© 2004 American Thoracic Society doi: 10.1164/rccm.2409006
SarcoidosisIn Search of the CauseImperial College of Medicine London, United Kingdom The cause of sarcoidosis has proved tantalizingly elusive. The characteristic granulomata suggest a response to persistent and poorly degradable antigen, or antigens, but, with the exception of beryllium and its salts, no exogenous cause has been consistently demonstrated. A time–space cluster of sarcoidosis has been reported on the Isle of Man in the United Kingdom (1), and seasonal clustering has been reported in Greece (2), Spain (3), and New Zealand (4), with the peak in patient numbers occurring in the spring months (in New Zealand between August and October). These observations are consistent with an infectious (or contagious) etiology. Mycobacteria and more recently Propionobacterium acnes and granulosum have been proposed as plausible candidates, but direct microbiological studies have not as yet provided consistent or strong supporting evidence (5). Most etiologic studies of sarcoidosis have used an epidemiologic approach, but the disease poses particular difficulties. Sarcoidosis can be present without clinical manifestations and often only be detected by a chest radiograph. This may, at least in part, account for the reported cluster of cases on the Isle of Man, a high proportion of whom were nurses at the island's main hospital. Similarly in Rochester, Minnesota "asymptomatic" sarcoidosis was more commonly detected in recent immigrants and health care professionals, both more likely to have more frequent chest radiographs (6). Because the disease is relatively uncommon, a case–referent study design is generally used, often with cases identified from a variety of clinical sources. The choice of an appropriate referent population—to provide, as a basis for comparison, an estimate of the distribution of exposure in the population from which the cases were drawn—becomes especially difficult. In this issue of the Journal (pp. 1324–1330), a very large case–referent study (ACCESS) of the etiology of sarcoidosis is reported (7). A total of 706 cases and an equal number of control subjects were matched on age, race, and sex. The study reports associations with agricultural employment and exposure to microbial aerosols or insecticides at work. As frequently observed before, but not readily explained, there was a negative association with cigarette smoking. Cases were recruited to ACCESS from academic medical centers, making it likely that these represented the more severe spectrum of the disease. Control subjects, not known to have the disease, were recruited by random digit dialing, with a high number of calls (on average 216) needed to recruit each control subject. This is unfortunate, reducing the likelihood of the chosen control subjects having a distribution of exposure representative of the source population of the study's cases, potentially diluting the strength of real associations. Confident inferences can most readily be drawn from epidemiologic studies with associations showing strong relative risks (e.g., > 2 implying more than 50% of cases are attributable to this cause). The absence of strong associations reported in the ACCESS study may also reflect a multifactorial etiology, with sarcoidosis possibly being an immune-mediated response to more than one extrinsic agent. The observed associations are, however, consistent with the findings of some other studies and have some degree of biological plausibility. For example, the associations within agricultural employment and exposure to microbial aerosols might support an infectious causative agent. Despite intensive efforts, this and other similar studies have largely failed in identifying an external agent or agents responsible for sarcoidosis. In this way, arguably, they expose the limitations of epidemiologic investigation. Epidemiology thrives on heterogeneity of exposure: the more widespread the external agent responsible for sarcoidosis in the environment, the less it explains the distribution of cases and the more these are determined by individual susceptibility. As Geoffrey Rose pointed out: "if everyone in the country smoked 20 cigarettes a day then clinical, case control and cohort studies alike would have led us to conclude that lung cancer was a genetic disease" (8). There is now evidence for important genetic influences in the development of sarcoidosis. A study in the United Kingdom found the relative risk of sarcoidosis in siblings of affected individuals to lie between 36 and 73 (9), and a previous report from ACCESS of cases in the United States reported an overall risk of 5.8, with 18 in whites and 2.8 in African Americans (10). There is also direct evidence for genetic susceptibility in sarcoidosis, including erythema nodosum and Löfgren's syndrome, a combination which in a study of British and Dutch patients was strongly associated with HLA DQB1*0201 (11). To be informative, future epidemiologic studies investigating putative external causes of sarcoidosis will probably need to take indicators of genetic susceptibility into account. This approach was used in investigation of the (protective) role of a farming childhood in the development of childhood respiratory allergies. Among Alpine village children, a strong association between polymorphisms in the TLR-2 gene (involved in the response to bacterial lipoprotein) and the prevalence of asthma and atopy among children raised on farms was found; no such association was seen among children from the same villages who had not been brought up on a farm (12). Only the consideration of both environmental and genetic factors allowed the identification of the relevant contribution of both. It is likely that further advances in the etiologic understanding of sarcoidosis, either through epidemiologic or microbiological study, will only be made if they incorporate the increasing understanding of those genetic factors that confer susceptibility to its development. FOOTNOTES Conflict of Interest Statement: A.N.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
