Non-invasive Approaches to Identify the Cause of Fatigue in Inflammatory Bowel Disease Patients.

Background information Inflammatory bowel disease (IBD) fatigue is, with diarrhoea, the most common medical symptom in patients with active Crohn's disease (CD) and the second commonest complaint after arthralgia in CD patients in remission. The IBD priority-setting partnership between the James Lind Alliance and the British Society of Gastroenterology has highlighted that IBD fatigue is a top 10 UK research priority.

The prevalence of fatigue ranges from 41-48% when in remission, increasing up to 86% when the disease is active. Chronic fatigue is associated with an impaired health-related quality of life in IBD and a reduced physical activity, which in paediatric cohorts plays a major role in normal growth development. In the absence of reversible clinical causes such as anaemia, inflammatory burden or nutrient deficiency, it is as yet unclear what causes fatigue in these patients. Present approaches are limited in offering clues to potential therapeutic approaches or accurate clinical evaluation for CD patients. This is likely due to a lack of a multicomponent approach and poor methodological sensitivity and reliability when assessing fatigue. A multidimensional concept of physical, cognitive and affective components have been suggested for IBD fatigue. Yet, it remains unclear how to treat IBD fatigue. Exercise therapy, psychosocial interventions and solution-focused therapy, are all suggested as treatment options but none have been shown to be convincingly effective.

Fatigue arises from peripheral and central processes, which are both equally important. While peripheral fatigue is ascribed to muscle aetiology, central fatigue is a progressive exercise-induced reduction in voluntary activation to the muscle due to failure of the central nervous system (CNS). Present approaches have been limited in failing to adequately quantify peripheral fatigue at a muscle level and central fatigue at the level of the CNS.

Peripheral fatigue: the need of muscle activity

Physical fatigue is linked to reduced muscle mass and quality (sarcopenia) and consequently reduced muscle function. Drivers of muscle mass loss include physical inactivity and anabolic resistance, which to a large extent blunt the stimulatory effect of protein nutrition on muscle protein synthesis. In keeping with this, patients with active CD show a significant decrease in expression of hypertrophy skeletal muscle signalling pathways, but no changes in the expression of atrophy signals.

Poor muscle quality is an important driver of metabolic dysfunction that is often overlooked as a driver of peripheral fatigue. In preliminary unpublished data, the investigators have shown for the first time that skeletal muscle of patients with active CD exhibits blunted glucose uptake in response to glucose feeding, suggesting diminished muscle metabolic quality in CD patients.

In fatigued IBD subjects, medium-to-large effect sizes for impaired aerobic fitness (VO2max), muscle strength and physical activity are seen in subjects with fatigue when compared to subjects with no reported history of fatigue and the normal population. Physical inactivity in quiescent CD disease could be an explanation underpinning these observations, pointing to inactivity secondary to disease burden being the aetiology of fatigue in CD, much as in chronic obstructive pulmonary disease (COPD). Physical inactivity has already been shown in CD in both paediatric and adult cohorts and has been significantly correlated to disease activity but is still prevalent in remission. Although this might imply circularity in the aetiology of fatigue in the context of physical inactivity, it is unlikely that these two factors have a simple direct cause and effect relationship. Indeed, when habitual physical activity is matched between volunteers, skeletal muscle fatigue is still most prevalent in CD, and predominantly in those subjects who complain of fatigue. Collectively, these data are indicative that both reduced cardio-respiratory fitness and peripheral muscle deconditioning (possibly linked to reduced muscle mitochondrial density) are contributing to early fatigue development in IBD patients who report symptoms of fatigue.

Central fatigue: the need of cardio-respiratory regulation

Skeletal muscle, the cardio-respiratory system and the CNS are interlinked. In preliminary data, it has been shown for the first time that it is possible to assess change of brain perfusion (or cerebral blood flow; CBF), an indirect measurement of haemodynamic measurement of brain neuronal activity, using functional MRI techniques, in relation to decreased cardiovascular fitness with age. These data have demonstrated that CBF measured directly using arterial spin labelling (ASL) or indirectly using blood-oxygen-level-dependent (BOLD) was reduced in older people with reduced cardiorespiratory fitness, and who are likely more fatigable than young volunteers. This suggests that the coupling between CBF and brain activity can be modulated by cardiac output, and thus be manipulated to improve brain function linked to central fatigue in CD patients.

This highlights the potential of non-invasive functional brain MRI techniques to determine the effect of brain perfusion during exercise on central fatigue in CD patients. Furthermore, these quantitative approaches could be used to determine whether improvements in cardiovascular fitness and appropriate downstream exercise or pharmacological interventions can alter central fatigue.

1.Aims and purpose of the proposed investigation

Hypothesis: The investigators hypothesize that the aetiology of IBD fatigue lies collectively with peripheral muscle deconditioning and brain central fatigue, the latter being secondary to a reduction in cardio-respiratory fitness and/or cardiac output, such that both central and peripheral fatigue contribute to fatigue development in IBD.

Aims:

To address this hypothesis, the investigators aim to recruit a cohort of fatigued and non-fatigued CD patients in remission and age-, gender- muscle mass-, and physical activity-matched healthy volunteers. Fatigue symptoms will be defined through relevant questionnaires at screening. The investigators will subsequently objectively quantify muscle strength of the knee extensors during a maximal voluntary isometric contraction, and fatigue development in the quadriceps muscle group. The latter will be defined as the loss in muscle peak-torque development over 20 consecutive isokinetic knee-extensions. Peripheral (muscle) and central (CNS) roles in fatigue aetiology will be investigated. Peripheral muscle deconditioning will be investigated by quantifying muscle phosphocreatine (PCr) re-synthesis during recovery from exercise using 31phosphorous MRS, while the potential of central fatigue will be examined through various measures of CNS perfusion and fractional brain oxygen extraction using MRI.

Detailed plan of investigation and scientific procedures/methodology

The investigators aim to recruit 32 patients with CD in remission, with and without a history of IBD fatigue (n=16 for IBD fatigue cohort and n=16 for no IBD fatigue cohort) within 36 months. These cohorts will be compared to 16 age-, gender- muscle mass-, and physical activity-matched healthy controls (from our existing Nottingham Digestive Diseases Biomedical Research Unit and MRC/ARUK databases) to undergo the experimental protocol outlined below. The investigators will specifically match for physical activity to investigate its potential primary effect on fatigue.

Objective 1 (Day 1): To determine muscle strength, fatigue and cardiorespiratory fitness.

Following consent, CD and healthy matched controls volunteers will perform 3 maximal isometric knee extensions of the quadriceps muscle group (each separated by 30s) to determine muscle strength. Following this, volunteers will perform 20 maximal single leg knee extensions on a Cybex isokinetic dynomometer (Cybex Norm, USA) at a constant angular velocity of 90 degrees per second to determine muscle torque development. Each maximal voluntary contraction will be initiated from a position of 90 degrees knee flexion and continued through to the point of full knee extension. After each contraction, the leg will be returned passively to the start position (approximately 1.1s) from which the next contraction will be immediately initiated. The outcome measure and the gold-standard definition of fatigue in this objective will be defined as the decline in peak torque generation for each contraction over the 20 knee extensions. Maximal voluntary contractions at this velocity has previously been shown to result in recruitment of both type I and type II muscle fibres of the quadriceps muscle group. Thirty minutes later, volunteers will undergo a continuous, incremental, supine exercise test using an MR compatible step ergometer (Ergospect diagnostic pedal). Exercise will commence at a low workload, which will increase every 3 min to determine peak VO2 during supine exercise using on-line expired gas analysis (Cosmed, Rome, Italy). Carbon dioxide (CO2) production and oxygen (O2) consumption will be measured throughout the test and used to measure peak oxygen uptake and respiratory exchange ratio (RER). These data will be normalized to body composition as measured by dual-energy X-ray absorptiometry (DEXA; Lunar Prodigy, GE Medical Systems, Bucks, UK) and will provide a gold-standard measure of cardiorespiratory (or aerobic) fitness.

Objective 2 (Day 2): To determine brain perfusion and central fatigue.

Recently, in collaboration with researchers at the Sir Peter Mansfield Imaging Centre at the University of Nottingham, the investigators have developed a protocol to quantify the cortical vascular response to exercise in volunteers and its modulation in ageing and cardiorespiratory fitness. Here, the same approach will be used to assess cortical vascular response to exercise in the brain regions associated with central fatigue of CD patients.

