Prevention of hepatocellular carcinoma by immunization*

/ / / ~~~~~~~~andup-to-date survey of the /bilan concis et fiable de la I /r[nzle to / / ~~present position in the se- /situation actuelle dans le I LII/ (1 / /te lected fields, and, over a domaine consid&re. Des ex-/ / / ~~~~~~~periodof years, will cover /perts couvriront ainsi suc-/ / / / ~~~~~~~manydiffierent aspects of /cessivement de nombreux / Xr + , , ~~~~~thebiormedical sciences /aspects des sciences bio-I e jo I l(t and public health. Most of medicales et de la santethe articles will be vrit-/ publique. La plupart de ces ten, by invitation, by ack- articles auront donc ett nowledged experts on the rediges sur demande par les subject. specialistes lesplus autorises. Bulletin ofthe World Health Organization, 61 (5): 731-744 (1983) © World Health Organization 1983 Prevention of hepatocellular carcinoma by immunization* A WHO MEETING The evidence for an association between the carrier state of hepatitis B virus infection and hepatocellular carcinoma (liver cell cancer) is now sufficiently strong to justify the use ofa vaccine against this infection as a means ofpreventing this cancer. Effective vaccines are available and have been tested in feasibility studies, and their use in field trials to test their effectiveness against the long-term risk of developing this cancer is now possible. At the present time, only limited quantities of the vaccine derivedfrom human plasma infected by the hepatitis B virus are available, and the development of other types of vaccine is discussed together with the necessary revision of the present requirements for vaccine production. As the studies to assess the prevention of this cancer would require the surveillance of subjects for some years, consideration is also given to the design of field studies with short- and medium-term objectives. EPIDEMIOLOGY OF LIVER CANCER Primary liver cancer includes hepatocellular carcinoma (relatively common), cholangio- carcinoma (rare), angiosarcoma (very rare), and other even rarer forms. These entities are not only pathologically distinct but also etiologically different. However, since hepato- cellular carcinoma is by far the most common of these cancers in virtually every country of the world, the descriptive epidemiology of primary liver cancer is very similar to that of hepatocellular carcinoma. Primary liver cancer is among the most common cancers in the WHO South-East Asia and Western Pacific Regions, and is perhaps the most common of all cancers in Africa, south of the Sahara (particularly in the south-east). Since these areas are heavily populated, * This article is based on the report of a WHO Meeting on Prevention of Liver Cancer, which was held in Geneva from 30 January to 4 February 1983. This report has been published in WHO Technical Report Series, No. 691, 1983. ' The participants at the meeting were G. Ada, W. G. van Aken, R. P. Beasley, N. Bhamarapravati, H. Brummelhuis, Chan So Ha, P. Coursaget, F. Deinhardt, A. Goudeau, M. R. Hilleman, Khin Maung Tin, S. Lagakos, Li Ho-Min, A. Linsell (Rapporteur), C. Loucq, J. E. Maynard, N. C. Nayak, K. Nishioka, G. T. O'Conor, Oon Chong-gin, R. Ryder, G. C. Schild, D. A. Shafritz, P. Smith, Sun Tsung-tang (Vice-Chairman), P. Toure, D. Trichopoulos, 0. S. Vyasov, G. P. Warwick, M. Yano, A. J. Zuckerman (Chairman); from WHO Secretariat-F. A. Assaad, T. A. Bektimirov, M. E. Chuwa, Y. H. Paik, F. T. Perkins, P. Sizaret, K. Stanley, J. Stjernsward (Secretary); from the International Agency for Research on Cancer (IARC) - N. Day, N. Mufloz, M. Parkin. Requests for reprints should be addressed to Chief, Cancer Unit, World Health Organization, 1211 Geneva 27, Switzerland. 4335 -731 732 A WHO MEETING primary liver cancer is among the ten most common cancers worldwide and it is estimated that there are at least 250 000 new patients with liver cancer each year. The actual incidence of the disease is at least 30 new cases per 100 000 population each year in some parts of Asia and Africa, whereas it is less than 5 new cases per 100 000 population per year in Australia and in most populations of Europe and North America. The incidence appears to be increasing with time in the majority of the low-incidence countries. Primary liver cancer is more common among males than females, the ratio being higher in the autopsy-diagnosed cases (median, around 5) than in the reported mortality rates (median, around 2). This difference is probably due to the confusion with other cancers that present with liver metastases. It is well established that the incidence of primary liver cancer increases with age, but in high-risk populations it also occurs in younger age groups. The disease shows a marked increased incidence in certain ethnic groups. Etiology of hepatocellular carcinoma Epidemiological data from case-control and cohort studies, and several laboratory investigations indicate that there is a consistent and specific causal association between hepatitis B virus (HBV) and hepatocellular carcinoma and that up to 800/o of such cancers worldwide are attributable to this virus. HBV is thus second only to tobacco among the known human carcinogens. Prospective studies As cohort or prospective studies have markedly strengthened the association between hepatitis B virus infection and hepatocellular carcinoma and have shown high relative risks, these studies will be examined in detail. Despite their obvious advantages, there have not been many such studies to determine long-term morbidity and mortality in relation to the various HBV markers. The high cost and the difficulties of long-term follow-up preclude such studies in many places. None the less, the several studies that have been carried out have all shown a marked increased risk for hepatocellular carcinoma among carriers of hepatitis B surface antigen (HBsAg), compared with persons with other HBV markers or with no evidence of previous hepatitis B infection. These studies have been of three types: (1) Follow-up of patients with chronic liver disease associated with HBV markers. Such studies have shown that HBsAg-positive patients with chronic hepatitis and/or cirrhosis have a much greater risk of developing hepatocellular carcinoma. (2) Retrospective investigation of HBV serum markers in patients participating in cohort studies. The patients were males of Japanese origin in Hawaii, blood donors in New York, and Eskimos in Alaska. A substantially higher rate of HBsAg was found in the patients who subsequently developed hepatocellular carcinoma than in the controls. (3) Prospective studies, in which the patients were unselected for disease but divided into different groups according to their HBV serum markers. This type of study is the most useful, but practical considerations make it difficult in most populations. However, two are already published and several others are currently in progress. In one study, with approximately 75 000 man-years of follow-up of about 22 000 middle-aged Chinese males in China (Province of Taiwan) among whom 157o were HBsAg carriers, a 223-fold excess risk for hepatocellular carcinoma among the carriers (compared with Chinese males with no evidence of being carriers) has been demonstrated.' This study clearly showed that a e BEASLEY, R. P. ET AL. Hepatocellular carcinoma and hepatitis B virus. A prospective study of 22 707 men in Taiwan. Lancet, 2: 1129-1133 (1981). PREVENTION OF HEPATOCELLULAR CARCINOMA 733 high risk for this carcinoma was associated with the HBsAg carrier state and not with other HBV markers. Mother-to-infant transmission and establishment of the HBsAg carrier state Mothers carrying HBV are sometimes highly infectious and in many parts of the world they transmit the infection to their newborn babies. Infection of infants is especially important because a proportion of these infants will become carriers. The number of carriers is inversely related to the age at infection, ranging from approximately 90% in newborns to 10% or less in adults. Infectivity is directly related to the presence of HBeAg in the mother's serum. When mothers are HBeAg-positive, approximately 95% of their new- born children are infected, usually in the perinatal period. The prevalence of HBeAg among HBsAg carriers, and thus the infectivity of mothers for their infants, varies markedly in different circumstances and geographical areas. In Asia, 30-50% of HBsAg- carrier women of childbearing age are HBeAg-positive, and perinatal infections may account for about half of the carriers in the population. Children of non-carrier mothers can be infected by contact with other children who have been infected by their carrier mothers. Thus, perinatal transmission appears to be the driving force in the maintenance of the high HBsAg carrier rates, among people from eastern Asia for example. The same rates prevail among the Chinese living outside China. In contrast, HbeAg positivity and perinatal transmission are uncommon in Caucasian populations and perinatal transmission is of intermediate frequency in mothers of west- Asian origin or of Afro-Caribbean origin. In Africa, mother-to-infant transmission is also important, but because HBeAg in carrier mothers is less frequent than in Asia, the infec- tion of infants occurs most commonly during early childhood. The timing and mechanism of transmission are of importance in developing strategies for immunization. It is important to note that mother-to-infant transmission during pregnancy, as opposed to the perinatal period, is rare. The few intrauterine infections that occur are probably due to occasional leakage of maternal blood into the fetal circulation, because it is known that the virus does not cross the intact placental barrier. Other etiological factors Several factors that may cause hepatocellular carcinoma, independently or in association with HBV infection, have been reported. These include other viruses that can cause hepatitis in man, but there is no evidence that they are involved in human liver cancer. There is a great deal of experimental evidence that aflatoxins, which are naturally occur- ring chemicals of fungal origin, are powerful liver carcinogens. It has been shown that people in some areas of the world are frequently exposed to food contaminated with afla- toxin and that a correlation exists between the level of such contamination and the appear- ance of hepatocellular carcinomas in those areas of the world where the incidence of these cancers is high. Evaluation of the carcinogenic risk of the aflatoxins for man has been reported,b and the laboratory and field evidence of the health hazards of the aflatoxins has been examined in detail and guidelines for health protection have been recommended.c No other mycotoxins have been incriminated in human liver cancer. Other chemicals that have been considered as liver carcinogens include nitrosamines, cycasin, safrole, tannic acid, tannins, and several pyrrolizidine alkaloids. The synthetic chlorinated hydrocarbons including the organochlorine pesticides (such as DDT) and the polychlorinated biphenyls have also been considered; they can cause liver cancer in experi- mental animals, but there is no evidence of hepatocarcinogenicity in man. bEvaluation of carcinogenic risk of chemicals to man. Lyon, International Agency for Research on Cancer, 1976 (IARC Monograph Series, Vol, 10). c Mycotoxins. Geneva, World Health Organization, 1979 (Environmental Health Criteria, Vol. 11). 