Through the community surveys in 2010 and 2013, we enrolled 264 households (254 unique households and 10 that participated in both years) (Table 1). Each enrolled household had a main household respondent; 87 persons from enrolled households participated as additional household respondents. Most of the 2013 respondents also participated in the 2013 follow-up survey (172/217 [79%] from the 2013 community survey and 18/21 [86%] from the 2013 bat hunter survey).

More than one quarter of enrolled households (72/264; 27%) had ≥1 household member who had ever participated in the bat festival (Table 1). Almost two thirds of enrolled households (168; 64%) had >1 household member who had ever had bat contact. Nearly two thirds of enrolled households (166; 63%) had >1 household member who had ever touched a bat. About one fifth of households had >1 household member who had ever been bitten (44; 17%) or scratched (56; 21%) by a bat. Nearly three quarters of households had >1 household member who had ever eaten a bat (188; 71%).

The enrolled households were composed of 2,112 persons, among whom 213 (10%) were reported to have ever had bat contact, 211 (10%) to have ever touched a bat, 52 (2%) to have ever been bitten by a bat, 66 (3%) to have ever been scratched by a bat, and 265 (13%) to have ever eaten a bat (Table 2). Of 254 main household respondents, 141 (56%) reported having ever had bat contact (Table 3). Factors significantly associated with bat contact included being male (OR 2.08, 95% CI 1.24–3.49), having ever participated in the bat festival (OR 20.17, 95% CI 6.09–66.82), having ever entered a bat cave or bat refuge (OR 31.45, 95% CI 7.45–132.73), having ever prepared a bat as food (OR 9.85, 95% CI 5.37–18.07), and having ever eaten a bat (OR 8.56, 95% CI 4.57–16.03).

Although more than half of participants with bat contact in the 2010 community survey, 2013 community survey, and 2013 bat hunter survey knew that animal bites are a mechanism of rabies virus transmission or that rabies is severe, they more often attributed dogs as being a rabies source (≥60%) than bats (≤3%) (Appendix 2 Table 1). About 50% of participants with bat contact in the 2010 and 2013 community surveys and 86% of participants in the 2013 bat hunter survey stated that they would do nothing if bitten or scratched by a bat. Among participants with bat contact in the 2010 community survey, 2013 community survey, and 2013 bat hunter survey, only 1%, 2%, and 5%, respectively, had ever received rabies vaccination. Furthermore, only 3%, 7%, and 5% of these participants, respectively, were aware that bats can cause diseases other than rabies.

More main household respondents with bat contact knew that animal bites are a mechanism of rabies virus transmission and that rabies is severe compared with those without bat contact (Table 3). However, knowledge about bats as a potential rabies source was low and not different among main household respondents with and without bat contact. There was no significant difference between main household respondents with and without bat contact regarding history of rabies vaccination and awareness that bats can cause diseases other than rabies. Study participants with bat contact in the 2010 community survey, 2013 community survey, and 2013 bat hunter survey infrequently reported knowledge of any illness as a result of bats or being in a bat cave (1%, 3%, and 0%, respectively) (Appendix 2 Table 1).

Among 170 main household respondents and additional household respondents in the 2013 community survey who participated in the 2013 follow-up survey, 23 (14%) had experienced a febrile illness within 90 days of the February 19, 2013, bat festival (Table 4). Factors such as having had any bat contact within the past 90 days, having touched a bat within the past 90 days, having been bitten by a bat within the past 90 days, having been scratched by a bat within the past 90 days, having participated in the bat festival within the past 90 days, and having entered a bat cave or bat refuge within the past 90 days were not significantly different between those with a febrile illness and those without.

Among 18 participants from the 2013 bat hunter survey who participated in the 2013 follow-up survey, 7 (39%) had experienced a febrile illness within 90 days of the February 19, 2013, bat festival. Mean age was significantly higher among those with a febrile illness compared with those without (61 years vs. 49 years; p = 0.048). The odds of having entered a bat cave or bat refuge within the past 90 days was significantly higher among those without a febrile illness compared with those with a febrile illness (p = 0.03). There were no other significant differences between those with a febrile illness and those without when analyzing the same characteristics (Table 4).

Of all study participants who underwent serologic testing, only 2 had lyssavirus neutralizing antibodies, both against rabies virus (Appendix 2 Table 2). Both denied recent encephalitis-like illness or having ever received rabies vaccine, but 1 reported prior bat contact. One of these respondents underwent repeat serologic testing for rabies virus neutralizing antibodies during the 2013 follow-up survey, and rabies virus neutralizing antibodies were still detectable.

We sampled 211 bats: 120 bats during September 2010 (112 Rousettus aegyptiacus, 8 Hipposideros gigas) and 91 during February 2013 (all R. aegyptiacus); none demonstrated clinical illness at time of capture. No R. aegyptiacus bats had neutralizing antibodies to Duvenhage virus; >50% had neutralizing antibodies to Lagos bat virus, Shimoni bat virus, and Mokola virus; and 1 had neutralizing antibodies to Ikoma lyssavirus (Table 5; Appendix 2 Table 3). Lyssavirus antigens were not detected in brain specimens from any of the 211 bats.

The occurrence of purposeful human interactions with bats, such as hunting for food (e.g., bushmeat), has been identified in several parts of the world and can pose a risk to human health through spillover of zoonotic pathogens from bats to humans (29–31). We therefore investigated bat and lyssavirus exposures among humans in an area of Nigeria that celebrates a biannual bat festival. Overall, we found that persons who interact with bats in this area are likely at risk for phylogroup II lyssavirus exposures, and public health precautions are warranted.

Although nearly two thirds of households enrolled in our study had >1 household members who had ever had bat contact, only about one quarter of households reported having >1 household members who had ever participated in the bat festival. This finding strongly suggests that a sizable proportion of the human population in the area has had bat exposures unrelated to the bat festival. Furthermore, 10% of persons living within households enrolled in our community surveys had previously had bat contact and 2% had been bitten by a bat. We do not know whether the bat contact and bat bites among these persons are related to participation in the bat festival or to interactions with bats from the festival caves. Because entry into the festival caves is allowed only during the bat festivals, we suspect that many of these persons have had interactions with bats that are not from the festival caves. Regardless, these person-level data on the prevalence of bat contact and bat bites are likely an underestimate of the true prevalence of bat contact and bat bites in the area; persons with a history of bat interactions might not have been available or were not referred by other household members so they were not enrolled in the study, or persons who have had such bat interactions might have failed to report them when responding to the survey.

We also found strong serologic evidence that lyssaviruses circulate among bats in the festival caves. We found neutralizing antibodies to Lagos bat virus, Shimoni bat virus, and Mokola virus in >50% of bats, which is higher than in some prior reports (17,32–34). All 3 of these lyssaviruses belong to phylogroup II. We did not detect lyssavirus antigen in brains of any seropositive bat that we captured, suggesting that these bats survived past exposure to a phylogroup II lyssavirus. We cannot be sure which phylogroup II lyssavirus predominantly circulates in this bat population, given potential serologic cross reactivity and because we did not isolate any lyssavirus from bats. However, we suspect Lagos bat virus because it has been documented in R. aegyptiacus bats before and because it was first isolated in a fruit bat in Nigeria, although we cannot rule out the possibility that a yet uncharacterized phylogroup II lyssavirus circulates among these bats (18,35).

Although some respondents reported a febrile illness after the 2013 bat festival, this finding was not associated with having recent bat contact or recent participation in the bat festival. We recommend caution in interpreting these findings. A variety of bat species, including R. aegyptiacus, which we identified in the festival caves, are known reservoirs for a range of potential pathogens, including filoviruses and coronaviruses (18,22,36,37). It is therefore plausible that at least some zoonotic pathogens are present in bats residing in the festival caves and that these pathogens can spill over into humans (16). Furthermore, the data we present on febrile illness are a snapshot from 2013, and given that excretion of virus in bats can be episodic, the risk of batborne infections may vary over time (23).

We did not find neutralizing antibodies to lyssaviruses in any person in the study, other than 2 persons who had neutralizing antibodies to rabies virus, perhaps reflecting prior rabies vaccination that was not recalled during the survey or abortive infection from bites of rabid dogs (5). Thus, we found no evidence of abortive phylogroup II lyssavirus infections among humans in this study, despite the high prevalence of neutralizing antibodies to phylogroup II lyssaviruses among bats in the festival caves and that many persons in the area frequently interact with bats. This result is perhaps not surprising. First, as previously explained, we suspect that many interactions with bats among the population are unrelated to the bat festival and unrelated to bats from the festival caves (although bat hunters who participated in the 2013 bat hunter survey, by definition, would have had interaction with bats from the festival caves). The data we present on the prevalence of neutralizing antibodies to phylogroup II lyssaviruses among bats are specific to bats from the festival caves and cannot be generalized to other bat populations in the area; the prevalence of these antibodies in other bat populations with which humans also interact might be lower than that for bats from the festival caves. Second, in the Amazon, where abortive lyssavirus infections have been documented, humans likely experience bat bites on a more continuous basis because of the predatory nature of vampire bats (5). In contrast, the bat festival in this part of Nigeria occurs at discrete times, leading to a lower frequency of bat bites and thus lower risk of lyssavirus exposure. Finally, the dates of the bat festivals vary each year and are determined based on traditional wisdom. Whether the bat festival timing, as determined by cultural leaders, implicitly accounts for periods of lower risk of batborne infections to festival participants warrants further investigation by an interdisciplinary team of biologists and anthropologists (23).

Our study has limitations. Accurate information on the distribution of communities in the area was limited, making it unclear whether persons we enrolled are representative of the area. We did not use a strict definition for febrile illness, nor could we verify the occurrence of a febrile illness; rather, we relied on retrospective, subjective reports. Our study did not have a robust method of identifying encephalitis-like illness and deaths that occurred between the initial data collection in 2013 and the 2013 follow-up survey, and we do not know what happened to participants who could not be located for the follow-up survey. Thus, we cannot draw conclusions on the ability of the predominant phylogroup II lyssavirus that circulates among bats in the festival caves to cause productive lyssavirus infections (rabies) in humans.

Emerging infectious diseases are on the rise around the world; most originate from animals (38). Although the source of the 2014–2016 Ebola outbreak remains unknown, it may have begun with a single spillover event involving initial bat contact (39), which underscores the health risks of interacting with bats without appropriate precautions. If we assume that the households we enrolled are representative of the Idanre area, then this part of Nigeria has high rates of bat contact and is at high risk for bat-related zoonoses. We therefore recommend that officials strengthen health security in the Idanre area, recognizing that an approach that bans hunting and consumption of bats is unlikely to be effective. Rather, a more productive approach will focus on harm reduction and community engagement. Specific recommendations include educating the population, particularly persons who participate in high-risk bat-related activities, about the health risks associated with bats and the ecosystem benefits provided by bats; providing preexposure prophylaxis for rabies and possibly other batborne disease (potentially even Ebola) for persons who participate in high-risk bat-related activities; and developing surveillance and outbreak response capacity in the area for syndromes such as febrile illness, encephalitis, and hemorrhagic fevers.