Recombinant hepatitis B vaccine and the risk of multiple sclerosis: A prospective study
C RainaMacIntyre, National Centre for Immunisation Research, Australia, NCIRS, Children's Hospital at Westmead and University of Sydney, Westmead, 2145, NSW, AustraliaRainaM@chw.edu.au
Heath Kelly, Damien Jolley, Helmut Butzkueven, Daniel Salmon, Neal Halsey, Lawrence H Moulton
Submitted November 11, 2004
Hernan et al [1] report an association between hepatitis B vaccine (HBV) and multiple sclerosis (MS) in a study mislabelled as “prospective”. Although the source data were recorded prospectively, this is a retrospective nested case control study. This introduces potential biases, such as ascertainment of HBV status when vaccinations were recorded in the notes but not in the computerized database.
The UK HBV program is targeted at people at high risk of infection. [2] However, uptake of HBV is sub-optimal in these groups [3,4], so that vaccinated individuals are those at highest risk. This means that HBV status was highly likely to be confounded by risk factors such as intravenous drug use (IVDU) and sexual practice, neither of which were adjusted for by Hernan. [1]
IVDU is associated with immune-mediated diseases, which may include MS. [5] The sub-analysis excluding the 2% of subjects with “risk factors” was inadequate, as it excluded only people whose records indicated a history of occupational risk, alcoholism, drug abuse or chronic renal failure. Most of this information on these and other critical variables was not in the database or medical records. The inclusion of influenza and tetanus vaccinations does not rule out bias, which could operate selectively for exposure to HB vaccination.
Only 163 of the 438 eligible cases were included in the study, creating the potential for bias in ascertaining cases, and in selectively determining onset and severity of symptoms for vaccinated cases. The principal analysis was based on only 11 vaccinated patients with MS. [1] This may explain an association between HBV and MS when analyzed by date of symptom onset, but not by date of diagnosis. The diagnosis was made in the past, independently of the study investigators, while the investigators determined date of symptom onset. Data for cases came from both the database and notes, whereas data on controls were obtained only from the database.
Upper respiratory tract infections exacerbate or trigger MS within a few weeks, so why would HBV trigger MS three years after vaccination? HBV is a subunit of the HB virus, which is not one of the many viruses previously implicated in MS. Using the Bradford Hill criteria for causation, the criteria of biological plausibility, consistency, coherence and dose–response are not met. The temporal association reported in this study requires one to assume a long lead-time, and is perhaps the strongest single argument in favor of a spurious association. The large body of negative evidence and the significant methodological weaknesses of the study indicate no need for change in vaccination policy.
References
1. Hernan MA, Jick SS, Olek MJ, Jick H. Recombinant hepatitis B vaccine and the risk of multiple sclerosis: a prospective study. Neurology 2004; 63:838-42.
2. Goldberg D, McMenamin J. The United Kingdom’s hepatitis B vaccination strategy – where now? Communicable Disease and Public Health 1998; 1:79-83.
3. Opaneye AA. Bashford J. Audit of hepatitis B immunization at the genitourinary medicine department in Middlesbrough, UK.International Journal of STD & AIDS. 13:268-70, 2002 Apr.
4. Dunn J, Shukla R, Neal K. Survey of neonatal hepatitis B vaccination in Leicestershire. Communicable Disease & Public Health 1999;2:218-9
5. Brosseau L, Philippe P, Methot G, Duquette P, Haraoui B. Drug abuse as a risk factor of multiple sclerosis: case-control analysis and a study of heterogeneity. Neuroepidemiology 1993; 12:6-14.
Hernan et al [1] report an association between hepatitis B vaccine (HBV) and multiple sclerosis (MS) in a study mislabelled as “prospective”. Although the source data were recorded prospectively, this is a retrospective nested case control study. This introduces potential biases, such as ascertainment of HBV status when vaccinations were recorded in the notes but not in the computerized database.
The UK HBV program is targeted at people at high risk of infection. [2] However, uptake of HBV is sub-optimal in these groups [3,4], so that vaccinated individuals are those at highest risk. This means that HBV status was highly likely to be confounded by risk factors such as intravenous drug use (IVDU) and sexual practice, neither of which were adjusted for by Hernan. [1]
IVDU is associated with immune-mediated diseases, which may include MS. [5] The sub-analysis excluding the 2% of subjects with “risk factors” was inadequate, as it excluded only people whose records indicated a history of occupational risk, alcoholism, drug abuse or chronic renal failure. Most of this information on these and other critical variables was not in the database or medical records. The inclusion of influenza and tetanus vaccinations does not rule out bias, which could operate selectively for exposure to HB vaccination.
Only 163 of the 438 eligible cases were included in the study, creating the potential for bias in ascertaining cases, and in selectively determining onset and severity of symptoms for vaccinated cases. The principal analysis was based on only 11 vaccinated patients with MS. [1] This may explain an association between HBV and MS when analyzed by date of symptom onset, but not by date of diagnosis. The diagnosis was made in the past, independently of the study investigators, while the investigators determined date of symptom onset. Data for cases came from both the database and notes, whereas data on controls were obtained only from the database.
Upper respiratory tract infections exacerbate or trigger MS within a few weeks, so why would HBV trigger MS three years after vaccination? HBV is a subunit of the HB virus, which is not one of the many viruses previously implicated in MS. Using the Bradford Hill criteria for causation, the criteria of biological plausibility, consistency, coherence and dose–response are not met. The temporal association reported in this study requires one to assume a long lead-time, and is perhaps the strongest single argument in favor of a spurious association. The large body of negative evidence and the significant methodological weaknesses of the study indicate no need for change in vaccination policy.
References
1. Hernan MA, Jick SS, Olek MJ, Jick H. Recombinant hepatitis B vaccine and the risk of multiple sclerosis: a prospective study. Neurology 2004; 63:838-42.
2. Goldberg D, McMenamin J. The United Kingdom’s hepatitis B vaccination strategy – where now? Communicable Disease and Public Health 1998; 1:79-83.
3. Opaneye AA. Bashford J. Audit of hepatitis B immunization at the genitourinary medicine department in Middlesbrough, UK.International Journal of STD & AIDS. 13:268-70, 2002 Apr.
4. Dunn J, Shukla R, Neal K. Survey of neonatal hepatitis B vaccination in Leicestershire. Communicable Disease & Public Health 1999;2:218-9
5. Brosseau L, Philippe P, Methot G, Duquette P, Haraoui B. Drug abuse as a risk factor of multiple sclerosis: case-control analysis and a study of heterogeneity. Neuroepidemiology 1993; 12:6-14.