734 A WHO MEETING Androgenic-anabolic steroids are associated with liver angiosarcomas, and steroid contraceptives with benign liver adenomas; there are a few reports that these are occasion- ally involved in the development of hepatocellular carcinoma. The pathology of hepatocellular carcinoma, including the role of alcohol, trace elements, and genetic factors, and the part played by cirrhosis in the etiology of this cancer have recently been reviewed.d EARLY DETECTION OF HEPATOCELLULAR CARCINOMA Hepatocellular carcinoma is a rapidly fatal disease but the biological behaviour of these tumours in Africa, in the Indian subcontinent, and other parts of Asia appears different. In Africa, these tumours grow rapidly and occur in early adult life. However, in Asia, some of these tumours grow more slowly. Rising and consistently high levels of alpha-fetoprotein (AFP) in the serum indicate underlying liver cancer in asymptomatic cases. Preliminary results from China and Japan, following resection of these asymptomatic tumours, are promising. However, before such screening is adopted widely in asymptomatic popu- lations, the value of the tests should be assessed by randomized trials in populations where the appropriate surgical treatment is readily available. In areas of high incidence of hepatocellular carcinoma, it may be of value to screen regularly, by measurement of serum AFP, persistent hepatitis B carriers in the older age groups. Family members of liver cancer patients, in whom HBsAg has been detected, could also be screened by AFP estimations. MOLECULAR BIOLOGY OF HEPATITIS B VIRUS A summary of research on the structure, genetic organization, and biological properties of hepatitis B virus has been published recently.e One of the most striking recent advances has been the finding of integrated HBV DNA in patients with chronic hepatitis and hepato- cellular carcinoma. This integration makes the elimination of HBV DNA in chronic carriers impossible but prevention may be attained by protection against primary infection. Integration of HBV DNA into human hepatocytes was first detected in a continuous cell line expressing HBsAg. The cell line was derived from a male HBV carrier with hepato- cellular carcinoma. In HBV carriers with or without histological evidence of liver disease, integration of HBV DNA may be diffuse in many sites or present in unique sites of the host genome. Most of these patients have HBsAg and anti-HBe and it has been proposed that continued HBsAg expression in these patients may be derived from integrated HBV DNA. However, some patients may have HBV DNA in their liver with no expression of HBsAg (latent viral infection), while others may express HBsAg or virus (persistent viral infection). All published studies are consistent with the interpretation that integration ofHBV DNA into the hepatocyte genome precedes the development of hepatocellular carcinoma by months or years. The exact time at which integration occurs, the relationship between integration and HBV serum markers and the frequency with which integration leads to the development of hepatocellular carcinoma require further study. From the epidemiological d MACKAY, 1. & OKUDA, K., ed. UICC Workshop on Hepatocellular Carcinoma (UICC Technical Report Series, No. 74) (in press). 'DEINHARDT, F. & GUST, I. D. Viral hepatitis. Bulletin of the World Health Organization, 60: 661-691 (1982). PREVENTION OF HEPATOCELLULAR CARCINOMA 735 evidence, it would appear that the HBV carrier state precedes by many years the develop- ment of liver cancer. Although these studies do not prove that HBV is oncogenic, the find- ing of integrated HBV DNA in many hepatocellular carcinomas, and in all such cancers of patients with HBV markers, is highly suggestive. This is true in spite of the fact that integrated HBV DNA may be found also in non-malignant liver cells during persistent infection. HEPATITIS B-LIKE VIRUSES IN ANIMALS Detailed consideration of hepatitis B-like viruses in animals will be of outstanding value in the study of the molecular biology of hepatitis B. These viruses share similar morpho- logical and biological properties and genomic organization with the human hepatitis B virus. They have been identified in eastern woodchucks (Marmota monax), ground squirrels (Spermophilus beecheyi), and certain ducks (Anas domesticus)! Some cross- reactivity exists between the viral proteins of different species, and the genomes show the same unique structure as in HBV. A viral DNA polymerase has also been identified in all three animal species. The surface and core antigen genes in the ground squirrel virus are in the same location and orientation as in the human virus, but there is little sequence homology over long stretches of DNA between woodchuck, ground squirrel, and duck viruses and human HBV DNA. These viruses do not infect other species, including chim- panzees or man. Although there is no histological evidence of liver disease caused by these viruses in ground squirrels and ducks, acute and chronic inflammatory changes as well as hepato- cellular carcinoma have been found frequently in woodchucks. In all but one case of such cancer in carrier woodchucks, the viral DNA was also integrated into the host cellular DNA at a number of discrete sites. In chimpanzees infected with HBV, both chronic persistent and chronic active hepatitis have been described. Why chimpanzee carriers of HBV do not develop hepatocellular carcinoma is not clear. Whether this simply reflects too short a period of time for these carriers to develop this cancer, a lack of unique HBV DNA integration, or other factors is