Steroid treatment given for community acquired pneumonia (CAP) has been looked at a number of times in the last decade, often in subtly different groups of patients with varying regimens of drug and dose. Almost all have shown benefit (eg Lancet 2011) . The assumption is that they reduce lung inflammation (proven CRP and IL-6 reduction) without causing significantly detrimental immune suppression, bleeding, or hyperglycaemia.
2 more recent studies now broadly support this:
1. Torres et al performed an RCT in 3 Spanish hospitals over 8 years. 120 patients with ‘severe’ CAP (CRP > 150mg/l) were given intravenous bolus of 0.5 mg/kg per 12 hours of methylprednisolone or placebo for 5 days, started within 36 hours of hospital admission.Treatment failure was the primary outcome.
(Early treatment failure was development of shock, new need for mechanical ventilation , or death within 72 hours. Late treatment failure defined as worse CXR, persistence of severe respiratory failure, or any early treatment failure criteria between 72 hours and 120 hours.)
There was less treatment failure in the steroid group (13 v 31%). Looking more closely it was the late treatment failure that was more affected, particularly the occurrence of septic shock. No statistically significant difference in hyperglycaemia.
Only 120 patients in 8 years raises concerns about recruitment and consequently external validity. Also there was no mortality difference between the groups – although the study was never properly powered to detect this it seems strange the dramatic benefit in treatment failure didn’t translate to survival.
Could this result be entirely explained by the haemodynamic benefits of steroid in the septic patient? Difficult to comment much given such small numbers (8 v 18 patients had ‘treatment failure’. Beyond this, it at least appears that methylprednisolone doesn’t do harm in this group.
2. A bigger study by Blum et al (which was to be called the STEP trial) examined CAP all-comers at 7 hospitals in Switzerland. Nearly 800 got either prednisolone 50mg for a week, or placebo. Strikingly more than two thirds of those initially assessed were ineligible (dementia, steroids for other reasons, GI bleeds).
The primary outcome here was return to clinical stability – a normal set of observations for 24hr. Antibiotic therapy was fairly loosely protocolised.
With this approach the intervention group stabilised quicker (3 v 4.5 days). Half of these patients had severe pneumonia (PSI class 4 or 5) but only around 5% were admitted to ICU again making it small numbers for us to extrapolate from. The treatment effect did appear stronger in those with sepsis.
Respiratory complications (intubation, empyema etc) were less common in the steroid group. However hyperglycaemia was again more prevalent in the steroid group though this association didn’t last beyond 30 days.
2 more encouraging studies that show a benefit of steroid in community acquired pneumonia. However, they used composite surrogate outcome measures – neither were powered for looking at mortality – and neither focused on the group that comes to the ICU. So for critical care per se there remains a hypothesis-generating signal that steroids might help. In addition a focused study would need to contend with the fact that many ICU cases get steroids for treating shock.
See also:
