Predictive Utility of a Short-term Improvement in Objectively Measured Physical Activity in Rheumatoid Arthritis (PUSIPA)

INTRODUCTION Chronic pain, fatigue, and decreased functional ability are challenging for patients suffering from Rheumatoid Arthritis (RA). Approximately half of this patient population report significant pain, extensive fatigue, and limitation in activities of daily living, including workability, despite comprehensive anti-inflammatory treatment. RA patients have an increased risk of cardiovascular disease similar to diabetes patients, and RA is associated with excess mortality. Physical activity (PA) reduces the cardiovascular risk through various pathways, including the enhanced activity of endothelial nitric oxide synthase, increased blood flow, and improvements in anti-oxidizing mechanisms. The fourth leading risk factor for death globally is physical inactivity, which is responsible for approximately 6% of deaths.

PA has been widely studied in RA, where both self-report questionnaires and performance-based assessments, such as accelerometry, have been applied. The reliability of accelerometry is well established; however, patient reporting of PA is subject to recall bias and biased estimation. Studies have demonstrated only a modest association between self-reported and performance-based measures of PA.

Several accelerometry studies of RA patients have been performed with partially conflicting results. Most studies have proven RA patients to be less physically active than healthy controls, and only a minority of RA patients seem to fulfil the World Health Organization (WHO) recommendations regarding the level of daily physical activity. However, some studies find no differences between RA patients and healthy controls. Several studies have examined differences in PA among RA patients regarding disease activity. They all found PA to be inversely correlated to disease activity or score on the Stanford Health Assessment Questionnaire Disability Index (HAQ-DI). However, most of these studies do not allow inferences about directions of the observed relationships given the study design applied. Only two studies (including n=18 and n = 30, altogether 48 patients) were prospective, leaving this question unanswered.

Rationale and Evidence-Based Research Concomitant fibromyalgia (FM) is a chronic pain syndrome characterized by widespread pain and tenderness, often accompanied by multiple symptoms from various organ systems. FM has been estimated to be present in between 15% and 20% of RA patients. As for RA, PA studies in FM patients assessed by accelerometry have concluded that FM patients are less physically active than healthy controls. Like RA patients, only a minority of FM patients fulfil WHO recommendations regarding daily physical activity, and a former study has shown that RA patients with concomitant FM have higher disease activity (DAS28-CRP) scores than sole RA patients. To our knowledge, no study has examined the impact of concomitant FM in RA patients regarding PA measured by accelerometry and the response to disease-modifying treatment.

Aims and Hypotheses The primary aim of this study is to investigate the prognostic value of immediate change in PA on treatment response among RA patients initiating their first biological treatment (according to Danish national treatment guidelines). The main hypothesis is that treatment response to biological therapy (defined as achieving DAS28-CRP<3.2 after 12-16 weeks, according to Danish national guideline) can be predicted from an initial higher increase in objectively measured PA, measured by accelerometer one week after first biological treatment onset compared to one week before biological treatment onset.

Secondary aims are to investigate whether the presence of concomitant FM (according to the 2016 criteria for FM), higher levels of CRP (above the median), and excessive fatigue (BRAF ≥ 6) affect the possible effect of being PA responder immediately after initiating biological treatment (i.e., testing the interaction between contextual factor and PA-group). The secondary hypotheses are that concomitant FM, lower CRP, and excessive fatigue are associated with a lack of improvement in objective measures of PA.

The investigators also aim to explore the association between objective measures of PA, measures of self-reported PA, functional ability, and fatigue on standardized disease-specific questionnaires. Finally, they want to examine whether any differences exist regarding PA measured one week after initiation of biological therapy compared to three months of biological therapy.

Objectives Our primary objective will be to compare the effect of an immediate significant improvement in the objectively measured physical activity, relative to individuals without, on the proportion of patients who subsequently achieve low disease activity (DAS28-CRP < 3.2) after at least 12 weeks in RA patients initiating their first biological treatment.

Furthermore, the investigators will explore whether the presence of:

• Concomitant FM (according to 2016 criteria) affects the clinical response (i.e., interaction between FM status and PA response)
• Higher levels of CRP (above the median) affect the clinical response (i.e., the interaction between CRP status and PA response)
• Excessive fatigue (BRAF ≥ 6) affects the clinical response (i.e., the interaction between fatigue status and PA response)

Other objectives include:

• Comparisons between objectively measured PA and self-reported PA (through a validated standardized questionnaire) on the ability to predict treatment response after at least 12 weeks
• To compare the effect of an immediate significant improvement in the objectively measured physical activity, relative to individuals without, on the proportion of patients who subsequently achieve ACR 20%, 50%, and 70% response
• To compare the effect of an immediate significant improvement in the objectively measured physical activity, relative to individuals without, on the proportion of patients who subsequently achieve enhancement related to patient-related outcome measurements, including HAQ-DI and VAS-scores of pain and fatigue.

Study design The study is designed as a prospective cohort study that follows RA patients from a week before the initiation of their biologic agent (baseline; t = 0 weeks) up to the first outcome assessment after at least 12 weeks from baseline.

