Materials & methods

Case selection and inclusion criteria

Between January 2020 and December 2024, a total of 390 cases involving 438 animals were submitted to Stichting Forensisch Dierenonderzoek (Forensic Animal Research Foundation). Of these, 346 animals (79.0%) underwent a complete necropsy, while the remaining cases were assessed through digital forensic analysis, which included the evaluation of photographic or video evidence provided by law enforcement authorities, animal welfare organizations, or members of the public when carcasses were unavailable or animals were still alive.

Inclusion criteria required that cases involve animals suspected of non-natural or abuse-related deaths, with submissions intended for potential legal evaluation. All cases were treated as having forensic potential, meaning they were processed using standardized internal protocols to preserve evidentiary integrity within our facility, including systematic documentation, photography, and sample collection for ancillary testing (e.g., toxicology).

However, the chain of custody prior to submission was not consistently documented by submitting parties, which may affect the legal robustness and interpretability of findings. Necropsies were nevertheless conducted under controlled forensic conditions to support reconstruction of events based on available evidence (e.g., injuries, decomposition stage, or contextual information), in line with established definitions of forensic science [28].

Cases were primarily submitted by law-enforcement authorities, municipal animal-welfare services, and non-governmental organizations, with occasional contributions from private veterinarians or individual citizens. Background information provided at submission was often limited, typically including species, and in some instances breed, sex, geographic location, and the suspected type of abuse. Ownership status (e.g., stray versus owned) was rarely indicated.

Dogs constituted the majority of submitted species (n = 183/438, 41.8%), followed by cats (n = 125/438, 28.5%) and birds (n = 58/438, 13.2%) across both necropsy and digital forensic cases (Fig. 1).

Fig. 1: A pie chart of the distribution of animals across submitted cases (N = 438), categorized in the categories dogs (orange), cats (yellow), birds (red), livestock (blue), rodents (pink) and wild animals (green)

Full-body radiography was performed in cases with suspected firearm involvement, and imaging findings were incorporated into the forensic reports. All necropsies and classifications were performed by a single qualified veterinary forensic scientist using standardized internal protocols. No second-examiner review or formal assessment of interobserver variability was conducted, which may affect reproducibility.

Necropsy procedure

A comprehensive postmortem examination was performed on each animal following a standardized forensic protocol. When possible, the breed was confirmed, and animals were scanned for subcutaneous microchips or examined for external identification such as ear tags. If sex and age were unknown, they were estimated during the examination.

Age estimation was based on dentition, including tooth eruption and wear, classifying animals into three age groups: juvenile (less than ten months), adult (eleven months to ten years), and senior (older than ten years). For companion animals, it was also assessed whether they were neutered.

The decomposition stage was assessed, when requested by the submitting agency, based on gross morphological characteristics using an internally defined operational scheme (Supplementary Material 2).

The external examination included systematic documentation of coat condition, body weight, dental and nutritional status, nail integrity, oral cavity, eyes and ears, neck mobility, torso condition, anogenital region, and any observable injuries or distinguishing features. Following the external examination, a systematic internal examination was conducted to assess organ condition, structural integrity, and potential pathological changes. The abdominal cavity was assessed for abnormalities, including the presence of blood, fluid, haemorrhages, or contusions. The intestinal tract, bladder, kidneys, adrenal glands, spleen, stomach, liver, and diaphragm were examined for structural or macroscopic pathological changes, as well as content where applicable. The thoracic cavity was inspected for blood, fluid, haemorrhages, or contusions. The heart was evaluated for macroscopic abnormalities. The lungs, oesophagus, and neck muscles were examined for any deviations from normal anatomy. Special attention was given to the trachea for the presence of fluid or foam. The skull and brain were also examined for abnormalities. Any additional findings were documented.

Photographic documentation was performed throughout the examination, and the packaging method at submission was recorded. Tissue and content samples were collected and preserved (e.g., formalin-fixed or frozen) for potential toxicological or other ancillary analyses; however, such analyses were only performed upon request by judicial authorities.

Determination of cause and manner of death

Cause and manner of death were determined based exclusively on gross necropsy findings. Histopathological examination and ancillary testing (e.g., toxicology, microbiology) were not performed within the authors’ institution. When additional analyses were indicated (e.g., for suspected poisoning or subtle lesions), tissues, organs, or contents (e.g., liver, stomach contents, blood) were collected and preserved (e.g., formalin-fixed or frozen) for retrieval by the submitting agency, which was responsible for external laboratory submission. Results from such analyses were not available to the authors and were therefore not incorporated into the dataset. Accordingly, all determinations should be interpreted as operational and presumptive classifications rather than definitive forensic diagnoses. Gross examination alone may not detect conditions such as intoxication, subtle trauma, asphyxia, infectious processes, or postmortem artifacts in all cases, and may therefore introduce a risk of misclassification.

The cause of death was defined as the primary pathological process considered to have led to death, based on macroscopic findings. The manner of death was categorized as natural, non-natural, accidental, or undetermined using a forensic veterinary framework. Non-natural deaths were further subdivided into the following categories: blunt force trauma, sharp force trauma, firearm injuries, asphyxia, bite injuries, fire/explosion injuries, suspected poisoning, and neglect and deprivation. Neglect and deprivation - including withholding of food, water, shelter, or veterinary care - were classified as non-natural deaths in accordance with established forensic veterinary frameworks [22, 29, 30].

Data analysis

All quantitative analyses were performed using descriptive and inferential statistics with a significance threshold of α = 0.05. Frequencies, proportions, and cross-tabulations were calculated for species, presumptive manner and cause of death, decomposition stage, and submitting agency.

For variables with missing or undetermined values, a complete-case approach was applied: cases were excluded only from analyses involving the variable for which data were absent, and retained in all other analyses. Missing data are reported as separate categories in descriptive summaries and their reasons are described in the relevant Results & Discussion sections.

To explore spatial and temporal trends, one-sample t-tests were conducted for each province, comparing the number of cases to the national mean. Logistic regression models were applied to assess year-to-year trends in cause of death categories, adjusted by Bonferroni correction. Chi-square tests and ANOVAs were used to explore associations between categorical variables (species, cause of death, decomposition stage, and submitting agency).

Given the absence of reliable population-based denominators (e.g., companion animal populations), geographic analyses reflect patterns in submitted cases rather than incidence rates.

Results are presented using bar charts, line graphs, pie charts, heat maps, and stacked area plots to visualize distributions across time, species, and geographic regions.