History

Zahorsky generally is given credit for the original description of roseola infantum. However, in his writings, he pointed out that the syndrome was described in earlier pediatric and dermatology texts. Altschuler observed that a British dermatologist, Willan, presented a description of the illness in his 1809 book On Cutaneous Diseases . The descriptions in the older literature did not separate the syndrome from the known exanthematous diseases (measles, rubella, and scarlet fever), an omission that Zahorsky corrected.

In 1921, Veeder and Hempelmann described the syndrome further and noted that leukopenia and relative lymphocytosis occurred. These investigators objected to the name roseola infantum, which in the past had been used to describe a large group of diseases with indefinite causes. They suggested the term exanthem subitum because it was “descriptive of the most striking clinical symptom, namely, the sudden, unexpected appearance of the eruption on the fourth day.” Currently, the term roseola is used most commonly to describe the syndrome.

From 1920 through 1940, many excellent clinical descriptions of the syndrome were published. From 1940 through 1988, articles relating to roseola were concerned with unusual manifestations and complications and attempts to recover an etiologic agent. In 1951 and 1954, Neva and associates noted the association of a roseola-like illness and infection with echovirus 16. This association was noted again in 1974. In 1988, Yamanishi and associates identified human herpesvirus–6 (HHV-6) in the blood of infants with roseola, and, since then, the association between this virus and the disease has been confirmed on many occasions. *

* References .

Epidemiology

In his original article, Zahorsky reported that roseola occurred most commonly in the fall. In his second article, he observed a year-round incidence ; in 1925, he pointed out that most cases occurred in the spring, summer, and fall. Breese noted that the greatest number of cases occurred in the summer and early fall. In contrast, 55% of Clemens’ cases occurred in February, March, and April; 16% were seen in October. In a review of 243 cases during a 10-year period, Juretic observed that the peak month was May. Juretic also reviewed the seasonal incidence in 10 other studies and found only minor variations by month. Prevalence was greatest in March, April, and October and least in December. One epidemic of roseola in a maternity hospital occurred in the summer, another epidemic in an infants’ home occurred in the fall, and a hospital outbreak occurred in the winter.

Roseola predominantly is an illness of young children. It occurs rarely in infants younger than 3 months or children older than 4 years. In a review of 1462 cases, the peak age range prevalence was 7 to 13 months of age; 55% of the cases occurred within the first year of life, and 90% occurred within the first 2 years of life. Occasionally cases have been seen in older children, adolescents, and young adults and in neonates and other infants younger than 6 months.

Although Faber and Dickey found twice as many girls as boys with the syndrome, the sex ratio in most large studies has been equal. Although three epidemics have been reported, and cases frequently occur in groups by season, most cases occur sporadically without known exposure. The syndrome, when seen sporadically, generally is considered to be noncontagious, but secondary cases have been reported occasionally. The incubation period range in epidemics is 5 to 15 days.

The attack rate of roseola has not been well studied. Berenberg and associates stated that roseola is the exanthem most commonly encountered in children younger than 2 years. Breese found that 16% of a group of infants he followed for the first 12 months of life had definite roseola. He estimated that 30% of children would have clinical roseola. Juretic looked at the frequency of roseola in 6735 children; the yearly attack rate during a 10-year period ranged from 1% to 10%, with a mean of 3.3%.

Etiology

In 1941, Breese reported vigorous attempts to isolate a filterable virus from three children with preeruptive roseola. These studies included extensive animal inoculations, but no viral agents were uncovered. In 1950, Kempe and associates reported the passage of the illness to a 6-month-old susceptible infant by the intravenous injection of serum from an 18-month-old child with preeruptive roseola. Febrile illnesses without exanthem also were produced in monkeys with serum and throat washings from a child with the syndrome. In similar experiments, Hellstrom and Vahlquist produced the syndrome in three children aged 6 to 9 days after the intramuscular administration of blood from typical roseola cases.

In electron microscopy studies, Reagan and associates observed uniform viruslike particles (100 to 110 nm) in the blood of an 18-month-old child with the syndrome. Febrile illness was produced in two monkeys after concentrated virus-containing material was inoculated.

