Study sites

Laos.

A total of 27 healthy captive Asian elephants, including 10 males (29.0 ± 16.1 years of age, range, 7–60 years) and 17 females (40.0 ± 8.8 years of age, range, 25–57 years) were included in this study (Table 1). From this population, 19 individuals were housed at the Elephant Conservation Center (ECC) in Sayaboury Province. The ECC consists of 530 hectares of mixed deciduous and dry dipterocarp forest. It is open to the public all year, but tourist activities are limited to remote observations only. Elephants spend the morning (0900–1200 hours) and afternoon (1330–1630 hours) socializing and foraging in different natural areas throughout the venue. At night (1800–0800 hours), elephants are restrained with a 40 m chain within different locations of the forest to allow natural foraging and prevent any conflict with surrounding farmers. The remaining eight elephants were located in the Nam Phouy National Protected Area (Nam Phouy NPA), which covers 191,200 hectares of forest [27] and shelters the second largest population of wild elephants in Laos [28]. Of these, four were part of a soft-release reintroduction program into the protected area with complete freedom to roam (‘released group’), unless intervention was deemed necessary for safety reasons – either to protect the elephants themselves or nearby communities. Once a month, physical health was evaluated by a veterinarian. Four other elephants were housed at Nam Phouy NPA, where mahouts (elephant keeper) employed traditional management techniques involving periodic release into the forest (‘traditional management group’). During the dry season, elephants were alternately tethered on 30–40 m chains or released with a dragging chain in wooded areas near the village. When tethered, they were relocated daily to refresh the feeding grounds and provide access to water. In the rainy season, when human resources were primarily allocated to planting crops, elephants were predominantly released into the forest for extended periods, monitored by mahouts through weekly searches.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of parameters related to the management of elephants at captive venues in Laos and Thailand.

https://doi.org/10.1371/journal.pone.0334550.t001

Sample collection

Blood and fecal samples were collected once or twice a month from February 2020 to April 2021 (Table 1). All samples from Lao elephants were collected and analyzed specifically for this study. In Thailand, comparative values were drawn from previously published data using comparable methodology [12,13,26]. Notably, both countries were under international travel bans due to COVID-19 from early 2020 through at least 2021; thus, samples represent data collected during the pandemic, when tourist activities were limited or absent.

Body condition scoring

Body condition was assessed monthly by scoring the back, ribs, and pelvic bone areas using a standardized photographic guide with detailed instructions [29]. Scores ranged from 1 (thinnest) to 5 (fattest), and assessments were made in 0.5-point increments for greater precision [15].

Serum glucose, insulin, G:I

Serum glucose was measured by automated standardized liquid chromatography-mass spectrometry at the International Medical Diagnostic Center in Vientiane, Laos (Roche Cobas C 311; Cobas® Mass Spec) or an automated glucose analyzer (Glucinet T01-149, Bayer, Barcelona, Spain) at Chiang Mai University in Thailand [12]. Serum insulin concentrations were measured by a solid-phase two-site bovine insulin enzyme immunoassay (EIA; Cat. No. 10−1,113−01; Mercodia, Uppsala, Sweden) validated for elephants [12,23]. Colorimetric responses were determined by a spectrophotometer at 450 nm filter with an Opsys MR Microplate Reader (TECAN Sunrise™ microplate reader; Salzburg, Austria) [12]. All samples were analyzed in duplicate; intra- and inter-assay coefficients of variation were <10 and <15%, respectively.

Lipid profiles

Serum TC, LDL, HDL, and TG were quantified by automated, standardized liquid chromatography-mass spectrometry Roche Cobas C311 (Cobas® Mass Spec) in Laos and an Automated Clinical Chemistry Analyzer (Sysmex; BX-3010, Sysmex Corporation of Japan) in Thailand [12].

Hormone processing and extraction

Fecal samples were extracted as described by Norkaew et al. (2018) [15]. Briefly, frozen fresh feces were thawed at room temperature (RT) and dried in a conventional oven at 60°C for ~24−48 hours and stored at −20°C until extraction. Mixed dry powdered feces (~0.1 g ± 0.01) were extracted in 5 mL 90% ethanol by boiling in a water bath (96°C) for 20 minutes and adding 95% ethanol as needed to keep from boiling dry. Samples were centrifuged at 2,500 x g for 20 minutes. Supernatants were transferred into a new glass vial, and the fecal pellets were centrifuged again with 5 mL 90% ethanol. The combined supernatants were dried under air in a 50°C water bath. Dried extracts were reconstituted in 1 mL assay buffer (0.0137 M Trizma base, 0.2 M Tris-HCl, 0.2 M NaCl, 0.2 M EDTA, 0.001% BSA, and 0.001% Tween 20; pH 7.5) and stored at –20°C until enzyme immunoassay (EIA) analysis.

Hormone analyses

Concentrations of fGCM were determined using a double-antibody EIA with a polyclonal rabbit anti-corticosterone antibody (CJM006) validated for Asian elephants [30] and described by Norkaew et al. (2018) [15]. The absorbance was measured at 450 nm by a microplate reader (TECAN Sunrise™ microplate reader; Salzburg, Austria). Assay sensitivity (based on 90% binding) was 0.14 ng/mL. Samples were diluted 1:4 for female elephant samples and 1:6–1:8 for male elephant samples in assay buffer for analysis. The inter-assay CV for high (30% binding) and low (70% binding) control samples was < 15%. Samples were reanalyzed if the duplicate CV was > 10%; thus, intra-assay CVs were <10%. Fecal data are expressed as ng/g dried feces.

Statistical analyses

Statistical analyses were performed using R version 4.1.2 [31]. Physiological measurements were collected monthly from most elephants, although some individuals were sampled twice within the same month. In those cases, the two values were averaged, so that each elephant contributed only one data point per month for each parameter. Repeated measures data were analyzed using Generalized Estimating Equations (GEE) to evaluate the effects of location and country, season, and sex on metabolic and lipid parameters, BCS, and fGCM concentrations. The analysis specifically examined: (1) the country effect; (2) the effect of management practices across the three main study locations (ECC, release/traditional management, and Thai tourist camps); (3) the influence of season (summer, winter, and rainy); and (4) the effect of sex on the outcome measures. To test whether seasonal effects differed by country, an interaction term of season and country was incorporated into the model. To assess sex differences, a separate model was fitted with a country and sex interaction. Individuals were treated as a random effect. GEE was fitted to marginal regression models using the geeglm() function from the geepack package in R, specifying an autoregressive correlation structure of order 1 AR (1) to account for within-subject correlation over time [32]. GEE models were followed by Bonferroni’s multiple comparisons of estimated marginal means post-hoc tests with p- value correction. The significance level for all statistical analyses was set at α = 0.05. Associations between BCS and metabolic hormones, lipid profiles, and fGCM, were further evaluated using aggregated data. For each physiological category, all parameters were consolidated, resulting in a single representative value per individual per category for the correlation analysis. Given the ordinal nature of BCS data and the presence of tied values, Kendall’s tau-b correlation coefficient was calculated using the cor.test() function in R (method = “kendall). To visualize the correlation trends, the geom_smooth() function from the ggplot2 package was also applied.

Metabolic profiles

Significant location and sex-based differences were observed in circulating glucose and insulin concentrations. Using comparable laboratory techniques, while most values were within normal ranges based on existing references [12,15,16,23,33], concentrations were overall higher in Thai elephants compared to those in Laos. The released elephants and those in the traditional management group within the Nam Phouy NPA had the lowest concentrations of metabolic markers (i.e., glucose, TC, HDL, and LDL), except for TG compared to all other groups, followed by elephants at the ECC. These metabolic differences may be related to diet. Elephants in Laos forage on a more natural/diverse diet throughout the year, especially the released/traditional management groups, and food supplements are not part of the diet. By contrast, the traditional diets of Thai elephants are highly managed, consisting of napier grass as the main fodder, with limited natural foraging opportunities, and feeding of high-calorie supplements like bananas and sugar cane by tourists. Even during the tourism ban, camp owners in Thailand continued to provide supplements, albeit in lower quantities [13], but apparently enough to continue having an effect on metabolic status. These results align with previous studies in black rhinoceros (Diceros bicornis) [34] and gorillas (Gorilla gorilla gorilla) [35], where higher insulin values were observed in captive-fed animals compared to free-roaming counterparts.

