Introduction
Vestibular schwannomas are benign tumours that develop from the sheath of the vestibulocochlear nerve.Reference Kentala and Pyykko1 Most vestibular schwannomas are unilateral, so-called sporadic, and mostly present with unilateral or asymmetric sudden or progressive sensorineural hearing loss, with or without persistent unilateral tinnitus. Problems with balance or other compression phenomena could be present depending on the size of the tumour and whether there is compression of cranial nerves or brainstem structures.Reference Kentala and Pyykko1,Reference Stangerup, Caye-Thomasen, Tos and Thomsen2 While tumour measurement methods can vary, the size of the tumour can be classified as small (less than 1.5 cm), medium (1.5–2.5 cm), large (2.5–4.0 cm) and giant (more than 4.0 cm).Reference Olson, Kalkanis and Ryken3 There can be differences in the management of patients with vestibular schwannomas due to local preferences and available resources; however, generally, for larger tumours, microsurgery is the preferred choice of treatment, with translabyrinthine and retrosigmoid approaches used for vestibular schwannomas with intracranial extension.Reference Olson, Kalkanis and Ryken3 Surgical removal of vestibular schwannomas usually causes acute vestibular symptoms.Reference Betka, Zvěřina, Balogová, Profant, Skřivan and Kraus4 These acute vestibular symptoms are a key factor in the post-operative recovery process; thus, targeting them is important in order to speed up vestibular compensation and post-operative recovery.
Previous studies have shown that pre-operative intratympanic gentamicin application can result in a substantial reduction in peripheral vestibular function in all semicircular canals.Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö5 Should intratympanic gentamicin be applied pre-operatively, it could prevent sudden loss of peripheral vestibular function and speed up vestibular compensation after vestibular schwannoma resection.Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö5 Based on the above concept and published works, some centres have used intratympanic gentamicin pre-operatively to improve recovery times post-operatively.
Despite the presence of some studies, there has been no systematic assessment of the level of evidence behind such a relatively new intervention. On these grounds, this systematic review aimed to perform an analysis of studies assessing the effect of pre-habilitation with intratympanic gentamicin on patients’ recovery following vestibular schwannoma resection.
Materials and methods
Search strategy and research question
We performed a systematic review with the research question being whether pre-operative intratympanic gentamicin injections improve the post-operative recovery following vestibular schwannoma resection. We used the databases Medline, PubMed, Frontiers, Cochrane Library, Cambridge Core and ScienceDirect to identify published studies on the pre-operative use of intratympanic gentamicin in patients with vestibular schwannomas. We used the key words ‘intratympanic gentamicin’, ‘vestibular schwannoma’, ‘preoperative’, ‘prehabilitation’ and ‘vestibular compensation’. The local Library Network facilitated a comprehensive literature search.
Our methodology was adapted from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) 2020 guideline for systematic reviews.Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow6 We used the ‘PICO’ (P-Population, I-Intervention, C-Control or Comparison, O-Outcome) framework to review the different components of the studies included, as mentioned in the Cochrane Handbook for Systematic Reviews.7
Eligibility criteria
We initially retrieved 17 studies. Overall, nine of these met the inclusion versus exclusion criteria (Figure 1).
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) flowchart. VS = vestibular schwannoma
Our inclusion criteria were studies that: (1) are clinical; (2) were published within the last 20 years; (3) were published in English or German language; and (4) involved the use of intratympanic gentamicin pre-operatively in vestibular schwannoma patients prior to vestibular schwannoma resection.
We excluded: (1) case reports; (2) conference abstracts; (3) studies that were published more than 20 years ago because pre-habilitation is a relatively recently described concept; and (4) studies with patients with vestibular schwannomas treated with pre-operative intratympanic gentamicin assessing factors not related to their post-operative recovery.
Using our inclusion and exclusion criteria, one potential study, by Tjernström et al.,Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö5 was excluded. While the objective in that study was to evaluate vestibular and auditory function after pre-habilitation with gentamicin in vestibular schwannoma patients, it reported the effect on function before patients underwent surgery, not after. Thus, a total of eight studies were included in this review.
Collected data
We collected the following listed data: (1) geographic origin and year of the study; (2) the presence of a control group; (3) level of evidence; (4) number of enrolled patients; (5) intratympanic gentamicin regimen and injection procedure; (6) outcome measures of effect of intratympanic gentamicin on recovery; (7) size of vestibular schwannoma (dimensions or Koos classification); (8) surgical approach; and (9) side effects from intratympanic gentamicin.
