Clinical epidemiological studies of the association between chronic inflammation, immune-modulatory therapies, and cancer

Abstract

Rheumatic diseases are chronic conditions that affect a substantial proportion of the adult population. Likewise, cancer is a major threat to public health, and the leading cause of death worldwide. Chronic inflammation is a key component in both rheumatic disease and cancer, and during the last decades, major treatment advances have been made in both of these fields. Due to the high prevalence of cancer in the age groups typically affected by rheumatic disease, it is a common comorbidity. Disentangling the association between rheumatic disease and cancer is complicated by the fact that cancer risk and prognosis can be affected by both aberrations related to host defense in rheumatic disease, as well as the treatment. For example, we know that risk of lymphoma is directly linked to disease severity in RA, but this does not preclude a further risk increase by RA treatment. In this thesis we capitalized on the rich data sources of Swedish national registers. Patients with rheumatic disease and matched comparators were identified, and by linkage to other registers, treatments, comorbidities, and other data were added. This allowed for comparisons by treatment and other characteristics within patient populations, and let us benchmark risks to that of the general population.

Studies I and II investigated the risk of cervical neoplasia in RA, and SLE, respectively. In Study I, the aim was to assess if there was an increased risk of cervical neoplasia in RA overall, and if TNFi-treatment increased this risk. In Study II we wanted to investigate the risk of cervical neoplasia in SLE overall, and if this risk differed between treatment-defined subgroups. We tried to separate the risk associated with the respective disease itself, from that of any potential risk carried by immunosuppressant treatment of RA and SLE. In these studies, we considered factors which were associated with the exposure and the outcome, and could act as confounders of the risk of rheumatic disease/treatment on cervical neoplasia. We found that there was an increased risk of cervical neoplasia overall in both RA and SLE, and that these risks were further increased in subsets treated with TNFi in RA, and other immunosuppressants in SLE, although the extent to which this was a direct effect of the treatments was hard to disentangle.

Study III investigated the risk of incident cancer, overall and by cancer site, in RA patients treated with TNFi and other bDMARDs. Five cohorts of RA patients initiating treatment with tocilizumab, abatacept, rituximab, and a first or second TNFi, were assembled, as well as a csDMARD treated cohort. With the exception of an increased risk of squamous cell skin cancer in abatacept-treated, there were no significant risk differences between bDMARD-, and csDMARD treated RA. We concluded that short- to medium-term use of tocilizumab, abatacept, rituximab, or TNFi drugs seems to be safe with regard to risks of incident cancer. Study IV investigated the association between RA and breast cancer, as well as anti-hormonal breast cancer treatment. In a matched cohort design, we replicated previous findings of a 20% decreased risk of breast cancer among women with RA. In a case-control design, we found that the risk of RA in women with breast cancer was also decreased. We found no evidence to support that anti-hormonal breast cancer treatment increased the risk of RA. Although we were able to take potentially important confounders into account, we could not disentangle the roots of the negative association between RA and breast cancer, which led us to conclude that it might be due to other shared factors.