Effect of Cardiovascular Training on Fitness and Perceived Disease Activity in People With Ankylosing Spondylitis

Authors


  • ClinicalTrials.gov identifier: NCT00913302.

Abstract

Objective

Several studies suggest that patients with ankylosing spondylitis (AS) have an increased risk of cardiovascular disease. This study aimed to evaluate the effects of a 12-week, individually monitored, with moderate heart rate level intensity cardiovascular training on cardiovascular fitness and perceived disease activity in AS patients.

Methods

Patients diagnosed with AS according to the modified New York criteria were randomized to either cardiovascular training or attention control. The training group performed 3 cardiovascular training units per week. All participants attended 1 weekly usual care flexibility training session. Attention control contained regular discussion groups on coping strategies. Adherence was self-monitored. Assessments were performed at baseline and after the intervention period of 3 months. Physical fitness was the primary end point, measured in watts using a submaximal bicycle test following the physical work capacity 75% protocol. All analyses controlled for sex, age, body mass index, baseline fitness and physical activity levels, and Bath Ankylosing Spondylitis Disease Activity Index (BASDAI).

Results

Of 106 AS patients enrolled, 40% were women and the mean ± SD age was 49 ± 12 years. A total of 74.6% of the training group reported exercising at least 3 times a week. At the 3-month followup, the fitness level in the training group was significantly higher than in the control group (mean ± SE 90.32W ± 4.52W versus 109.84W ± 4.72W; P = 0.001), independent of other covariates. The mean BASDAI total score was 0.31 points lower (P = 0.31) in the training group, reaching significance for the peripheral pain subscore (1.19; P = 0.01) but not for back pain or fatigue.

Conclusion

Cardiovascular training, in addition to flexibility exercise, increased fitness in AS patients and reduced their peripheral pain.

INTRODUCTION

Ankylosing spondylitis (AS) is a chronic inflammatory rheumatic disease that affects the spine and iliosacral joints and is also associated with extraspinal manifestations, such as peripheral arthritis, enthesitis, uveitis, or bowel inflammation ([1]). The Assessment of SpondyloArthritis international Society/European League Against Rheumatism recommendations for the management of AS recommend drug therapy, such as nonsteroidal antiinflammatory drugs, and in severe cases, tumor necrosis factor α (TNFα) inhibitors, in combination with spinal flexibility exercise ([2]). Supervised flexibility exercises in AS patients have been shown to be effective in improving spinal flexibility, physical function, and patients' well-being ([3]).

Several observational studies suggest that patients with AS carry an increased risk of cardiovascular disease ([4-7]). In addition to chronic inflammation, reduced physical activity because of disease activity, pain, and fatigue may further contribute to AS-related cardiovascular disease ([7, 8]). Regarding physical activity, however, exercise recommendations in AS guidelines focus on spine flexibility rather than cardiovascular training ([2]). It can be assumed that these recommendations are widely followed in AS exercise groups, meaning that cardiovascular training is not an established element in AS-specific exercise.

A secondary analysis based on the third Cochrane review updated in 2008 included 12 controlled trials on exercise interventions in AS patients (10 from the 2008 Cochrane review and 2 more recently published studies) and evaluated if the exercise programs in trials for patients with AS were intensive enough to be effective ([9]). Intensity of exercise programs was determined according to the recommendations of the American College of Sports Medicine (ACSM), which summarize the current evidence for effective exercise interventions ([10, 11]). Exercise interventions are effective if they achieve a physiologic response, such as increased flexibility, muscular strength, and cardiovascular fitness, which is determined by the frequency, intensity, and duration of corresponding exercises, as well as patient adherence ([12]). The secondary analysis could not determine conclusively the benefit of cardiovascular training because all exercise programs focused on flexibility training, while cardiovascular fitness components were part of only 5 exercise programs, with only one meeting the ACSM recommendations for cardiovascular training.

In order to better define the role of cardiovascular training in patients with AS, we developed this study to test whether training designed to improve cardiovascular fitness is successful in AS patients attending standard flexibility exercise classes. We chose Nordic walking (NW) as a simple cardiovascular exercise mode that may also attract inactive individuals. Further, we assumed that NW would be especially suitable for AS patients because of its controlled loading on joints and the spine and the enhanced spinal rotation supported by the use of the walking sticks ([12, 13]).

Box 1. Significance & Innovations

  • Ankylosing spondylitis (AS) patients carry an increased risk of cardiovascular disease. This is the first large trial with AS patients focusing on increased cardiovascular fitness.
  • Many exercise studies do not achieve sufficient therapeutic validity. This study achieved sufficient therapeutic dose, including sufficient adherence of the patients to the training program and the study protocol.
  • AS patients are able to perform intensive cardiovascular training without increasing disease activity.
  • This study provides a basis for studying long-term effects of cardiovascular training on further parameters, e.g., biomarkers.

PATIENTS AND METHODS

Study design

This randomized controlled trial was conducted at the Department of Rheumatology and the Centre on Aging and Mobility at the University Hospital of Zurich in close collaboration with the AS patient organization, the Swiss Ankylosing Spondylitis Association (Schweizerische Vereinigung Morbus Bechterew [SVMB]). In addition, an SVMB member (CM) was involved in the planning and conduct of the study, ensuring that the trial addressed patient-relevant outcomes and that all of the trial documents were understood by the target population and helping to recruit patients.

The local ethics committees approved the study and all participants gave written informed consent. The study has been registered with the International Clinical Trials Registry.

Participants

Participants diagnosed with AS based on the modified New York criteria ([14]) were recruited from the SVMB membership registry, which includes approximately 2,500 members, and from rheumatology outpatient clinics and private practice rheumatologists in German-speaking regions of Switzerland. Additional inclusion criteria were age >18 years and sufficient German language skills. Exclusion criteria were moderate to severe heart disease (functional New York Heart Association class III and IV) and inability to cycle on an ergometer bike. The most important reasons for nonparticipation were no interest and no time. Two recruitment and outdoor training periods, each from January to May and from August to October, respectively, in 2 subsequent years ensured similar weather conditions for outdoor NW training.

Randomization and procedures

Participants were randomized to either the training group (cardiovascular training and flexibility exercise) or the control group (attention control and flexibility exercise) by sequentially numbered, concealed treatment allocations prepared in advance by an independent statistician. We used a stratified block randomization procedure (block size 4) ([15]) with TNFα inhibitor treatment (yes/no) as the stratification variable. We stratified for treatment with TNF inhibitors because of their known effect on the secondary outcome (Bath Ankylosing Spondylitis Disease Activity Index [BASDAI]), which may influence the patients' ability to perform physical training.

In this trial, participants and physiotherapists who instructed the NW program were aware of the treatment assignment. However, the physiotherapists who performed baseline and followup assessments were blinded to the participants' group allocation.

Interventions

Cardiovascular training

The training group performed a 12-week supervised NW training for 30 minutes twice a week using individually monitored, moderate-intensity heart rate (HR) levels. Moderate-intensity HR ranges of 55–75% and 65–85% of the maximum HR (HRmax) were used for participants who reached less than 100W and those who reached at least 100W, respectively, in the baseline bicycle test ([10, 11]). The intensity range was adjusted if an individual exceeded the upper HR limit repeatedly or constantly during at least 20 minutes of activity. Participants with a low fitness status, i.e., having reached less than 100W in the test and not being able to perform the training in their individual lower fitness range for a sufficient duration of at least 20 minutes, were first asked to keep walking for at least 20 minutes, and if this was achieved, to perform NW within the intensity range. All participants in the training group were provided with the NW equipment and an HR monitor (Polar watch; Polar Electro Europe). The NW training was performed in small groups of 2–6 participants and was led by instructing physiotherapists. Furthermore, participants in the training group were asked to perform at least one additional unsupervised, but HR-monitored cardiovascular training, NW, or other endurance activity, e.g., outdoor or ergometer biking, to achieve at least 3 training units per week. All physiotherapists who instructed the cardiovascular training previously underwent a standardized 4-hour education session.

Attention control

Instead of the NW training, the control group was offered an attention control intervention consisting of monthly 2.5-hour discussion groups on coping strategies and techniques of mindfulness-based stress reduction led by a psychologist ([16]).

Standard flexibility exercise

All study participants received the current standard of care and attended a weekly 1-hour exercise group supervised by a physiotherapist, with focus on spinal flexibility as offered by the Swiss AS Association throughout Switzerland.

Training locations and adherence support

Every effort was made to offer a training location closest to the home or office of the participants and dates and times of group training most convenient for them to facilitate participation. One of the 2 weekly supervised NW training sessions could be attended on the same evening as the flexibility exercise class to reduce time constraints for the participants. All participants received an individual schedule and were asked to keep a diary of all of their supervised and unsupervised physical activities.

Assessments

The primary outcome in this trial was cardiovascular fitness, assessed with a submaximal bicycle test on an electrocardiogram (EKG)–equipped ergometer (Schiller-Resmed, cardiopulmonary diagnostics) following the physical work capacity 75% protocol (PWC75%) to estimate aerobic capacity (VO2 max) ([17]). A submaximal endurance test is considered more sensitive to change than a maximal endurance test ([18]). According to the PWC75%, heart rates at 55%, 65%, and 75% of the estimated age-related maximum HR were calculated and the watts produced at these heart rate levels were measured. Although heart disease (New York Heart Association class III and IV) was an exclusion criterion, simultaneous EKGs were applied in all participants for safety reasons during exercise testing. The cardiologist reviewed all of the EKGs and if identifying any abnormality, comparison with the patient's cardiovascular disease history was performed to decide whether the patient had to be excluded.

The secondary outcome was perceived disease activity assessed with the BASDAI on a 0–10 numerical rating scale (NRS; where 0 = none and 10 = very severe) ([19]). Predefined in our protocol, we assessed both the BASDAI total score and the subscales for spinal pain, peripheral pain, and fatigue.

Additional exploratory outcomes were 1) AS-specific functional health, assessed with the Bath Ankylosing Spondylitis Functional Index ([20]), Bath Ankylosing Spondylitis Metrology Index ([21]), and Bath Ankylosing Spondylitis Patient Global Score Index, including the patient's global assessment of disease activity, general pain, and nocturnal pain, all measured on a 0–10 NRS ([22]); 2) type, amount, and intensity of physical activity by use of the Office in Motion Questionnaire (OIMQ) ([23]), subsequently assigning metabolic equivalents to each reported activity ([24]), and an accelerometer (Actigraph, Manufacturing Technology); the small waist-mounted device is worn for 7 days, including a complete weekend, to calculate reliable average physical activity per day ([25, 26]), given as counts/minute (number of accelerations), minutes spent in moderate activity, and units of at least 20 minutes in vigorous activity per week, using the cut points defined by Swartz et al ([27]); 3) psychological status, using the German version of Hospital Anxiety and Depression Scale, which assesses anxiety and depression on two 7-item 0–3 scales (where 0 = no problems and 3 = severe problems) ([28]); 4) perceived general health using the EuroQol, applying a 0–100 visual analog scale (where 0 = worst health and 100 = best health) ([29]); and 5) laboratory data of disease activity (erythrocyte sedimentation rate and C-reactive protein [CRP] level) and metabolism (cholesterol and triglycerides). Further, we calculated the Ankylosing Spondylitis Disease Activity Score with the CRP level using parameters from the BASDAI and CRP level ([30]).

Sample size

We based our sample size calculation on the reported means ± SDs from a similar exercise trial in patients with rheumatoid arthritis (RA) ([31]). To achieve a minimum clinically meaningful difference of 20% for the primary outcome, 49 patients in each group were needed to achieve 90% power to detect this difference.

Statistical analysis

Statistical analysis was performed on an intent-to-treat basis. Analysis of variance (ANOVA) models were used to compare cardiovascular fitness levels, BASDAI total scores and subscores, and the exploratory outcome variables at followup. The crude models controlled for TNFα treatment status as a stratification variable and baseline level of the characteristic. The fully adjusted models additionally controlled for age, sex, body mass index, smoking status, and baseline levels of perceived disease activity, physical activity, and fitness. The data were analyzed using SAS statistical software, version 9.2 (2002–2008). All statistical tests were 2 sided, and the significance level was set at 0.05.

RESULTS

From a total of 185 AS patients who attended the information meetings about the study, 106 confirmed their participation and met the inclusion criteria. Forty-seven of these patients were enrolled in the first year and 59 were enrolled in the second year of enrollment (Figure 1). Baseline characteristics of the 2 groups were similar (Table 1), and formal statistical testing detected no significant differences. The participants had no history of heart disease, and no signs of coronary ischemia were observed in the EKGs. The cardiologist identified 3 patients with EKGs suggestive of left ventricular hypertrophy and reviewed these patients' cardiovascular histories, but no exclusions were performed. Although not assessed systematically, participants seemed to have a low cardiovascular risk: 19 patients were receiving antihypertensive medication, with 6 of them additionally receiving aspirin and/or a statin.

Figure 1.

Trial flow chart.

Table 1. Baseline characteristics of participants*
 Training group (n = 53)Control group (n = 53)P
  1. Values are the mean ± SD unless indicated otherwise. The chi-square test was used for categorical variables and nonparametric statistics (Wilcoxon's rank sum test) were used for all continuous variables. BMI = body mass index; TNF = tumor necrosis factor; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASG = Bath Ankylosing Spondylitis Patient Global Score Index; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; OIMQ = Office in Motion Questionnaire; METs = metabolic equivalents; HADS = Hospital Anxiety and Depression Scale; ASDASCRP = Ankylosing Spondylitis Disease Activity Score calculated with C-reactive protein (CRP) values.
Men, no. (%)34 (64)34 (64)1.00
Age, years50.1 ± 11.947.6 ± 12.40.29
Disease duration, median (range) years9 (0.5–45)8 (0.5–39)0.60
BMI, kg/m225.2 ± 4.125.5 ± 4.30.71
Smokers, no. (%)11 (20)16 (30)0.37
TNF treatment, no. (%)15 (28)16 (30)1.00
Fitness, W91.3 ± 37.4101.4 ± 45.50.21
Resting heart rate82.3 ± 11.482.1 ± 13.50.94
Heart rate at the end of the test130.5 ± 9.6132.7 ± 10.30.29
BASDAI (range 0–10)3.3 ± 1.93.6 ± 2.10.57
BASDAI fatigue (range 0–10)4.4 ± 2.45.0 ± 2.70.27
BASDAI neck/back/hip pain (range 0–10)3.8 ± 2.54.2 ± 2.70.65
BASDAI peripheral pain (range 0–10)2.2 ± 2.32.7 ± 2.60.36
BASG night pain (range 0–10)3.1 ± 2.82.9 ± 2.80.59
BASG pain (range 0–10)3.2 ± 2.03.5 ± 2.50.80
BASG disease activity (range 0–10)3.7 ± 2.34.2 ± 3.10.72
BASFI (range 0–10)2.4 ± 1.92.4 ± 2.10.92
BASMI (range 0–10)2.9 ± 2.12.8 ± 1.90.99
OIMQ, METs per week71.8 ± 39.178.4 ± 58.50.89
Accelerometer, no. of accelerations per day (counts/minute)336.3 ± 184.9370.5 ± 145.00.18
Accelerometer, moderate activity (minutes/day)145.9 ± 54.2170.4 ± 64.50.09
Accelerometer, vigorous activity units (≥20 minutes/week)11.1 ± 9.413.6 ± 12.90.50
HADS anxiety (range 0–21)6.9 ± 5.36.7 ± 4.50.92
HADS depression (range 0–21)5.2 ± 4.45.0 ± 4.50.70
EuroQol health score (range 0–100)64.5 ± 22.065.9 ± 21.20.75
ASDASCRP2.2 ± 0.82.3 ± 1.00.40
CRP level, mg/liter7.5 ± 9.86.4 ± 8.70.53
Cholesterol, mmoles/liter5.4 ± 0.95.3 ± 1.00.41
Triglycerides, mmoles/liter1.1 ± 0.61.4 ± 1.00.07

Based on the physiotherapists' protocols for group adherence and on the participants' diaries, 74.6% of the training group performed at least 3 training units per week (mean 3 trainings/week), i.e., 2 NW training sessions and 1 additional unsupervised cardiovascular training unit, but only 25% of the control group performed ≥3 trainings per week (mean 1 training per week).

In a few patients (n = 4) who were quite fit already at the beginning of the study and who exercised in the upper intensity range, i.e., 65–85% of the HRmax, the upper limit was increased to 90% of the HRmax during the intervention period; all of the less fit participants continued exercising in the lower intensity range during the intervention period, but at the end of the study, all of them achieved the required intensity and duration.

