Discussion

The first step in our guideline (Figure 1) is identifying patients for whom a diagnosis of necrotising otitis externa is suspected; although this may result in increased sensitivity of pick up, it ensures that cases are not missed. Lambor et al. stated that any diabetic patient presenting with otalgia and otorrhoea should be deemed to have necrotising otitis externa until proven otherwise.^{Reference Lambor, Das, Goel, Tiwari, Lambor and Fegade3} This concept has been adapted to define our patients with suspected necrotising otitis externa, along with findings from our retrospective case series and large-scale patient demographic studies in the literature.^{Reference Sylvester, Sanghvi, Patel, Eloy and Ying1,Reference Guerrero-Espejo, Valenciano-Moreno, Ramírez-Llorens and Pérez-Monteagudo4} This is the evidence used to constitute our diagnostic criteria. The diagnosis should also be considered in patients with a cranial nerve palsy, most commonly the VIIth cranial nerve, and signs of ear infection; however, this will often be a late presenting sign.^{Reference Soundry, Joshua, Sulkes and Nageris5} We also felt it prudent to include a failure to respond to one week of conservative otitis externa treatment as a ‘red flag’ sign to consider necrotising otitis externa.

Fig. 1. NHS Lothian guideline for the diagnosis and management of necrotising otitis externa (NOE). DM = diabetes mellitus; C&S = culture and sensitivity testing; CRP = C-reactive protein; HBA1c = haemoglobin A1c; Pip-Taz = piperacillin/tazobactam; IV = intravenous; qds = four times a day; tds = three times a day; CT = computed tomography; MRI = magnetic resonance imaging; OE = otitis externa; BM = blood glucose measurement; d/w = discussion with; po = orally; EUA = examination under anaesthesia; OPAT = out-patient parenteral antibiotic therapy; OPD = out-patient department; PICC = peripherally inserted central catheter

On clinical review of patients with suspected necrotising otitis externa, microsuction should initially be undertaken for adequate assessment of the ear canal. At this stage, swab samples of discharge should be sent for culture and sensitivity testing – both routine and fungal – to determine the organisms involved and to guide antibiotic choice. Tissue samples of any granulation should be obtained, and sent for both culture and sensitivity testing and histopathology. This is important in the potential identification of organisms that may be missed from superficial swab cultures. Furthermore, alternative diagnoses such as squamous cell carcinoma of the ear canal can, rarely, present similarly to necrotising otitis externa and must be excluded.^{Reference Zainuddin and Abdullah6} There is evidence that CRP is correlated with disease status (although our local audit findings do not support this) and therefore should be checked on admission.^{Reference Hopkins, Harris and Cuddihy7}

It is quoted that patients with good diabetic control do not progress to the more severe form of necrotising otitis externa with involvement of surrounding anatomical structures.^{Reference Lee, Lee, Seon, Jung, Lee and Choi8} Therefore, in patients with diabetes, particular attention should be placed on glycaemic control, with measurement of haemoglobin A1c and serial monitoring of blood glucose. Referral to the diabetic specialist team for review of diabetic control thus forms part of the management of necrotising otitis externa.

Regarding antimicrobial therapy, if there is a high level of clinical suspicion of necrotising otitis externa, empirical intravenous treatment should be commenced after microbiological sampling. Locally, piperacillin/tazobactam is the first-line therapy because of its efficacy against the most prevalent organisms in necrotising otitis externa, in particular P aeruginosa. Ceftazidime may be suitable for patients with a history of non-urticarial rash following penicillin use, although the risk of diarrhoea due to Clostridium difficile is greater. A true penicillin allergy requires elucidation through careful history taking; if necessary, patients can be referred for further investigations to our penicillin allergy clinic (the authors acknowledge this is not always available), or review on the ward by an infection specialist. All antibiotics should be reviewed following the analysis of culture and sensitivity findings.

From our retrospective data, a factor identified as having potential to improve patient outcomes was the presence of a named consultant otologist responsible for decision-making for individual patients. This enables a clear management plan to be followed and encourages accountability for a patient's care. Furthermore, such individuals are best placed to ensure the correct condition is being pursued. Under inexperienced care, the diagnosis can easily be confused with other conditions such as squamous cell carcinoma of the ear canal or external ear canal cholesteatoma.

The next area to consider is what happens when the patient has clinically suspected necrotising otitis externa but lacks radiological features of disease. As discussed, alternative diagnoses such as malignancy or cholesteatoma should be considered and ruled out. Otherwise, we would classify this under the ‘severe’ otitis externa category. These patients should be treated actively in the first instance, with reviews at 48-hour intervals to determine the response to treatment, as early necrotising otitis externa can be missed on imaging.^{Reference Sudhoff, Rajagopal, Mani, Moumoulidis, Axon and Moffat9} Experience has shown that if risk factors for necrotising otitis externa are present and the patient is under-treated, there is a potential for the condition to develop into osteomyelitis of the temporal bone.^{Reference Rubin Grandis, Branstetter and Yu10}

Those with diagnosed necrotising otitis externa undergo a two-week induction course of intravenous antibiotics, in line with current practice of treating osteomyelitis at other sites, although this concept has recently been challenged.^{Reference Li, Rombach, Zambellas, Walker, McNally and Atkins11} A comprehensive re-assessment is required at this stage, as suboptimal clinical recovery should prompt consideration of further tissue sampling, further scanning, and discussion with the infectious disease team or microbiologist regarding the appropriateness of the chosen antibiotics. At this time, the surgeons may want to consider whether local debridement or incision and drainage might be a useful adjunct to treatment.^{Reference Carfrae12} Indeed, a body of evidence suggests that samples from this deep debridement may harvest fungal organisms that have not been identified on previous cultures.^{Reference Gruber, Sela, Doweck, Roitman, Uri and Srouji13,Reference Mion, Bovo, Marchese-Ragona and Martini14}

Should a clinical improvement be observed, with no adverse clinical features present, and P aeruginosa susceptible to ciprofloxacin has been isolated, then an oral switch may be considered at this stage. We would advocate the prescription of a subsequent 4-week course of ciprofloxacin (750 mg) every 12 hours, unless renal impairment is present. Ciprofloxacin has excellent bone penetration; furthermore, this therapy allows the patient to be treated at home. Data from our audit investigating pseudomonal isolates support ciprofloxacin as a valid option in the majority of patients. The caveats are that this could change with ongoing antibiotic pressure and sub-therapeutic doses being administered topically. Patients also require counselling on the risk of tendon inflammation or rupture, neurological complications, risk of ruptured aneurysm,^{Reference Pasternak, Inghammar and Svanstrom15} and the increased risk of C difficile infection.¹⁶

Patients should be monitored clinically until recovery is ensured, with regular microsuction and examination of the external auditory canal. Subsequent to this, we advocate that patients attend follow up with their named otologist at four to six weeks to ensure clinical recovery is sustained. However, regarding relapse, local data appear to be reassuring, with only 1 out of 20 cases relapsing following completion of treatment.

Clearly, avoiding the progression of disease to severe necrotising otitis externa is essential. We classify severe necrotising otitis externa here as osteomyelitis of the petrous portion of the temporal bone, involving surrounding structures, and resulting in cranial nerve palsies, which, once present, are usually irreversible and life-altering. Indeed, any patient suffering from necrotising otitis externa with cranial nerve involvement would be subject to at least six weeks of intravenous antipseudomonal antibiotics. Furthermore, patients who have received two weeks of intravenous antibiotics without observation of sustained clinical recovery, or worse, an observation of disease progression, would also be considered for a further four weeks of intravenous therapy.

Current intravenous antipseudomonal antibiotics require frequent dosing, precluding the use of many out-patient parenteral antibiotic therapy services. Consequently, many patients are confined to hospital wards for this protracted period, which is detrimental to patient psychological health, as well as being a burden on a service where in-patient bed availability is increasingly under pressure. Our out-patient parenteral antibiotic therapy team has developed a service that facilitates the training of family or caregivers in administering intravenous antibiotics at home. This is offered in selected cases, whilst ensuring the patients receive regular review of their clinical status, particularly with regard to the appearances of the external auditory canal.

Necrotising otitis externa caused by invasive fungal infection, usually aspergillus species, is rare but must not be overlooked. These infections often develop from the mastoid or middle ear, and can be much more invasive. Consequently, identifying the pathogen early is essential. The ‘gold standard’ examination involves testing for the presence of hyphae in tissue samples. Recognition that a single isolate of mould or yeast from a swab must not be considered an indication for prolonged systemic antifungal treatment is essential. If invasive infection with aspergillus species is confirmed, then prolonged treatment (more than 12 weeks) with voriconazole is advised.^{Reference Mouas, Lutsar, Dupont, Fain, Herbrecht and Lescure17–Reference Patterson, Thompson, Denning, Fishman, Hadley and Herbrecht19} This should be overseen by the local consulting infection team, with levels assessment after 7 days or following any dose adjustment. Yeasts isolated from superficial swabs are not generally treated with antifungals; however, such a finding should prompt consideration of deep tissue sampling to determine any requirement for systemic antifungal therapy.

Evidence on topical antibiotics remains ambiguous and therefore our current guideline does not require specific topical treatment to be prescribed. Many ENT specialists advocate the use of topical treatment, and further work investigating the efficacy of this therapy is needed.

A recent systematic review was conducted by Phillips and Jones, which investigated the evidence for using hyperbaric oxygen therapy as an adjunct in necrotising otitis externa, and it was found to be equivocal.^{Reference Phillips and Jones20} Therefore, this has not been suggested as a therapeutic option in our guideline.

There is a paucity of evidence on surgical debridement, with much discrepancy between studies and with no common factors linking debridement to a positive outcome.^{Reference Carfrae12,Reference Singh and Bhardwaj21} We therefore concluded that this should be considered on a case by case basis by the patient's named otologist. It is reasonable to suggest that local soft tissue debridement or incision and drainage of abscesses would be a useful therapeutic adjunct, particularly in patients whose progress appears to have stalled.

Imaging modality remains a point of contention. The bulk of the evidence suggests that contrast-enhanced CT of the petrous portion of the temporal bone should be used as a gold standard for the detection of temporal bone osteomyelitis,^{Reference Hopkins, Harris and Cuddihy7,Reference Chawdhary, Pankhania, Douglas and Bottrill22} given its sensitivity, ease of access and cost-effectiveness. As a result, this is our first-line radiological investigation. However, there is evidence to suggest that CT may miss early stages of disease; therefore, we have reserved MRI for further clarification of disease extent in cases where clinical and radiological findings are incongruous.^{Reference Sudhoff, Rajagopal, Mani, Moumoulidis, Axon and Moffat9} From our own data, we have observed a case where the findings of both CT and MRI were negative, yet necrotic bone was excised on debridement. This episode precipitated a large-scale review of imaging modality and reporting. Good results are achieved with a dedicated ENT radiologist who has the opportunity to meet regularly with the ENT surgical department and who is given full clinical details (facilitated by radiology referrals being made by more senior members of the team).