Introduction
In the UK, aminoglycosides are considered as a first-line therapy for chronic suppurative otitis media (CSOM) patients with or without tympanostomy tubes.Reference Philips, Yung, Burton and Swan1 Chronic suppurative otitis media has an estimated global incidence of 0.47 per cent, and in the UK approximately 1.5 per cent of adults are estimated to have active CSOM.Reference Monasta, Ronfani, Marchetti, Montico, Brumatti and Bavcar2, Reference Ghosh, Panarese, Parker and Bull3 Serious complications occur if the infection is not effectively controlled, including hearing loss, intracranial spread of infection and venous sinus thrombosis.Reference Yorgancılar, Yıldırım, Gun, Bakır, Tekın and Gocmez4, Reference Dubey, Larawin and Molumi5 Topical antimicrobial therapy is the mainstay of treatment for CSOM; Cochrane reviews have shown that topical antibiotics clear discharge better than antiseptics or placebo, and no benefit is derived from systemic antibiotics.Reference Macfadyen, Acuin and Gamble6, Reference Macfadyen, Acuin and Gamble7
Ototopical aminoglycosides are also administered following tympanostomy tube placement. In the UK, approximately 30 000 patients undergo this procedure annually, and a significant proportion of these receive a statutory dose of aminoglycoside (in ear drops) during the procedure.Reference Sood and Waddell8 A Cochrane review found that topical antibiotics are effective in reducing post-tympanostomy otorrhoea, but it is unclear which agent should be used.Reference Syed, Suller, Browning and Akeroyd9 The Scottish Intercollegiate Guidelines Network recommend a single dose of topical antibiotic at the time of tympanostomy tube insertion; post-operatively, they advocate using the same antibiotic used to treat the infected ear, but do not comment on the merits of specific antimicrobials.10
Ototopical aminoglycosides
The ototoxic effects of systemic aminoglycosides are well documented and accepted, and these effects are also seen following ototopical aminoglycoside administration.Reference Helal11 Aminoglycosides applied to the middle ear diffuse across the round window membrane and can significantly concentrate in the inner ear.Reference Becvarovski, Bojrab, Michaelides, Kartush, Zappia and LaRouere12 A large amount of animal data are available showing ototoxicity from aminoglycosides applied directly to the middle ear.Reference Roland, Rybak, Hannley, Matz, Stewart and Manolidis13 Retrospective studies in humans have shown ototoxicity (mainly vestibular) resulting from topical gentamicin, and there are many published cases of cochlear and vestibular toxicity secondary to topical neomycin.Reference Bath, Walsh, Bance and Rutka14–Reference Matz, Rybak, Roland, Hannley, Friedman and Manolidis16 Concurrent use of many common medications (such as loop diuretics) may also augment the ototoxicity of aminoglycosides.
There are medicolegal implications in prescribing an ototoxic medication for use in the middle ear. Matthew et al. reviewed 15 years of clinical negligence claims in otology, through the National Health Service Litigation Authority, and found that 3.1 per cent of claims leading to damages were for ototoxicity following topical aminoglycoside treatment.Reference Matthew, Asimacopoulos and Valentine17 The Medical Protection Society advise caution in prescribing aminoglycosides in the presence of a grommet or perforation; they published a case of a patient suffering hearing loss following ototopical aminoglycoside, which was settled out of court.18
Ototopical quinolones
Topical quinolones have been used to treat ear infections since the 1980s and have minimal adverse effects.Reference Wai and Tong19 A Cochrane review has shown that topical quinolones are effective at treating CSOM.Reference Macfadyen, Acuin and Gamble6 Trials have also shown quinolones to be effective in cases of otorrhoea with tympanostomy tubes.Reference Heslop, Lildholdt, Gammelgaard and Ovesen20, Reference Dohar, Giles, Roland, Bikhazi, Carroll and Moe21
Unlike aminoglycosides, intratympanic ciprofloxacin does not cross the round window and is not found in the inner ear following application.Reference Becvarovski, Bojrab, Michaelides, Kartush, Zappia and LaRouere12 Macfadyen et al. did not report any ototoxicity in any of the trials reviewed.Reference Macfadyen, Acuin and Gamble6 Animal studies showed no change in hearing after ciprofloxacin was administered directly to the middle ear.Reference Gates22–Reference Daniel and Munguia25 A systematic review concluded that ototopical ciprofloxacin was safe when used in the presence of a non-intact tympanic membrane in adults and children.Reference Wall, Stroman, Roland and Dohar26 An Ovid and Pubmed database search using the keywords ‘safety’, ‘toxicity’, ‘topical quinolones’ and ‘topical ciprofloxacin’ did not reveal any articles that showed ototoxicity or other significant side effects following ototopical quinolone use.
