Articles

Adjunct prednisone therapy for patients with

community-acquired pneumonia: a multicentre,
double-blind, randomised, placebo-controlled trial
Claudine Angela Blum*, Nicole Nigro*, Matthias Briel, Philipp Schuetz, Elke Ullmer, Isabelle Suter-Widmer, Bettina Winzeler, Roland
Bingisser, Hanno Elsaesser, Daniel Drozdov, Birsen Arici, Sandrine Andrea Urwyler, Julie Refardt, Philip Tarr, Sebastian Wirz, Robert
Thomann, Christine Baumgartner, Hervé Duplain, Dieter Burki, Werner Zimmerli, Nicolas Rodondi, Beat Mueller, Mirjam Christ-Crain

Summary
Background Clinical trials yielded conflicting data about the benefit of adding systemic corticosteroids for Lancet 2015; 385: 1511–18
treatment of community-acquired pneumonia. We assessed whether short-term corticosteroid treatment reduces Published Online January

time to clinical stability in patients admitted to hospital for community-acquired pneumonia. 19, 2015
http://dx.doi.org/10.1016/
S0140-6736(14)62447-8
Methods In this double-blind, multicentre, randomised, placebo-controlled trial, we recruited patients aged 18 years or
See Comment page 1484
older with community-acquired pneumonia from seven tertiary care hospitals in Switzerland within 24 h of presentation.
*These authors contributed
Patients were randomly assigned (1:1 ratio) to receive either prednisone 50 mg daily for 7 days or placebo. The
equally to this work
computer-generated randomisation was done with variable block sizes of four to six and stratified by study centre. The Endocrinology, Diabetology
primary endpoint was time to clinical stability defined as time (days) until stable vital signs for at least 24 h, and and Metabolism, Department
analysed by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00973154. of Internal Medicine and
Department of Clinical
Research (C A Blum MD,
Findings From Dec 1, 2009, to May 21, 2014, of 2911 patients assessed for eligibility, 785 patients were randomly N Nigro MD, I Suter-Widmer MD,
assigned to either the prednisone group (n=392) or the placebo group (n=393). Median time to clinical stability B Winzeler MD, B Arici MD,
was shorter in the prednisone group (3·0 days, IQR 2·5–3·4) than in the placebo group (4·4 days, 4·0–5·0; hazard S Andrea Urwyler MD,
ratio [HR] 1·33, 95% CI 1·15–1·50, p<0·0001). Pneumonia-associated complications until day 30 did not di ffer J Refardt MD,
Prof M Christ-Crain MD), Basel
between groups (11 [3%] in the prednisone group and 22 [6%] in the placebo group; odds ratio [OR] 0·49 [95% Institute for Clinical
CI 0·23–1·02]; p=0·056). The prednisone group had a higher incidence of in-hospital hyperglycaemia needing Epidemiology and Biostatistics,

insulin treatment (76 [19%] vs 43 [11%]; OR 1·96, 95% CI 1·31–2·93, p=0·0010). Other adverse events Department of Clinical Research
(M Briel MD), and Emergency
compatible with corticosteroid use were rare and similar in both groups. Department (R Bingisser MD),
University Hospital Basel,
Interpretation Prednisone treatment for 7 days in patients with community-acquired pneumonia admitted to Basel,

hospital shortens time to clinical stability without an increase in complications. This finding is relevant from a Switzerland; Medical
University Clinic, Departments
patient perspective and an important determinant of hospital costs and efficiency. of Internal and Emergency
Medicine and Department of
Funding Swiss National Science Foundation, Viollier AG, Nora van Meeuwen Haefliger Stiftung, Julia und Endocrinology, Diabetology
Gottfried Bangerter-Rhyner Stiftung. and Clinical Nutrition,
Kantonsspital Aarau, Aarau,

Introduction testing a 7-day continuous infusion of hydrocortisone Switzerland (C A Blum,

P Schuetz MD, D Drozdov
Respiratory tract infections and pneumonia in particular versus placebo. A retrospective single-centre study 6 MD, B Arici, Prof B Mueller
are the third-leading cause of death worldwide. 1 Although including 308 patients suggested that the use of MD); Department of Clinical

outcome of community-acquired pneumonia improved corticosteroids was associated with decreased mortality. Epidemiology and
Biostatistics, McMaster
with the availability of antibiotics, this disorder still carries Two recent randomised placebo-controlled trials 7,8 University, Hamilton, ON,
a high risk for long-term morbidity and mortality. 2 Adjunct including 200–300 patients revealed controversial results. Canada (M Briel); Medical
therapeutic interventions could improve out-come of Whereas the first trial7 did not find any benefit of adjunct University Clinic, Kantonsspital

patients with this type of pneumonia. prednisolone, but an increased recurrence rate, the second Baselland/Liestal, Liestal,
Switzerland (E Ullmer MD,
In community-acquired pneumonia, an excessive release trial8 in which patients received intravenous H Elsaesser MD,
of circulating inflammatory cytokines can be harmful and dexamethasone over 4 days reported a significant Prof W Zimmerli MD); Medical
cause pulmonary dysfunction. Systemic corticosteroids reduction in length of hospital stay by 1 day. Two University Clinic, Kantonsspital

have anti-inflammatory effects, atten-uating the systemic systematic reviews9,10 and three meta-analyses11–13 Baselland/Bruderholz,
Bruderholz, Switzerland
inflammatory process in the disorder. 3 Therefore, adjunct concluded that adjunct corticosteroids in community- (P Tarr MD, S Wirz MD);
treatment with corticosteroids has been discussed since the acquired pneumonia might be beneficial, but a large, Department of Internal
1950s, when favourable effects of corticosteroids were adequately powered randomised trial is warranted. Medicine, Bürgerspital,
Solothurn, Switzerland
noted in pneumococcal pneumonia.4 More recently, a Therefore, we investigated the effects of short-term (R Thomann MD); Department
significant reduction of in-hospital mortality in patients prednisone versus placebo in patients admitted to hospital of General Internal Medicine,
with severe community-acquired pneumonia was noted in for community-acquired pneumonia with the primary Inselspital, Bern University
a small randomised trial5 (n=46) endpoint of time to clinical stability. Hospital, Bern, Switzerland