Setting and source of data RA patients treated in Esbjerg and OUH rheumatic outpatient clinics who fulfil nationally endorsed criteria for initiating biological treatment and where such treatment decisions have been made (by patients' treating rheumatologists) are consecutively invited to participate in the study. If acceptance is achieved and inclusion and exclusion criteria are satisfied, informed consent will be obtained, and patients will be enrolled in the study. Patients are planned to be evaluated regarding physical activity during one week prior to the onset of biological therapy and during the following week after the onset of biological therapy. After 12-16 weeks, study participants are again evaluated regarding physical activity and disease activity.

Participants Eligibility criteria: Patients with RA (> 18 years of age) who are to be treated with a biological agent and who can read and understand Danish. Exclusion criteria: patients who are physically impaired (patients depending on assistance devices for walking), pregnancy, patients who have received corticosteroids two weeks prior to and during the period of the movement measurement, and patients who by virtue of illiteracy or cognitive impairment are unable to complete the questionnaire.

Variables and Outcome(s) Clinical data include personal data, assessment for concomitant FM through the FM score according to 2016 criteria for FM, PA measurements by accelerometery, self-reported evaluation of PA, and disease activity scores including patient-reported outcome measures (PROMs), fatigue evaluation (by Bristol Rheumatoid arthritis fatigue multidimensional questionnaire [BRAF-MDQ], clinical assessments, and laboratory data. PA is measured by accelerometry, where patients' actual physical activity is quantified. The time the participant is lying, sitting, standing, walking, running, bicycling, and swimming is recorded for each activity by the sensor during the period of measurement. The percentage of time a subject exhibits MVPA, that is, the percentage of time a patient is in motion (i.e., walking, running, bicycling, or swimming), can thereby be derived. MVPA is measured before, one week after, and finally 3 months after initiation of biological therapy. The investigators will define the MVPA ratio MVPA AFTER / MVPA BEFORE, where "after" reflects time spent in MVPA one week after onset of biological treatment and "before" refers to time spent in MVPA one week before biological treatment onset. We define an MVPA ratio > the upper tertile of the cohort as representing a high MVPA ratio.

Predictors For the primary prognostic model, the primary exposure variable will be the MVPA ratio measured as MVPA AFTER / MVPA BEFORE, reflecting the time spent in MVPA 1 week after and before the onset of biological treatment. Patients will be dichotomized into two groups based on the upper tertile value of the MVPA ratio. Study participants with improvement corresponding to the upper tertile of MVPA ratio will be considered "PA Improvers" while other patients will be considered "PA Failures".

Data sources and measurements A sensor named "SENS motion" will be attached to the patient's left thigh to evaluate movement. The SENS motion is a small device that measures patients' actual physical activity. The time the subject is lying, sitting, standing, walking, running, bicycling, and swimming is recorded for each activity by the sensor during the period of measurement. In this manner, an overview of patients' actual physical activity is derived.

All enrolled patients are planned to be evaluated regarding movement during three periods lasting one week each. The "SENS motion" device will be attached to the patient one week prior to biological treatment onset defining period number one. Period number two is defined by the time of biological treatment initiation and during the following week. The SENS motion device will then be detached. At the 3-month follow-up, another SENS motion device is attached to the patient's left thigh, generating data regarding actual movement as earlier explained. Final data collection is completed 1 week after the scheduled visit to the rheumatic outpatient clinic, and the detachment of the SENS motion device is performed.

All study subjects will, before attachment of the Sens motion device, fill in the international physical activity questionnaire (IPAQ), which is a standardized and validated tool measuring PA. Patients will fill in the questionnaire before initiating biological therapy, when detachment of the Sens-motion device is performed the first time (after one week of biological treatment), and at the 3-month follow-up. Furthermore, all patients will, at enrolment and at 3 months follow-up, be evaluated for concomitant fibromyalgia according to 2016 criteria. Before and 12 weeks after the initiation of biological therapy, patients will fill in the standard questionnaires used for evaluating treatment response in RA patients, and DAS28-CRP calculations will be performed. The questionnaires include the HAQ-DI and VAS scores regarding pain, fatigue, and overall well-being. Furthermore, to evaluate the level of fatigue more thoroughly, patients will fill in the BRAF-MDQ simultaneously. Finally, all patients will, at inclusion, be evaluated regarding potential confounders as outlined earlier.

Based on the percentage of time a patient exhibits MVPA; that is the percentage of time a patient is in motion (i.e., walking, running, bicycling, or swimming) during one week before initiating biological therapy and during the following week after onset of biological therapy (i.e., MVPAAFTER / MVPA BEFORE) patients will be dichotomized in groups based as earlier explained.

Power and Sample Size Considerations

Based on earlier studies, the investigators expect that patients reaching low disease activity will show increased time of MVPA, however, it is unknown if this will occur within one week of treatment onset. As explained earlier, no similar studies have ever been performed, and no formalized power and sample size calculation was performed. However, it is assumed that, on average, two-thirds of RA patients are responding to biologics (independent of prognostic phenotype), corresponding:

ptotal = 2/3 ptotal = (w1 × p1 + w2 × p2) / (w1 + w2)

Anticipating almost all the patients (say p1=90%) with an immediate response on PA after 1 week (33% of the total sample), will have achieved low disease activity after 12-16 weeks, that corresponds to p2=55% will achieve low disease activity in the other group (67% of the sample). From this, it is estimated that there would be a reasonable statistical power (>80%) to detect a difference between the groups if we enroll 60 patients in total (i.e., corresponding to a n1=20 vs. n2=40). Due to the risk of missing data from the application of a new device, we decided to aim for including 100 patients in total.