Primary biliary cirrhosis (PBC) is an uncommon and enigmatic autoimmune disease, characterized by female predominance, autoantibodies to mitochondrial antigens (AMA), destruction of small bile ducts, and ultimately cirrhosis and liver failure. 1,2 Although genetic factors appear to be important in determining susceptibility to autoimmune diseases such as PBC, studies of a variety of autoimmune diseases in monozygotic twins have revealed concordance rates of only 10% to 50%. 1,2 Thus, environmental factors may trigger disease in susceptible individuals. Among the factors that have been suggested to play a role in the etiology of PBC are infectious diseases, 3 other autoimmune diseases, 4 and lifestyle factors, such as smoking, 5 and a high-fat diet. 6,7 Although many mechanisms have been proposed 3,7-9 by which these various factors may be involved in PBC, little epidemiologic research in the United States has been conducted; most published research consists of case reports and laboratory studies. The lack of conclusive information about the etiology of PBC is due, in part, to the methodologic difficulties of studying the disease and also to its relative rarity. Previous studies have been hindered by small sample sizes, unrepresentative samples, lack of a standard case definition, and lack of inclusion of appropriate comparison groups. 10 This report is unique because of its size and its use of two sets of controls, including siblings.
PATIENTS AND METHODSStudy Population. A total of 241 patients with PBC were identified from an electronic mail list server for an Internet support group for individuals with PBC. All people on the electronic mail list server were sent a questionnaire. Confirmation of the diagnosis of PBC was performed as described later. PBC cases who completed the survey were asked to provide names and addresses of their siblings to serve as a comparison group. The cases were later asked to provide the names and addresses of 3 friends who were the same sex, within 5 years of their age, and who did not have PBC to serve as a second comparison group. Siblings were sent the same packet of information as cases (see below), whereas friends were sent an abbreviated questionnaire, omitting the dietary questions. A total of 241 individuals with PBC, 261 siblings, and 225 friends were identified.Data Collection. The questionnaire for the study was developed using standardized questions drawn from the National Health and Nutrition Examination Survey (NHANES). 11 These questions have been tested for reliability and validity in a number of studies. 12 The initial survey instrument was pilot tested in approximately 25 individuals without PBC, after which appropriate modifications were made. The final survey included 209 questions, with many of them having several parts, for a total of 983 variables. The survey instrument included questions about demographics (age, education, gender, race/ethnicity), general medical history (including childhood illnesses, infectious illnesses, chronic diseases, surgeries, and m...