This is a randomized controlled trial which was conducted in accordance with the Helsinki declaration and approved by the research ethics committee of the Methodist University of Piracicaba-São Paulo-Brazil under protocol #13/11. All volunteers signed an informed consent form prior to their inclusion in the study.

Sixty-two patients from a Regional Specialty Outpatient Clinic diagnosed with COPD were invited to participate in the study. The inclusion criteria consisted of having a diagnosis of COPD, according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) recommendations (GOLD, 2022), and a low level of physical activity according to the International Physical Activity Questionnaire (IPAQ), defined as engaging in less than 600 MET-minutes of total physical activity per week or walking less than 30 minutes per day (Craig et al., 2003). Participants were excluded if they were not clinically stable; had experienced a COPD exacerbation in the past three months; had participated in regular physical training programs within the past six months; or presented with any of the following conditions: diabetes mellitus, endocrine disorders, liver diseases, other pulmonary diseases, associated inflammatory diseases, heart conditions, musculoskeletal or neuromuscular impairments that could prevent the execution of the experimental protocols, dermatological conditions (e.g., skin fistulas, infected wounds, skin mycoses, varicose ulcers), hypersensitivity to products used in pool water treatment, and/or hydrophobia.

Twenty-four male patients, aged between 50 and 75 years, met the inclusion criteria and accepted to participate, being randomly allocated into the control group (CG, n=12) or aquatic training group (ATG, n=12). Randomization was performed by a researcher who was not involved in the study. A random assignment website (https://www.sealedenvelope.com/simple-randomiser/v1/lists) was used to generate a sequence of numbers indicating group allocation (CG or ATG). The group assignments were printed on paper and placed into opaque envelopes according to the generated order. These envelopes were opened by the researcher responsible for the rehabilitation program only after the baseline assessments had been completed. Due to exacerbation of the COPD or personal reasons, 3 volunteers were excluded from the CG and 4 from the ATG (Figure 1). The CG was instructed to maintain the conventional clinical treatment, characterized by drug therapy, receiving instructions on healthy lifestyle, and only participated in the assessments. On the other hand, the ATG was submitted to an aquatic intermittent aerobic physical exercise program (Figure 1).

Figure 1 Figure 1 Inclusion criteria flowchart Figure 1 Inclusion criteria flowchart Source: the authors.

Pulmonary function

The lung function was assessed through spirometry tests using a portable spirometer (Easy oneTM, ndd Medizintechnik AG, Zurich, Switzerland), in accordance with the American Thoracic Society (ATS) and European Respiratory Society (ERS) guidelines. Forced vital capacity (FVC), forced expiratory volume in the first second (FEV1) and the FEV1/FVC ratio were obtained as absolute and relative values (% of the predicted) (ATS, 2002; Miller et al., 2005; Pereira et al., 2007).

Quality of life

Quality of life was assessed through the application of the Brazilian version of the Saint George Respiratory Questionnaire (SGRQ) (Camelier et al., 2006). This is a self-administered questionnaire that comprises three domains: 1) “Symptoms”: which provided information about discomfort due to respiratory symptoms; 2) “Activities”: which assessed changes in physical activity; and 3) “Impact”: which assessed the global impact on the patient’s activities of daily living and well-being. In addition, the total score was also quantified. The results were expressed in percentages ranging from 0 to 100%, a higher score meaning worse performance in each domain.

Functional Capacity

The functional capacity was assessed by the six-minute walk test (6MWT). The participants were requested to wear appropriate clothing (sportswear and trainers). Two tests were performed with a 30-minute interval and executed by two physical therapists. The repetition of two tests aimed at eliminating the learning effect and ensuring the reproducibility of the procedure (Hernandes et al., 2011). The best test was used in the data. The volunteers were instructed to walk at maximum speed as possible during the six minutes. During the walk, the volunteers were encouraged by the examiner through verbal stimuli at every minute, using the phrases recommended by the American Thoracic Society (ATS, 2002). The distance covered in meters was recorded at the end of the test. The blood pressure (BP), heart rate (HR), respiratory rate (RR), peripheral oxygen saturation (SpO2), and dyspnea and lower-limb (LL) fatigue score using the CR10 Borg scale were recorded at the beginning, during and at the end of the 6MWT. In case of desaturation (SpO2<90%), the test was interrupted, oxygen supplementation was provided, and another test was started (Borghi-Silva et al., 2010).

