Abstract
Local prevalences of individual diseases influence the prioritization of the differential diagnoses of a clinical syndrome of acute undifferentiated febrile illness (AFI). This study was conducted in order to delineate the aetiology of AFI that present to a tertiary hospital in southern India and to describe disease-specific clinical profiles. An 1-year prospective, observational study was conducted in adults (age >16 years) who presented with an undifferentiated febrile illness of duration 5–21 days, requiring hospitalization. Blood cultures, malarial parasites and febrile serology (acute and convalescent), in addition to clinical evaluations and basic investigations were performed. Comparisons were made between each disease and the other AFIs. A total of 398 AFI patients were diagnosed with: scrub typhus (47.5%); malaria (17.1%); enteric fever (8.0%); dengue (7.0%); leptospirosis (3.0%); spotted fever rickettsiosis (1.8%); Hantavirus (0.3%); alternate diagnosis (7.3%); and unclear diagnoses (8.0%). Leucocytosis, acute respiratory distress syndrome, aseptic meningitis, mild serum transaminase elevation and hypoalbuminaemia were independently associated with scrub typhus. Normal leukocyte counts, moderate to severe thrombocytopenia, renal failure, splenomegaly and hyperbilirubinaemia with mildly elevated serum transaminases were associated with malaria. Rash, overt bleeding manifestations, normal to low leukocyte counts, moderate to severe thrombocytopenia and significantly elevated hepatic transaminases were associated with dengue. Enteric fever was associated with loose stools, normal to low leukocyte counts and normal platelet counts. It is imperative to maintain a sound epidemiological database of AFIs so that evidence-based diagnostic criteria and treatment guidelines can be developed.
Introduction
Acute undifferentiated febrile illnesses (AFIs), such as malaria and dengue, cause considerable morbidity, mortality and economic burden to developing tropical nations. 1 Given the clinical confusion in distinguishing between AFIs, inappropriate use of antibiotics is rampant, frequently corroborated by improperly interpreted tests. 2 Evidence-based decision-making relies on quality information about the epidemiology of region-specific AFIs, of which very little is from South Asia. 1 This study aimed to delineate the regional aetiology of AFI and describe disease-specific profiles that would help clinicians reach diagnoses based on simple clinical evaluation.
Methodology
A prospective, observational study was conducted, following approval of the institutional review board, in a tertiary-care referral hospital (Vellore, South India) during January 2007–January 2008. Consecutive patients aged ≥16 years who had had a febrile illness for 5–21 days, with no evident focus of infection following initial clinical evaluation and who required hospitalization, were recruited after informed consent (study protocol – Figure 1). Immunocompromised patients were excluded. Diagnoses were assigned according to predefined criteria. Odds ratios were derived for clinical features associated with a given AFI compared to the remaining cohort.
Acute undifferentiated febrile illness
Results
Significant parameters on univariate and multivariate analysis for scrub typhus, malaria, dengue and enteric fever*
*Logistic regression models were not created for the small patient cohorts with leptospirosis, spotted fever and Hantavirus infection
†Adult respiratory distress syndrome (bilateral pulmonary infiltrates on chest X-ray; peak flow ratio < 200; normal central venous pressure)
‡The diagnosis of aseptic meningitis was assigned to patients diagnosed with scrub typhus in whom cerebrospinal fluid (CSF) analysis revealed a lymphocyte predominant pleocytosis (CSF leucocytes >5) and negative CSF cultures. Lumbar puncture was performed on patients with neck stiffness, altered sensorium or seizures, provided there were no contraindications for the procedure
AFI, acute undifferentiated febrile illness; SD, standard deviation; S. ALT, serum alanine transaminase
Discussion
AFI disease burden in South Asia
