Vaccination and type 1 diabetes mellitus

The massive reduction of invasive disease due to Haemophilus influenzae type b has been an outstanding example of the value of immunisation. However, as has been the case with several vaccines, memory of the devastation caused by the disease rapidly fades and is replaced by concerns about the safety of the vaccine. These are often unfounded, but untold damage can result.¹

Coincident with the increase in the number of vaccines given to children there has been a significant rise in the incidence of a number of poorly understood conditions, such as asthma, autism, and diabetes mellitus, has been noted in children in many countries.² Of course this temporal association does not prove a causal link, and the increase in incidence often predates the introduction of most vaccines. However, the temporal associations have raised questions, and such concerns should not be dismissed out of hand.

Classen and Classen have suggested that immunisation after the age of 2 months is associated with an increased risk of diabetes mellitus in rodents and humans.³ Latterly they have specifically implicated H influenzae type b vaccine as having this potential.⁴ One of the most extensive studies reported to date examined the effect of vaccination (excluding H influenzae type b vaccination, which had only been recently introduced) on the incidence of type 1 diabetes in Swedish children.⁵ No single vaccine, or combination of vaccines, was linked to an increase in diabetes, and children who had received a measles vaccine had a decreased risk (odds ratio 0.69, confidence interval 0.48 to 0.98). Last year Jefferson and Demicheli published a systematic review of the possible link between immunisation and diabetes mellitus.⁶ They found few studies that addressed the issue and none that confirmed a link, though no studies had examined an association with H influenzae type b vaccine. It is against this background that the paper by Karvonen et al in this week’s BMJ is particularly timely (p 1169).⁷

The authors have compared the incidence of diabetes mellitus in three groups of Finnish children. Children in cohort 1 were born between 1 October 1983 and 1 September 1985. They received the vaccines that were part of the routine schedule at the time, which did not include H influenzae type b vaccine. Cohorts 2 and 3 comprised children born between 1 October 1985 and 31 August 1987. The children in cohort 2 received the routine vaccines and H influenzae type b vaccine at 3, 4, 6, and 14-18 months. Those in cohort 3 received the routine vaccines and H influenzae type b vaccine at 24 months. Diagnoses of diabetes mellitus made by December 1997 were collected from a national register proved to have complete coverage. The H influenzae type b vaccine used was a conjugate using diphtheria toxoid (PRP-D). The authors found no statistical difference in the cumulative incidence of diabetes at the age of 10 between the cohorts. This was a well designed and very carefully conducted study whose methodology cannot be criticised, so we can be reassured about the validity of the findings.

In March 1998 a workshop was convened at the Johns Hopkins School of Public Health to address concerns about the relation between type 1 diabetes and immunisations.⁸ The workshop panel was drawn from many disciplines and considered evidence from various sources, including conflicting interpretations of the data from the study reported this week by Karvonen et al.⁴ The panel concluded that “selective vaccines are protective against type 1 diabetes in animals but the data in humans are inconclusive and no vaccines have been shown to increase the risk of type 1 diabetes in humans.”

Is this the end of the story? The conclusion to be drawn from today’s paper is reassuring, but it cannot be applied directly to the current situation because the H influenzae type b vaccine used in the 1980s has now been largely superseded by other conjugates, as the latter have been shown to be more effective. On the other hand, there is no reason to think that newer conjugates should behave any differently in respect of diabetes. There is a need for further data, and long term cohort studies are in place to examine the relation between various factors, including vaccines and the development of pancreatic islet cell autoantibodies and type 1 diabetes mellitus.

This study also illustrates how the same data may be interpreted in such a way as to reach conflicting conclusions.^4–8 The importance of personally forming one’s own judgment about the original data, rather than relying on someone else’s interpretation, cannot be overemphasised. This is particularly important in vaccination, where a misguided “scare” can result in the needless suffering and deaths of children whose parents have been understandably frightened into refusing a valuable vaccine.

Papers p 1169