Epidemiology of severe trauma in cats: An ACVECC VetCOT registry study

1 INTRODUCTION

Significant strides have been made in better understanding the causes and outcomes of trauma a in cats; however, few studies have looked at its epidemiology in a large population, particularly in North America.^1-4 Even less is known about the epidemiology, clinicopathological abnormalities, and outcomes of severe feline trauma cases. This lack of information continues to limit understanding of potential preventative steps. Gaining a better understanding may be an important step in improving outcomes, ensuring required resources are appropriately allocated, and providing more accurate prognostic information regarding trauma in this population.

Previous studies have consistently found traumatized cats to be young, with male and intact animals overrepresented.^1-7 Trauma is a common cause of mortality in cats and, in some countries, has been documented as the leading cause of death in the primary care setting.^{1-3, 8, 9} Mortality rates in the literature vary greatly with type and severity of trauma.^{1, 2, 10-12} The severity of trauma can be quantified using several scoring systems, including the Animal Trauma Triage Score (ATTS) and modified Glasgow Coma Scale.^{8, 13, 14} The ATTS has been shown to relate with mortality from various forms of trauma in smaller single-center studies in cats.^{6, 7, 10, 17} Recently, the ATTS has also been validated in a large multicenter population using the American College of Veterinary Emergency and Critical Care (ACVECC) Veterinary Committee on Trauma (VetCOT) registry.⁸

From this recent study,⁸ unpublished Youden's J statistic suggests that an ATTS of 3.5 is the optimal cut point for predicting mortality.¹ A smaller, unpublished analysis of 112 cats from the University of Minnesota evaluating ATTS versus survival to discharge using Youden's J indicated a cut point of 2.862 for prediction of in-hospital mortality (P < 0.0001).² In human studies, severe trauma is defined as a mortality risk >10% based on initial severity scores.^18-21 Based on these results, we posit that an ATTS ≥3 is consistent with severe trauma in cats.

The primary goal of this study was to identify demographic information and describe epidemiological factors and clinical abnormalities that differentiate cats with severe trauma, defined as an ATTS ≥3, from those with mild injury (ATTS 0–2). We hypothesized that cats with an ATTS ≥3 would have higher mortality than those with an ATTS of 0–2, along with increased lactate, decreased base excess, decreased PCV and total plasma protein (TPP), and increased morbidity, as demonstrated by increased rates of hospitalization and transfusion requirements. We also hypothesized that severely traumatized cats would be more likely to have suffered head and spinal injuries.

2 METHODS

This was a multicenter observational study using data entered into the ACVECC VetCOT registry. Data in the registry comprise cases input by a network of Veterinary Trauma Centers (VTCs) located in the United States, Canada, and the United Kingdom.²³ The VTC network consists of both university and private hospitals, and data are recorded on patients presenting for injuries associated with an external force, including blunt, penetrating, and crush injuries. The registry methodology, including variables collected, has been previously described.²³ Data entered into the trauma registry are deidentified, and only data generated via clinician-directed care are included. Variables that require funds to perform (eg, biochemical data and diagnostics) are considered optional data, and missing variables may result either because the test is not performed or not entered into the registry by VTC data entry personnel.

An application for use of the VetCOT registry data was submitted to and approved by the VetCOT registry subcommittee. Data were extracted for feline patients presenting to VTCs between April 1, 2017 and December 31, 2019. Although data are available as early as September 2013, for purposes of quality control and additional data capture, there were multiple changes to the registry questions on April 1, 2017. As such, only data captured on or after April 1, 2017 were included in this analysis. Cases were divided by ATTS as either low (ATTS of 0, 1, or 2) or high (≥3), with the high score group defined as severe trauma. Cats without an ATTS recorded were excluded. Any patient with body weight listed as 0 or of unknown sex was excluded from analysis in that category, but the remainder of the case information was retained for analysis. The registry was screened for possible data input errors reflecting biologically implausible results (eg, impossible negative number, or ionized calcium [iCa] >3 mmol/L [12 mg/dl]), which were excluded.

