Lower Handgrip Strength is Associated with Worse Quality of Life and More Fatigue and Disability in Long-term Colorectal Cancer Survivors

Colorectal cancer (CRC) treatment negatively affects anthropometric measures, which in turn can result, either directly or indirectly, in fatigue and a lower health-related quality of life (HRQoL). Since anthropometric measures are modiable, we examined which of these are associated with HRQoL, fatigue, and disability in long-term CRC survivors. In this cross-sectional study, body mass index (BMI), waist circumference, waist-to-hip ratio, body fat percentage, handgrip strength (HGS) and mid-upper arm muscle area of 155 stage I-III CRC survivors at 2–10 years after treatment were assessed during home visits (response rate 42%). HRQoL was measured by the EORTC QLQ-C30, fatigue by the Checklist Individual Strength (CIS), and disability by the World Health Organization Disability Assessment Schedule II (WHODAS-II). Associations between anthropometric measures and HRQoL, fatigue and disability were analyzed by confounder-adjusted linear regression. We conclude that HGS, as a measure of physical functioning in CRC survivors, was independently associated with HRQoL, fatigue and disability. Future prospective research should focus underlying and causal factors in the association between HGS and HRQoL after CRC treatment and on possible interventions to promote HRQoL outcomes through improving physical


Introduction
Worldwide, colorectal cancer (CRC) is the third most commonly diagnosed cancer in males (10%) and the second in females (9.2%) [1,2]. While the incidence of CRC has been increasing, morbidity and mortality rates have been decreasing. The increasing incidence is likely to be caused by aging of the population, while advanced screening methods and improved treatments lead to a steady increase in the number of CRC survivors [1][2][3][4]. Increasing number of patients surviving CRC not only makes oncological outcome important but also health-related quality of life (HRQoL), which might be negatively in uenced by long-lasting adverse effects of CRC treatment such as bowel dysfunction and a decrease in physical functioning [5]. In order to develop tailored interventions to improve quality of life in cancer survivors, it is essential to gather evidence about underlying modi able factors contributing to a decreased HRQoL in CRC survivors.
Growing evidence derived from various observational studies consistently shows a positive association between physical activity on the one hand and HRQoL, reduced fatigue symptoms, increased survival rate and decreased treatment side effects on the other hand in CRC survivors [6][7][8][9][10]. Poor dietary habits, obesity and decreased physical activity are often present in patients who are diagnosed with CRC [11]. Physical function might be further decreased as a result of cancer treatment, that negatively affects anthropometric measures, such as muscle strength or body composition [12,13]. Therefore, CRC treatment can result either directly or indirectly via physical function and the effect on anthropometric measures, in fatigue and a lower quality of life for treated patients, in the years after the end of treatment [14][15][16][17][18].
As anthropometric measures are modi able, it is important to identify which of these measures are associated with a better physical functioning and a better HRQoL in the years after the end of treatment. Current studies among CRC patients mainly focus on the association between certain anthropometric measures and patients' physical function instead of studying the direct relation with HRQoL [11,[19][20][21][22][23][24]. These prospective and cross-sectional studies have found that obesity in the period from 6 months until 10 years after CRC diagnosis was associated with a lower HRQoL in CRC survivors, in particular in patients with a lower level of physical functioning [11,[19][20][21][22][23][24]. Nevertheless, among CRC patients, associations between BMI and outcomes such as HRQoL and survival seem divergent since both a low and high BMI can be associated with worse HRQoL [7,[25][26][27].
Instead of BMI, other anthropometric measures, such as waist circumference, waist-to-hip ratio (WHR) and foursite skinfold measurement might be used for body fat and fat distribution [22,28,29]. These measures have been found to be associated with improved survival and HRQoL among patients with gastrointestinal cancer, with greater anthropometric measures predicting a poorer survival and HRQoL [27,30]. Previous studies have also shown that lower handgrip strength (HGS), as a measure of muscle weakness, is associated with a poor quality of life in frail elderly patients [31] and among lung and breast cancer patients [32][33][34].
This means that besides BMI, associations between other anthropometric measures and HRQoL after CRC treatment need to be studied in order to identify factors that can potentially be in uenced before, during and after cancer treatment, such as supervised speci c muscle training programs. Since studies investigating a direct association between anthropometric measures and patients' HRQoL in CRC survivors are scarce and often only include BMI as anthropometric measure, this study aims to examine the association between extensive anthropometric measures and HRQoL, fatigue, and disability in long-term CRC survivors.

Study Design And Participants
For this paper, data were used from the cross-sectional part of the Energy for life after ColoRectal cancer (EnCoRe) study, conducted in stage I-III CRC survivors who were 2-10 years post-diagnosis. Methods of the EnCoRe study, which consists of an ongoing prospective part and a cross-sectional part, have been described previously [35]. Eligible subjects, i.e. stage I-III CRC survivors diagnosed and treated between 2002 and 2010 at Maastricht University Medical Center+, the Netherlands, were preselected via the Netherlands Cancer Registry (NCR; managed by Comprehensive Cancer Centre the Netherlands) and eligible subjects were invited for participation by mail. Reasons for exclusion were death before the start of the study, stage IV disease, comorbidities obstructing successful participation such as Alzheimer disease or severe hearing disorders, and inability to understand the Dutch language (Fig. 1 as a measure of body fatness. Body fat percentage was determined based on measurement of the thickness (mm) of triceps, biceps, subscapular, and suprailiac skinfolds, using a skinfold caliper (Holtain LTD, Crymych, UK). Measurements were performed thrice per skinfold and the sum of the median thickness of the four skinfolds was used to calculate body fat percentage based on formulas provided by Durnin and Womersley [36].
In addition, several circumference measurements (cm) were performed in duplicate using a circumeter (type 05335, Premed, Bilthoven, NL), including mid-upper arm, waist, and hip circumference. The mean values of the duplicate waist and hip circumference measurements were used to calculate the waist-to-hip ratio as a measure of body fat distribution. Mid-upper arm circumference was used in combination with the triceps skinfold to calculate the mid-upper arm muscle area (cm 2 ) as a measure of muscle mass, using the following formula: 1/4π(mid-upper arm circumference -[triceps skinfold]π) 2 [37,38]. Finally, maximum isometric handgrip strength (HGS) of the dominant hand was measured with a validated Jamar hydraulic dynamometer (Sammons Preston Rolyan, Chicago, USA) [39] as a measure of general muscle strength [40]. Participants performed two HGS measurements with su cient rest in-between; the highest value was considered the maximum HGS (kg).

Health-related quality of life outcomes Cancer-speci c HRQoL was measured with the European Organisation for Research and Treatment of Cancer
Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30, version 3.0), which is a reliable and valid measure for HRQoL in patients with cancer [41,42]. For the subscales global health status/quality of life and physical, role, and social functioning, 100-point scores were calculated with higher scores indicating higher levels of quality of life or functioning [43]. Furthermore, disability was assessed by the 12-item version of the World Health Organization Disability Assessment Schedule II (WHODAS II) [44,45]. The WHODAS II has a good reliability and validity in different populations, including elderly individuals and breast cancer survivors [45][46][47]. A total disability score (100-point scale) was calculated, with a higher score indicating a higher level of disability [45]. In addition, fatigue was assessed through the Checklist Individual Strength (CIS), which was originally developed and validated in patients with chronic fatigue syndrome [48,49], but has also been applied in cancer survivors [50]. Individual items from the CIS were summed to calculate a total score for fatigue (scale: 20-140), with higher scores indicating higher levels of fatigue.

Other variables
The Short QUestionnaire to ASsess Health-enhancing physical activity (SQUASH) was used to measure habitual activity level and adherence to national physical activity guideline (> 150 min/week of moderate-to-vigorous physical activity) [35,51]. Participants were also requested to wear a tri-axial accelerometer on the anterior upper leg for 7 consecutive days to determine intensity, frequency, and duration of daily activities like walking, standing, and lying, as well as to determine sedentary time [52]. To measure their total energy intake (kcal), participants were asked to ll out a seven-day dietary record to assess consumption of all foods and drinks over a 1-week period. Socio-demographic characteristics (gender, age, education level) and smoking status were self-reported [11]. Presence of comorbidities was assessed during the study visits at participants' homes [53].
Primary data on patient, tumour and treatment characteristics (cancer stage, age at diagnosis, chemotherapy/radiotherapy treatment, and tumour subsite) were derived from the Netherlands Cancer Registry.

Statistical analysis
Patient characteristics were summarized with descriptive statistics for the total study population and for men and women separately. Multivariable linear regression models were used to analyse associations of the body composition parameters of interest (BMI, waist circumference, waist-to-hip ratio, body fat percentage, maximum HGS, and mid-upper arm muscle area) with global QoL, with physical, role and social functioning, and with fatigue and disability as HRQoL outcomes. All regression analyses were adjusted for a prede ned set of relevant confounders, which were selected based on our previously developed conceptual model for studying lifestyle factors and HRQoL in CRC survivors [35] and included age (years), gender, years since diagnosis, comorbidities (0/1/2 + comorbid conditions), chemotherapy treatment (yes/no), current smoking (yes/no), moderate-tovigorous physical activity (h/w), sedentary time (h/d), and total energy intake (kcal/d). Unadjusted and confounder-adjusted regression coe cients (B) with 95% con dence intervals (95% CI) re ecting associations between body composition parameters and HRQoL outcomes are presented. Values of P < 0.05 were considered to indicate statistical signi cance. All statistical analyses were performed using IBM® SPSS® Statistics, version 22.

Population characteristics
A total of 373 eligible CRC survivors were invited to participate, of whom 155 were recruited (response rate = 42%; Fig. 1). Participants were on average 3.9 years younger, had more often been diagnosed with rectal cancer (42% vs. 36%) and had more frequently received treatment with chemotherapy (52% vs. 43%) and radiotherapy (39% vs. 30%), than non-participants. No differences were observed in gender, years since diagnosis, and TNM stage. As shown in Table 1, a total of 96 male (62%) and 59 female (38%) CRC survivors were included, who were on average 70.1 years of age (standard deviation [SD] = 8.9) and 5.7 years since CRC diagnosis (SD = 1.8).
Of these, 80 survivors (53%) had a previous diagnosis of colon cancer, 64 (42%) rectum cancer and 7 (5%) rectosigmoid cancer. Seventy-one (46%) survivors were overweight, 44 (29%) obese, and 38 (25%) had normal weight; one participant was underweight. A total of 79 (51%) survivors reported two or more comorbid conditions.   Table 2.  Anthropometric variations in relation to HRQoL outcomes In unadjusted analyses (Table 2), no signi cant associations between BMI, waist circumference, waist-to-hip ratio and HRQoL outcomes were found. Nevertheless, each increase of 5% in body fat was signi cantly associated with a decrease in global quality of life score, physical functioning score, role functioning score, and an increase of fatigue symptoms and participants' disability score. After adjustment for patient and treatment characteristics, all associations between body fat and HRQoL outcomes were attenuated and no longer statistically signi cant. Additionally

Discussion
In this cross-sectional study among 155 patients who underwent surgical treatment for stage I-III CRC, we found that a lower HGS was independently associated with poorer scores of almost all speci c HRQoL domains 2 to 10 years after CRC treatment and that lower mid-upper arm muscle area was found to be independently associated with more fatigue symptoms. Although body fat and lower mid-upper arm muscle area were found to be associated with poorer scores in most HRQoL domains in unadjusted analyses, these associations attenuated after adjustment for differences in age, gender, years since diagnosis, comorbidities, chemotherapy treatment, current smoking, moderate-to-vigorous physical activity, sedentary time, and total energy intake. This indicates that the associations can be explained by confounding by the aforementioned factors.
A main nding was that lower HGS -as a measure of muscle mass and physical functioning -was independently associated with poorer HRQoL outcomes. Patients with a better HGS had a signi cantly higher score in global quality of life, physical functioning and social functioning. Moreover, a better HGS was associated with fewer complaints of fatigue and disability in patients after CRC treatment. As we adjusted for several relevant patient and treatment characteristics, these associations are probably not explained by confounding, that is, by the fact that CRC survivors with lower HGS are e.g. older, are more often female or smokers, have more comorbidities, received chemotherapy more often, have less moderate-to-vigorous physical activity and more sedentary time, and have a lower total energy intake. These cross-sectional results are in line with ndings of a previous cross-sectional study that showed a signi cant association between lower HGS and lower global quality of life in frail elderly people after adjustment for age and sex [54]. Also, in prospective studies among patients with breast cancer and advanced small non-cell lung carcinoma, a lower HGS before treatment was related to lower HRQoL after treatment [32,33]. Similar prospective studies investigating a comparable direct association between HGS and patients' HRQoL in CRC survivors are lacking.
We found that lower HGS and lower mid-upper arm muscle area were independently associated with more fatigue symptoms, although effect sizes of lower mid-upper arm muscle area were small. Previous crosssectional studies found a similar association between lower muscle mass at time of diagnosis and the occurrence of cancer-related fatigue symptoms among advanced cancer patients [55,56]. Pathophysiology behind this could be found in pro-in ammatory effects of cancer-related cytokine changes which causes changes in activity of the hypothalamus-pituitary-adrenal axis, enhancing fatigue symptoms on the one hand and alter muscle metabolism leading to lower muscle mass on the other hand [57,58]. Both atrophy and loss of skeletal muscle bers occurs mainly in type II, fast twitch glycolytic bers, resulting in relatively higher density of remaining type I, slow twitch oxidative bers with a supposed preservation of muscle endurance but a reduction in muscle strength [59]. As HGS is mainly a measure for muscle strength instead of muscle endurance, a reduced HGS can be expected in patients with a lower muscle mass.
Because of our cross-sectional design, it is hard to draw rm conclusions based on both our data and the current literature, as one may discuss the causality between body composition and physical functioning in CRC survivors. It can be hypothesized that cancer treatment-related effects alter body composition leading to a decrease in body strength which causes a loss of physical functioning. In addition, a loss of physical functioning as a result of cancer-related adverse effects may lead to an altered body composition and thereby loss of muscle strength [14]. The direction of a possible causal relation between body composition and physical functioning should be investigated in future longitudinal research. Moreover, a previous cross-sectional study of our research group found that a higher level of physical activity was associated with better HRQoL, physical function and wellbeing, and less CRC-speci c symptoms after treatment [9,60]. This emphasizes the importance of optimizing and maintaining a good body composition by encouraging physical activity among CRC survivors. Multimodal and customized re-and pre-habilitation programs may be future solutions for the improvement of cancer patients' physical and psychological resistance on the unfavourable consequences of cancer and its treatment [61,62].
Our study was strengthened by the use of validated measurements for collection of both anthropometric and HRQoL data, especially the standardized collection of several anthropometric measures. Notwithstanding, our study was limited by its cross-sectional design whereby the previously stated causality cannot be assumed and a possible association with pre-existing poor body measurements and a possible in uence of cancer treatment is lacking. Another limitation of our study was the possibility of selection bias as survivors with an already unfavourable body composition or an already existing lower HRQoL may have been less likely to participate, which could have decreased the magnitude of the associations we observed. Although associations between anthropometric measures and HRQoL were adjusted for age, gender, years since diagnosis, comorbidities, chemotherapy treatment, current smoking, moderate-to-vigorous physical activity, sedentary time and total energy intake, residual confounding might play a role.
Results of this cross-sectional study have shown that HGS in CRC survivors, as a measure of physical functioning, is independently associated with several HRQoL outcomes. Future prospective research should investigate underlying and causal factors in the association between HGS and HRQoL after CRC treatment.
These studies are necessary to investigate whether certain anthropometric measurements are prognostic factors Flow diagram of inclusion of individuals into the cross-sectional part of the EnCoRe study and analyses presented in this paper.