Functional status

The functional status was assessed using the modified version of the Pulmonary Functional Status and Dyspnea Questionnaire (PFSDQ-M). The PFSDQ-M consists of three domains: 1) influence of dyspnea on ADL, 2) influence of fatigue on ADL (5 general items and 10 specific items for each domain); and 3) change in ADL compared to the period prior to the disease (10 specific items). There are five general questions in the fatigue and dyspnea domains, which are informative and qualitative in nature and are not added to the questionnaire score. The participants describe how much dyspnea and fatigue interfere with the ten specific ADL items, scoring between 0 and 10 for each activity, being 0 (no interference), 1-3 (mild), 4-6 (moderate), 7-9 (severe) and 10 (very severe). In the third domain, the amount of change in ADL is reported in comparison to the period prior to the disease, choosing a value between 0-10 for each activity with 0 being as active as ever regarding this activity; 1-3 for small changes; 4-6 for moderate changes; 7-9 for extreme changes; and 10 for when no longer doing this activity. Subsequently, a partial score is calculated, ranging from 0 to 100 for each domain and a total score, which is composed of the sum of the partial scores of the three domains, ranging from 0 to 300. Higher values on the scale indicate greater limitations in ADL (Kovelis et al., 2008). A minimum change of 5 points in each component was considered as the minimum clinically important difference (MCID) (Regueiro et al., 2013).

London Chest Activity of Daily Living Scale

The level of dyspnea during the activities of daily living was assessed by the London Chest Activity of Daily Living Scale (LCADL), which is composed of four domains: 1) personal care, 2) domestic activities, 3) physical activities and 4) leisure activities, which includes 15 questions. The domains present items that receive a score ranging from 0 to 5, totaling a maximum score of 75 points, with higher values on the scale indicating greater limitations in ADL (Carpes et al., 2008). A minimum of 4 points change in the total score of the scale was considered as the MCID (Bisca et al., 2014).

Dyspnea

The dyspnea was also assessed using the Brazilian version of the Medical Research Council (MRC) Dyspnea Scale, which comprises five statements, with the patient reporting their subjective degree of dyspnea corresponding to how much it interferes with ADL as 1 (“only suffers from shortness of breath during intense exercise”); 2 (“suffers from shortness of breath when walking quickly or climbing a light slope”); 3 (“walks slower than people of the same age because of shortness of breath or has to stop to breathe even when walking slowly”); 4 (“stops walking to breathe after walking less than 100m or after a few minutes”); and 5 (“feels so short of breath that they no longer leave the house or when getting dressed”) (Kovelis et al., 2008).

BODE index

The BODE (B-body mass index; O-airflow obstruction; D-dyspnea; E-exercise capacity) mortality predictor index assesses the risk of mortality using the body mass index (BMI), FEV1, MRC scale score and the distance covered in the 6MWT. The score in the BODE index ranges from 0 to 10, with a higher score indicating greater severity (Celli et al., 2004).

Intermittent aquatic aerobic training protocol

The 8-week intermittent aquatic aerobic training, lasting 60 minutes, on alternate days, three times a week (American College of Sports Medicine Position Stand [ACSM], 1998; Nici et al., 2006) was performed based on a previously described protocol (Kim et al., 2010), adapted to the patients, which consisted of: 1) warm-up (10 minutes): stretching exercises and dynamic aerobic exercises; 2) aquatic physical conditioning: aerobic training consisting of aquatic exercises for the trunk, upper limbs (UL) and lower limbs (LL), with intensity prescribed at 4-6 on the CR10 Borg scale (Carvalho et al., 2009; Graef & Kruel, 2006; Horowitz et al., 1996), performed in an intermittent form, 1 minute of exercise and 1 minute of rest progressively increasing (Vogiatzis et al., 2002). The first eight sessions lasted 20 minutes, then the next eight sessions lasted 30 minutes and the last eight sessions lasted 40 minutes. 3) Cooling down (10 min): global stretching exercises. This training program was developed in a heated pool (32ºC). During all sessions, the patients were instructed to use pursed-lip breathing. Supplementary oxygen was offered if SpO2 levels were below 90%, during the exercise (Borghi-Silva et al., 2010; Moga, 2012). Prior to therapy, BP, HR, RR, SpO2, Borg dyspnea and Borg LL were measured and recorded. The sessions were carried out individually, and HR was monitored every 10 minutes of therapy using a FT1 model frequency meter (Polar® Electro Co.Ltda. Kempele, Oulu, Finland), SpO2 was measured by the wrist oximeter ONYX 9500 (Nonin®), and the subjective perception of the respiratory effort and LL effort by the Borg scale.

The data were analyzed using the SPSS 25.0 software (IBM, Amonk, USA). Normality of data distribution and homogeneity of variance were assessed using Shapiro-Wilk and Levene’s tests, respectively. A mixed twoway analysis of variance (ANOVA) was used to assess Group (CG vs ATG) and Time (Baseline vs Post) interaction. When interaction was significant, the Bonferroni’s post hoc pairwise comparison was performed, and the main effects results were disregarded. Partial eta squared (η²ₚ) was calculated to quantify the effect size for the ANOVA and interpreted as small (0.01), medium (0.06), or large (0.14). To assess the effect size for pairwise comparisons, Cohen’s d was calculated and classified as a small (0.2), medium (0.5), or large (0.8) effect, according to Cohen (1988). The significance level was established at 5% for all the analyses. Moreover, we analyzed the individual responses of each participant based on the available MCID values established in previous studies (Bisca et al., 2014; Regueiro et al., 2013).

