Introduction

High-grade glioma (HGG) is the most common primary cancer manifestation of the central nervous system [1]. Current treatment for HGG includes neurosurgical resection or biopsy, radiotherapy and chemotherapy [2]. Without a prospect of cure, patients experience a gradual neurological and functional decline [3, 4]. Still, the HGG prognosis has improved and long-term survival has been reported [5] which stresses the necessity of increased attention on health-related quality of life (HRQOL). For that reason it is imperative to identify and understand the mechanisms that contribute to the improvement or deterioration of HRQOL [6]. Physical deterioration has a negative impact on HRQOL, leading to anxiety and depression [7–9]. Other contributory factors that affect the level of HRQOL in patients with HGG patients are gender, tumor localisation, histological classification and reduced neuro cognitive function [10]. Further, depression is the most independent predictor of HRQOL in patients with brain tumors [11]. Symptoms of depression are found to be part of the symptom cluster that includes fatigue, sleep disturbance and cognitive impairment [12]. A higher symptom burden is associated with lower HRQOL [12, 13]. The prevalence of depression is higher among patients with brain tumors compared with other cancer patient populations [14]. D’Angelo et al. found the prevalence of depression among brain tumor patients to increase up to 1 year after surgery [15]. However, there is lack of knowledge based on longitudinal self-reported HRQOL studies including data on prevalence and severity of anxiety and depression among patients with HGG. The purpose of this study was to evaluate the physical activity level, Karnofsky Performance Status (KPS), HRQOL, prevalence and severity of anxiety and depression in a longitudinal study following patients with HGG for 1 year from the time of diagnosis. We hypothesize that KPS and functional status will decrease leading to feelings of anxiety and cases of depressions and as a result the HRQOL is expected to decrease over time. Moreover, 1-year survivors are expected to have the highest baseline KPS. In case of tumour progression patients are expected to have a risk of depression.

Materials and methods

Study design

The present quantitative study is based on patient reported outcome (PROs) data and is part of a larger project using mixed methods [16]. In a mixed methodology design, qualitative and quantitative data are collected in parallel and analyzed separately [17–19]. The qualitative data consist of interviews with patients and caregivers carried out at the same five test time points as the quantitative data are collected [20], the methodology being described in full elsewhere [16].

Participants and procedures

This study includes persons ≥18 years, newly diagnosed with HGG (WHO classification grade III/IV), KPS ≥60 at baseline to ensure independent responses to questionnaires and able to speak and understand Danish. The primary investigator (KP) and the clinical specialists assessed whether patients with cognitive deficits were able to participate in the study. Disease progression was not an exclusion criterion, and no patients were excluded during the study period. Patients were asked for study participation on the first postoperative day, and recruited with a consecutive sampling strategy from May to December 2012 at the Department of Neurosurgery, Rigshospitalet University of Copenhagen. All the included participants formed the primary sample, from which participants who completed all study requirements for the entire study period formed a sub-sample that allows for a complete longitudinal analysis during a 1-year follow-up. Written consent was obtained within 2 days prior to discharge from the hospital. The study is registered at the Danish Data Protection Agency (2007-58-0015/30-0758) and The Ethical Committee at the Capital Region of Denmark (H-2-2013-135) and carried out in accordance with the Declaration of Helsinki [28].

Measurements

Socio-economic conditions, KPS, disease and treatment variables were obtained at baseline from medical records and PROs. Repeated measurements were collected at 5 time points: test 1 after surgery and diagnosis (week 1), test 2 during radiotherapy (week 6), test 3 and 4 after treatment response scans and chemotherapy (week 28 and 40) and test 5 after standard treatment and response scan (week 52). Patients received the following three validated questionnaires by mail at the five test time points (after response scan) with written instructions on how to complete and return their response in an enclosed and addressed envelope.

  1. (1)

    Leisure-time physical activity level [21–23]. Self-reported physical activity levels are rated using four response options: I Almost completely inactive: reading, TV watching, etc., II Some physical activity less than 3 h per week (h/weak): riding a bicycle or walking for pleasure, III Regular activity: heavy gardening, running, etc. for (3 h/weak, IV Regular hard physical training: running, soccer etc. for more than 4 h/weak. Additionally, at baseline, patients reported retrospectively their level of activity at 3 months prior to diagnosis.