Plasma levels of ciprofloxacin are not measurable after the administration of ototopical quinolones in patients with or without an intact tympanic membrane.Reference Clarós, Sabater, Clarós and Clarós27–Reference Force, Hart, Plummer, Powell and Nahata29
Antimicrobial resistance
Concerns are frequently raised of increasing bacterial (particularly Pseudomonas aeruginosa) resistance to quinolones if they are more commonly prescribed. Whilst ciprofloxacin-resistant pseudomonas is increasingly isolated, there is no convincing evidence that ototopical quinolones cause this.Reference Weber, Roland, Hannley, Friedman, Manolidis and Matz30 Resistance is more likely to be due to systemic therapy, with the highest rates of quinolone resistance isolated in urine samples and sputum samples from patients with cystic fibrosis.Reference Blondeau, Suter, Borsos and Misfeldt31 Antibiotic resistance is commonly determined above a minimum inhibitory concentration; however, these values are of limited usefulness in guiding topical therapy, as concentrations many times higher are achieved compared to systemic therapy.Reference Weber, Roland, Hannley, Friedman, Manolidis and Matz30 This may explain why reported antimicrobial resistance and clinical success are often mismatched in studies utilising topical therapy. No studies have shown that ototopical ciprofloxacin is systemically absorbed.Reference Weber, Roland, Hannley, Friedman, Manolidis and Matz30
National and international guidelines
In 2004, the American Academy of Otolaryngology – Head and Neck Surgery published a consensus panel review advising that quinolones be considered as a first-line treatment in the presence of a tympanic membrane perforation or ventilation tube.Reference Roland, Stewart, Hannley, Friedman, Manolidis and Matz32 In 2005, a Canadian guideline stated that ‘gentamicin containing ear drops should not be used in patients with a non-intact ear drum’.Reference Rosser, Pennie and Pilla33 This was followed, in 2007, by similar position papers from the Australian and the New Zealand societies of otolaryngology head and neck surgery.Reference Black, Cousins, Chapman, Becvarovski, Coates and O'Leary34, Reference Gilbert, Dawes, Mahadevan, Baber and Hall35 The New Zealand guidance stipulates that if potentially ototoxic drops are used then the patient should be informed of the risk of ototoxicity, which is between 1:1000 and 1:10 000.Reference Gilbert, Dawes, Mahadevan, Baber and Hall35
In 2007, ENT-UK issued a consensus statement which urged caution in using topical aminoglycosides in the presence of a perforated tympanic membrane.Reference Philips, Yung, Burton and Swan1 However, ENT-UK stated that they could not recommend the use of quinolones as they are currently unlicensed in the UK. Of the 39 consultant otologists who formed the consensus panel, only 36 per cent supported the statement ‘unlicensed topical quinolones should be used instead of aminoglycosides’.Reference Philips, Yung, Burton and Swan1
Aims
This study aimed to review trials comparing the use of topical quinolones and aminoglycosides in patients with a non-intact tympanic membrane. The primary aim was to establish whether there is evidence of superiority in treating otorrhoea. The trials were also examined for reported antibiotic resistance, side effects or significant events following administration.
Materials and methods
Inclusion criteria
The review included randomised, controlled trials on human subjects, published in English, that compared topical aminoglycosides with topical quinolones for the treatment of CSOM or for post-tympanostomy prophylaxis.
Method
The review was conducted and reported with reference to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) statement.Reference Moher, Liberati, Tetzlaff and Altman36
A search of the PubMed, Ovid (including the Allied and Complementary Medicine Database, the Cumulative Index of Nursing and Allied Health Literature, Embase, and Medline) and Cochrane databases was carried out. Bibliographies of review articles were examined. The search terms used were ‘chronic suppurative otitis media’, ‘chronic otitis media’, ‘chronic discharging ear’, ‘grommets’, ‘ventilation tubes’ and ‘tympanostomy tubes’, together with ‘aminoglycosides’, ‘quinolones’, ‘fluoroquinolones’, ‘gentamicin’, ‘neomycin’, ‘framycetin’, ‘tobramycin’, ‘ciprofloxacin’, ‘ofloxacin’ and ‘levofloxacin’.
Duplicate articles were screened out from the list of results, which were then screened by title and abstract. During the next stage, full papers were reviewed; the articles were excluded if they did not meet the inclusion criteria or if they were of poor quality (aided by the Consolidated Standards of Reporting Trials ‘CONSORT’ checklist).
The search was up to date as of January 2015.
