Participants
This three parallel groups, randomized, controlled and assessor blind study was conducted in Sport Injuries and Corrective Exercises Department of Razi University and Rheumatology Department, Clinical Research Development Center, Imam Reza Hospital. Patients (all the patients were men) with 1988 modified New York criteria [15] for AS were enrolled in the study and assigned to three groups. In addition to their demographic characteristics including age, weight, height, and body mass index (BMI), they were also to recount the presence or absence of the main symptoms, time of diagnosis, and drug usage status. Exclusion criteria were the presence of prosthesis, hypertension, cardiovascular disease, chronic obstructive pulmonary disease, and exercising regularly during the previous three months [16]. Patients whose treatment regimens had changed during the last two months prior to the study were not included. Additionally, patients who were not vaccinated with two doses of Covid-19 vaccination were excluded. The included patients were asked not to use supplementary drugs or change the usual dosages throughout the study period. For a more accurate pain assessment, they were asked not to take any analgesics in the morning of the assessment day. The participants were fully informed about the nature and purpose of the study, and written informed consent was obtained from all subjects. It is notable that local ethics committee approval was obtained (IR.KUMS.REC.1399.1137), and the study was registered in Iranian Registry of Clinical Trials (IRCT20190426043377N3, registered on 22/05/2021, https://fa.irct.ir/user/trial/56058/view). patient recruitments were also started on 06/07/2021.
G ∗ power statistical software (Heinrich Heine University, Düsseldorf, Germany) was used to determine the sample size for a repeated-measures analysis of variance at 90.0% power, an effect size of 0.31 [17], and a 5.0% level of significance. As a result, the minimum number of patients required in the Aqua Stretch, aqua Pilates, and control group was estimated to be at least 13 patients. Dropping out rate was considered 20%. The patients were randomly assigned to three groups: aqua-based stretch (n = 17), aqua-based Pilates (n = 17), and a control group (n = 17) (Fig. 1). A software tool (Random Number Generator) was dedicated to simple randomly generating sequences, and the corresponding numbers were provided in sealed, opaque envelopes opened upon each patient’s agreement to participate in the research. Concealment of allocation was done by assigning the allocation task to a research assistant who was not involved in the recruitment process. Participants were randomly assigned to the Aqua Pilates and Aqua Stretch groups or no treatment (control) group in a 1:1:1 ratio. Eligible participants were informed of their allocation result by a project manager following the assessment of their baseline measurements. In this study, blinding of the outcome assessors and the statistician was maintained even though it was impossible to blind the participants and exercise coaches.
Interventions
All patients of the intervention groups were trained under the supervision of a Corrective Exercises and Sports Injuries specialist. He was a certified Pilates and aquatic therapy practitioner with 5 years of experience.
Aqua-based Stretch program
In total, 24 Aqua Stretch exercise sessions were held (four sessions per week), lasting for six weeks. Exercises were performed in a swimming pool at 32–33°c, and each session lasted 60 min. The program started with 10–15 min of poolside exercises including warming-up (forward, backward, and sideward walking and jogging). Self-wall side passive stretching exercises, executed by noodles, were done for 30-min during the first two weeks, followed by self-wall side active stretching exercises were performed in 35-min sessions for the second two weeks, and in 45-min sessions for the third two weeks. Each stretching exercise lasted 15 s, 30 s, and 45 s for the first, second, and third two weeks, respectively. Stretching exercises were performed in two series; the first series of stretches included stretching exercises performed on each joint [18, 19]. The second series of stretches was performed based on the pattern of muscle chains [20, 21]. The program ended with cooling-down including slow walking and passive stretching for 5–15 min.
Aqua-based Pilates program
In this study, a total of 24 sessions of aqua-based Pilates exercises was performed (four sessions per week), lasting for six weeks. The exercises were performed in a swimming pool at 32–33 °C, and each session lasted 60 min. The program started with 10 min poolside exercises including warming-up (five min of walking and five min of general active range of motion (ROM) and stretching). Poolside exercises were followed by 45 min of main exercises such as: breathing exercises, standing criss cross, T balance with noodle, standing spine stretch (3 set, 20 s hold), spine twists, standing single kicks (3 set, 10 repetition) for the first two weeks. Breathing, standing spine stretch and single side leg lift, plank on a noodle, side kick, hip work, single leg stretch (by cycling on a noodle), T balance exercises (3 set, 10 repetition) for the second two weeks. Finally, breathing and side kick exercises, side plank with one hand on a noodle, spine twist (3 set, 10 repetition), cat stretch and T balance in water perturbation (3 set, 20 s hold). The program ended with cooling-down, including slow walking and passive stretching for five min [22].
Meanwhile, the control group patients continued their usual activities and drug treatment (NSAIDs & Anti TNF).
