Monitoring of Sleep Indicators, Physical Activity, Pain, and Fatigue in Patients with Systemic Lupus Erythematosus and Relations among These Variables: A Pilot Study ()
1. Introduction
Poor sleep, fatigue, and pain are major health problems in patients with systemic lupus erythematosus (SLE) [1] [2] . The prevalence of sleep disorders ranges from 55% to 81% [3] , and fatigue is reported in 67% - 90% of patients with SLE [4] . To our knowledge, four studies on the quality of sleep in patients with SLE have been conducted within the past 10 years [3] [5] [6] [7] (Table A1). These studies focused on the predictors of the quality of sleep measured by the Pittsburgh Sleep Quality Index (PSQI), which focuses on the quality of sleep, such as sleep latency and sleep efficiency, within the previous month [3] . Factors associated with sleep quality were pain, SLE duration, comorbidity, medication, and side effects [3] [5] . In two small-scale studies, actigraphy was used to objectively measure the quality of sleep, such as time in bed and actual sleep [6] [7] .
Fatigue has become an important clinical outcome for patients with SLE [8] . Over 20 types of self-administered questionnaires have been used to measure fatigue [9] , but fatigue assessment remains difficult. Pain, medication, disease activity, and mental health are contributors to fatigue [1] [8] . Poor sleep quality, obesity, and reduced physical activity are also included in the risk factors for fatigue [4] .
Physical activity in patients with SLE is drawing attention in recent years as they are at high risk of cardiovascular disease [10] . Eight studies on physical activity or physical function in patients with SLE, conducted within the past 10 years, have been identified [11] - [18] (Table A2). Among them, five were cross-sectional studies on physical activity. Interestingly, these studies reported that patients with SLE tended to have longer sedentary hours than healthy controls or had lower activity levels than those recommended by the World Health Organization [16] [17] . Three small-scale intervention studies examined the effect of physical activity on quality of life and reported significant improvements in fatigue level, mental health, or physical functions [11] [12] [13] . The scales used to measure physical activity varied. ActiGraph GT3X was used in four studies to objectively measure physical activity. However, the ActiGraph GT3X’s monitoring period is short (1 - 7 days).
These monitoring studies used the mean values of indicators to examine the association between the main outcomes (sleep or physical activity) and factors of interest. A recent systematic review on sleep consistency (day-to-day variability) and health outcomes reported that higher sleep variability was associated with adverse health outcomes, such as mental and cardiometabolic health [19] . Day-to-day variability in sleep indicators in patients with SLE has not been documented, although they have a high prevalence of sleep problems. Understanding day-to-day fluctuations in these parameters provide useful information for symptom management in patients with SLE.
The aims of this pilot study were 1) to document the daily fluctuations in the device-monitored sleep indicators and physical activity in addition to the pain and fatigue levels in patients with SLE during 4-week and, and 2) to examine how the daily fluctuations in these variables affect one another.
2. Materials and Methods
2.1. Participants
Patients were not involved in the development of the research question, the design of the study, or the selection of outcome measures.
We recruited outpatients at two rheumatology centers (a university hospital and a prefectural hospital) in western Japan between September 2017 and May 2019. Primary physicians (rheumatologists) screened outpatients who met the following eligibility criteria: 1) were adults aged ≥ 20 years and were registered in the national SLE registry program, meeting the American College of Rheumatology classification criteria for SLE, and 2) had the ability to self-administer the questionnaires in Japanese. The exclusion criteria were as follows: 1) major comorbidity that would affect quality of life (QoL), such as terminal-stage cancer; and 2) overlap with other autoimmune diseases, such as rheumatoid arthritis or fibromyalgia. The first author contacted outpatients at the clinic to explain the research protocol and obtain written informed consent.
A sample size of 12 patients was recommended for a pilot study to investigate continuous variables [20] . As we planned to investigate multiple continuous variables and since the use of monitoring devices as well as daily reporting pain and fatigue scales were demanding for the participants, the sample size was set to 20 individuals.
