Corticosteroids as adjunctive therapy in the treatment of influenza

  • Protocol
  • Intervention



This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effectiveness and potential adverse effects of corticosteroids as an adjuvant therapy for influenza.


Description of the condition

Viral upper respiratory tract infections account for the majority of acute presentations to general practice (NICE 2008). Implicated pathogens vary by season and they are commonly due to influenza virus, respiratory syncytial virus and rhinovirus (HPA UK 2012). Considerable overlap exists in the clinical presentation of these infections (Bellei 2008; Eccles 2005) with the majority of affected individuals suffering from a mild, self limiting clinical illness characterised by fever, cough, sore throat, myalgia, headaches and gastrointestinal symptoms.

Influenza, in particular, is a significant cause of morbidity and mortality worldwide, with World Health Organization (WHO) estimates of one billion cases; three to five million cases of severe illness and 300,000 to 500,000 deaths annually (WHO 2008). Complications of influenza include primary viral pneumonia which may lead to adult respiratory distress syndrome (ARDS), secondary bacterial pneumonia and extrapulmonary manifestations including myocarditis and neurological sequelae. Severely ill patients may require prolonged hospitalisation and intensive care support; the associated mortality rate is high (Cheng 2012; Thompson 2003).

Seasonal influenza occurs annually during the winter months in temperate zones of both the Northern and Southern hemispheres and all year round in the tropics (Viboud 2006). Pandemic influenza (due to antigenic shifts of the virus) emerges unpredictably and infrequently with varying disease severity. The case fatality rates of the pandemics in 1918, 1957 and 1968 ranged from 0.2% to 3%, whereas the 2009 pandemic was similar in virulence to seasonal influenza, with an estimated case fatality rate of 0.03% (Donaldson 2010). However, young adults were disproportionately affected compared to seasonal influenza (Viboud 2010). The highest rates of hospitalisation were in children under 15 years of age (Kerkhove 2011).

Specific treatment options for influenza are limited to antiviral drugs such as neuraminidase inhibitors. A Cochrane Review of neuraminidase inhibitors for the treatment of influenza in adults and children reported that the duration of flu-like symptoms was shortened by these drugs. However, the review authors were unable to draw firm conclusions about the effects on complications of influenza, or transmission (Jefferson 2012). Not all reviews are as conservative as Cochrane Reviews (Arunachalam 2011; Hsu 2012) but all acknowledge the lack of robust data relating to outcomes of public health importance, such as severe complications, hospitalisation and death.

Currently, a Cochrane Review of statin use in the treatment of influenza is underway (Khandaker 2011). A Cochrane Review of Chinese medicinal herbs in the prevention and treatment of influenza concluded that the current evidence was insufficient to recommend its routine use (Chen 2007).

Description of the intervention

Corticosteroids are produced endogenously in the adrenal glands from cholesterol and regulated by the hypothalamic-pituitary-adrenal axis (Molenaar 2012). Synthetically derived analogues such as betamethasone, dexamethasone, hydrocortisone, methylprednisolone, prednisolone and triamcinolone are commonly used in clinical practice (BNF 2012). However, the evidence base for their use shows wide variation across the different conditions.

In the setting of severe sepsis and septic shock, high doses of corticosteroids in short courses were initially thought to be beneficial (Schumer 1976). However, several meta-analyses later demonstrated no overall benefit in mortality with this regime (Cronin 1995; Lefering 1995). A randomised, double-blind, placebo-controlled study of low-dose hydrocortisone and fludrocortisone (Annane 2002) demonstrated reduced mortality and a reduced need for vasopressor use at 28 days in the treated group. Several double-blind randomised studies later replicated these findings (Keh 2003; Oppert 2005). A systematic review in 2009 reported that low-dose corticosteroid use increased 28-day shock reversal, and reduced intensive care unit length of stay and 28-day mortality, with no associated increase in gastrointestinal bleeding rate, superinfection or neuromuscular weakness (Annane 2009).

When used for the treatment of bacterial meningitis, corticosteroids appear to reduce hearing loss and neurological complications (Brouwer 2010), while in tuberculous meningitis, an improvement in survival has been reported (Prasad 2008).

With relevance to respiratory virus infection, a Cochrane Review of systemic corticosteroid use in all-cause pneumonia found no mortality benefit but a reduction in time to resolution of symptoms (Chen 2011). Similarly, patients given systemic corticosteroids or inhaled corticosteroids in acute sinusitis were more likely to have a shorter time to resolution of symptoms (Venekamp 2011) or were more likely to experience a resolution of their symptoms (Zalmanovici Trestioreanu 2011), respectively. A review of corticosteroid use in croup found a lower symptom score at six hours, re-admission rate and length of stay in the treated group (Russell 2011). No benefits were seen in hospital admission rates, or length of stay in hospital, following systemic or inhaled corticosteroid use in acute viral bronchiolitis in infants and young children (Ricardo 2010). Intranasal corticosteroids used in the treatment of the common cold did not appear to reduce severity or length of illness, although the review was limited by small study numbers (Hayward 2012).

Due to a lack of robust evidence, corticosteroid use in influenza is currently inconsistent. Several studies investigating patients admitted to hospital and intensive care units during the 2009 influenza H1N1 pandemic reported that 30% to 50% had received corticosteroids (Brun-Buisson 2011; Chan 2011; Kim 2011; Martin-Loeches 2011). The World Health Organization consultation on human influenza A (H5N1) infection reported that 47% to 70% of patients received corticosteroids during the 2004 to 2005 outbreak in South East Asia (WHO 2005).

How the intervention might work

Viral replication and production of cytokines through activation of the host innate immune system are central in the pathogenesis of influenza infection (de Jong 2006). Elevated or excessive production of cytokines (hypercytokinaemia) correlates with symptoms and fever in acute influenza (Kaiser 2001). Comparisons between patients with mild and severe pandemic H1N1 influenza have revealed significantly higher levels of cytokines (especially interleukin-6) in the plasma of patients with severe disease (Yu 2011) and similar findings have been replicated in studies of severe seasonal influenza (Heltzer 2009). High levels of pro-inflammatory cytokines such as interleukin-1, interleukin-6 and tumour necrosis factor-α lead to activation of the hypothalamic-pituitary-adrenal axis (Chrousos 1995). The consequent production of corticosteroids is essential to maintain homeostasis during physiological stress through mechanisms such as inhibition of leucocyte trafficking and function, maintenance of epithelial integrity, regulation of vascular tone by inhibition of vasodilators (nitrous oxide) and increased sensitivity to vasopressors, and metabolic effects (Arafah 2006). However, up to 50% of patients with severe sepsis have evidence of adrenal dysfunction due to inadequate production of corticosteroids or impaired peripheral sensitivity to corticosteroids (Patel 2012). Administration of corticosteroids during severe influenza may attenuate this state of adrenal insufficiency.

Why it is important to do this review

There are no widely recognised large, good-quality randomised controlled trial (RCT) data on this topic, but it is not known if less robust trial data exist. Therefore, the authors believe that a systematic review of the evidence is warranted to identify and bring together what RCT data exist and, secondly, in view of the possible lack of robust RCT data, to identify non-RCT data and observational data that would provide further valuable clinical information to guide best practice. Such a review would also provide researchers with a stronger basis for a considered approach regarding further studies (including high-quality RCTs) in this area. If corticosteroids are found to be beneficial, there may be reductions in influenza-related hospitalisations, intensive care admissions, length of hospital stay and mortality. Conversely, if found to be harmful, the current unsystematic use of corticosteroids in influenza should be stopped. Either way, clarification of the role of corticosteroids in the management of influenza should lead to improvements in patient care and savings in healthcare costs.


To assess the effectiveness and potential adverse effects of corticosteroids as an adjuvant therapy for influenza.


Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs), quasi-RCTs, non-RCTs, prospective and retrospective cohort studies.