To study this mechanism in CD patients, the investigators propose to measure cardiac output and, cerebral and regional perfusion using Blood-Oxygen-level-dependent (BOLD) measures and arterial spin-labelling (ASL)-MRI as well as fractional oxygen extraction using TRUST-MRI respectively during supine exercise of the quadriceps with the MR compatible ergometer (Ergospect diagnostic pedal) at given values of relative VO2max (as measured in Objective 1).

This multimodal approach will determine a potential uncoupling between brain perfusion and neuronal activity and the role of central hypoxemia in central fatigue of IBD patients compared to healthy controls. Furthermore, these data will be correlated to cardiorespiratory fitness, muscle strength and fatigue, and will determine the predominant role of central processes in fatigue of IBD patients. If these brain vascular measures are associated with fatigue development, then this index of neurovascular coupling could benefit from exercise training in a similar way to muscle tissue.

Objective 3 (Day 2): To determine muscle deconditioning.

To maintain energy (ATP) during exercise, skeletal muscle regulates oxidative (mitochondrial) and non-mitochondrial production of ATP. In healthy individuals, the relative contribution of non-mitochondrial ATP production to total energy production is generally dictated by the exercise intensity, which determines the magnitude of accumulation of metabolic by-products that are directly linked to muscle discomfort and fatigue. In disease states, exacerbated fatigue and muscle pain has been attributed to reduced muscle blood flow and/or reduced mitochondrial mass.

To separate these two processes in CD patients, limb blood flow will be occluded with a cuff, whilst participants perform repeated plantar flexion exercise in a MRI scanner. This will drive PCr, (the available reserve of ATP) to zero. Specifically, participants with a lower limb cuff will perform 30 contractions over a 5-minute period using a MR compatible ergometer. Once the cuff is removed, they will undergo non-invasive and well-established in vivo 31P magnetic resonance (31P MRS) measurements on the gastrocnemius to determine PCr re-synthesis rate during resting recovery. 31P MRS data will be averaged, processed and quantified using in-house methods.

Since the rate of PCr re-synthesis is directly proportional to mitochondrial mass and oxygen delivery is not limiting during exercise recovery, this will allow us to measure muscle metabolic deconditioning in vivo.

Sample size

1. Cardiorespiratory fitness: The investigators have previously shown that the VO2max of a healthy volunteer is 3.65 (0.5) L/min, whilst the VO2max of a fatigued IBD patient is 1.99 (0.44) L/min and a non-fatigued patient is 2.43 (0.75) L/min. Assuming a power of 80% and α=0.05, 6 subjects in each group are needed to show a significant difference between the healthy volunteers and the non-fatigued cohort, and 13 patients in each group are needed to show a difference between the 2 CD cohorts.
2. Muscle Fatigue: Peak muscle Torque extension 180⁰/s (Nm) is lower in fatigued IBD vs non-fatigued IBD patients (60.73 vs 73.5, SD 12.3 respectively). Assuming a power of 90% and α=0.05, 16 CD subjects in each group are needed to show a difference between fatigued and non-fatigued CD cohorts. Importantly, the rate of muscle fatigue (defined as the gradient of force decrease from repetition 2 to 30 of an exercise protocol) is significantly greater in CD than healthy volunteers (5.2 Nm/min vs 1.3 Nm/min, SD 3. In this instance, assuming a power of 90% and α=0.05, 11 subjects in each group are needed to show a difference between CD and healthy volunteers.
3. Muscle conditioning: recent studies have shown that Pcr re-synthesis rate is 27.9 ± 0.6 s in lean sedentary subjects, whilst being 36.5 ± 2.4 s in Type 2 diabetes mellitus patients (T2DM). Thus, assuming a power of 100% and α=0.05, 3 subjects in each group are needed to show a difference between healthy volunteers and potentially fatigued cohort such as T2DM patients.
4. Anabolic resistance: Furthermore, our recent work on anabolic resistance has already shown significant differences in muscle quality between CD and healthy volunteers (significant difference observed n=5 patients). Percentage forearm fatigue in the dominant arm is 26.75%±11.1 and 18.3±7.8 in CD and controls respectively. Assuming a power of 80% and α=0.05, 11 subjects are needed in each arm.
5. MRI: It is not possible to power this exact MRI study as pilot data in these IBD cohorts has not been collected. However, the investigators have previously shown that brain perfusion in the young increases by 18±6% for 30% VO2max during a cardiorespiratory fitness test. Whilst in the elderly, it increases by only 3±4 %. Assuming a power of 100% and α=0.05, 10 subjects in each group are needed to show a significant difference between healthy volunteers (young) and a potentially fatigued cohort (elderly). Further, for fMRI BOLD analysis it is generally considered that a minimum of 12 subjects is needed in each cohort, with improved sensitivity of MR scanners (and thus higher temporal signal-to-noise ratio (SNR) now allowing improved individual subject analyses.

For the above reasons, the investigators feel that 16 subjects in each group (fatigued CD, non-fatigued CD and healthy volunteers) will suffice to show a difference in our chosen objective endpoints.

Data analysis Fatigue will be measured as the decline in peak torque development over 20 isokinetic knee extensions as described above. PCr re-synthesis rate and fractional brain oxygen extraction will be the primary endpoints in objectives 2 and 3.

31P MRS data will be analysis using standard NMR software (jMRUI Software for the clinical and biomedical MRS). Precisely, all spectra from the time series will be apodized (15-20 Hz Lorentzian shape). From these spectra, the signal intensities of 5 peaks will be determined after lineshaping. The pH will be determined prior to and at the end of the exercise protocol by measuring the frequency shift between the PCr and Pi, whilst in-house written code in MATLAB (MathWorks Inc., Matrix House, Cambridge, UK) will be used for further calculations of the PCr recovery rate. In this case, peak area intensities of the PCr peak from jMRUI will be used as input in the script to fit the data time courses using a mono-exponential function: PCr (t) = PCr (t0) + C (1 - e-t/τ); with PCr being the intensity level of the PCr peak, C a constant, τ [s] the rate constant for PCr recovery, and t [s] time, starting at the end of the exercise protocol (t0). The Reproducibility of 31P MRS measures of muscle energetics has been recently evaluated in trained human and show that coefficient of variation during exercise (27%) and at rest (8%) was relatively low, and difference were detectable with 15 volunteers. This confirms that 31P MRS approach provides reproducible and experimentally useful measures of PCr re-synthesis rate in human skeletal muscle.

BOLD MRI data will be analysed using standard neuroimaging packages (Statistical Parametric Mapping (SPM); Wellcome Department of Imaging Neuroscience and FMRIB Software Library software. Images will be motion-corrected, normalized to a standard Montreal Neurological Institute template to facilitate inter-participants averaging. For all data acquisitions, the investigators will record the vectocardiogram (VCG) and respiratory motion to allow data to be corrected for physiological noise. Contrast vectors will be formed to identify the CNS response to the exercise stimulus. The investigators will form statistical maps of those brain regions which show a decrease/increase in BOLD signal following the exercise intervention. These statistic maps will be used in contrasts by intervention and between participant groups. Results will be pooled in a second level, random effects group analysis (RFX) and active brain areas displayed at a false discovery rate corrected probability P < 0.05, k > 5. These analyses have been previously carried out and published by our group. ASL data will be motion corrected and modelled to quantify perfusion in ml/100g/min, and global and regional perfusion measures compared before and during the exercise intervention. TRUST data will be used to compute fractional oxygen extraction before and during exercise intervention. Data will be presented as mean +/- standard error of the mean (SEM). Binary comparisons will be carried out by t-tests while multiple comparisons will be carried out with an analysis of variance with a post-hoc Bonferroni correction in between the 3 groups. The effects of cytokines, vitamin D and disease duration will be investigated through a correlation co-efficient and post-hoc multivariate analysis.

The investigators will stratify subjects based on the quantitative outcome measures: PCr re-synthesis rate, cardiorespiratory fitness as measured through VO2max, decline in muscle output and change in regional perfusion and fractional brain oxygen extraction irrespective of their fatigue status. This will aid our understanding regarding which alterations in physiology and metabolism are best associated with specific fatigue symptoms. Doing this will allow objective quantification of a subjective cut-off for the IBD Fatigue scale, thereby better defining fatigue.