Topical antibiotics for chronically discharging ears with underlying eardrum perforations
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To assess the effects of topical antibiotics for chronically discharging ears with underlying eardrum perforation.
Background
Chronically discharging ears associated with underlying persistent eardrum perforations (chronic suppurative otitis media, CSOM) is one of the commonest causes of preventable hearing impairment in low and middle‐income countries (WHO 1998). It can follow untreated or inadequately treated acute otitis media, usually in the first five years of life (although often persists into adulthood), and is related to poor socio‐economic conditions.
What is CSOM?
Ear infections can involve the middle or outer ear (otitis media or otitis externa). CSOM is one of several types of otitis media and will be the focus of this review. The World Health Organization defines CSOM as "a stage of ear disease in which there is chronic infection of the middle ear cleft, a non‐intact tympanic membrane (i.e. perforated eardrum) and discharge (otorrhoea), for at least the preceding two weeks" (WHO 1998), although this could more strictly be considered a childhood definition. Perforations and infection can be in one ear (unilateral) or both (bilateral). A variety of underlying pathologies can cause CSOM, including: an acute episode of acute otitis media that has burst the ear drum and not settled within two weeks; a recurrent episode of acute otitis media in an ear with a perforation from a previous episode of acute otitis media; or an ear with a persistent perforation with active chronic otitis media with metaplastic changes to the mucosa of the middle ear and mastoid air cell system (Browning 2003, personal correspondence). In adults, the majority of patients are likely to have CSOM with a perforation that will not spontaneously heal.
For this review, we are considering all studies that included cases fitting under the broad WHO definition, irrespective of the underlying pathology, and for both adults and children.
What are the effects of CSOM?
Hearing impairment, aside from the discomfort from recurrent ear discharge, is the most frequent effect of CSOM. A school survey in Kenya found 63% of ears with CSOM had more than 30 decibels hearing loss, compared to only 3.4% of ears without outer or middle ear pathology (Hatcher 1995). Hearing impairment due to otorrhoea and a perforated eardrum will usually improve as the disease resolves. However, untreated CSOM may result in permanent hearing loss due to damage to the ossicles, which transmit sound vibrations from the eardrum to the cochlea. Because otitis media occurs mostly in children during pre‐school years, which is during the most dynamic phase of speech and language development, there is concern that the associated hearing deficits may result in speech and language delays or permanent learning disabilities, as well as disturbances in behaviour (Klein 2001).
In addition to hearing impairment (with its associated consequences), complications of otitis media can often result in death or severe disability, especially in low‐income countries. The infection may extend and spread to the head and neck structures and to the brain. Intracranial infections include meningitis, abscesses, hydrocephalus, or thrombosis of the lateral venous sinus (from suppuration within the mastoid causing clots occluding the lumen of the vessel) (Ludman 1997). Alternatively complications may be extracranial, such as facial paralysis, cholesteatoma (a destructive formation of layers of keratinising epithelium, accumulating in the middle ear and mastoid (Bluestone 1996), also described as 'active squamous (epithelial) chronic otitis media (Browning 1997)), labyrinthitis (extension to the labyrinth through the round window), or acute mastoiditis (spread of the infection to the mastoid air cells), which may spread further due to necrosis of the bony wall of the cells resulting in further life‐threatening complications (Ludman 1997, Dhillon 1999).
How much of a burden is CSOM?
Around 91% of the burden of otitis media (all types) and nearly all related deaths occur in low and middle‐income countries (World Bank 1993, WHO 2002). One review of school and community‐based studies reported a prevalence between 0.4 and 6.1 per cent in these countries (Berman 1995). Data from the World Health Organization and World Bank suggest the global burden of otitis media has dropped dramatically over the past 10 years, to approximately 6000 deaths (0.01% of all deaths) and 1,474,000 disability adjusted life years (DALYs) lost (0.100% of all DALYs) worldwide (WHO 2002). Most of these deaths are likely to be due to chronic otitis media and its complications, because acute otitis media is usually a self‐limiting infection (Acuin 2003b).
Although most of the background literature in this review relates to children, reliable and generalisable data for the global burden in children are not readily available; the WHO estimates therefore quoted here are for both adults and children.
What are the causes of CSOM?
