Article Text

Download PDFPDF

Thrombocytopenia is predictive of lethality in severe childhood falciparum malaria
  1. C Rogier1,
  2. P Gerardin2,
  3. P Imbert3
  1. 1Parasitology Unit, Institut de Médecine Tropicale du Service de Santé des Armées–IFR 48, Le Pharo, Marseille, France;
  2. 2Neonatal and Pediatric Intensive Care Unit, Hôpital Alfred Isautier, Saint-Pierre, Reunion Island, France
  3. 3Department of Infectious Diseases and Tropical Medicine, Hôpital d’Instruction des Armées Bégin, Saint-Mandé, France

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Moulin and others1 reported that thrombocytopenia was not a marker of severity in children suffering falciparum malaria. In a previously published study,2 we have shown a highly significant association between thrombocytopenia and either severity or prognosis in childhood falciparum malaria. To our knowledge, this association had not been noted3 or looked for4 before.

Our study took place from October 1997 to March 1999, in the paediatric department of the Hopital Principal in Dakar, Senegal, where clinical presentation, WHO criteria of severe malaria, and platelet count were prospectively recorded. Of 288 falciparum cases, 215 matched the 2000 WHO definition of severe malaria. Median platelet counts were lower (98 000/mm3 versus 139 000/mm3; p < 0.02) among severe cases than in mild cases, and in children who died (n = 26) than among those who recovered (68 500/mm3 versus 109 000/mm3; p < 0.002). In severe cases, children presenting with a platelet count <100 000/mm3 were more likely to die (20%, 22/110) than those with higher platelet count (3.8%, 4/105; odds ratio (OR) 6.31, 95% confidence interval (CI) 2.0 to 26.0; p < 0.0003). Moreover, multivariate analysis identified thrombocytopenia as an independent predictor of death (OR 13.3, 95% CI 3.2 to 55.1; p < 0.0001)—that is, when the effect of cerebral malaria, respiratory distress syndrome, severe anaemia, and other severe malaria criteria was taken into account.

The absence of an association between thrombocytopenia and clinical malaria severity in the samples studied by Moulin and others does not prove its non-existence. Moreover, their study suffers from several limitations that may explain the discrepancy between their and our conclusions. First, they have observed only four deaths among 234 malaria cases. Their survey was consequently powerless to establish any association between thrombocytopenia and lethality. The low level of lethality among severe cases (3.5%, 4/112) also suggests that clinical presentations observed in the Moulin et al study were less life threatening than in ours, even if 69 of their severe cases have been admitted in the same paediatric department where our study was conducted previously. It is well established that the case fatality rate varies according to clinical presentation5 and definition of severe falciparum malaria criteria.6 Thus any analysis of prognostic factors of malaria lethality must take into account the composition of the sampled cases in terms of clinical presentations, for example, analysing the different syndromes separately. In our study, the association between thrombocytopenia and lethality was significant among children with cerebral malaria or respiratory distress but not among children only presenting with severe anaemia, convulsions, or hypoglycaemia. It is possible that the composition of the severe cases sampled by Moulin and others may have confused the association. Age is another confounding factor that was not controlled in the Moulin et al study. We showed that age was associated with both lethality and thrombocytopenia.

Furthermore, the exposure of children to malaria may be not so different between the samples studied by Moulin and others. In large areas of Dakar, the level of transmission is very low, less than one infective bite every 10 years. It is thus likely that a significant proportion of the children from Dakar, and to a less extent from Libreville, had never been infected before by Plasmodium falciparum. Without reliable information about the exact location of their habitat, it is inappropriate to estimate their level of previous exposure to malaria in urban areas where the level of transmission is heterogeneous.

Finally, it must be stressed that in the studies of prognostic factors of lethality, the occurrence of death rather than substitute variables, for example, severe malaria criteria, is the variable of interest. Among children suffering severe falciparum malaria, thrombocytopenia (<100 000/mm3) should be considered as predictive of a fatal outcome, especially in those with cerebral malaria or respiratory distress.2