Promising prognostic accuracy of Sepsis-3 criteria for in-hospital mortality among patients with suspected infection presenting to the emergency department

Thank goodness for the Third International Consensus Definition for Sepsis and Septic Shock! (Yes, this is an odd statement…) They have done away with the terrible SIRS criteria and put the sepsis back in sepsis.

Sepsis has reverted to “a life-threatening organ dysfunction caused by a dysregulated host response to infection.” This is always what we had thought it to be… not the young person with man-flu.

SIRS has been criticised for having poor test characteristics so now we have SOFA and the easier qSOFA

qSOFA is (0-3):

• Hypotension (SBP less than 100)
• Alterred mental status
• Respiratory Rate >22

Yeeeeehaaaa... even I can remember 3 things!

The purpose of this study was to externally validate the qSOFA score among patients presenting to the ED and compare them to the old tools (SIRS, lactate, etc) to predict mortality.

This was a multicentre prospective cohort study mostly done in French ED’s that analysed 879 patients with sepsis. There was an overall 8% mortality rate.

How good was qSOFA?

Pretty darn good (if you believe this study to be valid).

If the qSOFA was less than 2, mortality was 2%. For 2 or greater it was 24%.

The area under the curve (AUC) is a plot of the true positives (sensitivity) on the y-axis and the false positives (1- specificity) on the x-axis. This is a great way to evaluate & compare diagnostic/prognostic tests that have continuous or ordinal values. An AUC of 0.5 would be a useless test (i.e. equal number of true to false positives). 1.0 would be perfect and of course does not exist.

What was the AUC for qSOFA?

0.80 (95%CI 0.74-0.85). This AUC would be considered excellent but certainly not perfect.

This was much better than SIRS which was marginal 0.65 (95% CI 0.59-0.70)

Of course, no study is without problems. This study used the worst qSOFA measurement during the patient's ED stay which increased the sensitivity. In the end, I doubt this tool is really that good.

Some aspects of qSOFA have face validity. I don’t need a decision instrument to tell me that patients with altered mental status and hypotension will do worse!

I also don’t know how this score will be in screening for sepsis. It strikes me as overly simplistic.

Sepsis is a complex and heterogeneous disease process. Clinical decision instruments have never been that useful in these circumstances. They often fail as no fancy regression analysis can account for all of the intricacies of complex disease and individual variability.

The good news is we can use experienced clinician judgement… go figure! In addition, you can incorporate aspects of SOFA in your decision making.  No doubt there will be more to come.

Covering: 

Freund Y, Lemachatti N, Krastinova E, et al, Prognostic Accuracy of Sepsis-3 Criteria for In-Hospital Mortality Among Patients With Suspected Infection Presenting to the Emergency Department. JAMA. 2017;317:301-308.