The causes and risk factors associated with CSOM are unclear, and few studies have examined these for CSOM. Instead authors have extrapolated results of studies for acute otitis media and otitis media with effusion to CSOM. However these studies often have conflicting findings, and there is no confirmed association between CSOM and the various host and environmental factors, which may differ to acute otitis media and otitis media with effusion. Despite this, some important factors that may be associated with CSOM include inadequate treatment (of CSOM and acute otitis media, for example with antibiotics), poor access to medical care, poor socioeconomic conditions, season, exposure to tobacco smoke, overcrowding, attendance at day care centres, lack of breastfeeding, poor nutrition or hygiene, altered immunity and underlying diseases (e.g. HIV/AIDS (Barnett 1992, Singh 2003), frequent upper respiratory tract infections, early onset of otitis media in the first months of life, and family history of otitis media. Some populations are at increased risk of developing CSOM, and have high rates reported, including certain ethnic groups (such as Native American tribes of Apache and Navajo, Australian Aborigines, and Inuits of Canada, Greenland and Alaska), and individuals with anatomical defects (e.g. cleft palate or submucous cleft), or Down's syndrome (Bluestone 1998).
What management approaches are there?
The aims of treatment are to stop the discharge (and to eradicate infection), to heal the tympanic membrane, improve hearing, prevent the common problems of recurrent or new infections, and to prevent potentially life‐threatening complications.
Strategies to prevent and treat otitis media are available, and may include pneumococcal vaccines to prevent acute otitis media (see Straetemans 2003).
There are a number of treatment options for CSOM, each of which will be considered in a separate Cochrane review in this series:
-
cleaning of the ear discharge by 'ear toilet' (dry mopping, ear wicking, gentle syringing, or suctioning)
-
systemic treatment (e.g. oral antibiotic preparations, or intravenous antibiotics)
-
topical treatment with antiseptics
-
topical treatment with antibiotics (THIS REVIEW)
-
systemic or topical steroids, with or without antibiotics (e.g. neomycin and polymyxin b) (antibiotic steroid combinations included in antibiotic Cochrane reviews above)
-
surgical treatment (especially for people who have CSOM with cholesteatoma or life‐threatening infections), involving, for example, surgical cleaning of the mastoid and middle ear, or repair of the tympanic membrane (by tympanoplasty)
A report of a WHO/CIBA Foundation Workshop held in 1996, recommends administration of topical (and/or systemic) antibiotics as well as dry mopping/wicking, since wicking alone is suggested to be ineffective (as found by Smith 1996) (WHO 1998). However, the WHO guidelines still currently recommend treating CSOM by using wicking to dry the ear alone (and a five‐day follow up).
A number of topical antibiotics have been used in the treatment of CSOM. However, concern exists regarding their ability to penetrate the middle ear and mastoid cavities as well as their activity against the causative bacteria (usually gram‐negative). There also remains controversy and uncertainty about the possible ototoxic effect, in particular of topical aminoglycoside antibiotics (by damaging the hair cells in the basal turn of the cochlea), particularly where the eardrum is not intact. The newer 4‐quinolone antibiotics (e.g. ciprofloxacin) are widely considered to be the 'gold standard' topical antibiotics, but are expensive and not generally available as ototopical preparations in many countries. Topical antibiotics may be superior to topical antiseptics, although this needs investigating, particularly as topical antiseptics are cheap and easily available, and are thus widely used in many low and middle‐income countries. We will compare topical antibiotics and topical antiseptics in this review; this comparison will not be investigated in the antiseptics review. We will also assess the effect of using of steroids in combination with topical antibiotics, which is common practice in many countries.
Objectives
To assess the effects of topical antibiotics for chronically discharging ears with underlying eardrum perforation.
Methods
Criteria for considering studies for this review
Types of studies
Individual randomised and quasi‐randomised controlled trials.
Cluster randomised controlled trials.
Types of participants
People of any age with a diagnosis of CSOM meeting the WHO definition.
Types of interventions
Intervention: topical (aural) antibiotics (all and individual)
Comparator: no intervention; placebo; other topical antibiotics with and without steroids; systemic antibiotics (all and individual); combination of topical and systemic antibiotics; antiseptics.
