HIV more than doubles the rate of deaths in Zambian children hospitalised with measles

HIV-infected Zambian children had over twice the risk of dying in hospital with measles than Zambian children who were not HIV-1-infected, according to the findings of an observational study published in the February 15th edition of Clinical Infectious Diseases. The HIV-infected children hospitalised with measles were more likely to be younger than 9 months than HIV-1-uninfected children.

Measles still remains a significant cause of childhood deaths in sub-Saharan Africa where 126,000 African infants die annually from measles despite the existence of an effective measles vaccine. Measles-associated deaths are due to increased susceptibility to secondary bacterial and viral infections such as bacterial pneumonia.

Risk factors for measles mortality range from lack of previous measles vaccination, young age, female sex, to malnutrition, particularly vitamin A deficiency. The role of HIV-1 in measles mortality remains unknown largely because the few studies of measles mortality were carried out in regions with no significant HIV-1 prevalence.

Since HIV-1 infection impairs both cellular and humoral arms of immunity, it is intuitive that HIV-1 contributes to measles mortality. The evidence for the role of HIV-1 in increased measles mortality comes from studies in the USA which are based on small numbers of children. A large study of Zambian children with confirmed measles virus (MV) infection lacked the power to detect any significant difference in hospital deaths between HIV-1-infected and HIV-1-free children.

In order to establish whether HIV-1 played a role in measles mortality in the same population of Zambian children, a team of US, British, and Zambian researchers extended the enrollment of infants by an additional 2.75 years and re-examined risk factors for in-hospital deaths.

The study site was the University Teaching Hospital in Lusaka, Zambia where children hospitalised with clinically diagnosed measles were enrolled between January 1998 and July 2003. The children came from three townships of Lusaka and were part of a measles vaccine study. Measles was confirmed by detection of antimeasles virus immunoglobulin M antibodies in plasma, and HIV-1 infection was confirmed by detection of plasma HIV-1 RNA.

Of 1474 enrolled children between January 1998 and July 2003, 1227 (83%) had confirmed measles and known HIV-1 infection status. Of the 1227 children with measles, 189 (15 %) had coinfection with HIV-1. HIV-1-infected children were more likely to have unconfirmed measles than HIV-1-uninfected children (18 versus 8 %; p

Almost one-third of the HIV-1–infected children with measles were less than nine months of age, the age of the first measles vaccination, compared with one-fourth of the uninfected children (p=.07). Mothers of HIV-1-infected children were more likely to have had more than 8 years of education than mothers of HIV-1-uninfected children. Among children aged more than 9 months old, 60 % of HIV-1-infected and 47 % of HIV-1-uninfected children had received measles vaccine (p = 0.008).

The overall measles case fatality rate among the 189 children with confirmed measles and HIV infection was 5.5 %. Death occurred during hospitalization in 23 (12.2%) of the HIV-1–infected children and 45 (4.3%) of the HIV-1–uninfected children (P <.001 measles.="" with="">

After adjusting for age, sex, and measles vaccination status, HIV-1 infection (odds ratio, 2.5; 95% confidence interval, 1.4–4.6), 8 years or less of maternal education (odds ratio, 2.4; 95% confidence interval, 1.2–4.8), and the presence of a desquamating rash (odds ratio, 2.2, 95% confidence interval, 1.3–3.6) were significant predictors of mortality due to measles. In this study the female sex was not a risk factor.

The strength of this study is that it is the largest prospective study of measles in a region of high HIV-1 prevalence in which both HIV-1 and MV was laboratory-confirmed. It therefore provides strong epidemiological evidence for the role of HIV-1 as an important risk factor in deaths of HIV-1-infected African children with measles.

A previous report from the same authors (Scott et al) demonstrated that HIV-infected infants born to HIV-seropositive Zambian women did not have maternally-acquired passive protection against MV and had an increased risk of measles before routine measles vaccination at 9 months.

In conclusion, HIV-1/MV coinfection more than doubled the rate of deaths in Zambian children hospitalised with measles. The policy implication is that public health authorities should invest greater efforts in reducing transmission of the MV in regions of high HIV-1 prevalence.