2.2. Instruments
2.2.1. Quality of Sleep Monitoring Device
The body vibrometer device (Nemuri SCAN, PARAMOUNT BED Co., LTD, Tokyo, Japan) objectively measured various sleep indicators. This device (78 cm × 24.5 cm, 1.5 cm high) was placed under the mattress to measure the following 11 sleep indicators [21] [22] : total sleep time [min] (total time scored as “sleep” relative to the time inside the bed), time in bed [min], sleep latency [min] (elapsed time from when the participant got into bed to the beginning of the first interval containing 10 min scored as “sleep” with not more than 1 min of wakefulness), sleep efficiency [%] (total sleep time divided by total time in bed), wake after sleep onset [min] (total time scored as “wake” during the period scored as “sleep”), frequency of leaving the bed [times], respiratory disorder index [count/hour], periodic body movement index [count/hour] (Number of times the amplitude of respiratory movement decays per hour of sleep), activity score [count/min] (the intensity and frequency of large body movements, excluding smaller movements, such as breathing and heartbeat), respiratory rate [count/min], and heart rate [count/min] [23] .
2.2.2. Physical Activities
Physical activity was measured using a 3-axis accelerometer (Medi-Walk MT-KTODZ, TERUMO Co., LTD, Tokyo, Japan). The participants were asked to wear the device, except when bathing or sleeping. The following three parameters measuring physical activities were used: the number of steps, mean-to-vigorous physical activity (MVPA) [min], and total daily energy expenditure [kcal].
2.2.3. Patient-Reported Outcomes (PRO)
1) PSQI [24] : The PSQI is an 18-item self-report scale of sleep quality during the previous month. The PSQI evaluates seven components of sleep quality: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction due to poor sleep. The PSQI score ranges from 0 to 21 points.
2) The Lupus Patient-Reported Outcome—Japanese version (LupusPRO) [25] : The LupusPRO is a disease-specific QoL scale that consists of 43 items (eight health-related QoL subcategories with 30 items and four non-health-related QoL subcategories with 13 items) related to the QoL of the SLE patient’s daily life during the last 4-week.
3) Short Form Health Survey-12 (SF-12) [26] : The SF-12 is a globally used, non-disease-specific QoL scale. It comprises the following eight domains: physical functioning, physical role, bodily pain, general health, vitality, social functioning, emotional role, and mental health.
2.3. Data Collection
This was a pilot study with a cross-sectional design. The data collection was labor-intensive and required the participants’ full cooperation in setting up a sleep monitoring device, wearing an accelerometer, and recording fatigue and pain levels for 4 weeks. To the best of our knowledge, this type of prospective data collection in patients with SLE has not been previously reported, and a small convenience sample is appropriate for the study objectives. The participants were asked to fill out a set of QoL questionnaires and mail them back to the first author at the beginning and end of the 4-week study period. Demographic data were obtained from the questionnaire, while clinical data were extracted from medical records. Rheumatologists evaluated the SLE Disease Activity Index 2000 (SLEDAI-2K) at the beginning and end of the 4-week study period. A flare was defined as an increase of ≥3.0 points in the total SLEDAI-2K score between the previous visit and survey data [27] .
Two types of devices were given to the participants to monitor their quality of sleep and physical activity. The first author explained to the patients how to use the devices and made follow-up phone calls to enquire about any problems related to them.
2.4. Data Analysis
Descriptive statistics were obtained for all variables. Imputation was not used for missing values. To examine day-to-day fluctuations in the monitored variables, the coefficient of variation (CV) was tabulated to determine changes in physical activity and sleep indicator scores. The CV was also obtained for self-reported daily pain and fatigue. The Shapiro-Wilk test was used to test the normal distribution of all variables, as the CV uses the mean as the denominator. Spearman’s rank correlation coefficients between all daily monitoring variables were calculated to explore the association between day-to-day fluctuations of some of these variables.
We grouped the participants into good and poor sleepers according to the PSQI values, using a cutoff point of ≥5.5 to categorize poor sleepers [24] . Statistical significance was set at p < 0.05. Correlation coefficients between the total PSQI score and subscales of the two QoL scales were obtained to examine sleep-related factors. JMP 15 software (SAS, Inc., Cary, NC, USA) was used for all statistical analyses.
3. Results
3.1. Participant Demographics
Outpatients with SLE from two rheumatology departments were screened for eligibility. A sample size of 21 outpatients was judged eligible by the primary physician; however, one patient declined to participate as she considered that it was difficult to use monitoring devices for 4 weeks. Thus, 20 eligible outpatients provided informed consent, completed self-administered questionnaires, and underwent sleep monitoring. Seventeen patients monitored physical activity for 4 weeks. Three participants failed to return the 3-axis accelerometer for measuring physical activities when returning the questionnaires and the body vibrometer device for measuring sleep indicators. The accelerometer automatically recorded null values daily, even when it was not used. Thus, physical activity data were replaced with null values by the time the researcher received the device.
Table 1 displays the demographic and clinical characteristics of the participants. Specifically, 80%, 60%, and 20% of the participants had at least a junior college education, part-time employment, and were unemployed, respectively.
![]()
Table 1. Demographic and clinical characteristics of the participants (n = 20).
SD: standard deviation; PSQI: Pittsburgh Sleep Quality Index; SLEDAI-2K: Systemic lupus erythematosus Disease Activity Index 2000.
Moreover, 60% and 50% of the participants had poor sleep quality, as measured by the total PSQI score, and had SLEDAI-2K < 4.0 points, indicating remission or a mild disease form, respectively. Approximately 75% of the participants used corticosteroids, while 40% used immunosuppressive drugs.
QoL measured using the SF-12 showed low mean scores on the general health and vitality subscales. Among the LupusPRO subscales, the social support and coping subscale scores were low (≤25).
3.2. Descriptive Statistics of 4-Week Monitoring of the Quality of Sleep Indicators, Physical Activity, Fatigue, and Pain
The descriptive statistics of the daily monitoring variables are presented in Table 2. The mean number of steps taken (range: 2287.0 - 8018.0) and MVPA duration (range: 0.0 - 19.0 minutes) showed substantial variability among participants. The majority of the participants had a mean sleep duration of <7 h. Sleep
![]()
Table 2. Descriptive statistics of daily monitoring of physical activity, quality of sleep, pain, and fatigue in this study.
SD: standard deviation; Min: minimum; Max: maximum; IQR: interquartile range; min: minute; CV: coefficient of variation; MVPA; moderate-to-vigorous physical activity; a: pain was recorded from 0 (none) to 100 (strong); b: fatigue was recorded from 1 (none) to 4 (strong).
efficiency for most participants exceeded 85%, a cutoff point for the lower normal range [28] , and sleep latency for most of the participants was within the normal range of 10 - 20 min [29] . The majority of the participants did not leave the bed most of the nights (0.8 times/night). The median pain score was also low (10 out of 100 points). Regarding CV, the indicator of daily fluctuation within the individual, the greatest CV among the 11 sleep indicators was the number of times leaving the bed (1.48), followed by “waking after sleep” (0.72), and sleep latency (0.59). Most other sleep variables had a CV of <0.3. The median CV of MVPA was twice as large as that of the number of steps (0.8 and 0.40, respectively). The median CV for the pain score was much higher than that for fatigue (0.74 and 0.28, respectively). The interquartile range of CV, an indicator of variability among individuals, was the highest for “frequency of leaving bed” (2.39), followed by pain (1.43) and MVPA (0.54). In short, day-to-day variability was high for “frequency of leaving bed,” pain, and MVPA within an individual as well as among individuals.
3.3. Association between Sleep Indicators, Fatigue, Pain, and Physical Activity
Table 3 shows statistically significant correlation coefficients between variables monitored daily in each participant as well as the mean and SD of these monitored variables.
The fatigue levels were moderately correlated with poor sleep quality indicators in all five participants (range: r = −0.40 to −0.52). Higher fatigue levels showed a moderate correlation with slower heart and/or respiratory rates in three participants (range: r = −0.40 to −0.50) and were correlated with elevated respiratory distress index scores in two participants (r = 0.40 and 0.61, respectively).
A longer duration of MVPA was significantly correlated with poor sleep outcomes, such as shorter sleep time (r = −0.50) (n = 1), longer sleep latency (r = 0.40 and 0.61) (n = 2), higher respiratory distress scores (range: r = 0.40 - 0.46) (n = 3), and higher heart rate (r = 0.60 and 0.62) (n = 2). In contrast, a higher number of steps was significantly correlated with better sleep outcomes, such as a lower heart rate (r = −0.41) (n = 1), respiratory distress index score, and activity score (r = −0.64, and −0.47) (n = 1). However, a higher number of steps was associated with worse pain (range: r = 0.52 - 0.63) and fatigue scores (range: r = 0.41 - 0.54) (n = 3).
Higher pain scores showed moderate correlations with higher physical activity levels in three participants and positive correlations with respiratory and/or cardiovascular system indicators in two. In contrast, negative correlations between pain scores and respiratory indicators or activity scores were observed in three participants. These statistically significant associations were even observed in participants with lower mean pain scores (<20 points). In short, 28-day monitoring revealed individual differences in the impact of daily fluctuations in physical activity level and/or fatigue on the quality of sleep.
![]()
Table 3. Correlation coefficients between factors related to daily fluctuations in fatigue in each participant and descriptive statistics (n = 20).
SLEDAI-2K, systemic lupus erythematosus disease activity index 2000; MVPA, moderate-to-vigorous; PSQI, Pittsburgh Sleep Quality Index; NA, not applicable; SL, sleep latency; RR, respiratory rate; HR, heart rate; letters in blue, sleep indicators; letters in red, pain and fatigue.
Figure 1 illustrates the daily fluctuations in the number of steps/100, pain score, and sleep time in minutes in identification number 5 (ID 5). The number of steps was converted to per 100 steps to use the same second X-axis as the pain score. The increased number of steps was job-related, and the peaks in the pain score seemed to correspond to the number of steps. To a lesser extent, sleep time tended to be lower when the number of steps peaked.
Higher fatigue levels were moderately correlated with higher activity levels (MVPA or the number of steps/day) in four participants (range: r = 0.41 - 0.54) and were associated with higher pain levels in three participants (range: r = 0.42 - 0.60). Negative correlations between the fatigue level and the respiratory/cardiovascular systems were observed in three participants (range: r = −0.40 - −0.50).
As a summary of the correlations between variables, a conceptual framework of the associations among the four daily measured factors is presented in Figure 2. Fatigue levels in the evening and physical activity levels during the day were significant predictors of sleep quality. Predictors of fatigue were pain and physical activity levels in four participants, and these three factors were correlated with each other in three participants. To a lesser extent, pain level was a predictor of the quality of sleep.
3.4. Association between Sleep Indicators, Fatigue, Pain, and Physical Activity
The total PSQI score at the end of the monitoring period was significantly and negatively correlated with four SF-12 subscales (role physical, role emotional, vitality, and mental health) (range: r = −0.57 to −0.67). It was also associated with three LupusPRO subscales (cognition, physical health, pain, and vitality) (range: r = −0.51 to −0.57) (Table 4). Regarding the daily monitored variables,
![]()
Figure 1. Daily fluctuations in the number of steps/100, pain score, and sleep time in minutes for ID 5.
![]()
Figure 2. Conceptual framework of association among the four factors measured daily. No. of steps, number of steps, MVPA, moderate-to-vigorous physical activity.
![]()
Table 4. Descriptive statistics of PSQI and QoL (SF-12 and LupusPRO) and correlation between PSQI and QoL (n = 20).
PSQI: Pittsburgh Sleep Quality Index; QoL: quality of life; SD: standard deviation; SF-12: short-form health survey-12; LupusPRO, lupus patient-reported outcome.
the total PSQI score was positively correlated with the mean and median fatigue levels (r = 0.62, p = 0.004; r = 0.65, p = 0.002, respectively) in the 4-week period and was positively correlated with the CV of the mean number of steps (r = 0.53, p = 0.036). In short, the total PSQI score was negatively moderately correlated with the QoL and positively moderately correlated with fatigue and the mean number of steps.
4. Discussion
The current study examined day-to-day fluctuations in sleep indicators, physical activity level, pain, and fatigue and explored the association between these monitored variables among patients with SLE. The changes measured by CV were large for the frequency of leaving the bed, duration of MVPA, pain score, and waking after sleep onset in minutes. In nearly half of the participants, fatigue and MVPA duration were major factors associated with sleep duration and cardiopulmonary indicators at night. Physical activity, fatigue, and pain were correlated in less than a quarter of the participants.
The correlations between the fatigue levels measured in the evening and poor sleep outcomes are at variance with previous studies on fatigue, which suggest that poor quality of sleep itself is one of the contributory factors to fatigue [1] [4] . However, previous studies using instruments, such as the PSQI and the daily fluctuations in sleep indicators, were not examined. Our cross-sectional data using the PSQI are in agreement with review findings [1] [4] , showing a negative association between the total PSQI score and the vitality subscale scores measured using SF-12 and LupusPRO.
The number of participants with significant correlations between pain, fatigue, and/or sleep quality is limited. Pain is regarded as a significant contributor to fatigue in patients with SLE as well as inflammation, corticosteroid use, and mental health issues [1] [4] [8] . Most participants had low mean pain scores, consistent with the fact that pain can be controlled more easily than fatigue. Detailed information on the use of pain medications was not available.
Corticosteroids are one of the contributors to fatigue [1] [4] , and most participants were on corticosteroids. Nevertheless, the medication dosage does not change daily and is unlikely to affect the day-to-day fatigue levels. All participants were in a non-flare state, and inflammation itself was unlikely to affect the day-to-day fatigue level.
The moderate correlations between physical activity level and the other monitored variables in our study suggest that increased physical activity was detrimental to the physical and mental health of some participants. None of our participants attained the goal recommended by the WHO [30] , and the MVPA duration in our study was one-third of that reported in North American studies [14] [16] . An exercise program is reported to be beneficial in reducing fatigue and improving mental health and cardiorespiratory capacity [31] , although the evidence level is low. Caution is necessary to apply study findings to the clinical setting because patients participating in randomized control trials may not represent those in real-world settings. In conclusion, our findings suggest the need for daily monitoring of physical activity, fatigue, and pain before exercise is recommended.
The number of steps taken seems to have an impact on the pain, fatigue, and quality of sleep indicators in some participants. This appears to affect sleep indicators that differ from MVPA duration. Our participants’ short duration of MVPA implies that most of the physical activity was of light intensity. Light-intensity physical activity is associated with reduced cardiometabolic risk factors and mortality [32] [33] . However, evidence for the appropriate level of light-intensity activity has not been established. Caution is necessary when increasing the number of steps because some participants showed positive correlations between the number of steps and increased pain, fatigue, and worse sleep indicators.
A moderate correlation between fatigue and physical activity level was observed, even in participants with low physical activity levels. This phenomenon may represent the concept of susceptibility to fatigue proposed by Cleanthous et al. [34] . Patients with SLE report a sudden onset of exhaustion after performing the minimal activity, such as taking a shower, which is defined as fatigue “disproportionate to the degree of activity or effort expended.” Further studies are needed to understand the patterns of exhaustion onset for symptom management.
Some of the sleep indicators used in our study are not well documented. For example, only small-scale studies have reported respiratory distress index and activity scores [35] [36] . The mean activity score in our study is similar to a variable used for adults with sleep disorders [36] and was much higher than that in young adults [35] . As these two sleep indicators show significant correlations with pain, fatigue, and physical activity levels, future studies should explore the potential causal associations between them.
The association between physical activity and cardiopulmonary indicators was an unexpected finding. Heart and respiratory rates were measured only when participants were lying in bed. Future studies are needed to monitor 24-h changes in vital signs to examine the effects of MVPA on the cardiovascular and respiratory systems using smartwatch-type monitoring devices.
Various sleep indicators in our study showed correlations with physical activity and/fatigue level. Genetic and environmental factors probably account for individual variability. Genetic research has revealed the link between the clock genes and multiple health outcomes such as sleep disorders, hypertension, and metabolic disorders [37] [38] . Further, the clock genes modulate melatonin, which in turn affects the inflammatory pathways in patients with rheumatoid arthritis [38] . Studies are underway to examine the impact of time of medicine that affects disease outcomes, such as cardiovascular events or joint diseases. [38] [39] . For patients with RA, bedtime administration of corticosteroid is reported to reduce joint pain by reducing inflammation. For patients with SLE, self-monitoring of sleep indicators, physical activity, and fatigue may help identify factors that affect the quality of sleep and fatigue for symptom management.
5. Conclusions
We examined the day-to-day fluctuations in physical activity, pain, fatigue, and sleep indicators. Daily changes measured by CV were high for the frequency of leaving the bed, followed by MVPA duration, pain score, and duration of awakening after waking at night. Fatigue was a significant predictor of poor sleep. Among the two physical activity indicators, MVPA duration was associated with poor sleep outcomes. A longer MVPA duration was significantly correlated with shorter sleep time and increased cardiopulmonary output at night.
In contrast, a higher number of steps was associated with better cardiopulmonary outcomes at night in two participants. An association between fatigue, pain, and physical activity level was observed in several participants. The current study illustrates the importance of daily monitoring of physical activity, sleep quality, pain, and fatigue levels to examine the relationship between these variables for symptom management after a recommended monitoring duration of 4 weeks.
6. Limitations of This Study
The major limitation of this study was the small sample size, and type II errors were expected. However, the obtained correlations among well-known QoL scale data are consistent with those of previous studies, showing that the total PSQI score was correlated with the SF-12 and the LupusPRO subscales related to mental and physical health [1] [8] . The fatigue scale was based on a 4-point Likert scale, which may limit the classification of fatigue variability. Nevertheless, higher mean and median fatigue levels were associated with higher total PSQI scores and were considered adequate for this pilot study.
Acknowledgements
We would like to thank all the patients who participated in this study and the staff at the two clinics that cooperated with the data collection. We also acknowledge Drs. Atsushi Kumanogoh and Yasushi Tanaka for supporting this project.
Funding
This study has been funded by JSPS KAKENHI Grant Numbers JP 16K15899 and 19K10944. This work was supported by JSPS KAKENHI Grant Number JP 1615899 and 19K10944.
Ethics Approval
This study was approved by the Ethics Committee of Hyogo Medical University and the two rheumatology centers (IRB no. 17012). At the initial contact with the participants, the first author provided a verbal and written explanation to confirm their willingness to participate. After obtaining written informed consent, the first author distributed a set of questionnaires and measurement devices.
Appendix
![]()
Table A1. Studies of quality of sleep in patients with SLE.
SLEDAI: Systemic Lupus Erythematosus Disease Index; SD: Standard Deviation; NA: Not Applicable; SLE: Systemic Lupus Erythematous; FM: Fibromyalgia.
![]()
Table A2. Studies of physical activity (PA) in patients with SLE.
SD: Standard Deviation; NA: Not Applicable; SLE: Systemic Lupus Erythematous; SLEDAI: Systemic Lupus Erythematosus Disease Index; RCT: Randomized Controlled Trial; RA: Rheumatoid Arthritis; MVPA: Moderate-to-Vigorous PA; WHO: World Health Organization. #SLEDAI-2K: Systemic Lupus Erythematosus Disease Index 2000.