Introduction

Brain tumours not only give rise to a range of neurological and physical deficits but also may result in personality and behavioural changes. Thus, the social impact of brain tumour has particular significance. Difficulties with personality and social behaviour have been consistently reported following damage to the frontal lobes [1–4] including lack of insight into these personality changes. Relatives of patients with brain tumours report personality change to be one of the factors associated with poor quality of life [5, 6]. Although personality change is a well-recognised symptom of brain tumour, there are very few studies specifically concerned with the impact of a brain tumour on personality and social dysfunction in daily life and within the family.

Increasingly, evidence suggests that different anatomical regions of the frontal lobes mediate different aspects of social behaviour [7]. Lesions in the orbitofrontal cortex are associated with emotional lability and mood disturbances [8, 9]. Social awareness can also be significantly affected, with patients showing reduced concern about or insight into the consequences of their behaviour and/or reduced empathy for the impact of their behaviour on others [10]. The orbitofrontal cortex also appears important in adhering to social rules and conventions; lesions in this area have been associated with disinhibited and socially inappropriate behaviour [11]. The dorsolateral cortex plays an important role in mediating executive cognitive functions, such as working memory, inhibition, flexibility, problem-solving, planning, goal-setting, initiation and strategy generation [12, 13]. The medial prefrontal cortex is associated with motivational aspects of behaviour; lesions here can result in apathy [14].

Personality and behavioural change can have devastating functional consequences resulting in reduced personal autonomy, unemployment and/or divorce. Executive difficulties with organising, initiating, directing, monitoring and controlling interpersonal behaviour can make it difficult to function personally and professionally [15, 16]. Despite what can be drastic personality changes, patients with frontal tumours may be unaware that their behaviour has changed or is socially inappropriate [17, 10].

Aims

The primary aim of this study was to measure patients and relatives assessment of behavioural and personality changes following diagnosis of the brain tumour and whether these changes are more prominent in the experience of patients with brain tumours involving the frontal cortex and their relatives. Identifying which groups of patients and/or relatives are more likely to report significant changes in personality and behaviour can enable targeted provision of appropriate support.

As the frontal lobes play an important role in self-awareness and self-monitoring, it was also expected that patients with frontal tumours would show reduced insight into their levels of social cognitive functioning compared to their relatives.

Methods

Sixty-six adult patients with focal frontal brain tumours and their relatives and 60 adult patients with focal tumours not involving the frontal cortex and their relatives were asked to participate. All participants were recruited from the Joint Multi-Disciplinary Neuro-Oncology Clinic. The inclusion criteria were (i) aged 18 years or above, (ii) able to give consent and (iii) having undergone surgery to confirm diagnosis.

There were no exclusion criteria for relatives, the majority (73.4 %) of whom were the spouse/partner of the patient, 14.3 % were the mother/father, 4.1 % sister of the patient, 4.1 % the son, and 4.1 % the nonrelative carer. Demographic and clinical background information was obtained for patients from their medical records (see Table 1). The majority of patients in both the frontal and nonfrontal tumour groups had low-grade tumours (75 and 81 %, respectively). Lesion localisation was determined on the basis of MRI scan and report from the senior neuroradiologist. The temporal and parietal lobes were the most common locations for nonfrontal tumours, each diagnosed in 33.3 % of patients, and 11.1 % had temporal-parietal tumours. All except two of the nonfrontal brain tumour patients had undergone surgery. Thirteen frontal tumour patients had radiotherapy, 7 of these patients also received chemotherapy. Of the 13 nonfrontal tumour patients that had radiotherapy, one also underwent chemotherapy. All participants gave written informed consent prior to participating in the study.

Table 1 Demographic and clinical features of patients
Full size table

Patients were asked to complete the Frontal Systems Behaviour Scale (FrSBe) and the Hospital Anxiety and Depression Scale (HADS). Relatives were asked to complete only the FrSBe. Questionnaires were posted to patients and relatives in separate envelopes along with stamped addressed envelopes for the return of completed questionnaires.

The Frontal Systems Behaviour Scale [18] is a 46-item behaviour rating scale designed to identify and quantify behaviours frequently reported following damage to the frontal systems of the brain. It consists of a self-rating form for completion by the patient and a family rating form completed by a relative who knows the patient well. Respondents are asked to rate the current frequency of behaviours (e.g. (apathy) ‘Speaks only when spoken to’; (disinhibition) ‘Talks out of turn, interrupts others in conversations’; (executive dysfunction) ‘Cannot do two things at once’) according to a 5-point Likert-type scale, where 1 = almost never, 2 = seldom, 3 = sometimes, 4 = frequently and 5 = almost always. Each rating form generates a total score derived from three subscales measuring apathy, disinhibition and executive dysfunction. The FrSBe provides a measure of behaviour both before (i.e. premorbid) and after brain damage.

The FrSBe provides normative data for age, gender and education level for individual subscale scores and for the total score. Raw scores were converted to T scores (linear transformations of the scores obtained in the normative sample, such that the distribution of FrSBe scores has a mean of 50 and standard deviation of 10). T scores at or above 65 represent ‘clinically’ significant levels of symptomatology, and T scores of 64 or less indicate ‘nonclinical’ levels. T scores of between 60 and 64 may be interpreted as an indication of borderline levels of impairment. For the purpose of the current study, borderline scores were classified as ‘clinical’ as having more than two levels of classification would result in insufficient power for the statistical analyses.

The Hospital Anxiety and Depression Scale (HADS) [19], was given to measure depressed mood, as this has been found to be common in patients with brain tumours and because of the behavioural similarities between the syndromes of apathy and depression [20, 21]. Although less studied in cancer populations, anxiety has also been shown to be prevalent among oncology patients [22]. The HADS is a 14-item self-report questionnaire and consists of an anxiety subscale (HADS-A) and a depression subscale (HADS-D), each containing 7 intermingled items. Cut-off scores are provided, enabling levels of anxiety and depression to be classified as normal, mild, moderate or severe. Scores of <8 classified as ‘normal’, scores 8–10 as ‘mild’, scores of 11–14 as ‘moderate’ and scores of 15–21 as ‘severe’. For the purpose of the current study, normal scores were classified as ‘nonclinical’ and mild, moderate and severe scores as ‘clinical’, as having more than two levels of classification would result in insufficient power for the statistical analyses.

Response rates for patients with frontal tumours and their relatives were 42 and 36 %, respectively, and for patients with nonfrontal tumours and their relatives, their response rates were 45 and 42 %, respectively. Thus, a total of 28 patients with frontal tumours and 24 of their relatives completed the questionnaires, of which there were 23 complete patient-relative pairs. A total of 27 patients with tumours not involving the frontal cortex and 25 of their relatives completed the questionnaires, of which there were 22 patient-relative pairs.

Statistical analysis

Paired t tests were conducted to compare premorbid with post-illness overall ratings on the FrSBe for the frontal and nonfrontal patient groups and the respective relatives groups. Hochberg [23] corrections were applied to control for type I errors in multiple comparisons.

Post-illness

One-way, independent groups, analyses of variance (ANOVA) was conducted to compare levels of apathy, disinhibition, executive function and overall frontal systems-related behaviour reported by patients and their relatives. To examine whether frontal patients showed reduced insight, one-way independent groups (relatives of patients with frontal tumours vs. patients with frontal tumours and relatives of patients with nonfrontal tumours vs. patients with nonfrontal tumours) ANOVAs were conducted to compare FrSBe ratings by patients and by their relatives for each frontal systems-related behaviour. Only data obtained from complete ‘patient-relative pairs’ were included in these analyses. Clinical vs. nonclinical levels on all the measures of the FrSBe for the patients and relatives groups were assessed by 2 × 2 chi-square tests. One-way, independent groups, ANOVAs were performed to compare levels of depression and anxiety reported by patients with frontal tumours and those with nonfrontal tumours. In addition, normative test data comparisons were used to classify patients’ HADS scores as ‘clinical’ or ‘nonclinical’, and 2 × 2 chi-square tests were performed to compare the number of patients with frontal tumours with the number of patients with nonfrontal tumours reporting clinical levels of anxiety and depression.

Apart from the paired t tests (where post-illness overall FrSBe score was predicted to be higher than premorbid score for all groups), all statistical tests were two-sided with statistical significance set at the 0.05 level.

Results

Premorbid vs. post-illness FrSBe rating scores

As expected, the total FrSBe scores for the frontal and nonfrontal patient groups and both relatives groups were significantly higher post-illness (see Table 2).

Table 2 FrSBe total mean rating scores (SD) premorbid and post-illness
Full size table

Post-illness FrSBe rating scores

Patients with frontal tumours reported significantly higher levels than patients with nonfrontal tumours on all measures: apathy, F (1, 53) = 4.44, p = 0.04; disinhibition, F (1, 53) = 4.90, p = 0.031; executive dysfunction, F (1, 53) = 4.10, p = 0.048; and total score (overall frontal systems-related behaviour), F (1, 53) = 5.46, p = 0.023 (see Table 3).

Table 3 FrSBe mean rating scores (SD) for patients (frontal and nonfrontal) and relatives (frontal and nonfrontal)
Full size table

Relatives of patients with frontal tumours reported significantly higher levels of apathy than relatives of patients with nonfrontal tumours (F (1, 47) = 4.44, p = 0.04). On all other measures, there was no significant difference (see Table 3).

The comparisons of ‘patient-relative pairs’ showed no significant difference on any measure between the frontal patients group and their relatives or between nonfrontal patients group and their relatives.

Clinical levels of frontal systems related behaviour

The number and percentage of patients and relatives in each group reporting clinical or nonclinical levels of frontal systems-related behaviours according to their FrSBe ratings are presented in Table 4.

Table 4 Number and percentage of frontal and nonfrontal tumour patients and their respective relatives reporting clinical or nonclinical levels of frontal systems-related behaviours on the FrSBe
Full size table

There was a trend for more patients with frontal tumours than patients with nonfrontal tumours to report clinical levels of all behaviours measured, and this reached statistical significance for disinhibition: X 2 = 5.973, df = 1, and p = 0.015 (see Fig. 1).

Fig. 1
figure 1

Percentage of patients with clinical levels of apathy, disinhibition and executive dysfunction as rated by patients and relatives

Full size image

Although there was a trend for more relatives of patients with frontal tumours than relatives of patients with nonfrontal tumours reporting clinical levels of apathy and disinhibition, this did not reach statistical significance (see Fig. 1).

Although there was a trend for more patients with nonfrontal tumours than their relatives to report clinical levels of disinhibition, this did not reach statistical significance, and there were no significant differences between these patients and their relatives on the other measures.

Depression and anxiety scores

Patients with frontal tumours scored significantly higher for depression than patients with nonfrontal tumours (F (1, 52) = 6.35, p = 0.015). There was no significant difference in anxiety scores (see Table 5).

Table 5 HADS anxiety and depression scores (means and standard deviations) and number and percentage of patients classified as having clinical and nonclinical levels of depression and anxiety
Full size table

There was no significant difference in the number of patients with frontal tumours and patients with nonfrontal tumours reporting clinical levels of depression, but significantly more patients with frontal tumours than patients with nonfrontal tumours reported clinical levels of anxiety (X 2 = 8.927, df = 1, p = 0.003).

Discussion

Pre and post-illness changes in frontal systems-related behaviour were analysed. As expected, patients with frontal tumours and patients with nonfrontal tumours and their respective relatives report significantly higher levels of overall frontal-system related behaviour post-illness.

Post-illness, patients with frontal lobe tumours reported significantly higher levels of apathy, disinhibition and executive dysfunction than patients with nonfrontal brain tumours. Normative comparisons revealed an overall trend for a greater percentage of frontal than nonfrontal patients to report clinical levels of apathy, disinhibition and executive dysfunction, which reached statistical significance for disinhibition. These findings indicate that the subjective experience of difficulties with personality and behavioural change is greater for patients with frontal tumours than patients with nonfrontal tumours and support the evidence that the frontal lobes play a key role in mediating various aspects of social functioning [4, 11]. Furthermore, patients with frontal brain tumours had significantly higher clinical levels of disinhibition suggesting that these patients had more problems with inhibitory control.

