Portal de Periódicos da CAPES

Risk Factors for Hepatocellular Carcinoma: A Historical Perspective

2021; Lippincott Williams & Wilkins; Volume: 18; Issue: S1 Linguagem: Inglês

10.1002/cld.1042

2046-2484

Yamini Natarajan, Hashem B. El–Serag,

Hepatitis B Virus Studies

Content available: Audio Recording Yamini Natarajan, M.D. Hashem B. El-Serag, M.D., M.P.H. Our understanding of the history of liver cancer is hampered by the difficulties encountered by its early investigators until the dawn of the 20th century CE. Identification of a liver lesion as a primary malignancy, which we would nowadays recognize as hepatocellular carcinoma (HCC), was ambiguous in previous eras because of the imprecise interpretations of the gross appearances and, until the beginning of the 19th century, a lack of microscopic anatomy. Here we share selected stories of the discoveries in liver cancer and highlight the combination of people and technology behind them, as well as how they used the foundations of study design to achieve their goals. We attempt to acknowledge and credit the legions of clinicians, pathologists, and scientists of yore who over the millennia have contributed their insight, creativity, and technical expertise to this field of hepatology. Descriptions of malignant tumors surfaced as early as the seventh to fourth centuries BCE and again in the second millennium BCE.1 In the earlier of these two eras, there were seeming allusions to the subject matter of the current historical essay in the Ramayana, the epic tale in Sanskrit that relates the travels of Rama, the legendary prince of the Kosala Kingdom, during his 14-year exile. Such allusions were also implicit in the second century BCE Rig Veda (Fig. 1), the oldest of Hindu religious texts (the four Vedas) that originated in ancient India and were composed in Vedic Sanskrit, an archaic form of that Hindu language.2 The ancient Egyptian medical papyri that date from approximately 3000 to 1500 BCE (Ebers, Smith, and Petrie) also contain some of the earliest accounts of afflictions of hepatological malignancy, especially the Ebers papyrus that was concerned particularly with tumors,3 but it is anyone's guess how reliably cancerous lesions were distinguished from inflammatory swellings and chronic ulcerations. In Western medical literature, Hippocrates of Kos (~400 BCE; Fig. 2) reputedly coined the term karkinoma (καρκίνωμα, from the Greek for "crab")4 that 600 years later was popularized in Roman medicine by his remote disciple, the Greek physician Galen (200 CE).4 Cancer, latinized from Karkinoma, entered the English language in the 17th century. What was, in effect, highly likely to have been HCC was first described in the second century CE by Galen's contemporary, Aretæus5 (Fig. 3), from Cappadocia in east-central Anatolia. This Cappadocian physician, who practiced in Alexandria in Egypt, even envisioned the sequence of pathological steps from liver inflammation (i.e., hepatitis) to the transformation of the liver from a swollen into a hardened state (probably cirrhosis) and ultimately to scirrhus, by which he probably meant cancer. The steatomata or hard tumors described in the liver at autopsy6 by Giovanni Battista Morgagni (1682–1771), who is credited with founding the discipline of pathologic anatomy, were almost certainly not HCC but metastases from stomach or pancreatic cancer, as surmised by Friedrich Theodor von Frerichs7 (Fig. 4), who is sometimes known as the Father of Modern Hepatology. In his classic 1860 A Clinical Treatise on Diseases of the Liver, which was translated in 1879 into English7 by Charles Murchison, M.D., L.L.D., F.R.S.,** Fellow of the Royal College of Physicians London, one-time President of the Pathological Society of London, among many other honors and appointments, staff physician to her Majesty's Bengal Army, and sole author, in his own right, of a textbook of liver disease.8 Frerichs acknowledges the discoveries related to the clinical presentation, prevalence, microscopic structure, and pathogenesis of liver cancer, by of a dozen or more legendary investigators in the 18th and 19th centuries. This pantheon of individuals included notables such as John Abercrombie, Gabriel Andral, Matthew Baillie, Giovanni Battista Bianchi, Vincent Alexander Bochdalek, Herman Boerhaave, Richard Bright, Friedrich Hoffmann, Thomas Hodgkin, Karl Freiherr von Rokitansky, Rudolf Virchow, and George Budd (Fig. 5), the author of the first textbook of liver disease in English.9 But it was the tuberculous friend and colleague of Laennec, Gaspard-Laurent Bayle in Paris, however, whom Frerichs credits with first publishing, in the 1812 Dictionnaire des Sciences Médicales, par une Société de Médecins et de Chirugiens,†† This is actually available for purchase online, a printed version scanned from the original. an accurate description