During 2010, we identified an outbreak and several sporadic cases of encephalitis caused by NiV infection. Two case-patients from the outbreak and 3 patients with sporadic cases had IgM against NiV in serum and NiV RNA in oropharyngeal swab samples by conventional and real-time reverse transcription PCR (14). We could not collect biological specimens from 3 probable outbreak case-patients, including the source case-patient; however, the onsets of illness of patients with confirmed and probable cases were within 3 weeks of each other in an area where NiV outbreaks have been repeatedly confirmed over the past decade (2,3,7–10,15,16). Clinical features of fever, evidence of brain involvement, and rapid progression to death were also consistent with previous NiV outbreaks (2,3,8–10,12).

The first 4 case-patients of the initial phase of the outbreak, and 6 of 8 of the patients with sporadic cases, apparently contracted NiV infection by drinking raw date palm sap contaminated with NiV by Pteropus bats, an exposure that has been linked to NiV infection in previous outbreaks (3,17).

The remaining 4 case-patients from the outbreak probably acquired NiV infection from physical contact with the source case-patient. Such person-to-person transmission has been observed in prior NiV outbreaks in Bangladesh (7–10). The 2 generations of transmission are reflected in the 2 peaks in the epidemiologic curve (Figure 2). Among the second-generation cases, a novel finding was the transmission of NiV from the corpse of the source case-patient to 2 persons who had contact with the corpse before burial. This is the most plausible transmission pathway, because they did not have known exposures to living persons with encephalitis and had no history of drinking raw date palm sap. Because NiV is found in the respiratory secretions of NiV case-patients (14), case-patient G may have had intimate hand and facial contact with the corpse’s respiratory secretions while performing ritual purification. Consistent with the culturally prescribed method of ritual bathing of a corpse, case-patient H did not wear a mask or gloves during cleansing of the corpse’s orifices. He only used 3 pieces of cloth and his bare hands, which then were almost certainly contaminated with NiV. Case-patient H also likely touched his face or nose during or after the ritual purification. Persons commonly touch their own faces subconsciously, and 1 videotaped observational study found that persons touched their own eyes, nostrils, and lips 16 times per hour during normal activities (18). During Muslim ritual bathing, water is poured on the body (19). Thus, the water may have become contaminated with NiV and came in contact with case-patient H’s clothes and body. Similar to other infectious diseases, including severe acute respiratory syndrome and measles, the transmission efficiency of individual NiV case-patients varies (17,20,21). Case-patient A was an unusually efficient spreader of NiV, perhaps because of an unusually high concentration of NiV in his oral secretions.

The dead bodies of all NiV-infected patients who are Muslim in Bangladesh have undergone the same process of ritual bathing, but to our knowledge, corpse-to-human transmission has not been previously. In other NiV outbreaks when NiV infection developed in family members, many persons had contact with the source case-patient during illness and when preparing the corpse, so we were unable to separately assess corpse-to-person transmission. This investigation suggests that occasional NiV transmission could occur during the Muslim ritual purification of a corpse before burial.

This study also documents the death of a physician in Bangladesh from NiV encephalitis after he cared for NiV-infected patients with encephalitis in the surveillance hospital. The physician’s colleagues and roommates did not report any history of his drinking raw date palm sap during the month preceding onset of illness. Although the physician had contact with oral secretions of several meningoencephalitis patients during the outbreak, the genetic sequence of NiV found in the physician was distinct from those of 2 hospitalized NiV-infected case-patients who were positive for NiV by reverse transcription PCR (14). Indeed, none of the 3 hospitalized patients with confirmed NiV infection was likely to have been the source of the physician’s infection. The duration between onset of illness of the physician and his contact with confirmed NiV case-patients was beyond the range of the 6- to 11-day incubation period for NiV (12,17). During the assumed time of exposure to NiV, he cared for patients in the adult medicine ward; some of them may have had NiV infections that were missed by hospital surveillance. However, we did not identify any patient who met the case definition for meningoencephalitis in that ward 6–11 days before onset of the physician’s illness. The clinical spectrum of human NiV infection in Bangladesh also includes patients who sought treatment with respiratory disease as the primary manifestation (12), and surveillance may have missed any NiV-infected persons on the ward with this clinical manifestation. Another line of evidence suggests that an unidentified NiV-infected patient was hospitalized on that adult medicine ward at that time. One patient with a sporadic case, who visited FMCH as a family caregiver, also provided care for several patients in the men’s medicine ward during the same days that the physician attended to patients on that ward. This case-patient may have come in close physical contact with the same unidentified NiV-infected case-patient as the physician.

During 2001, health care workers were infected by NiV in Siliguri, India. Among 66 infected persons, 45 case-patients were hospital staff or family caregivers attending to the patients, and 11 patients were infected from an unidentified, hospitalized index case-patient (22). However, during an NiV outbreak in Bangladesh in 2004, health care providers (using minimal personal protective equipment [PPE] and with substantial exposure to NiV case-patients) had no evidence of having acquired NiV infection (11). During 2010 in Faridpur, NiV was transmitted from person to person in community and hospital settings. The observed differences in risk for person-to-person transmission between outbreaks suggest that NiV strains may differ in their proclivity for person-to-person transmission.

Because NiV infection is not the major cause of acute meningoencephalitis in Bangladesh, and because most persons who contract NiV infection have died by the time a diagnosis is made, it is difficult to identify a strategy to prevent person-to-person transmission that could be consistently applied to NiV-infected case-patients. Strategies to reduce care providers’ exposure to respiratory secretions could prevent a broad array of saliva-transmitted infections, including NiV encephalitis. Prevention approaches to reduce corpse-to-person transmission of NiV and other potentially fatal respiratory secretion-transmitted viruses should focus on minimizing exposure to saliva and other bodily fluids from the body of a person who died of severe febrile illness. Wearing gloves and a mask during the handling and washing of a dead body before burial would not be feasible in low-income communities, where the annual total per capita spending on health is US $12 per person per year (23). Research to identify culturally acceptable cost-effective approaches that can be consistently implemented in low-income settings, for example, washing hands thoroughly with soap and water immediately after corpse contact, could save lives.

This report of nosocomial transmission of NiV to a health care worker in Bangladesh after caring for NiV-infected patients highlights the risk of working without PPE. Barriers to developing an appropriate prevention strategy for nosocomial transmission of NiV in hospitals in Bangladesh include the following: inadequate supplies of PPE for hospital staff, absence of isolation wards, absence of handwashing facilities in hospital wards and physicians’ rooms, and inadequate training and monitoring for infection control (24). Because saliva is the most likely vehicle for transmission of NiV among care providers, implementation of standard and contact precautions (25) that have been culturally and economically customized to fit this setting could reduce NiV transmission. As a first step, we recommend that handwashing stations be established and consistently supplied with soap and water in every ward of the hospital for health care workers and patient attendants. Second, because laboratory diagnosis for NiV infection is not available during the initial evaluation of patients with meningoencephalitis syndrome, during NiV season all hospitals in NiV infection–prone areas should admit patients with meningoencephalitis syndrome into an isolation room or ward and routinely provide gloves and masks for health care workers when they are caring for meningoencephalitis patients. Patient attendants could reduce their exposure to patient saliva and respiratory secretions by frequent handwashing and by avoiding sharing food and beds with patients.