Cancer risk after an inflammatory disease diagnosis
The setting
Rheumatologists have long suspected that patients with immune-mediated inflammatory disease carry a higher cancer risk. The suspicion had never been tested on Italian national data, so nobody could say how large the effect was, which cancers it involved, or when in the disease course it mattered.
What we did
- A nationwide retrospective cohort study on Italian Hospital Discharge Records. Adults hospitalised between 1 January and 31 December 2018, followed to 31 December 2023.
- The exposed group was patients with immune-mediated inflammatory disease, including rheumatoid arthritis and diffuse diseases of connective tissue. The unexposed group was patients admitted for injury and poisoning with no such history. Multivariable logistic regression, adjusted for demographics, geography, follow-up time and comorbidities.
- 356,022 patients: 54,896 exposed, 301,126 unexposed.
The interesting result is the shape, not the headline. Risk is highest in the first year and falls year on year. That pattern is hard to explain by medication, which accumulates, and easier to explain by inflammation, which early treatment reduces. The paper states it as a hypothesis, which is the correct strength.
What exists now
A published, peer-reviewed result. Every figure in the table below is in the paper and checkable by anyone with the DOI.
| Finding | Value |
|---|---|
| Cancer association in the IMID group | adjusted OR 1.32 (95% CI 1.27 to 1.38), p < 0.001 |
| Year one | adjusted OR 1.83, p < 0.001 |
| From year five | adjusted OR 1.20, p < 0.001 |
| Lung | adjusted OR 1.74, p < 0.001 |
| Leukaemia and lymphoma | adjusted OR 1.98, p < 0.001 |
| Bladder | adjusted OR 1.48, p < 0.001 |
| Melanoma | adjusted OR 1.48, p = 0.009 |
| Connective tissue disease vs rheumatoid arthritis | OR 1.53 vs 1.20 |
What it does not do
- It produces no individual risk score, and it cannot. The study used anonymised administrative records with no names, identifiers, addresses or telephone numbers.
- It is an association at population scale. It does not tell any clinician anything about any patient in front of them.
What has not been established
- Causation.
- Whether the inflammation hypothesis is right.
- Whether early anti-inflammatory treatment actually lowers cancer incidence, which would need a different study design entirely.
- Whether the pattern holds outside hospitalised populations, since the cohort is built from discharge records.
Governance
- Funded by Ricerca Corrente from the Italian Ministry of Health, allocated to AGENAS, approved by the Technical Health Committee on 3 December 2025.
- Ethics committee approval was not required and informed consent was waived, because the study used anonymised administrative data under the AGENAS monitoring mandate in the Ministerial Decree of 19 December 2022, Article 4.
- Data are available from AGENAS on request.
Where it goes next
The paper's own answer: find inflammatory biomarkers that predict which patients are in the high-risk window. The temporal pattern makes that a sharper question than it looks, because whatever drives the effect is strongest in the first year and fades.
What we are looking for
Co-investigators with access to biomarker data on an immune-mediated inflammatory disease cohort.
Get in touch →Reference
Giordani B, Pirtoli L, Putignano G, Giordano A, Marotto D, Baglio G. Cancer Association in Patients with Immune-Mediated Inflammatory Diseases: A Five-Year Nationwide Italian Cohort Study. Cancers 2026. doi:10.3390/cancers18061027