Intended for healthcare professionals

General Practice

Measles and rubella misdiagnosed in infants as exanthem subitum (roseola infantum)

BMJ 1996; 312 doi: (Published 13 January 1996) Cite this as: BMJ 1996;312:101

This article has a correction. Please see:

  1. Dereck R Tait, senior registrara,
  2. Katherine N Ward, senior lecturer/honorary consultanta,
  3. David W G Brown, consultant virologistb,
  4. Elizabeth Miller, consultant epidemiologistc
  1. a Department of Virology, Royal Postgraduate Medical School, University of London, London W12 0NN
  2. b Enteric and Respiratory Virus Division, Virus Reference Division, Public Health Laboratory Service Central Public Health Laboratory, London NW9 5HT
  3. c Immunisation Division, PHLS, Communicable Disease Surveillance Centre, London NW9 5EQ
  1. Correspondence to: Dr Ward.
  • Accepted 5 October 1995

Human herpesvirus-6 is widespread throughout human populations, with primary infection usually occurring in infants aged 40-60 weeks.1 Primary infection may be asymptomatic but where there is disease the classic presentation is of the common childhood illness exanthem subitum (roseola infantum).2 This begins with a sudden fever of about 40°C perhaps accompanied by febrile convulsions; the child appears surprisingly well with few other clinical findings apart from leucopenia. The fever resolves after three to five days coincident with the sudden appearance of a finely macular rose rash that may be transient but usually persists for two days. The exanthem is prominent over the thighs and buttocks, where each macule is sometimes surrounded by a fine halo. Despite detailed descriptions, however, exanthem subitum tends to be confused with measles and rubella.3 Two recent studies have shown that measles and rubella are themselves commonly misdiagnosed,4 5 with only 11% of clinical cases of measles in infants under 1 year being validated by laboratory tests.4 To investigate whether many of the misdiagnosed cases of measles and rubella in early childhood were in fact exanthem subitum, serum samples from children under 2 years old with rashes shown not to be measles or rubella4 5 were tested for laboratory evidence of recent primary human herpesvirus-6 infection.

Subjects, methods, and results

A single serum sample was collected from each of 103 children aged 10-120 weeks notified as having clinically diagnosed measles4 or rubella5 (67 and 36 children respectively). Samples were taken a mean of 30 days after the onset of illness and were known not to contain IgM specific for measles, rubella, and human parvovirus B194 5—the cause of “fifth disease,” another common childhood exanthem. An indirect immunofluorescence test for human herpesvirus-6 IgG was used to detect low avidity antibody.1 The results were compared by using logistic regression with those previously obtained with control samples from randomly selected children of the same age.1

Of the 103 children with rashes, 88 (85%) were seropositive for human herpesvirus-6, and of these 40 (39%) had low avidity antibody; in both cases the highest numbers were at about 50 weeks (figure). The proportion of serum samples with low avidity IgG was significantly higher in the children with rashes than in the historical controls (age adjusted odds ratio=3.68, 95% confidence interval 1.96 to 6.88, P<0.001). Among children aged under 1 year, 27 of the 54 (50%) who were seropositive had low avidity antibody; overall the proportion seropositive for human herpesvirus-6 was significantly higher in the study children than in the controls (54/63 (86%) v 47/94 (50%); age adjusted odds ratio=5.78, 2.44 to 13.7, P<0.001). Of the 52 samples from study children obtained within 30 days of illness, 29 (56%) contained low avidity human herpesvirus-6 antibody compared with only eight of the 43 samples (19%) taken after 30 days (P=0.0005; χ2 test with Yates's correction).


Antibody status and age distribution of the children with rashes and of the historical controls


It was impractical to identify primary human herpesvirus-6 infection by testing for IgM since this antibody may also be detected in recurrent infections. We therefore chose our antibody avidity test, which we had previously used successfully to detect primary infection,1 since low avidity antibody is invariably produced briefly after the first encounter with a particular antigen. The present results suggest that the test is most reliable within 30 days of onset of illness and therefore failure to detect low avidity antibody later does not necessarily exclude recent primary infection.

The higher proportion of study children than controls with low avidity human herpesvirus-6 antibody is evidence that the rash was in many cases exanthem subitum. This conclusion is strengthened by the observation that the age distribution of study children who had low avidity antibody was remarkably like that of exanthem subitum as described in Juretic's classic paper.3

The proportion of children under 1 year who were seropositive for human herpesvirus-6 was also higher in study children than controls, suggesting that in some cases, although the serum sample was taken too late for detection of low avidity antibody, the rash was nevertheless exanthem subitum.

This study confirms the importance of human herpesvirus-6 as a cause of rashes in young children and shows that many cases of exanthem subitum are misdiagnosed on clinical grounds as measles or rubella. Laboratory confirmation is essential to ensure the effectiveness of measles and rubella surveillance programmes.

We thank Dr P Farrington for statistical analysis and Mr D J Turner for excellent technical help.


  • Funding Action Research (S/P/2459) and the Department of Health.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract