Conservative management

The use of prophylactic antibiotics to reduce the risk of meningitis is controversial. In our series, patients with a CSF leak were no more likely to be given antibiotics than those without a CSF leak. Those with concurrent infection are treated with prophylactic antibiotics. However, there are well known arguments against their use, including avoiding the masking of early meningitis signs.^{Reference Patel and Groppo18} In contrast, many studies support the use of antibiotics in patients with concurrent infection or lacerations.^{Reference Brodie and Thompson4,Reference Nishiike, Miyao, Gouda, Shimada, Nagai and Nakagawa13} A 2011 Cochrane review that examined evidence from 208 patients from 5 randomised, controlled trials (RCTs) did not demonstrate any significant difference between antibiotic prophylaxis groups and control groups in terms of meningitis or ongoing CSF leak. This conclusion is also supported by non-RCT data.^{Reference Ratilal, Costa, Sampaio and Pappamikail19}

Surgical repair

The majority of CSF leaks close spontaneously within one week. The optimal timing of surgical repair is debated. Some centres advocate repair within 7 days, in light of evidence of a reduction in meningitis risk in this timeframe.^{Reference Brodie and Thompson4} In our centre, surgical repair is generally performed at around two weeks, following joint discussion with the neurosurgery department. Discussion points include the trial of a lumbar drain, and consideration of other factors such as: patient's symptoms, degree of leakage, fitness for surgery, status of hearing in the contralateral ear, brain herniation through the tegmen, patency of the ear canal and the location of the fistula. However, within the time frame studied, we had no cases that required surgical exploration.

Surgical techniques vary depending on the clinical scenario. For patients with profound SNHL, blind sac closure is recommended.^{Reference Kveton20–Reference Diaz, Cervenka and Brodie22} Patients with intact hearing and tegmen defects may undergo repair via a mini-craniotomy or an extradural middle cranial fossa approach.^{Reference Adkins and Osguthorpe17} Larger defects may be repaired using an intradural approach.^{Reference Glasscock, Dickins, Jackson, Wiet and Feenstra23} Cases may require neurosurgical assistance; it is therefore important that patients are managed at the appropriate centre. In patients with CSF leaks arising from the anterior skull base, a transnasal or bifrontal craniotomy approach may also be required.

Conservative management

Our results showed that 50 per cent of patients with delayed facial nerve palsy (10 out of 20) were administered high-dose steroids, compared to only one patient with immediate facial nerve palsy.

Results from the ‘CRASH’ (corticosteroid randomisation after significant head injury) trial (which involved more than 10 000 patients from across 49 countries) demonstrated that high-dose steroid use is contraindicated in patients with traumatic brain injury because of a significantly increased mortality rate.^{Reference Edwards, Arango, Balica, Cottingham, El-Sayed and Farrell24} Subsequently, the 2005 Cochrane review advised against steroid use in this patient population.^{Reference Alderson and Roberts25} The patients in the trial were randomised to a 2 g loading dose plus a 400 mg maintenance dose of methylprednisolone within 48 hours of injury. Although steroid use in the context of facial palsy within ENT is generally far more conservative (generally only doses of up to 60 mg of prednisolone), it is important to exercise caution when prescribing high-dose steroids in this patient population, based on this evidence.

Surgical management

It is largely accepted that patients with delayed facial nerve palsy will make a complete recovery.^{Reference Dahiya, Keller, Litofsky, Bankey, Bonassar and Megerian3,Reference Brodie and Thompson4} It has been reported that in some cases recovery is achieved in only 94 per cent of patients.^{Reference Turner26} One case report describes facial nerve transection during exploration of the nerve.^{Reference May27}

Injury to the facial nerve can be due to an avulsion injury or compression from comminuted bony fragments, and can occur anywhere between the internal auditory meatus and the vertical portion of the facial nerve. The area most commonly affected is the peri-geniculate region and the medial aspect of the middle ear where it is most susceptible to injury.^{Reference Fisch28–Reference Coker, Kendall, Jenkins and Alford30} A large series of 65 patients with trauma-related facial palsy (52 had immediate facial nerve paralysis) found that 66 per cent of injuries occurred at the peri-geniculate ganglion and 13 per cent of injuries resulted in a nerve gap.^{Reference Darrouzet, Duclos, Liguoro, Truilhe, De Bonfils and Bebear31}

Surgical repair of the facial nerve is controversial. Such treatment is dependent on a number of factors, including: the decision to operate, the timing of surgery and the surgical approach used (based on injury location). Fisch advocates repair as soon as possible, once more than 90 per cent Wallerian degeneration has been demonstrated on electroneuronography (ENoG, see discussion below).^{Reference Beck and Benecket32–Reference Shelton and Shelton34} A translabyrinthine approach may be utilised in patients with profound SNHL. Middle cranial fossa or transmastoid approaches may be used in patients with preserved hearing.^{Reference Shelton and Shelton34} Expected recovery rates following decompression are estimated to be around 55 per cent.^{Reference Johnson, Semaan and Megerian35} One series describing the conservative management of immediate facial nerve palsy reported good recovery rates in less than 60 per cent of patients. In practice, this may complicate decision making in cases that present to ENT beyond two to three weeks. The two most important factors in considering facial nerve decompression or repair are the severity of the injury and the delay of onset. In some patients, it is more appropriate to undergo facial nerve reanimation using dynamic or static techniques. This is usually performed within two years of injury and requires multidisciplinary input. Non-operative techniques include Botox® injections and physiotherapy.^{Reference Gordin, Lee, Ducic and Arnaoutakis36}

