In 2000, the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) worked with staff from Dhaka Medical College and Holy Family Red Crescent Hospital to initiate surveillance for dengue as part of an emergency response to an epidemic of dengue and dengue hemorrhagic fever. Physicians at Dhaka Medical College and Holy Family Red Crescent Hospital were trained in the clinical diagnosis and management of dengue and dengue hemorrhagic fever according to World Health Organization guidelines. Patients hospitalized with fever and in whom a physician suspected dengue were enrolled in the surveillance program. Clinical and epidemiologic information, as well as acute-phase serum specimens, were systematically collected from surveillance patients. Acute-phase serum specimens were assessed for dengue virus antibodies by using a commercial immunoglobulin (Ig)G and IgM capture enzyme-linked immunosorbent assay (ELISA) (PanBio Dengue Duo, PanBio Ltd., Brisbane, Queensland, Australia) (5). In addition, serum samples collected from a subset of patients during the first 5 days of illness were evaluated for dengue virus RNA by using reverse transcriptase–polymerase chain reaction (RT-PCR), as described previously (6). Serum samples from patients with no dengue infection shown by antibody or RT-PCR testing were retrospectively assessed for leptospirosis by using a real-time PCR that amplifies the LipL32 gene (7), a virulence factor that is conserved among pathogenic Leptospira strains (8). Microplate Leptospira IgM ELISA testing (PanBio Ltd.) was conducted on all Leptospira PCR-positive serum specimens of sufficient quantity (9).

Specimens from 1,297 patients hospitalized at Dhaka Medical College and Holy Family Red Crescent Hospital between January 1 and December 31, 2001, were evaluated for dengue infection by using capture ELISA; 55 acute-phase serum samples were additionally evaluated by using RT-PCR for dengue. A total of 938 (72%) patients were diagnosed with dengue fever by serologic tests (932 patients0, RT-PCR (3 patients), or both (3 patients). Acute-phase serum specimens from the 359 patients without laboratory evidence of dengue were evaluated for leptospirosis; 63 (18%) had Leptospira detected by using PCR. Sixty-one of the PCR-positive samples were tested for Leptospira-specific IgM; 18 (30%) showed positive results and 5 (8%) showed equivocal results.

Patients with leptospirosis diagnosed by using PCR were 6 to 70 years of age (mean 28, SD 13); 74% were male. Patients with leptospirosis were of similar age and sex as patients with dengue (Table 1). Patients with leptospirosis had less education and came from households with lower income than patients with dengue. The peak occurrence of leptospirosis was in October and November, shortly after the monsoon season in Bangladesh. This overlapped with the period of highest dengue activity (July through December) (Figure).

Patients with leptospirosis reported a slightly longer duration of fever than patients with dengue (Table 2). While most patients with dengue or leptospirosis had continuous fever, an intermittent fever was more likely with leptospirosis. Reports of rash were more common with dengue fever. Aside from fever and rash, the symptoms of patients with leptospirosis and dengue were similar (Table 2): headache, myalgia, nausea, and vomiting were most common.

The median temperature and heart rate at physical examination were higher in leptospirosis patients than in dengue patients. Evidence of bleeding, including petechial rash, positive tourniquet test result, and gum bleeding, were more common in patients with dengue, although they were also found in some patients with leptospirosis. Subconjunctival hemorrhage, which may have been confused with conjunctival inflammation, was more commonly reported in patients with leptospirosis. Hepatomegaly and jaundice were more common in leptospirosis patients, but this difference was not statistically significant.

On laboratory examination, total leukocyte counts were similar in patients with dengue and patients with leptospirosis; however, lymphocytes were more likely to be predominant in patients with dengue. Hemoconcentration and thrombocytopenia were associated with dengue fever.

Of the patients whose outcome was known, 3 (5%) patients with leptospirosis died, compared with 11 (1.2%) patients with dengue (p = 0.048). Antimicrobial therapy for leptospirosis was not provided, and data on the cause of death were not available. The patients who died of leptospirosis were younger than those who died of dengue (24 ± 6 years of age vs. 36 ± 9 years, p = 0.05).

This is the first description of disease caused by Leptospira in urban Bangladesh. Our findings indicate that leptospirosis causes serious febrile illness in the densely populated city of Dhaka.

Studies conducted in other dengue-endemic areas have shown that leptospirosis can be confused with dengue fever (10–12). Most clinical symptoms of leptospirosis patients in Dhaka were nonspecific and not distinguishable from symptoms associated with dengue fever or other viral illnesses. Although fever in leptospirosis patients was higher and of longer duration than in dengue patients, there is sufficient overlap of clinical findings to suggest that clinicians caring for patients in Bangladesh should maintain a high index of suspicion for both diseases, especially during the peak incidence seasons that follow the monsoons. Recognition of leptospirosis is especially important since antimicrobial agents can reduce its severity and duration (13).

Leptospirosis patients identified in Dhaka were impoverished and poorly educated. This may reflect more frequent exposure to environments contaminated with urine from rodents or other animals. In contrast, dengue patients came from households with higher incomes and levels of education. Whether these socioeconomic differences reflect differing patterns of disease can only be determined by future population-based studies, which may in turn shed light on optimal prevention strategies.

Most of the patients hospitalized with leptospirosis and dengue virus infection in Dhaka during the period of this study were male. A recent seroprevalence study in Bangladesh did not demonstrate a sex difference in dengue seropositivity (14), but leptospirosis has been reported predominantly in men in other regions (15). Further prospective research would be useful to better define the clinical spectrum and gender distribution of disease in Bangladesh.

This study has a number of limitations. Only acute-phase serum samples were obtained as part of the hospital-based dengue surveillance program in Bangladesh. Serologic diagnosis of leptospirosis with a single specimen obtained early in infection is limited; hence, we used Leptospira-specific PCR for diagnosis in our study population. Although this molecular technique is highly sensitive and specific for the presence of leptospiremia (7), more cases may have been detected through the use of microscopic agglutination testing on paired serum samples. Notably, less than one-third of the patients with a diagnosis of leptospirosis by PCR had detectable levels of Leptospira-specific IgM. This is likely due to the sampling of acutely ill patients before seroconversion. In support of this, leptospirosis patients who had detectable levels of IgM exhibited a trend toward longer duration of fever compared with those who did not have detectable levels (6.5 vs. 5.5 days, p = 0.12).

Some cases of leptospirosis resulted in death, and the case-fatality rate among leptospirosis patients was significantly higher than among dengue fever patients. The different case-fatality rates may be related, however, to the intensive training in dengue case management that occurred during this epidemic period, or to the lack of specific antimicrobial therapy for unrecognized cases of leptospirosis. Our findings underscore the need for greater awareness of leptospirosis in the Indian subcontinent, more data on its incidence in Bangladesh, and optimal treatment regimens for leptospirosis that can be applied in resource-poor settings.