Scenario

A chatty, 4 year-old girl is brought to the emergency department (ED) by her mother with a 3-day history of fever and loose stool. They returned from Nigeria 7 days earlier, having visited friends and relatives for the school holidays. Clinical examination is unremarkable, she is currently afebrile, and there are no signs of serious bacterial infection. A malaria blood film and histidine rich protein-2 (HRP-2) based rapid diagnostic test (RDT) are both negative. RDTs are for malaria antigens such as HRP-2 and lactate dehydrogenase (LDH).

The paediatric registrar discharges the patient but expresses concern regarding the ability of a single RDT and blood film to rule out malaria, following current guidelines, and insists she returns in 24 and 48 hours to repeat the tests. You wonder if this is really necessary.

Structured clinical question

In children returning from malaria-endemic areas with a history of fever do multiple blood films or RDTs improve the sensitivity and negative likelihood ratio for the diagnosis of malaria?

Summary of results

A recently published study by Houze et al recruited 1311 returning travellers to eight French hospitals. Five of the eight centres recruited children, though the number is not reported. More than half were born in malaria-endemic countries. The authors concluded that HRP-2 based RDTs are highly sensitive for P. falciparum, with false-negative results occurring in Pf-LDH tests only. The sensitivity of RDTs in detecting non-falciparum species was lower, and the authors concluded that one negative RDT must always be accompanied by a negative blood film to rule out malaria. This is the only peer-reviewed study to include children, and no paediatric specific data are provided. An abstract presented to the RCPCH reported on 104 returning child travellers to the Birmingham Children’s Hospital ED. A substantial majority (82%) of children had only one set of investigations performed and sensitivity for a combined LDH/HRP-2 RDT was reported as 100%. It is not clear how follow-up was achieved and whether or not children may have re-presented to another hospital for testing. The results have yet to be published in a peer-reviewed journal.

Other data from adult travellers support the finding that RDTs are highly sensitive in this setting. Rossi et al report a negative likelihood ratio of 0.04 for RDTs detecting P. falciparum. Four of 154 confirmed malaria diagnoses were falsely negative by RDT—these were found to contain extremely low parasite densities of 0.1%–0.4%. Two of approximately 2000 patients who tested negative by both RDT and blood film became positive for malaria within 1 week; each had uncomplicated malaria and recovered uneventfully. These more recent data affirm the results of a large meta-analysis including nearly 6000 adult travellers published in 2005 by Marx et al. Reference to the primary studies included in the Marx review yielded no additional data on children specifically. Of the 17 studies available in English, seven reported age ranges which included children but made no reference to the number of children or distinguished diagnostic performance in this age group. Four excluded children and the remaining six made no reference to the age profile of the participants. The most recent round of WHO product testing of RDTs suggests that accuracy has continued to improve since this study 12 years ago.3

The performance of RDTs in children in malaria-endemic regions was reported by Mtove et al who evaluated an HRP-2 based RDT in 965 children in Tanzania. The authors estimated sensitivity and specificity as 98.6% and 97.3%, respectively, compared with research blood film. With only limited loss to follow-up, the authors identified no missed cases of malaria. The population in question are more likely to be immune children rather than non-immune travellers.

Two systematic reviews by the Cochrane group report the diagnostic accuracy of RDTs for the diagnosis of P. falciparum and non-falciparum malaria in endemic countries.1 2 The reviews conclude that RDTs are highly sensitive when compared with microscopy. LDH-based tests were less sensitive than HRP-2 based tests for P. falciparum, and though RDTs for non-falciparum species have become increasingly sensitive, they perform less well than tests for P. falciparum. Although robust, the results should be interpreted with caution in our non-immune population of interest.

