Association of physical fitness components and health-related quality of life in women with systemic lupus erythematosus with mild disease activity

Objectives To study the association of different components of physical fitness [flexibility, muscle strength and cardiorespiratory fitness (CRF)] and a clustered fitness score with health-related quality of life (HRQoL) in women with systemic lupus erythematosus (SLE) and to analyze whether participants with high fitness level have better HRQoL. Methods This cross-sectional study included 70 women with SLE (aged 42.5; SD 13.9 years). The back-scratch test assessed flexibility, the 30-sec chair stand and handgrip strength tests assessed muscle strength, and the 6-min walk test (n = 49) assessed CRF. HRQoL was assessed through the 36-item Short-Form Health Survey (SF-36). Results Flexibility was positively associated with the physical function dimension and the physical component summary (PCS) (rpartial between 0.26 and 0.31; p<0.05), and negatively related with social functioning dimension (rpartial = -0.26; p<0.05). Muscle strength was positively associated with the physical function, physical role, bodily pain dimensions and the PCS (rpartial between 0.27 and 0.49; all p<0.05). CRF was positively associated with the physical function and bodily pain dimensions, and PCS (rpartial between 0.39 and 0.65; all p<0.05). The clustered fitness score was associated with the physical function (B = 17.16) and bodily pain (B = 14.35) dimensions, and the PCS (B = 6.02), all p<0.005. Patients with high fitness level had greater scores in the physical function, physical role, and bodily pain dimensions and the PCS, all p≤0.05. Conclusions Our study suggests that muscle strength and CRF are positively associated with HRQoL, while flexibility showed contradictory results. These findings highlight the importance of maintaining adequate fitness levels in women with SLE.

Introduction Systemic lupus erythematosus (SLE) is an autoimmune disease that may involve any organ or system and produce a wide spectrum of clinical manifestations [1]. The goals of SLE therapy are reducing disease activity, preventing organ damage and improving health-related qualityof-life (HRQoL) [2]. Although the spectrum of treatment options and the general knowledge of SLE have improved in recent decades (i.e. leading to similar survivals rates than the general population) HRQoL continues to be compromised in this population [3]. Evidence shows that patients with SLE present a clearly deteriorated HRQoL not only compared to healthy individuals [4][5][6][7] but also to other chronic diseases [8].
The determinants of HRQoL in SLE are complex but may include factors related to disease and treatment, health care provision [1], or psychological variables [9]. In addition to medical therapy, non-pharmacological therapies, such as exercise, seems to be an efficient and safe alternative to improve physical fitness [10], symptomatology and HRQoL of patients with SLE [11,12]. Aerobic training [13][14][15][16][17][18], as well as strength and flexibility training [13,18,19] have shown a variable efficacy in improving diverse dimensions of HRQoL in SLE. Noteworthy, exercise interventions are typically poorly described and provide insufficient information to warrant replication and a clear understanding on how the different components of physical fitness must be trained along an intervention period for SLE. Therefore, the optimal exercise intervention for SLE is still unclear [12].
Physical fitness is a strong health marker in the general population [20] and in other rheumatologic diseases including rheumatoid arthritis [21] and fibromyalgia [22][23][24], among others. It has been reported that patients with SLE present a reduced cardiorespiratory fitness (CRF) [4][5][6][7]25], muscular strength and functional capacity [26]. Functional aerobic impairment [6,25] and low levels of strength [26] in SLE are negatively correlated with perception of severity [6], fatigue [6,26], and age-related arterial stiffness [25]. Furthermore, a fitness level below the criterion-referenced standards could lead to a premature loss of independence [27]. In regards to HRQoL, this outcome has been also positively related to higher CRF [7,28] and muscle strength in SLE [26,29,30]. However, the scope of previous research is generally limited to physical dimensions ignoring the also relevant social and mental spheres of HRQoL [31,32]. As patient-reported outcomes are considered key variables in the assessment and management of patients with SLE [33], and fitness might be enhanced through exercise interventions [10], a more comprehensive characterization of the association of the different dimensions of HRQoL with physical fitness might inform the design of randomized trials and the development of therapeutic exercise protocols for this population.