Statistical analysis was performed using commercially available software.³ For each categorical variable, a chi-square test was used to compare between trauma groups (mild vs severe). Counts and proportions were also generated. Visual inspection of the data was used to assess normality and check for outliers. For each numeric variable, the Wilcoxon rank-sum test was used to compare between trauma groups (mild vs severe). A nonparametric test was used because variables showed evidence of skew and outliers. Summary statistics reporting median, Q1, Q3, and interquartile range (IQR) were reported. Statistical significance was set at P < 0.05. Tables and figures were created using commercially available statistical software.⁴ Post hoc mortality by individual ATTS was assessed to determine which corresponded to >10% mortality in this registry. Youden's J was calculated to determine the optimal ATTS cut point for predicting mortality.

3 RESULTS

3.3 Types of trauma

In this trauma population, blunt injury (n = 2218, 57.4%) was the most prevalent inciting cause, with 35.1% (n = 1353) having penetrating injury, and 7.5% (n = 288) having both penetrating and blunt injury. Cats in the higher severity group had a higher frequency of blunt trauma and the combination of both blunt and penetrating trauma, while penetrating trauma alone was less common. This difference in distribution was significant (P < 0.0001). These data are summarized in Table 1. Causes of trauma are summarized in Figures 1 and 2. In cases of blunt trauma, a fall from a height (24.0%), being struck by a vehicle (21.3%), and nonpenetrating bite wounds (7.6%) represented the most common classified causes. However, 22.6% had an unknown cause for the traumatic injury (ie, unobserved injury or animal return to caretaker injured after being missing), and 18.4% did not fall into one of the specified categories. Unknown trauma was more common in the lower severity score group (ATTS 0–2), as was falling from a height. In those with ATTS ≥3, being struck by a vehicle (35.9%) was far more common. In cases of penetrating trauma, the most common cause was bite wounds (43.9%) followed by an unknown cause of the penetrating injury, then caused by one of the specified categories, then lacerations from metal (5.9%) and laceration from glass (2.2%). Bite wounds were the most common cause of penetrating trauma in both ATTS severity groups.

4 DISCUSSION

For this study, we proposed an ATT ≥3 to define severe trauma in cats. Definitions of major (severe) trauma in people include illness severity scores with a >10% mortality rate.^18-21 While very little has been published in cats, previous work in dogs has identified an ATTS cut point of 5 as optimally predicting mortality.^{15, 16} In cats, unpublished Youden's J analysis from a previous study using the VetCOT registry⁸ suggested that an ATTS of 3.5 was the optimal cut point.^a A smaller, unpublished analysis of 112 cats from the University of Minnesota evaluating ATTS versus survival to discharge using Youden's J indicated a cut point of 2.862.^b In this study, an ATT of 3 corresponded with a >10% mortality risk, supportive of our dichotomy. Post hoc Youden's J suggests that the optimal ATTS for predicting mortality was 4. An ATTS of 3 had a very similar Youden's statistic, with improved sensitivity for predicting mortality, although with decreased specificity. This is higher than in previous studies and may suggest improved outcomes in this population. This trend should be followed over time.

Cats sustaining severe traumatic injury comprise a large proportion of feline trauma patients presenting to VTCs and have markedly worse outcomes than those less severely injured. In this population, approximately one third of cats were classified as severely injured, with a mortality rate of 41.5%, as opposed to a 3.4% mortality rate in the low injury score group. A comparison of our results to the previously reported registry summary of cases from 2013 to 2017 showed overall survival to discharge for traumatized cats (83.3% in our study vs 82.5%) remains similar.²² Given the high mortality rate, better understanding the features of this more severely injured population has the potential to improve outcomes. Early identification of feline patients with an ATTS ≥3 is important during the triage process, so that therapeutic and diagnostic resources can be appropriately allocated to immediately identify and correct life-threatening injuries.

Owing to the large size of the dataset, almost all variables examined exhibited statistical differences between ATTS groups, although these differences may not all be clinically significant. Similar to previous studies, the cats in this study were typically young, with intact animals, particularly males, overrepresented.^1-7 This was true in both ATTS groups, with no difference in age between groups. Cats with a high ATTS trauma were 0.3 kg lighter than those with a low ATTS, which may reflect the fact that smaller animals are more susceptible to severe injury. Intact animals represented a higher proportion in the higher ATTS group. This may be due to greater time spent outside, which can expose them to more opportunities for trauma, as well as delayed time between trauma and presentation to a veterinarian. However, housing status and where injuries occurred were not recorded in this population.

In this study, blunt trauma was the most common cause of trauma reported, consistent with previous studies.^{1, 2, 4, 22} Compared to 2013–2017, prevalence in the VetCOT registry of blunt trauma, penetrating trauma, and a combination of both also remain similar (57.4% vs 56.7%, 35.1% vs 39.1%, and 7.5% vs 4.2%, respectively).²² Vehicular trauma was a common cause of blunt trauma in this study, also consistent with previous studies.^{2, 4} Vehicular trauma has been previously associated with worse outcomes compared to other causes of trauma.² Measures that reduce exposure of cats to moving vehicles can therefore be one way to reduce mortality. Interestingly, falling from a height represents a large proportion of cases in this registry, as previously reported in the VetCOT registry summary.²² Such a high incidence has not been reported in other papers, which have reported an incidence between 9.7% and 13.9%.^{2, 4} As the VTCs have an urban skew, this may help to explain this high incidence. Falling from a height was more likely in the low ATTS group. Previous work on high rise syndrome found a 96.5% survival rate, very similar to the low trauma group's survival.²³ The severe trauma group had an 18.7% lower incidence of falls, but a 24.7% higher incidence of being struck by a vehicle.

In cases of penetrating trauma, bite wounds were the most common cause in both groups. This is consistent with previous studies.^{4, 7} The source of the bite wound (eg, other cat, dog, or wildlife) is not recorded in the registry. This represents an area for further study as it could allow targeted client education on bite prevention. Higher ATTS was not associated with an increase in penetrating wounds, which may reflect the fact that bite wounds are less likely to have greater systemic involvement, although this could differ on a case-by-case basis. The primary difference between trauma categories was a higher incidence of ballistic injury in the higher ATTS score group, although these were still rare. Mortality with gunshot injuries has previously been reported at 13.5% and correlated with ATTS.^{17, 24}

The higher ATTS group had a higher proportion of patients with head and spinal trauma. This is intuitive, given that one of the scoring categories for the ATTS is neurological function. Given the vital structures involved, a higher mortality rate is expected. In veterinary medicine, euthanasia due to prognosis and cost of prolonged care also contributes to mortality. Understanding factors affecting euthanasia and improving treatment of neurological injuries could significantly improve outcomes.

Both abdominal and thoracic POCUS were more likely to be performed in the higher trauma score group, although only about half of severe trauma cases were recorded as receiving one. The presence of free abdominal fluid was relatively uncommon in both trauma groups. When present, it was most commonly in only 1 site. As only the first abdominal POCUS performed is input into the registry, it is possible these patients subsequently developed free fluid. One limitation of the registry is that radiograph or advanced imaging findings are not noted, so it is possible that other imaging modalities were used. In veterinary medicine, POCUS has been best studied in dogs, where previous work showed a 27% incidence of free abdominal fluid after trauma, with half of these patients having free fluid at >2 sites. Additionally, 15% of dogs had a change in their fluid score on repeat exam.²⁸ Another canine study showed an incidence of free fluid at 40% in vehicular trauma.²⁹ A report on POCUS findings in cats with trauma has been published in abstract form and found a 10% incidence of free abdominal fluid on presentation, consistent with our findings.³⁰ Incidence of thoracic changes in cats has not been reported. Given that ultrasound has the potential to identify numerous important sequelae of trauma, including pulmonary contusions, pneumothorax, hemothorax, uroabdomen, and hemoabdomen, POCUS is indicated in this population to help identify degree of injury during triage and resuscitation.³¹

