Epidemiological aspects of sarcoidosis : risk factors and long-term consequences

Sammanfattning: Sarcoidosis is a systemic inflammatory disease of unknown etiology in which granulomatous lesions form mostly in the lungs and the lymphatic system of patients. Although more than a century has passed since sarcoidosis was first described, our understanding of its etiology and clinical course is limited. That is because epidemiological studies on large and representative patient cohorts have been lacking. The scope of this thesis was to examine aspects of sarcoidosis epidemiology using a linkage of large, nationwide health and administrative databases from Sweden complemented by clinical data. Six individual studies are included in this thesis; the first two dealt with risk factors for sarcoidosis, namely familial and infectious disease, and the rest with long-term debilitating patient outcomes: mortality, infection, and heart failure. In Study I, a case-control-family study, we estimated familial relative risks and the heritability of sarcoidosis. We found that having first-degree relatives with sarcoidosis increased the risk of being diagnosed with the disease by more than threefold. 39% of the susceptibility to sarcoidosis in the Swedish population was estimated to be attributable to additive genetic effects; the rest was due to non-shared (among siblings) environmental factors. Study II was a case-control study in which we estimated relative risks of sarcoidosis associated with having a history of infectious disease diagnoses. We showed that infectious diseases (commonly upper respiratory and genitourinary) diagnosed before sarcoidosis diagnosis were associated with a small increased risk of sarcoidosis in the future, a relative risk that did not vary markedly by latency period between infectious disease ascertainment and sarcoidosis diagnosis. These small relative risks could be easily explained away in analyses designed to test the robustness of these associations in the presence of reverse causation bias. In Study III, a cohort study, we followed individuals with sarcoidosis and general population comparators for all-cause death. We showed that there was an overall 61% increased risk of death associated with sarcoidosis. Stratification by treatment status around the time of sarcoidosis diagnosis approximating disease severity revealed a 2.3-fold higher risk of all-cause mortality compared to the general population in those treated while no risk increase was observed for untreated patients with sarcoidosis. Similarly, in Study IV, we followed individuals for a first or recurrent serious (hospitalized) infections. We observed a 1.8-fold higher risk of serious infection in sarcoidosis compared to the general population, which was even higher during the first two years since diagnosis and in individuals who were treated with an immunosuppressant around sarcoidosis diagnosis likely due to more severe or progressive disease at the time. In Study V, a target trial emulation, we compared six-month risks of infectious disease in initiators of methotrexate compared to azathioprine, two second line treatments for sarcoidosis. Six months after treatment initiation, a 43% lower risk of infectious disease was observed in the methotrexate compared to the azathioprine group. Study VI was a cohort study in which we examined the relative risk of heart failure and its predictors in sarcoidosis. We found a 2.4-fold increased relative risk of heart failure associated with sarcoidosis that was higher during the first two years since sarcoidosis diagnosis and in individuals without a history of ischemic heart disease. Diabetes, atrial fibrillation, and other arrhythmias were the strongest clinical predictors of heart failure diagnosis in sarcoidosis. Overall, findings from studies on risk factors in this thesis suggest that familial disease and genetics are important in sarcoidosis, albeit a larger contribution to the etiology of sarcoidosis is likely due to environmental factors. Among environmental factors, clinically identifiable infectious diseases are unlikely to be strong risk factors for sarcoidosis diagnosis. Future molecular and epidemiological studies on environmental triggers of sarcoid inflammation and disease should consider the issue of reverse causality owing to long preclinical disease in some patients. Studies on long-term patient outcomes in this thesis showed that sarcoidosis is not a ‘benign’ disease. Therefore, our quest to identify effective interventions and groups of patients to target should continue. If applied early, these measures can help alleviate some of the risks related to infection and heart failure, and improve life expectancy, especially in patients with severe or chronic disease.