Cohort studyindependent cross-checkfull text checked★ key study I-INTC-09 in-depth review
Weight-loss surgery and rheumatoid arthritis
Bariatric surgery and the incidence of rheumatoid arthritis - a Swedish Obese Subjects study.
Overview
Maglio C et al.. Rheumatology (Oxford) 2020 · PMID 31321442 · DOI 10.1093/rheumatology/kez275
Over a median of 21 years, people who had weight-loss surgery did not develop rheumatoid arthritis more often than matched controls.
Does the study support the specific claim? No · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Test bariatric surgery and RA incidence
- How was the study done?
Matched cohort
- How many people, animals or samples were studied?
4036 adults with obesity
- What was actually found?
No association
- How large was the observed effect?
HR 0.92 (0.59-1.46)
- What do the results mean in everyday life?
Surgery did not raise the risk of rheumatoid arthritis.
- What are the study's limitations?
Few cases; older surgery types; not randomised
- What does it say about Elawady's specific claim?
Contradicts 'surgery causes autoimmune disease' for RA
- Do the results apply to Germany and Europe?
High
- What can explicitly not be concluded from this study?
Effects of sleeve gastrectomy specifically; other autoimmune diseases
Scientific in-depth review
1. Full reference
- Authors
- Maglio C et al.
- Journal
- Rheumatology (Oxford)
- Year
- 2020
- Identifier
- PMID 31321442 DOI 10.1093/rheumatology/kez275
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Funded by Knut and Alice Wallenberg Foundation, NIDDK/NIH (R01DK105948), Swedish Research Council, Sahlgrenska ALF grants, German state and federal ministries. One author was partly salaried via an AstraZeneca grant until 2018; another received lecture fees from AstraZeneca and Johnson & Johnson; others none.
2. Research question
Does bariatric surgery change the incidence of rheumatoid arthritis?
3. Study design and population
- Study design
- Cohort study – Prospective, non-randomised, matched controlled intervention cohort (Swedish Obese Subjects, SOS); secondary analysis of an endpoint not predefined in SOS
- Population
Adults aged 37-60 years with obesity (BMI >=34 men, >=38 women) without prevalent RA, recruited at 25 surgical departments and 480 primary care centres
- Country
Sweden
- Period
Recruitment 1 Sep 1987 - 31 Jan 2001; register follow-up to 31 Dec 2016; median follow-up 21 years (range 0-29)
- Sample size
4036 analysed (2002 surgery, 2034 matched controls) after excluding 11 with prevalent RA from 4047 SOS participants; 92 incident RA cases
- Inclusion and exclusion criteria
Age 37-60 y, BMI >=34 (men) / >=38 (women); exclusions incl. prior gastric/duodenal ulcer surgery, earlier bariatric surgery, recent ulcer, malignancy within 5 y, recent MI, bulimia, substance abuse, psychiatric contraindications; prevalent RA (register) excluded
4. Methods and endpoints
- Exposure / intervention
Bariatric surgery at baseline: 19% gastric banding, 68% vertical banded gastroplasty, 13% gastric bypass (no sleeve gastrectomy)
- Comparator
Matched controls with non-standardised conventional obesity treatment (from intensive lifestyle to none); matching by sequential treatment assignment on 18 variables, not individual
- Endpoints
Incident RA identified by ICD-8/9/10 codes in the Swedish National Patient Register (inpatient and hospital-based outpatient care; no primary care); seropositive RA as secondary outcome
- Statistical methods
Kaplan-Meier, log-rank tests, Cox proportional hazards (adjusted for sex, age, BMI, obesity duration, ESR, CRP, smoking); ITT and per-protocol (285 controls who later had surgery censored); interaction surgery x sex; BMI change at 2 y as exposure
5. Actual results
- Results
92 incident RA cases (51 seropositive, 17 unknown serostatus). Adjusted ITT HR 0.92 (95% CI 0.59-1.46), P=0.74; per-protocol HR 0.86 (0.54-1.38), P=0.53; log-rank P=0.88 (ITT), P=0.93 (per protocol); seropositive RA log-rank P=0.52; no surgery x sex interaction (P=0.92). BMI change at 2 y: HR 0.96 (0.82-1.12) ITT. CRP per 10 mg/L HR 1.28 (1.06-1.55); smoking HR 1.63 (0.99-2.67) ITT, 1.69 (1.02-2.81) per protocol.