also not clear. Recently a full-length hepatitis virus RNA transcript was identified during the repli- cation cycle of the duck hepatitis virus (DHBV). It has been proposed that this transcript serves as an intermediate for replication of DHBV via a reverse transcriptase-like reaction. Therefore, hepatitis B-like viruses may have some properties similar to the retroviruses (RNA tumour viruses). This is an important new development which may explain how integration of HBV DNA occurs. Proviral DNA ofRNA tumour viruses integrates into the host chromosomal DNA as part of its normal replication cycle. Whether this reverse tran- scriptase activity is the same as or different from that observed in the human HBV remains unclear. HEPATITIS B VACCINES Efforts to grow hepatitis B virus in tissue culture have been unsuccessful and so far the only vaccines in use are derived from human plasma from carriers of the hepatitis B surface marker. f SUMMERS, J. & MASON, W. S. Properties of the hepatitis B liver viruses related to taxonomic classification. Hepatology, 2: 615-665 (1982). 736 A WHO MEETING Standardization and control The requirements for the production and control of hepatitis B vaccines derived from human plasma, which are of paramount importance for international progress towards the control of hepatitis B virus infection and related diseases, were formulated by a WHO Expert Committee in 1980.g Since that time, considerable additional experience has been gained in the production of this vaccine and a large number of doses have been ad- ministered. Consequently, there is now an urgent need for a revision of the present WHO requirements for hepatitis B vaccines derived from human plasma. Work is in progress in several countries on other types of hepatitis B vaccine (see below). These developments will call for the formulation of separate requirements for their manufacture and control. The use of products derived from human plasma and biologically complex material should be regarded as potentially dangerous and all possible measures should be taken to ensure the safety of vaccines using such source materials. A major problem in respect of vaccine control is that the chimpanzee remains the only available animal model for the detection of infectious hepatitis B virus. This primate is in limited supply and its use for routine vaccine control is likely to become increasingly restricted. The establishment of the consistency of production of the vaccine obtained by a given manufacturing process might be based on testing of the early consecutive batches of vaccine in susceptible human volunteers. However, whenever a new virus inactivation pro- cess is employed, it is important that, where possible, the efficacy of the process to inacti- vate hepatitis B virus should be examined in limited studies using chimpanzees. Chim- panzees may also be required in small numbers to establish the protective efficacy of the new hepatitis B vaccines. There is a need for international reference preparations of hepatitis B vaccine for use in the control of vaccine potency and purity in vivo and in vitro. Vaccines for the future Consideration must be given to the preparation of vaccines from sources other than human carriers with the object of improving safety and reducing cost. The search for and the development of these alternative sources of safe and potent vaccines should be encouraged. Development work is in progress in several countries on new hepatitis vaccines including hepatitis B polypeptide micelles. Another approach is to use a virus as a vector for DNA coding for the immunogen. Several research groups have replaced the early or late genes of SV40 with DNA coding for another protein. This has been done for HBsAg and reasonable yields obtained. Similarly the hepatocarcinoma cell line, derived from a hepatitis B carrier, produces HBsAg. However, because of the association with malignant tissue, this approach may not be acceptable. Cloning of the hepatitis B surface antigen gene in prokaryotes or in eukaryotes has resulted in the expression of surface antigen. This method is promising, provided that the necessary yields and purity of antigen can be obtained at a reasonable cost. Hydrophilic oligopeptides, representing an antigenic deter- minant of hepatitis B surface antigen, can be synthesized chemically in the laboratory but are of low potency, and a suitable carrier and/or adjuvant will need to be found if such synthetic material is to provide a practical basis for the production of vaccine; these are still only theoretical possibilities and are unlikely to be available for another 5-6 years. I Requirements for hepatitis B vaccine. In: WHO Technical Report Series, No. 658, 1981, Annex 4, pp. 131-156. PREVENTION OF HEPATOCELLULAR CARCINOMA CURRENT STATUS OF IMMUNIZATION STUDIES FOR PREVENTION OF PERINATAL TRANSMISSION Studies of hepatitis B immunoglobulin (HBIG) administered to newborns have been conducted, and without exception were successful in reducing the frequency of HBsAg carriage resulting from perinatal infection. The administration of HBIG immediately after birth to ensure maximum effectiveness is of considerable importance. Several additional doses given during the first year of life have also been shown to increase effectiveness. A reduction of the carrier state of between 4007o and 607o was achieved in several large studies. Although many infants receiving HBIG developed immunity, a significant propor- tion were nevertheless vulnerable to infection, once the passively acquired antibodies were no longer present. Many of these infants subsequently became infected as a result of continued exposure to their carrier Table 1. Studies of the immune response to HBV mothers, especially when the mothers had vaccine in newborn childrena HBeAg. Immunogenicity studies (Table 1) have HBV Anti-HBs + by demonstrated that more than 900%o of new- Placeb vaccine No. of 6 months borns develop antibodies by six months of source infants age in response to two doses of hepatitis B No. % vaccine. These findings are unusual, since Burundi Pasteur 66 62 94 the response of newborn infants to other China vaccines is generally poorer than that of (Province Merck 31 29 94 older infants and children. It should be of Taiwan) noted that passively acquired antibodies Senegal Pasteur 115 108 94 do not interfere with the response to South Africa Merck 73 67 92 immunization. a Excludes infants of HBeAg carrier mothers but includes Studies are now in progress in at least infants whose mothers are anti-e carriers, and infants who nineounties t evauatethe immuno- had passively acquired anti-HBs either from the mother ornine countries to evaluate th from administered HBIG. genicity and efficacy of HBV vaccine b Other studies in newborn children are in progress in alone or in combination with HBIG. Data several countries including China and the United States, but results were not available at the time of the preparation offrom most of these studies have not yet this report. been published, but the results from many of them are expected during the next two years. They involve basically two approaches: (1) use of HBV vaccine only, and (2) use of HBIG plus HBV vaccine. Most investigators expect optimal protection from the latter and numerous dose-time combinations are being evaluated (Table 2). A 90-99Vo protection against the carrier state was demonstrated in a study of 231 infants in Japan, who had been given HBIG at birth and (together with HBV vaccine) on three further occasions, the infants being followed up for at least 12 months. Studies in China (Province of Taiwan) have shown protection of more than 95%o of infants who received both HBIG and vaccine, and approximately 7507o of those given only the vaccine at the age of one week. GLOBAL STRATEGIES FOR THE PREVENTION OF HEPATITIS B AND LIVER CANCER USING IMMUNIZATION The only practical method of achieving widespread effective control of hepatitis B is by active immunization. Immunization strategies must take into consideration geographical 737 A WHO MEETING Table 2. Ongoing studies of immunization against perinatal transmission of HBV to infants of e-positive mothers Immunization schedules being compareda HBV Place vaccine Controls HBIG Vaccine HBIG Approx. Year when source only only plus number of results are vaccine infants ' expected Burma Merck Yes No Yes No 160 1984 China: Qidong Pasteur Yes Yes No Yes b 100 1983 Province Merck Yes H Yes Yesb 275 1983 of Taiwan Province Pasteur Yes No Yes Yes b 150 1983of Taiwan Federal Republic Merck H No No Yes 25 - of Germany Hong Kong Dutch Red Cross Yes No Yes Yes 120 1984 Italy Merck ? No No Yes 13 1984 Japan Kitasato H H No Yes 384 P Japan Kitasato H H No Yes 500 1983 Japan GrXeen Cross H H No Yes 350 1983 Thailand Merck ? No Yes No 20 - Sweden Merck ? No No Yes 2 1984 USA Merck H Yes No Yesb 124 1983 a H = Historical controls, i.e., data on controls obtained from previous studies. available. b More than one schedule being evaluated. ' Number enrolled up to January 1983. P = Published. ?, - = no information patterns in the prevalence of hepatitis B. The prevalence may be conveniently divided into three categories. In low endemic areas, such as North America, western Europe, and Australia, prevalence of HBsAg in asymptomatic carriers is 0.2-0.57o, while in areas such as eastern Europe, the Mediterranean region, and south-west Asia, the prevalence varies from 2% to 7%0. In high-prevalence areas, such as China, south-east Asia, and tropical Africa, rates of HBsAg carriage may be as high as 15%. Evidence of infection by HBV, as measured by the presence of anti-HBs in serum, shows a similar geographical distribution: 4-6%, 20-55%Mo, and 70-950o, respectively. In countries where hepatitis B is uncommon, infection rarely occurs before adult life. As the overall frequency of infection in the popu- lation increases, there is increasing tendency for the infection to occur at an earlier age, and in the areas of highest prevalence the infection occurs mostly in infancy and early child- hood. Within each country or geographical region, considerable differences in prevalence may exist between different ethnic and socioeconomic groups. Strategies for immunization are shown in Table 3. In areas of low endemicity, targeted immunization will probably be considered and, unless it is deemed essential to identify HBsAg-positive mothers and immunize their infants, the need for immunization products will be small. The approach for areas of intermediate and high endemicity could be similar, as the decision to embark on immunization by a national authority will be influenced by the logistics of health care delivery and the cost of immunization products as well as the preva- lences of hepatitis, the HBsAg carrier state, and hepatocellular carcinoma. The decision to immunize before or after exposure in these areas will depend on the results of the studies 738 PREVENTION OF HEPATOCELLULAR CARCINOMA outlined in the previous section, which will indicate whether or not vaccine alone is effective in preventing the infection and the carrier state of infants. Other factors to take into consideration are the cost and practical problems in carrying out wide- spread immunization of all infants, com- pared with the demands on the health care systems in the identification of HBsAg- positive mothers and the subsequent immunization