Therefore, the aims of this study were i) to examine the association of different components of physical fitness [flexibility, muscle strength and CRF] and a clustered fitness score with HRQoL in women with SLE and ii) to analyze whether participants with high fitness level have better HRQoL. We hypothesize that i) all physical fitness components, as well as the clustered fitness score, will be positively related to different physical, social and mental dimensions of HRQoL and that ii) participants with high fitness level have better HRQoL.

Design and participants
For this cross-sectional study, a total of 172 patients with SLE were invited to participate in the study through the Systemic Autoimmune Diseases Unit of the "Virgen de las Nieves" University Hospital and the "San Cecilio" University Hospital (Granada, Spain). After providing detailed information about the aims and study procedures, participants provided written informed consent. The inclusion criteria were: i) women aged 18-70 years), ii) with �4 SLE classification criteria according to the American College of Rheumatology [34], and iii) with a minimum medical follow-up of 1 year at the units and presenting clinical stability, defined as no changes in the systemic lupus erythematosus disease activity index (SLEDAI) and/or the treatment during 6 months before the study. Exclusion criteria were as follows: not being able to read, understand and/or sign the informed consent; presence of concurrent fibromyalgia; history of clinical cardiovascular disease in the last year, receiving a biological treatment or required doses of prednisone (or equivalent) greater than 10 mg/day in the last 6 months. The Research Ethics Committee of Granada reviewed and approved the study protocol.
The patients were evaluated on a single day at the "Virgen de las Nieves" University Hospital, where sociodemographic data and clinical history were obtained, anthropometric measures and body composition were assessed, and the fitness tests were carried out. Participants were instructed in how to fill the questionnaires related to HRQoL, depression, and fatigue at home and returned them throughout the same week. The questionnaires were collected and checked by the research team for potential non-response.

Measurements
Health-related quality of life (HRQoL). The HRQoL was assessed with the Spanish version of the Short-form 36 [36]. Upper-body flexibility was assessed with the back scratch test, where the cm between (positive distance) or overlap (negative distance) of the middle fingers behind the back was recorded twice for each arm. The best scores of dominant, and non-dominant side were used and the total mean was calculated. CRF was assessed with the 6-min walk test that measures the distance in meters that a participant is able to walk along a 45.7 m course within 6 minutes. Lower-body muscle strength was assessed with the 30-sec chair stand test, in which we registered the times an individual is able to rise, from a sitting position to a full stand position with arms folded across chest. Upper-body muscle strength was assessed with the handgrip strength test using a digital dynamometer (TKK 5101 Grip-D; Takey, Tokyo, Japan), as described elsewhere [36]. Participants performed the test twice (alternately with both 2 hands), with 1 min of rest between trials. The best score from both hands (dominant and non-dominant) were averaged and the total mean was calculated.
Anthropometry and body composition. Height was measured in cm using a stadiometer (SECA 222, Hamburg, Germany), and weight in kg with a bioimpedance device (InBody R20, Biospace, Seoul, Korea). The body mass index (BMI, kg/m 2 ) was calculated.
Depression. The Beck Depression Inventory-second edition (BDI-II) [37] is a 21-item self-report measure designed to assess depressive symptomatology. Each depressive symptom is rated from 0 (not present) to 3 (severe) according to how patients felt during the past 2 weeks. The BDI-II provides a single overall score (0-63) where higher score represents higher depressive symptomatology.
Fatigue. Fatigue severity was measured with the Spanish version of the Multidimensional Fatigue Inventory (MFI-S) that contains 5 scales. Each subscale consists of four items ranging from 4 to 20, with higher scores indicating greater fatigue. For the present study, only general fatigue dimension was included.
Sociodemographic variables and clinical history. All participants filled out a sociodemographic and clinical data questionnaire to collect information including age, educational level, occupational status, and SLE data (diagnostic criteria, year of diagnosis, time of evolution and treatments). The disease activity was measured by the SLEDAI [38].

Statistical analysis
Descriptive statistics were performed to examine sociodemographic and clinical characteristics of the sample. The continuous variables were presented as mean and standard deviation, and categorical variables with frequencies and percentages. In addition, we calculated the percentage of patients achieving the fitness standards proposed by Rikli and Jones [27] for the Senior Fitness Test Battery to maintain physical independence in later years. The values recommended for the lowest range of age (women between 60 and 64 years) of the fitness tests included in this study (30-sec chair stand and 6-min walk) were used. An independent t-test and χ2 test were used to compare sociodemographic and clinical features between participants with low and high fitness level.
To study the association between fitness and HRQoL, different approaches were conducted. First, partial correlations were used to assess the association of the different components of physical fitness with all dimensions of the SF-36 questionnaire while accounting for the following covariates: age, BMI, BDI-II total score and MFI general fatigue score. Because disease activity (SLEDAI) [39,40], systemic damage index (SDI) and disease duration (according to preliminary analyses) could be potentially related with HRQoL, additional adjustments for these variables were included in all the analyses. Next, a clustered fitness score was calculated as the weighted average (taking into account the number of tests assessing each component) of flexibility, muscle strength and CRF z-scores [(value-mean)/SD]. The association between the clustered fitness score (as a global measure of fitness including all components) and HRQoL was assessed with linear regressions analyses. All dimensions of the SF-36 questionnaire were entered in separate regression models as dependent variables and the clustered fitness score along with the covariates were introduced as independent variables using the 'enter' method.
Finally, for those dimensions of HRQoL significantly associated with fitness components in prior correlations analyses, further analysis of covariance (ANCOVA) compared the differences in SF-36 scores between patients with low and high fitness level. The median value of the sample for each fitness test [flexibility (1.35 cm), upper-body muscle strength (24.28 kg/cm 2 ), lower muscle strength (15 repetitions) and CRF (575 meters)] was used to consider high (�median value) or low (<median value) fitness.
The analyses were performed using Software SPSS version 21.0 and the statistical significance was set at p�0.05.

Results
The flow diagram of the participants included in this study is presented in Fig 1. A total of 77 patients met the inclusion criteria, agreed to participate and were assessed in 2 waves (49 women were assessed in October 2016 and 28 were assessed in February 2017). The evaluations were identical but the 6-min walk test was not included in the second assessment. Thus, after excluding participants with incomplete data, a total of 70 women were finally included in the study (except for those analyses including CRF, where only 49 women were included). The sociodemographic and clinical characteristics of the study participants are shown in Table 1. The sociodemographic and clinical variables according to fitness level (low and high) in all fitness tests are presented in Table 2.
The partial correlations of physical fitness variables and HRQoL dimensions are presented in Table 3. Overall, the strength of correlations ranged from weak (r partial = 0.2-0.39) to moderate (r partial = 0.4-0.6). Mean upper-body flexibility (back scratch test) was correlated with physical functioning, social function, and PCS (r partial between -0.26 and 0.28; all, p<0.05) and only non-dominant arm flexibility was significantly associated with those same dimensions (r partial between -0.26 and 0.31; all, p<0.05). Mean upper-body strength (handgrip test) was positively correlated with physical function, physical role, bodily pain and PCS (r partial between 0.27 and 0.49; all, p<0.05). The association of non-dominant upper body strength with HRQoL was greater compared to the dominant side for physical function (r partial = 0.48 vs. r partial = 0.47), physical role (r partial = 0.31 vs no significance) and PCS (r partial = 0.47 vs. r partial = 0.42), except for bodily pain (r partial = 0.26 vs. r partial = 0.34). Lower-body strength (30-sec chair stand) was positively associated with physical function, physical role, bodily pain, and physical component (r partial between 0.32 and 0.49; all, p<0.01). CRF (6-min walk test) was positively associated with physical function, bodily pain and physical component (r partial between 0.39 and 0.65; all, p<0.01). Additional adjustments for SLEDAI, SDI and disease duration were performed, obtaining similar results (except for the non-significant association between handgrip mean strength and physical role (p = 0.075) and bodily pain (p = 0.066), and between nondominant handgrip strength and bodily pain (p = 0.181) when considering this confounding variables).
The regression models assessing the association of the clustered fitness score with all HRQoL dimensions are shown in Table 4. The clustered fitness score was independently associated with physical function (B = 17.16), bodily pain (B = 14.35) and the PCS (B = 6.02; all p<0.005), and there was a non-significant trend with the physical role dimension (B = 8.67; p = 0.075). Further adjustment for SLEDAI, SDI and disease duration did not alter the significance of the results.

Discussion
The main findings of the present study suggest that muscle strength, and CRF are positively associated with different dimensions of HRQoL while flexibility showed contradictory results. The clustered fitness score was consistently associated with better scores in the physical  The reduced fitness level of the patients in the present study is in agreement with prior research [4][5][6][7]. We additionally observed that a 60% or less of our participants (aged, on average, 42.1 years), reached the level of lower-body strength and CRF proposed for women aged 60-64 to maintain physical independence [27]. These findings suggest that a considerable number of relatively young patients with SLE are already disabled or at risk of losing the physical capacity needed for daily living during their later life [27], stressing the importance of promoting exercise interventions that increase fitness level. Consistently, in further comparisons Physical fitness and health-related quality of life in systemic lupus erythematosus between groups of high and low fitness (according to the median value of the group) we detected mean significant differences in HRQoL of at least 12 points in single dimensions and 6.5 points in the PCS. Different thresholds have been proposed for minimal clinically important changes in the SF-36 in SLE, ranging from 5 to 10 points difference in single dimensions [41] and a minimum difference of 2.1 points for the PCS [42]. Considering the reduced fitness level used as cut-point for the comparisons, it could be argued that even moving from an extremely low to a relative low fitness level could be potentially related to relevant changes in HRQoL, yet this hypothesis needs to be confirmed in future intervention studies.
Cardiorespiratory fitness seems to be the fitness component that presented the strongest association with HRQoL, more specifically with physical function, bodily pain and PCS, beyond the non-significant association with the physical role dimension. Two previous studies examined the relationship of CRF with HRQoL in SLE [7,28]. Tench et al. [7] showed that a better peak aerobic capacity was associated with better physical function dimension. Additionally, those patients able to perform the treadmill test at a greater speed reported better physical function, physical role, and bodily pain [7]. Despite the use of different tools to assess CRF, our results generally concur with these previous findings. In other study, CRF assessed with the 6-min walk test was positively associated with physical function, social function, emotional role, and mental health in patients with SLE. As opposed to Balsamo et al.
[28] and our initial hypothesis, we observed no association of CRF with mental dimensions of HRQoL. It must be noted that, contrary to the study of that Balsamo et al.
[28], our analyses were adjusted for different psychosocial and clinical variables. Our results, however, are in agreement with studies in other conditions that observed associations of CRF with physical health rather than mental health [43]. Taking together our findings and those from previous research in SLE, it seems that there is a link between CRF and HRQoL, which is more consistent across the physical dimensions and that needs to be further explored for mental components.
Another major finding is that muscle strength was the component that presented correlations and significant differences in more dimensions of HRQoL. These associations were also roughly equal between dominant, non-dominant sides and mean values. However, for physical role the association with dominant grip strength was non-significant (p = 0.06). We also observed a slightly greater association of the mean strength which is in line with recent evidence stressing the relevance of including a relative measure of strength that considers both hands as a better indicator of health in comparison to dominant hand [44]. In patients with SLE, a positive association of dynamic maximum muscle strength (upper and lower limbs) with physical function and emotional role has been described [26]. Furthermore, Andrews et al. [29] reported that lower-body muscle strength (assessed by peak isokinetic knee torque and chair-stand time) was positively associated with physical functioning. This association was additionally confirmed in longitudinal research were a reduced lower extremity muscle strength predicted declines in physical function [45]. Our results not only corroborate prior research but also extend the findings to non-previously explored dimensions of HRQoL. Upper and lower extremity muscle strength were additionally associated with physical role, bodily pain and the PCS in this study which is in concordance with previous studies in other rheumatologic populations [22].
Literature regarding flexibility in SLE is merely descriptive [5,26] and no implications for health of this component in isolation has been stablished either in observational or intervention studies. In the present study, we found contradictory results: while there was a positive association of flexibility with physical function and physical component, there was a negative association with social function. This finding contrasts with the general positive association between all fitness components and HRQoL [46], and also with the benefits for mental health of flexibility interventions [47]. It is noteworthy, however, that flexibility showed the weakest correlations among all fitness components and that we did not find any differences in social dimension between groups of high and low flexibility. Also, the non-dominant limb presented stronger correlations compared to the dominant side. It is difficult to elucidate the nature of these findings as there is a lack of previous studies assessing how flexibility of each side of the body is related to health. For the descriptive values in our participants we could observe a 4.5 mean difference between dominant and non-dominant flexibility which is in line with prior evidence showing a reduced mobility in upper-body non-dominant side, probably related to usage [48]. It could be hypothesized that the relationship between flexibility and HRQoL may not be uniform across the whole spectrum of values but only relevant until the minimum level of mobility required for daily living activities from which greater flexibility would not represent greater benefits in HRQoL. Nevertheless, these exploratory findings need to be assessed in detail in future studies. Despite of these results, flexibility has recently gained attention due to its potential role in cardiometabolic health [49] and pain reduction in SLE [19] and its inclusion in exercise programs is currently encouraged [50]. Future interventions studies are needed to examine these contradictory findings regarding flexibility and to determine the adequate regimes of intensity, frequency, and progression for possible health benefits in patients with SLE.
Different hypothesis may explain the associations found between fitness and the HRQoL dimensions (mainly physical function, bodily pain and the PCS). Increasing fitness could improve the ability to perform daily living tasks assessed in physical function dimension [31] that typically requires a combination of CRF, strength and flexibility. However, as proposed by Balsamo et al. [28], it is also plausible that the relationship worked bidirectionally, with better HRQoL predicting better physical fitness. Muscle strength and CRF were the main contributors for the associations found between fitness and bodily pain, which is a frequently reported symptom in SLE [51]. Among the various causes of pain in SLE [51], exercise could act on reducing inflammation [52] or through an improvement of muscle oxygenation [53] that reduces peripheral and central sensitization [54,55], as suggested for other rheumatic conditions. Nonetheless, the scarce evidence in SLE regarding the changes on pain after exercise is inconclusive [15][16][17][18][19], and not necessarily related to changes in fitness [18,19]. Importantly, our findings also showed that the relationships found between fitness and HRQoL are independent of psychosocial and clinical aspects such depression, fatigue, or IMC, limiting the explanatory capacity of these frequent conditions concomitant to SLE. We must note that the correlations found between fitness and HRQoL ranged from weak to moderate. However, considering the multiple factors affecting HRQoL, the possible explanatory capacity of single variables is low and, therefore, these weak-moderate correlations were somehow expected and still of relevance. Also, we found a lack of association between physical fitness and mental domains. We hypothesized that physical fitness may be a greater determinant for the limitations assessed in the physical component while other factors such age [56], disease activity or duration [57,58] may be closer to mental health in SLE. Our findings in a relatively young with clinical stability sample would support this idea as we observed better scores for mental domains compared to physical domains when the values of the SF-36 questionnaire were compared to reference values of the Spanish population [59].
This study has limitations. The cross-sectional design does not allow establishing causal relationships. The sample size was relatively small and included patients with low disease activity. Consequently, our results may not be generalizable to the entire SLE population. In addition, the comparisons between "fit" and "unfit" individuals were performed according to an arbitrary cut point (median value). Because the Senior Fitness Test for older adults was used, our results in an adult population are not completely comparable with the reference standards. In future research, it would be of interest to assess other fitness components (lower body flexibility or agility) and the role of other clinical variables such us disease activity, concomitant therapy, or the numbers of the used drugs. Our findings might have implications for the development of exercise-based intervention studies, especially in those aimed to enhance specific components of physical fitness in women with SLE.

Conclusion
In summary, higher muscle strength, CRF, and global physical fitness were positively associated with different dimensions of HRQoL. The role of flexibility was contradictory and deserves further research. Patients with high fitness presented better scores in HRQoL, especially in those dimensions associated with physical health. Overall, there was no association of physical fitness with any of the mental-related dimensions of HRQoL. Future research is needed to understand the prognostic value of fitness and the extent to which increasing fitness through exercise might be linked to improvements in HRQoL.