Last week we saw the publication of the ProCESS study suggesting that protocolised treatment of sepsis makes little difference to outcome. A little earlier we heard that cooling to 33C vs 36C may not make much difference either.
Seriously, what the bally hell is going on with resuscitation these days? It seems as though interventions that we once trusted seem to make little difference and this is depressing, or is it?
There are clearly many reasons why ProCESS and TTM found few differences in their comparative therapies but what was interesting in recent trials and especially in the TTM trial was the huge difference in mortality as compared to historical controls.
The same may arguably be true of sepsis. Over the last 15 years we have seen, adopted, abandoned, reinvented, revised and reversed all sorts of treatments for severe sepsis. Have you given steroids, Protein C, starch solutions in that time? I certainly have, haven’t, have again, and then not so much. An ever changing world that seems to evade our search for the magic bullet treatment that will save our patients. The ebb and flow of treatments can create an illusion of status, a lack of direction, a lack of progress or success, but is this true.
This month there is an early publication in JAMA from Aus/NZ demonstrating a new therapy which you may have heard of, one that you’ve probably used and which you may use again and again. Not only is it familiar to you, it is arguably free, easy to access and widely available.
It also seems to work.
The Australian and New Zealand intensive care research centre have published longitudinal data on the mortality related to severe sepsis across 171 ICUs from 2000-2012. A total of 101,106 patients are included over 12 years. Read the abstract and full paper here.
This retrospective observational study is an excellent way of tracking trends for mortality for a specific condition, in this case Severe Sepsis.
Definitions are important so here SS is defined according to ACCP/SCM guidelines which should be valid, reliable and reproducible from database records. The details are well outlined in the paper with roughly a tenth of patients in the larger database of ICU admissions meeting criteria for SS.
The major results?
The principle outcome of mortality is hard to argue with, it’s important and relevant to patients, easy to measure and easy to track change over time.
Download the principle outcome graphs for mortality over time here.
So the headline results are that mortality from SS has fallen from 35% in 2000 to 18.4% in 2012. That’s a huge difference and if you want to express it as an NNT then it works out as 5.8. In other words for roughly every 6th patient admitted to ICU in 2012 as compared to 2000, we save a life and that’s fantastic. The trend appears to be linear with mortality steadily decreasing over this time. Mortality has almost halved over 12 years
Sub analyses looking at the effect of hospital size, location, level, admission source, APACHE III score or ICU type failed to show any statistical difference in the odds ratio of risk of death. The data is pretty consistent.
My only comment on the data is that the number of patients in the study by year group similarly changes year on year. In 2000 there were only 2708 patients with sepsis in the study as compared to 12512 in 2012. I am unsure why this is. Is it because of changing referral patterns or simply an increase in the number of units reporting into the study? If the former then this may have a significant impact on results as higher admission thresholds tend to increase mortality (as we have seen with comparisons between US and European mortality rates).
So what does this tell us?
One of my EM/CC colleagues summed this paper up as ‘we’re not always sure what we are doing, but it seems to be working’. Now I think that’s a little unkind as we are always working to the best of our knowledge but it has some truths in it. As with the principles of marginal gains mentioned on this blog, step changes are rare, difficult to find and often a false dawn. This study suggests that the reasons for change are sometimes difficult to pinpoint.
Whether the findings are due to changes in diagnostics, therapeutics, medical, nursing, physio or procedures is not clear and the authors findings that similar reductions in mortality for non-septic patients occurred over the same time period suggest that it is the overall package of ICU care that is important. Obviously my NNT of 5.8 is clearly hypothetical as no patient with severe sepsis can wait 12 years for treatment unless the next intervention is the invention of a time machine. We must do what we can for our patients right now according to the best available knowledge that we can muster.
So, I say to my critical care colleagues, chapeaux & well done, you may not always be able to know why you are doing a great job, but the stats suggest you are.
Time is indeed a great healer and you have done well.
2 thoughts on “JC: What has an awesome NNT of 5.8 for severe sepsis? St.Emlyn’s”
That’s great news, unless it is all due to changing definitions.
In the last 15 years we have discovered that we can make huge impacts in sepsis mortality by instituting a whole lot of care that actually doesn’t work. Because we can now do such wonderful things for sepsi the pressure is on to diagnose it early, rather than wait until it is staring us in the face. As a consequence, it is just possible, that we are now diagnosing with sepsis and admitting to ICU a heap of people who aren’t as sick as the people we used to diagnose with sepsis and admit to ICU and thus diluting the sample down.
I’d love to think that despite the fact that all these treatments turn out to not really do much, the thing that is really making a difference is the structured and aggressive approach to critical care that has developed over this time; that it is not the specific temperature target or the type of CVC but the well organised, close watching of the patient that is making a diffidence. I’d love to think that. I’m just not so sure and the huge increase in sepsis cases being treated in Australasian ICUs in that time without a huge increase in population does make me concerned that they are not the same patients that they used to be. Just like PE, and stroke.
Pingback: Rural Critical Care Pearls - KI Doc