Statistics from Altmetric.com
Two items in the review of cerebral malaria by Newtonet al 1 warrant comment. The first relates to corticosteroids in the treatment of cerebral malaria, and the second involves the definition of cerebral malaria.
In late December 1965, I arrived in Vietnam as a neurologist with the United States Army Medical Corps., and for the next 6 months I was the only neurologist serving United States forces in Vietnam. I was attached to the 93rd Evaluation Hospital, which opened only 2 weeks before my arrival. As recounted in a recent article about my Vietnam experience,4 two American soldiers with cerebral malaria died at the hospital during those 2 weeks, and the internal medicine specialists requested my involvement in all future cases. After reviewing the treatment protocols of the two soldiers, I decided that adding dexamethasone to the usual antimalarial regimen would be reasonable in patients with severe cerebral malaria. There was, I think, no literature at the time on the use of steroids; if there was, I certainly had no access to it in the combat zone.
During the next 10 months, we saw 19 patients with cerebral malaria, and they all recovered without any residual neurological dysfunction.5 Our success with steroids prompted all the United States medical units in Vietnam to adopt the practice. A decade after the end of the Vietnam conflict, Warrell et al,2 in a double blind trial, concluded that dexamethasone was both ineffectual and deleterious in cerebral malaria, and Hoffman et al 3 concurred. These two papers2 3 are those usually cited by tropical medicine specialists (Newton and Warrell, 1998; White, 1999. See Daroff4 for citations) in recommending against the use of steroids in adult patients with cerebral malaria. Given the importance of the papers2 3 buttressing the “no steroid” mandate, I will summarise them.
Warrell et al 2 studied Thai patients in a village hospital. There were two groups of 50 patients each; eight in the steroid treated group died, and nine of the untreated controls died. Hoffman et al 3 studied Indonesians in a provincial hospital with a mean age of 10.2 years, and found that four of 19 patients in both the treatment and control groups died. The two studies2 3 had mortality rates of 17%–21%, and neurological residua in the survivors, which is in sharp contrast to the experience with United States troops in Vietnam. As mentioned, we had no deaths or residual neurological dysfunction in our 19 patients5 and Dr Andrew Carr, the second neurologist in Vietnam, had an almost identical experience in 1966–7 using steroids in the sicker patients (personal communication). Blount (see Daroff4 for citation) treated 24 patients with cerebral malaria with steroids and had no deaths. The three United States army hospital experiences in Vietnam in 1966–7 included a total of 62 patients with cerebral malaria liberally treated with steroids, with no deaths. There were no neurological residua in our series or Carr's4; Blount did not comment on residua. This compares with the 17% death rate in the study by Warrell et al 2, 21% in that of Hoffman et al,3 and residual dysfunction in their survivors.
The cerebral malaria experience in the second world war is particularly relevant. Mortality rates in American troops were lower than in their Asian counterparts, and the rates in adults were lower than in children.5
Thus, I don't think that the conclusions of Warrellet al,2 Hoffmanet al,3 and Newtonet al 1 are applicable to western combat troops who are usually in excellent health before contracting malaria, and I continue to recommend steroids for an adult patient with cerebral malaria manifesting significant neurological dysfunction. I specify adults, as we lacked experience with children.
The definition of cerebral malaria is the other issue I have with Newton et al.1 We diagnosed cerebral malaria when a patient with confirmedPlasmodium falciparum parasitaemia displayed signs of cerebral dysfunction that could not be explained by hyperpyrexia or detectable metabolic abnormalities.5 The United States Army in the second world war also used this simple diagnostic criteria and experienced a similar occurrence rate. Our 19 patients with cerebral malaria were among about 1200 cases of falciparum malaria admitted during a 10 month period. This occurrence rate of about 1.6% is concordant with reports from United States hospitals in the second world war, which ranged from 1.2% to 2.3%.5
The neurological manifestations in our patients fell into five broad groupings.5 Disturbance of consciousness was the most common (eight patients), and ranged from extreme lethargy to coma; on awakening, the patients displayed the expected transient cognitive disturbances before gradually returning to normalcy. Four patients had delirium, three movement disorders (tremor, myoclonus, chorea), one had a unilateral cerebral hemispheric syndrome, and three had acute personality changes (manifesting in two as a paranoid psychosis and in the other as a delusional state). In each of the these three, psychometric testing showed organic dysfunction that ultimately normalised (Blocker et al 1968; Kastlet al 1968. See Daroff4 for citations). Previous studies, dating back to Anderson's 1927 monograph (see Daroff et al 5 for citation), reported similarly diverse manifestations in cerebral malaria.
Newton et al 1 commented about the loose definitions of cerebral malaria in the literature, often without evidence that confounding secondary causes of encephalopathy were excluded. We are confident that this criticism is not applicable in our series, nor others emanating from the medical experience among United States forces in Vietnam. Nevertheless, Newtonet al 1 suggested a strict definition of cerebral malaria, requiring “a deep level of unconsciousness” or coma. Such restricted criteria might be justified in studies comparing treatment protocols, but to insist that a cerebral condition lacks degrees of severity is contrary to the experience of any neurologist. The all or none definition of cerebral malaria is simply not acceptable.
White and Newton reply:
Daroff has queried the World Health Organisation (WHO) definition of cerebral malaria that we quoted in our paper.1-1 We pointed out that a patient with falciparum malaria with any impairment of consciousness or other sign of cerebral dysfunction should be treated as a medical emergency with parenteral antimalarial drugs. The point about the definition proposed by WHO,1-2 is to allow a direct comparison between studies in which the clinical syndromes are precisely defined
This point is well illustrated by Dr Daroff's comments on studies of corticosteroids in malaria; he compares his studies on cerebral malaria, in which the cerebral involvement was described in terms of “extreme lethargy”, “delirium”, or “stupor”1-3to those in whom cerebral malaria was strictly defined.1-4 1-5 It is difficult to compare the mortality in such disparate groups.
As for the continued use of corticosteroids in cerebral malaria, Daroff's recommendation is based on anecdotal experience during the second world war and Vietnam in the 1960s. There is little theoretical basis for using corticosteroids.1-6 The two double blind, randomised control trials of dexamethasone failed to demonstrate any benefit from corticosteroids in adults with a precise definition of cerebral malaria.1-4 1-5 Indeed the studies were associated with a significant increase in gastrointestinal bleeding,1-4 1-5 infections,1-4 and duration of unconsciousness.1-4
It is difficult to assess historical anecdotal evidence as there are many factors which may contribute, such as differences in the definitions of the clinical syndrome, drug resistance, and patient groups. Although the two randomised trials do not exclude a potential benefit, there is little substantial evidence to support the use. The onus is on the people who think that corticosteroids are beneficial, to provide more substantial data to support their conjecture, preferably as the results of a randomised trial.