Metabolic profile results were associated with BCSs. In Thailand, higher BCS was found to be correlated with elevated insulin concentration, whereas no such associations were observed in the Lao elephants. This discrepancy can be explained by the distribution of BCSs within each group. In Thailand, a substantial proportion of tourist elephants were classified as overweight or obese: nearly 45% were assigned a BCS of 4, and almost 30% were given the highest score of 5 at the onset of the study [26]. By the end of the study, 23% of Thai elephants still had a BCS of 4. By contrast, only 3.33% of elephants in Laos had a BCS of 4, with none scoring 5 at any time during the study. With so few overweight or obese individuals present in Laos, it was not surprising that metabolic correlations with BCS were not detected.

There were no sex differences in BCS among the Lao elephants, whereas in Thailand, females tended to have higher, though not significant, BCSs corresponding to higher insulin concentrations and lower G:I compared to males. The latter agrees with studies before the COVID-19 pandemic in Thailand [15,16,33] and could be explained by the fact that females tend to be the preferred choice for interactions with tourists, like feeding, as they are more docile and predictable than males [36]. Although there was no tourist feeding in Thailand during the COVID-19 pandemic, female elephants started the study considerably fatter. Within the first 2 months of the study, 91.3% of females in Thailand had BCS = 4 or 5, while only 27.3% of the males did. In both countries, male elephants exhibited lower insulin concentrations than females, which could be due to dietary restrictions as a traditional practice to regulate body weight to better control the time and intensity of musth periods [16]. This phenomenon aligns with findings from a human study, in which insulin levels decreased as a result of time-restricted feeding [37]. However, further research is needed to explore and validate this hypothesis, as it could also suggest potential differences in sex-related metabolic or hormonal regulation.

Lipid profiles and BCS

Location differences in relationships between BCS and lipids were also noted. Overall, BCS, TC, LDL, and HDL levels were higher in Thai compared to Lao elephants, which agrees with the observation that BCS remained relatively stable during the study in Laos, likely due to consistent food availability and exercise, even during the COVID-19 tourism ban. Sex differences were also apparent in the lipid panel, where males had higher TC and LDL than females in Thailand. However, these differences were not observed between male and female elephants in Laos. Previous studies in Sri Lankan elephants have reported higher TC in males than females, attributing these differences to physical activity, as exercise can reduce serum cholesterol [38]. However, in the present study, reported walking distances and hours spent on chains among the Thai population were similar between males (walking distance: 2.24 ± 0.21 km; chain hours: 23.70 ± 0.52h) and females (walking distance: 1.96 ± 0.10 km; chain hours: 22.49 ± 0.58h) [13,26], suggesting that other intrinsic or management-related factors may contribute to the observed metabolic differences.

Among Thai elephants, thinner individuals exhibited higher G:I,TC, HLD, and LDL concentrations. This phenomenon can be explained by the fact that animals with lower BCS have reduced fat reserves and may actively mobilize these stores to meet increased energy demands. During this metabolic process, fat deposits are broken down (lipolysis), releasing free fatty acids and other lipids into the bloodstream, which leads to elevated circulating lipid concentrations. This effect has been reported in pregnant and lactating dairy cows, where periods of negative energy balance, often due to increased metabolic demands, result in intensified fat mobilization and subsequently higher blood lipid levels such as non-esterified fatty acids [39]. In addition diets with low G:I had been shown to have benefits in terms of short-term glycemic control, weight, and adiposity [40]. In Laos, although the correlation was not significant, lower BCS was associated with higher HDL, a finding observed in humans, where an increase in physical activity was associated with an increase in HDL [41]. Lao elephants had regular walking routines that, in addition to naturalistic diets, could contribute to the increase in HDL concentrations and lower BCS. Therefore, these findings underscore the critical value of fostering active lifestyles within captive populations to support better health outcomes.

During the COVID-19 lockdown, Thai elephants experienced a marked reduction in physical activity, with some camps extending chaining periods for up to 48 contiguous hours, and chain lengths restricted to just 3 m during the day and 6 m at night [14,26]. In contrast, walking routines for elephants in Laos were unaffected by the absence of tourists. Elephants at the ECC continued to roam freely in the forests during the day, and at night were restrained with chains averaging 40 meters in length, allowing access to natural feeding grounds within an 80-meter diameter. Released elephants had unrestricted movement both day and night, while those under traditional management in the Nam Phouy NPA were only restrained occasionally.

