Stern Et Al-2017-Cochrane Database of Systematic Reviews
Stern Et Al-2017-Cochrane Database of Systematic Reviews
Stern Et Al-2017-Cochrane Database of Systematic Reviews
www.cochranelibrary.com
Anat Stern1 , Keren Skalsky2 , Tomer Avni2 , Elena Carrara3 , Leonard Leibovici2 , Mical Paul1
1 Division
of Infectious Diseases, Rambam Health Care Campus, Haifa, Israel. 2 Department of Medicine E, Beilinson Hospital, Rabin
Medical Center, Petah Tikva, Israel. 3 Infectious Diseases, Policlinico San Matteo Hospital, Pavia, Italy
Contact address: Anat Stern, Division of Infectious Diseases, Rambam Health Care Campus, Ha-aliya 8 St, Haifa, 33705, Israel.
[email protected].
Citation: Stern A, Skalsky K, Avni T, Carrara E, Leibovici L, Paul M. Corticosteroids for pneumonia. Cochrane Database of Systematic
Reviews 2017, Issue 12. Art. No.: CD007720. DOI: 10.1002/14651858.CD007720.pub3.
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
ABSTRACT
Background
Pneumonia is a common and potentially serious illness. Corticosteroids have been suggested for the treatment of different types of
infection, however their role in the treatment of pneumonia remains unclear. This is an update of a review published in 2011.
Objectives
To assess the efficacy and safety of corticosteroids in the treatment of pneumonia.
Search methods
We searched the Cochrane Acute Respiratory Infections Group’s Specialised Register, CENTRAL, MEDLINE, Embase, and LILACS
on 3 March 2017, together with relevant conference proceedings and references of identified trials. We also searched three trials registers
for ongoing and unpublished trials.
Selection criteria
We included randomised controlled trials (RCTs) that assessed systemic corticosteroid therapy, given as adjunct to antibiotic treatment,
versus placebo or no corticosteroids for adults and children with pneumonia.
Data collection and analysis
We used standard methodological procedures expected by Cochrane. Two review authors independently assessed risk of bias and
extracted data. We contacted study authors for additional information. We estimated risk ratios (RR) with 95% confidence intervals
(CI) and pooled data using the Mantel-Haenszel fixed-effect model when possible.
Main results
We included 17 RCTs comprising a total of 2264 participants; 13 RCTs included 1954 adult participants, and four RCTs included
310 children. This update included 12 new studies, excluded one previously included study, and excluded five new trials. One trial
awaits classification.
All trials limited inclusion to inpatients with community-acquired pneumonia (CAP), with or without healthcare-associated pneumonia
(HCAP). We assessed the risk of selection bias and attrition bias as low or unclear overall. We assessed performance bias risk as low
for nine trials, unclear for one trial, and high for seven trials. We assessed reporting bias risk as low for three trials and high for the
remaining 14 trials.
Corticosteroids for pneumonia (Review) 1
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Corticosteroids significantly reduced mortality in adults with severe pneumonia (RR 0.58, 95% CI 0.40 to 0.84; moderate-quality
evidence), but not in adults with non-severe pneumonia (RR 0.95, 95% CI 0.45 to 2.00). Early clinical failure rates (defined as death
from any cause, radiographic progression, or clinical instability at day 5 to 8) were significantly reduced with corticosteroids in people
with severe and non-severe pneumonia (RR 0.32, 95% CI 0.15 to 0.7; and RR 0.68, 95% CI 0.56 to 0.83, respectively; high-quality
evidence). Corstocosteroids reduced time to clinical cure, length of hospital and intensive care unit stays, development of respiratory
failure or shock not present at pneumonia onset, and rates of pneumonia complications.
Among children with bacterial pneumonia, corticosteroids reduced early clinical failure rates (defined as for adults, RR 0.41, 95% CI
0.24 to 0.70; high-quality evidence) based on two small, clinically heterogeneous trials, and reduced time to clinical cure.
Hyperglycaemia was significantly more common in adults treated with corticosteroids (RR 1.72, 95% CI 1.38 to 2.14). There were
no significant differences between corticosteroid-treated people and controls for other adverse events or secondary infections (RR 1.19,
95% CI 0.73 to 1.93).
Authors’ conclusions
Corticosteroid therapy reduced mortality and morbidity in adults with severe CAP; the number needed to treat for an additional
beneficial outcome was 18 patients (95% CI 12 to 49) to prevent one death. Corticosteroid therapy reduced morbidity, but not
mortality, for adults and children with non-severe CAP. Corticosteroid therapy was associated with more adverse events, especially
hyperglycaemia, but the harms did not seem to outweigh the benefits.
Outcomes Anticipated absolute effects* (95% CI) Relative effect of participants Quality of the evidence Comments
(95% CI) (studies) (GRADE)
* The risk in the intervention group (and its 95% conf idence interval) is based on the assum ed risk in the com parison group and the relative effect of the intervention (and its
95% CI).
CAP: com m unity-acquired pneum onia; CI: conf idence interval; RCT: random ised controlled trial; RR: risk ratio
rem ained signif icant in trials with low and unclear allocation concealm ent.
5 We downgraded quality f or inconsistency (I² = 74% f or this outcom e) and upgraded one level f or large ef f ect (RR = 0.32).
6 We downgraded quality f or risk of bias (analysis includes two low-quality trials).
7
We upgraded quality one level f or large ef f ect (RR = 0.41).
5
BACKGROUND fections, with estimates of HAP-associated mortality ranging from
20% to 50% (Chawla 2008; Rosenthal 2012). The rate of VAP
development is between 10% and 30% of patients receiving more
Description of the condition than 48 hours of mechanical ventilation (Rosenthal 2012; Torres
2010). The occurrence of VAP has a significant impact on patient
Acute pneumonia is a common and potentially serious illness. De-
outcomes; it is associated with substantial morbidity, significantly
spite significant advances in its aetiological investigation, antimi-
longer ICU stays, and a two-fold mortality rate compared with
crobial therapy, and improvements in supportive care, acute pneu-
similar patients who do not have VAP (Safdar 2005).
monia is still among the top 10 most common causes of death
among all age groups. In the USA in 2013, pneumonia was the
sixth-leading cause of death in people aged 65 years and over, and
the single most common cause of infection-related mortality (Xu Description of the intervention
2016). Corticosteroids (glucocorticoids, steroids) include steroid hor-
In the past, acute pneumonia was classified as community- mones that are naturally produced in the adrenal cortex of ver-
acquired pneumonia (CAP) and hospital-acquired pneumonia tebrates and their synthetic analogues. Corticosteroids are in-
(HAP) including ventilator-associated pneumonia (VAP), based volved in many physiological processes, including stress response,
on differences in aetiologic agents and antibiotic susceptibilities immune response and regulation of inflammation, carbohydrate
among these entities. Following increases in healthcare delivery metabolism, protein catabolism, blood electrolyte levels and be-
shifting to outpatient settings, an additional pneumonia cate- haviour. Synthetic derivatives of natural steroids include pred-
gory, healthcare-associated pneumonia (HCAP), has been defined. nisone, prednisolone, methylprednisolone, betamethasone, dex-
Healthcare-associated pneumonia occurs in non-hospitalised peo- amethasone, triamcinolone, and hydrocortisone. Corticosteroids
ple who have had extensive healthcare contact, defined as including are important components in the treatment of many inflamma-
one or more of: intravenous therapy, wound care, or intravenous tory, allergic, immunologic, and malignant disorders. Corticos-
chemotherapy over the previous 30 days; residing in a nursing teroids are administered by inhalation, orally, or intravenously
home or other long-term care facility; hospitalisation in an acute (Rhen 2005).
care hospital for two or more days over the previous 90 days; or Unfortunately, the therapeutic effects of corticosteroids are often
attending a hospital or haemodialysis clinic over the previous 30 accompanied by clinically significant side effects, most of which
days (Friedman 2002). are related to the dose and duration of therapy (Rhen 2005). Many
The aetiology of CAP varies by geographic region. However, Strep- adverse effects occur only with prolonged administration; most
tococcus pneumoniae is the most common cause worldwide (File short-term adverse events are reversible when the drug is discon-
2010). The overall incidence of CAP in adults is approximately tinued. Side effects observed with long-term, high-dose treatment
5.16 to 6.11 cases per 1000 persons per year and increases with include obesity with a special fat distribution pattern (e.g. face
age. There is seasonal variation, with more cases of pneumonia oc- swelling), immune depression, delayed wound healing, growth
curring during the winter months. About 20% of all people with retardation in children, hirsutism, diabetes, depressive disorders,
CAP are admitted to hospital (Niederman 2007), and approxi- Cushing’s syndrome, and osteoporosis (Oray 2016). Short-term
mately 10% to 20% of patients require admission to intensive care use of corticosteroids may be associated with hyperglycaemia, fluid
units (ICU) (Marrie 2007; Restrepo 2008). Thirty-day mortality retention, and hypertension, which are typically transient (Schäcke
in people who are hospitalised for CAP is approximately 10% to 2002). Neuropsychiatric side effects, ranging from insomnia and
12% and is higher in people admitted to the ICU (Musher 2014). irritability to mania, psychosis, and delirium, are commonly seen
HAP, VAP, and HCAP may be caused by a broad variety of with short-term corticosteroid use (Warrington 2006).
pathogens and can be polymicrobial. Common pathogens include
aerobic gram-negative bacilli and gram-positive cocci (e.g. Staphy-
lococcus aureus, including methicillin-resistant S aureus (MRSA)
and Streptococcus spp). Hospital-acquired pneumonia due to
How the intervention might work
viruses or fungi is significantly less common, except in people who Corticosteroids have been suggested for the treatment of different
are immunocompromised (Jones 2010). types of infections, including meningitis, tuberculosis, pneumo-
Precise incidence rates of HAP and HCAP are difficult to deter- cystis pneumonia, other bacterial pneumonia, and septic shock.
mine because of differences in local epidemiology and infection The theoretical advantages of corticosteroids differ for each infec-
control measures. Estimates of HAP incidence range from 5 to tion type (Chalmers 2010; Kalil 2016; Salluh 2008).
more than 20 episodes per 1000 hospitalisations (Chawla 2008), In septic shock, a condition of extreme physiological stress, sub-
accounting for up to 25% of all nosocomial (hospital-acquired) optimal cortisol production has been termed critical illness-re-
infections (Magill 2014; Torres 2010). Hospital-acquired pneu- lated corticosteroid insufficiency (CIRCI) (Marik 2008). The ma-
monia is the leading cause of death among hospital-acquired in- jor theoretical purpose of corticosteroids in sepsis is to restore bal-
trials included people with CAP who were treated in the commu-
Included studies
nity, HAP, or VAP. Inclusion was not limited by CAP pathogens
in all adult trials. Van Woensel 2003 included only young chil-
Details of the included studies are summarised in the dren who were mechanically ventilated due to respiratory syncy-
Characteristics of included studies table. tial virus lower respiratory tract infections (82 infants, of whom
We included 17 randomised controlled trials in this update, five of 41 had pneumonia). Two trials limited inclusion to children with
which were included in the 2011 review (Chen 2011). The studies M pneumoniae (Luo 2014; Wu 2014).
were performed between 1966 and 2014, but Marik 1993 did not A total of 2264 participants were randomised (1122 to the inter-
specify recruitment dates. Six of the studies were multicentre trials vention arm), of whom 1954 (86%) were adults (n = 13 trials,
including from two to seven centres, all in a single country (Blum mean age 69.8 years) and 307 (14%) were children (n = 4 trials,
2015; Confalonieri 2005; Meijvis 2011; Sabry 2011; Torres 2015; mean age 5.6 years, Luo 2014; Nagy 2013; Van Woensel 2003;
Van Woensel 2003), and 11 studies were single-centre trials. The Wu 2014). Of the adult trials, three reported that no people with
trials were conducted worldwide: eight in Europe, four in China chronic lung disease were included (Mikami 2007; Nafae 2013;
or Japan, three in the Middle East, and one each in South Africa Sabry 2011); in six trials reporting data between 11.2% and 35.9%
and Australia. of participants had chronic lung disease. A single trial reported
All included trials limited inclusion to participants with CAP, with that people with diabetes mellitus were not included (McHardy
or without HCAP, who were treated as inpatients. None of the
Corticosteroids for pneumonia (Review) 11
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
1972), and in five trials between 10.3% and 19.7% of participants Seven trials were funded by academic sponsors (Blum 2015;
had diabetes. Confalonieri 2005; Fernández-Serrano 2011; Meijvis 2011; Nagy
Six trials included only adults with severe pneumonia. Four trials 2013; Torres 2015; Van Woensel 2003), one trial was funded by a
defined severe pneumonia according to the Infectious Diseases So- pharmaceutical company (Snijders 2010), and one trial reported
ciety of America and American Thoracic Society guidelines (ATS/ receiving no funding (Sabry 2011). The remaining eight trials did
IDSA guidelines) for severe pneumonia, Mandell 2007, or ear- not report their funding source. We sought additional data for all
lier versions of these guidelines (Confalonieri 2005; El-Ghamrawy trials, which two authors supplied (Blum 2015; Torres 2015).
2006; Sabry 2011; Torres 2015). Marik 1993 defined severe pneu-
monia as presentation of three or more British Thoracic Soci-
ety (BTS) criteria for severe pneumonia (BTS guidelines 1987). Excluded studies
Fernández-Serrano 2011 defined pneumonia as the presence of We excluded six studies (see Characteristics of excluded studies
respiratory failure and extensive radiologically confirmed consoli- table). One trial that was included in the 2011 review was ex-
dations. cluded because it examined the use of inhaled corticosteroids
The intervention included oral prednisone in three trials (Blum (Cao 2007). Other reasons for exclusion were: not limited to peo-
2015; Luo 2014; McHardy 1972), and intravenous dexametha- ple with pneumonia (Van Woensel 2011), corticosteroids given
sone, hydrocortisone, or methylprednisolone in 13 trials. One to both treatment arms (Huang 2014), participants in the con-
trial used prednisone without limiting the administration route trol group were subsequently given either corticosteroids or in-
(Snijders 2010). The duration of corticosteroid treatment was travenous immunoglobulin therapy (Shan 2017), quasi-randomi-
10 days in one trial (Fernández-Serrano 2011), seven to 10 days sation (Wagner 1956), and one non-randomised trial (Montón
in one trial (Wu 2014), seven days in seven trials (Blum 2015; 1999).
Confalonieri 2005; El-Ghamrawy 2006; McHardy 1972; Nafae
2013; Sabry 2011; Snijders 2010), five days in three trials (Luo
2014; Nagy 2013; Torres 2015), and two to four days in four trials Studies awaiting classification
(Hatakeyama 1995; Meijvis 2011; Mikami 2007; Van Woensel One trial that assessed children with refractory M pneumoniae
2003). In one trial only one dose of corticosteroids was given pneumonia is awaiting classification (Lan 2015). We contacted the
(Marik 1993). Most adult trials used steroid doses equivalent to authors for clarification about study design and inconsistencies in
40 mg to 50 mg of prednisone per day. The comparator for cor- reported results.
ticosteroids was placebo in 11 trials and no treatment in six trials
(Confalonieri 2005; Luo 2014; McHardy 1972; Mikami 2007; Risk of bias in included studies
Sabry 2011; Wu 2014).
Risk of bias for all trials is summarised in Figure 2 and Figure 3.
Figure 2. Risk of bias graph: review authors’ judgements about each risk of bias item presented as
percentages across all included studies.
Children
early clinical failure was significantly lower in the corticosteroid
There were no reported deaths in the four trials that included arm compared to control (RR 0.32, 95% CI 0.15 to 0.7; I² =
children (risk difference (RD) 0.00, 95% CI -0.03 to 0.03; 4 trials, 74% for severe pneumonia, and RR 0.68, 95% CI 0.56 to 0.83; I²
266 children, Analysis 1.4). = 0% for non-severe pneumonia, random-effects model; Analysis
1.5). Heterogeneity in the subgroup of severe pneumonia resulted
from different magnitudes of benefit rather than from opposing
Secondary outcomes direct of effects. We graded the evidence for both analyses as of
high quality.
The dose of corticosteroid used in most adult trials reporting on
1. Early clinical failure early clinical failure was equivalent to 40 mg to 50 mg of pred-
nisone per day, for 5 to 10 days. Only Confalonieri 2005 reported
clinical failure in the subgroup of participants with septic shock.