Regarding the level of dyspnea during activities of daily living, as assessed by LCADL (Table 2), a significant Group*Time interaction was found for the Total Score (F1,15 = 41.4; p<0.0001, ηp² = 0.73) and the domains Physical (F_1,15 = 32.5; p<0.0001, ηp² = 0.68), Domestic (F_1,15 = 29.1; p<0.0001, ηp² = 0.66) and Personal (F_1,15 = 26.3; p<0.0001, ηp² = 0.64). No significant interaction or main effects were observed for the Leisure domain (F_1,15 = 3.8; p=0.08, ηp² = 0.19). Within group comparisons showed that ATG significantly improved total score (d=-1.48, p<0.0001), physical score (d=-1.64, p<0.0001), domestic score (d=-1.14, p<0.0001) and personal score (d =-1.12, p<0.0001) after the intervention. On the other hand, the CG did not present any significant changes (p>0.05). In the between group comparison, the ATG presented significantly lower values post intervention for the scores in the physical domain (d=-2.36, p<0.0001), personal domain (d=-1.08, p=0.04) and total score (d=-1.70, p=0.004), indicating an improvement in dyspnea in relation to the CG (Table 2).

Regarding pulmonary functional status, assessed by the PFSDQ-M (Table 2), a significant Group*Time interaction was found for Total Score (F_1,15 = 82.2; p<0.0001, ηp² = 0.81) and the domains Fatigue (F_1,15 = 62.5; p<0.0001, ηp² = 0.81), Dyspnea (F_1,15 = 39.2; p<0.0001, ηp² = 0.72) and Change in ADL (F_1,15 = 63.1; p<0.0001, ηp² = 0.81). Within group comparisons showed a significant difference in the total score (d=0.37, p<0.0001) and in all domains after the aquatic intervention, indicating an improvement in the impact of fatigue (d=-0.35, p<0.0001) and dyspnea (d=-0.5, p<0.0001) in activities of daily living, as well as in the perception of change in ADL (d=-0.28, p<0.04). On the other hand, the CG presented a significant increase, although with small effect on the total score (d=0.09, p=0.01), impact of fatigue (d=0.09, p=0.01) on ADL and the perception of change in ADL (d=0.004, p=0.004). There were no significant differences between groups for any of the variables concerning the PFSDQ-M (p > 0.05) (Table 2).

A significant Group*Time interaction was also found for the BODE index (F_1,15 = 7.93; p=0.01, ηp² = 0.35). After the training, the ATG presented a significant improvement in the BODE index (d= 0.52, p=0.001), which was not observed in the CG (d=0.05, p=0.69). No significant difference was found for the between group comparisons (p>0.05).

Finally, a significant Group*Time interaction was found for the MRC (F_1,15 = 21.3; p<0.0001, ηp² = 0.59, table 2). The ATG showed a significant improvement in dyspnea (d=-0.50, p<0.0001), assessed by the MRC, while no significant difference (d=-0.12, p=0.56) could be observed for the CG. Furthermore, the dyspnea observed in the ATG in the post-treatment period was significantly lower when compared to the dyspnea in the CG (d= 2.02, p=0.001, table 2).

Additionally, an individual analysis was carried out within each group to identify patients who achieved a clinically relevant improvement in the variables that presented MCID values reported in literature (Schünemann et al., 2003) (Figure 2). It was possible to identify clinically relevant improvement in 75% (6/8) of ATG patients in relation to the dyspnea domain of the PFSDQ-M; 50% (4/8) improved the fatigue and changes in ADL domains of the PFSDQ-M; and 87.5% improved the total LCADL score, indicating an improvement in dyspnea in ADL. In addition, patients classified as GOLD stages III and IV achieved the MCID across all the main outcomes (Figure 2).

Figure 2. Figure 2. Minimum clinically significant difference (MCID) delta chart. PFSDQ-M: Modified Pulmonary Functional Status and Dyspnea Questionnaire, SGRQ (Saint George Respiratory Questionnaire), LCADL (London Chest Activity Daily Living), BODE (B-body mass index; O- airflow obstruction; D-dyspnea; E-exercise capacity), and MRC (Medical Research Council). Figure 2. Minimum clinically significant difference (MCID) delta chart. PFSDQ-M: Modified Pulmonary Functional Status and Dyspnea Questionnaire, SGRQ (Saint George Respiratory Questionnaire), LCADL (London Chest Activity Daily Living), BODE (B-body mass index; O- airflow obstruction; D-dyspnea; E-exercise capacity), and MRC (Medical Research Council). Source: the authors.

Despite the promising results, the small size of the sample should be highlighted as a limitation of the study. Nevertheless, the analyses of the results were complemented by the effect size and the individual analysis in relation to the MDC, which are not size dependent. Furthermore, the inclusion of a group performing the same protocol on the land can also be seen as a limitation, as it would allow comparing the real advantages of training in the aquatic environment.