www.thelancet.com Vol 385 April 18, 2015 1511

 Articles

(C Baumgartner MD, Methods requiring more than 0·5mg/kg per day prednisone
Prof N Rodondi MD); Clinic of
Study design and participants equivalent, pregnancy or breastfeeding, and severe
Internal Medicine, Hôpital du
Jura, Site de Delémont,
This is an investigator-initiated, multicentre, double-blind, immunosuppression defined as one of the following:
Delémont, Switzerland (H randomised, placebo-controlled trial. Details of the trial infection with human immunodeficiency virus and a CD4
Duplain MD); and Viollier AG, design have previously been published. 14 In brief, cell count below 350 cells per μL, immunosuppressive
Postfach, Basel,
consecutive patients presenting with community-acquired therapy after solid organ transplantation, neutropenia
Switzerland (D Burki MD)
pneumonia were screened and enrolled at emergency below 500 cells per μL or neutrophils of 500–1000 cells
Correspondence to: Prof
Mirjam Christ-Crain, Clinic of
departments or medical wards in seven tertiary care per μL during ongoing chemotherapy with an expected
Endocrinology, Diabetology hospitals in Switzerland from Dec 1, 2009, to May 21, decrease to values below 500 cells per μL, cystic fibrosis,
and Metabolism, Department 2014, within 24 h of presentation. Inclusion criteria were or active tuberculosis. The conduct of the trial adhered to
of Internal Medicine and
age 18 years or older and hospital admission with the declaration of Helsinki and Good Clinical Practice
Department of Clinical
Research, University Hospital
community-acquired pneumonia defined by a new Guidelines, and ethical committees of all participating
Basel, Petersgraben 4, 4031 infiltrate on chest radiograph and the presence of at least hospitals approved the study before patient recruitment.
Basel, Switzerland one of the following acute respiratory signs and symptoms: All patients provided written informed consent.
mirjam.christ@usb.ch
cough, sputum production, dyspnoea, core body temp-
erature of 38·0°C or higher, auscultatory findings of Randomisation and masking
abnormal breathing sounds or rales, leucocyte count higher Eligible patients were randomly assigned (1:1 ratio) to
than 10 000 cells per μL or less than 4000 cells per μL. 15 receive either 50 mg of prednisone or placebo daily for 7
Exclusion criteria were permanent inability for informed days. Randomisation was done with variable block sizes of
consent, active intravenous drug use, acute burn injury, four to six and patients were stratified at the time of study
gastrointestinal bleeding within the past 3 months, known entry by study centre. Allocation was concealed with a
adrenal insufficiency, a condition prespecified computer-generated randomisation list, which
was centrally kept at the
2911 patients assessed for eligibility pharmacy of the main study centre.
Patients were randomly assigned to receive a prepared
set of study medication that contained seven tablets of 50
1504 did not meet eligibility criteria
241 informed consent not possible
mg prednisone or placebo. The placebo drug was
(dementia, disability) purchased from a local prednisone manufacturer
508 with immunosuppression (Galepharm AG, Küsnacht, Switzerland), which pro-duces
667 with indication for steroids
88 with gastrointestinal bleeding both prednisone and its corresponding placebo. The drugs
within the past 3 months were prepared before the initiation of the study and packed
into identical containers by the Pharmacology Department,
605 eligible, but declined to participate
University Hospital, Basel, according to the randomisation
list. Patients, treating physicians, investigators, and data
assessors were masked to treatment allocation.
802 randomised

Procedures
402 assigned to prednisone 400 assigned to placebo
After informed consent was obtained, baseline blood
samples were drawn and nasal swabs for virus multiplex
PCR were done. All other microbiological assessments
10 blinded post-randomisation exclusion 7 blinded post-randomisation exclusion were at the discretion of the treating physicians. Patients
started antibiotic therapy as soon as community-acquired
392 included in intention-to-treat analysis 393 included in intention-to-treat analysis
pneumonia was confirmed. Treating physicians chose the
empirical regimen according to the ERS/ ESCMID
guidelines adapted for Switzerland. 16,17 Most patients
30 protocol violations 27 protocol violations
started this regimen either with amoxicillin plus clavulanic
18 informed consent withdrawn 12 informed consent withdrawn
for study medication for study medication acid or ceftriaxone alone. In patients with clinical
6 application mistakes 8 application mistakes suspicion for legionellosis or in those requiring treatment
6 study medication stopped 7 study medication stopped
1 active glucocorticoid indication 4 active glucocorticoid indication
in the intensive care unit (ICU), the beta-lactam was
5 potential adverse event 3 potential adverse event combined with clarithromycin. Treatment was streamlined
and optimised according to the susceptibility pattern as
soon as a specific pathogen was known. Thereafter,
362 treated per protocol 366 treated per protocol
patients started receiving study medication, and we
monitored timing in relation to start of antibiotics. Study
Figure 1: Trial profile
No patient was lost to follow-up before reaching the primary endpoint. One patient in the nurses assessed patients for clinical stability every 12 h
prednisone group and three patients in the placebo group were lost to follow-up at 30 days. during hospital stay. All patients were

1512 www.thelancet.com Vol 385 April 18, 2015

 Articles

treated according to published community-acquired we assumed a mortality rate of 10% in the placebo group
pneumonia guidelines.18 Stewardship of antibiotic and 7·5% in the corticosteroid group over 14 days of
treatment duration by procalcitonin was encouraged by the follow-up with a proportion of 75% survivors being
study protocol according to Schuetz and colleagues. 19 clinically stable after 7 days in the corticosteroid group.
Baseline data included medical history items, relevant Estimating a decrease in the risk of non-stability after 1
comorbidities, clinical items of pneumonia, and all week in survivors by 25% through adjunct corticosteroids,
variables required for the calculation of the pneumonia we calculated that we needed a sample size of 800 patients
severity index (PSI).20 Routine laboratory tests of followed up for at least 14 days to achieve a statistical
inflammatory markers (procalcitonin, C-reactive protein power of 85%.
[CRP], white blood cell count) were done in both groups
on days 1, 3, 5, 7, and before discharge and included four
glucose measurements per day. Prednisone (n=392) Placebo (n=393)

Structured follow-up telephone interviews for secondary General characteristics

outcomes after discharge were done on day 30 and Age, years 74 (61–83) 73 (61–82)
included assessment of adverse events such as infections, Male sex 241 (61%) 246 (63%)
recurrent pneumonia, re-admission to hospital, new onset Clinical variables
diabetes or insulin dependence, and new onset Days with symptoms 4·0 (2·0–7·0) 4·0 (2·0–7·0)
hypertension. Temperature (°C) 37·6 (37·0–38·2) 37·6 (37·0–38·2)
Systolic blood pressure (mm Hg) 124 (110–140) 123 (110–140)
Outcomes Heart rate (beats per min) 84 (74–95) 82 (72–96)
The primary endpoint was time to clinical stability defined Respiratory rate (breaths per min) 20 (18–24) 20 (18–24)
as time (days) until stable vital signs for 24 h or longer. SaO2 (%) 95 (92–96) 94 (92–97)
Stable vital signs were temperature of 37·8°C or lower, Bacteraemia 39 (10%) 48 (12%)
heart rate of 100 beats per min or lower, spontaneous Confusion 22 (6%) 29 (7%)
respiratory rate of 24 breaths per min or lower, systolic CAP score (points)* 43 (30–60) 46 (29–63)
blood pressure of 90 mm Hg or higher (≥100 mm Hg for Laboratory values
patients diagnosed with hypertension) without vasopressor Procalcitonin (ng/mL) 0·52 (0·18–2·51) 0·50 (0·17–2·63)
support, mental status back to level before occurrence of C-reactive protein (mg/L) 159 (80·3–245) 164 (79·1–250)
community-acquired pneumonia, ability for oral intake, White-blood-cell count (cells per μL) 12 200 (8900–15 800) 11 900 (8700–15 600)
and adequate oxygenation on room air (PaO₂ ≥60 mm Hg Glucose (fasting morning, mmol/L) 6·3 (5·4–7·8) 6·5 (5·8–7·7)
or pulse oximetry ≥90%), which were based on current PSI score†
community-acquired pneumonia treatment
PSI class I 47 (12%) 45 (11%)
recommendations.15 Instability was defined if at least one
PSI class II 72 (18%) 69 (18%)
of these criteria were not met.
PSI class III 71 (18%) 95 (24%)
Secondary endpoints were time to effective discharge
PSI class IV 148 (38%) 132 (34%)
from hospital, recurrence of pneumonia, re-admission to
PSI class V 54 (14%) 52 (13%)
hospital, ICU admission, all-cause mortality, duration of
Total PSI score (points) 93 (63–115) 86 (65–110)
total and intravenous antibiotic treatment, disease activity
Comorbidities
scores specific to community-acquired pneumonia, 21
Diabetes mellitus (any type) 77 (20%) 78 (20%)
incidence of complications due to community-acquired
Insulin treatment 44 (11%) 35 (9%)
pneumonia (ie, acute respiratory distress syndrome,
Chronic obstructive pulmonary disease 73 (19%) 60 (15%)
empyema, persistence of pneumonia), side-effects of
corticosteroids (ie, rate of hyperglycaemia, hypertension, Heart failure 80 (20%) 62 (16%)

delirium, nosocomial infections, and weight gain), and Cerebrovascular disease 38 (10%) 31 (8%)
time to earliest possible hospital discharge. Renal insufficiency 125 (32%) 126 (32%)
For patients admitted to ICU we recorded length of ICU Neoplastic disease 29 (7%) 25 (6%)
stay, time to transfer to ICU, time to discharge from ICU, Liver disease 17 (4%) 12 (3%)
duration of vasopressor treatment, and duration of Co-infections‡ 45 (11%) 46 (12%)
mechanical ventilation. Antibiotic pretreatment 84 (21%) 95 (24%)

Data are median (IQR) or number (%), unless otherwise stated. SaO 2=saturation of oxygen. PSI=pneumonia severity
Statistical analysis index. *The CAP score is a disease-specifi c activity score for community-acquired pneumonia; it ranges from 0 to 100,
The statistical analysis was prespecified, and investigators 0 marking the worst, 100 the best score.21 †The PSI is a clinical prediction rule to calculate the probability of morbidity
and mortality in patients with community-acquired pneumonia;20 PSI risk class I corresponds to age ≤50 years, and no
who analysed the data were masked to treatment
risk factors (≤50 points), risk class II to <70 points, risk class III to 71–90 points, risk class IV to 91–130 points, and risk
allocation.14 The primary hypothesis of this trial was that class V to >130 points. ‡Nine skin infections, 43 urinary tract infections, 28 upper respiratory tract infections, ten
corticosteroids will reduce time to clinical stability in gastrointestinal tract infections, and one joint infection.
patients with community-acquired pneumonia without
Table 1: Baseline characteristics of enrolled patients
relevant adverse effects. On the basis of previous trials, 19,22

www.thelancet.com Vol 385 April 18, 2015 1513

 Articles
antibiotic (–1·21 to 0·27
treatment, days days)
4·0 0·
Intravenous (3·0– 5·0 (3·0– Difference – 01
antibiotic 6·0) 7·0) 0·89 days 1
(–1·57 to –
treatment, days 0·20) days)
58
CAP score* at 59 (41– (40– Difference 1·00 0·
day 5, 78) 74) (–5·23 to 75
7·2
1·00 Prednisone group points 3)
Placebo group 84
CAP score* at 83 (67– (72– Difference –1·00 0·
day 30, 88) 89) (–4·38 56
instability

0·75 HR=1·33 (1·15–1·50), p<0·0001 points to 2·38)

Data are median (IQR) or number (%) unless otherwise stated.

HR=hazard ratio. OR=odds ratio. ICU=intensive care unit. *The CAP
score is a disease-specifi c activity score for community-acquired
pneumonia. It ranges from 0 to 100, 0 marking the worst, 100 the
best score.21
0·50
of

Table 2: Overview of primary and secondary endpoints

The primary analysis followed

the intention-to-treat principle,
which means that patients were
Probability

analysed in the groups to which

they were randomly assigned,
0·25
independent of whether they took
the allocated treatment.14 The
See Online for appendix per-protocol population
focused on patients fully comp-
lying with the trial protocol. For
the primary endpoint, we
calculated an unadjusted hazard
0 ratio (HR) and 95% CI using Cox
proportional hazards regression
0 10 20 30
based on a
Time (days)
Number at risk
Prednisone group 392 37 12 6
Placebo group 393 63 16 5

Figure 2: Kaplan-Meier-curve of time to clinical stability

Prednisone Placebo (n=393) Regression analysis

(n=392)

HR, OR, or difference p value

(95% CI)

Primary endpoint
Intention-to-treat: time 3·0 (2·5–3·4) 4·4 (4·0–5·0) HR 1·33 (1·15 to 1·50) <0·0001
to clinical stability, days
Per-protocol: time to 3·0 (2·5–3·2) 4·4 (4·0–5·0) HR 1·35 (1·16 to 1·56) <0·0001
clinical stability, days
Secondary endpoints
Time to effective
hospital 6·0 (6·0–7·0) 7·0 (7·0–8·0) HR 1·19 (1·04 to 1·38) 0·012
discharge, days
Recurrent pneumonia 23 (6%) 18 (5%) OR 1·30 (0·69 to 2·44) 0·42
Re-admission to hospital 32 (9%) 28 (8%) OR 1·14 (0·67 to 1·93) 0·64
ICU admission 16 (4%) 22 (6%) OR 0·72 (0·37 to 1·39) 0·32
Time to ICU admission, 1 (1–1) 1 (1–1) HR 0·73 (0·38 to 1·38) 0·33
days
Time in ICU, days 3 (2–4) 3 (1–12) Difference –0·2 days 0·96
(–8·7 to 8·2)
Death from any cause 16 (4%) 13 (3%) OR 1·24 (0·59 to 2·62) 0·57
Time to death, days 8·0 (3·0–22·0) 9·0 (2·0–12·0) HR 1·23 (0·59 to 2·55) 0·59
Total duration of 9·0 (7·0–11·0) 9·0 (7·0–12·0) Difference –0·47 days 0·22
 for all analyses. This trial is registered with ClinicalTrials.
gov, number NCT00973154.

Role of the funding source

The funder of the study had no role in study design, data
collection, data analysis, data interpretation, or writing of
the report. The corresponding author had full access to all
binary outcome of achieving or not achieving clinical stability. Patients who died the data in the study and had final responsibility for the
before achieving clinical stability were censored at the day of death; all surviving decision to submit for publication.
patients not achieving clinical stability were censored at day 30. For the primary
endpoint, none of the patients was lost to follow-up. As a sensitivity analysis, the Results
primary analysis was repeated on the per-protocol population. As a further We enrolled 802 eligible patients in the trial and randomly
sensitivity analysis, a multivariable Cox proportional hazards model was fitted with assigned them to receive either prednisone or placebo
treatment group and prespecified potential confounders patient age and PSI score as (figure 1). After blinded post-randomisation exclusion of
independent variables.14 We did prespecified subgroup analyses (patient age, initial 17 patients retrospectively not meeting eligibility criteria,
CRP concentration, history of chronic obstructive pulmonary disease [COPD], PSI 392 patients were allocated to the prednisone group and
class, blood culture positivity) by including appropriate interaction terms in the 393 patients to the placebo group.
multi-variable Cox proportional hazards model.14 Baseline characteristics of the two groups were well
For all secondary endpoints, we calculated unadjusted and adjusted (for patient balanced (table 1). Median age of patients was 74 years,
age and PSI score) estimates of the effect size and corresponding 95% CIs using and 487 (62%) of 785 were men. Patients had a high
linear, logistic, or Cox proportional hazards regression (as appropriate). We burden of comorbidities including diabetes, chronic
analysed community-acquired pneu-monia scores using non-parametric linear obstructive pulmonary disease, chronic heart failure, and
models (quantile regression) due to substantially skewed distributions. 23 For all chronic renal insufficiency. About half the patients were in
time-to-event analyses of secondary endpoints, patients lost to follow-up were high-risk PSI classes IV and V. The appendix shows
censored at the time of lost contact; for all other analyses of secondary outcomes we microbiological aetiology of community-acquired
used complete case analyses. pneumonia, antibiotics given to patients, and supplemental
All reported CIs are two-sided 95% intervals, and tests were done at the two-sided data for clinical insta-bility variables at baseline.
5% significance level. We used STATA 12·1 (Stata Corp, College Station, Texas)
1514 www.thelancet.com Vol 385 April 18, 2015
 Articles