Data analysis and risk of bias
Due to the small number of patients, we did not compose the data; instead, we used a qualitative analysis. In order to ensure validity and reliability of the retrospective studies included in this paper, a tool was used to assess risk of bias in case–control studies (Appendix 1).8 We accepted the heterogeneity of the included studies as a limitation, given the overall absence of control trials and limited number of enrolled patients.
Results
A total of eight studies were included and qualitatively analysed (Figure 1). The findings and characteristics of the included studies have been summarised (Tables 1–3) to allow for overview and comparison.Reference Fellmann, Bächinger, Dalbert, Röösli, Huber and Wettstein9–Reference Amiraraghi, Gaggini, Crowther, Locke, Taylor and Kontorinis16
Study origin and tumour-related factors
All the included studies originate from European countries; interestingly, there were none from the USA or Asia, areas that traditionally contribute significantly to the literature in this specific field. This distribution of the included studies probably highlights geographic differences in practice. The available studies have relatively few patients, ranging from 4 to 39 patients in each group (examined and control groups). The age distribution did not significantly vary across the different studies. All patients selected in the studies had a vestibular schwannoma with a tumour size ranging from grade II to grade IV (based on the Koos grading system for vestibular schwannomas), with linear measurements being the ones reported. Tumours were resected mainly via the retrosigmoid or translabyrinthine approach. In the study by Fellmann et al.,Reference Fellmann, Bächinger, Dalbert, Röösli, Huber and Wettstein9 10 per cent of patients underwent a transtemporal approach; in the work by Magnusson et al.,Reference Magnusson, Kahlon, Karlberg, Lindberg and Siesjö10 the resection approach was not specified (Table 1). The level of evidence was predominantly level IV (Table 1).17
Included studies, with demographics, approach and vestibular schwannoma size
SD = standard deviation
Outcome measures and assessment methods
In terms of intervention and/or exposure, a sufficient dose of gentamicin to cause vestibular ablation was given to the patients in the gentamicin group, while the patients in the control group were managed without pre-operative gentamicin application. The Video Head Impulse Test and/or caloric testing were used to monitor vestibular function in patients. In certain studies, additional testing was carried out, which included the subjective visual vertical measurement, posturography and the Activities-specific Balance Confidence scale in Hrubá et al.,Reference Hrubá, Chovanec, Čada, Balatková, Fík and Slabý11 as well as the subjective visual vertical scale, optokinetic testing and Dizziness Handicap Inventory, General Behaviour Inventory and Glasgow Health Status Inventory in Balatkova et al.Reference Balatkova, Cada, Hruba, Komarc and Cerny12 and Čada et al.Reference Čada, Balatková, Chovanec, Čakrt, Hrubá and Jeřábek14 Tjernström et al.Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö13,Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö15 and Magnusson et al.Reference Magnusson, Kahlon, Karlberg, Lindberg and Siesjö10 also used posturography parameters (vibratory perturbation) in addition to the Video Head Impulse Test and caloric testing.
Control group
The patients for both the control group and the gentamicin group were selected based on similar inclusion and/or exclusion criteria in all studies, although the screening tests used varied. In general, among other criteria that each study had developed, inclusion criteria for the control group included patients with serviceable hearing, as defined by the studies, a large tumour size that was considered unsafe to wait for several months for surgery to take place, and lack of consent for intratympanic gentamicin. Most studies also had criteria based on vestibular function and postural performance prior to surgery, and all of the studies performed tests to assess vestibular function such as the Video Head Impulse Test. However, one of the studies, Čada et al.,Reference Čada, Balatková, Chovanec, Čakrt, Hrubá and Jeřábek14 did not have criteria based on vestibular function or postural performance prior to surgery. Another study, by Magnusson et al.,Reference Magnusson, Kahlon, Karlberg, Lindberg and Siesjö10 did not have a control group, which potentially limits the ability to confirm that the study results are due to manipulation of independent variables and not extraneous variables (Table 2).
Participant numbers and gentamicin regimens
vHIT = Video Head Impulse Test; SD = standard deviation
Outcomes
None of the studies reported any clinically significant side effects. With regard to whether there is an effect of pre-habilitation with gentamicin in vestibular schwannoma patients on the post-operative recovery process, four of the studies showed a statistically significant positive effect, while the remaining four studies showed better outcomes for the intratympanic gentamicin group, but without a statistically significant difference. In the latter four studies, a degree of benefit was reported, as some of the patients had shorter in-patient stay or exhibited fewer symptoms of postural impairment, albeit the results are not statistically significant, which might have been due to the small number of enrolled patients. In addition, three of the studies used posturography parameters to demonstrate a long-term positive effect (Table 3). One of the studies, Magnusson et al.,Reference Magnusson, Kahlon, Karlberg, Lindberg and Siesjö10 demonstrated an additional benefit of the use of intratympanic gentamicin in vestibular schwannoma patients with tumours compromising the cerebellum but with preserved vestibular function (Table 3).