of their children. The former will require large quantities of vaccine, and the latter a countrywide ante- natal and laboratory service, a sufficient supply of HBV diagnostic substances, and possibly also HBIG. An examination of national immunization schedules to deter- mine whether HBV vaccine can be given together with other vaccines, such as DTP or poliomyelitis vaccine, would also be needed. If HBV vaccine is to be ad- ministered as soon as possible after birth, its inclusion in neonatal care must be considered. Production of hepatitis B vaccine from donor plasma could form part of a coordi- nated production of plasma-derived prod- ucts, all of which are important for the prevention and control of hepatitis B (Table 4). With few exceptions, the purifi- cation processes used in the preparation of HBsAg and for the production of diag- nostic reagents and vaccine are similar. The transfer of technology to developing countries for the production of vaccines and reagents should be a priority. Table 3. Target groups for hepatitis B immunization Intermediate and high- Low-endemicity countries endemicity countries Pre-exposure Post-exposure Pre-exposure Post-exposure (High-risk Persons All infants Infants of groups) with acci- HBsAG- dental per- positive Health cutaneous mothers personnel exposure Dialysis Infants of patients HBsAg- positive Institution- mothers alized patients Sexual contacts of Drug acute addicts cases Male homo- sexuals Table 4. The use of blood products for the control of hepatitis B Therapeutic Diagnostic Products for products reagents immunization Blood derivatives HBsAG Hepatitis B immunoglobulin anti-HBs Hepatitis B vaccine anti-HBc HBeAg anti-HBe PROPOSALS FOR FUTURE STUDIES An evaluation of different methods of immunization is needed. Field trials should cover neonates and adults who are at risk of developing the hepatitis B carrier state. Comparative field studies to assess the effectiveness of immunization There is a need for field studies using HBV vaccines, involving several hundred subjects, before starting on large-scale immunization. These field studies are required in order to determine the optimal use of the vaccine, including timing of injections, and the value of HBIG. Since relatively small differences (85070 as opposed to 95%o) in the rates of effective- 739 A WHO MEETING ness of any two vaccine schedules will be of substantial practical importance for large- scale immunization programmes, it is vital that field studies should be designed to detect such differences. Thus, the following are essential: (1) Participants in comparative field trials should be randomized to ensure that the results are unbiased. If a double-blind randomization is practicable, this should be used. (2) Sample sizes should be large enough to detect with high probability any small differ- ences. For example, to detect an 85/o versus 95% difference in rates of effectiveness with high probability (0.90) using a 2-sided test, 175 subjects are needed in each treatment group. In studies with smaller numbers of subjects, differences of this magnitude could easily be undetected. Neonatal immunization Field trials should be initiated in populations in one or more developing countries where the prevalences of HBV infection, the HBV carrier state, and hepatocellular carcinoma are known to be high. Since the epidemiology of both this cancer and HBV infection show striking differences between high-risk populations in Africa, south-east Asia, and eastern Asia, it is necessary to initiate trials in at least two such areas. The trials should be designed to be capable of evaluating the effect of immunization on the risk of this cancer appearing many years in the future. Other factors influencing the choice of the site for such a trial would be the availability of medical services, background epidemiological data, and a commitment by the government of the host country to provide adequate support to ensure that the main aims of the project could be achieved. Trials could be initiated with vaccines prepared, tested for safety, and standardized according to the WHO requirements. It is anticipated that within two or three years after commencement of a trial, locally or regionally produced vaccines of the required standard should be available. Immunization of babies should be carried out as soon as possible after birth. During the trial, serological tests should be carried out: (1) If anti-HBs only is detected within the first 7 months after immunization, it will be concluded that immunity has been induced and can be expected to last up to 5 years. (2) If both anti-HBs and anti-HBc are detected at any time after immunization, it will be concluded that successful immunization has occurred followed by subclinical infection. Immunity would probably be lifelong. (3) If anti-HBc and HBsAg (with or without HBeAg) are detected, this indicates overt HBV infection and hence an immunization failure. (4) The presence of anti-HBc and the detection of HBsAg for more than six months, with or without HBeAg, would indicate not only immunization failure but development of the carrier state. Successful monitoring of immunization in this way would allow, within five years of the initiation of the project, a decision on whether prevention of infection and of the carrier state had occurred, the extent to which subclinical infection had occurred, and the duration of vaccine-induced immunity. On the basis of the results of the previously described feasibility studies (see page 737), it would be expected that, provided favourable conditions were present, immunization could induce a reduction of up to 95% of the naturally acquired infection. At the