Effect of a Home‐Based Exercise Program on Muscle Mass, Fatigue, and Health‐Related Quality of Life Among Patients with Oral Cancer: A Randomized Trial

ABSTRACT

Background

Patients with head and neck cancer (HNC) often experience deteriorated health‐related quality of life (HRQoL) and loss of body weight and muscle mass. The primary aim was to determine if a home‐based exercise program, aiming to increase physical activity (PA), could reduce muscle mass loss and fatigue.

Methods

Oral cancer patients were randomized to an exercise group (EG) or a control group (CG). The EG received a home‐based exercise program, and both groups were followed for 1 year. The Frändin/Grimby activity scale, bioelectrical impedance analysis, and HRQoL questionnaires were used to assess PA, muscle mass, and fatigue.

Results

In total, 119 patients were included. The EG increased their PA, but no differences in muscle mass were found between groups. A tendency toward less fatigue and better global QoL was observed in the EG.

Conclusion

Future studies are needed to explore the possible benefits of PA in patients with HNC.

1. Introduction

The most common tumor locations for head and neck cancer are the oral cavity and oropharynx. Patients with stage I and II oral cancer are often treated with surgery alone; in some cases, postoperative radiation is administered [1]. Stages III and IV oral cancer are advanced and generally treated with combination approaches such as surgery with or without reconstruction and postoperative radiation or chemoradiation. Surgical treatment can cause postoperative pain and functional impairments such as oral dysfunction, problems with chewing, and dysphagia. Common side effects of radiation include dry mouth, dysphagia, mucositis, pain, and trismus. Extensive tumor resection with reconstruction, visible scarring, loss of facial nerve function, and tooth loss can affect a patient's appearance and exert a negative psychological impact [2, 3, 4]. The tumor location and side effects from treatment, together with the disease itself, often cause impaired energy and nutrient intake, leading to weight loss and risk of developing malnutrition [5, 6, 7]. Weight loss in patients with HNC negatively influences survival, functional performance, and health‐related quality of life (HRQoL) [8, 9, 10, 11]. Loss of muscle mass is associated with the development of disease‐related sarcopenia and malnutrition, and studies in HNC patients treated with chemoradiotherapy (CRT) have revealed that lean body mass accounts for up to 70% of total weight loss [12, 13].

Physical activity (PA) during and after cancer treatment has been widely studied, especially among patients with breast, colon, or prostate cancer; the results have shown that PA during and after treatment can improve quality of life (QoL), fatigue, physical function, and survival [14, 15, 16, 17]. A meta‐analysis of 34 randomized controlled trials (RCTs) revealed that PA has a positive effect on HRQoL and physical functioning among cancer patients; furthermore, PA can also have a positive effect on muscle strength and muscle mass, but more studies are needed to confirm these effects [18]. Studies investigating the effects of PA interventions in HNC patients have several limitations, such as small sample sizes, heterogeneous study designs and outcome variables, and risk of bias [19]. HNC is a heterogeneous disease, and different treatment regimens are used depending on the tumor location and stage. To our knowledge, few studies have evaluated PA in patients with HNC at specific tumor sites [20, 21]. In a systematic review from 2021 in patients with HNC, 75% of the included studies reported that at least one outcome measure (most commonly aerobic capacity and fatigue) improved after a PA intervention [22]. To date, only a few RCTs have evaluated the effect of PA on muscle mass in patients with HNC [23, 24].

The hypothesis of this RCT was that an individualized home‐based exercise program, given by a physiotherapist, focused on increasing PA, would reduce muscle mass loss, reduce fatigue, and improve HRQoL among patients with oral cancer. Differences in change in fat‐free mass (FFM) and fatigue between the intervention and control group 4 months after inclusion were the primary outcomes. Secondary outcomes between the two groups at all measurement points were differences in self‐reported PA, change in FFM, weight change, prevalence of malnutrition, fatigue, and HRQoL.

2. Materials and Methods

2.1. Study Population and Study Design

Eligible patients were identified through the multidisciplinary tumor board, which is held weekly at Sahlgrenska University Hospital (Gothenburg, Sweden). The inclusion criteria were newly diagnosed carcinoma of the oral cavity with intended curative treatment and age ≥ 18 years. The exclusion criteria were previous treatment for HNC within the last 5 years, other cancers with ongoing treatments, inability to perform the exercise program, World Health Organization (WHO) performance status ≥ 2, cognitive impairments, or inability to understand the Swedish language.

All participants received oral and written information about the study and provided written informed consent to take part in the study. Patients who accepted were randomized to either the exercise group or the control group at a 1:1 ratio. The randomization was computer‐based and stratified by age, sex, body mass index (BMI), treatment regimen, tumor location, tumor stage, and WHO performance status. All procedures were performed in accordance with the Declaration of Helsinki, and the study was approved by the Regional Ethical Committee in Gothenburg, Sweden (registration number 103–18). Inclusion took place between August 2018 and December 2022. The study inclusion period was extended due to the COVID‐19 pandemic; for the same reason, enrollment in the study temporarily ceased between March and September 2020.

Measurements of PA and nutritional status were conducted by a registered dietitian at the time of enrollment and at 2, 4, 6, and 12 months after inclusion (Figure 1). At these time points, the patients also answered the HRQoL questionnaires. The dietitians who assessed the data were not blinded. The study follow‐ups were, if possible, planned according to the patients' clinical follow‐up visits to the Departments of Otorhinolaryngology or Oncology (Sahlgrenska University Hospital). At inclusion, the physiotherapist informed the patients in the intervention group about the exercise program during a face‐to‐face visit at the clinic. Follow‐up calls were then made by the physiotherapist 1 week and 1, 2, 4, and 6 months after enrollment in the study (Figure 1). If the patients did not answer despite repeated phone calls, they were contacted via text message (mobile).

FIGURE 1.

Timeline over the study period. Physiotherapist: One face‐to‐face visit at inclusion +5 telephone follow‐ups. Study assessment: Five visits with nutritional assessment, body composition measurement, and HRQoL questionnaires. Incl = Inclusion, m = months, w = weeks.

