Leptospirosis-associated Severe Pulmonary Hemorrhagic Syndrome, Salvador, Brazil

We report the emergence of leptospirosis-associated severe pulmonary hemorrhagic syndrome (SPHS) in slum communities in Salvador, Brazil. Although active surveillance did not identify SPHS before 2003, 47 cases were identified from 2003 through 2005; the case-fatality rate was 74%. By 2005, SPHS caused 55% of the deaths due to leptospirosis.

We report the emergence of leptospirosis-associated severe pulmonary hemorrhagic syndrome (SPHS) in slum communities in Salvador, Brazil. Although active surveillance did not identify SPHS before 2003, 47 cases were identifi ed from 2003 through 2005; the case-fatality rate was 74%. By 2005, SPHS caused 55% of the deaths due to leptospirosis.
L eptospirosis, a spirochetal zoonotic disease, is increasingly recognized as an important cause of hemorrhagic fever (1)(2)(3)(4). The classic presentation of severe leptospirosis, Weil's disease, is characterized by jaundice, acute renal failure, and bleeding. However, the 1995 Nicaragua outbreak raised awareness for leptospirosis as the cause of a severe pulmonary hemorrhagic syndrome (SPHS) (5). This syndrome, fi rst identifi ed in South Korea and People's Republic of China (6), is now reported worldwide (2). SPHS is associated with fatality rates of >50% and in certain settings, has replaced Weil's disease as the cause of death among leptospirosis patients (7)(8)(9)(10).
The factors responsible for SPHS and its emergence are not well understood. In Brazil, outbreaks of leptospirosis occur annually in slum communities during seasonal periods of heavy rainfall (11). Although SPHS is a frequently observed manifestation of leptospirosis in Rio de Janeiro and São Paulo (7,8,12), the occurrence of SPHS varies according to geographic region. In the city of Salvador (population 2.7 million), active surveillance did not detect SPHS among 1,786 leptospirosis cases identifi ed from 1996 through 2002 (unpub. data). However, in 2003 we identifi ed a patient in whom massive pulmonary hemorrhage and acute respiratory distress syndrome developed; subsequently, the number of cases with similar manifestations unexpectedly increased. We report the investigation of the emergence of SPHS in a setting where it was not previously observed.

