The quick sequential organ failure assessment score as a predictor of mortality in septic dogs

Imagine this clinical scenario...

You are a small animal veterinarian working at a referral hospital. You decide to review your processes for identifying which dogs with suspected or confirmed sepsis are at high risk of mortality. You are aware of the quick Sequential Organ Failure Assessment (qSOFA) to facilitate earlier interventions and improve patient outcomes, but do not know how effective it is as a predictor of mortality. You decide to consult the evidence to find out more.

The evidence

Four relevant papers were identified, all of which were retrospective cohort studies. Each paper examined the use of qSOFA to identify patients at higher risk of mortality, and used survival or non-survival as the outcome studied.

Donati et al. (2022) retrospectively reviewed the cases of 52 female canine patients referred to a veterinary hospital for treatment of pyometra between 2013 and 2019. Each patient was assigned a qSOFA score between 0 and 3.

Orlotani & Bellis (2021) retrospectively reviewed the cases of 267 dogs, admitted to an intensive care unit (ICU) between 2017 and 2018. 14 of the dogs were septic and 253 were non-septic. All patients were assigned a qSOFA score between 0 and 3.

Osgood et al. (2022) retrospectively reviewed the cases of 90 dogs admitted to a veterinary hospital between 2010 and 2019 with clinical signs of sepsis. The researchers searched digital medical records using the search terms ‘septic’ and ‘sepsis’. 45 dogs had Systematic Inflammatory Response Syndrome (SIRS), 38 dogs had severe sepsis, and 7 had septic shock. Patients were assigned a qSOFA score between 0 and 3.

Stastny et al. (2021) retrospectively reviewed the cases of 204 canine patients with a clinical diagnosis of sepsis at a university hospital between 2011 and 2018. 143 had septic peritonitis, 26 had septic soft tissue infections, 20 had pyometra, and 15 had pyothorax. Patients were assigned a qSOFA score between 0 and 3.

Limitations of the evidence

The overall strength of evidence was weak. All four papers were retrospective studies which is not a robust form of study design. Three papers (Donati et al., 2022, Orlotani & Bellis, 2021, Stastny et al., 2021) did not use a control group at all, while Osgood et al. (2022) used a control group of dogs diagnosed with non-infectious SIRS, rather than healthy dogs, which would have provided a more useful control variable. Three papers (Donati et al., 2022, Orlotani & Bellis, 2021, Stastny et al., 2021) also failed to compare the use of qSOFA to other illness severity scores, which makes it more difficult to evaluate the efficacy of qSOFA.

Three papers (Orlotani & Bellis, 2021, Osgood et al., 2022, Stastny et al., 2021) failed to include details of treatment the patients had received prior to being referred to hospital, which may have had an impact on scoring and clinical outcomes.

Donati et al. (2022) only studied dogs with pyometra, which makes its results less applicable to other forms of sepsis. SIRS criteria timing was also not recorded from presentation, but from the onset of infection.

Ortolani & Bellis (2021) studied a large overall cohort of dogs, but only 14/267 of those were septic, which limits the applicability of the study to this topic. The survival rate of this group was also high (71.4%). Additionally, in cases where septic patients were euthanised, it was not disclosed whether this was for financial or clinical reasons.

Osgood et al. (2022) excluded some patients from their cohort due to incomplete medical records, which may have led to selection bias. There was a further risk of selection bias owing to the use of both SIRS criteria and qSOFA. The fact that researchers used ‘septic’ and ‘sepsis’ as search terms to identify relevant cases could have meant some infectious SIRS patients were missed.

Stastny et al. (2021)’s study may have been susceptible to population bias due to 143/204 of the dogs having septic peritonitis. This risk was compounded by the exclusion of dogs with milder signs of sepsis. Dogs that were euthanised without treatment were also excluded from the study, which may have artificially decreased the mortality rate.

Summary of findings

Donati et al. (2022) found that the odds of mortality were 6.5 times higher in patients with a qSOFA score of 2 or higher, than in patients with a qSOFA score lower than 2. The overall survival rate recorded was 65% (34/52), and of the 18 non-survivors, 7 were euthanised for clinical reasons.

Ortolani & Bellis (2021) found that there was no significant difference in qSOFA score when comparing survivors and non-survivors. Of the dogs with sepsis, 10/14 survived and 4/14 died.

Osgood et al. (2022) found that the qSOFA score was not predictive of non-survival. 29/45 dogs with either severe sepsis or septic shock died, while 16/45 dogs with SIRS died. Of the 29 dogs with severe sepsis or septic shock that died, 17 were euthanised – 2/17 for financial reasons and 15/17 for poor prognosis.

Stastny et al. (2021) found that a qSOFA score of 2 or higher was associated with mortality, and that the qSOFA scores between survivors and non-survivors were significantly different. 63 patients were graded with a qSOFA score of 2 or higher, and 68.3% of them (43/63) did not survive.

Conclusion

The available evidence does not support the use of qSOFA as a standalone prognostic tool for dogs with sepsis. It may be a useful supplementary tool for early identification of at-risk patients but on the current evidence it should not replace clinical judgement when making care decisions. A low qSOFA score should not be used to delay investigations or treatments in isolation and should be used in conjunction with other observations. However, a higher score or an increase in score over time may be used to identify patients that are at risk of deterioration.

With all of this said, it must be noted that the overall strength of evidence is weak, and further research is required on this topic, such as the validation of qSOFA using canine rather than human parameters, and preferably within blinded, randomised control trials.

Disclaimer

The application of evidence into practice should take into account multiple factors, not limited to: individual clinical expertise; patient’s circumstances and owner’s values; country, location or clinic where you work; the individual case in front of you; and the availability of therapies and resources.