Importantly, while patients with frontal tumours reported significantly higher levels of all behaviours, and a trend was observed for more clinical levels of all frontal systems-related behaviours in patients with frontal tumours than patients with nonfrontal tumours, clinical levels of apathy and executive dysfunction were also reported in at least 40 % of patients with nonfrontal tumours. Thus, for a substantial proportion of patients with brain tumours, there are changes in personality and behaviour, regardless of tumour location. This supports the view that although the prefrontal cortex has a critical role in social functioning, the integrity of other regions of the brain is necessary for optimal functioning [24]. However, as disinhibition was reported in a relatively small proportion (24 %) of patients with nonfrontal tumours, this suggests that disinhibition is more specifically associated with damage to frontal areas of the brain. This lends support to a recent study examining the neural correlates of socioemotional disinhibition and executive function in older patients with neurodegenerative disease which reported a specific association between orbitofrontal areas and disinhibition [25].

Regarding relatives’ subjective experience of personality and behavioural change following brain tumour, there was a trend for the relatives of patients with frontal tumours to report higher levels of all frontal systems-related behaviour than relatives of those with nonfrontal tumours. Interestingly, while this trend was significant for all behavioural measures for the frontal patient group, for the relatives of frontal patients, this was only significant for apathy. Although the behavioural symptoms of apathy and depression can be similar, apathy is best characterised as a disorder of motivation rather than mood state per se. Thus, the disturbance of mood (e.g. sadness) which is a predominant feature of depression is either absent or only a minor feature in frontal systems-related apathy [18]. While the frontal brain tumour patients had a higher overall score on the depression subscale of the HADS, there was no significant difference in the number of patients reporting clinical levels of depression between the two patient groups. Therefore, the higher levels of apathy reported by the relatives of the frontal brain tumour patients are less likley to merely reflect higher levels of depression in these patients.

Although there was an overall trend for more relatives of patients with frontal tumours than relatives of patients with nonfrontal tumours to report clinical levels of apathy and disinhibition, these differences did not reach statistical significance. However, clinical levels of frontal systems-related behaviours were reported by at least 40 % of all relatives, suggesting that the subjective experience of brain tumour for a substantial proportion of relatives involves clinically significant difficulties with personality and behavioural changes, regardless of whether tumour location was frontal or nonfrontal.

It was expected that patients with frontal tumours would show reduced insight into their levels of personality and behavioural changes compared to their relatives. However, no significant difference was found on any behavioural measure between ratings by patients with frontal tumours and their relatives or between patients with nonfrontal tumours and their relatives. One explanation could be that as reduced insight into one’s behaviour is considered to be particularly associated with the orbitofrontal cortex [26], it may be that only a small number of our frontal group had tumours involving the orbitofrontal region. Alternatively, given that we found disinhibition levels to be significantly higher in our frontal patients implicating orbitofrontal involvement, this would support the contention of precise and dissociable networks between distinct regions and specific distant brain regions [27]. In other words, different parts of the orbitofrontal cortex are differentially involved in insight and disinhibition.

Patients with frontal tumours reported significantly higher levels of depression, although the rate of clinical levels of depression was not significantly greater than for patients with nonfrontal tumours. This however may reflect the relatively small number of patients with clinical levels of depression in both groups (frontal patients n = 8; nonfrontal patients n = 4). Depression is an important complication of primary cerebral glioma and is associated with reduced quality of life [20]. Our rate of depression in patients with frontal tumours was similar to the overall patient-rated measures reported in the review by Rooney et al. [20] (i.e. 28.5 vs. 27 %, respectively). Of the ten studies that used the HADS depression subscale, the mean scores in these studies ranged from 3.15 to 6.2. The mean score for patients with frontal tumours in our study was 6.1 which is higher than all but one of the studies reported by Rooney et al. [20]. For our nonfrontal patients, the mean score on the HADS depression subscale was lower than all but one of the studies reported by Rooney et al. [20]. Although there is currently no consistent evidence that tumour location and depression are associated, the study by Wellisch et al. [28] reported that frontal lobe tumour location was independently associated with major depressive disorder using DSM-IV criteria. Our findings lend support to a possible relationship between tumour location and depression. However, larger studies investigating the frequency of depression in patients with primary brain tumours are required to elucidate the potential association of tumour location in the aetiology of brain tumour-associated depression.