Commentary

Malaria remains one of the most common imported infections in the UK, causing an estimated 1500 infections per year, of which 10% occur in children.4 Both adult and paediatric guidance is to perform three blood films if there is clinical suspicion of malaria before ruling out the diagnosis, though supporting evidence for this is elusive.5 6 An early study found initial blood films to be falsely negative in 7% of returning adults infected with any of the four plasmodium species7 and the potential for non-immune patients to deteriorate rapidly with falciparum malaria warrants an appropriately conservative diagnostic approach. With the advent of highly sensitive malaria diagnostics, however, we propose that one set of negative tests, consisting of microscopy and RDT, is enough to rule out the diagnosis in well-appearing febrile children returning from a malaria-endemic area.

Expert microscopy of thick and thin blood films remains the gold standard for malaria diagnostics and allows reliable species identification as well as quantification of parasitaemia. However, more than a quarter of all cases of malaria are seen in centres which see fewer than 10 cases per year.8 In this case, malaria RDTs offer reliable diagnostic accuracy, particularly for the diagnosis of P. falciparum. Paediatricians ought to know whether and which RDTs are being used in their local laboratory. Ten years ago, a survey of laboratories in the UK found that RDTs were being used for only 31% of samples in normal working hours and 44% of samples during on call hours, with the remainder being analysed by microscopy alone.9 There are no more recently published data on RDT availability in UK laboratories. RDTs are less sensitive for non-falciparum species; therefore combining epidemiological knowledge with awareness of local tests in use is particularly important. Both RDT and locally performed blood films should be corroborated by blood films sent to a reference laboratory.

The ideal study to answer our questions would demonstrate the ability of serial malaria blood films and RDTs taken from children returning to the UK from an endemic area to rule out malaria. In the absence of data on children returning from malaria endemic areas, we report the results of diagnostic accuracy studies in adults returning from endemic areas and of studies of malaria diagnostics in children in endemic areas. Children living in endemic areas build up immunity on repeated malarial infection and may have asymptomatic parasitaemia. In such areas, where the prevalence of asymptomatic parasitaemia is high, children present later and with higher parasite counts potentially giving rise to a higher reported sensitivity of diagnostic testing. WHO evaluations of RDTs at low parasite counts of 200/μL provide reassurance for paediatricians in the UK who may be treating child travellers presenting with low level parasitaemia.

The published evidence supports a role for HRP-2 based RDTs in the diagnosis of malaria in children and adults in endemic and non-endemic areas. Interestingly, in some areas of the world, malaria parasites which fail to produce the HRP-2 antigen and thereby evade the HRP-2 antigen test appear to have a survival advantage. In Peru, Gamboa et al found 41% of tested parasites lacked the pfhrp2 gene, leading to false-negative RDTs.10 HRP-2 deletions have also been found in Brazil, Mali, India and Senegal.11–14 False-negative malaria RDTs may also occur in the setting of extremely high parasite densities. Luchavez et al observed a ‘prozone’ effect when testing HRP-2 RDTs against blood samples with varying levels of parasitaemia.15 Blood samples containing the highest parasite densities began to show a fainter test line, though the line never entirely disappeared. This is of lesser concern to clinicians in the UK, where non-immune children are likely to present symptomatically at lower parasite counts. These observations highlight the importance of simultaneous microscopic examination of a blood film.

We have illustrated a somewhat unlikely scenario in which other important causes of fever are confidently excluded. This is frequently not the case, and in febrile children returning from the tropics there may be other important reasons to consider further testing, a period of observation or empirical treatment. A geographically appropriate differential diagnosis including typhoid, dengue and other tropical causes of fever should be considered.16 Additional testing including a full blood count, inflammatory markers and culture or serological testing may all provide evidence for malaria or an alternative non-malarial febrile illness.

Despite the limitations of the available data, we propose that in the well-appearing febrile child returning from the tropics, a negative malaria screen including a single negative RDT combined with a negative blood film is sufficient to exclude malaria. Appropriate safety netting with a planned telephone consultation or with advice to return if fever re-emerges would safely reduce unnecessary attendance and investigation in such children. Should the child be persistently febrile or appear unwell, repeat malaria testing may be appropriate, but attention should shift to identifying other treatable causes.