Outcome measures
The patients were assessed in terms of pain, spinal mobility, physical function, and quality of life. Evaluations were performed before (week 0) and after (week 6) treatment. The research assessor was blinded since they assessed the patients’ clinical assessment parameters prior to their assignment to groups. In addition, spinal mobility of the subjects was assessed by Spinal Moues and physical function was evaluated by The Bath Ankylosing Spondylitis Functional Index (BASFI) and 40-m walking Test (MWT). Pain was also assessed by 0–10 cm visual analog scale (VAS). Finally, quality of life was evaluated by Ankylosing spondylitis quality of life (ASQoL) questionnaire.
Pain
In this study, the VAS was used to measure the intensity of perceived pain. The mentioned tool is a ruler in the form of a horizontal strip with the length of 10 cm, on which one end is zero (no pain) and the other end is 10 (excruciating pain). The ruler also has two qualitative and quantitative sides. In the present study, the patients were asked to mark the qualitative side of the ruler based on their perceived pain intensity. Following that, the marked point was recorded by the researcher as the pain rate of patients [23]. The ICC (95% CI 0.96 to 0.98) of this scale was reported to be 0.97 [24].
Functional capacity
The BASFI is composed by ten questions elaborated to determine the degree of functional limitation in patients with AS. The participants were asked to indicate their level of ability within the last week on a 10-cm horizontal line (from easy to impossible). The BASFI sum score was calculated by adding all scores from questions 1–10 and then dividing them by 10 (ICC was reported 0.94) [25, 26]. In addition, the 40-MWT was used to assess functional capacity [27]. For this purpose, the duration of walking a distance of 40 m was recorded with a stopwatch in seconds. A shorter duration indicated greater physical fitness (ICC = 0.99) [28].
Health quality
The ASQOL questionnaire is an instrument specific for patients with AS [29]. The tool consists of 18 items requesting a yes or no response to questions. All item scores are summed to give a total score or index which may range from 0 (good QoL) to 18 (poor QoL). The correlation coefficient for test–retest reliability of the ASQOL questionnaire was 0.85 (95% CI 0.80–0.89) [30].
Spine posture evaluation
A spinal mouse device (model: 3.32, made in Switzerland) was used to measure the spinal curves angles at the sagittal plane (ICC: 0.63–0.97) [31]. The spinal mouse is a computer-assisted device, with which spinal shape and mobility is assessed with surface-based measurements. In this test, the participants stood in front of the assessor with bare feet, and the C7 and S2 landmarks were marked. In the next stage, the angle of the thoracic kyphosis and lumbar lordosis in the upright position were measured by placing the mouse device on the C7 vertebra and pulling down the vertebral spinous. The flexion and extension range of motion (ROM) was also measured in the upright position. Moreover, thoracic kyphosis, lumbar lordosis, and sacral inclination were measured in the next stage. Thoracic kyphosis refers to forward curvature of the thoracic spine in the sagittal plane, from the first thoracic vertebra (T1) to the 12th thoracic vertebra (T12) and the sum of the 11 angles from T1/2 to T11/12. On the other hand, the lumbar lordosis angle is the sum of the six angles from T12/lumbar 1 to lumbar 5/sacral 1 (S1) in the curvature throughout the lumbar spine (T12-S1). Finally, sacral slope (SS) is defined as the angle between the horizontal and the sacral plate, drawn along the segmented endplate of the sacrum. In this study, the positive values obtained from the measurements were attributed to kyphosis, whereas the negative values indicated lordosis [32].
Statistical analysis
Statistical analysis was performed in SPSS version 22.0.0 (IBM Corporation, Armonk, NY, USA). In addition, means and standard deviations were calculated for all dependent variables. The Shapiro-Wilks test was applied to test the normality of quantitative data, and One-way ANOVA test was used to assess differences between the groups at baseline and compare the mean differences of variables among groups. Moreover, repeated measure ANOVA was employed to determine the main and interaction effects of Time (pretest vs. posttest) and Group (Aqua Stretch, Aqua Pilates, and control) on the outcome measures. Bonferoni test was used for post-hoc analysis when necessary. Notably, paired sample t-test was used to investigate the within-group comparisons in case of significant interaction effects. Furthermore, one-way ANOVA was used to evaluate the between-group comparisons. It is notable that P ≤ 0.05 was considered statistically significant.
Based on the 95% confidence interval in the current study, the minimal clinically important difference (MCID) was assessed using the distribution-based approach and the reliable change index (RCI), while the minimal detectable change (MDC) was evaluated using the standard error of measurement (SEM) and the following equations [33]:
$${\text{SEM}} = {\text{SD}}_{{{\text{pre}}}} \surd {1} - {\text{r}}_{{{\text{test}}}}$$
$${\text{MDC}}_{{\% {95}}} = {1}.{96} \times \surd {2}\;{\text{SEM}}$$
$${\text{MCID}}_{{{\text{RCI}}}} = {1}.{96} \times {\text{SD}}_{{{\text{pre}}}} \left\{ {\surd [{2} \times ({1} - {\text{r}}_{{{\text{test}}}} )]} \right\}$$