Adults No trials reported on this outcome by pathogen or among partic-
Six trials reported the outcome of early clinical failure (1324 par- ipants with and without COPD. Meta-regression analyses could
ticipants). The definitions used in the trials varied and are reported not be fitted due to the paucity of studies reporting on the out-
in Table 1. There was a significantly lower rate of early clinical fail- come and participant characteristics.
ure in participants treated with corticosteroids compared to con-
trols (RR 0.40, 95% CI 0.23 to 0.7; random-effects model). We
found substantial heterogeneity in this analysis (Analysis 1.5, I² =
Children
72%). We downgraded the quality of the data for this outcome to
moderate due to inconsistency, despite the large effect. The outcome of early clinical failure was reported in two studies
(88 children) with definitions provided in Table 1, and was sig-
nificantly lower in the corticosteroid group compared to the con-
trol group (RR 0.41, 95% CI 0.24 to 0.70; I² = 25%, fixed-effect
Subgroup analyses model; Analysis 1.6). We graded this analysis as high quality.
Five trials including 419 participants provided data regarding early In an analysis by allocation concealment among adults and chil-
clinical failure in participants with severe pneumonia, and two dren combined, early clinical failure was significantly lower in the
trials including 905 participants reported this outcome in partic- corticosteroid arm both for trials with low risk of bias (RR 0.59,
ipants with non-severe pneumonia. In both subgroups the rate of 95% CI 0.37 to 0.94) and trials with unclear risk RR 0.27, 95%
Children Children
Four trials reported time to clinical cure, of which three studies One trial including children with bacterial pneumonia, Nagy
(225 children) evaluated bacterial pneumonia and one included 2013, and one trial including children with viral pneumonia, Van
41 children with respiratory syncytial virus pneumonia. In one Woensel 2003, reported lengths of hospital stay. We did not pool
trial (Nagy 2013), mean and SD were estimated from the median results due to significant heterogeneity, with a significant benefit
and IQR. In the three trials evaluating bacterial pneumonia, time to steroids in the trial on bacterial pneumonia (MD -4.70 days
to clinical cure was significantly shorter in the corticosteroid arm with steroids, 95% CI -7.50 to -1.90) and no difference in the
compared to the control arm (MD -1.57 days, 95% CI -2.55 to trial on viral pneumonia (Analysis 1.15).
-0.60 days; I² = 80%, random-effects model; Analysis 1.9). Van
Woensel 2003, which evaluated viral pneumonia, showed no effect
7. Duration of ICU stay for participants admitted to the ICU
of corticosteroids on time to clinical cure. Here we also accepted
the study definitions for time to clinical cure, which were not For adults who were admitted to the ICU (342 participants from
compatible with our protocol definitions (Table 1). 8 studies), ICU stay was significantly shorter in the corticosteroid
group compared to control (MD -1.88 days, 95% CI -2.96 to
-0.81; I² = 46%, fixed-effect model; Analysis 1.14). Five trials
3. Development of respiratory failure not present initially provided median values (range/IQR) for this outcome, and means
The need for new non-invasive or invasive mechanical ventilation, (SD) were calculated from these values.
not present at onset of pneumonia, was reported in four adult
trials (1030 participants) and was significantly lower in the corti-
costeroid arm compared to the control arm without heterogene- 8. Pneumonia complications not present initially
ity (RR 0.40, 95% CI 0.20 to 0.77; fixed-effect model; Analysis Nine trials reported pneumonia complications, including eight
1.10). adult trials (1573 participants) and one trial in children (59 chil-
dren). Most trials defined this outcome as a combination of lung
abscess, empyema or pleural effusion. One trial defined complica-
4. Development of shock not present initially tions as septic shock or acute respiratory distress syndrome (ARDS)
The development of shock not present initially was reported in (Nafae 2013), and two trials evaluated this outcome without pro-
six adult trials (five including participants with severe pneumo- viding a definition. The rate of pneumonia complications was sig-
nia) and was significantly lower in the corticosteroid arm without nificantly lower for the corticosteroid arm when compared to con-
heterogeneity (RR 0.18, 95% CI 0.09 to 0.34; fixed-effect model; trols (RR 0.58, 95% CI 0.40 to 0.84; I² = 43%, fixed-effect model;
Analysis 1.11). Analysis 1.16).
REFERENCES
References to studies included in this review Luo 2014 {published data only}
Luo Z, Luo J, Liu E, Xu X, Liu Y, Zeng F, et al. Effects
Blum 2015 {published and unpublished data} of prednisolone on refractory Mycoplasma pneumoniae
Blum CA, Nigro N, Briel M, Schuetz P, Ullmer E, Suter- pneumonia in children. Pediatric Pulmonology 2014;49(4):
Widmer I, et al. Adjunct prednisone therapy for patients 377–80.
with community-acquired pneumonia: a multicentre,
Marik 1993 {published data only}
double-blind, randomised, placebo-controlled trial. Lancet
Marik P, Kraus P, Sribante J, Havlik I, Lipman J, Johnson
2015;385(9977):1511–8.
DW. Hydrocortisone and tumour necrosis factor in severe
Confalonieri 2005 {published data only} community-acquired pneumonia. A randomised controlled
Confalonieri M, Urbino R, Potena A, Piattella M, Parigi P, study. Chest 1993;104(2):389–92.
Puccio G. Hydrocortisone infusion for severe community-
acquired pneumonia: a preliminary randomised study. McHardy 1972 {published data only}
American Journal of Respiratory and Critical Care Medicine McHardy VU, Schonell ME. Ampicillin dosage and use
2005;171(3):242–8. of prednisolone in treatment of pneumonia: co-operative
controlled trial. British Medical Journal 1972;4(5840):
El-Ghamrawy 2006 {published data only} 569–73.
El-Ghamrawy AH, Shokier MH, Esmat AA. Effects of
low-dose hydrocortisone in ICU patients with severe Meijvis 2011 {published data only}
community-acquired pneumonia. Egyptian Journal of Chest Meijvis SC, Hardeman H, Remmelts HH, Heijligenberg
Diseases and Tuberculosis 2006;55:91–9. R, Rijkers GT, van Velzen-Blad H, et al. Dexamethasone
and length of hospital stay in patients with community-
Fernández-Serrano 2011 {published data only}
acquired pneumonia: a randomised, double-blind, placebo-
Fernández-Serrano S, Dorca J, Garcia-Vidal C, Fernández-
controlled trial. Lancet 2011;377(9782):2023–30.
Sabé N, Carratalà J, Fernández-Agüera A, et al. Effect
of corticosteroids on the clinical course of community- Mikami 2007 {published data only}
acquired pneumonia: a randomized controlled trial. Critical Mikami K, Suzuki M, Kitagawa H, Kawakami M, Hirota N,
Care 2011;15(2):R96. Yamaguchi H. Efficacy of corticosteroids in the treatment of
community-acquired pneumonia requiring hospitalisation.
Hatakeyama 1995 {published data only}
Lung 2007;185(5):249–55.
Hatakeyama S, Tachibana A, Suzuki K, Okano H.
Treatment of aspiration pneumonia with low-dose Nafae 2013 {published data only}
methylprednisolone and antibiotics. Japanese Journal of Nafae RM, Ragab MI, Amany FM, Rashed SB. Adjuvant
Thoracic Diseases 1998;33(1):51–6. role of corticosteroids in the treatment of community-
Corticosteroids for pneumonia (Review) 22
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
acquired pneumonia. Egyptian Journal of Chest Diseases and added to standard treatment with intravenous azithromycin
Tuberculosis 2013;62(3):439-45. for refractory Mycoplasma pneumoniae pneumonia in
Nagy 2013 {published data only} children. World Journal of Pediatrics 2017 Jan 27 Epub
Nagy B, Gaspar I, Papp A, Bene Z, Nagy B Jr, Voko Z, et ahead of print]. DOI: 10.1007/s12519-017-0014-9
al. Efficacy of methylprednisolone in children with severe Van Woensel 2011 {published data only}
community acquired pneumonia. Pediatric Pulmonology Van Woensel JB, Vyas H, STAR Trial Group.
2013;48(2):168–75. Dexamethasone in children mechanically ventilated for
Sabry 2011 {published data only} lower respiratory tract infection caused by respiratory
Sabry NA, Omar EE. Corticosteroids and ICU course of syncytial virus: a randomized controlled trial. Critical Care
community acquired pneumonia in Egyptian settings. Medicine 2011;39(7):1779–83.
