Clinical Investigations in Canine Sepsis
ACVIM 2008
Jamie M. Burkitt, DVM, DACVECC
Davis, CA, USA


Sepsis is the systemic inflammatory response to infection, and carries a published mortality rate of up to 71% in the dog.1 Sepsis secondary to bacterial infection appears to be more common in dogs than viral or fungal sepsis. Sepsis secondary to parvoviral enteritis is also common, though it is usually unclear whether the sepsis is truly viral in origin, or secondary bacterial sepsis due to loss of gut barrier function.

A large body of literature on septic syndromes in humans has been published in the last twenty years, but relatively little information has been available about naturally occurring sepsis in the dog. As the veterinary critical care community expands and as clinical veterinary patients begin to be used as models for human disease, interest in the clinical course of canine sepsis is increasing. More literature about clinical sepsis in dogs has become available in the last ten years.

The Clinical Presentation of Canine Sepsis

In 1992, members of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee published guidelines for the identification and classification of humans with septic syndromes.2 This group of experts identified abnormalities in body temperature (T°), heart rate (HR), respiratory parameters, and the leukogram as signs used together to define systemic inflammation in humans. In 1997, Hauptman and others3 evaluated rectal T°, HR, respiratory rate (RR), white blood cell count (WBC), percent immature neutrophils (%imm), platelet count, and serum glucose concentration as markers of clinical canine sepsis. They found that the presence of at least two of the following (a-d) was 97% sensitive and 64% specific for sepsis in the dog: a) rectal T° < 100.6 or > 102.6; b) HR > 120; c) RR > 20; d) WBC < 6,000/µL or > 16,000/µL, or > 3%imm. However, since publication of Hauptman's results, others have used different, though similar, criteria to identify sepsis in dogs,4,5,6 probably in an attempt to increase specificity for the purpose of study subject inclusion. Hence, no consensus definition of clinical sepsis exists for the dog (or for any other veterinary species) at this time. A 2005-06 survey of ACVECC and ACVIM diplomates, veterinarian members of the Veterinary Emergency and Critical Care Society, residents in approved emergency and critical care residency training programs, members of the Student Veterinary Emergency and Critical Care Society, and attendees at the 2006 ACVECC Sepsis in Veterinary Medicine Course yielded some information regarding identification criteria for canine sepsis, but provided no consensus.7 Clinical canine sepsis is likely best identified using some combination of the seven parameters evaluated by Hauptman and others. However, due to lack of more specific criteria, it seems reasonable to consider sepsis a possibility in any systemically ill dog displaying signs consistent with the syndrome (e.g., clinical signs of vasodilation, tachycardia, tachypnea, fever or hypothermia, hypoglycemia, hypotension resistant to fluid loading).

Etiologies of Canine Sepsis

The inflammatory pathway that creates the clinical systemic response to sepsis is the same regardless of underlying infectious agent or primary infected organ. Therefore, systemic inflammatory signs in a dog with fungal pneumonia are basically clinically indistinguishable from those in a dog with bacterial cellulitis.


Much of the veterinary literature focused on naturally occurring canine sepsis regards dogs with parvoviral enteritis; parvovirus is indeed an important underlying etiology of clinical canine sepsis. There are also a number of studies regarding the surgical management of septic peritonitis in dogs, the vast majority of which must have been septic. However, none of these studies were designed to evaluate sepsis per se, so most don't confirm systemic inflammation in the study population. One study including 84 dogs with septic peritonitis by Bentley and others8 found that positive cultures of abdominal effusion most commonly yielded purely gram-negative bacterial organisms (21/54 cultures; 39%), whereas 18/54 (33%) grew both gram-positive and gram-negative bacteria, and only 15/54 (28%) grew only gram-positive bacteria. The most commonly isolated bacterial organisms in that study were, from most common, E. coli, Enterococcus spp., Clostridium spp., Staphylococcus spp., and Enterobacter cloacae. No statistical associations between bacterial organism or Gram status and outcome were reported.

Less information is available regarding etiologic organisms in other populations of dogs with sepsis. A recent investigation of twenty dogs with culture-confirmed sepsis from heterogeneous anatomic sources found that 37% grew only gram-negative bacterial organisms on culture, 37% grew both gram-positive and gram-negative organisms, and 26% grew only gram-positive organisms.4 The most common isolate was E. coli, followed by Streptococcus spp. and Enterococcus spp. In a group of 47 prospectively-identified dogs with confirmed sepsis from heterogeneous anatomic sources that were admitted to the Small Animal Intensive Care Unit at the University of California, Davis, we found that 45/47 (96%) had bacterial sepsis and 2/47 (4%) had fungal sepsis. Of the dogs with bacterial sepsis, 21% had purely gram-negative cultures, 42% had mixed gram-positive and gram-negative cultures, and 37% had purely gram-positive cultures; E. coli was by far the most commonly cultured organism.9

Primarily Infected Site

Since such a large proportion of the clinical canine sepsis literature has historically focused on parvovirus and septic peritonitis, the gastrointestinal tract and other intra-abdominal organs are by far the most commonly reported anatomic sites of infection. In their investigation of twenty septic dogs, de Laforcade and others reported infections of the peritoneal cavity in 7/20 (35%), reproductive tract in 5/20 (25%), and respiratory system/pleural space in 4/20 (20%).4 In 47 septic dogs we found the most commonly infected anatomic sites were the peritoneal cavity (n = 14; 30%), the subcutaneous tissues (11; 23%), and the lung (9; 19%).9

Associated Morbidities

Multiple Organ Dysfunction Syndrome

Sepsis is associated with distant organ dysfunction--dysfunction or failure of organ(s) not affected at the onset of sepsis--likely due to tissue damage from microangiopathy, cytokines, and reactive oxygen species. This distant organ dysfunction is called multiple organ dysfunction syndrome (MODS) or multiple organ failure (MOF). In order to be considered MODS, organ dysfunction must be severe enough to require therapeutic intervention to maintain homeostasis; patients who develop MODS are considered to have severe sepsis.2 Simple sepsis is associated with a relatively low mortality rate in people; however, it is well established that survival is significantly worse (as poor as 50% or worse) in humans with severe sepsis. In humans, mortality is directly related to the number of failed organs, and reaches 100% with four organs failing. The incidence of MODS in dogs with sepsis is unknown. We recently found MODS in 21/47 septic dogs (45%). Distant organ dysfunction was seen more commonly in dogs with a surgical disease process compared to those with a non-surgical septic source, and was associated with death. As with in humans, death was predicted by an increasing number of failing organs. Fourteen of the 21 (67%) dogs with distant organ dysfunction died or were euthanized for grave prognosis.9

Septic Shock

Septic shock is defined as sepsis with fluid-refractory hypotension associated with signs of hypoperfusion.2 Vasodilation and vascular hyporesponsiveness are likely due to endothelial dysfunction and vascular smooth muscle insult from cytokines, tissue hypoxia, and reactive oxygen species. Despite decades of effort in the laboratory and clinic, septic shock still carries a high mortality rate of 50-60% in humans. The mortality rate appears to be higher in dogs, in which mortality may approach 100%, though adequate data are lacking in the species. Ten of 47 (21%) septic dogs in our investigation developed septic shock, all of which died or were euthanized for impending death.9 Vasopressors are required for patients that develop hypotension despite adequate fluid loading. Recent data indicate that human patients that remain hypotensive despite vasopressor therapy may benefit from low-dose hydrocortisone therapy, which may shorten or reverse the episode of septic shock.10,11 A multi-center trial is currently ongoing to study the effects of low-dose hydrocortisone in dogs with septic shock.


Severe sepsis and septic shock carry mortality rates of up to 50-60% in humans despite aggressive therapy. Human mortality from septic shock varies significantly with certain factors; for instance, increasing age, certain ethnicities, male gender, presence of co-morbidities, pulmonary or renal failure, and fungal etiology are each associated with worse survival in people.12 Reported mortality in canine septic syndromes varies from as low as 29%13 to as much as 71% for dogs that develop sepsis as a post-operative complication.1 Mortality from septic syndromes in dogs likely varies by certain predispositions, underlying etiology, and organ failures as it does in humans, though no investigations have as yet proven this. One recent investigation found that Doberman Pinschers and Rottweilers had higher TNF-α production after lipopolysaccharide stimulation in vitro than mixed-breed dogs,14 suggesting that certain dog breeds may have predispositions for development of more severe sepsis than others. One investigation found significantly greater mortality in septic dogs that had a poor increase in serum cortisol concentration after a standard ACTH stimulation test.6 In the population of 47 septic dogs in our Intensive Care Unit, hypoglycemia, total number of failing organs, and the presence of septic shock were each associated with death or euthanasia for severe illness.9 Further investigation is warranted to determine factors that may predict mortality in septic dogs.

Future Directions

In an effort to compile data required to form a consensus on septic syndromes in veterinary species, Otto and others have generated a free online sepsis registry open to all veterinarians and veterinary technicians. The URL is: for those who wish to join the group and contribute.7 One multi-center trial evaluating feline sepsis has been performed to date,15 and another multi-center trial in dogs with septic shock is currently underway. Multi-center and multi-specialty collaboration will be essential in the timely collection of meaningful data.


1.  Hardie EM, et al. J Am Anim Hosp Assoc 1986;22(1):33-41.

2.  ACCP/SCCM Consensus Conference Committee. Crit Care Med 1992;20(6):864-74.

3.  Hauptman JG, et al. Vet Surg 1997;26:393-7.

4.  de Laforcade AM, et al. J Vet Intern Med 2003;17:674-9.

5.  Beck JJ, et al. J Am Vet Med Assoc 2006;229(12):1934-9.

6.  Burkitt JM, et al. J Vet Intern Med 2007;21:226-31.

7.  Otto CM. J Vet Emerg Crit Care 2007;17(4):329-32.

8.  Bentley AM, et al. J Vet Emerg Crit Care 2007;17(4):391-8.

9.  Burkitt JM, et al. Unpublished data.

10. Annane D, et al. JAMA 2002;288(7):862-71.

11. Sprung CL, et al. N Engl J Med 2008;358(2):111-24.

12. Annane D, et al. Lancet 2005;365(1):63-78.

13. Staatz AJ, et al. Vet Surg 2002;31:174-80.

14. Nemzek JA, et al. J Vet Emerg Crit Care 2007:17(4):368-72.

15. Costello MF, et al. ACVECC Postgraduate Course 2006: Sepsis in Veterinary Medicine; 41.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Jamie Burkitt, DVM, DACVECC
University of California
Davis, CA

MAIN : VECCS : Canine Sepsis
Powered By VIN