Several clinicopathological abnormalities were statistically different between trauma groups. While PCV and TP were both significantly lower in the higher ATTS group, these differences are likely too small to be clinically useful. In spite of this small initial difference, the higher trauma score group did have a higher incidence of RBC-containing transfusions, suggesting that ongoing monitoring for progressive blood loss is indicated in this population, particularly prior to anesthetic events or procedures. Only a limited number of patients had laboratory data entered into the database. Because laboratory testing and additional diagnostics are at clinician discretion and also depend on owner finances, this may skew the cases where data are collected. PCV was performed in only 47.7% of cases with a high ATTS; lactate was performed in 40% of these cases. As bloodwork results can be input if collected within 24 hours, differences between pre- and postresuscitation values, such as PCV/TPP, cannot be assessed. Clinicopathological abnormalities in feline trauma have been examined in only a few papers. One single-center study looked at 43 cats with dog bite wounds and found increased lactate and decreased iCa in more severely affected cats.⁷ Ionized calcium did not differ between groups in this study, which contrasts with previous studies that have found hypocalcemia to be a feature of critical illness, including trauma.^{4, 7} The reason for this discrepancy is not clear, but as most blood samples are presumed to be collected near hospital admission, hypocalcemia may occur later during treatment. A prospective investigation of clinicopathological abnormalities in feline trauma is needed.

The low incidence of transfusion in this study population is unexpected. One study of pelvic trauma in cats found an 18.8% incidence of transfusion, with requirement for transfusion positively correlating with ATTS.⁶ The previously reported registry data (from 2013 to 2017) showed an incidence of 2.6%.²² To the authors’ knowledge, no other literature on feline transfusion frequency in a trauma population has been produced. The low incidence in these patients contrasts with a 36% incidence in 1 study of dogs hospitalized due to trauma.²⁵ Given the move toward resuscitation of hypovolemic human trauma patients with component blood products, better understanding resuscitation practices for severely injured feline trauma patients is an important next step.^{26, 27} Given the lower frequency of free abdominal fluid in this study compared to what has been reported in dogs, it is possible that hemorrhage is less common, but this requires further investigation.

Cats in the higher ATTS group had a higher frequency of surgical procedures performed in an operating room. Given that bite wounds and lacerations were the most common cause of penetrating trauma in both groups, wound repair may account for many of the procedures performed in the emergency room. The registry does not record the type of procedure performed, but procedures such as laceration repairs might be expected to occur in the emergency room, while more major abdominal or thoracic radiographs might be expected to occur in operating rooms. Future work assessing factors such as injury distribution, damage control versus definitive surgery, and timing of surgery could be helpful in guiding clinical decision-making.

The limitations of this study include the potential for data input errors into the registry. However, the large number of data points would be expected to minimize the effects of any such outliers. Some diagnostics may have been performed and not accurately recorded in the registry. Finally, while severe trauma cases present to all veterinarians, case distribution and outcomes may differ at VTCs. Nevertheless, the results reported here highlight key differences in cases of severe feline trauma and serve as an important starting point for further studies.

In conclusion, severe trauma, as quantified by an ATTS ≥3, is common in cats and has a high mortality rate. When compared to a lower severity score group (ATTS 0–2), significant differences exist in this population, including an increased frequency of blunt force trauma (particularly vehicular trauma), head and spinal trauma, and clinicopathological changes. These differences provide targets to help improve future outcomes. Relatively low incidences of POCUS evaluation and transfusions also merit further investigation.

CONFLICT OF INTEREST

Dr. Hall is the VetCOT and registry subcommittee chair and recused herself from the application process for VetCOT registry data use. The other authors declare no conflict of interest.

OFFPRINTS

Offprints will not be available.