Outcome measures and side effects
DHI = Dizziness Handicap Inventory; VS = vestibular schwannoma; vHIT = Video Head Impulse Test; GBI = Glasgow Benefit Inventory
Risk of bias
We assessed the risk of bias using five domains (see Appendix 1 for additional details): (1) can we be confident in the assessment of exposure to gentamicin?; (2) can we be confident that cases had developed the outcome of interest (i.e. that intratympanic gentamicin had an effect on recovery) and controls had not?; (3) were the cases (those patients who were exposed to gentamicin and developed the outcome of interest) properly selected?; (4) were the controls (those patients exposed to gentamicin who did not develop outcome of interest) properly selected?; and (5) were cases and controls matched according to important prognostic variables, or was statistical adjustment carried out for those variables?
With respect to the assessment of exposure to gentamicin and outcome measures, we identified an increased risk of bias if evidence of exposure and outcome of interest were acquired by subjective methods, as was the case in many papers. However, reasonable steps had been taken across the included studies to provide independent validation of the results. In the selected studies, objective tools of measurement used included diagnostic tests such as the Video Head Impulse Test and caloric stimulation for the neuro-otologic examination of the vestibulo-ocular reflex. Therefore, although some studies used subjective measurements, they combined more than one test to validate their results. This reduced the level of bias. There was, however, as noted above, significant heterogeneity.
With respect to patient selection, the studies investigated specific groups; any control groups were appropriately matched. In one of the studies, Balatkova et al.,Reference Balatkova, Cada, Hruba, Komarc and Cerny12 the selection was not strictly random due to ethical concerns. Patients with serviceable hearing were preferentially assigned to the control group, and those with large tumours that surpassed Koos grade 4 (defined by a large tumour with brainstem and cranial nerve displacement) had the choice to willingly undergo gentamicin treatment because the probability of hearing preservation for them was low. In addition, the sample sizes for all the studies are relatively small, which can possibly result in a type II error and limit the randomisation process. As such, there was probably a low associated risk of bias.
With regard to the risk of bias linked to the matching of cases and controls according to important prognostic variables, there was statistical adjustment carried out for important prognostic variables across all the studies included in this review; hence, in this domain the risk of bias was low for all the studies.
Discussion
Main outcome
This review systematically assessed the effect of vestibular pre-habilitation with intratympanic gentamicin in vestibular schwannoma patients post resection. Eight studies met the criteria and were included. While the outcome measures varied, the results in four of the eight studies show that there is a statistically better post-operative performance in the gentamicin group than in the control group. The remaining four studies still demonstrated better outcomes but without reaching the level of statistical significance. As such, the studies showed that there is either a beneficial effect (four studies with statistical and four without statistical significance), or no negative effect without associated side effects.
It has been well demonstrated that a major post-operative concern in vestibular schwannomas is post-operative vestibular symptoms, which incapacitate the patient by prolonging recovery times. However, by employing vestibular ablation with intratympanic gentamicin prior to resection, it is possible to avoid this. We recognise that stronger evidence, conducting studies with larger cohorts and standardised outcome measures, is still required; however, the existing evidence, at least partially, supports intratympanic gentamicin as pre-habilitation in patients with vestibular schwannomas undergoing resection. Given the number and the limitations of the existing studies, this outcome should be implemented in clinical practice with caution.
Special considerations
There are a few potential limitations relating to the risk of bias in the included studies. The outcome measures used in the included studies involved a variety of questionnaires and subjective tests. As the studies concerned used other questionnaires and tests to supplement the outcome measures listed earlier, the risk of bias is mitigated. In terms of selection bias, the sample sizes for all the studies are relatively small, which can have an effect on the observed outcomes. However, a moderating factor is that there was statistical adjustment carried out for important prognostic variables across all the studies included in this review. Additionally, it is worth accepting that the number of vestibular schwannoma resections is relatively limited; thus, obtaining large numbers, involving control groups, can be challenging.