same time, the incidence of the carrier state would be reduced by the same extent. To be successful, field trials in developing countries will need considerable external fund- ing. Possible sources for such funds include private foundations, international agencies, 740 PREVENTION OF HEPATOCELLULAR CARCINOMA cancer agencies, and bilateral agreements between WHO Member States. A donor country in particular will need to know that the recipient country fully supports the field trial proposal and will provide the necessary national support, at least for long enough to see whether the trial has been successful or not. Adult immunization There is evidence that most HBsAg carriers acquire the condition as a result of early infection. However, there is no evidence that the carrier state arising from an (albeit more rare) late infection, perhaps in adult life, is less likely to progress to hepatocellular carci- noma. Therefore, if a group of male adults could be identified at the peak age for this cancer risk, who are negative for HBV markers and are nevertheless at a sufficiently high risk for this infection, then this group may be suitable for a randomized trial to assess quantitatively the effectiveness of the vaccine in preventing this cancer. In any group pre- liminarily identified for this purpose, it will be important to determine the rate at which hepatitis-negative adults become antigen carriers. This determination must be part of the initial feasibility studies. Only if this rate is sufficiently high would an immunization trial of reasonable size have adequate statistical power. Large-scale studies to evaluate the effectiveness of immunization in preventing hepato- cellular carcinoma Before mass immunization campaigns are implemented, it is desirable to undertake one or more large-scale trials, each of which would probably involve tens of thousands of indi- viduals. It would also be desirable to conduct similar studies in different areas of the world, because the long-term effectiveness of immunization may vary with factors affecting HBV transmission. These studies will also be useful for assessing the logistic problems associated with the establishment of mass immunization campaigns and their evaluation in different parts of the world. Design of immunization campaigns There are good grounds to suppose that immunization of children with hepatitis B vaccine at birth, or shortly thereafter, will confer long-term protection against the develop- ment of hepatocellular carcinoma. Nevertheless, it would be undesirable to consider immunization on a widespread scale without, at the same time, organizing studies designed to establish whether or not the vaccine produces a reduction in the incidence of the carrier state and subsequently cancer. Although the main health benefit expected from large-scale immunization programmes in infants might be a reduction in mortality from chronic liver disease and hepatocellular carcinoma, this is unlikely to occur for some years to come. Earlier endpoints, such as carrier rates after 5 and 10 years, should therefore be incorporated into the study design. The best method to ensure that a protective effect of immunization could be measured would be to use randomized controlled trials. In this way, it would be possible to ensure that the groups being compared, both immu- nized and unimmunized, would be similar with respect to all possible risk factors for this cancer, except for immunization. Such a randomized trial would present ethical problems if sufficient vaccine were available to treat all newborn children. At present, and probably for the next several years, however, this is not the position as the vaccine is expensive and supplies are limited. 741 A WHO MEETING Randomization In large-scale immunization trials, randomization of individuals will often be imprac- ticable, and it would be better to use randomization at the community, village or hospital level, provided that there are a reasonable number of "communities" (at least 10) in the programme. For example, suppose an infant immunization programme is being planned for a region consisting of twelve communities but that, in the first year of the programme, there is only enough vaccine for six of the communities. Then six of the twelve communities can be randomly selected to receive the vaccine in the first year, and the newborn from the other six communities in that year can serve as controls. During the course of the study, the vaccine may become much cheaper and more widely available. In these circumstances, ethical considerations may dictate that all infants should be given the vaccine. This may not jeopardize the interpretation of the study, since many persons would already have become carriers as infants or young children by the time of general immunization campaigns and would remain at a high risk for hepatocellular carci- noma, and thus serve as controls. In areas where the vaccine is not allocated on a random basis, it is recommended that careful records should be maintained of all those who are immunized so that an eventual assessment may be made of the effectiveness of the vaccine against this cancer. Identification and follow-up of subjects It is vitally important that, for the satisfactory interpretation of the results, individuals should be correctly identified as having been originally in an immunized or unimmunized group. The logistic problems to ensure this can be considerable, especially in some develop- ing countries. In some societies, children are not named until they have reached a specific age or their names are changed at that time. Date of birth is often uncertain, and it may be necessary to record additional