To the editor—Hojo et al. described a tumor growth-promoting effect of the immunosuppressive drug cyclosporin A (CsA) in immunodeficient mice1. The effect was attributed to the upregulation of the cytokine transforming growth factor beta by CsA. CsA is used not only in transplant medicine to prevent rejection but also in autoimmune diseases. An unwarranted effect of CsA in the induction, growth or behavior of malignant tumors would be of the utmost importance for clinicians and patients.

In the medical field of rheumatology, we also have feared an increased frequency of malignancies in CsA-treated patients and in particular in patients with rheumatoid arthritis (RA), an autoimmune disease with an intrinsically increased risk of malignant lymphomas. To address this concern, we undertook a retrospective controlled cohort study of all RA patients who participated in CsA clinical trials in the Netherlands between 1984 and 1992 (ref. 2). Every index patient was matched (for age, sex and duration of disease) with two control patients who had never been exposed to CsA. Neither index patients (n = 208) nor control patients (n = 415) had a history of malignancies (according to the Pathological Anatomical National Automated Archives, a database of all histological examinations in the Netherlands since 1974). No increased risk of malignancies was found, either in general or for site-specific cancers. In fact, the relative risk of developing cancer for the CsA-treated patients (relative risk, 0.41; 95% confidence interval, 0.19–0.89) suggested a protective rather than a tumor-promoting effect. This effect did not disappear after correction for many potential confounders.

Challenged by the findings of Hojo et al., we assessed whether the observed effect of CsA was dependent on CsA dose or CsA treatment duration, and how the observed incidence of cancer in our study cohort compared with the expected incidence in the general population. The protective effect by CsA was dependent on the duration of treatment—patients who had used CsA for more than 1 year had approximately 400% less chance of developing a malignancy than patients who had used CsA for 1 year or less. The standardized incidence ratio (compared with the normal population) for malignancies was 0.82 (0.36–1.62) in the subgroup of RA patients that were treated with CsA for more than 1 year, compared with 1.98 (1.37–2.69) in the control group. This suggests that RA patients who are on long-term treatment with CsA have a risk of developing malignancies that is equal to the risk in the general population, and lower than the risk in other RA patients.

These epidemiological results suggest that in the absence of further evidence, the tumor-promoting effects of CsA seen in the laboratory setting should not be extrapolated to the clinical situation.

Table 1 Cases of cancer per 1,000 follow-up years with respect to treatment duration with cyclosporine A
Full size table