Meta-analysisindependent cross-checkfull text checked★ key study I-ORG-10 in-depth review
Meta-analysis: removing ovaries at hysterectomy
Long-term outcomes of hysterectomy with bilateral salpingo-oophorectomy: a systematic review and meta-analysis.
Overview
Hassan H et al.. Am J Obstet Gynecol 2024 · PMID 37364803 · DOI 10.1016/j.ajog.2023.06.043
Linked to less breast cancer but more heart disease, dementia and other conditions in younger women; effect on death was inconsistent.
Does the study support the specific claim? Partly · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
BSO outcomes
- How was the study done?
Meta-analysis
- How many people, animals or samples were studied?
n/a
- What was actually found?
Trade-offs
- How large was the observed effect?
CVD HR 1.18; dementia HR 1.70
- What do the results mean in everyday life?
Removing healthy ovaries in young women has costs and some benefits
- What are the study's limitations?
Observational studies, few per outcome, fixed-effect pooling, no study-level risk-of-bias assessment
- What does it say about Elawady's specific claim?
Supports risks; mortality unclear
- Do the results apply to Germany and Europe?
Applicable
- What can explicitly not be concluded from this study?
Causation; harm after age 50; absolute risk
Scientific in-depth review
1. Full reference
- Authors
- Hassan H et al.
- Journal
- Am J Obstet Gynecol
- Year
- 2024
- Identifier
- PMID 37364803 DOI 10.1016/j.ajog.2023.06.043
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Cancer Research UK Catalyst Award CanGene-CanVar (C61296/A27223); one author NIHR Cambridge BRC. Authors report no conflict of interest.
2. Research question
Long-term outcomes of bilateral salpingo-oophorectomy (BSO) at hysterectomy vs hysterectomy with ovarian conservation or no surgery
3. Study design and population
- Study design
- Meta-analysis – Systematic review and fixed-effect meta-analysis of observational studies (PROSPERO CRD42022349899), updating Evans et al. 2016; GRADE per outcome
- Population
Women with hysterectomy plus BSO for benign indications (excluding cancer treatment, gender affirmation, high ovarian-cancer risk); studies from USA, Europe, Australia, Canada, Asia
- Country
UK authors (University of Cambridge); included studies international
- Period
Search January 2015 - August 2022 (plus studies from Evans et al. review and hand search); English only
- Sample size
38 studies (12 prospective cohorts, 24 retrospective cohorts, 1 case-control, 1 case-cohort); sample sizes 2900 to >5 million; most meta-analyses based on 2-6 studies
- Inclusion and exclusion criteria
Observational studies comparing hysterectomy with BSO vs hysterectomy alone or no surgery; follow-up >=3 years; excluded cancer-treatment or gender-affirmation surgery, high-risk women, short-term outcomes
4. Methods and endpoints
- Exposure / intervention
Hysterectomy with bilateral salpingo-oophorectomy (stratified by age at surgery <45/<50 vs >50)
- Comparator
Hysterectomy with ovarian conservation, or no surgery (subgroups by reference category)
- Endpoints
Cancer (ovarian, breast, colorectal, thyroid, kidney, all), CVD, CHD, stroke, CVD risk factors, dementia, depression, anxiety, fractures, all-cause and cause-specific mortality
- Statistical methods
Inverse-variance fixed-effect meta-analysis of adjusted HRs; I2 and chi-square; outlier exclusion when I2 >50%; GRADE
5. Actual results
- Results
Ovarian cancer HR 0.11 (0.09-0.15). Breast cancer BSO <45: 0.78 (0.73-0.84; GRADE high); >50: 1.05 (0.96-1.15). Colorectal cancer <45: 1.27 (1.10-1.47; I2 72%, driven by Mytton; excluding it 1.11, 0.94-1.31); >50: 1.25 (1.04-1.51). All cancers <50: 0.88 (0.83-0.94). Total CVD <50: 1.18 (1.11-1.25; GRADE very low); >50: 1.04 (0.96-1.12). CVD mortality <50: 1.24 (1.08-1.43). CHD <50: 1.17 (1.10-1.25; moderate); >50: 1.21 (1.05-1.39). Stroke <50: 1.20 (1.10-1.31). Vs no surgery: hypertension 1.13 (1.06-1.20), hyperlipidaemia <45: 1.44 (1.25-1.65), diabetes 1.16 (1.09-1.24), dementia <45: 1.70 (1.07-2.69; 2 Mayo cohorts; unstratified 1.36, 0.99-1.85; GRADE very low), depression 1.39 (1.22-1.60). Hip fracture no association. All-cause mortality BSO <50: 1.22 (1.16-1.28; I2 85%; after excluding outliers 1.19, 1.12-1.25, I2 59%; GRADE moderate); >50: 0.98 (0.91-1.05; GRADE moderate).
- Absolute risks
not reported in full text (relative hazard ratios only)
- Authors' conclusion
Hysterectomy with BSO at any age reduces ovarian cancer and at a young age breast cancer risk, but this should be balanced against possible increases in CVD, colorectal cancer and all-cause mortality in young women; more evidence on HRT and better-adjusted studies needed.
6. Strengths
- Registered protocol; update with meta-analysis of a prior review
- Adjusted HRs; age-at-surgery stratification (reduces heterogeneity)
- Subgroups by reference category (hysterectomy vs no surgery)
- GRADE per outcome
- Explicit discussion of confounding, detection and misclassification bias
7. Weaknesses
- Fixed-effect pooling despite substantial heterogeneity for several outcomes (I2 up to 92%)
- No formal per-study risk-of-bias tool (only GRADE)
- Few studies per outcome (often 2-3); publication bias not assessable (authors)
- English only; new search limited to 2015-2022
- Confounding by indication and comorbidity, especially vs no surgery; strongest associations (mortality, colorectal cancer, dementia) from studies not adjusted for comorbidity (authors)
- Neuropsychiatric estimates from two Mayo Clinic cohorts only, not adjusted for parity
- Text/table inconsistencies (dementia vs depression HRs swapped in text; CHD age strata swapped)
- Overlap with Parker (A-ORG-13) and Mayo cohorts (A-ORG-14/15); not independent confirmation
8. Risk of bias
- Orienting tool
- AMSTAR 2
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- high
- Rationale
Changed from some_concerns: critical weaknesses in items 9 and 15 and fixed-effect pooling under heterogeneity mean overall confidence would be rated 'low' to 'critically low' by AMSTAR 2; underlying evidence is observational with GRADE mostly low/very low except breast cancer (high) and CHD, all-cancer and mortality (moderate). Overlaps with A-ORG-13/14/15. Structured single-reviewer judgement from full text; a formal assessment would require two independent reviewers and the protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 2 protocol registered (critical) | low | PROSPERO CRD42022349899. |
| Item 4 comprehensive search (critical) | some concerns | PubMed, Web of Science, Embase plus prior review and hand search; English only; new search 2015-2022. |
| Item 5/6 duplicate selection and extraction | some concerns | Two screeners with arbiter; duplicate extraction not described. |
| Item 7 excluded studies justified (critical) | some concerns | PRISMA flow only in main text; no list of excluded full texts visible. |
| Item 9 risk of bias of included studies (critical) | high | No study-level risk-of-bias tool; RoB only within GRADE. |
| Item 11 appropriate meta-analytic methods (critical) | some concerns | Fixed-effect models used even with high heterogeneity; outlier exclusion post hoc; stratification by age and reference group appropriate. |
| Item 13 RoB considered in interpretation (critical) | low | Confounding, detection and misclassification bias discussed; GRADE reported. |
| Item 15 publication bias (critical) | high | Not assessed (too few studies). |
| Item 16 COI | low | Funding and COI reported. |
Full-text review (25 October 2026): Added study composition (38 studies), eligibility, full table of pooled HRs by age and reference group incl. ovarian cancer 0.11, all-cause mortality <50 1.22 vs >50 0.98, GRADE ratings, text/table inconsistencies, funding/COI; AMSTAR 2 items added; bias changed some_concerns -> high.
9. Assessment of the statistical interpretation
Relative associations are modest (HR about 1.1-1.4) and age-dependent; harms concentrate in BSO before 45-50 years, while after 50 there is no association with all-cause mortality (HR 0.98). Dementia evidence (HR 1.70) rests on two Mayo cohorts with very low GRADE. Strong protection against ovarian cancer (HR 0.11) and breast cancer before 45 must be weighed.
10. Relevance to the specific claim
Supports C-ORG-09/11 with trade-offs (lower breast cancer); mortality evidence inconsistent.
11. Independent contrary findings
12. Transferability to Germany and Europe
Applicable.
13. Overall methodological judgement
Best available summary; supports risk-benefit framing of elective BSO in younger women, but does not show harm for women after 50 and most harm estimates are low/very low certainty.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | high |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Effect of HRT; salpingectomy alone; comorbidity-adjusted estimates; absolute risks.
15. Linked fact checks
- Bilateral oophorectomy at hysterectomy for benign disease increases all-cause mortality, coronary disease and lung cancer — roughly one extra death per nine operations (Parker 2013, Nurses' Health Study). – Studies with mixed or unclear results
- Oophorectomy before menopause raises the risk of cognitive impairment or dementia (HR 1.46), more so the younger the age at surgery (Rocca 2007). – Studies with mixed or unclear results