Quality of life outcomes in acoustic neuroma: systematic review (2000–2021)

The majority of acoustic neuroma (AN) outcome studies examine medical outcomes. An alternative is to examine how AN, its treatment, and complications impact patients’ quality of life (QOL). A systematic review for AN was conducted using MEDLINE, PUBMED, and EMBASE. Manuscripts were limited to human participants, written in English, and published from 2000 to 2021. Of 426 unique citations, only 48 examined QOL outcomes. Risk of bias was assessed using Downs and Black’s Quality Assessment Index. Sixteen studies (33.3%) compared AN patients to normative/control data, 15 (31.25%) compared QOL outcomes between treatment groups, 8 (16.7%) examined changes in QOL over time within a treatment modality, 7 (14.6%) examined QOL in relation to a specific sample characteristic, and 2 (4.2%) used non-standard methods to describe patients experiences. QOL was worse post-surgery and/or radiotherapy compared with healthy controls and active surveillance and tied to symptoms experienced. Study quality was reasonable despite expectable limitations due to the nature of the population. AN treatments, particularly surgical, may result in pain and nerve damage. Whilst not common, these can significantly impact QOL domains, including social and emotional wellbeing and social participation. Surgical teams should prepare patients for this possibility and ensure adequate community follow-up with specialist physical therapies, audiology, ophthalmology, and psychology to mitigate these effects should they occur.

there is little postoperatively recovery time [8,10]. However, there are also short-and long-term risks of SRS/ SRT including radiation toxicity, which can damage neural structures resulting in facial weakness, facial numbness or dysesthesia, and impaired hearing and balance [11][12][13]. Hydrocephalus can also occur as a late complication post-radiosurgery and may require ventricular shunting [14]. Conservative management (i.e. active surveillance) involves serial imaging to monitor growth and is seen as suitable in patients with smaller tumours (< 20 mm), in older patients, or those with comorbidities that preclude more invasive treatment [15][16][17][18]. This treatment protocol is based on ANs slow growth rate and static presentation in many patients with minimal symptoms [19,20].
Most AN outcome studies focus on medical outcomes such as cranial nerve function and recurrence rates. Whilst important, these do not reflect how AN diagnosis, or its treatment/s, impact the lives of those affected. An alternative approach is to examine impact upon patients' quality of life (QOL). The World Health Organization's defines QOL as an individual's perception of their position in life in the context of their own culture and value systems which spans physical health, psychological state, level of independence, social relationships, and their relationships to their environment [21].
In a systematic review on QOL outcomes of AN, Gauden et al. [22] concluded, based upon 47 studies published until 2010, that (1) patients with AN experience reduced QOL compared to population norms, (2) those with larger tumours may experience lower QOL than those with smaller tumours, (3) surgery may improve QOL in patients with large tumours after a temporary decline in physical dimensions of QOL and, (4) conservative treatment and radiosurgery have no effect on QOL. Furthermore, comparison indicated that whilst QOL did not differ significantly between those who had surgery and conservative management, more mixed results were found when comparing surgical and SRS treatment, though these results were confounded by differential tumour size.
A further systematic review [4] compared QOL in relation to treatment options and control populations in 39 studies . Microsurgery in retrospective studies had a negative effect on QOL, but this improved in prospective studies, though never to the degree of controls. Most studies suggest that SRT had little or no negative impact on QOL. The authors note few studies allowed "fair" comparison between treatments as the microsurgical group had immediate morbidity from surgery and recovery, whilst SRT has insignificant immediate effects but may develop toxicities over time. Importantly, allocation to treatment takes into account tumour size and patient demographics and wishes, introducing bias.

Aim
Surgical and non-surgical AN treatments have continued to improve over the last 10 years. The aim of this paper was to provide an updated systematic review of literature on QOL outcomes in patients with AN, with a focus on recent publications (i.e. .

Search strategy
This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The systematic review involved searches of databases MEDLINE, PUBMED, and EMBASE with search term Acoustic Neuroma as a keyword. Inclusion criteria comprised adults with AN where QOL was a study's outcome variables. The search was limited to publications with human participants, original research (both observational and experimental), in English, and published from 2000 to 2021.
Four hundred twenty-six unique citations were identified after duplicates were removed (see Fig. 1). Titles and abstracts of each citation was screened by the first author. Of those remaining, a further148 were removed as the abstracts indicated the papers did not measure QoL, leaving 284 papers. Review of full manuscripts revealed that only 34 papers referred to QOL as an outcome variable. After reviewing reference lists of these 34 papers, an additional 14 papers not identified in the original search (as they contained Schwannoma rather than AN as key words) were identified which met inclusion criteria, leading to final inclusion of 48 papers. Any conflicts that arose during eligibility assessment were resolved through consensus with the authors. Each of the included papers obtained participant consent and was approved by an appropriate ethics committee.