Treatment effects

Primary outcome

At the 3-month followup, both the minimally adjusted and the fully adjusted ANOVA models showed a significant benefit in fitness level in the training group compared to the attention control group. In the minimally adjusted model, the mean ± SE fitness level in the training group was 107.98W ± 3.98W compared to 87.78W ± 3.87W in the control group (P = 0.0004). In the fully adjusted model, mean ± SE fitness level in the training group was 109.84W ± 4.72W compared to 90.32W ± 4.52W in the control group (95% confidence interval [95% CI] 9.18, 31.24; P = 0.001), the difference therefore being independent of the covariables (Table 2).

Table 2. Primary and secondary outcomes at the 3-month followup*
 Minimally adjusted ANOVA modelaMultivariate adjusted ANOVA modelb
Exercise groupControl groupPExercise groupControl groupP
  1. Values are the mean ± SE. ANOVA = analysis of variance; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index.
  2. aAdjusted for tumor necrosis factor α (TNFα) treatment (stratification variable) and baseline level of the characteristic.
  3. bAdjusted for age, sex, body mass index, smoking status, baseline physical activity, baseline perceived disease activity, TNFα, and baseline level of the characteristic.
  4. cBaseline values are given in Table 1.
  5. dSignificant at 0.001.
  6. eSignificant at 0.05.
Primary outcome      
Fitness level, Wc107.98 ± 3.9887.78 ± 3.870.0004d109.84 ± 4.7290.32 ± 4.520.001d
Secondary outcome      
BASDAI total (range 0–10)c3.07 ± 0.203.35 ± 0.200.332.84 ± 0.243.15 ± 0.230.31
BASDAI fatigue (range 0–10)3.73 ± 0.324.29 ± 0.320.223.64 ± 0.394.39 ± 0.370.13
BASDAI neck/back/hip pain (range 0–10)3.31 ± 0.334.15 ± 0.300.073.25 ± 0.383.98 ± 0.360.13
BASDAI joint pain (range 0–10)2.05 ± 0.312.74 ± 0.300.111.32 ± 0.342.36 ± 0.330.02e

At baseline, 2 patients in each group were not able to perform the PWC75% test because of their low fitness level. At followup, no patients in the training group, but 6 patients in the control group, were not able to fulfill the PWC75% test protocol, which we considered related to the training effect.

The OIMQ and the accelerometer were administered before and shortly after the intervention period in order to assess the participants' actual physical activity. Therefore, the results may not reflect their amount of physical activity during the intervention period, but rather their usual physical activity. On average, this seemed to be the same in both groups after the study conclusion, although the NW participants had performed substantially more physical activity during the intervention period.

Secondary outcome

There was no difference between the 2 groups in the BASDAI total score at the 3-month followup (Table 2). For the subscores, we found a significantly lower level of peripheral pain in the training group (mean ± SE 1.32 ± 0.34) compared to the control group (mean ± SE 2.36 ± 0.33 [95% CI −1.89, −0.18]; P = 0.02), while the subscores for fatigue and neck/back/hip pain were not different between the groups.

Additional exploratory outcomes

There were no significant differences between the 2 groups regarding the exploratory end points, either in functional measures or in biomarkers (Table 3). With respect to the Bath indices, functional limitations were generally low.

