Watch for cryptococcal disease in children with HIV, South African doctors warn

Clinicians in low- and middle-income countries should give greater consideration to the possibility that infants and children with HIV are presenting with symptoms of cryptococcal disease, according to a report from South Africa published in the advance online edition of AIDS.

A total of 361 (2%) of more than 16000 incident episodes of cryptococcosis reported to the surveillance programme in South Africa over a three-year period were in children, the largest number described to date, Susan T Meiring and colleagues report in a prospective, population-based comparative review of adult and paediatric cases.

Cryptococcosis among children was less frequent than among adults. However, contrary to previous findings, it was not uncommon. In 2007, the authors note, among children under one year of age incidence was comparable to that of haemophilus influenza meningitis, a common cause of bacterial meningitis, with both at 4 per 100,000 population.

Over 90% of all cases were HIV-infected and over two-thirds of both children and adults were severely immunocompromised (CD4 cell counts under 50 cells/mm³) at the time of diagnosis, with only 26% of the children on ART.

While infection cannot be prevented because the Cryptococcus fungus is ever-present in the environment, cryptococcal disease can be prevented. It is believed that infection occurs through breathing in infectious particles. Disease usually happens later in life when severely immunocompromised. Cryptococcal disease presents chiefly in the form of meningitis, causing neck stiffness, headache, fever, changes in mental status and eventual coma and death, if not diagnosed and treated.

The authors stress the importance of getting HIV-infected children diagnosed early and on to ART treatment promptly before they become severely immunocompromised, so putting them at increased risk for developing cryptococcal disease.

South Africa’s Department of Health’s 2012 National Strategic Plan includes plans for screening for cryptococcal antigenaemia in patients with CD4 cell counts under 100 cells/mm³, followed by pre-emptive treatment. A phased implementation is in development. The laboratory screening process will include patients of all ages and so may improve early diagnosis in children.

Many studies have described cryptococcus, a fungus, as a cause of life-threatening disease among HIV-infected adults. In the United States, before effective antiretroviral treatment, approximately 6 to 10% of HIV-infected adults, notably those with CD4 cell counts under 100 cells/mm³, were infected with cryptococcus. Incidence in the general population ranged from two to seven cases per 100,000 people.

In contrast, the authors cite a 2004 study in South Africa estimating an incidence in the general population of 15.6 cases per 100,000 people, yet 14 cases per 1000 among adults living with HIV. But little is known about its incidence and epidemiology in children.

There is little evidence to explain why cryptoccocosis is seen more frequently in immunocompromised adults than in immunocompromised children. Previous studies have shown that, while children become infected after the age of two, it generally affects older children at a median age of 9.8 years. However, some cases have been reported among newborns suggesting vertical transmission of the infection.

The authors chose to compare the incidence and epidemiology of laboratory-confirmed cryptococcal disease in South Africa between children and adults diagnosed from 1 January 2005 until 31 December 2007.

Case definition was determined by a positive India ink test, cryptococcal antigen test or cryptococcal culture.

Additional clinical data were obtained from people who were diagnosed at 21 enhanced surveillance sites (ESSs) across South Africa.

The authors found that, in 2007, incidence among the general paediatric and adult populations was 1 and 19 cases per 100,000 people, respectively.

Incidence among HIV-infected children (47 per 100,000) was lower than among HIV-infected adults (120 cases per 100,000); however, there was a peak incidence among 10 to 14 years-olds (517 cases per 100,000), an unexpected finding. This may reflect, the authors note, an underestimation of the true HIV prevalence in this age range.

Clinical data from ESS was available for 23% (84/361) of paediatric cases and 28% (4378/15831) of adults. While over 90% were HIV-infected, only one in four children was on ART at the time of diagnosis.

The authors suggest these findings may reflect inadequate uptake of the prevention of mother-to-child transmission programme in South Africa in the early years. Implementation has significantly improved so incidence may be considerably lower for children born after 2007, they add.

Additionally among HIV-uninfected inidividuals there was a higher proportion of children with cryptococcosis, compared to adults with cryptococcosis.

Multivariate analysis showed that, compared to adults, children were more likely to be male, diagnosed on blood culture, infected with cryptococcosis gattii, started on amphotericin B and admitted to hospital for a longer stay.

However, children were less likely than adults to be discharged from hospital on fluconazole maintenance therapy (83 compared to 93%, p=0.007). The authors suggest this may have been due to the paediatrician either not knowing the treatment guidelines (as it is seen as an adult-onset disease) or the paediatric fluconazole suspension not being available.

Diagnosis on blood culture, the authors note, suggests a possible chance finding rather than confirmation of a clinical suspicion of the disease.

Even though more children than adults were prescribed amphotericin B, the recommended antifungal drug for managing cryptococcal disease in South Africa, less than a third of adults and children got the drug.

Median age among children at the time of diagnosis was seven (range 0 to 14 years). However, in contrast to previous findings, the authors found there were two time periods of highest incidence: first among those under one year of age, then a second sustained peak in incidence among children aged five to 10 years of age. 

The authors suggest that “since most paediatric HIV infection is vertically transmitted and cryptococcosis is an AIDS-defining illness, this bimodal distribution may reflect the occurrence of cryptococcal disease among rapid and slow HIV progressors”.

Limitations include the fact that only laboratory confirmed disease was reported, so the true burden among adults and children may have been underestimated. Many may have been too ill to access health care services or specimens were not taken for diagnosis.

Clinical data were only available at ESSs, leading to the potential for bias. As an observational study, clinical data sets were incomplete. Insufficient information precluded differentiating between episodes of cryptococcosis resulting from immune reconstitution inflammatory syndrome (IRIS).

The authors conclude that their findings have clearly shown that “the diagnosis of cryptococcosis must be considered in the paediatric HIV-infected population, especially among those found to be severely immuno-compromised.”