Results
The initial search identified 1351 articles. Eighty-one full text papers were examined. Nine trials of sufficient quality, which included 1116 patients, were found to meet the inclusion criteria.Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37–Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
Study characteristics
The trials included were published over a 13-year period, from 1995 to 2008. They varied widely, making meta-analysis impossible.
The population and setting varied extensively across the studies. Seven trials focused on CSOMReference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37–Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 and two investigated post-tympanostomy prophylaxis.Reference Morpeth, Bent and Watson44, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45 Two studies were community-based, examining a paediatric age group (ages 1–14 and 1–16 years) and included only Aborigines in Australia.Reference Leach, Wood, Gadil, Stubbs and Morris38, Reference Couzos, Lea, Mueller, Murray and Culbong39 The trials investigating post-tympanostomy prophylaxis were carried out in the USA, in large ENT departments, and included only paediatric patients (aged 7–132 and 5–146 months).Reference Morpeth, Bent and Watson44, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45 The remaining trials were carried out in a variety of areas: Spain,Reference Miro41 Hong Kong,Reference Tong, Woo and Hasselt40 Jordan,Reference Nawasreh and Fraihat42 TurkeyReference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 and Israel.Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37 Patients in these trials were treated in a hospital ENT setting and included a wider age range (14–71, 9–62, 9–65 and 19–73 years).
Follow up was limited in most of the studies. Three of the CSOM trials followed patients up to the end of the treatment course of 10 days;Reference Miro41–Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 the other studies followed patients up for 21, 30, 84 and 90 days.Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37–Reference Tong, Woo and Hasselt40 The post-tympanostomy prophylaxis trials followed patients up for 10–14 and 21 days.Reference Morpeth, Bent and Watson44, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
Of the CSOM trials, six included ciprofloxacin,Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37–Reference Couzos, Lea, Mueller, Murray and Culbong39, Reference Miro41–Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 one included ofloxacin,Reference Tong, Woo and Hasselt40 two included gentamicin,Reference Nawasreh and Fraihat42, Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 two included Sofradex® (dexamethasone, framycetin and gramicidin),Reference Leach, Wood, Gadil, Stubbs and Morris38, Reference Couzos, Lea, Mueller, Murray and Culbong39 two included Otosporin® (hydrocortisone, neomycin and polymyxin)Reference Tong, Woo and Hasselt40, Reference Miro41 and one included tobramycin.Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37 The post-tympanostomy trials compared Cortisporin® (neomycin, polymyxin and hydrocortisone) with ciprofloxacin or ofloxacin respectively.Reference Morpeth, Bent and Watson44, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
The definition of CSOM was comparable through all the studies (defined as chronically discharging ears with a non-intact tympanic membrane).
Study bias
Of the CSOM trials, four were double-blindedReference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37–Reference Tong, Woo and Hasselt40 and three were not blinded.Reference Miro41–Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 Of the non-blinded trials, one was using different packaging and dosing schedules so abandoned blinding,Reference Miro41 but the remaining two gave no explanation.Reference Nawasreh and Fraihat42, Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 Of the two post-tympanostomy prophylaxis trials, both were double-blinded at the time of otoscopy.Reference Morpeth, Bent and Watson44, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
All studies were described as randomised; however, only three included adequate information on how randomisation was conducted.Reference Leach, Wood, Gadil, Stubbs and Morris38, Reference Couzos, Lea, Mueller, Murray and Culbong39, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
Although the study by Miro did match numbers of patients who had tympanostomy tubes or had undergone middle-ear surgery between the two treatment groups, the author did not comment on any difference in outcome between these patients and those who had not undergone these interventions.Reference Miro41
Fradis et al. randomised patients' ears separately if bilateral disease was present;Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37 all other studies randomised each patient to one treatment. In all studies, the patients included had all undergone some treatment previously, but a treatment-free period prior to entering the study was standard (range, 3–14 days). In the study by Leach et al., all patients had previously received treatment with aminoglycosides, which had failed; this may have made it more likely that ciprofloxacin would be more effective.Reference Leach, Wood, Gadil, Stubbs and Morris38
Otorrhoea resolution
Two studies showed a significantly higher clinical cure rate with quinolones compared to aminoglycosides.Reference Couzos, Lea, Mueller, Murray and Culbong39, Reference Tong, Woo and Hasselt40 A double-blind, randomised, single-centre trial by Tong et al. revealed that significantly more patients treated with ofloxacin had resolution of otorrhoea compared to those treated with Otosporin (93 per cent vs 71 per cent, p = 0.04).Reference Tong, Woo and Hasselt40 Couzos et al. conducted a larger trial (111 patients completed treatment), which again was double-blind and randomised.Reference Couzos, Lea, Mueller, Murray and Culbong39 This multicentre trial compared ciprofloxacin with Sofradex in a paediatric aboriginal population and found a cure rate of 76.4 versus 51.8 per cent respectively (p = 0.009, absolute difference of 24.6 per cent, 95 per cent confidence interval (CI) = 15.8–33.4 per cent). They also recorded rates of perforation healing, but found no difference between the groups.Reference Couzos, Lea, Mueller, Murray and Culbong39
Miro and the Spanish ENT Study Group conducted the largest study in this series.Reference Miro41 They compared ciprofloxacin and Otosporin in a multicentre trial. They initially recruited 328 patients, but finished with 232. This was a well-constructed trial, with good information on power calculations and randomisation protocols. However, the trial was not blinded because of differences in the application of the drug (single vs multiple dose bottles and different dosing schedules). The study was designed and powered to prove the equivalence of ciprofloxacin to Otosporin in treating CSOM. They showed a resolution in otorrhoea of 91 per cent in the quinolone group and 87 per cent in the aminoglycoside group (absolute difference of 4 per cent, 90 per cent CI = 4.8–8.86 per cent), and concluded that ciprofloxacin was at least as effective as Otosporin in their population.
