A Useful Boost in the Treatment of Community Acquired Pneumonia?
Written and Presented by Dr Will Dean
Steroids have come in and out of vogue for many of the diseases we see and, in general, we are supportive of their use on NCCU where appropriately indicated. However, there is clear harm associated with their use, trauma, and TBI being commonly cited examples, however there is perhaps an argument for revisiting some of these questions. The harm largely arose from increased incidence of hospital acquired infections – we are arguably better now at infection control but we also have more drug resistant organisns.
Pneumonia has perhaps seen the most well defined swings around the use of steroids over the last decade or two and maybe represents the ideal disease model to study the effects of steroids in critical illness, provided of course that the inclusion criteria are tight enough to exclude the various forms of chronic chest diseases exacerbated by recurrent acute infections. In any case, the question of whether to use steroids in severe pneumonia perhaps now has an answer.
We have all seen and treated many patients with community acquired pneumonia (CAP) and are aware of the high mortality associated with this condition. In fact case mix data from ICNARC suggests an in-hospital mortality of 40% for patients admitted to critical care with pneumonia – it is something we have to get right.
Steroids have been proposed as an adjunct in the treatment of community acquired pneumonia – dampening of the inflammatory cascade may act to reduce the damage caused by the host response. Evidence for their efficacy has been sought for many years in RCTs and other studies (for example from 2011) and the trend has been towards benefit, albeit non-significant. A now key study published in 2015 (JAMA, cautious editorial) appears to provide the robust evidence needed, and a timely meta-analysis (editorial and associated review) supported the use of steroids in severe CAP.
However, less than a year later this week’s study, another meta-analysis (correspondence), is less positive about the role of steroids in CAP. Briefly, the authors found that whilst the data trended towards a survival benefit, especially in severe CAP, no significant reduction in mortality could be seen. Benefit was also found in a number of secondary outcomes, namely length of hospital and ICU stay, with a reduction in IV antibiotic treatment length. There were no increases in adverse events and in fact steroid use appeared to reduce the incidence of ARDS. Moreover, and perhaps better illustrated by Siemieniuk et al., the 6% rise in hyperglycaemia events should not be underestimated, an adverse effect we are keen to avoid in our NCCU population.
The difference between these studies opens an interesting discussion around the fragility faced by systematic reviews. Ultimately the difference between these analyses resulted from inclusion of different studies and different statistical analysis. The inclusion of studies from 1956 and 1972 by Siemieniuk shifted the balance towards statistical significance. It is difficult to argue that these studies can be representative of clinical practice today. This excellent editorial advises caution in interpretation the benefit of steroids in CAP.
The real issue is this – surprisingly, pooling of all available data revealed just 347 patients from RCTs assessing steroid effect on severe CAP – why is this number so low?
Some other questions to think about…
Why don’t we routinely give steroids in the treatment of severe CAP?
What types of pneumonia do we see on NCCU and are these studies generalisable to our population?
How much of this data is CAP specific or are we actually analysing the role of steroids in sepsis?
If we agree steroids are beneficial, what steroids should be used and at what dose?
Are all meta-analyses the gold standard of evidence we hold them to be?