In other species, moderate exercise is known to improve lipid homeostasis [42,43], while frequent, varied feedings, rather than a few large meals, can help suppress hunger and lower serum insulin [44]. As megaherbivores, Asian elephants are adapted to spend much of the day foraging on a wide variety of plant materials [45–47], and disruptions to this natural pattern, such as those observed in Thai tourist camps, are likely to contribute to the observed metabolic derangements. The Thai elephant’s low dietary diversity, reliance on sugar-rich supplements, and limited feeding frequency likely compounded the negative effects of inactivity, explaining the higher BCSs and less favorable lipid and metabolic panels compared to Lao counterparts, underscoring the protective effects of naturalistic management [8,23].

fGCM and BCS

High BCSs have been linked to limited exercise, feeding of high-calorie treats, and higher fGCM concentrations in Thailand [15,16,33] and U.S. zoos [23,48]. By contrast, in the present study, the Lao population presented with lower BCSs and higher fGCM concentrations compared to the Thai population. Although the negative correlation between fGCM and BCS in the Lao elephants only approached significance, it suggests that as body condition decreased, fGCM concentrations tended to increase, a finding that has been reported for free-ranging Asian elephants in India [49]. Low BCSs could produce more physiological stress due to resource scarcity or other environmental challenges [49]. Thus, both low [50–53] and high [54–56] BCSs may be a stressful event for the body. These results highlight that both exercise and diet diversity are critical for maintaining healthy metabolic and lipid profiles in captive elephants, and that management practices mimicking natural conditions can promote better physiological outcomes.

fGCM and seasonality and management factors

Seasonal peaks in fGCM were observed during the rainy (May – October) and winter (October-February) seasons in the Thai population, aligning with another Thai study conducted during the pandemic [26]. Elephants in Myanmar also exhibit higher fGCM concentrations during the monsoon season (May-September) [57]. In addition, these data also align with pre-pandemic findings in Thailand where fGCM concentrations peaked only during the winter during the high tourist season [15,16], suggesting that natural environmental factors could play an important role the in adrenal activity. Furthermore, since no seasonal variation was observed among the Lao population, it is possible that the elevated adrenal activity observed in Thai elephants is also influenced by management conditions. In Thailand, elephants are often restrained with short chains in non-naturalistic environments, which can increase stress responses during extreme weather events such as heavy rain or strong winds, as they have limited options to seek shelter, find comfort areas, or take refuge near other individuals. Providing elephants with more options to respond to adverse weather could be an important consideration for future management practices. In addition to seasonal variation, numerous other factors can influence GC concentrations. Aversive situations may elicit increases in concentrations, but they can also be increased by events considered to be neutral or pleasant [58]. Behaviorally, the ECC elephants rarely displayed behaviors commonly associated with stress, like stereotypies, whereas those are observed at high rates in Thailand tourist camps (57%) [59]. Rather, ECC elephants are more socially active, as both males and females have daily social interactions, especially compared to Thailand tourist camps. During the COVID-19 pandemic, the ECC initiated a social program for bulls to create all-male groups like those in the wild. Introducing new individuals can elicit stimulatory adrenal responses, especially during the initial months of the introduction [60–62]. During these social encounters, affiliative interactions were performed more frequently than aggressive and submissive behaviors and accounted for over 79.7% of the interactions (López Pérez et al., in prep). Thus, the elevated fGCM concentrations observed in ECC elephants, likely reflect the physiological arousal and adaptive stress response associated with increased social activity and the challenges of group restructuring, rather than distress. However, due to COVID-19 restrictions, several ECC mahouts chose to return to their hometowns, which may have compounded these effects. Studies demonstrate elephants respond more efficiently to familiar handlers, requiring weeks to months to acclimate to new caretakers [63,64]. Thus, higher fGCM at the ECC compared to Thailand and other Lao populations may reflect a multifactorial outcome of proactive social enrichment and pandemic-driven management disruptions, rather than a simple welfare deficit.