In the intention-to-treat analysis, median time to clinical

Prednisone Placebo Regression analysis
stability was significantly shorter in the prednisone group (n=392) (n=393)
(3·0 days; IQR 2·5–3·4) than in the placebo group (4·4
OR (95% CI) or p value
days; IQR 4·0–5·0) with an HR of 1·33 (95% CI 1·15– difference (95% CI)
1·50, p<0·0001; figure 2 and table 2). Results for the per-
Incidence of pneumonia-associated complications until day 30
protocol population were similar (table 2) and were
Complications due to community- 11(3%) 22 (6%) 0·49 (0·23 to 1·02) 0·056
confirmed by the adjusted analysis (appendix). We noted acquired pneumonia, any
no evidence of effect modification in different prespecified Acute respiratory distress syndrome 0 1 (<1%)
subgroups based on median age, initial median CRP Empyema 1(0·3%) 5 (1%)
concentration, previous history of COPD, severity of Respiratory failure, intubation 1(<1%) 6 (2%)
community-acquired pneumonia defined by the PSI score Persistence of pneumonia 6(2%) 5 (1%)
(I–III vs IV–V), or blood culture positivity (appendix). We Mortality associated with community- 5(1%) 7 (2%)
noted no significant effect modification in post-hoc acquired pneumonia*
analysis of patients with or without sepsis at randomisation Incidence of adverse events compatible with corticosteroid use until day 30
(data not shown). However, we noted a trend towards a Weight change, kg –1·0 –1·0 Difference 0·34 (–0·56 0·46
larger treatment effect in patients with sepsis. (–3·0 to 1·0) (–3·0 to 0·4) to 1·25),
Median time to effective discharge from hospital was Adverse events, any 96 (24%) 61 (16%) 1·77 (1·24 to 2·52) 0·0020
shorter in the prednisone group than in the placebo group In-hospital hyperglycaemia needing 76(19%) 43 (11%) 1·96 (1·31 to 2·93) 0·0010
(table 2). The total duration of antibiotic treatment did not new insulin treatment
differ between groups, but duration of intravenous New insulin dependence at day 30 5(1%) 1 (<1%)
antibiotic treatment was lower in the prednisone group New hypertension at day 30 6(2%) 2 (1%)
than in the placebo group. Rates of recurrent pneumonia, Delirium 5(1%) 2 (1%)
re-admission to hospital, and ICU admittance were similar Gastrointestinal bleeding 3(1%) 4 (1%)
in both treatment groups. All-cause mortality at day 30 did Nosocomial infections 13(3%) 14 (4%)
not differ between groups. The community-acquired Other adverse events until day 30
pneumonia scores at day 5 and at day 30 did not di ffer Any 20(5%) 34 (9%) 0·57 (0·32 to 1·00) 0·052
between groups. Falls with fracture 0 4 (1%)
CRP concentrations were significantly lower in the Cardiac decompensation 5(1%) 10 (3%)
prednisone group than in the placebo group on days 3, 5, Cardiac event 6(2%) 3 (1%)
and 7 (data not shown); PCT levels did not differ between Acute stroke 2(1%) 2 (1%)
treatment groups (data not shown). Thrombembolic event 0 3 (1%)
Overall, complications associated with community- Other 9(2%) 12 (3%)
acquired pneumonia (ie, acute respiratory distress
syndrome, empyema, respiratory failure with intubation, Data are median (IQR) or number (%) unless otherwise stated. OR=odds ratio. *Mortality associated
with community-acquired pneumonia was defi ned as death from community-acquired pneumonia or
persistence of pneumonia, and mortality associated with death from complications due to community-acquired pneumonia.
community-acquired pneumonia) tended to be lower in the
prednisone group than in the placebo group (table 3; Table 3: Complications and adverse events

appendix).
Incidence of any adverse events compatible with to be lower in the prednisone group than in the placebo
corticosteroid use was higher in the prednisone group than group. The prednisone group had a higher rate of
in the placebo group (table 3). This finding was due to a inhospital hyperglycaemia needing insulin treatment than
higher rate of in-hospital hyperglycaemia needing insulin did the placebo group, whereas other adverse events
treatment in the prednisone group than in the placebo compatible with corticosteroid use were rare and similar in
group. The rates of new need for insulin treatment at day both groups.
30 were low in both groups. The incidence of other Pulmonary and circulating inflammatory cytokine
adverse events compatible with corticosteroid use was concentrations are increased in patients with community-
small and similar in both groups. acquired pneumonia, serving as effective mechanism for
the elimination of invading pathogens. Although initially
Discussion beneficial, an unrestrained inflammatory condition might
In this trial, a 7-day treatment with prednisone in patients be detrimental.3 Accordingly, non-survivors of community-
with community-acquired pneumonia led to a reduction in acquired pneumonia show persistently increased
time to clinical stability of 1·4 days, to an overall reduction concentrations of circulating inflammatory cytokines over
of length of hospital stay of 1 day, and to a reduction in time.24 Corticosteroids are the most potent anti-
duration of intravenous antibiotic treatment of 1 day. This inflammatory drugs, with which favourable effects were
effect seemed to be valid across all PSI classes and reported in patients with pneumococcal pneumonia from as
independent of age. Incidence of pneumonia-associated early as 1955.4 Our findings support the hypothesis that
complications until day 30 tended administration of corticosteroids modulates the
www.thelancet.com Vol 385 April 18, 2015 1515
 Articles

