Subjective and objective levels of physical activity and their association with cardiorespiratory fitness in rheumatoid arthritis patients

Sixty eight RA patients (age = 55 ± 13 years, body mass index: 27.8 ± 5.4 kg/m², median of disease duration = five (two to eight) years) participated in this study. The current medication of all studied participants appears in Table 1. The data presented herein are part of a randomised controlled trial study with 100 participants (register number: ISRCTN04121489); in this study we present the baseline data from the participants who provided complete data for objective and subjective PA as well as VO2max data (68 out of 100). All patients fulfilled the revised American College of Rheumatology classification criteria for RA [23] and were recruited from Russells Hall Hospital (Dudley Group NHS Foundation Trust, UK). The study protocol and the main trial were approved by Birmingham East, North and Solihull Research Ethics Committee. All participants provided informed consent after verbal and written information was presented to them about all procedures involved in the project. Each participant visited the laboratory (Clinical Research Unit) on two separate occasions, seven days apart. During the first visit, participants were provided with a questionnaire pack, including the Health Assessment Questionnaire (HAQ) [24] to complete at a time convenient for themselves. During the same visit, an accelerometer (GT3X, ActiGraph, Pensacola, FL, USA) was provided to the participants to be worn over the subsequent seven consecutive days in order to assess objective PA. In addition, clinical disease activity and physical function were assessed as described below. During the second visit, participants returned the accelerometer and questionnaire pack and completed an assisted long form of the International Physical Activity Questionnaire (IPAQ) [15], in order to evaluate their subjective PA levels. This was then followed by the assessment of cardiorespiratory fitness, with a VO2max test.

Height was measured with a Seca Stadiometer 208 while body mass index and body composition were evaluated via bioelectrical impedance (Tanita BC418-MA, Tokyo, Japan).

Subjective PA was assessed using the long form of the IPAQ [15] which has previously been used in the case of RA patients [19]. The IPAQ measures the level of PA across four domains, that is, leisure time PA, domestic and gardening activities, work-related PA, and transport-related PA. In each domain, the duration (in minutes) and frequency (days) of PA including sitting, walking, moderate and vigorous PA are self-reported.

Daily objective PA and sedentary time (ST) were measured using the ActiGraph accelerometer (GT3X, Pensacola, FL, USA). Accelerometry is widely used for assessing individual levels of daily activity/time spent sedentary across different age groups ranging from children [25] to older people [26], as well as in the case of the general population from different countries [27,28]. The validity and reliability [29] of the GT3X has been examined against a previous accelerometer model [30] and across different testing situations including comparisons between mechanical and real life settings [19].

Movements recorded by the GT3X are converted into ‘counts’ within each epoch. These counts are calculated in relation to time spent at different activity intensities according to different cut-off points [27]. Daily averages were calculated for different activity levels including ST, light, moderate, vigorous and moderate-to-vigorous physical activity (MVPA). The cut-off points applied in the present study stemmed from Troiano and his colleagues’ research involving more than 6,000 participants [27] (Sedentary 0 to 99 counts per minute (CPM), Light 100 to 2,019 CPM, moderate 2,020 to 5,998 CPM and vigorous 5,999 CPM and above, MVPA 2,020 CPM and above). In the present study, the accelerometers were initialized in 60 second epochs replicating previous research protocols involving RA patients [17]. Data were screened and to be included in the analyses, the participant had to have valid data for a minimum of 10 hours per day [13] and for at least four days [31].

Participants in the current study were advised to wear the accelerometer on their waist at all times during a seven-day period, including their sleeping time if they did not find this uncomfortable. It was recommended to the participants that they should only take the device off during showering, bathing or engaging in any other water activities. If the participants engaged in any other water activity which was not possible to record on the accelerometer, they were requested to report the activity type, intensity and duration when they returned the device to the researchers. However, this was not the case for any of the current study participants.

The VO2max test was performed using a breath-by-breath system (Metalyzer 3B, Leipzig, Germany) on a HP Cosmos mechanical treadmill (Nussdorf-Traunstein, Germany). After identifying a convenient speed for each individual patient (which was normally 2 mph), the speed increased by 0.5 mph every minute until it reached 4 mph while inclination remain constant at 1%. Thereafter, there were only increases in inclination by 1% every 30 seconds until volitional exhaustion.

Normal distribution of the variables was assessed via the Kolmogorov-Smirnov normality test. To test our first hypothesis, Bland-Altman plots [32] were used to evaluate the agreement between the different methods of assessing PA (that is, subjective versus objective). Furthermore, given the non-normal distribution of both the subjective (IPAQ) and objective (GT3X) PA data, the Wilcoxon signed-rank test was used to test differences between these two methods for ST, light, moderate and MVPA levels.

To test our second hypothesis, correlations between VO2max with subjective and objective PA at different levels were examined via Spearman correlations. Furthermore, Steiger’s Z test [33] was used for analysing the differences between the correlations of subjective PA versus VO2max, and objective PA versus VO2max at ST, light, moderate, vigorous and MVPA levels. All statistical analyses were conducted using SPSS version 21 for windows package (SPSS, Chicago, IL, USA) with the level of significance set at P <0.05.