The team conducted this study during 2011–2014 in Rajshahi and Rangpur Districts, Bangladesh. Surveillance physicians from Rangpur Medical College Hospital and Rajshahi Medical College Hospital identified suspected case-patients (Table), recorded clinical histories and home addresses, and collected whole blood samples from each. The serum was separated from each sample, stored in liquid nitrogen at the hospital, and then transported to IEDCR (3). The laboratory team at IEDCR tested serum samples for NiV IgM and IgG by using IgM-capture and indirect IgG enzyme immunoassays (19,21).

On the basis of type of exposure, we categorized suspected case-patients as primary or secondary case-patients, and these two groups were further categorized as case-patients with probable or laboratory-confirmed NiV infection (Table). To identify a potential NiV infection cluster, surveillance physicians asked each of the admitted suspected case-patients and their caregivers present in the hospital about other sick persons or persons in their communities who had recently died with similar symptoms. We defined a NiV infection cluster as >2 suspected meningoencephalitis case-patients living within 30 minutes’ walking distance from each other who had onset of similar illnesses within 3 weeks of one another or had epidemiologic linkages to one another (22).

After identifying laboratory-confirmed NiV infection cases (Table), the team reviewed hospital records, took preliminary information from the surveillance physicians, and visited case-patients from each cluster within a month of case confirmation. We limited this study to clusters in which no case-patients had a history of drinking fresh date palm sap. In the community, the team used a structured questionnaire to interview surviving case-patients and the friends, relatives, and neighbors of deceased case-patients as proxy respondents. The team collected information related to exposure and signs and symptoms of illness to determine routes of NiV transmission for each case-patient.

The team visited the households of case-patients and used culturally appropriate approaches to build rapport and trust with the community (23). The team conducted in-depth interviews and group discussions with surviving case-patients and the family caregivers, friends, and neighbors of the deceased to explore the exposure histories of each case-patient by using an open-ended interview guide. Open-ended questions allowed the interviewers to obtain new insights about the outbreak. Good rapport with community members helped the team collect information on potentially sensitive issues, such as alcohol consumption (24), which is prohibited among the majority Muslim population of Bangladesh. The preliminary data collected during this study suggested that the case-patients might have consumed tari before their illness onset. Therefore, the team also interviewed 5 date palm sap harvesters and conducted 3 group discussions with community members to learn about tari production, consumption, and selling practices in the affected communities. The team also collected and analyzed whole blood samples from surviving case-patients and from nonpatients (defined as persons in the community who drank tari with the case-patients but did not experience any symptoms) by using the same methods described above.

Three clusters were identified, consisting of 14 case-patients (9 with confirmed NiV infection, 5 with probable NiV infection). Eight of the 14 case-patients were primary case-patients (3 with confirmed NiV infection, 5 with probable NiV infection). All 6 of the secondary case-patients had confirmed NiV infection. Among the 14 case-patients, 7 had illness onset during January–March 2011 in Rangpur District (first cluster), 3 had illness onset in February 2012 in Rajshahi District (second cluster), and 4 had illness onset during January–February 2014 in Rangpur District (third cluster). Eight drank tari before their illness onset, whereas 6 only had exposure to other case-patients. None of the primary case-patients had any history of drinking fresh date palm sap or exposure to sick humans or animals. All of the illnesses began with fever. All 8 of the primary case-patients had altered mental status followed by loss of consciousness and death. The median duration from illness onset to death was 6 days. All 6 secondary case-patients survived. The median age of all 14 case-patients was 32 years. All the primary case-patients were male, and all the secondary case-patients were female. None of the non–case-patients had NiV IgM or IgG detected in their serum samples.

Figure

Figure. Timeline of illness onset in persons with primary and secondary cases of Nipah virus infection that occurred in 3 clusters, Rangpur and Rajshahi Districts, Bangladesh, 2011, 2012, and 2014. Asterisks indicate...

In the 2011 Rangpur cluster, case-patients A, B, C, and G drank tari regularly in the evenings, and the sap to make tari was collected from a single village (Figure). Case-patients A and G were family friends and lived in the same village. Case-patients B and C were also related and lived in the same village. Case-patients D, E, and F were family caregivers of case-patient C and provided close-contact care during his illness. All 3 family caregivers experienced illness within 2 weeks of case-patient C’s illness onset.

In the 2012 Rajshahi cluster, case-patient H was a tari producer and harvester of date palm sap. Case-patient I was a neighbor and cousin of case-patient H, and both drank tari regularly in the evenings. They drank tari together from the same pot 6 days before their illness onsets. Case-patient J was a family caregiver of patient H and provided close-contact care during his illness. She experienced fever 11 days after case-patient H’s illness onset.

In the 2014 Rangpur cluster, the respondents reported that case-patients K and L were family friends and had drunk tari every day in the evenings before their illness onsets (Figure). Case-patient L became ill after 5 days after case-patient K, and case-patient L did not have any contact with case-patient K while case-patient K was ill. The tari they drank was collected from the same village where the 2011 Rangpur cluster was identified. Case-patients M and N shared the same bed with case-patient L at home; case-patient N was a family caregiver to case-patient L, providing close-contact care for him at home and in the hospital. Case-patient M had onset of fever 12 days later, and case-patient N had onset of fever 14 days after case-patient L’s illness onset.

The laboratory, clinical, and qualitative findings in this study suggest that the 14 case-patients in the 3 clusters we investigated were infected with NiV. The primary NiV case-patients identified in the clusters drank tari regularly in the evenings before their illness onsets, and none of them had a history of fresh date palm sap consumption or any exposure to other NiV case-patients, which were the main transmission pathways for NiV infection identified in previous outbreak investigations in Bangladesh (9,26). Moreover, none of the case-patients had exposure to sick animals, another possible pathway for NiV transmission reported in studies conducted in Malaysia and Singapore (27).

Tari is a date palm sap product. Because tari fermentation was a continuous process and date palm sap was fermented inside the tari pots while they were hanging in the trees, some date palm sap added to the tari might technically be fresh sap. However, the primary case-patients probably did not consume fresh date palm sap because tari was collected from 8 a.m. to noon but the primary case-patients drank tari only in the evening, which suggests that all the sap they consumed was at least partially fermented by the time of consumption. Findings from this investigation suggest that drinking tari is a potential source of NiV infection in Bangladesh. Investigators in India had similar findings during an outbreak reported in 2007 in West Bengal near the border with Bangladesh (4). They reported that drinking fermented date palm sap possibly contaminated with bats excreta and secretions was the source of NiV infection for the index patient, which further supports our assertion that NiV infection from drinking tari is plausible (4). Moreover, 2 clusters in this study were traced back to the same tari production village in Rangpur, further strengthening the conclusion that tari was the mode of transmission.

Previous studies have shown that fruit bats frequently lick the date palm sap and occasionally urinate inside collection pots (14). In our investigation, the reported evidence of bats visiting date palm trees, the presence of bat excreta inside tari pots, the reported use of the same pot for several days without cleaning, and the accumulation of sap from multiple pots into 1 pot suggest that sap is probably contaminated with bat urine or saliva during collection and fermentation. NiV can survive up to 4 days in bat urine and at least 1 day in sap contaminated with bat urine when kept at an average temperature of 19°C (28). A study to determine viability of NiV in artificial palm sap contaminated with NiV (strain Bangladesh/200401066) found no statistically significant reduction in NiV titers for at least 7 days when kept at a temperature of 22°C (29). Generally, enveloped and thus lipophilic viruses like NiV are susceptible to alcohol. A 60%–70% alcohol solution is recommended for sterilizing contaminated objects (30). A study conducted in India showed that tari derived naturally from fermenting date palm sap contains 5%–8% alcohol and has a pH of 4.5–6.0 (12). This 5%–8% alcohol concentration might not have been high enough or sufficiently distributed throughout the tari to sterilize the NiV, thus allowing persistence of viable virus and transmission of NiV to tari consumers during the winter months, when the ambient temperature ranges from 15°C to 28°C (21).

In Bangladesh, family caregivers commonly provide close-contact care to hospitalized patients (31). Infected patients often shed the virus through body secretions and excretions and can contaminate foods and surfaces, including bed rails, bed sheets, and towels (32,33). Close contact is the most likely route through which family caregivers became infected (26). Family caregivers identified in this study had direct contact with primary case-patients and their body secretions.

The findings of our study are subject to limitations. First, some participants may have been reluctant to report production, consumption, and selling of tari because these are illegal activities in Bangladesh. Initially, some of the date palm sap harvesters and tari producers were reluctant to share information with us. However, the social scientists on the investigation team built rapport and trust with the respondents, which encouraged the respondents to share sensitive information (23). Thus, the hesitance to disclose behaviors related to tari probably did not have a substantial effect on our findings. Second, no control group was available. Without a comparison group, we were unable to determine if our primary case-patients were more likely than other persons residing in these villages to drink tari. However, given the absence of evidence that the primary case-patients had other contact with bats, sick animals, or persons with NiV infection, consumption of tari appears to be the most likely transmission route. Third, we interviewed family members and friends of the deceased case-patients as proxy respondents to ascertain case exposures. However, during outbreaks of fatal diseases such as NiV infection (with a case-fatality rate >70%), there is no alternative to this approach (20). Since 2003, we have interviewed proxy respondents for case-exposures in every NiV-associated outbreak investigation conducted in Bangladesh (3,9,19). Fourth, because of the delays in investigation of the 2011 cluster, our definition for a confirmed cases of NiV infection was based on the presence of NiV IgM or IgG in serum samples. Confirming a case based on the presence of IgG is reasonable, however, because all the family caregivers from the 2011 Rangpur cluster had illness onset within 2 weeks after contact with a case-patient (19).

Because harvesting date palm sap for tari production is similar to harvesting it for consumption of fresh date palm sap, the intervention of using bamboo skirts to cover the shaved part of the date palm tree and the sap collection pots to prevent bat contact and possible NiV introduction is worth exploring. The use of bamboo skirts is already a successful, affordable, and culturally acceptable method to prevent bat access to date palm sap, and this strategy could also be used to prevent NiV transmission from tari consumption (34,35). In addition, tari harvesters from ethnic minority communities have limited access to mass media because of their ethnic, religious, and linguistic minority status in Bangladesh. Efforts should be made to raise their awareness about strategies that interrupt bat access to date palm sap. At this time, we are not aware of any studies that have tested the survival of NiV in tari. As a next step, we recommend testing NiV survival in tari at different levels of alcohol concentration.

All 3 of the clusters of NiV infection that we investigated were linked to drinking tari. Drinking tari might also be a route of exposure for other batborne viruses. A total of 55 newly described viruses from 7 virus families were recently identified in urine and saliva from Pteropus spp. bats in Bangladesh (36), suggesting that these bats could also contaminate tari with other viruses that could cause disease in humans. Date palm sap is harvested for fermentation in many areas where Pteropus spp. bats and other fruit bats are native, including Australia, Asia, and Africa (37–40). Consumers of fermented drinks and other date palm products that are harvested using similar processes as in Bangladesh might be at risk for NiV infection and other batborne diseases (12,13,36).