Types of outcome measures
Primary
-
Resolution of CSOM at 2 to 4 weeks and after 4 weeks, according to the following findings:
a) No report of otorrhoea
b) Disappearance of discharge on otoscopy
c) Healing of the eardrum perforation on otoscopy
d) Time to resolution of CSOM according to any other definition of resolution made by the authors, including improvement in mucosal appearance
Secondary
-
Time to resolution of CSOM as defined above
-
Improvement in hearing threshold, as measured by audiometry at 2 to 4 weeks, and after 4 weeks
-
Time to re‐appearance of discharge and perforation after its previous resolution
-
Adverse Events that
a) are fatal, life‐threatening, require inpatient hospitalisation or prolongation of existing hospitalisation, or result in persistent or significant disability/incapacity, such as permanent hearing loss, tinnitus or vertigo (Karbwang 1999; UMC 2003 (the Uppsala Monitoring Centre, http://www.who‐umc.org/))
b) result in withdrawal or discontinuation of treatment
c) any other adverse events, such as ear pain, ear canal reactions and transient dizziness
Where outcomes (resolution of discharge, healing of the tympanic membrane, and hearing threshold) are reported at several time‐points within the ranges above, we will take the last reported result.
Search methods for identification of studies
An independent search will be carried out by the Trials Search Co‐ordinator of the Cochrane ENT Group. Both reviewers will select trials independently (at least for electronic searches), and we will contact authors for clarifications from them whenever needed.
We will attempt to identify all relevant studies regardless of language or publication status (published, unpublished, in press, and in progress).
We will use the following search terms: otitis media, otitis media suppurative, suppuration, suppurat*, purulen*, pus, chronic disease(s), chronic*, persist*, chronic* otorrh*, persist otorrh*, chronic* discharg*, persist* discharge, csom, mastoiditis, perforation, antibiotics, antibiotic*, antibacterial*.
We will search the Cochrane ENT Disorders Group trials register (code SR‐ENT), and the Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library (Issue 3, 2003), for relevant trials up to August 2003. Full details of the Cochrane ENT Disorders Group methods and the journals handsearched are published in The Cochrane Library in the section on Collaborative Review Groups.
We will also search the following electronic databases using the search terms provided above, in combination with the search strategy for identifying trials developed by The Cochrane Collaboration (Clarke 2003).
(1) MEDLINE (1966 to August 2003)
(2) EMBASE (1974 to August 2003)
(3) LILACS (www.bireme.br; 1982 to August 2003)
(4) AMED
(5) CINAHL (1982 to August 2003)
(6) OLDMEDLINE (1958‐1965)
(7) PREMEDLINE
We will search the following potential sources of trials:
-
Metadatabase of registers of ongoing trials, mRCT, accessible via the Internet: http://controlled‐trials.com/mrct/
-
ENT Disorders Group trials register for any relevant abstracts from conference proceedings
-
Previously published (systematic) reviews, including, but not limited to: 'Interventions for chronic suppurative otitis media' (Acuin 2003a), in The Cochrane Library, Issue 2, 2003; 'Chronic suppurative otitis media', in Clinical Evidence, May 2003 (Acuin 2003b) and 'Systematic Review of Existing Evidence and Primary Care Guidelines on the Management of Otitis Media (Middle Ear Infection) in Aboriginal and Torres Strait Islander Populations, March 2001 (Couzos 2001)
-
Contacting of organisations and individual researchers working in the field of otitis media (including authors of published trials and other experts who may know about additional trials)
-
Contacting pharmaceutical companies (see published notes for list of companies contacted) to locate additional studies, unpublished data, confidential reports, and raw data of published trials.
-
Reference lists of all articles/trials identified by the above methods (includes searching of bibliographies for relevant citations)
Data collection and analysis
Eligibility Assessment
We will independently review the titles and abstracts of articles for a preliminary assessment of the potential relevance of those identified during the search.
We will retrieve the full papers for all potentially relevant studies. We will independently assess their eligibility to be included in the review using an eligibility form based on the stated inclusion criteria. We will scrutinise each trial report for multiple publications from the same data set. Where outcomes are not reported, we will contact the author of the paper for this information, as the outcome(s) may have been collected but not reported. We will exclude studies that do not meet the inclusion criteria and state the reason in the 'Characteristics of excluded studies'. We will resolve disagreements through discussion, or by consulting the Co‐ordinating editor of the Cochrane Ear, Nose and Throat Disorders Group (Martin Burton, MB). Where necessary, we will contact the study authors for clarification.
Assessment of methodological quality
We will independently assess the methodological quality of all the trials identified as eligible for inclusion. We will report the level of agreement between the two reviewers. We will resolve disagreements through discussion, or by contacting another person at the Cochrane Ear, Nose and Throat Disorders Group editorial base where necessary. Where necessary, we will contact the study authors for clarification. We will display the assessment of methodological quality in an additional table and describe it in the section 'Methodological quality of included studies'.
We will assess the methodological quality of the trials in terms of allocation concealment, generation of allocation sequence, blinding and inclusion of randomised participants. We will classify allocation concealment, generation of allocation sequence, and inclusion of randomised participants as adequate, inadequate and unclear as outlined by Juni 2001. Blinding will be classified as double blind, single blind, or open. More details are outlined in the published notes.
Data collection
We will independently extract data of study characteristics, including methods, participants, interventions, and outcomes, and record these on standard forms. In studies where data are insufficient or missing, we will contact the authors of the original studies for additional data and/or verification of methods, to clarify any uncertainties about the data and the way in which they were collected, and to try to obtain missing data. When trialists provide unpublished data or conduct additional analyses, Carolyn Macfadyen will send them a draft copy of the review for comment, and negotiate permission to use unpublished data.
Where outcomes are reported in number of ears only, we will also attempt to obtain the values for number of participants. We will check the data and resolve any discrepancies by referring to the trial report, through discussion, or involving a third party at the editorial base.
Where possible we will extract data to allow an intention‐to‐treat analysis (i.e. the analysis should include all the participants in the groups to which they were originally randomly assigned). If the number randomised and the numbers analysed are inconsistent, we will calculate a percent loss‐to‐follow‐up and report this information in an additional table. For binary outcomes, we will record total number of participants and number with the event in each group of the trial. For continuous outcomes, for each group, we will extract the number of participants, and the arithmetic means and standard deviations. If the data are reporting using geometric means, we will extract standard deviations on the log scale. We will extract medians and ranges and report them in tables.
Data Analysis
We will follow the statistical guidelines in the Cochrane ENT Group Guidelines for Reviewers (CochraneENTGuideline, updated November 2000). In accordance with the ENT Group requirements, two reviewers should enter data independently into Review Manager 4.2, which we will use to analyse the data.
We will pool data where appropriate. In studies that have enrolled people with otitis externa, draining surgical cavities or acute otitis media, as well as CSOM, we will only include the results for CSOM participants where available.
For binary data, we will calculate a weighted treatment effect across all trials, expressed as an adjusted relative risk and its 95% confidence intervals. For continuous data, we will calculate the weighted mean difference and its 95% confidence intervals where the outcome is measured on the same scale. Where data have been reported using medians and ranges, or there is evidence of skewed data, we will report medians and ranges (dividing mean by the standard deviation; results of <1.64 indicate a positive skew, as does a standard deviation greater than the mean). If continuous data have been reported using geometric means, we will combine the findings on a log scale and report on the original scale.
Where significant heterogeneity is present, and it is appropriate to pool data, we will use the DerSimonian & Laird random effects model.
The primary analysis will be of all eligible studies. If heterogeneity is found, we will explore whether this is explained using subgroup analyses or meta‐regression. We will explore age (under 16, and adults 16 years or older), associated mopping (dividing studies into those with some form of ear toilet, and those without), co‐interventions (dividing studies into those with treatment comparison alone, and those with treatment comparison in combination with other co‐interventions), and methodological quality (initially excluding studies of the poorest quality to see if the homogeneity improves).
We will conduct a secondary sensitivity analysis with all studies that meet the inclusion criteria, including quasi‐randomised designs. We will perform sensitivity analysis on the trials, particularly according to methodological quality (notably adequate concealment), and also by trial design (e.g. cluster randomisation), to test for heterogeneity in the results and to investigate the cause of any skewed data. We will display the results for each sensitivity analysis according to the subgroups within each methods category.
The sensitivity analysis will include the following, as outlined in the statistical guidelines in the Cochrane ENT Group Guidelines for Reviewers (CochraneENTGuideline updated November 2000).
(1) Repeat the analysis excluding unpublished studies (if any).
(2) Repeat the analysis excluding studies of the lowest quality (already done if there is heterogeneity).
(3) If there are one or more very large studies, we will repeat the analysis excluding these, to investigate how much they dominate the results.
(4) Repeat the analysis excluding studies where people with CSOM are only a subgroup of the participants included in the study, for example, those that enrol people with otitis externa, draining surgical cavities or acute otitis media, as well as CSOM.
We will investigate publication bias using funnel plots. We will also visually examine forest plots, in conjunction with the chi‐squared test, using a 5% level of statistical significance, to investigate whether the asymmetry may be caused by other factors such as heterogeneity and quality. We will examine for heterogeneity in all meta‐analyses and comment on it. Where present, we will not generally quote the overall effect estimate, but we will attempt to investigate it. If this is not possible, we will quote the random effects (DerSimonian & Laird) and the fixed effects (Mantel‐Haenszel) models.