Table 3. Additional exploratory outcomes*
 Minimally adjusted ANOVA modelaMultivariate adjusted ANOVA modelb
Exercise groupControl groupPExercise groupControl groupP
  1. Values are the mean ± SE. All baseline values are given in Table 1. ANOVA = analysis of variance; BASG = Bath Ankylosing Spondylitis Patient Global Score Index; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; OIMQ = Office in Motion Questionnaire; METs = metabolic equivalents; HADS = Hospital Anxiety and Depression Scale; ASDASCRP = Ankylosing Spondylitis Disease Activity Score calculated with C-reactive protein (CRP) values.
  2. aAdjusted for tumor necrosis factor α (TNFα) and baseline level of the characteristic.
  3. bAdjusted for age, sex, body mass index, smoking status, baseline physical activity, baseline fitness level, Bath Ankylosing Spondylitis Disease Activity Index score, TNFα treatment, and baseline level of the characteristic.
BASG night pain (range 0–10)3.00 ± 0.262.63 ± 0.270.322.26 ± 0.322.64 ± 0.310.34
BASG pain (range 0–10)3.25 ± 0.293.39 ± 0.280.743.13 ± 0.353.20 ± 0.330.88
BASG disease activity (range 0–10)4.05 ± 0.353.74 ± 0.340.523.87 ± 0.403.59 ± 0.380.59
BASFI score (range 0–10)2.49 ± 1.772.41 ± 1.700.732.53 ± 0.212.40 ± 0.200.63
BASMI score (range 0–10)2.64 ± 0.253.02 ± 0.240.272.40 ± 0.282.66 ± 0.280.46
OIMQ, METs per week51.37 ± 6.2755.40 ± 6.390.6552.94 ± 7.2458.03 ± 7.510.61
Accelerometer, no. of accelerations per day (counts/minute)333.35 ± 22.16335.34 ± 20.440.95373.98 ± 30.00349.87 ± 24.950.45
Accelerometer, moderate activity (minutes/day)141.17 ± 8.55147.81 ± 7.940.26149.14 ± 11.58146.58 ± 9.530.83
Accelerometer, vigorous activity units (≥20 minutes/week)7.45 ± 1.259.50 ± 1.320.268.74 ± 1.779.36 ± 1.520.74
Resting heart rate81.10 ± 1.9581.57 ± 1.860.8677.89 ± 1.9380.27 ± 1.910.33
Heart rate at the end of the test128.30 ± 0.59128.37 ± 0.590.93128.92 ± 0.69128.70 ± 0.680.80
HADS anxiety (range 0–21)6.27 ± 0.356.58 ± 0.340.526.58 ± 0.406.63 ± 0.390.93
HADS depression (range 0–21)5.10 ± 0.314.48 ± 0.300.165.18 ± 0.374.33 ± 0.350.07
EuroQol health score (range 0–100)64.24 ± 3.0563.01 ± 2.930.7761.97 ± 3.3165.19 ± 3.130.44
ASDASCRP2.26 ± 0.152.16 ± 0.150.622.43 ± 0.172.23 ± 0.180.38
CRP level, mg/liter6.27 ± 1.084.95 ± 1.070.397.49 ± 1.224.93 ± 1.300.12
Cholesterol, mmoles/liter5.66 ± 0.155.90 ± 0.160.285.66 ± 0.185.93 ± 0.190.24
Triglycerides, mmoles/liter1.30 ± 0.121.42 ± 0.130.481.23 ± 0.141.35 ± 0.160.52

DISCUSSION

To our knowledge, this is the first controlled trial to test the effect of cardiovascular training in addition to standard flexibility exercise on patients with AS. The results demonstrate that an appropriately designed and conducted cardiovascular training that meets recommended standards leads to significantly improved cardiovascular fitness in AS patients. NW as a strategy to improve cardiovascular fitness in AS patients was well tolerated by the participants who, despite having a chronic inflammatory condition, were able to exercise frequently, on moderate heart rate level intensity, and over a longer period without increasing their disease activity or pain. This supported high adherence to the intervention, which is key in regard to exercise studies. The clear association between exercise participation rates and achieved fitness level increased the credibility of the diary protocols and demonstrated adherence to the study protocol. The fact that resting HR did not decrease to the expected extent in the training group may be due to the relatively short training period. Furthermore, the decrease in HR after a course of exercise training has been described mainly in patients with established coronary heart disease or chronic heart failure ([32, 33]).

Regular physical activity and aerobic exercise training are related to a reduced risk of coronary events in healthy individuals ([34-36]) and subjects with coronary risk factors ([37]) and established coronary artery disease ([32]). Therefore, physical activity and aerobic exercise training are recommended by international guidelines for primary and secondary cardiovascular disease prevention ([38, 39]).