The Study
The Oswaldo Cruz Foundation and State Secretary of Health of Bahia have conducted active surveillance for leptospirosis since March 1996 in the metropolitan region of Salvador. According to health secretary protocols, suspected cases are referred to the state infectious disease hospital. The study team consecutively identifi ed case-patients who were admitted to this hospital and met the clinical defi nition for severe leptospirosis (11). Study participants were enrolled according to informed consent protocols approved by the Oswaldo Cruz Foundation and Weill Medical College of Cornell University. After the fi rst case was identifi ed on May 1, 2003, the study team prospectively identifi ed SPHS from October 2003 through December 2005 by evaluating patients 5 days a week for fi ndings of massive pulmonary hemorrhage (hemoptysis >300 mL or aspiration of fresh blood after endotracheal intubation, which did not clear with suctioning) and respiratory insuffi ciency (respiratory rate >30 per min or use of supplemental oxygen therapy). Medical records of patients hospitalized between January 2000 and September 2003 were reviewed to identify cases that may not have been previously recognized.
A data-entry form was used to extract information from medical charts. Patients or family members were interviewed to obtain information on demographics and risk exposures in the household and workplace. Blood samples were collected during hospital admission, on day 4 or 5 of hospitalization, and 2 weeks after the initial sample was obtained. Laboratory-confi rmed diagnosis of leptospirosis was defi ned as a 4-fold rise in titer between paired samples or a single titer of >800 in the microscopic agglutination test (MAT) (11) or a positive result on the immunoglobulin M (IgM) ELISA (Bio-Manguinhos, Rio de Janeiro, Brazil) (13 Among the 47 SPHS case-patients, 7 (15%) and 20 (42%) had pulmonary hemorrhage and respiratory insuffi ciency, respectively, at the time of hospitalization (Table  1). Pulmonary hemorrhage was identifi ed in 19 (40%) patients only after endotracheal intubation (Table 2). Except for respiratory insuffi ciency, hemoptysis, and oliguria, SPHS case-patients had similar clinical manifestations to those of non-SPHS patients at the time of initial evaluation (Table 1). However, 7 (15%) and 16 (34%) of the 47 SPHS patients did not have signs of jaundice and acute renal insuffi ciency, respectively. Respiratory failure developed in all SPHS patients. Acute lung injury was documented in 25 (76%) of the 33 patients for whom arterial blood gas measurements were obtained (Table 2). Although patients received supportive care with mechanical ventilation (94%), dialysis (53%), and packed erythrocyte transfusion (60%), case-fatality rate for SPHS was 74% and signifi cantly high-er than that (12%) for non-SPHS leptospirosis (Table 1). By 2005, SPHS was the cause of 55% of the deaths among leptospirosis patients (Figure).
The annual incidence of SPHS was 0.43 cases per 100,000 population, based on the 32 patients who resided within Salvador. The overall incidence of severe leptospirosis was 4.65 cases per 100,000 population (341 cases). SPHS cases occurred during the same winter period of seasonal rainfall during which non-SPHS leptospirosis cases were identifi ed. All SPHS case-patients were residents of slum settlements from which non-SPHS case-patients were identifi ed before and after the appearance of SPHS. Most SPHS case-patients were adults (mean age 37.6 ± 19.4 years) and male (70%) (

Conclusions
There has been growing recognition of the importance of leptospirosis as the cause of SPHS (1-3). Yet except for a few outbreak situations (4-6), it remains unclear whether increased reporting represents enhanced detection of an under-recognized manifestation (3) or de novo emergence of this disease form. This investigation identifi ed the appearance of SPHS in a region in which urban leptospirosis is endemic and active surveillance was in place. Prior underrecognition of this syndrome was unlikely because clinicians were aware of SPHS's occurrence in other Brazilian cities (7,8,12). Laboratory confi rmation of leptospirosis was obtained for 79% of the cases, indicating that SPHS was due to this disease rather than to other causes of hemorrhagic fever. Three years after the fi rst case was identifi ed, SPHS accounted for 19% of hospitalizations and 55% of the deaths from leptospirosis.
Our fi ndings underscore the diffi culties in identifying SPHS, even in the setting of heightened surveillance. The fatality rate was 74% among SPHS patients despite aggressive supportive care. Clinical parameters that could differentiate patients at risk of acquiring SPHS were not found during initial evaluation, thereby hampering attempts to implement timely triage procedures. Finally, pulmonary hemorrhage was only identifi ed in 40% of the patients at the time of endotracheal intubation. Recognition of SPHS will therefore need to rely on a high index of suspicion in patients who have acute respiratory insuffi ciency.
That we did not identify environmental risk exposures for acquiring SPHS suggests a role for pathogen or hostspecifi c factors. In Thailand, the recent sustained outbreak of leptospirosis was due to the widespread introduction of a Leptospira clone (14). Our fi ndings suggest that the agent  for SPHS in Salvador was L. interrogans serovar Copenhageni, which was also the cause of non-SPHS leptospirosis in this setting (11,15). Further isolation studies and genotyping analyses are needed to determine whether the appearance of SPHS is due to introduction or emergence of a clone that has enhanced virulence, within this serovar. Gender-specifi c factors, whether risk activities or host-susceptibility determinants, may have contributed to acquiring SPHS because women who with leptospirosis had twice the risk for this disease form. A large proportion of the world's slum population resides in leptospirosis-endemic regions. Research is needed to elucidate the factors responsible for transmission of SPHS so that effective prevention can be identifi ed.