of primary cancer of the liver. Bayle had pointed out that bona fide liver cancer, as judged by anatomic and behavioral similarity with cancer of the breast and other organs, had been described previously using inexact descriptive terms, such as steatomata, white bodies, tubercles, or scirrhus. Frerichs's catalog of the gamut of macroscopic and microscopic anatomy of primary liver cancer rivals that found in any modern textbook of liver pathology. However, detailed classification of HCC did not really occur until the late 1800s, with gross descriptions by Hanot and Gilbert10 and Eggel,11 the 1911 microscopy of Yamagiwa12 and Goldzieher and von Bokay,13 and the histopathological analysis of 100 cases (among 48,900 autopsies) published in 1954 by Edmondson and Steiner.14 In earlier accounts, the distinction between primary and metastatic disease was not always possible, and the renowned pathologist Virchow muddied the waters by his dictum in 1862 that "organs commonly affected by metastases are rarely the site of primary neoplasia."15 It was soon shown that liver metastases are 25-fold more common than primary cancer,16, 17 but that likely consequence of the liver's extensive blood and lymphatic supply does not preclude primary hepatic carcinogenesis. Confusion was also spread by the many synonyms for primary liver cancer that had been devised over the years, such as cancer trabeculaire,10 adenocarcinoma,18 malignant adenoma,19 carcinoma hepatocellulaire,13 and many others, including hepatoma, the last of which has been used for hepatocellular tumors that are benign, intermediate between benign and malignant, and both by no less an oncology eminence than James Ewing.20 The uncommon fibrolamellar carcinoma variant of HCC (FLC) that occurs in adolescents and young adults was originally called eosinophilic HCC with lamellar fibrosis, but was not published in detail until the 1980s,21, 22 although it may have been encountered inadvertently as early as 1956.23 FLC has unique epidemiological,24 clinical,21 radiological,22 histopathological,21, 23 and molecular genomic signatures.25 The prognosis of FLC is far more favorable than HCC,24, 25 and there is the added potential for distinct genetic-based molecular therapy.25 In the 20th century, advances in molecular biology brought forth new knowledge regarding the structure of the liver, the role of infectious pathogens, such as hepatitis B virus (HBV) and hepatitis C virus (HCV), the discovery of aflatoxin and other environmental toxins, behavioral risk factors, such as alcohol and obesity, and constitutional genetic diseases in causing liver disease, usually but not exclusively cirrhosis, and HCC. The pathogenic mechanism by which cirrhosis fosters HCC, an association noticed by Sabourin in 1881,26 will be covered in another essay later in this series, by Drs. Atoosa Rabiee and Lopa Mishra. The constitutional genetic disorders that predispose to HCC, glycogen storage disease I, Alagille syndrome, hereditary tyrosinemia, some of the familial cholestatic syndromes, neurofibromatosis, ataxia telangiectasia, and Fanconi anemia are discussed fully elsewhere,27, 28 as are hereditary hemochromatosis and other iron-overload states.28, 29 Establishing causality is a difficult endeavor. HBV, initially known as the "Australia antigen," discovered and characterized as an HBV surface antigen (HBsAg) by Baruch Blumberg in 1969, which will be discussed in Dr. Robert Gish's essay in a forthcoming issue in this series, was first evaluated in in vitro and animal studies, which revealed a relationship between infection and genesis of HCC via immune-mediated mechanisms.32 Furthermore, results of early case-control studies showed only a weak association between HBsAg, using an older and now known to be less accurate HBsAg test, and HCC.33 Subsequently, a landmark cohort study (1981) using radioimmunoassays and enzyme immunoassays, which detect much lower levels of HBsAg, demonstrated a markedly increased incidence of HCC in chronic HBV-infected men compared with non-chronically infected control subjects.34, 35 There is an enthralling story behind the cohort study. R. Palmer Beasley (Fig. 6) was an epidemiologist at the University of Washington School of Public Health. He was well known for demonstrating mother-to-child transmission in HBV. Despite his fame, his interest in the link between HBV and HCC was met with scepticism. Beasley wrote, "The prospect for funding seemed bleak in view of the overwhelming general belief that aflatoxins had already been virtually established as the etiology of HCC."33 However, Beasley and fellow epidemiologist (and wife) Dr. Lu-Yu Hwang evaluated the options of conducting a large enough cohort study with a follow-up duration long enough to arrive at precise HCC risk estimates among people with and without chronic HBV. At that time, there