Medical Matters: The difficulties and diversities of diagnosing sepsis
Like many doctors, I have seen a number of cases of sepsis in my professional lifetime. Broadly defined as a life-threatening disorder that arises when the body’s response to an infection injures its own organs and tissues, sepsis can lead to shock, multiple organ failure and death.
But it is a hugely variable clinical entity. I have seen healthy children become incredibly unwell within minutes as meningococcal infection literally races through their bodies in front of you. And I have also looked after some patients in their 80s whose blood pressure drops like a stone as they lose consciousness due to septic shock, but who defy your expectation that they will die by rallying unexpectedly.
Although first described by Hippocrates, sepsis remains a poorly understood medical condition. In Ireland, the sad case of Savita Halappanavar, who died in University Hospital Galway in 2012, brought sepsis into the public domain. Her demise after a severe infection complicated an incomplete miscarriage when she was 17 weeks pregnant may have grabbed international public attention for other reasons but a superbly conducted inquest gave us an unprecedented insight into the many unknowns that still surround sepsis.
Writing in the Lancet recently, a distinguished group of experts acknowledge that the present consensus definition, which includes a long list of probable signs of sepsis, is inadequate: “One approach to this issue has been the introduction of threshold decision-making, in which diseases are defined as being present when one or more predefined criteria are breached . . .” they write. “However, this approach relies on availability of a clinical or biochemical biomarker that can provide the so-called gold standard criterion for diagnosis. Unfortunately, no satisfactory biomarkers exist that can be used to diagnose sepsis, unlike disorders such as hypertension or diabetes mellitus, which have clear-cut, internationally accepted criteria.”
Increased levels of the chemical marker troponin in the blood of a person with chest pain? They have an acute coronary syndrome and may be on their way to having a heart attack. A hairline fracture on the ankle X-ray of someone who fell? Even though clinically you might have put your money on a bad sprain rather than a fracture, there is no denying the evidence of a sound diagnostic test.
However, sepsis remains a clinical syndrome without a reliable biomarker to prove or disprove the diagnosis. In the past 20 years, biomarkers have been the subject of intense research to try to identify molecules that might be early indicators of sepsis, predictors of risk, or likely response to treatment. Despite this major effort we still rely on relatively non-specific tests such as white blood cell counts and C-reactive protein levels to give us some sense of how severe each case of sepsis might be.
This does not surprise the authors of the Lancet review. “In view of the complexity of the pathogenesis of sepsis, the diversity of clinical presentations, and the heterogeneity of patient populations, to believe that one biomarker will prove to be ideal is arguably unrealistic,” they say, arguing this is more likely to be achieved by using a combination of disease markers.
Progress on the treatment front has also been frustrating. We have no specific anti-sepsis treatments so that mortality rates of up to 50 per cent persist in the case of septic shock.
Commenting on research efforts in the past 20 years, the experts note, “an unavoidable if uncomfortable truth is that despite many hundreds of clinical trials, involving tens of thousands of patients, and costing hundreds of millions of dollars, not one so-called novel drug has come to the market and passed the test of time by being incorporated into routine clinical practice”.
A breakthrough will occur. In the meantime, early recognition of the condition remains our best chance of beating sepsis.