Since the advent of modern diagnostic virology in the early 1950s, numerous viral agents have been recovered from children with roseola. In 1951, Neva and associates studied an epidemic exanthematous illness (i.e., Boston exanthem) caused by echovirus 16, in which many of the illnesses were characteristic of roseola. In 1954, Neva observed additional cases of roseola-like illness associated with echovirus 16 infection. In 1974, Hall and colleagues reported four additional echovirus 16 infections with clinical manifestations of roseola. The reporting of roseola in Rochester, New York, nearly doubled during the time of echovirus 16 activity in the area. Roseola-like illnesses that also have been associated with enteroviruses are caused by coxsackievirus A6, A9, B1, B2, B4, and B5 and echovirus 9, 11, 25, 27, and 30. Outbreaks of roseola that occur in the summer and fall probably are caused by enteroviral infections.

In addition to enteroviruses, adenovirus types 1, 2, 3, and 14 and parainfluenza type 1 virus have been recovered from children with roseola. Saitoh and associates detected rotavirus capsomeres in fecal specimens of nine children with roseola. In contrast to these findings, Gurwith and colleagues studied fecal specimens from five children with roseola, and in none were viral particles identified. One of 13 children in this study did develop antibody to rotavirus around the time of illness, however. In addition to the occurrence of roseola associated with numerous natural viral infections, its pattern (i.e., fever and then rash with defervescence) was observed frequently in recipients of Edmonton B measles vaccine.

In 1988, Yamanishi and associates isolated HHV-6 from four infants with roseola, and all four had significant titer increases for this virus. Shortly after this finding was reported, several other investigators noted similar findings. The implication from these studies, as suggested by the various investigators, is that HHV-6 is the cause of roseola. This viewpoint overlooks or ignores the past experience in which other viral agents have been associated with the clinical syndrome. Subsequent studies indicate that HHV-6 is a major cause of roseola and the cause of acute febrile illness without exanthem in infants. *

* References .

In a study of 1653 infants and young children with acute febrile illnesses, Hall and colleagues found that 160 (9.7%) had primary HHV-6 infections; 27 (17%) of the children who were infected with HHV-6 had roseola. Zerr and associates identified 80 children with primary HHV-6 infections, and, of these, 30% had roseola. In the same population-based study, three of 80 (4%) children without primary HHV-6 infection also had roseola. In a study of clinical roseola, Okada and associates found that 81% had serologic evidence of HHV-6 infection and that 8% had an echovirus-18 infection. In a study of roseola in Italy, Braito and Uberti found serologic evidence of HHV-6 infection in only 30% of the cases. In 33% of the remaining cases, they attributed the illnesses to another infectious agent. In 1994, Hidaka and associates estimated that 73.5%, 10.2%, and 16.3% of their roseola cases were caused by HHV-6, HHV-7, and other viruses, respectively. In a study of HHV-6 infections in young Brazilian children with rashes it was noted that only 21% had typical roseola. In summary, HHV-6 is an important cause of roseola but not the only cause.

Pathophysiology

The pathophysiology of roseola is unknown. Watson, in the pre–HHV-6 era, suggested that roseola is not an infection caused by one particular pathogen but is instead the result of an immunizing reaction against many different viruses. He also suggested that the rash is caused by the neutralization of virus in the skin at the end of the period of viremia.

Viremia is common in HHV-6, HHV-7, enteroviral, and adenoviral infections; thus a reasonable conclusion (as originally suggested by Watson ) is that the rash in roseola is related to an immunologic event resulting from the virus that is localized in the skin. Why the pattern of fever and then rash with defervescence is so clearly age dependent is unknown. Most of the viruses that in the past have been associated with roseola cause other exanthematous manifestations in older patients.

Clinical Presentation

The basic clinical pattern of roseola is a febrile period of 3 to 5 days, defervescence, and the appearance of a rash that persists for 1 to 2 days. Because the syndrome is caused by many different viruses, the illness apparently may be associated with numerous other signs and symptoms. The major manifestations have been reviewed elsewhere.

Illness usually occurs with the apparent abrupt onset of fever. Slight irritability and malaise occur frequently, but, more commonly, the child’s temperature is taken because a parent notices that the child feels warm. The temperature usually is in the range of 38.9°C to 40.6°C (102.2°F to 105°F). Despite the high fever, the child usually is active, alert, and generally unfazed. The fever is constant or intermittent, with its greatest degree occurring in the early evening. Restlessness and irritability occur with higher temperatures. The usual duration of fever is 3 to 5 days, but it has persisted for 9 days. The temperature most often returns to normal by crisis, but, in some cases, temperature “lysis” occurs over the course of 24 to 36 hours.

