Letters to the Editor
Indian Pediatrics 2001; 38: 1318-1322
Hepatitis B in India: A Review of Disease Epidemiology
Their key messages were that the prevalence of chronic HBV infection in India may be below 2% and that currently available data are insufficient to objectively evaluate the utility of the vaccination program(1). We feel that neither message is justified from their review. Moreover, the review itself was selective, incomplete and flawed; hence the conclusions are neither valid nor reliable. The paper is likely to mislead policy makers and pediatricians, about the usefulness of and need for HBV immunization. We reiterate the urgent need to include HBV vaccine in India’s Universal Immunization Program (UIP), and unequivocally endorse the recommendations to this effect by the World Health Organization (WHO), the Indian Academy of Pediatrics (IAP) and the Indian National Association for the Study of Liver Diseases (INASL).
The Purpose of HBV Immunization
We question the statement that "the goal of immunization program against hepatitis B is to reduce the incidence of, and possibly eliminate hepatocellular carcinoma and chronic liver disease by reducing the number of HBV carriers in the population". Immunization will not prevent the develop-ment of chronic liver disease (including carcinoma) in those who are already chronically infected, nor will it cure chronic infection in order to reduce its prevalence in the community. The primary aim of immunization is to protect the individual from infection by HBV, so that acute hepatitis, fulminant hepatitis, chronic infection and its sequelae will be prevented. As the risk of infection may begin during infancy itself, immunization must also start in early infancy. Although infection is mostly asymptomatic in infancy and childhood, the risk of chronic infection is highest in infancy. The purpose of including HBV vaccine in UIP is to give all infants the benefit of protection. Thus, there is no fundamental difference between HBV immunization and other childhood immunizations.
It is due to the undue emphasis given by Lodha and colleagues to the prevention of chronic liver disease that they tend to rely upon the rate of prevalence of HBV carrier state as the sole guiding criterion for recommending immunization. Thus, they suggest that India should be reclassified as a ‘low prevalence’ (<2%) country from the current classification as ‘intermediate level’ (2–7%), thereby suggesting that the urgency or need to include HBV immunization in UIP would lessen or disappear. On the other hand, if immunization is to prevent infection, then what we need to understand is the lifetime risk of infection with HBV, as well as the age–dependent frequency of infection. The rate of carrier state is dependent on the incidence of infection and particularly on its age distribution. If the incidence is high but the age of infection is in adults, then acute hepatitis and fulminant hepatitis will be more frequent, but the rate of chronic infection would be lower than if the infection occurred in infancy and childhood. If infection occurs frequently in infancy and childhood, then acute disease will be less frequent, but carrier state and chronic liver disease will be more frequent. Therefore, the prevalence rate of carrier state is only one factor in the decision making process for immunization, but not the sole factor. The review did not attempt to examine available data to quantify the lifetime risk of HBV infection and the age distribution. Hence the review does not help in assessing the magnitude of the problem, which could be prevented by immunization.
What is actually different for HBV infection from some other vaccine preventable infections is the wide age range when the first experience with HBV may take place, the wide age range when clinical disease may appear and the wide range of clinical disease itself. While many vaccine-preventable infections are generally associated with childhood, primary tuber-culosis, Bordetella pertussis infection, diphtheria, tetanus and measles may occur during childhood, during adolescence or even later. Childhood immunizations (and boosters in some cases) will give protection over wide age range; this is also true of HBV immunization. Tetanus is particularly relevant here. Infant immunization is not because of high incidence of tetanus in infancy or early childhood, but on account of the fact that the opportunity for lifelong prevention presents itself in infancy. Infant immunization does not prevent neonatal tetanus, for which another strategy is applied, namely maternal immunization. In countries where all mothers are fully immunized, there is no further need to immunize them during pregnancy. On similar lines, giving HBV vaccine to infants will only prevent vertical transmission some 20 or more years hence. During this interval, infants of carrier mothers need to be immunized immediately after birth.
