Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease which predominantly affects women in the fifth and sixth decades of life. In addition to biochemical evidence for cholestasis, the presence of serum antimitochondrial antibody (AMA) is considered a diagnostic hallmark based on its occurrence in approximately 90% of patients.1 For asymptomatic patients at diagnosis, the cumulative risk for developing symptoms is 50% over 5 years.2 Subsequently, the risk for disease progression appears to be increased among symptomatic versus persistently asymptomatic individuals.3 Despite the therapeutic benefits associated with ursodeoxycholic acid,4 a number of patients will develop end-stage liver disease resulting in death or the need for liver transplantation.5 Based on the measurable risk for disease burden, a number of studies describing the clinical epidemiology of PBC have been reported.6 From investigations originating in Europe, Australia, and the United States, the annual incidence and prevalence rates of PBC vary between 2 to 24 and 19 to 402 cases per million population, respectively.7–12 Experience to date is consistent with a stable rate of new diagnoses over time with prevalent cases primarily representing patients with early-stage disease.12
Deciphering the pathogenesis of PBC, however, continues to be a scientific challenge. Perhaps the only certainty, amid several proposed mechanisms, relates to the consensus-driven hypothesis that PBC develops from an interaction between environmental factors and inherited genetic predisposition. Evidence already exists suggesting that the environment contributes to the development of PBC. Large case-control studies have shown that a history of smoking and urinary tract infection are significantly associated with the presence of PBC.13, 14 Moreover, familial studies have provided evidence that PBC aggregates in biological relatives.15 For example, the reported relative risk ratio for a sibling (i.e., λs) of a patient with PBC to develop the disease or to have detectable AMA in serum is 10.5-fold15 and 18-fold (K. Lazaridis, unpublished observations), respectively, compared with the general population. Despite these observations, the pragmatic contest for investigators of this disease is to dissect those environmental and genetic factors that contribute to the development of PBC.
To date, there has been only one investigation examining the association between a specific environmental exposure and risk for PBC.16 In this study, an estimated 88% of patients with PBC shared a particular reservoir as their water source in a defined area of Sheffield, England. No other obvious geographical features explained this distribution. However, the variation in source and composition of a particular water supply over time will make it difficult to determine the exact nature of this association in other populations. In contrast, the technique of geographical analysis may be used to identify variations in disease risk by location within spatial boundaries and whether clustering of cases is present within specific areas. Using this approach, Prince et al., identified the presence of strong variations in risk and clustering of patients with PBC in Northern England supporting the notion that environment strongly influences the occurrence of PBC.17
Several limitations exist, however, in using geographic analysis for the description of disease. The major obstacle is a lack of precise identification for the potential environmental risk factor associated with spatial variation or clustering of PBC cases. To this extent, one could further raise questions about the rarity of the exposure and required selectivity of effect on women as compared to men. For cases which appear to be clustered, the existence of populations with active migration patterns could also be the reason for variation rather than exposure to a specific agent. Finally, the use of non-systematic methods for selecting zip codes to represent random sampling of controls is highly dependent on population intensity by zip code.18
In this issue of HEPATOLOGY, Ala et al.,19 utilize methods based on geography and population density to study the prevalence and putative clustering of patients with PBC near superfund toxic waste sites (SFS), in New York City (NYC). These sites have been designated as dangerous toxic waste locations by the Department of Environment and Conservation. Indeed, previous studies have shown that thyroid disease was highly prevalent among females residing close to SFS in NYC.20 In this cluster analysis, the authors included PBC patients who received orthotopic liver transplants (PBC-OLT) as well as those not listed for liver transplantation (PBC-MSSM). In addition, patients transplanted for PSC (PSC-OLT) were used as a control group. The investigators compared the prevalence ratio of both PBC-OLT and PSC-OLT for each zip code (n = 174) and borough (n = 5) of NYC with regard to proximity or density of SFS. Specific software was utilized to define the proposed clusters of PBC-OLT and PBC-MSSM cases. The authors reported that the prevalence ratio of PBC-OLT was significantly higher in zip codes containing or in proximity to SFS when compared to zip code boundaries distant from SFS (1.225 vs. 0.670 respectively, P = .025). That was not the case for the PSC-OLT cases of the study. In addition, at the borough level, Staten Island had the highest prevalence ratio of PBC-OLT cases (1.54) and density of SFS (3.15/100 sq. miles). Based on these data, Ala et al., conclude that exposure to toxins, likely related to SFS, may be a risk factor contributing to PBC pathogenesis.
We have to accept that despite its limitations, this is the first study addressing the possible geographic clustering of PBC cases in the United States. Moreover, it shows the gathering of PBC cases within or in proximity to known sites of environmental toxins. However, several clinical questions remain unanswered. What percent of PBC-OLT patients residing within or adjacent to SFS are women, have thyroid disease and/or exposure to smoking or prior urinary tract infection? Why is it that only a small percent of PBC-OLT individuals reside close to an SFS? Finally, what might be the environmental element contributing to the development of PBC in patients not living in proximity to an SFS?
One of the primary methodologic concerns of the study relates to the mobility of the PBC cases in NYC. The residential zip code at the time of liver transplant waiting list registration was used for case ascertainment. Consequently, the PBC-OLT patients may have been living at several locations prior to diagnosis of PBC or listing for OLT. This limitation exists despite the effort of the authors to minimize the potential bias. Another concerning issue relates to the fact that statistical significance was detected only for PBC-OLT patients and not for cases with early disease (i.e., PBC-MSSM). Therefore, this raises the question as to whether the exposure to an SFS influences the risk for disease progression rather than the pathogenesis of PBC itself. Other important issues that are not addressed include: (1) the duration of time at a particular residential address when listed for OLT, (2) the dose-response relationship between a potential exposure from an SFS and the risk for PBC, and (3) whether any relationship exists between exposure, duration, and severity of disease. Determining whether geographic proximity to an SFS affects disease progression following diagnosis compared to disease incidence would be a topic of great interest for additional study.
An ideal method to study the outcome of environmental risk factors on health and disease would include the identification of large geographically-based cohorts, the prospective collection of exposure data over time, and the determination of strength of association between selected exposures and disease once it develops in specific individuals. This approach has been successful in elucidating the pathogenesis of several major causes of death and morbidity including coronary artery disease and cancer.21 While this approach may be tremendously difficult to apply for a condition as rare as PBC, the ability to better identify individuals at-risk for PBC may allow for studies of this nature to be conducted in the future.
Robust methodologies are currently available to screen the human genome for genetic variation or other inherited elements that might predispose to the development of PBC. To this end, it may be more appropriate to think in terms of “susceptibility genes” and “sensitivity environmental factors” that interact continuously in selected individuals to cause PBC. Accepting these concepts is essential to conduct future epidemiological and translational studies which attempt to untangle the processes resulting in PBC.