  2. (2)

    The Hospital Anxiety and Depression Scale (HADS) [24, 25]. The HADS is a validated measurement tool for detecting symptoms of anxiety and depression in patients estimating psychological well-being. The questionnaire is divided into two sub-scales for anxiety and depression with seven questions each all; 14 questions are rated on a four point scale representing the degree of distress [0 = none, 4 = unbearable]. The responses from HADS are presented as separate scores for anxiety and depression higher scores indicating greater likelihood of depression or anxiety. Recommended cut-off values for both scales are as follows: below 7 is a normal score, scores of 8–10 indicate mild cases, scores between 11 and 15 reveal moderate cases, while 16 and above indicate severe cases of anxiety and/or depression. A score of 11 or more on either scale is considered a definite case of anxiety and/or depression.

  3. (3)

    The Functional Assessment of Cancer Therapy, General and Brain (FACT-G and FACT-Br) [26, 27]. The FACT-G is a validated multidimensional questionnaire that measures HRQOL in cancer patients. FACT-Br (version 4) is a 50-item measure that includes the 27 items from FACT-G and the brain sub-scale (23 items) to assess HRQOL in brain tumor patients. It comprises 4 sub-scales: physical well-being (PWB), social/family well-being (SWB), emotional well-being (EWB) and functional well-being (FWB), on a 5-point Likert scale ranging from 0 (not at all) to 4 (very much). Moreover, the brain cancer sub-scale (BrCS) reflects issues specific for brain cancer. The highest possible score is 28 for the PWB, SWB and FWB subscales, 24 for the EWB subscale and 184 for the BrCS. FACT-Br total score is the sum of the 4 sub-scales (of the core instrument). The FACT-G total score provides a summary of the overall HRQOL across the group of patients [range 0–108] and the FACT-Br Trial Outcome Index [range 0–132] is a summary index of physical/functional outcomes. The higher the score, the better HRQOL.

Procedures for data entry and audit program were established to ensure quality control in manual keying of data [28, 29].

Statistical analyses

The KPS, the HADS sub-scales, the FACT-Br sub-scales and the responses to the ordinal items of the leisure time physical activity scale were analysed separately. Categorical variables are reported as frequencies and percentages, while continuous variables are reported as mean and standard deviations (s.d.) using a significance level of p < 0.05. HADS and FACT-Br data refer to the normative values [30, 31].

For the sub-sample, differences between baseline mean scores (anxiety, depression, FACT-Br sub-scales, FACT Br Trial Outcome Index) and 1-year follow-up assessments were examined for significance (p < 0.05) using paired t-tests. Further, plots of HADS and FACT-Br (n = 30 and n = 18) are presented as supplementary material.

We estimated rank correlations for the primary sample between the FACT-Br sub-scales, the sub-scales of the HADS, age, occurrence of WHO grade III tumour, and leisure time physical activity with linear mixed models using standardized dependent and independent variables providing an estimate of the correlation that takes the clustered nature of the data into account. The statistical analysis was performed with SAS statistical software, version 9.3.

Results

Sample

A total of thirty patients are included in the primary sample of the study (Fig. 1). All patients received radiotherapy and/or chemotherapy. Patients with KPS <60 were excluded [n = 35: aphasia (n = 13), severe cognitive impairment (n = 9) such as problems with memory, language and thinking [32], severe neurological impairment (n = 9) to a degree where inpatient treatment is required and psychotic/severe stress reactions (n = 4) requiring acute psychological treatment]. Participants who completed all study requirements during the 1-year study period form a sub-sample (n = 18). Medical and demographic characteristics for study patients are outlined in Table 1 according to the primary sample and the sub-group. The longitudinal response rate to the PRO questionnaires ranged from 60 to 97 %. Missing data were due to neurological decline or death.

Fig. 1
figure 1

Study Flow. KPS Karnofsky Performance Scale; HADS Hospital Anxiety and Depression Scale; FACT-Br the functional assessment of cancer therapy, brain cancer; PA physical activity

Full size image
Table 1 Medical and demographic characteristic for study patients
Full size table

Functional status and physical activity n = 30

At baseline, 93 % of the patients had a KPS ≥70 (Table 2).The most profound decline (20 %) occurred during oncological treatment (between test 2 and test 3), when only 63 % of the patients had a KPS ≥70. At test five 53 % of the patients still had a KPS ≥70, and 23 % a KPS ≥90 as compared to 63 % at baseline. Most patients (75 %) were physically active more than 3 h per week (h/weak) 3 months prior to the diagnosis, and 9 % kept the same level of physically at follow-up 1 year later (Table 3).

Table 2 Karnofsky performance status
Full size table
Table 3 Physical activity levels
Full size table

Symptoms of anxiety and depression n = 30

According to the cut-off points of HADS anxiety scale, eleven patients (39 %) reported moderate or severe symptoms of anxiety at baseline (Table 4). This proportion decreased over time and at test 4 and 5 only one patient exceeded the cut-off level for anxiety (≥11). This trend was also seen for the mean anxiety score that was highest at baseline (mean 8.5 ± 5.2) and lowest at test 4 (mean 4.4 ± 3.5).

Table 4 HADS: anxiety and depression
Full size table

Concerning depression, the HADS scale showed moderate depression in four patients at baseline. At test 4 and 5, only one patient reported moderate depression. Mean depression scores were highest at baseline (mean 6.1 ± 3.5) and lowest at test 5 (mean 3.7 ± 3.1). A total of eight out of the 30 patients (26.6 %) reported scores indicating depression at least at one test time point during the 1-year study period (Supplementary Fig. 1), 87.5 % of them having tumor progression (see Supplementary Fig. 1 and Table 4). Compared to the normative values [30] our data were higher for anxiety (8.5 ± 5.2) at baseline compared to the normative value (of 6.14 ± 3.76) and for depression (6.0 ± 3.9) at baseline and test 3 compared to the normative value (of 3.68 ± 3.07).

Health-related quality of life n = 30

FACT-Br scores at each time point are shown in Table 5 along with normative values of the general U.S. adult population (N = 1075) [31]. FACT-G Total was lowest at baseline (71.9 ± 14.3) and highest at test 5 (83.4 ± 17.1). The same trend was seen for all sub-scales except the SWB sub-scale and the Brain Cancer sub-scale that appeared unchanged. For the SWB sub-scale the patients scored higher at baseline compared to the normative US values (p < 0.0001), whereas for the PWB, EWB, FWB, and the FACT-G Total the patient scores were lower than the normative sample.

Table 5 The functional assessment of cancer therapy, brain cancer
Full size table

Sub-sample analysis n = 18

Analysis of the development throughout the entire 1-year study period of KPS, physical activity levels, HADS and FACT-Br scores was performed in the sub-sample group of patients (n = 18) who survived until study completion. The KPS baseline data show that all the patients had a KPS ≥70, which also was the case 5 weeks later during radiotherapy at test 2, and 78 % still had a KPS ≥70 at the end of the study (Table 2). They reported a decrease of physical activity, where 89 % were active less than 3 h/weak at end of study (Table 3).

Mean anxiety was highest at baseline (mean 7.8 ± 5.7) and lowest at test 4 (mean 4.5 ± 3.9) (Table 4). There was a decrease in mean anxiety from test 1 to test 5 (p = 0.0095). No significant change in depression scores from test 1 to test 5 was identified (p = 0.07). Four individuals (22.2 %) obtained scores indicating depression at least at one test time point during the 1-year study period (Supplementary Fig. 1).

The EWB subscale of the FACT-Br shows a significant increase (p = 0.0023) from baseline (13.8 ± 6.7) to test 5 (18.4 ± 4.5). No significant changes were identified for the other sub-scales (Table 5).

The impact of anxiety

The correlation analysis showed that level of anxiety was closely related with (1) overall HRQOL (FACT-G Total) [correlation −0.60 with 95 % confidence limits (Cl) at −0.74 to −0.45] and (2) emotional well-being (correlation −0.68 with 95 % Cl at −0.81 to −0.55). Less closely associations were established between the variables of (3) physical and functional outcomes (PWB, FWB, FACT-Br TOI), and (4) the sum of the four sub-scales of FACT-Br (FACT-Br-Total), (5) the leisure time physical activity level and (6) the performance status (KPS) (results not shown).

The impact of depression

The depression subscale of the HADS showed close correlations to (1) the FACT-Br Trial Outcome Index (correlation −0.55 with 95 % CI at −0.69 to −0.42), (2) the FACT-G [correlation −0.64 with 95 % CI at −0.78 to −0.50], and (3) the FACT-Br total score [correlation −0.60 with 95 % CI at −0.72 to −0.48]. Regarding the individual components of the FACT-Br the association was closely to functional well-being and less close to social well-being (results not shown).

Discussion

As expected we found that KPS and physical activity levels decreased during the study period. Nonetheless anxiety and emotional well-being improved over time for 1-year survivors of HGG, with a close mutual relationship. The hypothesis that the sub-group would have a higher KPS than the primary sample was confirmed. In case of tumour progression data indicate a high prevalence of depression. Newly operated HGG patients experience early functional limitations [33], as indicated by the observation that 71 % of our study participants were completely inactive or physically active less than 3 h/weak at this time. Physical activity level did not improve among the 1-year survivors. An early and ongoing functional decline occurs, especially for the primary sample, while the sub-group had a higher KPS value. A lower tumor grade did not explain this as four out of five patients diagnosed with a WHO grade III tumor belonged to the primary sample. Exercise behaviour is found to be a strong independent predictor of survival [34]. Nevertheless, only a limited number of international intervention studies have investigated the possible benefit of physical training among HGG patients [35]. Maintaining independence for as long as possible is a priority of the patients [20] and early physical rehabilitation is recommended [36], but additional research is warranted in order to clarify the value of training in this group of patients.

This study suggests that patients with HGG report emotional distress and anxiety during the post-surgical period. This is explained by the poor prognosis and an unpredictable future [20, 37, 38]. However, the individual variations show that some patients with baseline scores within the normal range experience depression at later test time points. Fox et al. found the prevalence of depression to be 95 % at a mean of 46 months after being diagnosed with HGG [12]. Kilbride et al. identified among 13 patients, four who suffered from anxiety and five who were depressed during the interval between surgery and radiotherapy [39]. Our study found that eight [26 %] out of 30 patients [the primary sample] and four [22 %] out of 18 patients [sub-sample] reported cut-off levels indicating depression at least at one time point. Further, impairment of the physical condition and fatigue [9, 40–43] causes severe limitations of daily activities and reduce QOL [44], being followed by distressful emotional and psychological reactions such as depression [45–49].

Finally, brain tumour histology and grade of malignancy affect the production and release of biological factors that cause depressive symptoms [50]. Our study showed strong associations between depression and the variables of physical and functional outcomes suggesting that patients are at risk for unrecognised depression throughout the trajectory. In contrast, our sub-group analysis [n = 18] suggests a decreasing tendency for depression over time [p = 0.07]. This indication is supported by studies identifying patients with HGG that report an acceptable QOL during the oncological treatments [51, 52]. Still, a few cases of moderate depressions are identified at each test time point. It is difficult to distinguish between depression, ‘understandable’ sadness [53] or vegetative symptoms of depression such as loss of interest, emotion and energy [54]. Systematic screening and quantification for levels of depression and anxiety is imperative [55] and recommended in this group of patients, especially in case of physical decline and tumour progression. However, instruments should not be used as a substitute for in-depth conversations but rather function as guidance for clinicians that facilitate the dialogue with the patient [56].

The present results obtained with the FACT-Br identified an improved emotional well-being throughout the HGG trajectory for the sub-sample who completed the 1-year of follow-up. There may be various explanations for this result. First, patients might adapt to their HGG illness in a way that facilitates control [5] or a better coping despite symptom progression [57]. Second, and similar to previous studies [58, 59] the SWB score was higher than the normative values at all test points showing that patients experience increased social support from friends and families, which might lead to an improvement of the HRQOL. Third, previous study confirms that hope is a source of strength that helps individuals to remain active [60, 61], brings purpose and meaning to life, encourages a positive attitude and improves psychological well-being [62]. EWB was related to the level of functional status [63], and the 1-year survivors had higher KPS values compared to the primary sample throughout the study, which could be the reason for the increase in the EWB. Our findings indicate that there is a need for interventions that improve symptom management as patients report a number of brain cancer related difficulties. It is possible that a higher level of HRQOL can be achieved and maintained among HGG patients through an early palliative approach with the aim to improve life planning [64].

Whether a prolonged survival time or improvement of HRQOL should be prioritized relies on individual preferences and needs [65, 66]. Therefore, it is crucial to involve the patients actively in the treatment decisions of depression as several therapeutical options can be provided [38]. Patient involvement has the potential to contribute to a better adaption process to the HGG illness [67]. Eventually, this can lead to a positive effect on HRQOL among terminally ill HGG patients [6].

Study limitations

The literature has identified differences in levels of anxiety, depression and HRQOL with respect to gender [68], age and KPS [10], while the role of tumor localisation is inconclusive [69–71]. Our study examined a limited number of relationships. Whether, age, gender, education and tumour progression are variables that may have an effect on the results cannot be estimated due to small sample size. Drop-out of patients was predominantly being due to neurological decline or death. A cognitive phenomenon known as response shift refers to a change in the meaning of the participants’ values and/or perception over a period of time [72]. Therefore subjects might change their response to PROs measures over time, not only because of change in the evaluated study items, but also because they might have changed their perception of the significance of HRQOL [73].

Conclusion

Patients with HGG experience a functional and physical decline, with a negative impact on the level of physical activity. Patients’ physical and functional outcomes were closely associated to the prevalence of depression. Nonetheless anxiety and emotional well-being were closely related and both improved over time for 1-year HGG survivors. Explanatory factors can be social support from friends and families and the higher performance status among the 1-year HGG survivors.