Although rickettsial fevers are being increasingly reported from the Indian subcontinent, the incidence is unknown. 3 Our study showed a high scrub typhus proportion (47.5%), probably due to an epidemic which occurred during the study period, a low disease awareness and, consequentially, a higher referral rate. Up to 80% of reported malaria cases in southern/south-eastern Asia are from India, with the majority from states such as Orissa and Andhra Pradesh. 4 Malaria accounted for 17.1% of AFIs in our study. Dengue fever incidence has been estimated at 14% among AFIs in a rural population-based southern Indian study and 48% in a hospital-based study in urban northern India. 2 The study dengue case numbers, though low (7%), comprised severe cases as evidenced by the high case fatality rate (25%), possibly due to a referral bias. Salmonella, the most common bloodstream bacterial infection in southern Asia, 1 accounted for a tenth of AFIs in a north Indian study 2 – similar to our cohort (8%). Incidence rates for leptospirosis in our study were lower, with predominantly milder non-icteric forms, than that observed in centres with higher rainfall. The sero-prevalence of Hantavirus in South India is documented, though the incidence of clinical disease is unclear. 5
Respiratory disease
Respiratory symptoms, signs and abnormal chest radiography were the most common in patients with scrub typhus. Pulmonary involvement, commonly interstitial pneumonitis with possible vasculitis, leading to acute respiratory distress syndrome (ARDS), occurs in up to 55% of scrub typhus patients. 6 Scrub typhus, malaria and dengue contributed 75.8%, 9.7% and 2.9%, respectively, of all patients with ARDS in this cohort. A much higher incidence of ARDS in scrub typhus (24.9%) was documented in our cohort than previously reported. 6 Falciparum malaria associated ARDS is documented in 2.1–11.4% of Indian in-patients, the risk being higher among pregnant and non-immune individuals. 7 The pathophysiology of ARDS in malaria and dengue occurs as a result of endothelial injury, increased alveolar permeability and fluid overload.
Hepatic and renal disease
The predilection of Orentia tsutsugamushi for the liver sinusoidal epithelial cell results in mild elevations in hepatic transaminase levels in the majority of patients (70.1% in our study), with relatively mild elevations in alkaline phosphatase and bilirubin. Hepatic injury in malaria causes marginal rises in hepatic transaminases with significant mixed hyperbilirubinaemia due to intravascular haemolysis, hepatocyte dysfunction and bile stasis. In contrast, studies (including ours) have shown that significantly elevated hepatic transaminase levels are common in dengue infections. Normal serum aspartate transaminase (AST) levels are a strong negative predictor for dengue haemorrhagic fever (DHF). 8 Renal failure was seen most commonly in falciparum malaria (38.2%) followed by scrub typhus (19.6%), dengue (17.9%) and leptospirosis (16.7%).
Haematological involvement
Leukocytosis is seen in scrub typhus and leptospirosis, though it is not an invariable feature of scrub typhus. Normal/low leukocyte counts are evident in malaria, dengue and enteric fever. Thrombocytopenia is integral to the presentation of malaria, with up to 70% of patients with falciparum malaria exhibiting this. 9 Marked thrombocytopenia, overt bleeding and haemoconcentration secondary to plasma leak favour DHF/dengue shock syndrome (DSS). Thrombocytopenia in scrub typhus is generally mild.
Central nervous system (CNS) involvement
In this study, 74.6% of the patients with aseptic meningitis and 80% of patients with seizures had scrub typhus. Altered sensorium, including coma, mainly occurred in scrub typhus (53.6%) and falciparum malaria (18.8%). Cerebral malaria, documented in up to 70% of complicated falciparum malaria cases, was uncommon in our cohort. CNS involvement, commonly encephalitis presenting with altered sensorium and seizures, has been documented in 1–25% of dengue admissions. 10 In our study, 7.1% of dengue cases had aseptic meningoencephalitis.
Limitations
The majority of our patients presented late and required hospitalization due to multisystem involvement or complications. Extrapolating this data to patients with mild, shorter duration AFI in the community would be inaccurate.
Conclusion
Scrub typhus contributes a significant, hitherto under-recognized, disease burden in southern India. Respiratory manifestations, including ARDS, aseptic meningitis, mildly elevated hepatic transaminases and leukocytosis, characterize scrub typhus. Eschar detection and a therapeutic response to Doxycycline clinch the diagnosis. A hepato-renal syndrome constituting mixed hyperbilirubinaemia with marginally elevated hepatic transaminases, splenomegaly, renal failure and thrombocytopenia suggests malaria. Dengue, especially DHF/DSS, is characterized by marked thrombocytopenia, leukopenia, high transaminases and overt bleeding. Loose stools with low/normal leukocyte counts suggest enteric fever. Spotted fever (in patients with rash), anicteric leptospirosis and Hantavirus infection are important considerations in South India. Region-specific epidemiological databases of AFI need to be created so that evidence-based diagnostic criteria and treatment guidelines can be developed.