Pharmacology & Pharmacy 2011;2:73–81. DOI: 10.4236/ Wagner 1956 {published data only}
pp.2011.22009 Wagner HN Jr, Bennett IL Jr, Lasagna L, Cluff LE,
Snijders 2010 {published data only} Rosenthal MB, Mirick GS. The effect of hydrocortisone
Snijders D, Daniels JM, de Graaff CS, van der Werf TS, upon the course of pneumococcal pneumonia treated with
Boersma WG. Efficacy of corticosteroids in community- penicillin. Bulletin of the Johns Hopkins Hospital 1956;98
acquired pneumonia: a randomized double-blinded clinical (3):197–215.
trial. American Journal of Respiratory and Critical Care
References to studies awaiting assessment
Medicine 2010;181(9):975–82.
Torres 2015 {published and unpublished data} Lan 2015 {published data only}
Torres A, Sibila O, Ferrer M, Polverino E, Menendez R, Lan Y, Yang D, Chen Z, Tang L, Xu Y, Cheng Y.
Mensa J, et al. Effect of corticosteroids on treatment Effectiveness of methylprednisolone in treatment of children
failure among hospitalized patients with severe community- with refractory Mycoplasma pneumoniae pneumonia and its
acquired pneumonia and high inflammatory response: a relationship with bronchoalveolar lavage cytokine levels.
randomized clinical trial. JAMA 2015;313(7):677–86. Zhonghua Er Ke Za Zhi 2015;53(10):779–83.
Van Woensel 2003 {published data only}
Van Woensel JB, van Aalderen WM, de Weerd W, Jansen References to ongoing studies
NJ. Dexamethasone for treatment of patients mechanically
ventilated for lower respiratory tract infection caused by NCT01283009 {published and unpublished data}
respiratory syncytial virus. Thorax 2003;58(5):383–7. NCT01283009. Extended Steroid in CAP(e) (ESCAPe).
clinicaltrials.gov/ct2/show/NCT01283009 (first received
Wu 2014 {published data only}
21 January 2011).
Wu YJ, Sun J, Zhang JH, Feng LL. Clinical efficacy of
adjuvant therapy with glucocorticoids in children with lobar NCT02618057 {published and unpublished data}
pneumonia caused by Mycoplasma pneumoniae. Zhongguo NCT02618057. Effects of oral steroid in Mycoplasma
Dang Dai Er Ke Za Zhi [Chinese Journal of Contemporary pneumoniae pneumonia. clinicaltrials.gov/ct2/show/
Pediatrics] 2014;16(4):401–5. NCT02618057 (first received 8 November 2015).
NCT02735707 {published and unpublished data}
References to studies excluded from this review NCT02735707. Randomized, embedded, multifactorial
Cao 2007 {published data only} adaptive platform trial for community-acquired
Cao LF, Lu YM, Ma HG, Ma M. Budesonide inhaling pneumonia (REMAP-CAP). clinicaltrials.gov/ct2/show/
auxiliary therapy after Mycoplasma pneumoniae infection of NCT02735707 (first received 11 December 2015).
children. International Journal of Respiration 2007;27(8): NCT03121690 {published and unpublished data}
567–9. NCT03121690. The applicability of different scoring
Huang 2014 {published data only} systems and use of steroids in the treatment of hospital
Huang L, Gao X, Chen M. Early treatment with acquired pneumonia. clinicaltrials.gov/ct2/show/
corticosteroids in patients with Mycoplasma pneumoniae NCT03121690 (first received 10 April 2017).
pneumonia: a randomized clinical trial. Journal of Tropical
Pediatrics 2014;60(5):338–42.
Additional references
Montón 1999 {published data only} Annane 2004
Montón C, Ewig S, Torres A, El-Ebiary M, Filella X, Rañó Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D,
A, et al. Role of glucocorticoids on inflammatory response Kupfer Y. Corticosteroids for severe sepsis and septic shock:
in nonimmunosuppressed patients with pneumonia: a pilot a systematic review and meta-analysis. BMJ 2004;329
study. European Respiratory Journal 1999;14(1):218–20. (7464):480.
Shan 2017 {published data only} Annane 2009
Shan LS, Liu X, Kang XY, Wang F, Han XH, Shang YX. Annane D, Bellissant E, Bollaert PE, Briegel J, Confalonieri
Effects of methylprednisolone or immunoglobulin when M, De Gaudio R, et al. Corticosteroids in the treatment of
Corticosteroids for pneumonia (Review) 23
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
severe sepsis and septic shock in adults: a systematic review. File 2010
JAMA 2009;301(22):2362. File TM Jr, Marrie TJ. Burden of community-acquired
pneumonia in North American adults. Postgraduate
Atkins 2004
Medicine 2010;122(2):130–41.
Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y,
Flottorp S, et al. GRADE Working Group. Grading quality Fine 1997
of evidence and strength of recommendations. BMJ 2004; Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld
328(7454):1490. LA, Singer DE, et al. A prediction rule to identify low-
risk patients with community-acquired pneumonia. New
Avni 2015 England Journal of Medicine 1997;336(4):243–50.
Avni T, Shiver-Ofer S, Leibovici L, Tacconelli E, DeAngelis
G, Cookson B, et al. Participation of elderly adults in Friedman 2002
randomized controlled trials addressing antibiotic treatment Friedman ND, Kaye KS, Stout JE, McGarry SA, Trivette
of pneumonia. Journal of the American Geriatrics Society SL, Briggs JP, et al. Health care-associated bloodstream
2015;63(2):233–43. infections in adults: a reason to change the accepted
definition of community-acquired infections. Annals of
Brouwer 2015 Internal Medicine 2002;137(10):791–7.
Brouwer MC, McIntyre P, Prasad K, van de Beek D.
Corticosteroids for acute bacterial meningitis. Cochrane GRADEpro GDT 2015 [Computer program]
Database of Systematic Reviews 2015, Issue 9. DOI: GRADE Working Group, McMaster University.
10.1002/14651858.CD004405.pub5 GRADEpro GDT. Version (accessed prior to 30 August
2017). Hamilton (ON): GRADE Working Group,
BTS guidelines 1987 McMaster University, 2015.
British Thoracic Society. Community-acquired pneumonia
Higgins 2011
in adults in British hospitals in 1982-1983: a survey of
Higgins JP, Green S, editor(s). Cochrane Handbook
aetiology, mortality, prognostic factors and outcome. The
for Systematic Reviews of Interventions Version 5.1.0
British Thoracic Society and the Public Health Laboratory
(updated March 2011). The Cochrane Collaboration,
Service. Quarterly Journal of Medicine 1987;62(239):
2011. Available from handbook.cochrane.org.
195–220.
Horita 2015
Chalmers 2010
Horita N, Otsuka T, Haranaga S, Namkoong H, Miki M,
Chalmers JD, Singanayagam A, Akram AR, Mandal P,
Miyashita N, et al. Adjunctive systemic corticosteroids for
Short PM, Choudhury G, et al. Severity assessment tools
hospitalized community-acquired pneumonia: systematic
for predicting mortality in hospitalised patients with
review and meta-analysis 2015 update. Scientific Reports
community-acquired pneumonia. Systematic review and
2015;5:14061.
meta-analysis. Thorax 2010;65(10):878–83.
Jones 2010
Chawla 2008 Jones RN. Microbial etiologies of hospital-acquired bacterial
Chawla R. Epidemiology, etiology, and diagnosis of pneumonia and ventilator-associated bacterial pneumonia.
hospital-acquired pneumonia and ventilator-associated Clinical Infectious Diseases 2010;51(Suppl 1):S81–7.
pneumonia in Asian countries. American Journal of Infection
Control 2008;36(4 Suppl):S93–100. Kalil 2016
Kalil AC, Metersky ML, Klompas M, Muscedere J, Sweeney
Comprehensive Meta-Analysis V3 [Computer program] DA, Palmer LB, et al. Management of adults with hospital-
Biostat. Comprehensive Meta-Analysis V3.3.07. Version acquired and ventilator-associated pneumonia: 2016
(accessed prior to 30 August 2017). Englewood (NJ): clinical practice guidelines by the Infectious Diseases Society
Biostat, 2014. of America and the American Thoracic Society. Clinical
FDA CAP industry guidance Infectious Diseases 2016;63(5):e61–111.
