Journal of the New Zealand Medical Association, 17-October-2008, Vol 121 No 1284
Are antibiotics indicated as an initial treatment for patients with acute upper respiratory tract infections? A review
Bruce Arroll, Tim Kenealy, Karen Falloon
Upper respiratory tract infections (URTIs, syn common cold) are the most common acute illness worldwide;1,2 it is estimated that adults have 2 to 3 colds per year. They also represent one of the most common causes of antimicrobial use and a frequent reason for prescribing antibiotics in ambulatory practice and primary care.2,3
In this setting, antibiotics are considered to be overprescribed,4 broad-spectrum varieties tend to be overused,5 and primary care physicians (in some jurisdictions) treat over half of these colds with antibiotics.6
Despite this frequent antibiotic usage, the vast majority of URTIs have a viral aetiology. The most common virus is rhinovirus when symptoms are mild and influenza or parainfluenza viruses when symptoms are more severe, particularly when accompanied by muscle aches and fatigue.7 Most patients tend to recover spontaneously in 1–2 weeks without the need for antibiotics. However, there is a small risk of serious complications (e.g. otitis media) with all the upper respiratory tract infections and this in part influences clinicians in their desire to prescribe antibiotics.8
Patients’ perceived expectations and other non-clinical factors can also influence the decision to prescribe antibiotics.9
Due to the fact that URTIs are such a prevalent condition worldwide, it is important to obtain an estimate of the effectiveness of antibiotics. If they are ineffective there is a concern that the widespread use of antibiotics is not only a poor utilisation of health resources but also that they cause adverse effects and hasten development of resistant strains of bacteria.10,11
A recent study out of Europe has warned that with current prescribing practices we risk losing the miracle drugs of the 20th Century.12 The study investigated outpatient antibiotic use in 26 countries in Europe and demonstrated correlations which showed that resistance to microorganisms (S. pneumoniae, S. pyogenes, and E. coli) increased with consumption of antibiotics. It also noted that in most countries there was a growing use of the newer (i.e. broad-spectrum) antibiotics, such as combination amoxicillin and clavulanic acid, the new macrolides and quinolones.
A subsequent study looked into this relation further by investigating two macrolides—azithromycin and clarithromycin—which are among the medications commonly used in respiratory tract infections.13 Their results showed that (inappropriate) macrolide use is the single most important driver of the emergence of macrolide resistance in vivo.
This systematic review aims ascertain the effectiveness of antibiotics as the initial treatment for patients with upper respiratory tract infections. In order to assess the benefits and harm of antibiotics in upper respiratory tract infections we systematically examined the antibiotic versus placebo literature and conducted a meta-analysis in terms of the outcomes of general improvement and adverse effects.
A Medline search covering the years 1966 to January 2007 was undertaken using the following MeSH terms: common cold, respiratory tract infections (upper), random allocation, double blind method, or single blind method.
A similar search strategy was used on EMBASE and a search of the Cochrane Controlled Trials Register. The Family Medicine Database was searched through the Canadian College of Family Physicians Library in London Ontario including a search on FAMLI Vol. 1, 1980 to Vol. 13, 1993 (this database was discontinued in 1993). We also searched the reference lists of relevant trials, review articles, and textbook chapters. Authors of the studies finally retrieved by the above method were approached in writing to enquire about unpublished or unretrieved articles.
Other investigators have approached the British Pharmaceutical industry for papers on respiratory tract infections but received no unpublished material (T Fahey, personal communication).
The International Classification of Health Problems in Primary Care (ICHPPC-2) defines a URTI as an illness with evidence of acute inflammation of nasal or pharyngeal mucosa and the absence of other specifically defined respiratory infections e.g. streptococcal tonsillitis, laryngitis, bronchitis, pneumonia, asthma, and hayfever.14
Lower respiratory tract signs were accepted in patients with the above symptoms so long as the majority of patients in the study did not have these signs and that pneumonia was ruled out. The reviewers recognised that there would be some undetermined overlap with the alternative diagnoses of bronchitis and pharyngitis. However, this was not necessarily a disadvantage as it reflected the real life situation faced by practitioners making decisions on the use of antibiotics.
All randomised controlled trials of antibiotic versus placebo as initial treatment for acute upper respiratory tract infection.
Studies were excluded:
We used Cochrane review manager v4.2 software for the analysis. A fixed effects method was used for all analyses except for side-effects. Here a random effects model was used as a Chi squared test indicated heterogeneity of the findings.16
Eight trials were eligible for inclusion in our review.17–24 All eight studies were double blind evaluations comparing antibiotic with placebo.
The proportion of drop outs ranged from 0% in the study by Hoagland et al (1950) to 6.5% in the Kaiser et al study (1996).
Six of the studies provided a global measure of improvement form day 1 to day 7 from onset of treatment. These are summarised in Figure 1, which shows a non-significant benefit from antibiotics compared with placebo, with a relative risk of 0.89 (95%CI 0.77–1.04).
Two studies did not contribute to this analysis.17,19 The study by Gordon et al (1974) expressed their results as p values with placebo being better at relief of symptoms than ampicillin (p=0.05) and no significant difference for placebo versus erythromycin nor placebo versus penicillin. The study by Howie and Clark (1970) found no significant benefit for a number of outcomes including illnesses with no purulent spit, with purulent spit of 10 days or more, cough, spit and purulent nasal discharge.19
Only two of the studies had outcomes as measured by the effects of antibiotics on specific symptoms.19,22 Unfortunately the denominator for one of these was based on episodes of illness rather than individual patients which did not allow the data to be added to the other studies.19 This study by Howie and Clark did analyse side effects by individual and hence this data could be used in the combined analysis.
Four studies reported adverse effects and the summary relative risk was significantly increased 2.71 (95%CI 1.08–6.83) (Figure 2); a random effects model was used due to heterogeneity. The adverse effects in the adult patient studies was 4.06 (95%CI 2.34–7.04) with no significant heterogeneity using the fixed effects model. It would appear that mixing the child and adult results created the heterogeneity.
One study in this review found a significant benefit for antibiotics in the subset of patients (20% of whole group) who had positive nasopharyngeal aspirates for three respiratory pathogens H. influenzae, M. catarrhalis, and S. pneumoniae; p=0.001.20 Another study that found an overall benefit from antibiotics may have had a large number of patients with streptococcal tonsillitis but the presentation of the findings made this difficult to determine.24
The results of this meta-analysis show that there is no benefit in terms of overall improvement in prescribing antibiotics as an initial treatment for patients with upper respiratory tract infections. The number of well conducted studies is small for such a common condition. The suggestion that there may be some benefit in terms of resolution of nasal discharge has been reviewed elsewhere.25
Antibiotics were associated with more than twice the risk of side effects than were placebos. This was similar to the side effects in the meta-analysis of antibiotics versus placebo for children with otitis media.26 There are some possible explanations for the heterogeneity in the side effects data. The two studies that were statistically significant were conducted in adults.19,20 The duration of treatment was identical while there were some differences in the antibiotics.
The Howie and Clark (1970) paper reported on demethylchlortetracycline, the Kaiser et al (1996) paper reported on amoxycillin with clavulanic acid, while the Taylor et al paper (1977) in children22 reported on cotrimoxazole and amoxycillin. The side effect rate for the cotrimoxazole was 24% which was higher than the 15% for amoxycillin while the placebo rate was 19%. The fact that the side effect rate for amoxycillin was lower than that for placebo may explain the heterogeneity.
There is no gold standard definition of upper respiratory tract infection and the diagnosis is made on clinical grounds. Although the range of inclusion criteria appear wide, we believe that the majority of patients were suffering from viral upper respiratory tract infections. In effect, we accept the clinical judgement of the doctors examining the patients from the various studies. Only two studies reported the presence of lower respiratory tract findings.
One of these studies found lower respiratory tract signs in 13% of under-2 year olds and in 13% of those over 6 years of age.17 In the other study, 43% of patients had auscultatory evidence of more extensive peripheral airways disease.22 We assumed that most of these signs were due to bronchospasm.
Two of the studies included in this meta-analysis may underestimate the benefit of antibiotics. One study reported findings at 24 hours which would be regarded as being too soon to show any effect for antibiotics treatment but the results have been included here for the sake of being systematic.18 The other study used 15 mg of tetracycline or equivalent given three times daily which would be considered sub-therapeutic today.23 However both studies were included as they met the pre-stated inclusion criteria.
We excluded studies with more than 7% of patients with streptococci found on throat swab but kept the paper by Gordon et al (1974) which had 2 of 89 patients with cultures of beta haemolytic streptococci.17 Other studies did not test for throat bacteria. The issue of bacterial involvement is a concern for reviews of bronchitis, sore throats and upper respiratory tract infections. There is only one overlap between our study and the bronchitis Cochrane review by Fahey et al (2006)1 and that is the study by Howie et al 1970.2
The sore throat review by Del Mar (1992) included the Taylor et al (1977) study that is in our review.3,4 The role of bacteria in upper respiratory tract infections either as a causal factor or as a complication is highlighted in the study by Kaiser et al (1996). They make a case for the role of three bacteria, H. influenzae, M. catarrhalis, and S. pneumoniae, and it would be helpful to see this work repeated in another centre.
Only two of the studies were conducted in general practice.19,22 The other studies were at a military base18 used research clinic staff,23 at a casualty clinic17,20 and hospital outpatients.21 If the assumption is made that patients can self refer to these secondary care settings and use them as primary care providers then our results are generalisable to the wider primary care setting.
Antibiotics appear to have no benefit in the initial treatment of acute upper respiratory tract infections. The implications for practice are that prescriptions of antibiotics should not be given as an initial treatment as they will not affect overall improvement and carry the risk of adverse effects.
Physicians prescribing antibiotics are encouraged to consider the appropriateness of their use in order to minimise the potential ecological side effect of antibiotic-resistant organisms. Further research is needed on the role of pathogenic bacteria from nasopharyngeal aspirates in upper respiratory tract infections.
Note: A electronic version of this paper is available at the Cochrane website www.cochrane.org This will be updated over time and represent the most recent evidence on this topic.
Competing interests: BA is a PHARMAC Seminar Committee member and has received support from Astra Zeneca (UK) and Sanofi-Aventis (NZ).
Author information: Bruce Arroll, Professor and Head of Department; Timothy Kenealy, Associate Professor; Karen Falloon, Academic Registrar; Department of General Practice and Primary Health Care, University of Auckland, Auckland
Acknowledgements: The Charitable trust of the Auckland Faculty of the Royal New Zealand College of General Practitioners for a grant to fund the literature search.
Correspondence: Bruce Arroll, Dept of General Practice and Primary Health Care University of Auckland, Private Bag 92019, Auckland, New Zealand. Fax: +64 (0)9 3737624; email firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals