|Does this patient have malaria?
|Taylor SM, Molyneux ME, Simel DL, Meshnick SR, Juliano JJ
The authors concluded that splenomegaly or hepatomegaly presence increased the likelihood of malaria in patients from endemic areas, but individual clinical findings were of limited use. Clinical assessment was diagnostically useful in returning travellers, but patients still required laboratory testing. Given the potential for bias, unclear quality of the studies and variation between them, these conclusions should be interpreted cautiously.
To assess the accuracy of clinical signs and symptoms for the diagnosis of malaria (Plasmodium) in residents of, and travellers returning from, endemic areas.
MEDLINE and EMBASE were searched between 1950 and July 2010 for English language publications. Search terms were reported in an online supplement (see URL for additional data). In addition, reference lists of retrieved articles were manually searched.
Studies that used using thick or thin blood smears (diagnostic criterion standard) to detect malaria parasites in patients suspected of having acute malaria (endemic and imported) were eligible for inclusion. Eligible studies had to report sufficient details on the frequency of a number of pre-determined clinical signs and symptoms to allow the construction of 2x2 tables. Studies that reported only specific malaria presentations (e.g. malarial anaemia) were excluded, as were studies of species-specific malaria, with the exception of Plasmodium falciparum (P. falciparum) in returning travellers. Case reports were also excluded.
Most studies in endemic areas were of febrile children in sub-Saharan Africa, but some studies also included febrile adults. The prevalence of illness varied between countries. Studies of returning travellers were conducted in Africa, Asia, Latin America/Caribbean, and the Pacific. The use of malaria prophylaxis in travellers was unclear. Most infections were attributable to P. falciparum or P. vivax, but some were attributable to P. malariae or P. ovale. Symptoms included fever, vomiting, chills/rigors, nausea, headache, joint pain, dyspnoea, cough, and diarrhoea. Signs included pallor, jaundice/icterus, splenomegaly, and hepatomegaly. Laboratory findings included total bilirubin level and platelet count.
Two reviewers independently screened studies for inclusion. Discrepancies were resolved through discussion.
Assessment of study quality
The authors did not state that they assessed the quality of the included studies. However, included studies were assigned a level of evidence score according to pre-specified criteria (reported in a separate online appendix, see URL for additional data).
Two reviewers independently extracted the frequency of each clinical sign and symptom from patients in endemic area, and returning travellers and the laboratory findings for returning travellers, into 2x2 tables to estimate sensitivity and specificity, along with 95% confidence intervals (CIs). Where insufficient data were reported, primary authors were contacted. Discrepancies were resolved by consensus.
Methods of synthesis
Univariate or bivariate random-effects models were used to calculate summary likelihood ratios (LRs) for each clinical sign and symptom. Statistical heterogeneity among all studies and studies of children only was assessed using the I2 statistic.
For endemic studies, separate analyses were undertaken in adults only and children only for each sign and symptom, except for pallor, which differed substantially between adults and children, and in its definition. Summary scores for pallor were combined for both adults and children.
The c index (or area under the curve) was also used to assess the whole range of scores to identify the most diagnostically informative clinical finding.
Results of the review
Forty-seven studies were included in the review; 14 studies of endemic malaria (n=15,559 patients, range 168 to 2,984) and 33 studies of imported malaria, although only seven studies on imported malaria were included in the meta-analysis (n=34,113 patients; range 60 to 30,811). Only four studies of endemic malaria described blood smear quality control.
Patients in endemic areas : No individual symptom was shown to be useful in accurately diagnosing the presence of malaria. Clinical signs were more useful; they were the presence of splenomegaly in all participants (summary LR 3.3, 95% CI 2.0 to 4.7; eight studies), hepatomegaly in all participants (summary LR 2.4, 95% CI 1.6 to 3.6; five studies) and pallor in children only (summary LR 2.0, 95% CI 1.2 to 2.8; seven studies). There was evidence of statistical heterogeneity for all summary likelihood ratios (I2 ranged from 51 to over 70%). The absence of any signs or symptoms were less useful in predicting the absence of malaria. Three studies assessed combinations of clinical factors in endemic areas and reported different findings for different areas.
Returning travellers: Fever or a history of fever was shown to be useful in accurately diagnosing the presence of malaria (summary LR 5.1, 95% CI 4.9 to 5.3; one study; n=30,221) and the absence of fever useful in predicting the absence of malaria (summary LR 0.12, 95 %CI 0.10 to 0.15; one study). Clinical signs were more accurate in diagnosing malaria; presence of splenomegaly (summary LR 6.5, 95% CI 3.9 to 11.0; three studies), jaundice or icterus (summary LR 4.5, 95% CI 1.7 to 12.0; one study), and pallor (summary LR 2.8, 95% CI 1.7 to 4.6; one study). Total bilirubin levels exceeding 1.2mg/dL (i.e. hyperbilirubinaemia) accurately diagnosed malaria (summary LR 7.3, 95% CI 5.5 to 9.6; one study), as did platelet counts below 150x103/μL (i.e. thrombocytopenia; summary LR 5.6, 95% CI 4.1 to 7.5; three studies). Normal platelet counts or bilirubin levels were also accurate in predicting the absence of malaria. The absence of other signs and symptoms were less useful in predicting the absence of malaria.
Findings from the remaining studies (case series and prospective studies) on returning travellers were reported in the review.
The presence of splenomegaly or hepatomegaly increased the likelihood of malaria in patients from endemic areas, but individual clinical findings were of limited use and could not reliably exclude malaria; combinations of clinical findings may be useful to stratify risk in patients. Clinical assessment could be diagnostically useful in returning travellers, but all patients still required laboratory testing.
The review question was clear and was supported by broad criteria. The literature search was limited to two databases and restricted to English language publications, so potentially relevant studies may have been missed. Selection of studies and data extraction were undertaken in duplicate, minimising the potential for reviewer error and bias.
The design of the studies included in the meta-analysis was unclear; study quality was not formally assessed. However, the authors did highlight the paucity of high quality studies for returning travellers. The authors highlighted clinical heterogeneity between study populations, and there was some evidence of statistical heterogeneity. Therefore, it was unclear whether pooling of the results was appropriate. The authors acknowledged that only a small number of studies described blood smear quality control. They also acknowledged the uncertainty regarding generalisability of the findings to countries other than sub-Saharan Africa.
Although the authors’ conclusions appeared to reflect the evidence, the potential for review bias, the unclear quality of the included studies, and considerable heterogeneity among studies mean that these conclusions should be interpreted with caution.
Implications of the review for practice and research
The authors did not state any implications for research or practice.
One author was supported by National Institutes of Health, National Centre for Research Resources, grant number KL2RR025746.
Taylor SM, Molyneux ME, Simel DL, Meshnick SR, Juliano JJ. Does this patient have malaria? JAMA 2010; 304(18): 2048-2056
Subject indexing assigned by NLM
Adult; Child, Preschool; Endemic Diseases; Female; Hepatomegaly /etiology; Humans; Malaria /complications /diagnosis /epidemiology; Predictive Value of Tests; Splenomegaly /etiology; Travel
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.