Although we found no significant difference in the overall scores on the HADS anxiety subscale between the frontal and nonfrontal patient groups when the scores were classified as clinical and nonclinical, significantly more frontal tumour patients than nonfrontal tumour patients had clinical levels of anxiety. Indeed, over 71 % of the frontal tumour patients reported clinical levels of anxiety compared with 30.7 % of the nonfrontal tumour patients.

Very few studies have investigated the presence and characteristics of anxiety in patients with primary brain tumours. However, in a recent study, Arnold et al. [22] reported 48 % of their 363 patients with primary brain tumours had current generalized anxiety disorder according to the responses on the Modified Brief Patient Health Questionnaire. This rate of anxiety is concordant with the overall rate of anxiety of our combined frontal and nonfrontal brain tumour patients which was 51.9 %. Anxiety and depression have been found to be negatively associated with all aspects of quality of life [29]. The high rates of depression and anxiety in our patients with frontal brain tumours highlights the importance of assessing the presence of neuropsychiatric illness in patients with brain tumours. The routine use of screening instruments for depression and anxiety in patients with brain tumours would therefore inform more effective detection and appropriate treatment intervention. The HADS has been considered to be a more accurate screening measure of depression in glioma as it minimises somatic symptoms [21].

The neurobehavioural changes following brain tumour reported by a substantial percentage of both relatives of and patients with brain tumours in our study highlight the need for support and management for relatives as much as patients. As personality change is one of the factors associated with poor quality of life [5] and over 75 % of the patients in our study have low-grade tumours and therefore relatively long survival, potential treatment is particularly important. Neurobehavioural changes are challenging for relatives, and not understanding the nature of these changes has been associated with more difficulty coping with them [30]. As a first step, information regarding potential neurobehavioural changes including how to recognise, understand, predict and manage behaviours that are causing difficulties following brain tumour is likely to help both patients and their relatives “make sense” of these often challenging behaviours. In addition, better access to clinical neuropsychologists with expertise in cognitive impairment and neurobehavioural changes working within multidisciplinary neuro-oncology teams to provide support and to develop and implement coping strategies would help patients and their families manage the neurobehavioural sequelae of brain tumour and improve their quality of life.

There are several limitations to this study. First, the sample size is relatively small which restricts the extent to which the findings can be generalized. Second, given the small sample size and multiple analyses, there is an increased risk of type 1 errors, although Hochberg corrections were applied to paired t test analyses to control for type 1 errors in multiple comparisons. Third, as this was an exploratory study with a relatively small sample, we did not attempt to investigate the relationship between a range of variables including type and grade of tumour, education level and potential treatment effects on behavioural changes. Future research with larger samples to establish whether particular variables are associated with greater neurobehavioural change in patients with brain tumour is required to inform care provision.

Summary

Little attention has been given to the neurobehavioural changes in patients with brain tumours. These results indicate that the subjective experience of a substantial proportion of patients with frontal tumours as well as patients with nonfrontal brain tumours has clinical levels of difficulties with personality and behavioural changes, which impact on social functioning. The present study also reveals high rates of depression and anxiety in patients with primary brain tumours. Routine assessment of depression and anxiety will facilitate evidence-based management of depression and anxiety in patients with primary brain tumours and ultimately improve quality of life. With growing recognition on the impact brain tumours have on social functioning, it will be important to establish effective methods of care and support for these patients and their relatives.