were no large databases, patient registries, or national electronic health record repositories. After noticing, serendipitously, that the leading cause of death among subjects in a prospective study of cerebrovascular disease was liver cancer, they embarked on a prospective cohort study of 22,707 male civil servants attending a Government Employment Clinic in Taiwan. The incidence rate of HCC in HBsAg-positive men was 1158 per 100,000 person-years versus 5 per 100,000 person-years in uninfected men. This translated to a relative risk for HCC of 223 in HBsAg-positive compared with HBsAg-negative men. Use of a cohort study design revealed the temporal trends in HBV-infected individuals. The study results also demonstrated the challenges of executing a prospective study; when US foreign policy shifted to view Taiwan as part of China, support initially waned. However, through funding via the US National Institutes of Health, Beasley was able to carry on with his work that subsequently garnered him numerous awards, perhaps the most colorful being the 1999 Prince Mahidol Award for Medicine for his discovery of the link between HBV and HCC, which was presented to him by King Bhumibol and Queen Sirikit of Thailand at the Grand Palace in Bangkok (Fig. 7). The causal association between HBV and HCC was nailed by another dramatic intervention: HBV vaccination in Taiwan. HBV vaccines were developed in the early 1980s. Taiwan's mass vaccination program against HBV was launched in 1984, and Prof. Mei-Hwei Chang and his group reported in a landmark article that the incidence and mortality rates for liver cancer declined in successive birth cohort generations between 1981 and 1994 of children 6 to 14 years of age in Taiwan.36 The aflatoxin story37 that had the potential for unraveling the molecular pathogenesis of HCC generally did not live up to expectations.38 Aflatoxins, the bisfuranocoumarin mycotoxin products of the Aspergillus flavus and parasiticus molds that contaminate peanuts, soybeans, grains, and cassava root in tropical areas, are indeed potent carcinogens in animals via some of their metabolites, and this includes hepatocarcinogenesis in rodents, nonrodent species, and nonhuman primates.39 Aflatoxin-associated HCC in humans, however, turns out to result from a synergism with HBV in individuals who have genetically determined variants of aflatoxin metabolism that result in an inactivating mutation of codon 249 of the p53 tumor suppressor gene.40, 41 Initially identified as posttransfusion "non-A non-B hepatitis," HCV affects millions worldwide.42 Antibodies for HCV were first discovered in 1989 by Michael Houghton, Qui-Lim Choo, and George Kuo of Chiron Corporation. Houghton described working for 6 years using a variety of molecular techniques. When they did isolate clone 5-1-1, which would eventually be known as the HCV antibody, the "Eureka" moment was in fact a gradual and extended one that required 6 months of confirmatory testing.41 (The history of HCV43 has been reported elsewhere in this series, by Harvey Alter, Patrizia Farci, Jens Bukh, and Robert Purcell in an essay entitled "Reflections on the History of HCV: A Posthumous Examination".44) Next, researchers from Hospital Clinic Provincial at the University of Barcelona, Spain, showed a high prevalence of these HCV-associated antibodies in patients with HCC,45, 46 and subsequent case-control studies showed an association between HCV and HCC. Akin to the HBV story, cohort studies were required but, fortunately, did not have to start from scratch; several investigators took advantage of preassembled cohorts with precollected blood that could be tested for HCV antibodies. Initially, well-publicized, population-based cohort studies of non-A non-B hepatitis failed to show a meaningful increase in mortality or HCC risk, whereas hospital- and clinic-based cohort studies showed a major increased risk.47-50 It was later understood that the risk for HCC in patients is related to long-standing infection and cirrhosis formation.51 With biological plausibility and temporal precedence established for an HCV-to-HCC causal pathway, covariation was still needed. Covariation requires the causal variable to vary systematically with the outcome variable that it is assumed to cause. However, with no effective vaccine (unlike HBV) and initial interferon-based treatment of HCV having low efficacy and high side effects (a combination that led to limited use), one had to wait a couple of decades. The advent of direct-acting antivirals ushered in excitement regarding widespread treatment of HCV, and larger studies with longer follow-up eventually showed a decreased risk.52-54 Alcohol was thought to have medicinal value since ancient Egypt; it was known as the water of "immortality" in the Middle Ages. However, in 18th-century England, when a surplus of