On March 20, 1998 the Johns Hopkins School of Public Health held a vaccine safety work shop in Baltimore, MD to review all available information.  Numerous participants from WHO, CDC, Nat Center for Health Statistics, FDA, National Diabetes Data Group and many others attended this prestigious gathering.  John Classen, MD was allowed to present his case.  After other presentations were made, and a careful evaluation was made by the best leading experts, these are their conclusions:

4. Do the available data indicate an association between childhood immunizations and type I diabetes? In animals? In humans? If so: (a) Do childhood vaccinations protect against type I diabetes? (b) Do childhood vaccinations increase the risk of type I diabetes?

The available data indicate that some vaccines have a protective effect on the risk of developing diabetes in animals, but the data in human studies are inconsistent and inconclusive with regard to a protective effect. There is no evidence that any vaccines have increased the risk of type I diabetes in animals or humans.

Animals. Administration of bacille Calmette-Guerin (BC) vaccine to infant BB rats protects against development of diabetes mellitus . These findings have been confirmed in several other laboratories. In addition Dr. Noel Maclaren reported that administering Freund's complete adjuvant (which contains killed Mycobacterium butyricum organisms) is protective, but incomplete Freund's adjuvant (without the mycobacteria) is not. BCG and Freund's complete adjuvant induce cell-mediated immune responses.

Humans. Dr. J. Barthelow Classen presented ecologic studies of selected human populations supporting lower rates of type I diabetes mellitus in populations where BCG is often given.
Dr. Lawrence Moulton presented a more extensive review of the incidence of type 1 diabetes mellitus by BCG vaccine use. The rates in countries where BCG is routinely given at birth or at 1 to 3 months of age are generally lower than the rates where BCG is not given or given at >1 year of age. Preliminary data from a multiple regression analysis suggest that these differences decrease after adjustment for distance from the equator, per capita gross national product, child mortality and per capita caloric intake. Ecologic studies may be used to generate hypotheses regarding possible factors associated with disease, but ecologic studies do not demonstrate causal relationships. Several other factors could explain the observed differences in diabetes incidence including genetic differences in populations and increased exposure to immune modulating infections early in life in tropical climates.

Do childhood vaccinations protect against type I diabetes?
A study in Sweden revealed that children with type I diabetes were less likely to have received measles vaccination than comparable children without diabetes, but no association was noted for BCG, smallpox, tetanus, pertussis, rubella or mumps vaccines. Additional studies are indicated to determine whether measles or other vaccines reduce the incidence of diabetes.

A case-control study of children with diabetes in Canada revealed no significant differences in likelihood of having received BCG vaccine as compared 17 with controls. Because BCG vaccine is given to al most all infants at birth in China, Dr. LaPorte noted that the marked variability in the incidence of type I diabetes within China is additional evidence against a major effect of BCG on diabetes incidence.

Do childhood vaccinations increase the risk of type I diabetes? Dr. Classen has hypothesized that childhood vaccinations might increase the risk of type 1 diabetes, and the recent introduction of several new vaccines might explain increases in the incidence of diabetes in several populations.  .. However, the panel concluded that the analytic methods were incorrect and a careful analysis of data from 10 years of follow-up has revealed no significant differences in the incidence of type I diabetes mellitus in children who received one vs. four doses of Hib vaccine (J Tuemilehto, personal communication).

Dr. LaPorte presented data demonstrating a global increase in the incidence of type I diabetes mellitus in recent years that cannot be explained by improved surveillance. Because the incidence of type I diabetes mellitus has increased in countries with and without introductions of new vaccines into the immunization schedule, the data do not support the hypothesis that vaccines affect the risk of diabetes mellitus.

Dr. LaPorte obtained additional preliminary data from other diabetes investigators who have conducted case-control studies. None of their studies revealed significant differences in rates of receiving any vaccine in children with type 1 diabetes as compared with controls. Dr. Patricia Graves presented preliminary data from a prospective cohort study (DAISY) of children at high risk for developing diabetes mellitus.    There was no indication that children in the cohort who developed antibodies to islet cell antigens to date had different early childhood immunization histories than the controls. The timing and number of doses of polio, DTP, Hib or measles, mumps and rubella vaccines were no different between case and control groups. There was no difference in the proportion receiving Hib, receiving hepatitis B vaccine (HBV) at birth rather than later, or the mean age of first HBV.

Dr. Frank De Stefano presented preliminary data from the Vaccine Safety Datalink project which reveal a possible lower risk of type I diabetes mellitus in children who received HBV at birth, but the number of cases is low and the confidence intervals are very wide. A larger case-control study is planned. No association between HBV and diabetes in children was noted in a study in New Zealand.

5. Is there evidence to indicate that changing the routine childhood immunization schedule would increase or decrease the risk of developing type I diabetes? No changes in childhood immunization schedules for any vaccines are indicated at this time.