Pulmonary rehabilitation (PR) is a cost-effective intervention to manage people with chronic obstructive pulmonary disease (COPD), enhancing several health-related domains.1,2 A key aim of PR is to promote physical activity (PA),2 as PA can improve the COPD prognosis by decreasing the likelihood of exacerbations, hospitalizations and death.3-5

However, adopting and sustaining physically active behaviour after PR is highly challenging, and PR benefits are usually lost within 6 to 12 months after its completion.2,6 Effectiveness of post-PR maintenance programmes remains controversial, as study designs vary significantly, with systematic reviews reporting inconsistent findings.6-10 The optimal maintenance strategy capable of sustaining PR benefits is still unknown.

Engagement in regular long-lasting PA will most likely determine the maintenance of PR benefits.6,11 Nevertheless, to promote sustained physically active behaviour, interventions should be accessible and tailored to individual needs, goals and preferences, rather than assuming a “one size fits all” approach.10,11 Technology-based programmes, like those using social media, smartphone applications, websites or telemonitoring, have emerged as promising maintenance strategies to promote PA.10,12-14 However, their success still depends on patient digital literacy and access to technology, which remains a challenge for a considerable proportion of patients.15-17 Within the array of maintenance programme possibilities, both home-based and community-based approaches seem to be safe, feasible, and may provide different PA modalities that can be personalised to patient needs and preferences.18-21 Moreover, community-based maintenance programmes can also embrace several PA facilitators, such as: peer support, social connections, sense of inclusion within the community, supervision, enjoyment, accessibility and appropriate environmental conditions and infrastructures.19,22-27 In fact, lack of these elements in PA programmes has been suggested as reasons behind the nonadherence to a physically active lifestyle.5,19,24

Our study introduced an unprecedent programme by allowing people with COPD to select from various community-based PAs (e.g., senior or pool exercise classes) whichever best met their preferences and needs. The novelty of our study is, therefore, to unravel the role of these community resources to personalise PA and, ultimately, optimise outcomes in COPD.

We aimed to assess the efficacy and effectiveness of a six-month personalised community-based PA programme (PICk UP) on PA levels, six-months after completing PR, in people with COPD. As a secondary aim, effects on sedentary behaviour, functional exercise capacity and performance, peripheral muscle strength, balance, symptoms, emotional state, health-related quality of life, occurrence of exacerbations of COPD and healthcare usage were also evaluated. This study hypothesized that the PICk UP programme, targeting preferences and needs of participants with COPD, and based on an alliance among patients, health professionals and community, would be a solution to promote PA and prolong PR benefits.

Among the 107 participants assessed for eligibility (between February 2020 and July 2022), 61 fulfilled the inclusion criteria and were randomised, 32 to the EG and 29 to the CG. Number of participants randomised per site location were as follows: 13 in Aveiro, 27 in Estarreja, 13 in Oliveira do Bairro and 8 in Montemor-o-Velho. Characteristics of participants among sites were similar, except for differences in the BPAAT within the ITT set (P = 0.044) (Tables A1 and A2 from appendix). Overall, 49 (80 %) individuals with COPD completed the six-month follow-up (EG:25; CG:24; between January 2022 and January 2023). The per-protocol analysis included 33 participants (EG:14; CG:19) for the accelerometer outcomes and 43 (EG:20; CG:23) for the remaining outcomes. The CONSORT flowchart is shown in Fig. 1.

Fig. 1.

CONSORT flowchart of the PICk UP trial.

(0.82MB).

Legend: COPD: chronic obstructive pulmonary disease; ITT: intention-to-treat analysis; PA: physical activity; PR: pulmonary rehabilitation.

Baseline characteristics (post-PR) for the ITT (Table 1), per-protocol (Table A3 from appendix) sets were balanced among groups except for pack-years (greater in the CG) and body mass index (greater in the EG). Thus, these variables were included as covariates in ITT and per-protocol adjusted models. In the subset of the accelerometer data analysis, participants adherent to the proposed intervention presented higher PR attendance rates compared to non-adherent ones (Table A4 from appendix). Therefore, PR attendance rate was also included as a covariate in the adjusted model for the accelerometer data per-protocol analysis.

Of the 20 participants from the EG that enrolled in the community-based PAs (14[7–51] days between randomisation and PA integration), six (30 %) tried more than one type of community-based PAs. Specifically, 11 tried senior exercise classes, nine tried gym, six tried pool exercise classes and two tried Chi Kung. Following the initial experimental month, seven (35 %) enrolled in gym, six (30 %) in senior exercise classes, three (15 %) in pool exercise classes, two (10 %) in Chi Kung and two (10 %) opted for a mix of pool exercise classes and gym. On average, community-based PAs consisted of 36[28–48] sessions throughout the six-month period. Physiotherapists supervised 1[1–4.3] PA session and contacted participants via phone call 2[2–3] times, thus, spending approximately 6 h with each participant of the EG (4:30 h collecting data, 1 h supervising the integration in the community-based PA and 30 min scheduling assessments and providing support).

Except for sedentary behaviour, COPD assessment test, and the activities sub-scale of the SGRQ, PR improved significantly all the measured outcomes (Table A5 from appendix). