Is primary biliary cirrhosis rare or common? The truth lies somewhere in between

Since 1851, when the first cases of primary biliary cirrhosis (PBC) were described, epidemiology has evolved to an indispensable tool in unfolding hidden sites of PBC 1. The very first serious attempt to carry out an epidemiological study of PBC was conducted in 1974 by Hamlyn and Sherlock 2. At that time, the view of Dame Sheila Patricia Violet Sherlock was that PBC is a very rare disease with almost a uniformly fatal outcome and no treatment to modify its course. Hence, a mortality survey would serve as an unbiased tool for the estimation of PBC in the general population 2. One hundred and sixty-five death certificates of PBC were issued over a five-year period (1967–1971) in England and Welsh leading to a crude death rate of 1.2 per million population per year for women reaching a peak of 4.1 in the 55–64 age group 2. During the last three decades, the natural history of PBC has been significantly modified. Sheila Sherlock's contribution was also vital in Deborah Doniach's pioneer work on developing a serological test for the accurate detection of antimitochondrial antibodies (AMA), the hallmark autoantibody of the disease 3, 4. We both met Sheila Sherlock and discussed her views about the diagnostic value of this test. Her strong view up to her death was that PBC is to be questioned in the absence of AMA. The widespread application of AMA testing has transformed the diagnosis of PBC and our perception of the natural history of this disease. The recognition of asymptomatic PBC which now comprises over 60% of the affected cases is largely based on the presence of high titer AMA in the absence of signs and symptoms of overt disease 5. The use of ursodeoxycholic acid, which constitutes the only approved agent for PBC, has been shown to improve survival in patients with stage I and II disease, similar to that reported in the general population6. In addition, the advent of transplantation has significantly improved survival rates. Over the last few decades, a number of epidemiological studies, mostly case-population studies, have substantially changed the geoepidemiology of PBC, pointing towards an increased prevalence and incidence of the disease. Studies such as those originated from Newcastle in North-East England and Minnesota in the US have estimated a prevalence of approximately 40 cases per 100 000 inhabitants, leading to the conclusion that PBC is not as rare as was considered in the past 7, 8. However, the prevailing notion that the prevalence of PBC follows a pattern of north–south gradient has been questioned recently. There is no study up to now that estimated the prevalence of PBC in a whole country, the only exception being that of Iceland (an island of 318 000 inhabitants), where the prevalence of PBC reached 64.4 per 100 000 women 9. In this issue of the Journal, Boonstra and colleagues present data from the largest up to the present population-based study conducted in Holland. The size of the catchment area which encompasses 6 adjacent provinces with 157 municipalities is profoundly the biggest strength of this study 10. The significance of the obtained data becomes more evident if we consider that the area studied comprises 50% of the Dutch population and corresponds approximately to 6 million people over the age of 20 10. Conducting population-based studies has always been challenging especially for diseases considered to be rare, with heterogeneous clinical presentation and complex case definition. In order to achieve the most precise estimation of the incidence and prevalence of a disease, one should aim towards accurate definition of the numerator and denominator populations. The correct definition of the population at risk, which comprises the denominator, is of immense importance. Before Boonstra's work, all but few studies largely referred to small sample sizes that prevented from accurate assessment of the prevalence and incidence rates of PBC. Among the exceptions, the Newcastle study investigated an estimated population of over 2 million people in UK during the period 1987–1994 7 and the Australian study conducted in Victoria referred to a population of just less than 5 million people between 1990 and 2002 11. While the former study reported one of the highest ever incidence and prevalence rates (3.22 and 34 per 100 000 inhabitants and 94 per 100 000 women over 40 years, respectively), the latter showed one of the lowest prevalence rates ever (5.1 per 100 000 inhabitants). The Dutch study combined meticulous case-finding and case-ascertainment strategies and identified 992 PBC patients in 44 hospitals leading to a mean incidence of 1.9 in women and a point prevalence of 13.2 per 100 000 inhabitants 10. These rates are similar to those reported in the neighbouring Sweden and Finland 12, 13, but much lower to the rates reported in Iceland 9. Both incidence and prevalence are reported to have been increased significantly during the 8-year period of the study 10. In agreement to the data from the Dutch study, two recent studies from Spain and Israel have also pointed towards an increase in the incidence and prevalence of PBC during a study period of 12 and 20 years, respectively 14, 15. In contrast, three studies from USA, Canada and Iceland showed that PBC prevalence rose over time, while incidence remained stable 8, 9, 16. Whether differences in the geoepidemiology of PBC are real or can be attributed to disparity in physician's expertise and accessibility to healthcare facilities is still questionable. The most likely explanation of the observed discrepancies appears to be the heterogeneity in the methodology applied among studies. Nevertheless, the multiple case-finding methods (personal registries of physicians, hospital discharge registries, positive AMA results and death certificates) used since the late 90s 7 have set some standards and considerably improved the quality of case-finding and case-ascertainment 17. Thus, attempts towards more accurate definition of cases has probably minimized errors in the estimation of the numerator and prevented from misdiagnosis. The common belief of an increase in the prevalence of PBC over the years is in agreement with reports of an increase in the prevalence of autoimmune diseases over the last 40 years. Environmental factors have been considered important for this raise, but the exact mechanisms which are responsible for the induction of these diseases are poorly understood. A recent study spanning a 17-year time period in the well-defined geographical area of Newcastle, where prevalence and incidence of PBC have been reported to be among the highest worldwide 7, provided evidence of spatial variation and clustering in the incidence of PBC. This was associated with variations in social and demographic factors and suggested that certain aspects of deprivation are involved in its aetiology 18. These data support previous indications of an aetiological role for geographically dependent environmental triggers 19. Along this line, Boonstra et al. 10 have shown geographical differences in incidence rates in 2 of the 157 municipalities in the study region. These differences were mainly attributed to age differences rather than the influence of environmental factors. The authors conclude that the increase in incidence and prevalence rates is real, as patient registration, diagnostic approach and therapy remained unchanged over these years, and net growth was not caused by a decrease in number of deaths. Information of disease stage, the presence of concomitant diseases, liver-related complications including hepatocellular carcinoma is lacking making further analysis impossible. The genetic/ethnic background of the patients is unknown and speculations concerning a possible relation of sex differences and genetic influences and in the development of the disease cannot be made 20. Presentation of such data could have added more strength to the study. Boonstra et al. 10. have also attempted to collect and analyze a 10-item questionnaire on information regarding family history, smoking habits and reproductive factors. Other risk factors and comorbidities have not been investigated. This subanalysis included 464 patients with PBC and just 128 controls from the outpatient clinics of four of the 44 hospitals. The number of controls for this type of studies is rather small and cannot be compared to those participated in the large epidemiological studies conducted in USA, UK and France. Boonstra et al. have failed to identify risk factors associated with PBC. These results contrast those of former studies that have reported, for example, an association of cigarette smoking with PBC. In conclusion, the study by Boonstra et al. is the biggest conducted so far in continental Europe. It may have methodological flaws, but still provides clear messages which cannot be overlooked. First, the prevalence and incidence of the disease are rising over time. In addition, PBC is not as rare as we were thinking in the past; it appears that the disease is not as common as previously speculated. As for now, ‘the truth lies somewhere in between’. Conflicts of interest: The authors do not have any disclosures to report.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.