information, such as the names of siblings and their birth order, to be sure of subsequent identification. For the trial to be evaluated adequately, it is important to be able to identify members of the unimmunized group, as distinct from non- participants, and it will certainly not be sufficient to identify only those individuals who were given the vaccine. Even if the participants in the study are identified in a satisfactory manner at birth, it may be necessary to conduct a census of the study population at regular intervals, and to institute some method of continuous monitoring, in order to detect any possible side- effects of the vaccine and other relevant disease events, and possibly also to reimmunize individuals whose immunity has waned. Migration is common in some places and this may deplete the study population at a rather crucial period. It will probably not be feasible to follow up individuals once they have left the study area, and so the possibility of migration will need to be taken into account in the planning and evaluation of the study. Associated studies The logistics and cost of setting up a long-term hepatitis B vaccine trial may be formidable. If such a trial is started, however, it will provide excellent opportunities for a wide range of associated investigations aimed at understanding the natural history of HBV infection and associated chronic liver disease. It seems likely that the worldwide distri- bution of hepatocellular carcinoma cannot be explained solely in terms of the prevalence of HBV infection and that there may be cofactors playing an important part in the etiology of 742 PREVENTION OF HEPATOCELLULAR CARCINOMA this cancer. Exposure to aflatoxin is a strong candidate for such a role, and this might be studied on a longitudinal basis during a vaccine trial. RESUME PREVENTION DU CARCINOME HEPATOCELLULAIRE PAR LA VACCINATION Le cancer du foie est l'un des cancers les plus communs dans le monde et un des plus fre- quents dans les pays en developpement. Les preuves de l'intervention du virus del'hepatite B (HBV) dansl'etiologie du cancer primitif du foie sont fondees sur l'observationepide- miologique et geographique d'une forte association entre l'infection parl'hepatite B et cette forme de cancer, de meme que sur des resultats recents d'etudes de biologie molecu- laire qui ont revele l'integration de I'ADN viral dans le genome del'hote. Ces preuves sont maintenant suffisamment solides pour justifier l'utilisation d'un vaccin contre cette infec- tion, comme moyen de prevenir le cancer primitif du foie. Il a estime qu'environ 80Oo des cancers du foie sont attribuables ce virus. Parmi les cancerogenes humains connus,1'HBV n'est donc depasse que par le tabac. De l'immunoglobuline specifique de l'hepatite B (HBIG) a administree des nouveau-nes et on a observe une reduction de 75% 90% del'etat de porteur dans plusieurs etudes de grande envergure. Des enquetes sur l'immunogenicite ont prouve que plus de 90% des nouveau-nes presentent des anticorpsa l' age de six mois enreponse deux doses de vaccin contrel'hepatite B. On dispose de vaccins efficaces qui ont eprouves dans des etudes de faisabilite et leur utilisation dans des essais pratiques, en vue de controler leur efficacite contre le risque A long terme de developper ce cancer, est mainte- nant possible. Al'heure actuelle, il n'existe que des quantites limitees de vaccin obtenu partir de plasma humain infecte par le virus del'hepatite B.I1 faut examiner les moyens de preparer des vaccins partir de sources autres que les porteurs humains afin d'accroitre lasecurite et dereduire le couft. Les strategies vaccinales doivent prendre en consideration la distribution geographique del'hepatite B. Dans les regions de faible endemicite, la vaccination doit viser les groupes haut risque tels que le personnel de sante, les malades sous dialyse, les drogues et les personnes ayant subi une exposition percutanee accidentelle, par exemple les nourrissons de meres porteuses du virus. Dans les zonesoful'endemicite est intermediaire ouelevee, la decision de proceder la vaccinationdependra de la logistique de la distribution des soins de sante et du couft des vaccins, aussi bien que de la frequence del'hepatite, del'etat de porteur et du carcinomehepatocellulaire. Pourdecider si l'on vaccinera tous les nour- rissons ou seulement les enfants de meres qui sont porteuses, on se fondera sur lesresultats d'etudes effectuees sur le terrain et sur des considerations pratiques telles que le cofut, les exigences auxquelles le systeme de soins de sante doit faire face, et la possibilite d'admi- nistrer du vaccin anti-HBV enmeme temps que d'autres. Des mesures internationales doivent etre prises en vue de planifier et de lancer un certain nombre d'essais pratiques dans des populations parmi lesquelles on sait que l'infection par le virus del'hepatite B,l'etat du porteur et le cancer du foie ont desfrequenceselevees. faut entreprendre deux typesd'etudes: des etudes comparatives sur le terrain pour deter- miner l'utilisation optimale du vaccin, notamment la chronologie des injections et la valeur de1 'HBIG, et des etudes de grande envergure en vued'evaluerl'efficacite de la vaccination pour la prevention du carcinomehepatocellulaire. Si le principal avantage pour la sante 743 744 A WHO MEETING qu'on peut attendre de programmes de vaccination a grande echelle chez les nourrissons etait une reduction de la mortalite par maladie chronique du foie et par carcinome hepato- cellulaire, 1'evaluation exigerait la surveillance des sujets pendant un certain nombre d'annees; mais il faut egalement envisager, dans la conception des etudes pratiques, des objectifs 'a court et 'a moyen terme.