As seen in Table 1, most of the studies utilised retrosigmoid and/or translabyrinthine approaches for vestibular schwannoma patients, except Fellmann et al.Reference Fellmann, Bächinger, Dalbert, Röösli, Huber and Wettstein9 in which 10 per cent of patients underwent a transtemporal approach. None of the studies focused specifically on differences between the approaches and response to gentamicin pre-habilitation. However, given the documented outcomes, it appears that intratympanic gentamicin can help post-operative recovery regardless of the approach. While one could argue the potential hearing-preserving character of a retrosigmoid approach and any potential cochleotoxicity of gentamicin, this was not really assessed. Additionally, the effect of gentamicin on hearing was not specifically addressed in the included studies, because this was not the main focus of any of the papers that we examined.
Overall, we found that the included papers offered a good overview of how patients recover following vestibular schwannoma resection with and without pre-operative intratympanic gentamicin. All studies demonstrated good post-operative outcomes, with the main difference being speed of recovery, which was faster for the gentamicin group. There was significant heterogeneity of the outcome measures; thus, the term ‘better post-operative outcome’ cannot be interpreted uniformly throughout the available literature. Essentially, while most patients will recover eventually, the ones treated with pre-habilitation are more likely to recover sooner. This has been linked primarily to the gradual effect of gentamicin in ablating vestibular function, compared to the sudden effect of surgery through labyrinthectomy or vestibular nerve resection or a combination of both.Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö5,Reference Fellmann, Bächinger, Dalbert, Röösli, Huber and Wettstein9,Reference Magnusson, Kahlon, Karlberg, Lindberg and Siesjö10,Reference Tjernström, Fransson, Hafström and Magnusson18 The work by Amiraraghi et al.Reference Amiraraghi, Gaggini, Crowther, Locke, Taylor and Kontorinis16 showed, using Video Head Impulse Tests, the milder effect of surgery on the contralateral side in patients from the gentamicin group, providing a reasonable explanation of the positive effect of pre-habilitation.
Limitations of evidence
Most limitations were related to the study design and primarily to the small size of the included cohorts. This was addressed through the risk of bias assessment. A few aspects of clinical significance are worth noting. The gentamicin regimen is not homogeneous across the studies (Table 2). The concentrations, number of injections and intervals between each dose vary significantly. In addition, one of the studies, Tjernström et al.,Reference Tjernström, Fransson, Kahlon, Karlberg, Lindberg and Siesjö13 did not specify an interval between injections or injections and surgery. This is not necessarily a limitation because no adverse effects were reported and the goal was to ablate vestibular function, which the studies did successfully, regardless of the regimen used. Many studies primarily used the Video Head Impulse Test as a screening tool and to assess vestibular function after administration of gentamicin injections; it was also used following surgery in the study by Amiraraghi et al.Reference Amiraraghi, Gaggini, Crowther, Locke, Taylor and Kontorinis16 Four studies additionally used caloric testing, which has the limitation of assessing only the lateral semicircular canal; the Video Head Impulse Test has been shown to be more capable than caloric testing in assessing the function of each of the six canals individually and has been used more frequently since 2009.Reference Halmagyi, Chen, MacDougall, Weber, McGarvie and Curthoys19 However, as the studies used both caloric stimulation and the Video Head Impulse Test, the accuracy of the results was not compromised.
Implications of results for practice, policy and future research
Four of the eight studies demonstrated clinically significant statistical differences between the gentamicin group and the control group. The remaining studies showed, although not statistically significant, evident benefits post-operatively when vestibular schwannoma patients were pre-habilitated with intratympanic gentamicin. Three of the eight studies even indicate long-term benefits using vibratory posturography data. While we accept the limitations of the existing studies, it appears that ablation of vestibular function achieves compensation before surgery and speeds up recovery. There is some risk of hearing loss with intratympanic gentamicin injections,Reference Tarnutzer, Bockisch, Buffone, Huber, Wettstein and Weber20 but this risk must be weighed against the benefit of recovery. This is especially so if patients already have significant hearing loss or are undergoing translabyrinthine surgical resection of their vestibular schwannomas, because the gentamicin would not add any additional morbidity, particularly if no hearing-preservation surgery is being considered. Nevertheless, if hearing-preserving vestibular schwannoma resection is considered, the risk of hearing loss due to gentamicin should be thoroughly considered; this could potentially be a contra-indication for pre-habilitation with gentamicin.
Conclusion
Given the published evidence, the results support the rationale of pre-habilitating patients with vestibular schwannomas with intratympanic gentamicin prior to surgical resection. This conclusion should be adopted with caution in clinical practice because more prospective studies and larger sample sizes are needed to support it and strengthen the developing evidence.
Acknowledgement
The authors would like to greatly thank the Librarians of our Library Network for their help with a comprehensive literature search.
Competing interests
None declared
Appendix 1. Risk of bias assessment tool