Risk of bias
Risk of bias was assessed using Downs and Black's [23] Quality Assessment Index which was designed to assess quality in randomised and non-randomised studies of health care interventions. Scores range from 0 to 31 with higher scores indicating better quality. Scores are calculated from nine items on reporting, three on external validity, seven on bias in measurement, six on bias in subject selection, and one on statistical power. Scores are described as "excellent" (score 24-28), "good" (score [19][20][21][22][23], "fair" (14)(15)(16)(17)(18), or "poor" (≤ 13).

Main findings
Whilst some overlap between groups was present, overall, 16 studies (33.3%) made comparisons of AN patients  Table 1.
Other measures related to QOL included the Beck Depression Inventory, Rosenberg Self-Esteem Scale and a visual analogue rating of QOL used in one study; the European Quality of Life-5 dimensions in one study; and the Zung Self-Rating Anxiety and Depression scales that were used in another study. Four studies did not use standardised assessment but reflected on QOL using a survey of patients' overall experiences [25], conducted thematic analysis of patient focus group transcripts [70], invited responses to questions tied to the WHO framework of impairments, disability and Handicap [71], or administered a survey developed to capture post-operative QOL [31].

Comparison to norms/controls
Of the 16 studies which compared AN to a non-AN comparison group, the AN group comprised post-surgical patients in 12 studies (one limited to post-surgical vertigo and one to translabyrinthine approach), active surveillance patients in one study, and mixed treatment groups in the remaining studies (1 = surgical vs SRS, 2 = surgical vs SRS vs active surveillance). The majority (10/16) of studies examined AN in relation to normative data, with one comparing to three different normative data sets. An additional study also compared to three normative data sets alongside two sets of surgical data. Of the remaining studies, one compared AN to age-and gender-matched controls, one to a non-patient control group who visited the same hospital, and the remaining three studies made  Sig lower scores than all three normative data sets for roleemotional, physical functioning, social functioning, and vitality Sig lower than one study on body pain and 2 studies for mental health Compared to surgical norms: Lower vitality compared to both surgical groups, whilst comparison to one study also showed higher body pain, but worse mental health, physical function, and social functioning In the only study to compare active surveillance to ageand gender-matched controls (patients at the same clinic with similar symptoms but not AN), QOL scores were comparable [46]. Two additional studies included active surveillance participants, one of which [43] also reported no difference between this group and controls. The final study did not report comparisons based upon treatment group [45].
Of those studies which included various treatment groups, Kelleher [43] found worse QOL in those who underwent surgery than norms for social function and physical role limitations (SRS group was too small to conduct analyses). Myrseth [44] also found significantly lower SF-36 social functioning and physical role limitation scores than norms, with deviations below norms significantly greater for surgery than SRS for physical functioning, physical role limitations, and emotional role limitations. Overall, GBI score and scale scores for general and psychosocial health were significantly lower post-surgery than post-SRS. As noted previously, Myrseth [45] did not report findings by treatment group.
Of the remaining studies, which contrasted post-surgical AN patients to normative data, the most consistent finding was of patients having lower QOL than normative data, often across all SF-36 scales [36,40]. Of particular note, physical role limitations were lower than controls in five studies [31-33, 39, 42], and in comparison to two out of three normative data sets by a final study [38]. A few studies also reported patients producing higher SF-36 scores than norms, particularly for vitality [34,35,38], emotional role limitations [31,38], and general health [35,38]; one particular study reported all three [38].

Links to patient characteristics
Of the seven studies that linked QOL to specific patient characteristics, two described participants in relation to membership in local AN associations, two studies examined QOL in those under active surveillance, and the remaining three examined QOL in relation to symptom outcomes and demographics post-surgery.
In the two studies reporting on QOL in members of local AN associations, Prummer et al. [24] report that members were younger, had larger tumours, included more women, were more likely to have had surgery or radiation, and to report symptoms (hearing loss, tinnitus, dizziness, headache, facial paralysis) than non-members. As would then be expected, members had significantly lower scores on the PANQOL. In a description of QOL in British AN association members, Broomfield and colleagues [25] found that the most common symptoms at diagnosis were hearing loss, tinnitus, and imbalance. In asking about overall life quality related to treatment, the largest proportion felt a little worse, followed by those who were unchanged, a lot worse, a lot better, and a little better.
In the studies of active surveillance, Oddon et al. [26] examined the impact of vertigo or dizziness on QOL and found poorer QOL on SF-36 (Social Functioning, Role-Emotional) and PANQOL scale (Balance, Energy) when either of these was present. Lloyd [30] also found that dizziness alongside increased age predicted physical QOL in those under active surveillance. Vertigo was also linked to reduced QOL in the treatment comparison study by Breivik [52].
In the remaining three studies looking at QOL within surgically treated groups, all authors reported worse QOL than norms [27][28][29]. Al-Shudifat [27] found that age and gender did not impact QOL, but tumour size correlated significantly with physical and mental aspects of QOL, whilst Lasseletta [29] linked worse QOL to postoperative pain but not facial functioning.

QOL change within treatment modality
Eight studies examined changes in QOL over time within a particular treatment modality; two examined changes from diagnosis to follow-up in those under active surveillance (one included those treated during follow-up), two examined changes with surgical intervention, and four examined changes pre versus post-SRS.
The two studies of active surveillance had contrasting results. Godefroy et al. [51] reported non-significant decreases in QOL to follow-up except for social function which had a non-significant improvement; whereas Breivick et al. [52] reported slight but significant improvements in vitality and emotional role limitation over time but reductions in social functioning. This later study also reported reduced QOL was related to presence of vertigo and tinnitus.
In those who underwent surgery, Fahy et al. [53] report no statistically significant association pre versus post-surgery or in relation to changes in tinnitus. Turel [47] found reductions in all areas of QOL (compared to norms) preoperatively, followed by improvements in these at 1 year and again at 2 years post-surgery.
Across the four studies of changes in QOL with SRS treatment, two [48,50,54] report QOL similar to norms, with no changes up to 18 months post-treatment [54]. One of these specified no relation of QOL to age, gender, tumour size, or radiation dose but a significant relationship between auditory and vestibular symptoms and reduced QOL [48]. The remaining paper [49] found increased mental health issues between diagnosis and pre-treatment, which were not present post-treatment, and an increase in pain related QOL and decreases in general health and social functioning from pre to post-SRS.

Comparison between treatments
Of the 15 studies that compared QOL outcomes between treatment modalities, six compared surgery, SRT, and/ or SRS and active surveillance; two compared surgery and SRS; two compared surgery and active surveillance; one compared surgery, SRS or SRT, surgery +SRT, and active surveillance; and one compared surgery, active surveillance, SRS, active surveillance+SRS, and active surveillance+surgery. Finally, two studies compared translabyrinthine and retrosigmoid surgical approaches, and the final study compared four approaches to facial nerve repair. Examining these last three studies, on the SF-36, whilst one study reported no differences in QOL in relation to surgical approach [66], the other reported that retrosigmoid surgery resulted in significantly higher social functioning, role limitations due to emotional issues and mental health, whilst translabyrinthine surgery was associated with better outcomes in terms of body pain and vitality [64]. Whilst Lin [67] reported only two significant findings, with patients who underwent end-to-side interposition nerve repair reporting significantly better QOL than those in end-to-end group for physical function and role limitations due to physical difficulties.
One of the studies with surgery and active surveillance groups [55] did not look at QOL post-treatment but whether QOL impacted treatment decision making, finding no significant association with QOL, voicerelated QOL, depression, or self-esteem. Of those studies which contrasted QOL following different treatments, four reported no significant differences. All four studies included surgical and SRS/SRT groups, with three including active surveillance and one also including surgery followed by SRS. The one study to contrast SRS to SRT [69] noted no significant differences in QOL. In the only study which included those diagnosed but with treatment not allocated as a group [68], those in active surveillance had the best QOL. Of note, those who had undergone both surgery and SRS and those who had not yet had treatment allocated had the worst QOL scores.
In the remaining studies, there was a trend towards significantly worse QOL post-surgery compared to active surveillance, whose QOL was maintained over time [63]. Another pre-post comparison study similarly reported that whilst QOL did not change for active surveillance and SRT groups; post-surgery, there was a significant improvement in QOL [59]. This difference possibly reflects use of different measurement tools (SF-36 vs GBI) and in timing of measurement with the latter being assessed on average < 3 years post, whilst the former had a delay of up to 15-years post-treatment.
The remaining studies had mixed findings, though active surveillance generally had better outcomes. Pollock [12] reports that only surgery related to significant reductions in QOL scores at 3 (Physical function, Role-Physical, Body pain) and 12 months (Physical and Body pain), whilst there was no change in those who had SRS. Robinette [57] reported that those post-SRT had significantly better QOL than those post-surgery or remained in active surveillance, but that QOL in the SRT group reduced as time elapsed post-treatment. In contrast, Sandooram [63] reported that those undergoing SRS had similar QOL to those undergoing surgery (whose QOL reduced post-treatment), which was poorer QOL than those in active surveillance whose QOL remained about the same.

Alternative QOL measurement
The final two studies examined QOL in AN samples using unstandardized methods of measurement. Brooker et al. [70] conducted the only qualitative study which spoke specifically to QOL. The thematic analysis of focus groups of patients with various treatments (surgery with/ without SRS/SRT and active surveillance) and from 1 to 11 years post-treatment revealed that individuals were most concerned about physical symptoms (hearing loss, balance, fatigue, tinnitus, facial paralysis); they presented with anger/frustration, depression, anxiety/uncertainty, and issues with body image but also reflected gratitude, relief, and psychological benefit associated with less severe physical symptoms and having a 'good treatment outcome' . Social interactions were seen as being impacted by balance and hearing impairments and social factors were important to overall adjustment.
In contrast, Bateman and colleagues [71] used a questionnaire to ask about impairment, handicap, and disability as defined by the WHO framework 1-3 years post-surgery. In relation to handicap, social isolation emerged as a strong theme, with over one third of individuals expressing reluctance to attend large social gatherings, and a smaller proportion finding impacts on employment outcomes important.

Study quality
As can be seen in Table 1, the Quality Index (QI) for the studies ranged from 9 to 24 with an average QI of 17.81 across all of the studies. One study was of excellent quality, whilst the largest proportion of studies were of good quality (n = 26; 54.17%), 16 (33.3%) were of fair quality, and the remaining five (10.42%) were of poor quality. As might be expected given differences in design, studies within different areas produced different QIs. The lowest mean QI (12.0; poor range) was produced by the studies which used alternative approaches to measurement of QOL. This is not unexpected given that both studies were exploratory in nature, did not contain control groups, and did not group participants in terms of either treatment received or explore confounders. This was followed by studies looking at sample characteristics (15.14; fair), comparisons to norms (18.38; fair to good), changes within a treatment (19.25; good), and comparisons between treatments (20.07; good). Characteristics of the population contributed to lowered scores across study types. For example, as would be expected, none of the studies provided power calculations (which contributes 5 points to the total), involved randomisation to treatments, or blinding of participants to treatment. Across the board, studies had good levels of reporting, provided good descriptions of their samples and treatments of interest, used appropriate statistics, and showed no evidence of data dredging.

Discussion
This study demonstrates continued interest in AN outcomes beyond physical symptoms but also QOL and provides an update to previous systematic reviews [4,22]. Study quality was generally good despite expected limitations due to the nature of the population, which does not allow randomisation or blinding or participants. To summarise the present findings: (1) when compared to normative controls, QOL of those with AN is similar if they are under active surveillance. This group typically produced significantly better QOL scores than those who underwent surgery, and there was some indication of improvements in QOL from diagnosis to follow-up.
(2) Whilst the greatest proportion of studies comparing QOL between treatment modalities found no significant differences, across study types, those who underwent surgery tended to produce worse QOL than normative data, or those undergoing SRT/SRS or active surveillance, and also show reductions in QOL pre-to post-surgery which were not present in those undergoing SRT/SRS or active surveillance. Evidence for changes in QOL following SRT/SRS was mixed. (3) There was consistency across study types that presence of vertigo and dizziness negatively impacts on QOL regardless of treatment received and that pain (whilst uncommon) also reduces QOL post-surgery; however, the findings for demographic factors such as age, gender, and tumour characteristics were variable across studies. (5) Those who belong to local AN associations may not be representative of AN as a whole, being less likely to remain in active surveillance and more likely to report symptoms and have poorer QOL.

