The study was based in Semarang, a large harbor city in central Java. Consecutive outpatients were recruited at 2 primary healthcare centers and hospitalized patients at a governmental referral center (Dr. Kariadi University Hospital, Department of Internal Medicine). All eligible AUF patients (>5 years of age) were included who met the following criteria: fever >38°C (central) for <14 days with no apparent other disease. After informed consent was obtained, a blood sample was taken. A convalescent-phase sample was drawn after ≈14 days. The study was approved by the local medical ethical committee.

A specific microimmunofluorescent antibody (IFA) assay for Rickettsia spp. was performed in Marseille, France, by using whole-cell antigens of O. tsutsugamushi, R. japonica, R. heilongjiangensis, R. slovaca, R. honei, R. conorii subsp. indica, Rickettsia ATI, R. helvetica, R. felis, R. typhi, and R. prowazekii . The assay results were considered positive when 1) antibody titers were >256 for immunoglobulin (Ig) G and >64 for IgM, or 2) seroconversion was observed, or 3) a >4-fold increase in titers between the acute-phase and the convalescent-phase serum specimen was detected. Serologic analysis for leptospirosis was performed in Semarang, Indonesia. Crosschecks and PCR were performed in Amsterdam, the Netherlands. Convalescent-phase samples were screened with the LeptoTek Dri Dot (bioMérieux, Marcy l’Etoile, France). All positive samples were tested by the microscopic agglutination test (MAT) and IgG ELISA (8).

Additionally, a real-time PCR specifically targeting the secY gene of pathogenic Leptospira spp. (9) was performed on all samples. For the MAT, a panel of 31 serovars was used containing 28 pathogenic and 3 nonpathogenic serovars. For patient samples tested by ELISA or MAT, a titer >320 on a single sample was considered positive; also considered positive were those samples that showed seroconversion or a >4-fold increase in titers between paired samples, as well as any patient sample with a positive PCR, irrespective of the serologic result. All samples were run in parallel.

From February 2005 through February 2006, 137 AUF patients were included: 67 hospitalized patients and 70 outpatients. A convalescent-phase sample was available for 106 (77%) patients. The main symptoms were headache (85%), myalgia (70%), nausea (64%), cough (44%), and abdominal pain (38%).

Murine typhus and leptospirosis were found to cause AUF in this clinical series (Appendix Table). In total, 9 patients (7%) had evidence of an acute infection with R. typhi; none showed a rash. Murine typhus could be diagnosed in 6 (9%) of 67 hospitalized patients; 3 (4%) of 70 outpatients had acute murine typhus. Another 9 (7%) patients showed inconclusive R. typhi serologic results. One patient showed evidence of a past infection with R. typhi (IFA IgG/IgM titer 128/0 in both serum specimens). Evidence for acute infection with O. tsutsugamushi or SFGR was not found.

Leptospirosis was diagnosed in 13 (10%) of 137 patients; results for 2 of these patients were positive only by PCR. Eleven leptospirosis patients were recruited in the hospital; 2 patients were recruited outside the hospital. Consequently, the percentage of AUF caused by leptospirosis in hospitalized patients was 16% and in outpatients 3%. The most frequently identified serogroup by MAT was Bataviae (5 cases). No dual infections were detected; however, 3 (23%) leptospirosis patients showed titers in the R. typhi IFA assay.

We report that murine typhus and leptospirosis are important causes of AUF in Semarang, Indonesia. A previous study from rural Thailand identified both diseases in 2.8% and 36.9% of AUF cases, respectively (6,7). In Vientiane, the capital of Laos, R. typhi was reported to cause fever in 9.6% of investigated persons, results that closely resembling our data (10). Unfortunately, leptospirosis was not investigated.

In the present study, we expected leptospirosis to be a cause of AUF because the Dr. Kariadi University Hospital admits ≈50 severe cases each year. These cases were not included in the study because of the high clinical suspicion of leptospirosis on admission with jaundice, azotemia, and/or bleeding. A definite diagnosis of murine typhus and leptospirosis co-infections could not be made, but in 3 cases this scenario was plausible. We did not find evidence for scrub typhus, which we expected, because O. tsutsugamushi transmission occurs primarily in rural areas (11). Although SFGR have been reported in Southeast Asia and proof for their presence in Indonesia is accumulating (2,12), these rickettsia were not identified as a cause of AUF in the present study.

From an epidemiologic point of view, Semarang, Indonesia, seems to encompass environmental circumstances that are prerequisites for R. typhi and leptospirosis transmission. Previous studies have shown that murine typhus is particularly prevalent in tropical port cities where rats are abundant (13,14). In the Indonesian urban situation, R. rattus and R. norvegicus rats are likely to be the main hosts harboring R. typhi–infected X. cheopsis fleas (12,15). These rats are also likely to be the maintenance hosts for pathogenic Leptospira spp. in Indonesia. In fact, the identified serogroups are commonly associated with rats.

Although serologic analysis might be hampered by cross-reactions and test sensitivity issues, we believe that our data are representative for the area. The chosen cut-off values are unlikely to cause false-positive results in a disease-endemic setting. In regard to leptospirosis serologic analysis, we used a wide panel for the MAT. This panel included serovars recommended by the World Health Organization and serovars that were previously isolated in Indonesia. Moreover, most serogroups were represented in our panel, and cross-reactions are likely to detect missing serovars.

Because of nonspecific clinical features, both diseases are difficult to diagnose on clinical grounds only. Misdiagnosis can lead to aberrant use of antimicrobial drugs and other pharmaceuticals. Therefore, rapid, cheap, and reliable diagnostic tests are needed to support clinical decision making.