corn led to increased liquor production and its excess consumption resulted in the so-called Gin Plague55‡‡ A term favored by US author Thomas Coraghessan Boyle, in his 1981 novel Water Music.56 , there followed the degradation of the working classes (as depicted in William Hogarth's renowned political cartoon, Gin Lane; Fig. 8) and a surge in cases of cirrhosis. In a companion cartoon, Hogarth echoed the widespread contrasting belief that beer was healthful. In the 1800s, Thomas Addison and Matthew Baillie separately described the association between alcohol and steatosis/cirrhosis, but the link to HCC was less clear.57 This school of thought did not gain momentum until the mid-20th century.58 Frank Lee59 evaluated patients with cirrhosis who sought care in The London Hospital§§ The London Hospital (where the current Series Editor was born) was the London Infirmary prior to 1757, the Royal London Hospital after 1990, and since 1994 has been joined with St Bartholomew's Hospital and the London Chest Hospital to become the Royal Hospitals Trust. from 1912 to1963. HCC occurred in 30% of patients with alcoholic cirrhosis versus 11% of patients with nonalcoholic cirrhosis. Subsequently, population-based cohort studies showed an increased risk in patients with alcoholic cirrhosis.60, 61 The World Health Organization reported a relationship between alcohol and numerous cancers, including those affecting the liver, oral cavity, and esophagus. In Europe and North America, where there was a lower prevalence of viral hepatitis, alcohol was determined only during the 1970s to be one of the biggest risk factors for cirrhosis and HCC.58 The alcohol-HCC association suffered from generally low estimates of risk, especially compared with the dramatic values reported for HBV and HCV, there being a lack of a reliable and easy-to-measure marker of alcohol use in epidemiological studies, and the lack of a dramatic intervention that would illustrate covariation concepts previously discussed. Therefore, the alcohol-HCC association had to survive the peeling off of HBV and HCV from the mix of risk factors. Only recently (i.e., 2010 onward) have we started to see reports highlighting alcoholic liver disease as becoming a rising cause of HCC in the United States.62-64 The history of diabetes and obesity as possible risk factors for HCC is still being written. During his studies on insulin and the development of diabetes, J. von Mering showed in 1873 that increased hepatic steatosis arose in patients with diabetes.57 The search for a link to HCC grew as obesity rates increased in the late 1990s and early 2000s. The risk for obesity and diabetes has been evaluated in two case-control studies from Europe in the 1990s that reported an increased risk for cirrhosis and liver cancer with obesity and diabetes.65, 66 Results of studies in the United States examining the US veteran population confirmed this.67, 68 This risk remained when patients were examined prospectively, even after adjusting for alcohol use.69 During this time, awareness of a new entity, nonalcoholic fatty liver disease (NAFLD), led to the study of this syndrome as the driver for HCC in this population. Similar to the HCV story, results of initial NAFLD cohort studies with what appeared to be adequate follow-up, but perhaps underpowered, did not show any association with HCC; indeed, in one study, not a single case of HCC occurred. Apparently, the duration of follow-up in these studies was not decent enough to show an effect, knowing what we know now about the long duration required for cirrhosis and HCC to develop. It took larger, prospective cohort studies to demonstrate that patients with NAFLD have an increased incidence of cirrhosis compared with the general population. But it really took the technological advance of the availability of high-quality "big data" and the ability to analyze it to muster large retrospective cohort studies with large sample sizes that are outside the reach of what traditional, prospective research could ever do. Further studies are needed to elucidate whether this risk remains outside of that associated with cirrhosis per se.70, 71 Accurate descriptions of disease burden drive or subdue further curiosity and subsequent research. Palmer Beasley attributed the dearth of research into the HBV/HCC association to ignorance about HCC in other countries where it was a big problem. Here, it is worth pointing out that vast discrepancies are seen globally in the prevalence of HCC between countries (Fig. 9), which is sometimes referred to as a "geographical gradient" between regions (Fig. 10). Little wonder that the enormity of the scourge that is HCC in the less-developed nations and poor populations was unappreciated in the industrialized parts of the world. Incidentally, Fig. 10 also shows dramatically the sex discrepancy in HCC incidence. Part or most of the ignorance about this disease was a function of lack of reliable and easily accessible disease registries. Reports in the 1980s in the United States described HCC as follows: "Age-adjusted liver cancer death rates have been stable over time except for nonwhite males among whom there has been a 45% increase in liver cancer mortality between 1958 and 1975."72 Secular trends in HCC were examined initially in the 1990s. One of the authors (H.B.E.