each 1 SD increase in PHQ-9 predicted a 1.2-point decrease in FSFI (p = 0.005)
• PHQ-9 and FSFI correlated inversely at r = −0.39 (p < 0.001)
• Body-image disturbance had a partially PHQ-9-mediated effect on FSFI
• Three trajectory clusters emerged, with postpartum dysfunction rates spanning 28.6% to 89.7%
• A combined psychosocial model achieved AUC 0.9, sensitivity 0.8, specificity 0.8 in predicting dysfunction

Study Quality

Prospective longitudinal design with six measurement points is the right tool for this question, and the use of validated instruments (FSFI, PHQ-9, WHOQOL-BREF) is appropriate. The trajectory cluster analysis adds real value over a simple pre/post comparison.

The central limitation is variance explained: the multivariable model accounts for only 19% of FSFI variance (R² = 0.19), which means 81% of what drives postpartum sexual function remains unmodeled. An AUC of 0.9 for the psychosocial prediction model sounds impressive but is difficult to evaluate without knowing whether it was internally or externally validated — in a single-center sample of 187, overfitting is a genuine concern. The PHQ-9 and other psychosocial scales were collected postpartum, not at baseline during pregnancy, which limits any causal inference about depression causing dysfunction versus both arising from the same postpartum stressors.

Red Flags

• Single center (Timișoara, Romania) with 187 eligible participants — generalizability is uncertain
• Full analysis conducted on the 90 women with documented dysfunction, not the entire cohort — selection effect baked in
• Funding and COI both unlisted; impossible to assess industry influence
• Psychosocial predictors measured postpartum only; no pregnancy-baseline psychological data to establish temporal precedence
• R² = 0.19 despite the headline framing of depression as the key driver
• No correction described for multiple comparisons across six time points and multiple predictors
• AUC 0.9 with no mention of external validation or cross-validation procedure in abstract

Strengths

• Six-point longitudinal design maps the full arc from early pregnancy through 12 months postpartum — rare in this literature
• Three validated instruments administered consistently across time points
• Trajectory cluster analysis identifies heterogeneity rather than treating women as a homogeneous group
• Competes depression directly against obstetric variables (delivery mode, breastfeeding, pelvic floor), which is the useful clinical question
• Explicitly reports effect sizes and confidence intervals, not just p-values

Verdict

This paper earns attention for its design and its direct comparison of psychosocial vs. obstetric predictors — the finding that depression outweighs delivery mode is clinically plausible and the longitudinal architecture supports it more than a cross-sectional study would. But the single-center sample of 90 dysfunctional women, the missing funding/COI disclosures, the unvalidated prediction model, and the modest R² all demand replication before the psychosocial framing gets baked into clinical narratives. It's a solid hypothesis-generator from a regional cohort, not a practice-changing study.