Even with the prescription of newer and more potent antimicrobial agents, mortality due to CAP has remained relatively constant over time. Corticosteroids have been proved to block several arms of the inflammatory cascade. Available evidence so far, derived from a few, mainly small randomized controlled trials, suggests that corticosteroid treatment might be beneficial in terms of survival in patients with CAP [6, 9, 10, 34–36].
Most studies showing a beneficial effect of corticosteroids in patients with CAP only included severe pneumonia, as defined with a PSI class IV and V [11, 34, 37, 38]. In these patients, reviews and meta-analyses [13–16] pointed to a possible benefit of corticosteroids on either length of stay or mortality without evidence of more complications. However, insufficient data on non-severe CAP are available. Recently, two randomized double-blind controlled trials including 200 to 300 patients with all severities of CAP found controversial results: the first study using 40 mg of prednisolone for seven days showed no effect on time to clinical stability, with a higher rate of late failure in the corticosteroid group, defined by recurrence of signs and symptoms of pneumonia after 72 hours of admission after an initially beneficial response to treatment . The second study using 4 days of 5 mg of dexamethasone intravenously showed a significantly shorter duration of hospital stay in patients treated with corticosteroids [6, 12]. In view of these controversial findings, a large prospective and adequately powered trial should conclusively determine the risks and benefits of adding corticosteroids to the treatment of patients with CAP.
In the current trial, we will include all severity levels of CAP (PSI I to V) requiring hospitalization. This will lead to a lower mortality rate than observed in previous studies only including severe CAP patients, and our study will not be powered to show a survival benefit. We therefore chose as primary endpoint time to clinical stability, which is a less stringent but still clinically relevant endpoint. By reducing the time to clinical stability and consecutively also length of stay (LOS), nosocomial complications like infections, thrombembolic events, worsening of pre-existing frailty or delirium may be prevented. This leads to better allocation of resources at the hospital and to a marked reduction of costs . While Meijvis et al. looked at LOS , our primary endpoint will be time to clinical stability, as defined by IDSA/ATS guidelines . As we have previously observed, LOS itself in community-acquired pneumonia may be confounded and prolonged by medical or organizational problems unrelated to CAP [39–41]. We believe that by measuring time to clinical stability, we will be looking at a clinically relevant parameter, as patients are able to switch to oral antibiotics and ready for discharge once they reach stability.
Inclusion of all patients with CAP will increase the generalizability of the findings and allow retrospective evaluation in exploratory analyses as to which patients indeed profited most from corticosteroid treatment.
In our study, we administer an oral dose of 50 mg prednisone for 7 days without tapering. We opted for an oral formula, as the bioavailability of oral prednisone is excellent, it may be ground and is therefore applicable in patients not able to swallow, such as in patients on mechanical ventilation or older-aged patients. Furthermore, by choosing a uniform application formula, further bias, like the ability to blind and to secure provision of blinded study medication 24 hours on 7 days per week, was prevented.
The ideal choice and duration of corticosteroid treatment in critically ill patients and patients with CAP is currently disputed. Hydrocortisone is the natural corticosteroid and was used in some studies in critically ill patients, partly together with a mineralocorticoid . Most patients in our study, however, will not suffer from critical illness and will be hospitalized outside the ICU on the wards where oral intake of treatment is preferred. Since hydrocortisone is only available in 10 mg tablets, patients would have to take 20 tablets a day, which does not appear feasible. Methylprednisolone and prednisolone have also been used in some studies including patients with CAP and ARDS [10, 12, 43, 44], putting it forward as an alternative formulation for our study, as adequate oral formulations are available. Prednisolone is the pharmacologically active metabolite of prednisone, which was used in this study; both are used interchangeably except in patients with severe liver insufficiency, where prednisolone is preferred. Dexamethasone was not recommended by guidelines for use in septic shock and ARDS at the time of study design . However, a recent study showed a reduced length of hospital stay in CAP patients treated with dexamethasone as compared to placebo , also putting it forward as a possible type of corticosteroid in patients with CAP. The additional use of fludrocortisone as in the study by Annane et al.  is currently considered optional . Furthermore, the optimal duration of glucocorticoid treatment and mode of discontinuing (tapering or not) is not known and heavily disputed [45, 46].
In high-dose glucocorticoid studies, which used as high as 1 to 3 g hydrocortisone per day, short-term complications, such as co- and re-infections, led to excess mortality [47, 48]. In so-called low- or physiological-dose glucocorticoid studies with 100 to 300 mg hydrocortisone-equivalent per day, the main adverse effects included hyperglycemia  and rebound pneumonia . Although confusion, co-infection and gastrointestinal side effects were reported in association with glucocorticoid treatment, the incidence was not significantly higher in the steroid group . While some authors advise five days of glucocorticoid treatment in critically ill patients without tapering [49–51], others recommend tapering of glucocorticoids to avoid a rebound of inflammatory markers with consecutive rebound pneumonia [37, 45, 46, 52]. However, it has been shown that glucocorticoids given > seven days lead to a worse clinical course measured by length of stay, clinical stability and mechanical ventilation, and more systemic complications when looking at shock and cardiac arrhythmia . From COPD studies, there is evidence that stopping glucocorticoids without tapering is safe without an increased recurrence rate . For these reasons, we decided against tapering of the glucocorticoid dose in order not to prolong glucocorticoid exposure.
In conclusion, this current large and adequately powered randomized trial will determine the risks and benefits of adding 50 mg of prednisone for 7 days to the treatment of hospitalized patients with CAP. Our hypothesis is that corticosteroid treatment will result in a reduced time to clinical stability in CAP compared to placebo treatment.
The results of this trial may only be valid for the specific type, dose and duration of glucocorticoid studied.
This study is powered for clinical stability, but not for mortality; at least 4,000 patients would be needed to be powered to detect a clinically relevant difference in mortality.
Furthermore, we also included lower severity CAP patients, which might weaken the effect of glucocorticoids. A conclusion for ICU patients might not be possible as the number of these patients included will be low. There is an ongoing ICU multicenter study (clinicaltrials.gov number NCT01448109) aiming for the goal of 4,000 patients that will hopefully answer this question for ICU patients with CAP.