-S.) attributes his interest in examining HCC in the United States in the mid-1990s to a combination of personal factors (his own need to branch out from his postdoctoral research topic of esophageal cancer, his prior training in epidemiology and SAS [previously known as the Statistical Analysis System] programming), external chatter (anecdotal remarks and observations made by his Albuquerque, NM, Division Chief at the time, Denis McCarthy), and crucially, the recent availability of electronic databases, mostly unknown to clinical researchers. In North America, there had been little mention of the trends of HCC, likely because HCC was a rare/unusual tumor. The emergence of population-based cancer registries, such as the SEER (Surveillance, Epidemiology, and End Results) program of the National Cancer Institute of the US National Institutes of Health, the availability of vital statistics containing cause-specific mortality, and computerized repositories of electronic medical records from national health care systems, such as the Veterans Health Administration, have facilitated accurate and timely generation and reporting of secular trends in cancer. In 1999, El-Serag and Mason73 reported an increase in the average incidence of HCC in the United States from 1.4 to 2.4 per 100,000 (Fig. 11), which resulted in increased morbidity and mortality. Epidemiological studies that defined risk factors for HCC have paved the way to introducing and testing preventive care: HBV vaccine, HCV treatment, and alcohol-cessation programs. The history of how these risk factors were detected took a common pathway. Initial observations led to case-control studies; then researchers conducted prospective/longitudinal cohorts to evaluate HCC risk. Epidemiological evidence directed studies in basic and translational science to evaluate mechanisms and physiology. As we look to the future, it is important to remember the path taken for scientific advancement in the past, especially because epidemiological "age-period-cohort" models predict worrisome increases in HCC incidence in many, but admittedly not in all, countries studied.74 The increasing prevalence in obesity and metabolic-associated fatty liver disease gives no reason for complacency. The hallmark of this series of essays on the History of Hepatology is to have contributions from authors who are universally recognized authorities on the subjects of each manuscript and have invariably been responsible for many of the relevant discoveries and publications. Measured by these criteria, I could not have made a better choice than inviting Hashem El-Serag to chronicle the evolution of our understanding of the epidemiology of HCC, aided with consummate skill and charm by his junior faculty colleague at the Michael E DeBakey VA Medical Center and Baylor College of Medicine, Yamini Natarajan. Initially I thought that Dr. El-Serag's Arabic family name سراج was pronounced SEERAJ, and as such would be translated into English as "Bright Lantern," which seemed appropriate given his upward of 170 publications on HCC (out of a total of approximately 500) that have illuminated the field. Unfortunately, my unfamiliarity—in truth, utter ignorance—with diacritics (Al Tashkeel) in Arabic, those graphic symbols or glyphs that are added to letters to indicate pronunciation, specify that Serag is actually SARRAJ, which is the occupational name "saddle maker," itself a noble occupation. Whereas Dr. El-Serag would be the first to admit that he is not accustomed to the genre of this series, the final result compares more than favorably with the other offerings. In this essay, Drs. Natarajan and El-Serag lead us skillfully through the Byzantine HCC literature, where a lack of diagnostic methods confounded investigators of old, leading to early misinterpretation of liver lesions, a bewildering proliferation of nomenclature, and misunderstanding of pathogenesis. It did not help that investigators in the "first" world paid little heed to what was dismissed as a "third" world disease. This hapless situation was rescued in the 20th century by the application of the tools of molecular and cellular biology, histopathology, and immunobiology, and new epidemiology techniques involving the advanced computer-assisted analysis of huge population datasets, so-called big data, in which the senior author excels and illuminates our way.