FDA. Guidance for industry. Community-acquired Kaplan 2002
pneumonia - developing antimicrobial drugs Kaplan V, Angus DC, Griffin MF, Clermont G, Scott
for treatment. www.fda.gov/downloads/Drugs/ Watson R, Linde-Zwirble WT. Hospitalized community-
GuidanceComplianceRegulatoryInformation/Guidances/ acquired pneumonia in the elderly: age- and sex-related
UCM123686.pdf (accessed prior to 11 August 2017). patterns of care and outcome in the United States. American
Journal of Respiratory and Critical Care Medicine 2002;165
FDA HAP/VAP industry guidance
(6):766–72.
FDA. Guidance for industry. Hospital-acquired bacterial
pneumonia and ventilator-associated bacterial pneumonia: Krinsley 2003
developing drugs for treatment. www.fda.gov/downloads/ Krinsley JS. Association between hyperglycemia and
Drugs/GuidanceComplianceRegulatoryInformation/ increased hospital mortality in a heterogeneous population
Guidances/UCM234907.pdf (accessed prior to 11 August of critically ill patients. Mayo Clinic Proceedings 2003;78
2017). (12):1471.
Corticosteroids for pneumonia (Review) 24
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Lee 2010 Moher 2009
Lee GE, Lorch SA, Sheffler-Collins S, Kronman MP, Shah Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA
SS. National hospitalization trends for pediatric pneumonia Group. Preferred reporting items for systematic reviews and
and associated complications. Pediatrics 2010;126(2): meta-analyses: the PRISMA statement. Annals of Internal
204–13. Medicine 2009;151(4):264–9.
Lefebvre 2011 Musher 2014
Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching Musher DM, Thorner AR. Community-acquired
for studies. In: Higgins JP, Green S editor(s). Cochrane pneumonia. New England Journal of Medicine 2014;271
Handbook for Systematic Reviews of Interventions Version (17):1619–28.
5.1.0 (updated March 2011). The Cochrane Collaboration, Niederman 2007
2011. Available from handbook.cochrane.org. Niederman MS. Recent advances in community-acquired
pneumonia: inpatient and outpatient. Chest 2007;131(4):
Lim 2009
1205–15.
Lim WS, Baudouin SV, George RC, Hill AT, Jamieson C,
Le Jeune I, et al. Pneumonia Guidelines Committee of the Oray 2016
BTS Standards of Care Committee. BTS guidelines for the Oray M, Abu Samra K, Ebrahimiadib N, Meese H, Foster
management of community acquired pneumonia in adults: CS. Long-term side effects of glucocorticoids. Expert
update 2009. Thorax 2009;64(Suppl 3):iii1. Opinion on Drug Safety 2016;15(4):457–65.
Restrepo 2008
Magill 2014
Restrepo MI, Mortensen EM, Velez JA, Frei C, Anzueto A.
Magill SS, Edwards JR, Bamberg W, Beldavs ZG, Dumyati
A comparative study of community-acquired pneumonia
G, Kainer MA, et al. Emerging Infections Program
patients admitted to the ward and the ICU. Chest 2008;133
Healthcare-Associated Infections and Antimicrobial Use
(3):610–7.
Prevalence Survey Team. Multistate point-prevalence survey
of health care-associated infections. New England Journal of Review Manager 2014 [Computer program]
Medicine 2014;370(13):1198–208. Nordic Cochrane Centre, The Cochrane Collaboration.
Review Manager 5 (RevMan 5). Version 5.3. Copenhagen:
Mandell 2007 Nordic Cochrane Centre, The Cochrane Collaboration,
Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, 2014.
Campbell GD, Dean NC, et al. Infectious Diseases Society
Rhen 2005
of America. Infectious Diseases Society of America/
Rhen T, Cidlowski JA. Antiinflammatory action of
American Thoracic Society consensus guidelines on the
glucocorticoids - new mechanisms for old drugs. New
management of community-acquired pneumonia in adults.
England Journal of Medicine 2005;353(16):1711–23.
Clinical Infectious Diseases 2007;44(Suppl 2):S27–72.
Rhodes 2017
Mandell 2015 Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M,
Ellison RT 3rd, Donowitz GR. Acute pneumonia. In: Ferrer R, et al. Surviving Sepsis Campaign: International
Mandell GL, Bennett JE editor(s). Mandell, Douglas and Guidelines for Management of Sepsis and Septic Shock:
Bennett’s Principles and Practice of Infectious Diseases. 8th 2016. Critical Care Medicine 2017;45(3):486–552.
Edition. Vol. 1, London: Churchill Livingstone, 2015:
Rosenthal 2012
823–46.
Rosenthal VD, Bijie H, Maki DG, Mehta Y, Apisarnthanarak
Marik 2008 A, Medeiros EA, et al. International Nosocomial Infection
Marik PE, Pastores SM, Annane D, Meduri GU, Sprung Control Consortium (INICC) report, data summary of 36
CL, Arlt W, et al. American College of Critical Care countries, for 2004-2009. American Journal of Infection
Medicine. Recommendations for the diagnosis and Control 2012;40(5):396–407.
management of corticosteroid insufficiency in critically ill
Ruan 2014
adult patients: consensus statements from an international
Ruan SY, Lin HH, Huang CT, Kuo PH, Wu HD, Yu CJ.
task force by the American College of Critical Care
Exploring the heterogeneity of effects of corticosteroids on
Medicine. Critical Care Medicine 2008;36(6):1937–49.
acute respiratory distress syndrome: a systematic review and
Marrie 2007 meta-analysis. Critical Care 2014;18(2):R63.
Marrie TJ, Shariatzadeh MR. Community-acquired Safdar 2005
pneumonia requiring admission to an intensive care unit: a Safdar N, Dezfulian C, Collard HR, Saint S. Clinical and
descriptive study. Medicine 2007;86(2):103–11. economic consequences of ventilator-associated pneumonia:
Marti 2015 a systematic review. Critical Care Medicine 2005;33(10):
Marti C, Grosgurin O, Harbarth S, Combescure C, Abbas 2184–93.
M, Rutschmann O, et al. Adjunctive corticotherapy for Salluh 2008
community acquired pneumonia: a systematic review and Salluh JI, Póvoa P, Soares M, Castro-Faria-Neto HC,
meta-analysis. PLoS ONE 2015;10(12):e0144032. Bozza FA, Bozza PT. The role of corticosteroids in severe
Corticosteroids for pneumonia (Review) 25
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
community-acquired pneumonia: a systematic review. Research Methodology 2014;14:135. DOI: 10.1186/
Critical Care 2008;12(3):R76. 1471-2288-14-135
Schäcke 2002 Wan 2016
Schäcke H, Döcke WD, Asadullah K. Mechanisms involved Wan YD, Sun TW, Liu ZQ, Zhang SG, Wang LX, Kan
in the side effects of glucocorticoids. Pharmacology & QC. Efficacy and safety of corticosteroids for community-
Therapeutics 2002;96(1):23–43. acquired pneumonia: a systematic review and meta-analysis.
Chest 2016;149(1):209–19.
Siemieniuk 2015
Warrington 2006
Siemieniuk RA, Meade MO, Alonso-Coello P, Briel M,
Warrington TP, Bostwick JM. Psychiatric adverse effects
Evaniew N, Prasad M, et al. Corticosteroid therapy for
of corticosteroids. Mayo Clinic Proceedings 2006;81(10):
patients hospitalized with community-acquired pneumonia:
1361–7.
a systematic review and meta-analysis. Annals of Internal
Medicine 2015;163(7):519–28. Wood 2008
Wood L, Egger M, Gluud LL, Schulz KF, Jüni P, Altman
Torres 2010 DG, et al. Empirical evidence of bias in treatment effect
Torres A, Ferrer M, Badia JR. Treatment guidelines and estimates in controlled trials with different interventions
outcomes of hospital-acquired and ventilator-associated and outcomes: meta-epidemiological study. BMJ 2008;336
pneumonia. Clinical Infectious Diseases 2010;51(Suppl 1): (7644):601–5.
S48–53.
Wunderink 2015
Virkki 2002 Wunderink RG. Corticosteroids for severe community-
Virkki R, Juven T, Rikalainen H, Svedström E, Mertsola acquired pneumonia: not for everyone. JAMA 2015;313
J, Ruuskanen O. Differentiation of bacterial and viral (7):673–4.
pneumonia in children. Thorax 2002;57(5):438–41.