hospital stay from 16 to 11 days. The small patient number

Panel: Research in context
might have been the reason for the non-significant results. 26
Systematic review
We systematically searched PubMed using “community- Our results are in contrast to another randomised trial
acquired pneumonia” and “corticosteroids” for randomised, including 213 patients7 showing that prednisolone once
placebo-controlled trials or meta-analyses investigating the daily for a week did not improve clinical cure at day 7 nor
effects of corticosteroids as adjunctive therapy for length of hospital stay compared with placebo. However,
community-acquired pneumonia. Language restrictions when analysing these data without including patients
were not imposed. The last search was done on Dec 2, directly admitted to the ICU, a significant reduction in time
2014. Recent systematic reviews and meta-analyses 9–13 to clinical stability as well as an improvement in length of
suggest a possible benefi t of adjunct corticosteroids for hospital stay of 1 day was noted. 31 This finding is
community-acquired pneumonia in severely ill patients, but unexpected, since it challenges the recent view that
with moderate disease severity, evidence is weak. The only corticosteroids are beneficial for patients with severe
two randomised placebo-controlled trials including 213 and community-acquired pneumonia in the ICU, but not
304 patients with pneumonia of any severity yielded indicated in patients with milder community-acquired
controversial results. One study7 did not fi nd any benefi t of pneumonia outside the ICU.32 In our study, results
adjunct corticosteroids, but an increased recurrence rate, remained similar with or without patients admitted to the
the other8 reported a reduction in length of hospital stay by ICU. It must be taken into account that patients in our
1 day. cohort presented with community-acquired pneumonia of
lower severity than that in previous studies. 7,8 More
Interpretation evidence on patients with severe community-acquired
Our trial with 785 patients shows that a 7-day course pneumonia in the ICU setting is expected by 2016 after
of 50 mg oral prednisone daily shortens time to completion of the ADRENAL trial (ClinicalTrials.gov,
clinical stability by 1·4 days in patients admitted for NCT01448109).
community-acquired pneumonia of any severity. Irrespective of the inclusion or exclusion of patients with
Furthermore, time to hospital discharge and duration direct ICU admission in the analysis, the study by Snijders
of intravenous antibiotic treatment are reduced by 1 and colleagues7 suggested more recurrences in the
day without an increase in complications associated prednisolone group (20 [19%] of 104) than in the placebo
with community-acquired pneumonia. When adding group (10 [9%] of 109), raising concern about the
our study to previous evidence, a meta-analysis now occurrence of a rebound of inflammation after initial
shows a signifi cant reduction of length of hospital suppression by corticosteroids. 7 Our larger study could not
stay (appendix). However, hyperglycaemia has to be confirm this finding, since rates of recurrence of
anticipated, and the usual contraindications for pneumonia and re-admission to hospital were similar in the
corticosteroids must be taken into consideration. prednisone and placebo groups. A clinically relevant
rebound phenomenon with an increased frequency of
immune response and thereby shortens time to clinical recurrent pneumonia seems therefore unlikely. The
stability and length of hospital stay. Importantly, symptom recurrence rate in our study was about 5% in both groups;
resolution, reduction of length of hospital stay, and considering the small absolute number of recurrences in
reduction of intravenous antibiotic therapy are relevant the study by Snijders and colleagues, their finding of
clinical goals in the treatment of patients with community- slightly more recurrences with corticosteroids is probably
acquired pneumonia. The anti-inflammatory e ffect of attributable to chance.
prednisone treatment has a significant beneficial e ffect on Other pneumonia-associated complications in our study
these goals. This is not only relevant from a patient such as incidence of empyema, acute respiratory distress
perspective, but is also an important determinant of syndrome, or respiratory failure were low and tended to be
hospital costs and efficiency. higher in the placebo group than in the prednisone group.
Our results confirm data of various clinical trials
(panel),4–8,25–30 systematic reviews,9,10 and meta-analyses11–13 Hyperglycaemia needing insulin treatment during
showing a beneficial effect of corticosteroids in hospital admission was higher in the prednisone group than
community-acquired pneumonia. They are consistent with in the placebo group, similar to the study by Meijvis and
a randomised trial8 comparing intravenous dexamethasone colleagues.8 This increased hyperglycaemia did not,
with placebo in 302 patients with community-acquired however, affect the clinical outcome and did not prolong
pneumonia outside the ICU, showing a similar reduction in hospital stay. Moreover, the number of patients with new
length of hospital stay of 1 day as observed in our study. A need for insulin treatment at day 30 was low. Other
small randomised trial with open-label design in 31 adverse events compatible with corticosteroid use in our
patients with community-acquired pneumonia of any study were rare, and the incidence did not di ffer between
severity comparing pred-nisolone for 3 days with placebo groups. We assume that the short exposure to
reported a reduction in corticosteroids, when they