Types of participants

We will include:

  1. clinically diagnosed influenza or influenza-like illness (ILI), defined as fever, cough, symptoms of upper respiratory tract infection (coryza, sore throat) and constitutional symptoms (headache, myalgia) of acute onset; and/or

  2. microbiologically confirmed influenza through sampling of the respiratory tract (nasal swabs, throat swabs, bronchoalveolar lavage, etc).

We will not place restrictions on age or influenza type (i.e. influenza A, influenza B, H1N1, H5N1). We will consider studies conducted in out-patient settings as well as hospital settings for inclusion.

We will exclude studies addressing the effects of corticosteroids in ARDS, sepsis, pneumonia, sinusitis or any other respiratory infection, apart from influenza or ILI. (Note: this does not exclude studies of patients with pneumonia, ARDS, sepsis or other respiratory tract complications consequent on influenza or ILI).

Types of interventions

  1. Corticosteroids versus placebo.

  2. Corticosteroids plus supportive therapy versus supportive therapy alone.

  3. Corticosteroids plus antiviral therapy versus antiviral therapy alone.

  4. Different doses/durations of a given corticosteroid in the context of 1, 2 and 3.

  5. Treatment regimens including combinations of corticosteroids in the context of 1, 2 and 3.

Types of outcome measures

Primary outcomes
  1. For studies of hospitalised patients:

    1. number of deaths at 30 days following admission (30-day mortality);

    2. rate of admission to intensive care units.

  2. For studies in the community setting:

    1. rate of hospitalisation;

    2. time to resolution of symptoms;

    3. 30-day mortality.

Secondary outcomes
  1. For studies of hospitalised patients:

    1. hospital re-admission rate at 30 days post-discharge;

    2. number and nature of adverse events secondary to corticosteroid use, such as incidence of gastrointestinal bleeding, hospital-acquired infections and metabolic complications (e.g. hyperglycaemia, hypernatraemia);

    3. proportion of patients requiring mechanical ventilation;

    4. length of stay in hospital.

  2. For studies in the community setting:

    1. number and nature of adverse events secondary to corticosteroid use, such as incidence of gastrointestinal bleeding, hospital-acquired infections and metabolic complications (e.g. hyperglycaemia, hypernatraemia).

Search methods for identification of studies

Electronic searches

We will search the Cochrane Central Register of Controlled Trials (CENTRAL, latest issue) which contains the Cochrane Acute Respiratory Infections Group's Specialised Register, MEDLINE (1950 to present), CINAHL (1981 to present), EMBASE (1980 to present), LILACS (1982 to present) and Web of Science (1985 to present).

We will use the following search strategy to search CENTRAL and MEDLINE. We will combine the MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials (Lefebvre 2011). To identify observational studies we will use the Scottish Intercollegiate Guidelines Network (SIGN) filter (SIGN 2011). We will adapt the search strategy to search CINAHL, EMBASE, LILACS and Web of Science.


1 Influenza, Human/
2 exp Influenzavirus A/
3 exp Influenzavirus B/
4 or/1-3
5 exp Adrenal Cortex Hormones/
6 corticosteroid*.tw,nm.
7 adrenal cortex hormon*.tw.
8 (adren* cortic* adj1 (hormone* or steroid*)).tw.
9 adrenocortic*.tw,nm.
10 corticoid*.tw,nm.
11 glucocorticoid*.tw,nm.
12 hydroxycorticosteroid*.tw,nm.
13 exp Steroids/
14 steroid*.tw,nm.
15 (hydrocortisone* or prednisolone* or prednisone* or dexamethasone* or methylprednisolone*).tw,nm.
16 or/5-15
17 4 and 16

There will be no date, publication or language restrictions.

Searching other resources

We will search the Controlled Trials Registry ( for ongoing clinical trials. We will scrutinise the bibliographies of included studies in order to further identify relevant trials. We will also search the last three years of three major infectious diseases conferences (Interscience Conference on Antimicrobial Agents and Chemotherapy, European Society of Clinical Microbiology and Infectious Diseases and Asia Pacific Society of Infection Control) to identify potentially eligible studies.

