pain recognition

A novel approach to pain recognition in donkeys

Gabriela Olmos
Faith A. Burden
Presentation date

Pain due lesions and clinical conditions is one of the main welfare concerns of the more than 42 millions donkeys that presently exist in the world. Yet, the knowledge to gauge pain in donkeys is lacking, misunderstood and/or not validated (Ashley, 2005).

Pain (yes/no/uncertain) and its severity (VAS; no pain=0 to worst pain=100mm) was assessed in 403 donkeys’ ante-mortem (ATM) and post-mortem (PTM). Also behaviours/signs (BS) and pain related lesions (PRL) were assessed ATM and PTM, respectively. Using principal component analysis the more than 53 BS and 238 PRL observed were narrowed to 58 biologically meaningful component or groups (14 BS and 44 PRL components, respectively). Components were used as risk factors in multiple regression analysis to identify which BS and/or PRL are commonly used in clinician’s (veterinary/pathologist) decision making process to determine whether a donkey ‘is’ (i.e. ATM) or ‘was’ (i.e. PTM) in pain and its severity (mild to severe). Furthermore, multiple correlations were made to understand which BS relate significantly with specific PRL and how.

A cross tabulation between pain ATM and PTM, where pain related lesions are used as a quasi-gold standard of pain assessment; identify that 2 in 10 donkeys are wrongly assumed as in NO-PAIN. Moreover, only 43% of the donkey observations are used by clinicians to make their opinion on donkey pain and its severity (i.e. 7 BS and 18 PRL components were significantly associated with pain as stated by clinicians). Yet, multiple correlations showed 20 plausible biologically meaningful relationships between BS and PRL; some currently not used by clinicians.

This methodology, previously successfully used in humans (Gregory, 2010) is novel to donkey veterinary medicine and warrants further research to consolidate findings. Nonetheless, the achieved correlation list of behaviours vs. pathologies is a significant work with valid applications in donkey pain identification and prognosis.


F. H. Ashley, A. E. Waterman-Pearson, H. R. Whay (2005) Behavioural assessment of pain in horses and donkeys: application to clinical practice and future studies, Equine Vet Journal, 37(6), 565-575.

N. G. Gregory (2010) Relationships between pathology and pain severities: a review. Animal Welfare 19, 437-448.

A novel approach of pain recognition and assessment in donkeys: initial results

Gabriela Olmos
Ayin Q. Alvarado-Arellano
Nicole du Toit
Faith A. Burden
Neville G. Gregory
Presentation date

This paper proposes an approach to use pain-relevant pathologies to enhance our understanding of the clinical and behavioural signs of pain in donkeys and outlines initial results of this ongoing investigation.

The methodology is summarized as follows. Trained veterinary clinicians examined live donkeys under two situations: A) before being euthanized due to a terminal illness or reduced quality of life (n= 347 sedentary donkeys in UK; DU) or B) when about to be slaughtered in an abattoir (n=164 working donkeys in Mexico; DM). The animals that represent populations in terms of age (years) for DU and DM respectively, were: <5 = 0.5%, 31.7%; 5-15 = 4%; 67.7%; 16-20 = 6%, 0.6%; >20 = 89.5% and 0%. For sex were: stallions 0.5%, 44%; geldings 52.2%, 8%; females 47.3%, 48% for DU and DM respectively. The body condition was: <2 =18.5%, 44.7%; 2.5 – 3 = 56.5%, 54%; >3.5 = 25%, 1% for DU and DM respectively and the girth was: 115 cm ±SD 9.6, 112cm ±SD 7.3 for DU and DM respectively.

The clinical examination (CE) included the oral mucosa, heart & respiratory rate, rectal temperature, plus an evaluation of 6 demeanours and 47 behaviours/signs that could relate to pain. At this point an overall pain visual analogue score was derived (VAS 0cm = no pain to 10cm = the worst pain). At post-mortem (PM), lesions/pathologies were noted, grouped by system-organ/tissue, ranked (mild, moderate, severe) and classified according to the following potentially painful pathologies: 1) trauma, 2) inflammation, 3) over-distension 4) perforation/ rupture, 5) stripping/ ulceration, 6) adhesions, 7) swelling, 8) exposure of sub-chondral bone. From these observations a second VAS was produced.

Raw correlations from the two populations showed that donkeys given a higher VAS at CE and PM presented with a greater severity of lesions in more systems as well as a higher heart rate at CE than those donkeys with a lower PM VAS. Moderate to severe pain identified CE was often recognized as severe pain at the PM stage. These initial observations show promise, and so further analysis will be done to test the relationships between pain indicators and pain pathologies.

Published as conference proceedings
Publication date
ISBN (13-digit)
Subscribe to pain recognition