Information from three patients in our series who underwent surgical exploration was retrieved. In the first case, intra-operative findings showed bony spicules embedded in the geniculate ganglion. These spicules were removed, but facial function remained poor, with synkinesis affecting all muscle groups. The patient underwent a trial of Botox injections, but recovery remained limited and the patient was referred for consideration of muscle transfer.

In the second case, there was a severed nerve in the descending portion, and the patient underwent a sural nerve interposition graft at five months post-injury. Subsequent canalplasty was performed in light of intractable granulation tissue within the fractured floor of the external ear canal. Facial nerve recovery was also poor and the patient was referred to the plastic surgery department for consideration of facial nerve reanimation.

In the third case, the patient had probable immediate House–Brackmann grade VI bilateral facial nerve palsies. At four weeks post-injury, the patient underwent electromyography (EMG), which revealed ‘profuse spontaneous activity in facial nerve muscles bilaterally without recordable motor unit action potentials under voluntary control’. At two months post-injury, the patient underwent left-sided facial nerve decompression (vertical portion to first genu). The incus was found pushed in toward the first genu and was decompressed (plus ossiculoplasty). Marginal bilateral improvement was achieved at four months post-operation, on the operated and non-operated side. The patient has since made further recovery, with House–Brackmann grades improving to grade II bilaterally.

In all of the above cases, intervention was delayed beyond the two-week window because of co-existent injury, and ENoG was not used.

Figure 2 demonstrates evidence of ossicular dislocation, a fracture through the facial nerve canal and pneumolabyrinth in a patient who fell from a height of 20 feet (6 metres).

Fig. 2. Axial, non-contrast computed tomography scan of a patient who sustained a fall from 20 feet (6 metres), showing a non-otic capsule sparing fracture of the right temporal bone. The thin white arrow indicates an absent incus, the black arrow demonstrates a fracture line through the facial nerve canal, and the thick white arrow shows pneumolabyrinth.

Role of nerve conduction studies

Electroneuronography and EMG are objective tests of facial nerve function, and are helpful in guiding treatment and determining prognosis following facial nerve injury.^{Reference Mannarelli, Griffin, Kileny and Edwards37} In short, ENoG is an evoked EMG; it records nerve conduction distal to the site of injury, by measuring muscle response to an electrical stimulus.^{Reference Gantz, Rubinstein, Gidley and Woodworth38} It is able to distinguish between first-degree injuries (i.e. neuropraxia) and injuries that have developed Wallerian degeneration (second-degree to fifth-degree injuries). The speed of the onset of Wallerian degeneration is proportional to the severity of injury. Fisch established that Wallerian degeneration was complete by 3–5 days in fifth-degree injuries (complete nerve transection), but took less than 21 days in second-degree injuries (axonotmesis).^{Reference Fisch39}

It is generally accepted that patients with a positive ENoG test result who have more than 90 per cent Wallerian degeneration are more likely to develop incomplete recovery. This evidence predates the House–Brackmann facial nerve grading system, making it difficult to compare the quoted positive predictive value of 91 per cent with figures reported in more recent studies.^{Reference House and Brackmann40} Gantz et al. has demonstrated 100 per cent recovery to House–Brackmann grades I or II in cases where ENoG studies showed a negative result.^{Reference Gantz, Rubinstein, Gidley and Woodworth38}

It is advised that EMG be performed at 10–14 days post-injury to allow for abnormal healing or resting potentials to occur. In traumatic cases where there is commonly a delay in performing nerve conduction studies (because of the initial treatment of immediately life-threating injuries), an EMG may be more suitable than ENoG. However earlier investigation should be conducted when possible, as a delay of 10–14 days may impact nerve recovery. If any electrical activity is demonstrated on EMG then nerve transection is ruled out.^{Reference Mannarelli, Griffin, Kileny and Edwards37}

The decision to perform nerve conduction studies varies widely in the UK. This may be because of surgeon preference or a lack of equipment availability. Some surgeons are reassured that a negative ENoG finding and voluntary muscle activity are sufficient to indicate that a patient will have a high probability of regaining normal or near-normal function, and therefore does not require surgical exploration or repair.^{Reference Gantz, Rubinstein, Gidley and Woodworth38} Other surgeons rely on high-quality imaging alongside a convincing history when deciding whether to operate, thus potentially limiting its value in the acute setting. At our centre, we are more likely to rely exclusively on imaging to guide decisions regarding surgery, particularly when the clinical case is unclear.