Data collection and analysis

Selection of studies

Two authors (CR, WSL) will independently review all the trials retrieved using the search strategy described above. We will do this in two stages; firstly, by analysis of study titles and abstracts and secondly by analysis of the full text of the articles. We will resolve disagreements at any of these stages through discussion with a third review author (JVT). We will use a standardised proforma to apply the inclusion criteria against each study.

Data extraction and management

Two review authors (CR, JLB) will independently extract data using a standardised proforma specifically adapted for this review. We will extract the following data.

  1. Characteristics of study: general information, methodological details including 'Risk of bias' criteria for RCTs and the Newcastle-Ottawa Scale for non-randomised trials and comparative observational studies.

  2. Characteristics of participants: demographics, co-morbidities, numbers in study.

  3. Characteristics of intervention: type of steroid, route of administration, dose, timing of corticosteroid use (early versus late) and duration of treatment.

  4. Characteristics of outcome measures: deaths, complications of influenza, length of stay data, adverse outcomes of interventions and rates of hospitalisation.

Assessment of risk of bias in included studies

Two authors (CR, JLB) will independently assess methodological quality using the Cochrane 'Risk of bias' tool for RCTs (Higgins 2011). We will assess the following criteria.

  1. Sequence generation - was the method of generation of the randomisation sequence adequate?

  2. Allocation concealment - was the method of allocation concealment adequate? It will be considered 'low risk of bias' if the assignment could not be foreseen.

  3. Blinding - were participants, clinicians and outcome assessors blinded with regards to the intervention given?

  4. Incomplete outcome data - how many participants were lost to follow-up in each treatment group and were reasons for losses adequately reported?

  5. Incomplete outcome data - were all participants analyzed in the groups to which they were originally randomised (intention-to-treat (ITT) principle)?

  6. Selective outcome reporting (for RCTs only) - were all the primary outcomes listed in the study protocol that are relevant to this review, reported in the study?

  7. Other potential sources of bias.

We will rate trial quality as 'low risk of bias', 'high risk of bias' or 'unclear risk of bias'. We will resolve any disagreements by discussion amongst all review authors to reach a consensus. We will use the Newcastle-Ottawa Scale for non-randomised and comparative observational studies, and apply the validated 'star system' (Higgins 2011) to judge studies on three broad perspectives:

  1. selection of study groups;

  2. comparability of groups;

  3. ascertainment of outcome.

Measures of treatment effect

We will calculate risk ratios (RR) with corresponding 95% confidence intervals (CI) for dichotomous data such as mortality, intensive care admission rate, mechanical ventilation rate, 30-day re-admission (hospitalised patients) or hospital admission rate (community managed patients), and incidence of adverse events. We will use medians and interquartile ranges for continuous data that are not normally distributed such as length of stay. We will use mean difference (MD) or standardised mean difference (SMD) with corresponding 95% CI in the event of normally distributed continuous data.

Unit of analysis issues

We anticipate that the individual participant will be the unit of analysis for RCTs. However, if we identify any cluster-randomised controlled trials, then we will analyze the study allowing for the level of randomisation.

Dealing with missing data

We will analyze data on an ITT basis. For dichotomous outcomes, we will assess the effect assuming participants with missing data had a poor outcome. We will not use any form of imputation for participants with missing continuous outcome data. We will consult the CONSORT type flow chart (Schulz 2010) of participants through the study if available. Where no flow chart is available, we will look for information in the text of the results to determine whether all participants included in the study have been analyzed. Where there is ambiguity, we will contact the trial authors to seek further information.

Where there are missing data relating to results, for example, measures of dispersion, we will contact the trial authors of the study to request more information.

Assessment of heterogeneity

We will use the I2 statistic to assess heterogeneity across studies (RCTs, non-RCTs, comparative observational studies). We will consider a value greater than 75% as a substantial amount of heterogeneity between the studies and we will not pool data for meta-analysis; we will report the findings from the included studies individually.

Assessment of reporting biases

We will assess funnel plots for publication bias (small study bias) if we have a sufficient number of included trials (> 10 studies).

Data synthesis

One review author (CR) will enter data into Review Manager (RevMan 2012) and two review authors (CR, JLB) will independently summarise data. We will present data for RCTs separately from other types of studies (non-RCTs, comparative observational studies).

Where the interventions and populations are similar, we will perform a random-effects meta-analysis to pool data from the studies due to the potential for inherent biases in the studies. We will perform separate pooled analyses for RCTs, non-RCTs and comparative observational studies, where possible.

For non-RCT data, we will extract tabulated data, crude estimates or adjusted estimates of effect from the studies. We will use adjusted estimates of effect in preference to minimise potential confounding between the treatment groups. We will use similar meta-analysis methods to pool data from non-RCTs as described for the RCTs.

Subgroup analysis and investigation of heterogeneity

We will carry out a subgroup analysis in the following areas.

  1. Daily corticosteroid dose (low versus high; low dose defined as hydrocortisone ≤ 300 mg, dexamethasone ≤ 12 mg, prednisolone ≤ 75 mg, methylprednisolone ≤ 60 mg) (Annane 2004).

  2. Timing of corticosteroid use (early versus late; early defined as < four days of onset of symptoms and late ≥ four days).

  3. Duration of corticosteroid course (short versus long course, short course defined as < five days and long course ≥ five days) (Annane 2004).

  4. Adult versus child population (adult defined as ≥ 16 years).

  5. Route of administration (intravenous, oral).

  6. Seasonal influenza versus pandemic/outbreak influenza.

Sensitivity analysis

Where there are sufficient studies, we will perform sensitivity analyses to assess the effect of study design on primary and secondary outcomes using stratification.


We wish to thank the following people for commenting on the draft protocol: Vinod Singh, Noorin Bhimani, Harri Hemilä, Rashmi Das, Sree Nair and Lubna Al-Ansary.

Contributions of authors

Chamira Rodrigo (CR) was involved in writing the protocol with supervision from Wei Shen Lim (WSL).
Jonathan Nguyen-Van-Tam (JVT) provided expertise on the background section of the protocol.
Jo Leonardi-Bee (JLB) provided expertise on the methods section.

Declarations of interest

Chamira Rodrigo - none known.

Jonathan Nguyen-Van-Tam: The University of Nottingham Health Protection Research Group is currently in receipt of research funds from GSK. The group has recently accepted an unrestricted educational grant for influenza research from F. Hoffmann-La Roche. Research on influenza funded by an unrestricted educational grant from Astra Zeneca is also underway. The aforementioned funding received from GSK, F. Hoffmann-La Roche and Astra Zeneca did not support any aspect of this work. JSN-V-T has received funding to attend influenza related meetings, lecture and consultancy fees and research funding from several influenza antiviral drug and vaccine manufacturers. All forms of personal remuneration ceased in September 2010, but departmental funding for influenza-related research from GlaxoSmithKline, F. Hoffmann-La Roche and Astra-Zeneca remains current. He is a former employee of SmithKline Beecham plc. (now GlaxoSmithKline), Roche Products Ltd, and Aventis-Pasteur MSD (now Sanofi-Pasteur MSD), all prior to 2005, with no outstanding pecuniary interests by way of shareholdings, share options or accrued pension rights.

Jo Leonardi-Bee is a co-applicant of an Educational Grant from Roche to carry out further research in the area of pandemic influenza. Dr. Leonardi-Bee will be using this to carry out a systematic review and individual patient meta-analysis of the evidence (published and unpublished) of the impact of antiviral use on public health outcomes for 2009 pandemic influenza A/H1N1. This systematic review has been registered with PROSPERO (International prospective register of systematic reviews).

Wei Shen Lim has received an unrestricted research grant from Pfizer and an unrestricted donation of Binax test kits from Alere in support of a study in pneumococcal pneumonia which is unrelated to the submitted work.