Evaluation of body mass index as a prognostic indicator from two rough‐toothed dolphin (Steno bredanensis) mass strandings in Florida

Abstract Rough‐toothed dolphins (Steno bredanensis) are a common mass stranding species in Florida. These large stranding events typically include a small number of sick or injured individuals and a much larger number of healthy individuals, making rapid triage essential. Little data exist on rehabilitation outcomes, and historically, successful outcomes are limited. Furthermore, very little data exist on the feeding habits and dietary needs of this species. This study compared morphology and body mass index (BMI) in two rough‐toothed dolphin mass stranding events in Florida: August 2004 (n = 36) and March 2005 (n = 32). The two groups were significantly different in morphologic measurements, with age and gender‐adjusted intake BMI significantly (p < .01) different (2004 = 0.34 ± 0.02; 2005 = 0.41 ± 0.02) between groups. Ten animals from 2005 had weights tracked throughout the rehabilitation process and demonstrated an initial drop in BMI followed by an increase and a plateau prior to release. When comparing initial BMI by stranding outcome, individuals that were rehabilitated and released had a significantly (p = .03) higher BMI than individuals who were euthanized. However, there was no difference between dolphins that died of natural causes (p = .56) and animals successfully rehabilitated. Analysis of BMI can be a useful marker in triage during a stranding, when resources are limited to identify individuals most likely to survive, as well as in determining the appropriate body condition for release. The data reported here can provide guidance on evaluating the nutritive status on this uncommon species that would otherwise be difficult to obtain among wild populations.


| R ATIONAL /BACKG ROUND
Rough-toothed dolphins (Steno bredanensis) are members of the family Delphinidae. They are typically found in pelagic waters throughout warmer temperate and tropical regions of the world. In the United States, there are three major recognized stocks: one in Hawaii, one in the Northern Gulf of Mexico, and one in the Western North Atlantic (NOAA Fisheries, 2012. As a pelagic species, they are more prone to stranding than their coastal counterparts, as they are naïve to navigating the conditions of inshore waters (Geraci & Lounsbury, 2005). In addition, rough-toothed dolphins exhibit exceptionally strong social bonding with free-ranging social groups typically consisting of up to 50 individuals. Due to the highly communal nature of many marine mammals, individuals are inclined to remain in close proximity to one another (Jefferson, 2009;Kuczaj & Yeater, 2007;Leatherwood & Reeves, 1983;Lodi, 1992;Würsig, Jefferson, & Schmidly, 2000). were successfully released back into the wild (NOAA, unpublished data). Therefore, there is a clear need to be able to identify healthy individuals during triage in order to improve the outcomes associated with mass strandings of this species.
Examining "body condition" is one way of selecting the best candidates for rehabilitation. It is often defined as the amount of energy held in an individual's lipid stores (Pitt, Larivière, & Messier, 2006).
These fat stores are a reliable reflection of an individual's foraging effort and success (Aguilar & Borrell, 1990). Further, morphometrics can be used as an indicator of the nutritive condition not only for the individual, but also for the entire population (Hart, Wells, & Schwacke, 2013).
Applying length-to-weight comparisons as an objective assessment of health is becoming common practice in cetaceans (Perrin, Dolar, Chan, & Chivers, 2005), though further research is required to improve precision and species-specific reference ranges. There is currently no consensus on the best morphometric index with which to calculate body condition, let alone what healthy ranges look like (Kershaw, Sherrill, Davidson, Brownlow, & Hall, 2017). In a recent study, the length-to-girth ratio for stranded common dolphins (Delphinus delphis) was found to be significantly different between animals with postrelease success and failure; whereas, the length-to-weight ratio was not found to be a good prognostic indicator .
The lack of standardized morphometric ranges is a significant limitation on the clinical care and rehabilitation of rough-toothed dolphins.
With limited resources, funding, and time veterinarians and responders often need to make decisions regarding triage, logistics, rehabilitation facilities, and animal survival prognosis based on limited clinical data. A species-specific range for body mass index (BMI) is a useful guide during triage, when determining if an animal in rehabilitation is receiving adequate nutritional support, and/or if an animal is ready for release into the wild. Therefore, the objectives of this study were to compare BMI data from two rough-toothed dolphin mass stranding events for differences and determine the usefulness of BMI during triage, through rehabilitation in an effort to identify a suitable body condition level, and to potentiate increased postrelease success. Weights for animals were obtained using a digital hanging scale.

| ME THODS
From the scale, chains or straps supported a sling which was tared and subsequently held either a living or deceased animal. Some carcasses were suspended on a hook by the dorsal fin to get a weight reading depending on equipment availability.
Standard protocols were used to ensure that morphologic measurements were taken in a consistent manner. Length was taken by measuring the straight line distance from the tip of the snout to the notch of the flukes; while blubber thickness was measured along the maximum girth, anterior to the dorsal fin insertion, transecting at dorsal midline, midlateral, and ventral midline areas (Geraci & Lounsbury, 2005).
The first stranding event occurred on August 6, 2004; 37 roughtoothed dolphins were initially found on a beach on Hutchinson Island in St. Lucie ( Figure 1). During the response animals in hand were assessed, five were given dorsal fin roto-tags before impending severe weather forced rescuers to release the animals offshore and take shelter. Nearly 3 hr later, 36 animals including the five-tagged individuals were found restranded on a nearby beach while one was recovered alive on August 8. Of the 37 original animals, 30 were humanely euthanized on the day of the stranding. The remaining animals were moved into rehabilitation on August 9, where one died spontaneously 3 days later, three others died spontaneously over 2 months, leaving three which were successfully rehabilitated and released (Table 1)

