Building a biopsychosocial model of cancer-related fatigue: the BIOCARE FActory cohort study protocol

Two hundred women newly diagnosed with breast cancer by an oncologist and satisfying inclusion and exclusion criteria (Table 2) will be informed of the study protocol. If interested, patients will receive an information letter from their oncologist and will be given the opportunity to ask any questions pertaining to it. After a 48 h period of reflection, study coordinators will phone the patients to confirm their participation, obtain a written consent form and schedule their first visit.

This prospective multicentric (Clinique Victor Hugo-Centre Jean Bernard, Le Mans, France and Centre Hospitalier du Mans, Le Mans, France) study will consist of an 18 month follow-up of a women cohort with breast cancer. Four visits will be scheduled (the study timeline is presented in Fig. 1). Detailed patient assessments will be performed before treatment begins (visit 1: diagnosis); during the week after completion of the first-line treatment, or 6 months after diagnosis, whichever comes first (visit 2); and then 12 and 18 months after diagnosis (visit 3 and 4, respectively). This study has been approved by the French ethics committee of human research CPP SUD EST VI (IDRCB: 2019-A02525–52) and will be performed according to the Declaration of Helsinki. Furthermore, this study protocol has been written in accordance with the SPIRIT guidelines (SPIRIT Checklist provided in Additional file 1) and is registered in a database (ClinicalTrials.​gov, NCT04391543, May 2020). In order to longitudinally assess the evolution of biopsychosocial dimensions of CRF, the same procedure will be followed for each visit. Each session will last 1.5 h and will be composed of anthropometric data collection, a semi-structured interview, cognitive tests, postural control tests, neuromuscular fatigability tests and a cardiorespiratory fitness test. Clinical and biological data will be collected during medical follow-ups. In addition, participants will be asked to wear an actigraphy device (received by mail one week prior to each experimental session) and to complete questionnaires to assess psychological aspects and quality of life.

Body height, body mass, lower limb fat mass (using impedancemetry) and calf circumference (CC) will be measured. CC will be recorded to assess the dominant lower limb volume using the truncated cones technique [39] by dividing the lower limb volume into a series of segments. Then, the lower limb volume will be used to normalize maximal strength of participants. CC under 31 cm will serve as a clinical indicator of sarcopenia [40].

To investigate the social dimension of CRF, a semi-structured interview will be conducted (duration ~ 30 min). Demographic (e.g., age, gender, socio-professional category, residence area, income) and sociability (e.g., family, friendly, professional) will be addressed. The interview will be introduced by an open question on a typical week to then develop the sociability. The entire interview will be audio recorded, fully transcribed and then will be the subject of a lexicographical treatment for rating quantity and intensity of sociability (on a scale from − 5 to 5). In order to build a typology of patients’ sociability, researchers will also assess the level of social ties’ perception. Lastly, family, friendships, professional sociabilities and income will be considered in the SEM analysis.

The Montreal Cognitive Assessment (MoCA), the Trail Making Test (TMT) and the Stroop test will be completed by participants to assess cognitive functions. The MoCA is separated into several tasks including visuospatial/executive functioning, naming, memory, attention, language, abstraction, delayed recall and orientation. For example, the participants will be asked to connect numbers and letters in a defined sequence, count backwards from 100 by increments of 7, draw different figures and perform word associations. The MoCA total score (/30) is obtained by summing the scores from each item; a cut-off score below 26 is an indicator of mild cognitive impairment [41]. The TMT, evaluating processing speed and cognitive flexibility, consists of connecting numbers in ascending order (Part 1) and then in connecting numbers and letters in both ascending and alphabetical order (Part 2), as fast as possible [42]. Time needed to complete each part and number of mistakes will be recorded to assess performance. The Stroop test is based on color recognition (i.e., red, green, yellow and blue) with interference in three parts, each composed of 6 lines with 4 items per line (colored sticker or word). The first part is composed of colored stickers (control), the second has colored words (low interference) and the last part has color-words printed in a color not denoted by the name (high interference). For each part, participants will be advised to name the ink color for each word, regardless of the semantic content, as quickly and accurately as possible. The time to complete each part is measured to attribute an interference score used to assess response inhibition, which has been related to cognitive functioning in daily life [43]. The TMT and the Stroop test will be implemented on Inquisit software (Millisecond Software, LCC, Seattle, USA). For SEM analysis, MoCA total scores (/30), TMT time (s) and Stroop interference scores (%) will be considered.

Participants will be asked to stand up barefoot, as still as possible, during 60 s on a pressure distribution measurement platform (FDM-S, zebris Medical GmbH, Isny, Germany). Feet position will be standardized using markings to keep both heels spaced apart by 10 cm and an angle of 15° between both feet. Participants will be advised to keep arms alongside the body during recording. Four conditions composed of two trials of a simple task and two trials of a dual task (counting backwards by 2 from a number close to 100) with eyes open or closed will be counterbalanced and interspersed with a 2 min rest period. In eyes open conditions, participants will be asked to stare at a visual marker placed 2 m in front at eye level. For the dual task, participants will be asked to count as fast as possible whilst remaining as immobile as possible. The numbers of answers and errors will be recorded. Temporal analysis will be performed in order to investigate anterior-posterior and medial-lateral center of pressure (COP) distance, mean and maximal COP velocities and 95% confidence ellipse area. In addition, the frequency domain measures (relative power in < 0.5 Hz; 0.5–1.5 Hz; > 1.5 Hz frequency bands) will be computed. 95% confidence ellipse area (mm²) will be considered in the SEM to represent the physical function latent variable, while maximal COP velocities (mm/s) will be considered representative of the emotional function latent variable, as it is a relevant hallmark of depression-related psychomotor retardation [44].

Throughout this test, participants will remain in prone position on a patient table, with a fully-extended knee and an ankle angle of 90°, their foot securely blocked at the metatarsal level in a custom–made device enabling isometric strength measurement with a load cell (LSB350, Futek, Irvine, USA). Firstly, optimal electrical stimulation intensity will be determined to set the supramaximal intensity used during the subsequent neuromuscular assessments by progressively increasing the current (from 20-mA to 200-mA, with a 20-mA increment) until there is no further increase in the evoked isometric twitch response. The last intensity obtained will be further increased by 20% to ensure stimulus supramaximality. All electrical stimulations will be delivered with a constant current (Digitimer DS7A-H, Hertfordshire, UK) using square-wave stimuli of 200 μs duration with a maximal voltage of 400-V and via rectangular self-adhesive electrodes (5 × 10 cm, Compex). The cathode will be placed over the gastrocnemii (~ 5 cm distal to the popliteal fossa) and the anode over the soleus (~ 10 cm proximal to the calcaneus). Secondly, participants will accomplish a standardized warm-up (10 isometric contractions of 4 s at 50% of their maximal strength with 4 s of recovery in between), followed by 2-min of rest and then by the maximal voluntary contraction (MVC) measurement (two MVC of 4 s separated by 2 min; if the difference between these MVCs is superior to 5%, a third one will be performed). The higher peak force produced will be considered as the MVC in non-fatiguing conditions. Next, pre-fatigue neuromuscular functions will be assessed on a third MVC, using a 100-Hz doublet during the force plateau and a stimulation sequence on the relaxed muscles beginning 2 s after the end of contraction: a 100-Hz doublet, a 10-Hz doublet and a simple stimulation, interspersed by 3 s. The fatiguing exercise will be composed of 62 isometric MVC in ankle plantar flexors with the dominant leg (Fig. 2). Each MVC will last 4 s with 1 s rest; duty cycle will be ensured using a metronome with visual and sound signals. To avoid pacing strategies, participants will not be informed of the time remaining or the number of MVC performed [45]. Investigators will use verbal encouragements for participants to contract as strong as possible during MVC. The post-fatigue neuromuscular functions will be assessed on the last MVC of the fatiguing exercise (62nd) with the same stimulation procedure as in pre-fatigue condition. On the 60th MVC, neuromuscular functions will also be tested but only with two 100-Hz doublets (during the force plateau and on the relaxed muscles 2 s after the MVC end).

Peak force occurring during each 4-s MVC set will be recorded and the force-time relationship asymptote (F_A; expressed in percentage of the MVC in non-fatiguing conditions) will be used to represent a neuromuscular fatigability threshold, above which fatigability drastically sets in [20, 46, 47]. Then, voluntary activation (VA) and evoked forces by a 100-Hz doublet (Db100) will be determined using the interpolated twitch technique [48] and expressed in pre-post fatiguing test differences, normalized to pre-fatiguing test values. F_A (% MVC), VA (%) and Db100 (%) will be considered in the SEM.

After a five-minute recovery period, a blood drop (10 μL) will be taken from the earlobe and analyzed extemporaneously (within 15 s) to obtain a resting blood lactate concentration ([Lactate]_b) (Lactate Scout 4, EKF diagnostics, Cardiff, UK). Then, participants will perform a submaximal incremental exercise on a cycle ergometer (model 928E, Monark, Varberg, Sweden), with saddle and handlebar heights adjusted for each patient. Throughout the test, participants will be encouraged to maintain a constant cadence (approximately 60 rpm) and heart rate (HR) will be recorded with a heart rate monitor (HRM-Dual, Garmin, Olathe, USA). After a 3-min standardized warm-up at a rate of perceived exertion (RPE) of 8–9 (Borg 6–20 scale), the test will start at 20 W. Every 2 min [Lactate]_b will be measured as previously described and then intensity will be increased by 10 W steps. Just before the end of the step, participants will indicate RPE (Borg 6–20 scale). Exercise will be stopped as soon as i) [Lactate]_b ≥ 4 mmol. L^− 1; ii) participant RPE > 15; or iii) participant will no longer be able to maintain the 60 rpm cadence. Participants will be able to interrupt the exercise when they wish to do so, particularly in the case of nausea, chest pain or dyspnea. This test will be used to determine the physiological and biomechanical parameters (i.e., HR, power output associated with the first lactate threshold (LT1; defined as the first inflection point in the lactate concentration ([Lactate]_b) curve). Power output at LT1 (W) will be considered in the SEM.

As experimental sessions will accompany medical follow-ups, clinical and biological data will be also recorded. Cancer stage and details related to treatments (e.g., type, duration and dose) will be recorded in a medical file by their oncologist. Venous blood samples will be collected by medical staff. Plasma will be obtained after a 10 min centrifugation at 4 °C, then divided into aliquots and stored at − 80 °C until analysis. Following assays will be performed on plasma for the assessment of inflammation (IL-6, TNFɑ, IL-8, IL-1β). Abdominal computed tomography (CT) images at the L3 level will be taken by medical staff on two occasions (Visit 1 and Visit 2). The surface of the muscular tissues will be selected according to the CT Hounsfield unit and normalized to stature in order to calculate the lumbar skeletal muscle index (LSMI). A cut-off value (LSMI < 38.5cm²/m²) will be used to characterize sarcopenia [49]. Cancer stage and treatment; sarcopenia; and IL-6, IL-1β, TNFɑ (pg/mL) concentrations will be considered in the SEM.

One week prior to each experimental session, participants will receive by post self-assessment questionnaires about quality of life (EORTC QLQ-C30), CRF (EORTC QLQ-FA12), coping strategies (Brief Cope) and anxiety-depression symptoms (The Hospital Anxiety and Depression Scale - HADS). Participants will be asked to complete questionnaires, alone in quiet conditions, and bring it back the day of the experimental session. The subsequent instructions will be provided: “Please answer all questions yourself by circling the number that best applies to you. The information provided will remain strictly confidential. Take as much time as necessary. There are no “right” or “wrong” answers.”

The EORTC QLQ-C30 Fatigue scale score (FA item) will be used to assess the general degree of CRF (ranging from 0 to 100; with higher levels indicating a greater degree of CRF), with a threshold for clinical importance of 39 [50]. EORTC QLQ-FA12 subscales will be used to assess physical, cognitive and emotional dimensions of CRF [51]. The HADS subscale scores (/21) will be used to assess anxiety and depression, respectively [52]. In the SEM, general degree of CRF (%); dimensions of CRF (%) (physical, cognitive and emotional); coping scores (/8); and anxiety and depression scales (/21) will be considered.

One week prior to each experimental session, participants will be asked to wear a portable device (eTact®, BodyCap, Caen, France) containing a tri-axial accelerometers enabling to record every acceleration from the body (sampling frequency: 25 Hz, measurement range: 0.1-2G, sample measurement: 1 min). They will be advised to wear the device continuously for 7 days (days and nights) regardless of daily activities. Sleep characteristics will be obtained using the eTact® Analysis software (BodyCap, Caen, France), which assess the individual sleep quality based on the actigraphy data collected over selected periods. Considered sleep parameter will be sleep time (i.e, duration of effective sleep period) and sleep efficiency (i.e., percentage of time spent sleeping during the rest period). From the software, total activity duration during wake period will be calculated as the accumulated time in each activity intensity band (i.e., low, moderate and intense). In addition, participants will be asked to complete a sleep diary (including estimated sleep onset and offset hour) during the actigraphy measurements. Total activity duration (min) and sleep efficiency (%) will be considered in the SEM.

A sample size of n = 200 patients has been chosen on the basis of resource constraints [53]. 200 correspond to a 15% inclusion rate of the patients meeting the inclusion criterion during a 24 months inclusion period and including a 20% potential drop-out rate. Considering this sample size, a sensitivity power analysis performed with G-power (F tests family, linear multiple regression, R² increase) to estimate the smallest effect size we could detect with 90% power regarding our main independent variable (i.e., CRF) at the level of analysis with the lowest power (i.e., level 2, between-person level). This analysis, with a power of 0.90, an α level of 0.05 and 6 predictors (i.e., 6 latent variables), revealed that the smallest effect size we could detect with 90% power, at this level of analysis, is f² = 0.09, which correspond to a small to medium effect.

Hypotheses will be addressed by constructing a biopsychosocial model of the relationships among biological, physical, emotional, cognitive and social dimensions of CRF using multilevel SEM. SEM is particularly well-suited to this type of analysis because it simultaneously accounts for multiple interactive relationships among variables, easily handles multiple sources of variance, and permits testing of hypothesized directional relationships [54, 55]. Consequently, the results of SEM analyses facilitate inferences regarding the relationships among variables [56]. Relations between the manifest variables and their underlying latent constructs in a hypothesized biopsychosocial model predicting CRF are presented in Table 3. The model will be tested with the maximum likelihood method using Lavaan R package for SEM [57]. A primary cross-sectional analysis will be completed using data collected at each visit, and then subsequently by using the longitudinal data set. The models’ fit was assessed by examining the minimum discrepancy (CMIN/DF), the probability level (p-value), the Bentler-Bonett normed fit index (NFI), the comparative fit index (CFI), the Tucker–Lewis-Index (TLI), and the root-mean-square error of approx- imation (RMSEA). A satisfactory model fit is indicated by a CMIN/DF ratio below 2.00 [58], a p-value over 0.05 [59], a NFI over 0.95 [60], a TLI over 0.90 [61], a CFI over 0.93 [62], and a RMSEA below 0.05 [63]. All non-significant paths were deleted according to the methods described by MacCallum [64].