Only the immunized individuals will be protected from HBV infection due to immunity. When many are prevented from infection, the transmission chain will slow down, and some of the un-immunized will get the benefit of reduced exposure to infection. This is not the primary goal of an immunization program but a spin-off benefit. It is called the ‘herd effect’ of immunization.
If we desire the early reduction of incidence of infection in age groups beyond infancy or childhood, then ‘catch up’ immunization will be needed at the desired ages. Italy introduced immunization at 12 years of age and in infancy. In 12 years, all below 24 years of age will be protected and immunization at age 12 will be stopped. In India, the desire for catch up immunization is exemplified by the unexpected popularity of HBV immunization camps and campaigns for a wide age range. This is a case of ‘let the people lead and the leaders may follow’. We ‘experts’ may continue to debate, but the people seem to know what they want.
The True Prevalence of HBV Carriers in India
The relatively restricted review of Lodha and colleagues and the earlier more thorough review by INASL, of the prevalence of chronic HBV infection, reveal two pieces of undisputed information(2). HBV infection is very widespread in the country and there are wide variations in the prevalence rates according to socioeconomic levels (e.g., voluntary versus replacement versus professional blood donors), geography (southern India versus northern) and ethnic groups (tribal versus others). Obviously, the rates vary with age and gender also. There-fore, the ‘true prevalence’ can be ascertained only if the entire population is tested for chronic infection. What is practical is to estimate the average prevalence, using all available data, and not restricted on artificial criteria as done by the reviewers. The prevalence estimation by the reviewers is more flawed than those of others in the past. Indeed, the reviewers themselves rightly presented the pitfalls of such an exercise: "for the assessment of prevalence ... the sample selected should be representative of the whole population and should be adequate. There were no reports which had the study subjects who were representative of the Indian population. Although blood donors and ante-natal women provide convenient samples, the results from these groups cannot be extra-polated to general population." Yet, in spite of this insight, they ignored their own logic and went on to assess prevalence on a selected small sample of the available studies. While it is true that they described their method of selection and analysis, accurate description does not represent adequacy for reliability and validity.
The review included one study on blood donors, from Vellore. The prevalence of HBsAg among voluntary donors was 0.7%. We had established routine screening of all donor blood for HBsAg from 1973(3). Voluntary donors found to be positive once would not come again for testing. Therefore, only the first study from a center is likely to be representative of the group, not subsequent ones. Thus, the 0.7% has no validity for inclusion for assessing prevalence rate in the subset of general population represented by the voluntary donors. Replacement donors, on the other hand, are constantly changing and replenished, and are more likely to be representative of the adult able-bodied segment of patient-relatives, and likely to be undergoing the test for the first time. However, those already known to be positive would have self-selected for exclusion. This illustrates the need for greater understanding of the sampling errors in the various studies if we are to refine the estimated prevalence from what is previously published.
In February 2000, Indian Pediatrics published an excellent paper on ‘community studies on prevalence of HBsAg’ among children(4). Each sample was tested twice(4). While Lodha and colleagues complained about the lack of such studies, it was inexcusably excluded from their review. In Rajamundry the prevalence was 3.3% and in Bangalore 4.2%(4). The extensive review by INASL had arrived at a consensus figure of 4.7% as the national average for carrier state(2). Presenting their new estimate of <2% without confronting the earlier and widely quoted estimate is unfair both to the readers and the earlier reviewers.
The Deficiencies in the Method of Analysis
We have described above the deficiencies of the review process itself. The analysis and conclusions are also open to question. The spread of prevalence data cited by them included two studies with < 1%, 6 studies with 2–2.6%, 3 with 3.1– 3.8%, 3 studies with 4.4–4.9%, and one each with 7% and 15.7%. The authors picked the number 3 as the prevalence, but it has no scientific basis. They had already stated that voluntary blood donors or pregnant women would not be representative of the population. The social stratum represented by voluntary donors would be a minority in the country and women are less likely than men to become chronic carriers. Therefore, if the prevalence rates of these two groups are to be used for extrapolation and projection, then both deserve upward correction.
The authors used another assumption for the analysis, namely that the sensitivity and specificity of the tests would be 98%. These indices were suggested by an expert panel for the purpose of assessing the cost-benefit of different tactics of HBV vaccination(5). In the real world of laboratory testing the sensitivity may be lower, missing persons with very low levels of antigenemia. Post-transfusion hepatitis B does rarely occur in persons transfused with blood found to be negative for HBsAg. The specificity index may indeed be higher since the analyte is a foreign antigen, unlike in the case of tests for an antibody, which may show cross-reactions with antibodies of other antigens. If sensitivity is lower and specificity higher than 98%, then the derived prevalence will be an under-estimate of the true prevalence.
The Utility of HBV Immunization Program
The reviewers imply that their focus was on those aspects of epidemiology which are essential for decision-making about the use of HBV vaccine in UIP. They conclude that the magnitude of the problem of HBV had been probably overestimated in India. One of their key messages is that "further data regarding the true extent of the problem is required so that the utility of the vaccination program can be evaluated objectively." Although not stated explicitly, their tone is against the introduction of HBV vaccine under UIP, until better data are obtained. They certainly have not endorsed the recommendation to include it under UIP. The word chosen in this context is "utility" which raises question about the very usefulness of the vaccine against the backdrop of allegedly "overestimated" prevalence rate.
While more data and more accurate data are always welcome, we already have more than enough information for justifying the utility of the vaccine under UIP. Let not perfection be the enemy of the good that we can do. If someone says that accurate data are not available for recommending routine immunization, we pose the question: at what level of incidence or prevalence will the recommendation be made? This has to be stated first.
We know already that HBV infection is geographically widespread in India. Even though there are geographic and other variations in the risk of infection, almost everyone is potentially at risk of getting infected. The modes/routes of infection include mother-to-child, person-to-person, nosocomial, and sexual. We do not need to accurately quantify the incidence of infection or the prevalence of carrier state to recommend it to all at risk. The affluent families are already using the vaccine to protect their children. It is the less privileged families that depend on UIP for the vaccine. The risks of infection and carrier state are more among them than in the more affluent. Therefore, it is the responsibility of the Government to ensure that they are protected through the UIP. The vaccine is no longer very expensive; today 3 doses may be obtained for a Government immunization programme at less than 100 Rupees. The Government should not consider this as an item of expenditure, but as an investment for development, since preventing disease and disability is an integral part of development.
Lodha and colleagues cited a paper on cost-effectiveness of HBV vaccination in the United States, where the virus carrier rate is very much lower and the cost of vaccination very much higher than in India(5). The conclusion of that paper was: "In conclusion, HBV vaccine is a cost-effective intervention that demands increased attention and emphasis. We recommend a two pronged strategy. First, all pregnant women should be screened before delivery, and newborns of mothers who test positive (HBsAg+) should be given HBIg and vaccination. Second, all adolescents should show proof of complete HBV vaccination at entry to middle or junior high school, just as children entering elementary school must have proof of vaccination against polio, measles and other contagious childhood diseases"(5).
We already know that in our country over 1% neonates are at risk of vertical infection. Therefore, pregnant women must be screened for HBV infection wherever possible, and infants born to virus carriers must be given immediate post-birth immunization. Where screening of mothers is not feasible, every attempt should be made to begin immunization as soon after birth as possible. When even this is not feasible, the infant should be started on immunization at the time of other UIP immunizations. The HBV vaccine is safe and effective. Its utility is unquestionably evident from innumerable studies in India and also from HBV vaccination experience in over 100 countries in the world. There is no excuse whatsoever, epidemiological or economic, to delay the inclusion of HBV vaccine in India’s UIP.
T. Jacob John,