Quality of life and AN
Whilst active AN treatments can have positive outcomes, they can result in a wide range of long-term deficits including reduced hearing, vision, facial movements, vertigo, dizziness and imbalance, and pain. Whilst these are uncommon, for those that are impacted, they significantly impact on self-perception, identity, well-being, social participation, and QOL. As noted above, when compared to normative controls, QOL of those with AN is similar if they are under active surveillance, whilst those who undergo surgery tend to produce lower QOL scores. This is not unexpected given these individuals have undergone major surgical intervention and often a lengthy recovery. Patients should be prepared prior to treatment initiation for the possibility of these symptoms and supported postoperatively should they arise. It should be noted here that in addition to the possibility of major surgery with potential for complications driving reduced QOL, the factors contributing clinical treatment decisions will likely have resulted in between group differences pre-treatment. For example, those in active surveillance tend to have smaller tumours and fewer symptoms than those who undergo surgery. In one study reviewed, Breivik et al. [52] followed 186 patients under active surveillance, 74 of whom underwent treatment during the follow-up period, reporting that treatment resulted in small but significant worsening in QOL related to physical functioning and emotional role limitations. Sandooran [63] found surgery after surveillance led to QOL scores similar to surgery without an initial surveillance period, whilst in contrast delayed SRS had better QOL outcomes compared to SRS.

Treatment differences
Whilst differences between treatments, particularly between surgery and active surveillance, are present, the findings are inconsistent. Those who underwent surgery tended to produce worse QOL scores than other group (i.e. norms, SRT/SRS, surveillance) and also show reductions in QOL pre-to post-surgery. It is important to recognise the difference in patient diagnostics, characteristics, and comorbidities across groups. For example, decision-making between treatment modalities needs to consider long-term benefits and risks as well as posttreatment QOL. Whilst those with active surveillance were spared worsening QOL, as is noted above, this group differs significantly, and up to 40% of patients will eventually require surgery or other active treatment [47]. Indeed, the most consistent finding was that, regardless of treatment received (i.e. surgical, SRS, surveillance), the presence of auditory and vestibular symptoms negatively impacted QOL.

Clinical implications
For many, active AN treatments are necessary and result in good physical and QOL outcomes. Whilst validated QOL questionnaires offer useful normative comparisons, more qualitative research is needed to provide an in-depth, richer understanding of the post-treatment services needed by AN patients who have residual deficits. Brooker et al. [70], for example, identified a number of factors not covered by typical QOL measures, such as presence of anger/frustration, body image, gratitude/relief, and impact of perceptions of treatment outcomes. In addition, Prummer and colleagues [24] suggest patients do not access peer support in the UK unless they have significant impairments, with a trend towards younger patients with larger tumours and more physical symptoms seeking support. It is well established in cancer care that treatment burden, outcomes and QOL can be improved by social supports [72]. Peer support, either 1:1, support groups, online support, or national associations can offer social, emotional, and practical assistance and contribute to improved QOL long-term. This systematic review suggests patients should be offered peer support to maximise QOL post-diagnosis and treatment.

Limitations and future directions
Whilst a systematic database search was conducted, it is possible that some papers of relevance were missed. Whilst we accepted all observational and experimental studies available with a QOL outcome, our ability to draw conclusions on QOL between treatment modalities is limited by the non-randomisation to treatments due to clinicians' need to make clinical decisions for patient's best outcomes. Future studies should focus on qualitative studies asking patient opinions on their QOL, its relationship to physical symptoms, and what support they would like from healthcare providers. Also, quality of the research could be improved by blinding assessors to treatment allocation, which was reported by only two studies. 31,61 Whilst the impact of potential confounders (age, tumour size) were addressed in many studies, as seen in Table 1, the length of follow-up was often variable. The impact of this potential confounder should be addressed in future either statistically or through initiating regular assessments post-diagnosis (e.g. alongside annual scans), regardless of treatment allocation. This would allow comparisons of patients diagnosed across time but with more uniform follow-up periods post-treatment.

Conclusions
Active AN treatments, whilst necessarily to preserve health, can leave some patients with long-term physical deficits. These can have a significant impact across a wide range of QOL domains beyond the physical impairment and into social and emotional wellbeing and participation in society. Surgical teams should adequately prepare patients for this prior to treatment and ensure patients receive adequately community follow-up in terms of specialist physical therapies, audiology, ophthalmology, and psychological follow-up as well as encouraging patients to join support groups and AN associations to learn from and engage the support of peers. Quality-of-life should be a consideration in treatment decisions in view of the differences across modalities.