1516 www.thelancet.com Vol 385 April 18, 2015

 Articles

are not tapered off during several days but stopped by santésuisse and the Gottfried and Julia Bangerter-Rhyner Foundation.
abruptly after 7 days, has a favourable effect on PS is supported by the Swiss National Science Foundation (SNSF
Professorship, PP00P3_150531 / 1). NR was supported for this study by a
corticosteroid-associated complications. grant from Inselspital Bern, University Hospital and by a grant from the
We did not note any effect modification in di fferent Swiss National Science Foundation (SNSF 320030-138267 and SNSF
prespecified subgroups based on median age, initial 320030-150025). BM, CAB, and PS were supported for this study by the
median CRP, previous history of COPD, severity of research funds from the Department of Endocrinology, Diabetology and
Metabolism, Medical University Clinic of the Kantonsspital Aarau and the
community-acquired pneumonia defined by the PSI score, “Argovia Professorship” of the Medical Faculty of the University of Basel.
or blood culture positivity. In a post-hoc analysis, we noted We gratefully thank the staff of the emergency departments and
a trend towards a larger treatment effect in patients with medical wards of all participating hospitals in supporting this study.
sepsis. Furthermore, a positive treatment response to Furthermore, we thank the many supporters, study and laboratory
personnel at all participating centres who have made this trial
prednisone was not restricted to patients with a specific possible, especially Cemile Bathelt, the study nurse who has
aetiological diagnosis, but was seen irrespective of the coordinated the trial-related tasks during the 5-year trial duration at
identified microorganism. the main study centre. We thank Werner Albrich, Nina Hutter,
Our study has several strengths as compared with Maria Candela, Helga Schneider, Katharina Regez, Ursula Schild,
Eva Grolimund, Alexander Kutz, Martina Bally, Deborah Steiner,
previous studies. First, it is the largest and thus most Anna-Christina Rast, Stefanie Schwarz, Anoush Razzaghi,
conclusive randomised placebo-controlled trial of Ingeborg Schnyder, Judith Siegenthaler, Carla Walti, Cornelia Mueller,
corticosteroids in community-acquired pneumonia Lukas Burget, Katharina Timper, Stefanie Meyer, Sonja Schwenne,
including more than 800 patients; this gave the study Merih Guglielmetti, Kristina Schumacher, Nathalie Christa Schwab,
Danielle Krebs, Cornelia Krismer, Manuel Blum, Christina Wirth,
sufficient power to show a difference in time to clinical Eveline Hofmann, Sebastian Ott, and Khadija M’Rabet Bensalah for
stability and length of hospital stay. Second, patients with local study and patient management, Fausta Chiaverio, Renate Hunziker,
all severity classes of community-acquired pneumonia Ursina Minder, and Christoph Noppen for laboratory support, and Patrick
Simon, Clinical Trial Unit, University Hospital Basel, for database
were included, representing daily routine in patients
support. We also thank the members of the Data Safety and Monitoring
admitted to hospital with community-acquired pneumonia. Board for their valuable time, especially Marc Donath and Benjamin
However, the sickest patients (eg, patients in the ICU and Kasenda. We thank Christian Schindler for statistical advice and Nicole
patients with sepsis) were under-represented. In these Salvisberg for administrative help.
patients, the current guidelines of the Surviving Sepsis
Campaign2 regarding the indication for glucocorticoids References
1 WHO. The top 10 causes of death. Geneva: World Health Organization, 2014.
should be applied. Finally, we used an oral dose of http://www.who.int/mediacentre/factsheets/ fs310/en/index.html (accessed Oct 22, 2014).
prednisone, which is easier to apply to patients with 2 Dellinger RP, Levy MM, Rhodes A, et al. Surviving Sepsis Campaign: international guidelines for
community-acquired pneumonia compared with management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39: 165–228.
intravenous doses of dexamethasone, applied in an earlier 3 Busillo JM, Cidlowski JA. The five Rs of glucocorticoid action during inflammation: ready, reinforce,
repress, resolve, and restore. Trends Endocrinol Metab 2013; 24: 109–19.
trial.8 4 Wagner HN, Bennett IL, Lasagna L, Cluff LE, Rosenthal MB, Mirick GS. The effect of
Our study has also limitations. First, we exclusively hydrocortisone upon the course of pneumococcal pneumonia treated with penicillin.
included patients admitted to hospital, which precludes a Bull Johns Hopkins Hosp 1955; 98: 197–215.
5 Confalonieri M, Urbino R, Potena A, et al. Hydrocortisone infusion for severe community-acquired
generalisation to patients in ambulatory care. Second, our pneumonia: a preliminary randomized study. Am J Respir Crit Care Med 2005; 171: 242–48.
study was not powered for mortality; therefore, mortality data 6 Garcia-Vidal C, Calbo E, Pascual V, Ferrer C, Quintana S, Garau J. Effects of systemic steroids in
have to be acknowledged with care. Third, our primary patients with severe community-acquired pneumonia. Eur Respir J 2007; 30: 951–56.
endpoint, time to clinical stability, has limitations, since it is a 7 Snijders D, Daniels JM, de Graaff CS, van der Werf TS, Boersma WG. Efficacy of corticosteroids in
community-acquired pneumonia: a randomized double-blinded clinical trial. Am J Respir Crit Care Med
combined endpoint including several parameters; however, it 2010; 181: 975–82.
is a well accepted and frequently used endpoint in patients 8 Meijvis SC, Hardeman H, Remmelts HH, et al. Dexamethasone and length of hospital stay in patients
with community-acquired pneumonia.33 Finally, with community-acquired pneumonia: a randomised, double-blind, placebo-controlled trial. Lancet 2011;
377: 2023–30.
corticosteroid-induced hyperglycaemia might have led to
9 Chen Y, Li K, Pu H, Wu T. Corticosteroids for pneumonia. Cochrane Database Syst Rev 2011; 3:
unblinding in some patients. CD007720.
Contributors 10 Siempos, II, Vardakas KZ, Kopterides P, Falagas ME. Adjunctive therapies for community-acquired
CAB, MB, BM, and MC-C designed the study and wrote the protocol. pneumonia: a systematic review. J Antimicrob Chemother 2008; 62: 661–68.
CAB, NN, EU, IS-W, BW, RB, HE, DD, BA, SAU, JR, PT, SW, RT, CB, 11 Lamontagne F, Briel M, Guyatt GH, Cook DJ, Bhatnagar N, Meade M. Corticosteroid therapy for
HD, and NR recruited patients for the study and participated in acute lung injury, acute respiratory distress syndrome, and severe pneumonia: a meta-analysis of
randomized controlled trials. J Crit Care 2010; 25: 420–35.
coordination. CAB, NN, PS, MB, and MC-C had access to all the data,
analysed the data, were responsible for the decision to submit the report, 12 Nie W, Zhang Y, Cheng J, Xiu Q. Corticosteroids in the treatment of community-acquired pneumonia
and drafted it. PS, RB, WZ, PT, DB, NR, BM, and MC-C participated in in adults: a meta-analysis.
PLoS One 2012; 7: e47926.
coordination, gave financial and staff support, and critically revised the
report. All authors read and approved the final report.
Declaration of interests
We declare no competing interests. 1517

Acknowledgments
This study was supported by a grant by the Swiss National Foundation
(PP0P3_123346) to MC-C and the Nora van Meeuwen Häfliger Stiftung and the
Gottfried Julia Bangerter-Rhyner Stiftung. Nasopharyngeal PCR is supported
entirely by Viollier AG, 4002 Basel, Switzerland. MB is supported

www.thelancet.com Vol 385 April 18, 2015

 Articles

13 Shafiq M, Mansoor MS, Khan AA, Sohail MR, Murad MH. Adjuvant 24 Fernandez-Serrano S, Dorca J, Coromines M, Carratala J, Gudiol F,
steroid therapy in community-acquired pneumonia: a systematic review and meta-analysis. J Hosp Med Manresa F. Molecular inflammatory responses measured in blood of
2013; 8: 68–75. patients with severe community-acquired pneumonia.
14 Blum CA, Nigro N, Winzeler B, et al. Corticosteroid treatment for Clin Diagn Lab Immunol 2003; 10: 813–20.
community-acquired pneumonia—the STEP trial: study protocol for a randomized controlled trial. Trials 25 Marik P, Kraus P, Sribante J, Havlik I, Lipman J, Johnson DW.
2014; 15: 257. Hydrocortisone and tumor necrosis factor in severe community-acquired
15 Niederman MS, Mandell LA, Anzueto A, et al. Guidelines for the pneumonia. A randomized controlled study. Chest 1993; 104: 389–92.
management of adults with community-acquired pneumonia. Diagnosis, assessment of severity, 26 Mikami K, Suzuki M, Kitagawa H, et al. Efficacy of corticosteroids in
antimicrobial therapy, and prevention. Am J Respir Crit Care Med 2001; 163: 1730–54. the treatment of community-acquired pneumonia requiring hospitalization.
16 Woodhead M, Blasi F, Ewig S, et al. Guidelines for the management Lung 2007; 185: 249–55.
of adult lower respiratory tract infections—full version. 27 Sabry NA, Omar EE-D. Corticosteroids and ICU course of community
Clin Microbiol Infect 2011; 17 (suppl 6): E1–59. acquired pneumonia in Egyptian settings. Pharmacol Pharm 2011; 2: 73–
17 Laifer G, Flückiger U, Scheidegger C, et al. Management of 81.
community acquired pneumonia (CAP) in adults (ERS/ESCMID guidelines adapted for Switzerland). 28 Fernandez-Serrano S, Dorca J, Garcia-Vidal C, et al. Effect of
http://www.sginfch/ssi-home/ guidelines/ssihtml (accessed Dec 15, 2014). corticosteroids on the clinical course of community-acquired pneumonia: a
18 Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases randomized controlled trial. Crit Care 2011; 15: R96.
Society of America/American Thoracic Society consensus guidelines on the management of community- 29 McHardy VU, Schonell ME. Ampicillin dosage and use of
acquired pneumonia in adults. Clin Infect Dis 2007; 44 (suppl 2): S27–72. prednisolone in treatment of pneumonia: co-operative controlled trial. Br
19 Schuetz P, Christ-Crain M, Thomann R, et al. Effect of procalcitonin- Med J 1972; 4: 569–73.
based guidelines vs standard guidelines on antibiotic use in lower respiratory tract infections: the ProHOSP 30 Nafae RM, Ragab MI, Amany FM, Rashed SB. Adjuvant role of
randomized controlled trial. JAMA 2009; 302: 1059–66. corticosteroids in the treatment of community-acquired pneumonia. Egypt J
20 Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low- Chest Dis Tuberc 2013; 62: 439–45.
risk patients with community-acquired pneumonia. 31 Snijders D, de Graaff CS, van der Werf T, Boersma WG. Effect of
N Engl J Med 1997; 336: 243–50. excluding ICU-admission on clinical outcomes in a randomized control trial
21 El Moussaoui R, Opmeer BC, Bossuyt PM, Speelman P, studying the effect of corticosteroids in patients hospitalized with CAP. Eur
de Borgie CA, Prins JM. Development and validation of a short Respir J 2012; 40 (suppl 56): P1734.
questionnaire in community acquired pneumonia. Thorax 2004; 59: 32 Meduri GU, Bell WA, Confalonieri M. Glucocorticoid treatment in
591–95. community-acquired pneumonia without severe sepsis: no evidence of
22 Christ-Crain M, Jaccard-Stolz D, Bingisser R, et al. Effect of efficacy. Am J Respir Crit Care Med 2010; 181: 880–82.
procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: cluster- 33 Garin N, Genne D, Carballo S, et al. Beta-lactam monotherapy vs
randomised, single-blinded intervention trial. Lancet 2004; 363: 600–07. beta-lactam-macrolide combination treatment in moderately severe
23 Chamberlain G. Quantile regression, censoring, and the structure of community-acquired pneumonia: a randomized noninferiority trial. JAMA
wages. In: Sims CA, ed. Advances in Economics Sixth World Congress. Cambridge: Cambridge University Intern Med 2014; 174: 1894–901.
Press, 1994: 171–209.

1518

www.thelancet.com Vol 385 April 18, 2015