We have shown that cardiovascular training is safe and feasible in patients with AS and increases their fitness level independent of their initial fitness level. Whether this benefit translates into the prevention of cardiovascular disease needs to be tested in a larger longer-term clinical trial. Notably, we measured biomarkers of cardiovascular health as exploratory end points (cholesterol and triglycerides), which did not improve significantly in the training group over the control group, possibly also because of the short followup period. Unfortunately, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol have not been analyzed separately. Therefore, we could only speculate about potential changes in the HDL to LDL ratio with an increase of protective HDL levels and a corresponding decrease of LDL levels after the training course. Our initial assumption was that NW training may also improve fatigue, as assessed by the BASDAI. Fatigue is one of the predominant symptoms of AS patients and has been shown to be influenced by physical activity ([40]). However, although many NW participants reported feeling more energy after the training and some cardiovascular training studies in RA showed positive effects on fatigue, even in rather small samples ([41, 42]), we could not demonstrate changes in the BASDAI total score or in the fatigue subscore in this group. Disease activity in our sample was generally low because we recruited patients already treated according to current standards, including the availability of TNF inhibitors, and the majority of our patients were not classified as fatigued ([43]), which may partly explain our results.

Interestingly, however, participants in the training group reported a significant improvement in the peripheral pain subscale over the control group, which addresses another important end point in patients with AS. Studies for effects of cardiovascular training on pain in musculoskeletal diseases are scarce; only 2 early studies in RA ([31, 44]) found that the dynamic exercise groups significantly decreased the number of clinically active joints. The 2006 Cochrane review evaluating 6 trials on dynamic exercise therapy for treating RA, however, concluded that there were no positive effects, but also no negative effects, on pain ([45]).

Our trial has several strengths. First, it was the largest exercise trial in AS patients thus far, and in addition to providing an intervention based on the current and established ASCM recommendations, we chose a training strategy that was feasible and well tolerated by participants. Further, we guaranteed high-quality outcome assessments by blinding the assessment physiotherapists to the treatment allocation at baseline and followup. Also, our study was appropriately powered to detect a statistically significant difference for the primary outcome. Moreover, the attention control intervention lends credibility to the training effect of NW.

There are also limitations to our study, including a limited sample size for the secondary and exploratory end points tested and the short followup period of 3 months. Further, no efforts were made to support the participants' achieved fitness levels. However, support is usually necessary to keep patients on their achieved physical activity levels ([46]). Physiotherapists' coaching of early-stage RA patients, in terms of 2 face-to-face meetings and subsequent regular telephone calls, has been shown to successfully support their physical activity adherence over 1 year ([47]). However, after another year without any support, no differences in physical activity levels were present compared to the controls ([48]). It seems that even distance coaching is effective to keep people active and to maintain exercise effects.

In summary, the improved cardiovascular fitness and the significant improvement in BASDAI peripheral pain support the inclusion of cardiovascular training as an additional exercise strategy for people with AS. Future research should test the long-term effects of NW on cardiovascular health in people with AS.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Niedermann had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Niedermann, Muggli, Dagfinrud, Hermann, Tamborrini, Ciurea, Bischoff-Ferrari.

Acquisition of data. Niedermann, Muggli, Hermann, Tamborrini, Ciurea, Bischoff-Ferrari.

Analysis and interpretation of data. Niedermann, Sidelnikov, Dagfinrud, Ciurea, Bischoff-Ferrari.

Acknowledgments

The authors would like to thank Rebecca Lang and Daniela Zenger from the SVMB for the organizational and administrative work, and René Bräm, CEO, for his support throughout the study. The authors also thank Dr. Fabienne Matthier, psychologist, Department of Psychosocial Health, University Hospital Zurich, for providing the mindfulness-based stress reduction program to the control participants; Ursula Abt and Esther Hartmann, physiotherapists, for educating the physical therapists providing the NW training; all physiotherapists hired by the SVMB for providing the training and flexibility exercise interventions with great commitment; all physiotherapists employed at the Department of Rheumatology, Institute of Physiotherapy, University Hospital Zurich, for conducting the assessments, and their superiors for making them available for the study; Dr. Stephen Ferrari, Ferrari Data Solutions, for data management support; Dr. Urs Mäder, Johanna Hänggi, and Corinne Aebischer from the Federal Office of Sports for their advice and handling of accelerometer data; and Michelle Schmocker for reading data. Finally, the authors thank the European League Against Rheumatism Standing Committee of Health Professionals for awarding the primary investigator with an educational visit grant to Professor Kare B. Hagen and Dr. Hanne Dagfinrud, Oslo, in the preparation stage of this study.

Ancillary