2.2. Treatment

All patients were treated according to national guidelines [25]. The first option of treatment for all patients was surgery, alone or with radiotherapy (RT). Larger tumors were treated with extensive surgery with or without reconstruction followed by RT, and nonresectable tumors were treated with RT or CRT if possible. RT was given with intensity‐modulated radiation therapy (IMRT) to a total tumor dose of 68 Gy, 6 fractions/week. Additional treatment when needed was concomitant chemotherapy (CT), weekly cisplatin, or antibody treatment with cetuximab if CT was contraindicated. All patients were under nutritional surveillance at the Departments of Otorhinolaryngology or Oncology according to the clinics' routine. The clinical dietitian individually met all patients at the time of diagnosis in connection with the multidisciplinary tumor board, and frequent clinical follow‐ups were performed according to the patients' nutritional needs. Nutritional treatment included individualized counseling, protein‐rich oral nutritional supplement (ONS), and enteral (EN) and/or parenteral nutrition (PN), which were primarily introduced reactively when indicated (e.g., when oral food intake was insufficient).

2.3. Intervention with Progressive Resistance Training and Aerobic Activity

At enrollment, a physiotherapist informed each patient in the intervention group about the aim of the intervention during a clinic visit. At the visit, instructions were given on how to perform four muscle strengthening exercises: squats, heel raises, hip thrusts, and push‐ups. Each exercise had four levels of difficulty, which were individualized; for example, the easiest level of push‐ups was performed while standing against a wall, and the hardest level was performed on the floor on toes. Patients performed 3 sets of 10–15 repetitions of the four strengthening exercises and were told to perform the resistance training and additional aerobic activity (chosen by the patient), of at least moderate intensity, for at least 30 min per day, preferably 5 to 7 days a week for a 6‐month period. This corresponds to the recommendation of at least 150 min of PA per week for both aerobic and muscle strengthening activity [26]. Patients who were already active were encouraged to increase the intensity and/or duration. The resistance training was not recommended to be performed during the first 2 weeks after surgery, but the patients were encouraged to take shorter walks at a leisurely pace. To increase motivation, patients received a diary in which they were instructed to mark each day they performed the resistance training and aerobic activity.

During the follow‐up calls, the physiotherapist answered questions, motivated the patients to continue being physically active, and instructed patients on how to intensify aerobic activity and/or increase the difficulty level of the resistance training, if possible. Patients in the control group received general information about the study and information about the potential health benefits of being physically active during and after cancer treatment. In accordance with clinical practice, patients who underwent neck dissection were instructed about a specific postoperative exercise program for shoulder mobility by a physiotherapist.

2.4. Physical Activity Assessment

PA was analyzed by grading the patients on the Frändin/Grimby activity scale, ranging from 1 to 6 (hardly any PA to hard or very hard exercise regularly) [27]. The patients also answered three categorical questions from The National Board of Health and Welfare (Socialstyrelsen) about the amount of time (minutes) spent per week on PA (at least 10 min at a time of moderate intensity, e.g., walks, bicycling, or gardening), the amount of time (minutes) spent per week on physical exercise (at high intensity, which makes you short of breath, e.g., running, gymnastics, or ball sports), and the amount of time (hours) spent per day on sedentary behaviors (excluding the number of hours of sleep) [28, 29]. The results from questions 1 and 2 were weighted together into a common measure: activity minutes per week. The median value of the categories was used for summing. When the results were weighed together, the time spent on physical exercise was counted twice. Patients were then categorized into two groups depending on whether they reached the recommendation of ≥ 150 activity minutes per week [28].

2.5. Body Composition and Nutritional Assessment

Body weight (kg) was measured to the nearest 0.1 kg, height (cm) was measured to the nearest 0.5 cm, and BMI (kg/m²) was calculated. Weight loss (kg) at inclusion was based on patients' self‐reports. At the other four follow‐up assessments, weight loss was based on weight at the time of inclusion. To evaluate body composition, Nutribox (Data Input, Lindenberg 782 343 Pöcking, Germany), a 4‐electrode, 50‐kHz, bioelectrical impedance analysis (BIA) device, was used. BIA evaluations were conducted with patients in the supine position on the patient's right side, using two surface electrodes on the hand and two on the foot [30]. The equation by Dey et al. [31] was used to calculate FFM:

$$\begin{array}{c}\mathrm{FFM}\left(\mathrm{kg}\right)=11.78+\left(0.499\times H^2\left(\mathrm{cm}\right)/R\left(\mathrm{\Omega }\right)\right)\hfill \\ +\left(0.134\times \text{Weight}\left(\mathrm{kg}\right)\right)+\left(3.449\times \mathrm{sex}\right)\hfill \end{array}$$

where H = height, R = Resistance, and sex = 0 for females and 1 for males. The calculated FFM was used as a surrogate measure of skeletal muscle mass, and the fat‐free mass index (FFMI) was calculated by dividing the FFM (kg) by the height² (m). The cutoffs for reduced muscle mass were FFMI < 17 (kg/m²) for men and < 15 (kg/m²) for women [32]. Enteral feeding tubes and the use of ONS were registered, and blood samples were drawn to analyze C‐reactive protein (CRP) levels. Malnutrition was diagnosed according to the GLIM criteria (Global Leadership Initiative on Malnutrition) [32]. Patients with an intake < 50% of the estimated energy requirement for > 1 week and all patients with enteral tube feeding or parenteral nutrition were considered to have reduced food intake. A CRP level > 5 mg/L was used as the cutoff for inflammation [32, 33].

2.6. Health‐Related Quality of Life

HRQoL was studied using three questionnaires from the European Organization for Research and Treatment of Cancer (EORTC), which were completed by the patients at the time of enrollment and at the four follow‐ups. The EORTC QLQ‐C30 [34] is a well‐established questionnaire developed for all cancers and includes both scales and single‐item measures: five functional scales, three symptom scales, a global QoL scale, and six single items. The EORTC QLQ‐HN35 [35] incorporates problems and symptoms common to HNC patients and consists of 7 scales and 11 single items. The EORTC QLQ‐FA12 [36] measures cancer‐related fatigue and consists of three symptom scales and two single items. Most of the questions are answered on a four‐point Likert scale that is converted with linear transformation to a final score ranging from 0 to 100. A higher score on a functional scale or for the global QoL scale represents a high level of function, but a higher score on a symptom scale represents a high level of symptoms or problems [37]. Differences of ≥ 10 points between the two groups indicated a clinically significant difference in HRQoL [37, 38]. Fatigue was analyzed via the fatigue scale from the EORTC QLQ‐C30 and the three scales for physical, emotional, and cognitive fatigue from the EORTC QLQ‐FA12. From the EORTC QLQ‐C30, global QoL; physical, role, emotional, and cognitive function; pain; and appetite loss were analyzed, and from the EORTC QLQ‐HN35, local pain, ability to swallow, dry mouth, and sticky saliva were analyzed.

2.7. Statistical Analysis

Continuous variables are presented as the mean and standard deviation (SD). Mean difference is accompanied by a 95% confidence interval (CI). Categorical variables are presented as numbers and percentages (%). Changes in means were calculated from baseline values. For comparisons between the exercise group and the control group, the t test was used for continuous variables, the Mantel–Haenszel chi‐square test was used to compare ordered categorical variables, the chi‐square test was used to compare nonordered categorical variables, and Fisher's exact test was used to compare dichotomous variables. The Mann–Whitney U test was used to compare HRQoL between the two groups. The prevalence of malnutrition was determined according to the GLIM criteria, which are based on dichotomous variables and are presented as numbers and percentages. Kaplan–Meier analysis was used to estimate overall survival. The statistical analyses included only patients who were assessed at the different time points. Additionally, for the HRQoL questionnaires, at least half of the items on a scale needed to be answered to qualify for analysis. The threshold for statistical significance was a two‐sided p value of 0.05. Analyses were performed via SAS version 9.4 statistical software.

Power analyses were performed with the Mann–Whitney U test, including one for the primary outcome of FFM and one for the primary HRQoL outcome of fatigue. To detect a difference of 2.8% (SD 3.0%) in change in FFM between the groups, with a power of 90%, a total of 54 patients were required to complete the 4‐month follow‐up. The change in FFM is presented as the mean percentage with a 95% confidence interval (CI) for the mean difference. For a power of 90% to detect a difference of 20 points (SD 24.6) between the groups for the EORTC QLQ‐C30 fatigue scale, 70 patients were required at the 4‐month follow‐up.

3. Results

3.1. Patient Characteristics

A total of 119 patients were randomized, but 11 withdrew from the study or died before the first follow‐up. Therefore, 108 patients were included in the analyses (Figure 2). At baseline, the age of the participants ranged from 25 to 83 years, with a mean age of 67 years. A total of 48% of the participants were male (Table 1). The tongue was the most common tumor site in both groups. Most of the patients were diagnosed with stage I–II disease (74% vs. 69%) and were treated with surgery alone (72% vs. 55%). More than 90% of the patients in both groups had a WHO performance status of 0, and a majority of the participants were considered overweight or obese. No significant differences in baseline characteristics were found between the two groups (Table 1). During the follow‐up period, few patients discontinued the study or died (Figure 2), and the 11 patients who withdrew from the study before the first follow‐up were mainly women (90%), and a majority (55%) of the patients were diagnosed with advanced cancer. Additionally, there were also missed follow‐up assessments, mainly due to COVID‐19 but also due to delayed doctor appointments and patient illnesses.

FIGURE 2.

Flow chart of inclusion, exclusion, randomization and follow‐ups in the study.

TABLE 1.

Patient characteristics at baseline by randomization: Exercise group and control group.

Variable	Total (n = 108)	Exercise group (n = 50)	Control group (n = 58)	p
Age (years)	67.1 (12.0)	66.6 (13.0)	67.5 (11.2)	0.69
Sex
Female	56 (52%)	26 (52%)	30 (52%)
Male	52 (48%)	24 (48%)	28 (48%)	1.00
Living situation
Alone	36 (33%)	15 (30%)	21 (36%)
Together	72 (67%)	35 (70%)	37 (64%)	0.64
Smoking
Nonsmoker	55 (51%)	29 (58%)	26 (45%)
Previous smoker	39 (36%)	15 (30%)	24 (41%)
Current smoker	14 (13%)	6 (12%)	8 (14%)	0.38
WHO performance status
0	102 (94%)	47 (94%)	55 (95%)
1	6 (6%)	3 (6%)	3 (5%)	1.00
Tumor site of the oral cavity
Tongue	60 (56%)	28 (56%)	32 (55%)
Gingivomandibular	18 (17%)	8 (16%)	10 (17%)
Gingivomaxillar	9 (8%)	4 (8%)	5 (9%)
Buccal	7 (6%)	4 (8%)	3 (5%)
Retromolar	7 (6%)	2 (4%)	5 (9%)
Floor of mouth	4 (4%)	2 (4%)	2 (3%)
Palate	3 (3%)	2 (4%)	1 (2%)	0.94
Stage a
I‐II	77 (71%)	37 (74%)	40 (69%)
III‐IV	31 (29%)	13 (26%)	18 (31%)	0.72
Treatment received
Surgery	68 (63%)	36 (72%)	32 (55%)
Surgery and Radiotherapy	30 (28%)	12 (24%)	18 (31%)
Chemoradiotherapy	4 (3.5%)	1 (2%)	3 (5%)
Surgery and Chemoradiotherapy	6 (5.5%)	1 (2%)	5 (9%)	0.21
Weight (kg)	75.2 (15.2)	77.1 (17.4)	73.6 (12.9)	0.25
Weight loss during 6 months before inclusion
Yes	38 (35%)	19 (38%)	19 (33%)	0.71
BMI (kg/m2)
< 18.5	3 (3%)	2 (4)	1 (2)
18.5–24.9	40 (37.0%)	19 (38.0%)	21 (36.2%)
25–29.9	48 (44%)	18 (36%)	30 (52%)
> 30	17 (16%)	11 (22%)	6 (10%)	0.72
FFMI (kg/m2)	19.0 (2.6)	19.2 (2.8)	18.9 (2.4)	0.63
Female	17.5 (1.4)	17.6 (1.4)	17.4 (1.4)	0.57
Male	20.7 (2.5)	20.9 (3.0)	20.6 (2.0)	0.68
Overall survival
One‐year survival	96%	98%	95%	0.38
Two‐year survival	95%	96%	95%	0.76

Note: For continuous variables, the mean (SD) is presented, and for categorical variables, number (%) is presented. The Kaplan–Meier analysis was used to estimate overall survival. For comparison between the two groups Fisher's exact test was used for dichotomous variables, the Mantel–Haenszel chi square test for ordered categorical variables, chi square test for nonordered categorical variables and t test was used for continuous variables.

Abbreviations: BMI, body mass index; FFMI, fat free mass index.

^a Staging according to the Union for International Cancer Control TNM classification of malignant tumors 7th edition.

3.2. Self‐Reported Physical Activity

Compared with those in the control group, significantly more patients in the exercise group increased their PA according to the Frändin/Grimby 6–level activity scale at the 2‐, 4‐, and 6‐month follow‐up assessments (Table 2). During the study, 30%–40% of patients in the intervention group and slightly more than 50% of patients in the control group remained at the same activity level on the Frändin/Grimby scale as at baseline (Table 2). When the six categories on the Frändin/Grimby scale were compared, a significant difference was observed at the 6‐month follow‐up, with more patients in the control group reporting hardly any activity or mostly sitting (Figure 3). When the questions from the National Board of Health and Welfare (Socialstyrelsen) were analyzed, patients in the control group reported a significantly greater amount of time spent performing PA at baseline (Table 3). At all four subsequent follow‐up assessments, compared with the control group, the exercise group significantly increased their time spent on PA (Table 3). Approximately 50% of patients in the intervention group increased time spent performing PA at all follow‐ups, compared with 25% in the control group. These numbers were reversed when comparing the number of patients that reduced their PA. No significant differences were observed for changes in the amount of time spent on physical exercise or sedentary behavior (Table 3). A comparison of the number of patients who performed at least 150 min of activity per week revealed a significant difference in favor of the control group at baseline before the intervention started (Figure 4). However, at the 2‐month follow‐up, compared with those in the control group, significantly more patients in the exercise group reported being physically active for 150 min or more. At the three remaining follow‐ups, no differences were found (Figure 4).

TABLE 2.

Change in physical activity according to the Frändin/Grimby 6‐level activity scale by randomization: Exercise group and control group.

Visits	Variable	Exercise group	Control group	p
Baseline	Frändin/Grimby activity scale	n = 50	n = 58
	1. Hardly any physical activity	5 (10)	0 (0)
	2. Mostly sitting	11 (22)	7 (12)
	3. Light physical exercise (2–4 h/week)	16 (32)	25 (43)
	4. Moderate exercise (1–2 h/week)	12 (24)	21 (36)
	5. Moderate exercise (3 h/week)	3 (6)	5 (9)
	6. Hard or very hard exercise	3 (6)	0 (0)	0.15
2 months	Change Frändin/Grimby activity scale	n = 49	n = 54
	Reduced activity level	16 (33)	22 (41)
	Unchanged activity level	19 (39)	29 (54)
	Increased activity level	14 (28)	3 (5)	0.025
4 months	Change Frändin/Grimby activity scale	n = 42	n = 44
	Reduced activity level	13 (30)	18 (41)
	Unchanged activity level	14 (33)	23 (52)
	Increased activity level	16 (37)	3 (7)	0.011
6 months	Change Frändin/Grimby activity scale	n = 44	n = 47
	Reduced activity level	9 (20)	15 (32)
	Unchanged activity level	14 (32)	24 (51)
	Increased activity level	21 (48)	8 (17)	0.0086
12 months	Change Frändin/Grimby activity scale	n = 44	n = 48
	Reduced activity level	7 (16)	11 (23)
	Unchanged activity level	19 (43)	26 (54)
	Increased activity level	18 (41)	11 (23)	0.091

Note: For categorical variables, number (%) is presented. Change is calculated from baseline. For comparison between the two groups, the Mantel–Haenszel chi‐square test was used for ordered categorical variables. The outcome variables were significant if p < 0.05 provided in italics.

FIGURE 3.

Comparison of the Frändin/Grimby 6‐level activity scale [27] in percentage between the physical activity group and the control group at inclusion and 2, 4, 6, and 12 months after enrollment. The Mantel–Haenszel chi‐square test was used for comparison between the two groups. A significant difference was found only at the 6‐month follow‐up. The outcome variables were significant if p < 0.05 provided in italics. [Color figure can be viewed at wileyonlinelibrary.com]

TABLE 3.

Change in time (minutes) per week spent on physical activity and physical exercise and time (hours) per day spent sedentary, by randomization: Exercise group and control group.

Visits	Time spent on physical activity				Time spent on physical exercise				Time spent sedentary
	Variable	Exercise	Control	p	Variable	Exercise	Control	p	Variable	Exercise	Control	p
Baseline	Daily exercise	n = 50	n = 58		Physical exercise	n = 50	n = 58		Sedentary	n = 50	n = 58
	0 min	2 (4)	1 (2)						Almost the entire day	1 (2)	3 (5)
	< 30 min	4 (8)	2 (3)		0 min	23 (46)	29 (50)		13–15 h	1 (2)	1 (2)
	30–60 min	6 (12)	6 (10)		< 30 min	11 (22)	8 (14)		10–12 h	8 (16)	10 (17)
	60–90 min	8 (16)	2 (3)		30–60 min	4 (8)	10 (17)		7–9 h	15 (30)	10 (17)
	90–150 min	11 (22)	9 (16)		60–90 min	6 (12)	3 (5)		4–6 h	18 (36)	26 (45)
	150–300 min	5 (10)	16 (28)		90–120 min	2 (4)	0 (0)		1–3 h	7 (14)	7 (12)
	> 300 min	14 (28)	22 (38)	0.022	> 120 min	4 (8)	8 (14)	0.93	Never	0 (0)	1 (2)	1.00
2 months		n = 47	n = 51			n = 47	n = 51			n = 47	n = 51
	Reduced	11 (23)	27 (53)		Reduced	9 (19)	16 (31)		Reduced	10 (21)	14 (28)
	Unchanged	16 (34)	13 (26)		Unchanged	18 (38)	20 (39)		Unchanged	22 (47)	24 (47)
	Increased	20 (43)	11 (22)	0.0029	Increased	20 (43)	15 (29)	0.11	Increased	15 (32)	13 (26)	0.39
4 months		n = 42	n = 44			n = 42	n = 44			n = 42	n = 44
	Reduced	10 (24)	23 (52)		Reduced	7 (17)	10 (23)		Reduced	11 (26)	16 (36)
	Unchanged	10 (24)	12 (27)		Unchanged	16 (38)	19 (43)		Unchanged	21 (50)	18 (41)
	Increased	22 (52)	9 (21)	0.0012	Increased	19 (45)	15 (34)	0.29	Increased	10 (24)	10 (23)	0.48
6 months		n = 44	n = 47			n = 43	n = 47			n = 44	n = 46
	Reduced	10 (23)	23 (49)		Reduced	8 (19)	13 (28)		Reduced	15 (34)	12 (26)
	Unchanged	12 (27)	13 (28)		Unchanged	12 (28)	17 (36)		Unchanged	16 (36)	24 (52)
	Increased	22 (50)	11 (23)	0.0033	Increased	23 (54)	17 (36)	0.12	Increased	13 (30)	10 (22)	0.99
12 months		n = 44	n = 48			n = 44	n = 48			n = 44	n = 48
	Reduced	9 (21)	25 (52)		Reduced	6 (14)	13 (27)		Reduced	16 (36)	16 (33)
	Unchanged	16 (36)	12 (25)		Unchanged	15 (34)	16 (33)		Unchanged	17 (39)	27 (56)
	Increased	19 (43)	11 (23)	0.0030	Increased	23 (52)	19 (40)	0.11	Increased	11 (25)	5 (10)	0.43

Note: For categorical variables, number (%) is presented. Change is calculated from baseline. For comparison between the two groups, the Mantel–Haenszel chi‐square test was used for ordered categorical variables. The outcome variables were significant if p < 0.05 provided in italics.

FIGURE 4.

Comparison of the proportion of patients with < 150 min of activity and ≥ 150 min of activity per week between the physical activity group and the control group at inclusion and at 2, 4, 6, and 12 months after enrollment. Fisher's exact test was used for comparisons between two groups. Significant differences were found at baseline and at the 2‐month follow‐up. The outcome variables were significant if p < 0.05 provided in italics. [Color figure can be viewed at wileyonlinelibrary.com]

3.3. Changes in Fat‐Free Mass, Body Weight, and Malnutrition

The reductions in FFM and body weight during the study period were small in both groups, and no significant differences were found between the groups (Table 4). In the intervention group, the highest percentage of FFM loss was observed at 12 months (−2.6%), and the highest percentage of body weight loss was observed at 6 months (−2.7%). In the control group, the greatest loss in both FFM (−1.8%) and body weight (−2.9%) was observed at the 6‐month follow‐up. At 12 months, the change in the mean FFM (%) in both groups differed from the median change in FFM: −1.4% (range: −21.5; 6.0) in the exercise group and −1.3% (range: −10.3; 20.2) in the control group. One extreme value in each group had a large effect on the mean value at this time point. The number of patients considered to have reduced muscle mass ranged from 2 to 7 patients in the exercise group and from 5 to 8 patients in the control group. There was no significant difference in the prevalence of malnutrition between the two groups. The highest prevalence of malnutrition was observed at the 4‐month follow‐up, with 14% and 18% in the intervention and control group, respectively (Table 4). During the study period, approximately 80% of the patients in both groups used protein‐rich ONS, and 20% in the exercise group and 30% in the control group used an enteral feeding tube at some point.

TABLE 4.

Fat‐free mass, body weight, and malnutrition by randomization: Exercise group and control group.

	Variable	Exercise group	Control group	p	Mean difference and 95% CI
Baseline		n = 50	n = 58
	FFM, kg	52.8 (10.2)	51.3 (9.3)	0.45	1.4 (−2.3; 5.2)
	FFMI Female, kg/m2	17.6 (1.4)	17.4 (1.4)	0.57	0.2 (−0.5; 0.9)
	FFMI Male, kg/m2	20.9 (3.0)	20.6 (2.0)	0.68	0.3 (−1.1; 1.7)
	Number of patients with reduced muscle mass a	6 (12%)	6 (10%)	1.00	−2 (−16; 12)
	Body weight, kg	77.1 (17.4)	73.6 (12.9)	0.25	3.5 (−2.5; 9.4)
	Number of patients with malnutrition b	0 (0%)	1 (2%)	1.00	2 (−4; 7)
2 months		n = 49	n = 54
	FFM, kg	52.2 (10.1)	50.7 (9.1)	0.41	1.6 (−2.2; 5.4)
	Change FFM, %	−1.3 (3.9)	−1.1 (3.5)	0.79	−0.2 (−1.7; 1.3)
	FFMI Female, kg/m2	17.3 (1.4)	17.2 (1.4)	0.79	0.1 (−0.7; 0.9)
	FFMI Male, kg/m2	20.6 (2.9)	19.9 (2.1)	0.32	0.7 (−0.7; 2.1)
	Number of patients with reduced muscle mass a	6 (12%)	8 (15%)	0.87	3 (−12; 18)
	Body weight, kg	75.7 (16.5)	71.7 (13.0)	0.17	4.0 (−1.8; 9.8)
	Change body weight, %	−2.0 (3.0)	−1.9 (3.4)	0.85	−0.1 (−1.4; 1.1)
	Number of patients with malnutrition b	4 (8%)	5 (9%)	1.00	1 (−12; 14)
4 months		n = 41	n = 41
	FFM, kg	51.9 (9.8)	51.4 (8.6)	0.81	0.5 (−3.6; 4.6)
	Change FFM, %	−1.2 (4.2)	−1.0 (4.5)	0.89	−0.1 (−2.0; 1.8)
	FFMI Female, kg/m2	17.2 (1.0)	17.5 (1.5)	0.51	−0.3 (−1.1; 0.5)
	FFMI Male, kg/m2	20.6 (2.8)	19.7 (2.2)	0.24	0.9 (−0.7; 2.5)
	Number of patients with reduced muscle mass a	2 (5%)	5 (12%)	0.43	7 (−6; 23)
	Body weight, kg	74.3 (16.4)	72.9 (12.1)	0.67	1.4 (−5.0; 7.7)
	Change body weight, %	−2.1 (4.9)	−1.8 (4.5)	0.75	−0.3 (−2.4; 1.7)
	Number of patients with malnutrition b	6 (14%)	8 (18%)	0.81	4 (−12; 21)
6 months		n = 44	n = 48
	FFM, kg	51.5 (9.6)	50.8 (9.2)	0.72	0.7 (−3.4; 4.9)
	Change FFM, %	−1.2 (3.9)	−1.8 (4.1)	0.48	0.6 (−1.1; 2.4)
	FFMI Female, kg/m2	17.2 (1.1)	17.2 (1.4)	0.99	0.0 (−0.8; 0.8)
	FFMI Male, kg/m2	20.5 (2.4)	20.0 (2.4)	0.55	0.5 (−1.1; 2.1)
	Number of patients with reduced muscle mass a	5 (13%)	7 (17%)	0.86	4 (−13; 20)
	Body weight, kg	74.3 (16.5)	70.9 (12.1)	0.27	3.4 (−2.7; 9.4)
	Change body weight, %	−2.7 (7.2)	−2.9 (5.7)	0.89	0.2 (−2.5; 2.9)
	Number of patients with malnutrition b	6 (14%)	8 (17%)	0.91	3 (−13; 19)
12 months		n = 44	n = 48
	FFM, kg	52.3 (9.4)	50.8 (9.4)	0.45	1.5 (−2.5; 5.5)
	Change FFM, %	−2.6 (5.5)	−0.9 (5.5)	0.15	−1.7 (−4.1; 0.6)
	FFMI Female, kg/m2	17.1 (1.4)	17.2 (1.8)	0.84	−0.1 (−1.1; 0.9)
	FFMI Male, kg/m2	20.5 (2.3)	20.1 (2.3)	0.55	0.4 (−1.0; 1.9)
	Number of patients with reduced muscle mass a	7 (16%)	8 (18%)	1.00	2 (−15; 19)
	Body weight, kg	76.0 (16.8)	71.8 (13.0)	0.18	4.2 (−2.0; 10.4)
	Change body weight, %	−2.3 (8.0)	−2.8 (6.8)	0.77	0.5 (−2.6; 3.5)
	Number of patients with malnutrition b	0 (0%)	2 (4%)	0.54	4 (−5; 14)

Note: For continuous variables, the mean (SD) is presented and for categorical variables, number (%) is presented. The mean difference is presented with a 95% confidence interval. Change is calculated from baseline. For comparison between the two groups, a t test was used for continuous variables, and Fisher's exact test was used for dichotomous variables. Missing values: 2 months: control group FFM n = 2, 6 months: exercise group FFM n = 5; control group FFM n = 6, 12 months: exercise group FFM n = 1; control group FFM n = 4.

Abbreviations: CI, confidence interval; FFM, fat‐free mass.

^a Cut off values reduced muscle mass: FFMI < 15 kg/m² for females and < 17 kg/m² for men.

^b Malnutrition diagnosis according to the GLIM criteria [32].

3.4. Health‐Related Quality of Life

The results from the selected scales from the EORTC QLQ–C30, QLQ‐HN35, and QLQ‐FA12 questionnaires are shown in Table 5. At the 2‐month follow‐up, the patients in both groups reported most problems with fatigue. The reduction in the fatigue scale (QLQ–C30) was clinically significant compared to baseline within both groups. There was a 6‐point smaller change in fatigue (QLQ–C30) and a 9‐point smaller change in physical fatigue (QLQ–FA12) in favor of the exercise group; however, these differences were not significant. At the 4‐month follow‐up, problems with fatigue were still more pronounced than they were at the time of diagnosis but were not clinically significant in either group. At the remaining follow‐up assessments, further reductions in fatigue were observed (Table 5).

TABLE 5.

Comparison of health‐related quality of life (EORTC QLQ‐C30, QLQ‐HN35, and QLQ‐FA12) between the exercise group and control group.

Variable	Baseline			2 months			4 months			6 months			12 months
	Exercise	Control	p	Exercise	Control	p	Exercise	Control	p	Exercise	Control	p	Exercise	Control	p
	n = 49	n = 58		n = 47	n = 51		n = 41	n = 44		n = 43	n = 47		n = 43	n = 48
	Mean	Mean		Δ	Δ		Δ	Δ		Δ	Δ		Δ	Δ
EORTC QLQ‐C30
Global QoL	65.6 (21.6)	66.2 (21.1)	0.78	−3.0 (23.3)	−10.3 (19.5)	0.041	1.8 (22.4)	−7.8 (17.0)	0.033	3.7 (22.0)	−0.53 (22.0)	0.39	7.0 (23.4)	5.9 (21.7)	0.74
Physical functioning	91.4 (9.7)	91.4 (12.6)	0.46	−7.2 (13.8)	−11.4 (14.8)	0.29	−5.6 (14.0)	−7.4 (10.7)	0.38	−3.7 (12.7)	−6.4 (10.1)	0.21	−4.5 (12.8)	−2.3 (7.3)	0.24
Role functioning	85.4 (25.4)	86.8 (23.9)	0.73	−19.1 (34.7)	−24.0 (30.7)	0.54	−0.81 (33.7)	−15.9 (28.1)	0.030	−2.3 (33.3)	−8.2 (28.6)	0.56	−3.9 (39.7)	3.8 (25.3)	0.32
Emotional functioning	69.7 (19.0)	66.4 (22.0)	0.64	6.6 (20.8)	4.8 (16.5)	0.43	9.6 (23.0)	13.3 (20.0)	0.38	7.4 (20.3)	12.5 (23.4)	0.26	14.0 (20.2)	14.1 (20.9)	0.79
Cognitive functioning	81.0 (22.6)	80.7 (21.6)	0.82	2.5 (20.0)	1.0 (17.1)	0.76	3.7 (21.3)	3.8 (13.9)	0.97	3.5 (22.0)	1.8 (16.8)	0.89	5.4 (17.7)	7.6 (15.0)	0.45
Fatigue	23.6 (19.5)	24.7 (22.7)	1.00	−10.6 (21.6)	−16.6 (20.9)	0.21	−6.8 (26.2)	−9.3 (25.6)	0.67	−5.0 (24.9)	−3.6 (22.2)	0.84	−2.6 (24.9)	−0.93 (18.2)	0.99
Pain	21.4 (23.3)	20.7 (24.0)	0.79	−5.0 (27.8)	−9.5 (30.1)	0.18	−3.7 (29.2)	−12.5 (27.2)	0.19	−1.6 (27.7)	−4.6 (26.6)	0.58	−2.3 (29.0)	3.1 (20.0)	0.49
Appetite loss	11.6 (18.7)	12.6 (21.5)	0.99	−6.5 (29.5)	−10.5 (36.2)	0.57	−8.9 (35.8)	−10.9 (41.6)	0.76	−8.5 (36.4)	−8.5 (29.1)	0.79	−0.8 (29.5)	2.1 (24.5)	0.86
EORTC QLQ‐H&N35
Local pain	24.1 (21.1)	26.0 (21.2)	0.74	−4.3 (24.2)	−2.6 (23.2)	0.64	−1.0 (28.3)	−7.8 (31.0)	0.61	5.7 (27.1)	0.41 (21.5)	0.27	4.3 (29.3)	4.9 (17.3)	0.79
Swallowing	7.8 (14.2)	7.6 (16.1)	0.29	−11.5 (18.4)	−11.4 (27.5)	0.83	−3.9 (17.2)	−15.2 (32.7)	0.18	−2.1 (17.6)	−8.2 (23.8)	0.19	1.2 (11.3)	−3.0 (20.2)	0.025
Dry mouth	21.1 (35.2)	23.0 (31.3)	0.44	−17.0 (40.5)	−14.7 (30.2)	0.59	−22.0 (38.4)	−17.4 (36.3)	0.26	−20.2 (41.2)	−15.6 (38.6)	0.40	−22.5 (39.0)	−19.4 (33.6)	0.61
Sticky saliva	13.6 (25.4)	14.4 (21.7)	0.57	−13.5 (23.7)	−17.0 (36.1)	0.87	−20.3 (37.2)	−24.0 (37.3)	0.90	−15.5 (27.6)	−17.7 (35.3)	1.00	−9.3 (32.0)	−11.1 (29.4)	0.70
EORTC QLQ‐FA12
FA12 Physical	28.4 (22.0)	27.0 (23.3)	0.61	−6.2 (25.9)	−15.7 (20.7)	0.063	1.5 (24.9)	−7.1 (22.8)	0.19	1.4 (20.3)	−1.3 (19.8)	0.53	4.8 (25.3)	1.4 (18.0)	0.33
FA12 Emotional	19.7 (21.1)	25.1 (26.5)	0.32	−1.7 (22.6)	−0.44 (20.5)	0.75	1.1 (23.0)	2.8 (22.0)	0.79	1.6 (23.7)	4.7 (25.8)	0.64	5.2 (22.5)	4.9 (22.0)	0.79
FA12 Cognitive	13.3 (19.2)	15.8 (24.9)	0.93	4.3 (19.2)	0.33 (19.0)	0.30	5.7 (20.6)	3.8 (22.4)	0.59	1.6 (20.8)	2.1 (21.6)	0.56	4.3 (22.2)	3.8 (17.9)	1.00

Note: For continuous variables, the mean (SD) is presented at baseline and the change in mean (SD) is presented at the follow‐ups. Change is calculated from baseline values. For comparison between the two groups, the Mann–Whitney U test was used for continuous variables. EORTC QLQ‐C30: The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire‐Core 30. EORTC QLQ‐H&N35: The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire‐Head and Neck 35. EORTC QLQ‐FA12: The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire‐Fatigue 12. At baseline a higher score for the global quality of life scale and the function scales reflects a better function, and a higher score for the symptom scales represents more problems. At the follow‐ups a positive change in score reflects improved function and symptoms, and a negative change represents impaired function and symptoms. > 10 points are considered a clinically relevant difference. Δ = change. The outcome variables were significant if p < 0.05 provided in italics.

The results from the other scales revealed statistically significant differences in global QoL at the 2‐ and 4‐month follow‐up assessments in favor of the exercise group, but the differences were less than 10 points (Table 5). At 2 months, both groups reported a clinically relevant decline in swallowing function of 11 points, and at the 4‐month follow‐up, compared with the exercise group, the control group reported clinically significantly more swallowing difficulties. At the 4‐month follow‐up, the control group had a significantly greater decline in role functioning and tended to have more problems with pain. There was a clinically significant decrease in problems with dry mouth in both groups at all follow‐ups and a decrease in problems with sticky saliva at all time points except at 12 months, but no differences were observed between the two groups. At 6 and 12 months, no relevant differences were found between the groups (Table 5).

4. Discussion

In this RCT, patients with oral cancer were given an individualized home‐based exercise program for 6 months, and muscle mass, fatigue, and HRQoL were assessed over a one‐year period. Significantly more patients in the intervention group reported engaging in PA for 150 min or more per week at the 2‐month follow‐up, and the intervention group also reported a significantly larger increase in PA than the control group did during the entire study period. The decrease in muscle mass and body weight was small in both groups, and no effect on muscle mass was established when the exercise group was compared with the control group. The exercise group tended to have fewer problems with fatigue and significantly better global QoL and role function.

The individualized home‐based exercise program was designed to meet the current recommendations for PA [26]. Instructions were given face‐to‐face at the start of the study, but the exercise was unsupervised, home‐based, and carried out with body weight as resistance. The intention of a home‐based program with a higher frequency of training sessions and shorter duration was to ease participation, facilitate regular training, minimize workload (e.g., extra visits, travel), and be flexible for patients. During the 6‐month intervention period, the physiotherapist contacted the patients 5 times by telephone, and no training session was monitored. In the study by Samuel et al. [39], which examined 148 HNC patients, an exercise intervention during and after treatment with CRT significantly improved both functional capacity and HRQoL and reduced fatigue. Similar to our exercise program, they had five training sessions per week with shorter duration, but the training sessions were supervised for the first 7 weeks, and thereafter, patients were monitored once a week for 4 weeks of home‐based exercise. Other interventions with supervised exercise for patients with cancer have shown advantages in terms of HRQoL, fatigue, and physical function [40, 41]. In contrast, a study on exercise preferences in patients with HNC has shown that most patients want advice and instructions face‐to‐face, with a majority preferring unsupervised exercise and a flexible schedule [42]. Additionally, in another study on preferences in HNC survivors, the most popular exercise type was walking; patients preferred to exercise primarily at home or outdoors, and one of the most frequently reported exercise barriers was fatigue [43]. Understanding HNC patients' preferences and barriers to exercise is important, as these factors may also affect participation and adherence.

In our study, self‐reported PA increased significantly in the exercise group during the intervention period, and the increase was also maintained 12 months after inclusion, which is an important finding since studies in patients with HNC have shown a reduction in PA and low PA levels after treatment [44, 45]. In the present study, subjective methods of assessing PA were used, which introduce a risk of bias. Few studies have used objective methods to assess PA in patients with HNC, and two of those studies that used accelerometry reported low levels of PA before treatment [46, 47]. At baseline, the control group reported greater PA, which resulted in similar PA levels in both groups at the follow‐ups, despite the exercise group increasing their self‐reported PA. Adherence to the resistance training could not be assessed because few exercise diaries were submitted or were not used properly. These findings could explain why no differences in muscle mass loss were found between the two groups. Another explanation could be that the mean decrease in body weight and muscle mass did not exceed 3% in either of the two study groups, and the prevalence of malnutrition varied from 0% to 18% during the study period. The small mean loss in body weight and muscle mass in both groups made it difficult to detect differences, and significant weight loss and prevalence of malnutrition are often more common in patients with HNC [48, 49, 50]. The small average loss in weight and muscle mass could be explained by the fact that most patients underwent surgery as monotherapy, allowing for faster recovery, and that all patients received active nutritional treatment during and after treatment, including contact with a dietitian at diagnosis and protein‐rich ONS. In the RCT by Nett et al. [51], the author reported that postoperative intervention with nutritional counseling and protein‐rich ONS reduced weight loss among patients with oral cancer. Similarly, other studies with nutritional interventions and increased protein intake have also shown reductions in the loss of body weight and muscle mass [52, 53].

Few studies, especially RCTs, have examined the effect of PA on muscle mass in patients with HNC with different PA interventions, tumor sites, and treatments [23, 24]. Similar to our results, no effect on reducing muscle mass loss in patients with HNC was observed between groups in the study by Capozzi et al. [24] after either a 12‐week immediate or a 12‐week delayed lifestyle intervention (during or after treatment). In contrast, the DAHANCA 25B randomized clinical trial reported a positive effect of 12 weeks of progressive resistance training, initiated either 8 or 20 weeks after the end of treatment, on lean body mass and muscle strength [23]. During treatment, patients with HNC face several challenges, such as increasing nutritional difficulties and inflammation‐driven muscle depletion, which can make exercise less effective in terms of reducing muscle mass loss. The treatment burden may also affect adherence; therefore, posttreatment exercise may be more beneficial for muscle mass recovery.

Our study revealed that compared with the control group, the exercise group tended to experience less fatigue and had better global QoL and role function. In two other studies where the exercise intervention was performed during treatment with CRT, the exercise group experienced significantly less fatigue and improved QoL compared to the control group [20, 39]. In the previously mentioned DAHANCA 25B trial, where the training program was initiated after treatment, QoL improved in the early exercise group; however, fatigue was reduced in both the early and delayed exercise group without any established differences between the groups [23]. This suggests that regardless of when exercise is introduced, it is possible to reduce fatigue and improve QoL. Thus, early initiation of exercise during treatment should be preferred to prevent severe deterioration, maintain well‐being, and promote recovery.

The strengths of the study were that it was randomized, included exclusively patients with oral cancer, and had few dropouts during the follow‐up period. However, some limitations must be addressed. Patients were not stratified according to PA level at diagnosis, which caused the randomization to fail in balancing the two study groups at baseline. The study participants and outcome assessors were not blinded, which may have inflicted performance bias. Too few exercise diaries were submitted, or the diaries lacked information about the training and could therefore not be used to assess adherence to the exercise program. Finally, the number of missed follow‐ups was higher due to the COVID‐19 pandemic, and more patients than expected needed to be included.

5. Conclusion

In conclusion, this RCT revealed that patients who received a home‐based exercise program increased their PA. The decrease in muscle mass was small, and no differences in muscle mass changes were found between the groups. Compared with the control group, the exercise group tended to have less fatigue and had better global QoL. Currently, PA programs are not part of usual care for patients with HNC, although positive effects have been established, and more research is needed. We suggest that future studies, preferably RCTs, should include larger sample sizes of HNC patients to investigate the effects of exercise on muscle mass, fatigue, and HRQoL. Assessment of PA levels may be valuable in identifying patients in need of exercise interventions, and both objective and subjective methods of assessing PA need to be evaluated. The timing of when it is most beneficial to introduce an exercise intervention also needs to be explored further, and PA interventions in the HNC population can be designed to offer more individualized options to meet different preferences and increase adherence.

Conflicts of Interest

The authors declare no conflicts of interest.

Wallmander C., Bosaeus I., Silander E., et al., “Effect of a Home‐Based Exercise Program on Muscle Mass, Fatigue, and Health‐Related Quality of Life Among Patients with Oral Cancer: A Randomized Trial,” Head & Neck 48, no. 3 (2026): 680–695, 10.1002/hed.70063.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy or ethical considerations.