| Statistical analysis
Body mass index was calculated using methods published by Kershaw et al. (2017) comparing morphometric equations for body mass index in small cetaceans suggesting that normalizing the data can be achieved using a simple calculation:

| RE SULTS
To test whether the animals in this study exhibit sexual dimorphism, a two-sample t test was performed with data from all individuals by age class. These animals did not appear to be sexually dimorphic (subadults p = .81; adults p = .18), and as such, male and female individuals were grouped together in subsequent statistical calculations.
A total of 68 individuals had complete demographic and morphometric data and were included in the statistical analysis ( According to necropsy reports, Cetacean Data Records, and notes from the stranding events themselves, three individuals were described as being "underweight" from the 2004 stranding; 20 from 2005 were "underweight" while five were described as "healthy weight." This is, however, a largely subjective measurement and subject to observer bias. Notes from the 2004 stranding event were very sparse, with no other information on apparent body condition besides the three "underweight" individuals, leaving 33 animals unknown. The

| D ISCUSS I ON
In this study, we compared BMI across stranding outcomes for rough-toothed dolphins. There was evidence of differences by geographic location, suggesting that the two populations (Atlantic and Gulf of Mexico) may be more distantly related than previously assumed. Further, there was no sexual dimorphism evident in either population (p = .99), indicating that differences in morphology were not caused by the sex of the stranded individuals (it should be noted; however, that several studies have found evidence of sexual dimorphism in this species Miyazaki & Perrin, 1994;West, 2002 (Barco et al., 2012).
Calculating the BMI for wild, non-stranded individuals requires the capture and weighing of free-swimming animals. Several methods have been tested for estimating nutritive status in the wild; comparing total length to maximum width has shown promising results in estimating both reproductive and nutritive status in large freeswimming cetaceans (Perryman & Lynn, 2002;Pettis et al., 2004).
Physical indicators can also be utilized to assess overall health and nutrition. In domestic mammals, including cows, horses, and dogs, there are specific morphologic points (e.g., spinal processes, ribs, hips) that can be evaluated to create a general, qualitative measure (Eversole, Browne, Hall, & Dietz, 2009  Studies on blubber distribution have shown that, during periods of starvation, blubber is mobilized almost exclusively from the thorax while tailstock blubber is unaffected, suggesting that blubber caudal to the anus functions exclusively in locomotion and is metabolically inert (Koopman, Pabst, McLellan, Dillaman, & Read, 2002). In harbor porpoises (Phocoena phocoena), thoracic blubber thickness in emaciated animals was found to be half of that in healthy animals, while tailstock blubber showed no difference in thickness. As such, there is no "one size fits all" definition of body condition; species-specific descriptions are necessary to most accurately predict an individual's overall health status.
One morphological parameter, dorsal blubber thickness, was not statistically different between the two mass stranding events. In delphinids, the dorsal blubber thickness is largely unimpacted in cases of acute emaciation (NOAA, personal communication), suggesting that the two populations were not in a state of chronic emaciation in which fat stores along the epaxial musculature would be mobilized.
The significant difference between the lateral and ventral blubber thicknesses suggests these measurement sites could be reference points for determining body condition. Thinning of the lateral blubber layer (i.e., thoracic body wall), results an increased visibility of the ribs with declining body condition defining emaciation (Joblon et al., 2014).
Using morphometrics as positive prognostic indicator for determining potential rehabilitation success could have a significant impact on the triage and rehabilitation process, when resources are sparse, and efforts are best utilized for animals with the greatest chances of being released. A 2014 study suggested that failed animals in rehabilitation had significantly higher length-to-girth ratios that those that were released , indicating that thinner animals were less likely to survive. In this study, only two parameters in the 2005 population proved to be connected to survival: a smaller intake length (p = .038) and a smaller intake weight (p = .03). This indicates that younger, smaller animals had the greatest chances of surviving the initial stranding event as well as succeeding in rehabilitation. In a mass stranding event, the younger naïve animals tend to be following the older individuals, suggesting that they may be in better health than larger animals that are stranding due to health problems. Intake length and intake weight were release was set at 4 weeks, a duration that only three individuals were proven to have successfully achieved, all 10 of the tagged released animals were determined to be successfully rehabilitated (Wells et al., 2008).
Body mass index comparisons found in this study may be utilized in rehabilitation settings to help determine when individuals are at a healthy weight and potentially ready for release. According to data from the 2005 stranding event, consequential rehabilitation and postrelease satellite monitoring, individuals found to be clinically stable, in good body condition and deemed ready for release with a BMI >0.53 were successfully rehabilitated. For the 2004 animals, according to postrelease satellite monitoring, only animals with a BMI above 0.5 had adequate transmission data to be considered successfully released (Manire & Wells, 2005 their diet was reconfigured to include less fat and more protein, with an overall increase in the total weight of fish given per day, resulting in improved BMI (Figure 3).
In the future, consistent collection of weight and length mea-

ACK N OWLED G M ENTS
The examination or treatment of animals was authorized through

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
All contributing authors have approved this work for publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
The marine mammal stranding demographic and morphological data