Xu 2016
Walters 2014 Xu J, Murphy SL, Kochanek KD, Bastian BA. Deaths: Final
Walters JA, Tan DJ, White CJ, Gibson PG, Wood- Data for 2013. National Vital Statistics Reports 2016;64(2):
Baker R, Walters EH. Systemic corticosteroids for acute 1–119.
exacerbations of chronic obstructive pulmonary disease.
Cochrane Database of Systematic Reviews 2014, Issue 9.
References to other published versions of this review
DOI: 10.1002/14651858.CD001288.pub4
Chen 2011
Wan 2014 Chen Y, Li K, Pu H, Wu T. Corticosteroids for pneumonia.
Wan X, Wang W, Liu J, Tong T. Estimating the sample Cochrane Database of Systematic Reviews 2011, Issue 3.
mean and standard deviation from the sample size, DOI: 10.1002/14651858.CD007720.pub2
∗
median, range and/or interquartile range. BMC Medical Indicates the major publication for the study
Blum 2015
Outcomes Primary outcome: time to clinical stability defined as time (days) until stable vital signs
for 24 hours
Mortality outcome definition: 30-day all-cause mortality
Other relevant outcomes:
• Clinical failure: defined as number of participants not reaching clinical stability
on day 5
• Need for invasive MV
• Need for ICU transfer
• Length of ICU stay
• Time to clinical cure: defined as time to stable vital signs for 24 hours or longer
• Length of hospitalisation
• Pneumonia complications: defined as empyema rates
• Superinfections
• Adverse events: any adverse event, GI bleeding, hyperglycaemia, neuropsychiatric
adverse events, and cardiac adverse events
Full agreement between outcomes in registry vs results: only primary outcome defined
in registry
Full agreement between outcomes in methods vs results: yes, except for timing of CAP
scores defined only in results
Risk of bias
Allocation concealment (selection bias) Low risk Allocation was concealed with a prespec-
ified computer-generated randomisation
list kept centrally at the pharmacy of the
main study centre
Blinding of outcome assessment (detection Low risk Outcome assessor was blinded.
bias)
All outcomes
Incomplete outcome data (attrition bias) Low risk Intention-to-treat data presented.
All outcomes
Selective reporting (reporting bias) High risk Only primary outcome reported in registry,
differences in outcome description between
methods and results
Other bias Low risk Sample size calculation without early stop
Confalonieri 2005
Outcomes Primary outcome: improvement in PaO :FiO (PaO :FiO 300 or 100 increase from
study entry) and mortality
Mortality outcome definition: 28 days, in hospital, and 60-days all-cause mortality
Other relevant outcomes:
• Clinical failure: defined as number of participants not achieving PaO
:FiO
improvement ≥ 100 from study entry at day 8
• Development of shock not present at randomisation
• Length of ICU stay
• Time to clinical cure: defined as duration of mechanical ventilation
• Length of hospitalisation
• Pneumonia complications: defined as lung abscess rates
• Superinfections
• Adverse events: GI bleeding and cardiac adverse events
Full agreement between outcomes in registry vs results: no registry
Full agreement between outcomes in methods vs results: no
Risk of bias
Random sequence generation (selection Low risk Randomisation schemes were generated in
bias) blocks of 10 for each participating site by a
central randomisation centre
Allocation concealment (selection bias) Low risk The randomisation assignment provided to
the recruiting centre in sealed envelopes
Incomplete outcome data (attrition bias) Low risk All participants accounted for, except 2 par-
All outcomes ticipants who met the exclusion criteria af-
ter randomisation
Selective reporting (reporting bias) High risk No agreement between outcomes reported
in methods and in results
El-Ghamrawy 2006
Risk of bias
Blinding of participants and personnel Unclear risk Placebo use without mention of blinding
(performance bias) and no description of the placebo
All outcomes
Selective reporting (reporting bias) High risk No agreement in outcomes between meth-
ods and results
Risk of bias
Incomplete outcome data (attrition bias) Low risk Intention-to-treat data presented.
All outcomes
Selective reporting (reporting bias) Low risk Full agreement between outcomes in reg-
istry, methods, and results
Other bias Low risk Sample size calculation without early stop
Hatakeyama 1995
Risk of bias
Selective reporting (reporting bias) High risk No registry, outcomes not defined in meth-
ods.
Luo 2014
Outcomes Primary outcome: not specifically defined, implied: duration of dyspnoea and hypoxia
Mortality outcome definition: 7-days all-cause mortality
Other relevant outcomes:
• Clinical failure: defined as number of participants with no infiltrate resolution at
day 7
• Time to clinical cure: defined as time to resolution of hypoxaemia
• Superinfections
• Adverse events: any adverse event and adverse events requiring discontinuation of
treatment
Full agreement between outcomes in registry vs results: no registry
Full agreement between outcomes in methods vs results: no
Risk of bias
Selective reporting (reporting bias) High risk No agreement in outcomes between meth-
ods and results
Risk of bias
Random sequence generation (selection Low risk Randomisation by a random number gen-
bias) erator
Selective reporting (reporting bias) High risk Outcomes not defined in methods, no reg-
istry.
McHardy 1972
Full agreement between outcomes in methods vs results: outcomes not defined in meth-
ods
Risk of bias
Random sequence generation (selection Unclear risk Randomisation method not clearly speci-
bias) fied.
Allocation concealment (selection bias) Low risk Allocation concealed by sealed and num-
bered envelopes.
Incomplete outcome data (attrition bias) Unclear risk There were postrandomisation dropouts,
All outcomes but numbers are not provided. Number
randomised provided, but number evalu-
ated not clear
Selective reporting (reporting bias) High risk Outcomes not defined in methods, no reg-
istry.
Meijvis 2011
Outcomes Primary outcome: length of hospital stay in days until hospital discharge or death
Mortality outcome definition: 30-day all-cause mortality
Other relevant outcomes:
• Need for ICU transfer
• Length of ICU stay
• Length of hospitalisation
• Pneumonia complications: defined as empyema or pleural effusion
• Superinfections
• Adverse events: GI symptoms, hyperglycaemia, and cardiac adverse events
Full agreement between outcomes in registry vs results: no, more outcomes in results
Full agreement between outcomes in methods vs results: yes
Risk of bias
Random sequence generation (selection Low risk Computer-generated random number ta-
bias) ble
Selective reporting (reporting bias) High risk Not all outcomes defined in registry
Other bias Low risk Sample size calculation without early stop
Mikami 2007
Risk of bias
Random sequence generation (selection Unclear risk Randomisation method not specified.
bias)
Allocation concealment (selection bias) Unclear risk Allocation concealment method not speci-
fied.
Selective reporting (reporting bias) High risk No registry, no agreement between out-
comes in methods and results
Other bias Low risk Sample size calculation without early stop
Nafae 2013
Risk of bias
Blinding of participants and personnel High risk Single-blind; only participants were
(performance bias) blinded.
All outcomes
Selective reporting (reporting bias) High risk No registry, outcomes not specified in
methods.
Nagy 2013
Risk of bias
Allocation concealment (selection bias) Unclear risk Allocation concealment method not clearly
specified.
Blinding of participants and personnel High risk The control group was given placebo, but
(performance bias) according to the methods description there
All outcomes was no blinding
Selective reporting (reporting bias) High risk Outcomes not defined in registry.
Other bias Low risk Sample size calculation without early stop
Sabry 2011
Interventions • Intervention: IV hydrocortisone (loading dose of 200 mg, followed by 12.5 mg/h)
versus placebo
• Day start: not specified
• Planned duration: 7 days
• Follow-up: 8 days
Outcomes Primary outcome: improvement in PaO :FiO (PaO :FiO > 300 or ≥ 100 increase
from study entry) and Sepsis-related Organ Failure Assessment (SOFA) score by day 8
and the development of delayed septic shock
Mortality outcome definition: 8 days
Other relevant outcomes:
• Clinical failure: defined as number of participants not achieving PaO
:FiO
improvement ≥ 100 mm Hg compared to study entry, evaluated at day 8
• Development of shock not present at randomisation
• Time to clinical cure: defined as time to weaning from mechanical ventilation
• Pneumonia complications: defined as lung abscess or ARDS
• Superinfections
• Adverse events: GI symptoms, GI bleeding, and cardiac adverse events
Full agreement between outcomes in registry vs results: no, more outcomes in results
Full agreement between outcomes in methods vs results: yes
Risk of bias
Selective reporting (reporting bias) High risk Not all outcomes presented in results were
defined in registry
Other bias Unclear risk No sample size calculation. Early stop not
reported.
Snijders 2010
• Follow-up: 30 days
Risk of bias
Incomplete outcome data (attrition bias) Low risk Intention-to-treat data presented.
All outcomes
Selective reporting (reporting bias) High risk Not all outcomes defined in registry.
Other bias Low risk Sample size calculation without early stop
Outcomes Primary outcome: treatment failure (composite outcome of early treatment failure -
within 72 hours, late treatment failure between 72 and 120 hours, or both early and late
treatment failure)
Mortality outcome definition: in-hospital mortality
Other relevant outcomes:
• Clinical failure: defined as number of participants with radiographic progression
(increase of ≥ 50% of pulmonary infiltrates compared with baseline), persistence of
severe respiratory failure (ratio of PaO
to fraction of inspired oxygen < 200 mm Hg or with respiratory rate ≥ 30 breaths/
min in participants not intubated), development of shock, need for invasive MV not
present at baseline, or death after 120 hours
• Development of shock not present at randomisation
• Need for invasive MV
• Need for ICU transfer
• Length of ICU stay
• Time to clinical cure: defined as when the following values were achieved for all
parameters: temperature of 37.2 °C or lower, heart rate of 100 beats/min or lower,
systolic BP 90 mm Hg or higher, and arterial oxygen tension of 60 mm Hg or higher
when the participant was not receiving supplemental oxygen
• Length of hospitalisation
• Superinfections
• Adverse events: GI bleeding, hyperglycaemia, and neuropsychiatric adverse events
Full agreement between outcomes in registry vs results: only primary outcome defined
in registry
Full agreement between outcomes in methods vs results: yes
Risk of bias
Incomplete outcome data (attrition bias) Low risk Intention-to-treat data presented.
All outcomes
Selective reporting (reporting bias) High risk Not all outcomes defined in registry.
Other bias Low risk Sample size calculation without early stop
Risk of bias
Allocation concealment (selection bias) Low risk Trial medication prepared in advance in the
pharmacy centre, where the concealed ran-
domisation list was kept until the study was
completed
Incomplete outcome data (attrition bias) Low risk 3 children excluded postrandomisation,
All outcomes but exclusion was justified
Selective reporting (reporting bias) Low risk No registry, full agreement between out-
comes in methods vs results
Other bias Low risk Sample size calculation without early stop
Risk of bias
Selective reporting (reporting bias) Low risk No registry, full agreement between meth-
ods and results
Other bias Unclear risk No sample size calculation. Early stop not
reported.
Cao 2007 Examined the use of inhaled corticosteroids, not systemic corticosteroid treatment as specified in our inclusion
criteria
Shan 2017 Participants in the control group were subsequently given either corticosteroids or intravenous immunoglobulin
therapy
Van Woensel 2011 Population included people with and without pneumonia, and outcomes were not reported separately for
participants with pneumonia
Lan 2015
Methods Unclear, presented as randomised, however the methods section described selection of children for different man-
agement strategies
Outcomes Treatment failure reported, however results in abstract and in text are inconsistent
Notes Awaiting clarification from authors; if they are unreachable, we will exclude the study
NCT01283009
Interventions Methylprednisolone
Notes Completed
NCT02618057
NCT02735707
Trial name or title Randomized, embedded, multifactorial adaptive platform trial for community-acquired pneumonia
(REMAP-CAP)
NCT03121690
Trial name or title The applicability of different scoring systems and use of steroids in the treatment of hospital acquired pneu-
monia
Methods Randomised
Interventions Prednisone
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Mortality - adults 11 1863 Risk Ratio (M-H, Fixed, 95% CI) 0.66 [0.47, 0.92]
1.1 Low risk of bias for 6 1594 Risk Ratio (M-H, Fixed, 95% CI) 0.80 [0.54, 1.19]
allocation concealment
1.2 Unclear risk of bias for 5 269 Risk Ratio (M-H, Fixed, 95% CI) 0.36 [0.19, 0.71]
allocation concealment
2 Mortality - adults, severe 9 995 Risk Ratio (M-H, Fixed, 95% CI) 0.58 [0.40, 0.84]
pneumonia, by allocation
concealment
2.1 Low risk of bias 5 771 Risk Ratio (M-H, Fixed, 95% CI) 0.72 [0.46, 1.13]
2.2 Unclear risk of bias 4 224 Risk Ratio (M-H, Fixed, 95% CI) 0.34 [0.17, 0.68]
3 Mortality - adults, non-severe 4 868 Risk Ratio (M-H, Fixed, 95% CI) 0.95 [0.45, 2.00]
pneumonia, by allocation
concealment
3.1 Low risk of bias 3 823 Risk Ratio (M-H, Fixed, 95% CI) 0.95 [0.44, 2.06]
3.2 Unclear risk of bias 1 45 Risk Ratio (M-H, Fixed, 95% CI) 0.96 [0.06, 14.37]
4 Mortality - children 4 266 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.03, 0.03]
4.1 Bacterial pneumonia 3 225 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.03, 0.03]
4.2 Viral pneumonia 1 41 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.09, 0.09]
5 Early clinical failure - adults 6 1324 Risk Ratio (M-H, Random, 95% CI) 0.40 [0.23, 0.70]
5.1 Severe CAP 5 419 Risk Ratio (M-H, Random, 95% CI) 0.32 [0.15, 0.70]
5.2 Non-severe CAP 2 905 Risk Ratio (M-H, Random, 95% CI) 0.68 [0.56, 0.83]
6 Early clinical failure - children 2 88 Risk Ratio (M-H, Fixed, 95% CI) 0.41 [0.24, 0.70]
6.1 Bacterial pneumonia 2 88 Risk Ratio (M-H, Fixed, 95% CI) 0.41 [0.24, 0.70]
7 Early clinical failure - adults 8 1412 Risk Ratio (M-H, Random, 95% CI) 0.40 [0.26, 0.63]
+ children by allocation
concealment
7.1 Low risk of bias 4 1164 Risk Ratio (M-H, Random, 95% CI) 0.59 [0.37, 0.94]
7.2 Unclear risk of bias 4 248 Risk Ratio (M-H, Random, 95% CI) 0.27 [0.14, 0.52]
8 Time to clinical cure - adults 9 1322 Mean Difference (IV, Random, 95% CI) -1.83 [-2.45, -1.21]
9 Time to clinical cure - children 4 Mean Difference (IV, Random, 95% CI) Subtotals only
9.1 Bacterial pneumonia 3 225 Mean Difference (IV, Random, 95% CI) -1.57 [-2.55, -0.60]
9.2 Viral pneumonia 1 41 Mean Difference (IV, Random, 95% CI) 1.70 [-2.50, 5.90]
10 Need for mechanical ventilation 4 1030 Risk Ratio (M-H, Fixed, 95% CI) 0.40 [0.20, 0.77]
- adults
11 Development of shock - adults 6 415 Risk Ratio (M-H, Fixed, 95% CI) 0.18 [0.09, 0.34]
12 Need for ICU transfer - adults 4 1164 Risk Ratio (M-H, Fixed, 95% CI) 0.73 [0.45, 1.18]
13 Length of hospitalisation - 9 1658 Mean Difference (IV, Random, 95% CI) -2.91 [-4.92, -0.89]
adults
14 Length of ICU stay - adults 8 342 Mean Difference (IV, Fixed, 95% CI) -1.88 [-2.96, -0.81]
15 Length of hospitalisation - 2 Mean Difference (IV, Random, 95% CI) Totals not selected
children
15.1 Bacterial pneumonia 1 Mean Difference (IV, Random, 95% CI) 0.0 [0.0, 0.0]
15.2 Viral pneumonia 1 Mean Difference (IV, Random, 95% CI) 0.0 [0.0, 0.0]
Corticosteroids for pneumonia (Review) 56
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
16 Pneumonia complications - 9 1632 Risk Ratio (M-H, Fixed, 95% CI) 0.58 [0.40, 0.84]
adults + children
17 Secondary infections - adults 7 1533 Risk Ratio (M-H, Fixed, 95% CI) 1.19 [0.73, 1.93]
18 Secondary infections - children 3 225 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.03, 0.03]
18.1 Bacterial pneumonia 3 225 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.03, 0.03]
19 Any adverse events - adults 3 1028 Risk Ratio (M-H, Fixed, 95% CI) 1.21 [0.99, 1.47]
20 Hyperglycaemia - adults 7 1578 Risk Ratio (M-H, Fixed, 95% CI) 1.72 [1.38, 2.14]
21 Gastrointestinal bleeding - 7 1190 Risk Ratio (M-H, Fixed, 95% CI) 0.91 [0.40, 2.05]
adults
22 Neuropsychiatric side effects - 4 1149 Risk Ratio (M-H, Fixed, 95% CI) 1.95 [0.70, 5.42]
adults
23 Adverse cardiac events - adults 5 1249 Risk Ratio (M-H, Fixed, 95% CI) 0.60 [0.32, 1.13]
24 Any adverse events - children 2 Risk Difference (M-H, Fixed, 95% CI) Subtotals only
24.1 Bacterial pneumonia 2 117 Risk Difference (M-H, Fixed, 95% CI) 0.0 [-0.05, 0.05]
ADDITIONAL TABLES
Table 1. Study definitions for clinical failure and time to cure
