During March 2004–May 2007, we sampled P. hypomelanus bat colonies from several roost sites in Pulau Kapas and Pulau Perhentian on the northeast coast of Terengganu, Pulau Tioman on the southeast coast of Pahang, and Pulau Pangkor on the northwest coast of Perak. We also sampled P. vampyrus bats in the states of Perlis; Perak (Teluk, Memali, and Lenggong); Terengganu (Kampung Alor Lek, Setiu, and Kuala Berang); Pahang (Tanjung Agas, Pasir Panjang, and Ganchong); and Johor (Benut and Kesang) (Figure 1). We determined these sites using published information on pteropid roost locations (15), reports from the Malaysia Department of Parks and Natural Resources authority (PERHILITAN), information from local hunters and residents, and field observations. A complete list of observed roost sites was published by Epstein et. al. (16). Colonies were selected for this study on the basis of observed presence of bats, accessibility for bat capture, and roost size (e.g., number of bats). A sample size of 35 bats was targeted at each location or for each sampling effort to be able to detect a NiV-seropositive bat given a minimum prevalence of 10% with 10% precision at a confidence level of 95% (20).

Permission for the study was granted by PERHILITAN, and Institutional Animal Care and Use Committee approval was obtained from the Wildlife Trust Institutional Animal Care and Use Committee (New York, NY, USA). A special permit to trap and humanely kill the bats was obtained from PERHILITAN. Bats were captured nonrandomly by using 2 methods: opportunistic sampling of hunted bats (hunters were not solicited or incentivized to hunt bats for this study) and mist nets.

Hunted bats were sampled at the site of hunting, and bats were attributed to the location of the nearest known roost. Blood samples were collected by cardiac puncture, and the kidneys were harvested. The blood, urine, and kidney samples were processed as described (18). Bats captured in mist nets were extracted from the nets immediately after capture and anesthetized before sampling by using medetomidine, a combination of medetomidine/ketamine (21), or isoflurane gas. Blood (3 mL) was collected from the cephalic vein or brachial vein and placed into serum-separator tubes (Vacutainer; Becton Dickinson, Franklin Lakes, NJ, USA). Sterile cotton swabs were used to collect oropharyngeal and urogenital samples. All samples were stored in liquid nitrogen at −190°C) and transported to the Veterinary Research Institute in Ipoh, Perak, Malaysia.

For each bat hunted or captured, information on the date of sampling, species, location, sex, reproductive status, and estimated age was recorded. We assigned each bat an age category of adult (secondary sexual characteristics visible/reproductive), juvenile (no observable secondary sexual characteristics, not pregnant, dental wear characteristics), or pup (dependent and attached to its dam) (3). We categorized the reproductive status as pregnant, carrying a pup (adult females with attached pup), nursing (adult females without attached pup but with evidence of lactation), or dry (not in any of the previous categories) using published criteria (18). The total number of bats sampled for each pteropid species is shown in Tables 1 and 2. Bats captured in mist nets were released at the site of capture after recovery from anesthesia.

We performed descriptive analysis to describe the seroprevalence on the basis of species of bats and locations. Differences among positive antibody titers (range <8–1,025) between species were tested by using the Mann-Whitney U test. Comparison of the seroprevalence rates between the 2 pteropid species was performed by using the χ² test. When there were sufficient data to perform the analysis, several factors were examined for its association to NiV seropositivity by using the χ² test.

We performed a longitudinal study on a population of P. hypomelanus bats in Pulau Tioman in which 50 bats from the same colony were captured by using mist nets and sampled approximately every 6–8 weeks during January 2004–October 2006. Monsoonal rains prevented access to the island of Pulau Tioman during December–February, which extended the interval between some sampling points. Captured bats were tagged by using a thumb band or implantable microchip (Avid Identification Systems, Inc., Norco, CA, USA) and unique identification numbers to ensure that sampling was not repeated on the same bat. The bats were anesthetized by using ketamine and xylazine (22). Data, blood, and swab samples were collected as described in the previous section.

We investigated associations between serostatus (positive or negative) and each of the hypothesized risk factors (sex, age, reproductive categories, time) using the χ² test or Fisher exact test when the χ² test was not appropriate. Our previous study (18) suggested a correlation between the serologic status of pups and their dams. Therefore, serologic data for P. hypomelanus pups were excluded from further analysis. Logistic regression with generalized estimating equations was used to analyze longitudinal data to control for the effect of clustering, assuming that bats sampled between various times were from the same population. To compare the categories across sexual and reproductive maturity, we reclassified the data on the basis of sex (male and female) and reproductive maturity (juvenile, adult, dry, pregnant, carrying a pup, and nursing) and renamed the variable the sexual and reproductive maturity factor. The logistic regression analysis includes the sexual and reproductive maturity categories and sampling time for the bats. All hypothesis testing was 2-sided, with α = 0.05, and was performed by using SPSS version 19 (SPSS Inc., Chicago, IL, USA).

NiV-neutralizing antibodies were detected in bats at all locations (Table 1). Overall, 82 (32.8%) of 253 bats were seropositive. We found no significant difference in seroprevalence rates in bats from the 5 states (p = 0.213, by Fisher exact test). NiV-neutralizing antibody titers ranged from <8 to 1,024 (median <8), and among seropositive bats, the median titer was 64. All culture and PCR results were negative. Among pups, juvenile, and adult bats, 0% (0/1), 25% (14/56), and 35.2% (68/193), respectively, were seropositive (Table 1). Age, sex, and female reproductive status were examined for their effects on serostatus of P. vampyrus bats. Univariate analysis showed that age and sex were not associated with seroprevalence of NiV. However, among nursing bats, a higher risk for NiV seropositivity was observed than in other adult females (Table 3).

Of 119 plasma samples collected from island sites, 2 were not used because of inadequate amounts of plasma. Of the remaining 117, 13 (11.1%) were seropositive. NiV antibodies were detected in 13 (13.8%) of 94 bats from islands off the east coast (Pulau Perhentian, Pulau Kapas, Pulau Tioman) of Peninsular Malaysia. Titers ranged from <8 to 256 (median <8); among seropositive bats, the median titer was 32. Samples from Pulau Pangkor (n = 24) were negative for NiV-neutralizing antibodies. All culture and PCR results were negative. Pups were not captured, but among juvenile and adult bats, 0% (0/31) and 11.1% (13/73), respectively, were seropositive (Table 2). Univariate analysis showed that sex was not associated with NiV seropositivity. Stratified data for various reproductive categories were sparse and lacked power for meaningful analysis (Table 3).

The antibody titer difference between P. vampyrus and P. hypomelanus bats was significant (p<0.001, by Mann-Whitney U test. The difference in seroprevalence for NiV between P. hypomelanus (13/117, 11.1%) and P. vampyrus (82/253, 32.8%) bats was significant (χ² 19.54, p<0.001), and risk for a seropositive reaction to NiV was 3.9× higher for P. vampyrus bats than for P. hypomelanus bats.