Attempts have been made to differentiate serious bacterial infections from viral infections.6,33,40,43,48,92,93 Clinical and laboratory methods have been employed.92,93 The
38
definitive diagnosis of occult bacteraemia is by blood culture. However, high costs, delay in diagnosis and institution of specific antibiotic treatment, risk of contamination and false negative results are all significant limitations in the developing world. 92 The inadequate levels of skilled man-power and facilities for proper microbiological analysis of samples outside tertiary institutions are additional problems. The attendant sequelae of occult bacteraemia and the limitations of blood culture have led to the search for alternative methods to correctly identify patients with bacterial and parasitic infections from others with viral
infections.6,8,44,47,50,92,93,97
In the tropics, the co-existence of bacteraemia with other infections remains a problem
7,46,93 and no practical methods, clinical or laboratory, have been shown to distinctively distinguish malaria from bacteraemia or LF from other infections. 90 The clinical features of coryza, diarrhoea, seizures, or coma have not distinguished between bacteraemia and malaria.7 Clinical signs of hepatomegaly, severe anaemia, and a combination of hepatosplenomegaly and severe anaemia were significantly of a higher prevalence amongst children with bacteria-malaria coinfection, but could not distinguish between malaria and bacteraemia.93,98
The need for practical methods of differentiating the child with bacteraemia from those with relatively benign viral infections or malaria has led to the evaluation of several clinical indices. These indices include age, association with febrile seizures, height of temperature, toxic appearance (severity of illness) and initial diagnosis
Although occult pneumococcal bacteraemia is associated with febrile seizures, 49 a febrile seizure episode is not an independent risk factor for occult bacteraemia. 44 Teele et al 44 found no significant difference between the prevalence of occult bacteraemia associated with febrile seizures (6.6%) and the prevalence of pyrexia of at least 40oC without seizures (5.0%).
Convulsions associated with malaria are common, and may be due to fever, cerebral involvement or metabolic derangements such as hypoglycaemia. 10, 39 In parts of the world where Plasmodium falciparum malaria is holo- and hyper- endemic, cerebral malaria (and other severe forms of the disease) occurs in children within the 6 months -
39
5 years age bracket, with a frequency of occurrence as high as 16% of all cases of malaria.
4 Differentiating febrile
convulsions from convulsions due to cerebral malaria is difficult on clinical grounds 10, 39, as there are no uniformly defined criteria for the diagnosis of cerebral malaria. The preponderance of malaria as a cause of febrile convulsions in most reports from the tropics may, therefore, be accounted for partly by the inclusion of cases of cerebral malaria. This may partly explain the high morbidity and mortality associated with febrile seizures, in contrast to the benign outcome in temperate regions.
The clinical case definition of malaria as fever without an obvious focus is highly sensitive, but poorly specific.39 Intermittent pyrexia, with a wide range of temperature, may be due to malaria, bacteraemia, meningitis or LF. No studies on the relationship between the degree of temperature and the prevalence of LF were found. Also, it has been argued that, in most
infections or inflammatory diseases, fever patterns may contribute little to the diagnosis.93, 99 The degree and pattern of fever may be difficult to rely on because of the widespread indiscriminate use of antipyretics and the problem of pre-treatment with antibiotics which, singly or together, are common in the tropics and can distort the clinical presentation of severe
bacterial infections, especially meningitis. 93, 98
The acutely ill, toxic appearance may be a feature of any number of acute febrile
illnesses.8,39,44,92,93 As a result of this non specificity, an acutely ill appearance or the absence of localizing signs may not differentiate between LF, malaria, bacteraemia or even meningitis.
In the tropics, a presumptive diagnosis of malaria could be made in the presence of CAWF and no localising signs of infection including typical signs of meningeal irritation. Several studies reveal the inaccuracy of this practice.93 For example Hendrickse et al 39 reported that the clinical diagnosis of malaria is unreliable,although, a significant association between the combination of fever, convulsions, hepato-splenomegaly and acute malaria with increasing age beyond infancy was observed. Also Sowunmi et al 100, in a review of 1345 infants, with acute fever in whom a presumptive diagnosis of malaria was made, reported that only 30% had malaria parasitaemia. Even at periods of peak malarial transmission approximately 50% of presumptive malaria diagnoses were inaccurate.100 This correlates well with the findings of Hendrickse and his colleagues who reported
40
malaria parasitaemia in only 36.8% of 500 children clinically diagnosed as having malaria.29
A case has, therefore, been made for the laboratory confirmation of malaria parasitaemia as a necessity for making a diagnosis of malaria. 39, 101 Those against this view argue that the presence of parasitaemia is not synonymous with clinical malaria. 93 An added complexity is the reported co-morbidity of malaria with bacteraemia observed in some studies.93, 98 Clearly, a confirmed diagnosis of malaria would not rule out bacteraemia or other differentials at that.
No combination of risk factors (clinical / laboratory) has been found to be completely reliable in predicting the presence of occult bacteraemia.8,39,97 Bacteraemia frequently follows URTI and diarrhoea. 93,97,102 Malaria 10,39,93 and meningitis 13,48,93, like occult bacteraemia, are associated with diarrhoea, vomiting, convulsions and cough. These symptoms may also be the clinical presentation of LF. 22, 90
III. LABORATORY PARAMETERS IN DIFFERENTIATING BETWEEN MALARIA,