Blum 2015 Number of participants not reaching clinical stabil- Time to clinical stability defined as stable vital signs
ity on day 5 for ≥ 24 hours
Clinical stability defined as stable vital signs for ≥ Stable vital signs were temperature of ≤ 37.8 °C,
24 hours. Stable vital signs were temperature of ≤ heart rate of ≤ 100 beats per min, spontaneous res-
37.8 °C, heart rate of ≤ 100 beats per min, spon- piratory rate of ≤ 24 breaths per min, systolic blood
taneous respiratory rate of ≤ 24 breaths per min, pressure of ≥ 90 mm Hg (≥ 100 mm Hg for partic-
systolic blood pressure of ≥ 90 mm Hg (≥ 100 mm ipants with hypertension) without vasopressor sup-
Hg for participants with hypertension) without va- port, mental status back to level before occurrence
sopressor support, mental status back to level be- of community-acquired pneumonia, ability for oral
fore occurrence of community-acquired pneumo- intake, and adequate oxygenation on room air (PaO
nia, ability for oral intake, and adequate oxygena- ≥ 60 mm Hg or pulse oximetry ≥ 90%)
tion on room air (PaO ≥ 60 mm Hg or pulse
oximetry ≥ 90%)
Confalonieri 2005 Number of participants not achieving PaO :FiO Time to weaning from mechanical ventilation
improvement ≥ 100 mm Hg compared to study
entry, evaluated at day 8
Luo 2014 Number of participants with no infiltrate resolution Time to resolution of hypoxaemia
at day 7
Nafae 2013 Number of participants with no improvement at Time to weaning from mechanical ventilation
day 7 (improvement definition not provided)
Nagy 2013 Number of participants not improving based on Time to fever resolution
clinical and radiological status on day 7
Sabry 2011 Number of participants not achieving PaO :FiO Time to weaning from mechanical ventilation
improvement ≥ 100 mm Hg compared to study
entry, evaluated at day 8
Snijders 2010 Number of participants with clinical failure at day Time to clinical stability defined as when all 4 of the
7. following criteria were met: improvement of cough
Clinical failure defined as: persistence or progression and shortness of breath, temperature < 37.8 °C for
of all signs and symptoms that developed during the at least 8 hours, declining serum C-reactive protein
acute disease episode after randomisation, or the de- levels, and adequate oral intake and gastrointestinal
velopment of a new pulmonary or extrapulmonary absorption
infection, or the deterioration of chest radiography
after randomisation, or death due to pneumonia, or
the inability to complete the study owing to adverse
events
Torres 2015 Number of participants with treatment failure be- Time to clinical stability defined as when all of the
tween 72 hours and 120 hours after treatment ini- following criteria were met: temperature ≤ 37.2 °C,
tiation. heart rate ≤ 100 beats/min, systolic blood pres-
Treatment failure defined as radiographic progres- sure ≥ 90 mm Hg, and arterial oxygen tension ≥
sion (increase of ≥ 50% of pulmonary infiltrates 60 mm Hg when the participant was not receiving
compared with baseline), persistence of severe respi- supplemental oxygen. In participants receiving oxy-
ratory failure (PaO :FiO < 200 mm Hg, with res- gen therapy at home, stability was considered to be
piratory rate ≥ 30 breaths/min in participants not achieved when oxygen needs were the same as be-
intubated), development of shock, need for invasive fore admission
mechanical ventilation not present at baseline, or
death
3 March 2017 New search has been performed A new team of authors updated this review.
3 March 2017 New citation required and conclusions have changed We included 12 new studies in this update (Blum 2015; El-
Ghamrawy 2006; Fernández-Serrano 2011; Hatakeyama
1995; Luo 2014; Meijvis 2011; Nafae 2013; Nagy 2013;
Sabry 2011; Snijders 2010; Torres 2015; Wu 2014). We
excluded one previously included study, Cao 2007, and ex-
cluded five new trials (Huang 2014; Montón 1999; Shan
2017; Van Woensel 2011; Wagner 1956). One trial is await-
ing classification (Lan 2015).
We found that corticosteroids reduce mortality and mor-
bidity in adults with severe community-acquired pneumo-
nia and morbidity in adults and children with non-severe
community-acquired pneumonia
HISTORY
Protocol first published: Issue 2, 2009
Review first published: Issue 3, 2011
CONTRIBUTIONS OF AUTHORS
SOURCES OF SUPPORT
Internal sources
• No sources of support supplied
External sources
• Cochrane Acute Respiratory Infections Group, Australia.
• We prospectively rewrote a protocol for this update, before starting the update.
• We defined new objectives for this review not identical to the 2011 review objectives. Specifically, we broadened the efficacy
assessment to include mortality and morbidity, and we did not assess relative adrenal insufficiency and dose-effect relationships, which
were defined as objectives in the original review.
• The inclusion and exclusion criteria for this update are similar to the original 2011 review with a few differences. Differing from
the 2011 review we did not exclude studies including participants with immunosuppression, tuberculosis, acute schistosomiasis,
fungal or parasitic infections, or chemotherapy and radiotherapy, as we believed these can be pooled together with other pneumonia
patients. We excluded studies including neonates and people with HIV and Pneumocystis pneumonia, as we believed these represent
different entities and require separate consideration.
• There were several differences in the types of interventions assessed in this update compared to the 2011 review. First, we
specified corticosteroid therapy to include only systemic administration and exclude inhaled corticosteroids. The mechanisms of
action of these two interventions are different and cannot be pooled. We included only trials comparing corticosteroids to placebo or
no treatment and excluded trials in which corticosteroids were given to both treatment arms, as the question of the review is on the
efficacy of corticosteroid therapy.
• We assessed all-cause mortality as our primary outcome, as this is the currently recommended outcome for assessment in severe
infections and the only outcome that will change practice. We added several secondary outcomes that were not collected in the 2011
review, including early clinical failure, length of hospitalisation, and pneumonia complications. We also specified the outcome of
adverse effects and added outcomes of specific adverse effects including superinfections, adverse effects requiring discontinuation of
corticosteroids, hyperglycaemia, gastrointestinal bleeding, and neuropsychiatric and cardiac adverse events. The outcomes assessed in
our review are aligned with contemporary guidance for outcome assessment in clinical trials of pneumonia (FDA CAP industry
guidance; FDA HAP/VAP industry guidance; Mandell 2007), and we believe that evidence summaries should address these relevant
Corticosteroids for pneumonia (Review) 60
Copyright © 2017 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
outcomes. We changed the outcome definition of ’mortality’ to ’all-cause mortality’ because we believe that ’all-cause mortality’ more
clearly specifies that we collected all death cases including pneumonia-related cases and cases not related to pneumonia. To note, we
revised the outcomes during the writing of the protocol and before starting the update.
• We changed subgroup analyses to address the contemporary relevant clinical questions. We performed subgroup analyses based
on different patient characteristics (pneumonia severity, comorbidities, pathogen, etc.) and added meta-regression analyses when
subgroup analysis was not applicable.
• We re-applied inclusion and exclusion criteria and re-extracted all data in duplicate, since we did not agree with the study
selection and outcome extraction in a sample of tested articles, and since ’Risk of bias’ methodology has developed since the 2011
review.
INDEX TERMS