A little bit of measles does you good

Measles still kills 800 000 children in developing countries every year,¹ although immunisation has substantially reduced the number of deaths. Immunisation lowers mortality primarily by reducing the incidence of measles, but it may also lower mortality by increasing the age at which children are infected and by reducing the severity of infection in immunised children and their contacts.² Morever, the vaccine itself may reduce mortality from conditions other than measles.

Epidemiological research has shown two important characteristics of measles: the severity of clinical illness is largely determined by the infecting dose, and, surprisingly, mild infection and standard doses of Schwarz vaccine substantially reduce mortality from conditions other than measles.³ Children infected with a large dose of measles virus have a shorter incubation period, more severe disease, and a higher mortality. Children who are infected outside the home (primary cases) have milder disease than secondary cases (who are infected in the household with, on average, a larger dose of virus).² This can result in an amplification effect, where each generation of cases becomes progressively more severe; conversely, if index cases are mild or there are only a few generations of cases, perhaps because of immunisation, mortality will be low.²,³

What is the evidence for the remarkable hypothesis that standard doses of Schwarz vaccine reduce mortality from conditions other than measles? Firstly, measles causes only 10% of child mortality, but the vaccine reduces mortality in developing countries by at least 30%.⁴ Secondly, immunised children who have not had measles have a much lower mortality than unimmunised children who have not had measles.³,⁴ This reduction in non-measles mortality is greater in girls than in boys.⁵

In developed countries measles vaccine is usually given at 12-15 months of age because seroconversion rates are higher at that age than in younger children. However, in developing countries many children die from measles before they are 12 months old, so measles vaccine is usually given at 6-9 months. In 1990 the World Health Organisation recommended that high doses of the Edmonston-Zagreb vaccine should be given at the age of 6 months,³ because this gave much higher seroconversion rates than standard doses of Schwarz vaccine given at 6 months. However, this recommendation was rescinded when it was found that girls given the high titre Edmonston-Zagreb vaccine had a higher mortality than girls who had received the standard Schwarz vaccine.

The higher mortality was not due to vaccine failure: the girls did not have more measles, and they did not have a higher mortality than unimmunised children. The explanation seems to be that high titre Edmonston-Zagreb vaccine did not protect against mortality from conditions other than measles (an effect that is more marked in girls than boys).³ A little bit of vaccine does you good—but a lot of vaccine is not so good.

When standard doses of Schwarz vaccine are given at 4-8 months of age seroconversion rates are lower than after vaccination at 9 months and more children get measles. However, case fatality rates are lower in the excess cases, and the protection against non-measles deaths occurs earlier, so total mortality is lower with immunisation at 6 months despite the lower seroconversion rate.⁶

Severe measles has a high fatality rate, so it is not surprising that many studies have found that children who have measles have a higher mortality than children who do not have it. However, many of the children who do not get measles have been immunised, which reduces their mortality from diseases other than measles. Compared with unimmunised children who have not had measles, unimmunised children who have measles as primary cases (with a small innoculum) have a lower mortality, but secondary cases (with a larger innoculum) have a similar or higher mortality.⁷ A little bit of measles does you good—but a lot is bad.

These observations suggest two important conclusions: when measles occurs after immunisation this does not necessarily imply total vaccine failure, and the effects of a new vaccine cannot therefore be assessed solely by antibody responses and protection data. Vaccine trials will provide more useful information if they concentrate on mortality rather than laboratory evidence of seroconversion and clinical illness.

We have the ability to eradicate measles.¹ However, there is strong evidence that measles vaccine protects against death from conditions other than measles, so it might be sensible to continue to give measles vaccine to children in developing countries even if we eradicate the disease.