- Absolute risks
Incident RA in 47/2002 (2.3%) of the surgery group vs 45/2034 (2.2%) of controls over median 21 years
- Authors' conclusion
No association between bariatric surgery and RA incidence over up to 29 years.
6. Strengths
- Long follow-up (median 21 years) with nationwide register linkage
- Matched controls on 18 variables; ITT and per-protocol analyses
- RA in Sweden diagnosed by rheumatologists, increasing register validity
- Prespecified multivariable adjustment for known RA risk factors
7. Weaknesses
- Not randomised (ethics boards did not allow randomisation in 1987); surgery group had higher BMI, CRP and smoking at baseline
- RA was not a predefined SOS endpoint; authors state the study may be underpowered (92 events)
- Outcome from register only (no primary care), possible missed cases; serostatus unknown in 17
- No sleeve gastrectomy (banding, vertical banded gastroplasty, bypass only)
- Only one autoimmune disease examined
8. Risk of bias
- Orienting tool
- ROBINS-E
- 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
- some concerns
- Rationale
Single-reviewer structured ROBINS-E judgement from full text; not a formal assessment (would require two independent reviewers and a protocol). Main limitation for the linked claim is low power and the absence of sleeve gastrectomy rather than bias.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Self-selection into surgery; matching on 18 variables and adjustment for main RA risk factors, but residual imbalance (surgery group higher BMI, CRP, smoking) and unmeasured confounding possible. |
| Exposure measurement | low | Surgery is a documented, unambiguous exposure; per-protocol analysis handles later crossover of controls. |
| Selection of participants | low | Start of follow-up aligned with surgery date for both groups; only 11 prevalent RA excluded. |
| Post-exposure interventions | some concerns | 285 controls later underwent surgery; non-standardised control treatment. |
| Missing data | low | Register-based follow-up with near-complete coverage; 924 died/emigrated and were censored; serostatus missing in 17 affects only secondary analysis; CRP available for 3693/4036. |
| Outcome measurement | some concerns | ICD codes in a patient register not covering primary care; not validated by chart review, though rheumatologist diagnosis required; outcome assessment not blinded but register-based. |
| Selection of reported result | low | Both ITT and per-protocol results and seropositive subgroup reported; null result reported transparently. |
Full-text review (25 October 2026): Added absolute risks (2.3% vs 2.2%), procedure mix (no sleeve gastrectomy at all), recruitment period, inclusion criteria, register limitations, author-stated underpowering, funding/COI; full ROBINS-E domains (overall unchanged: some concerns).
9. Assessment of the statistical interpretation
Wide CI (0.59-1.46) compatible with roughly 40% lower to 46% higher risk; absolute incidence nearly identical (2.3% vs 2.2%). No signal of increased RA risk, but the study cannot exclude a moderate effect in either direction.
10. Relevance to the specific claim
Directly tests whether bariatric surgery increases one major autoimmune disease; no increase was found. Mostly older procedures, so indirect for sleeve gastrectomy.
11. Independent contrary findings
None for RA.
12. Transferability to Germany and Europe
Swedish data; directly European.
13. Overall methodological judgement
Best available direct evidence identified against an RA increase after bariatric surgery.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | low |
| Directness | indirect |
| Reproducibility | unknown |
| Consistency with the overall evidence | unknown |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Data for sleeve gastrectomy (not represented in SOS) and for other autoimmune diseases; adequately powered studies.
15. Linked fact checks
- Sleeve gastrectomy causes autoimmune diseases and fibromyalgia. – Studies contradicting or limiting the claim