Mild cough and coryza are seen frequently in cases occurring in the winter and spring. Headache and abdominal pain are reported in older children, mainly in the summer and fall. Vomiting and diarrhea occur infrequently.

On initial physical examination during the febrile period, most children appear to be happy, alert, and playful. With high temperatures, some children are irritable; occasionally a child appears to be sick, which suggests more serious illness, such as meningitis or septicemia. Examination within the oral cavity frequently reveals one or more abnormalities. Mild inflammation of the pharynx and tonsils occurs most commonly. Occasionally small exudative follicular lesions are noted on the tonsils. In other cases, small ulcerative lesions on the soft palate, uvula, and tonsillar pillars are observed. Usually the lesions on the soft palate consist of only erythematous macules and maculopapules, presumably because of lymphoid hyperplasia.

Mild injection of the tympanic membranes occurs commonly. Enlargement of the suboccipital, posterior cervical, and postauricular lymph nodes is a common finding, but the degree is not remarkable.

Berliner noticed that children with roseola had palpebral edema. He suggested that the “heavy eyelids” or “droopy” or “sleepy” appearance resulting from this edema was diagnostic of the syndrome before the appearance of the rash. Bulging of the anterior fontanelle also has been observed in roseola.

Appearance of the rash in roseola usually coincides with the subsidence of fever, but it may occur after an afebrile interlude of several hours to 2 days. When defervescence occurs by lysis, onset of the exanthem can occur before the temperature has returned entirely to normal. By definition, it is incorrect, however, to call an illness roseola if the fever and rash are truly concomitant.

Zahorsky originally described the rash as morbilliform, but his use of morbilliform was not the same as is used today (measles-like, erythematous, maculopapular with confluence). The rash is erythematous and macular or maculopapular, and the lesions are discrete. The lesions are 2 to 5 mm in diameter and blanch on pressure. Frequently individual lesions are surrounded by a whitish ring. The rash is most prominent on the neck and trunk, but the proximal extremities and the face also may be affected. Although they have been reported, pruritus and desquamation usually do not occur. The rash usually persists for 24 to 48 hours. In occasional cases, well-documented rashes have been observed to appear and resolve within 2 to 4 hours. Yoshida and associates described a 7-month-old boy with HHV-6 infection and typical roseola initially. On the ninth day of illness, vesicular lesions appeared on the face and limbs, however. These lesions persisted for 12 days.

Except for the white blood cell count, routine laboratory studies are of little use in roseola. The total white blood cell count usually is low. Early in the febrile period, high counts occasionally are found, however. The total count reaches its nadir by the third to sixth day of illness and then gradually returns to normal over the ensuing 7 to 10 days. During the same time frame, the percentage of lymphocytes increases from a normal value of about 50% to 60% to 80% on days 3 to 10 and then returns to normal over the next 7 days. Frequently extreme counts in the range of 3000 cells/mm ³ with 90% lymphocytes are found, which raises the consideration of a granulocytic defect. Huang and Lin note that sterile pyuria may occur in infants with roseola. They noted that these cases can be differentiated from bacterial urinary tract infections by the presence of leukocytosis.

Clinical Complications

The most important complications of roseola are convulsions and other neurologic symptoms. *

* References .

Möller also reported cerebrospinal fluid evaluations in 29 cases of roseola and febrile convulsions. In six instances, the pressure was elevated; in two, there were 5 white blood cells/mm ³ ; and in another instance, there were 9 white blood cells/mm ³ . In most other cerebrospinal fluid examinations, the findings have been normal, but mild pleocytosis with mononuclear cells has been identified occasionally. A surprising number of cases of encephalitis associated with roseola have been reported, and residua have been common. Hemiplegia has occurred after illness, and permanent paresis and mental retardation have occurred in some affected patients. The syndrome of inappropriate secretion of antidiuretic hormone has been reported in roseola associated with HHV-6 infection. Facial nerve palsy and Guillain-Barré syndrome also have been noted after HHV-6–induced roseola.

Thrombocytopenic purpura was noted in one report in five children with roseola; all of these patients recovered. In a more recent study, Hashimoto and colleagues noted five children with thrombocytopenia during the acute phase of roseola caused by HHV-6 infection. Their data suggested that the thrombocytopenia was due to bone marrow suppression, rather than immune-mediated peripheral consumption. A 14-month-old girl developed a generalized eruptive histiocytoma with rapid progression and then resolution after roseola.