Leach et al. conducted their study in another paediatric aboriginal population.Reference Leach, Wood, Gadil, Stubbs and Morris38 This randomised, double-blind, single-centre trial compared ciprofloxacin and Sofradex in a group of 97 children who had persistent otorrhoea despite treatment with topical aminoglycosides. The findings showed resolution of otorrhoea in 70 per cent of the quinolone group and in 72 per cent of the aminoglycoside group (absolute difference of 2 per cent, 95 per cent CI = −20–16 per cent). The authors concluded that there was no clinical difference between the two treatments in the study population.Reference Leach, Wood, Gadil, Stubbs and Morris38
A further two studies showed a statistically significant difference in microbiological cure in favour of quinolones over aminoglycosides (87.5 per cent vs 30 per cent, p < 0.001, and 88 per cent vs 30 per cent, p < 0.001).Reference Nawasreh and Fraihat42, Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 These studies were unblinded and neither published specific data on clinical outcomes.
The study by Fradis et al., which randomised 60 ears to ciprofloxacin, tobramycin and placebo groups, was inadequately powered and therefore did not produce any statistically significant results.Reference Fradis, Brodsky, Ben-David, Srugo, Larboni and Podoshin37
Post-tympanostomy primary outcomes
Both post-tympanostomy prophylaxis studies failed to show a statistically significant difference between the quinolones and aminoglycosides in reducing otorrhoea. In the study by Poetker et al., a control group which received no antibiotic prophylaxis was included, and a significant difference was seen between this group and both antibiotic groups (85.1 per cent clear in the control group vs 94.5 per cent in the Cortisporin group (p = 0.01) and 91.9 per cent in the ofloxacin group (p = 0.04)).Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45 Poetker et al. also showed a significant improvement in blocked tube rates between the control group and antibiotic groups, but there was no significant difference between the quinolone and aminoglycoside groups.Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45
Secondary outcomes
All studies published details on bacteriology. P aeruginosa was cultured in all trials and was the most common in four of the six relevant studies.Reference Leach, Wood, Gadil, Stubbs and Morris38, Reference Couzos, Lea, Mueller, Murray and Culbong39, Reference Nawasreh and Fraihat42, Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 Staphylococcus aureus was also cultured in all studies and was the most common in two of them.Reference Tong, Woo and Hasselt40, Reference Miro41 Streptococci and enterococci were also documented, but were much less common.
Four papers provide details on antibacterial sensitivities; globally, there was more resistance to aminoglycosides than quinolones. The study by Miro described the highest resistance to quinolones, with 16 per cent reported resistance to ciprofloxacin; aminoglycoside resistance was higher, however, with 22 per cent of cultures resistant to neomycin.Reference Miro41 In a study by Nawasreh and Fraihat, 52.3 per cent of pre-treatment bacteria were resistant to gentamicin, but none were resistant to ciprofloxacin.Reference Nawasreh and Fraihat42 Resistance to gentamycin developed in a further 29 per cent of patients post-treatment, whereas ciprofloxacin resistance remained at 0 per cent. Similarly, Tutkun et al. showed 0 per cent pre- and post-treatment resistance to ciprofloxacin, compared to 60 per cent pre-treatment and 70 per cent post-treatment resistance to gentamicin.Reference Tutkun, Ozagar, Koc, Batman, Uneri and Sehitglu43 Leach et al. alone reported a small increase in ciprofloxacin-resistant pathogens after treatment, from 0 to 2 per cent.Reference Leach, Wood, Gadil, Stubbs and Morris38
With regard to safety, across all the studies, only one patient was reported as having a significant change in pure tone audiometry values.Reference Miro41 The patient was documented as having a new all frequency hearing loss after receiving aminoglycosides, but the study does not give any more details. No other significant side effects were recorded in any study. However, minor short-term side effects, including pruritus, vertigo and bitter taste, were recorded in several studies in both treatment arms.Reference Couzos, Lea, Mueller, Murray and Culbong39–Reference Miro41
Discussion
This systematic review revealed nine good quality trials. There was no evidence that aminoglycosides were superior to quinolones in treating CSOM. Quinolones were effective at resolving otorrhoea and in accomplishing microbiological eradication. Quinolones were also as effective as aminoglycosides in reducing otorrhoea and tube blockage in patients with tympanostomy tubes.
The main limitation of the evidence base was a wide variation in population and location. No trials have been carried out in the UK. It is therefore unclear whether the outcomes would be comparable in a British population, which may have different bacteriology and patient factors. The numbers included in the trials are also relatively small (the largest comprised 232 patients) and contain possible sources of bias as discussed above.
Since the most recent relevant Cochrane review, in 2005,Reference Macfadyen, Acuin and Gamble6 there have been two new trials.Reference Leach, Wood, Gadil, Stubbs and Morris38, Reference Poetker, Lindstrom, Patel, Conley, Planary and Link45 More importantly, that review excluded preparations without steroids. The question asked in this paper is more clinically relevant, as it seeks to inform the choice made by clinicians when choosing a medication for a patient with a non-intact tympanic membrane. In UK practice, aminoglycoside-steroid combinations are commonplace, whereas quinolones are rarely available in a combination form. This review included four trials that examined aminoglycosides combined with steroids; these were not commented on in the Cochrane review. The current review has shown that quinolones without steroids are equal to or better than aminoglycosides with steroids. The extra benefit to be gained from combining a steroid with a quinolone needs further research.
The ENT-UK consensus document did not explore the efficacy of topical quinolones as they are not licensed.Reference Philips, Yung, Burton and Swan1 Concern about increasing the prevalence of ciprofloxacin-resistant bacteria is cited in the ENT-UK document as another barrier. However, there is no convincing evidence of this. In the studies reviewed here, resistance to ciprofloxacin was low and there were no significant changes after treatment; conversely, there were high levels of aminoglycoside resistance, and this increased following exposure. Furthermore, the high levels of resistance reported in some texts do not compare to the efficacy of quinolones in treated CSOM; this may suggest that the resistance is partial and can be overcome by the high concentrations achieved topically.Reference Ghosh, Panarese, Parker and Bull3, Reference Poole46
The General Medical Council (GMC) advises consulting the British National Formulary for guidance in prescribing medication. However, the British National Formulary is unclear in regard to prescribing therapy for CSOM. The British National Formulary states that ototopical aminoglycosides are contraindicated in the presence of a tympanic membrane perforation but that they are used by specialists.47 It describes quinolone ear drops as ‘an unlicensed but effective alternative’.
The GMC states that drugs should usually be prescribed within the terms of their licence.48 However, the prescription of unlicensed medications is advised if ‘there is no suitably licensed medicine that will meet the patient's need’.48 An application for licensing of medication is not driven by clinical need but by market forces; a manufacturer needs a licence to import and market a drug.49 Therefore, if there is not sufficient demand (and therefore profit) for a drug, then a company will not wish to market the drug and no licence will be applied for.
A search of non-UK national formularies revealed licences for topical quinolones for use with a non-intact tympanic membrane. In Australia, ear drops containing quinolones are licensed specifically for treating CSOM in adults and children.50, 51 Ciprodex® (ciprofloxacin-dexamethasone) is licensed in Canada for patients aged six years or older with non-intact tympanic membranes, and in the USA ciprofloxacin is licensed for ototopical use in three formations (Cipro® HC, Ciprodex and Cetraxal®).52, 53
Conclusion
None of the available evidence from randomised trials has shown that aminoglycosides are superior to quinolones in treating CSOM. Topical quinolones do not carry the risk of ototoxicity that aminoglycosides do. Furthermore, they are equal or more effective in treating CSOM, or when used as prophylaxis post-myringotomy. They should therefore be considered as a first-line treatment for these patients, as is already the case in many countries around the world. If licensing is truly a barrier to the adoption of this policy, perhaps it is time that the UK's ENT community began to take a more proactive approach in calling for quinolones to be licensed for ototopical use.
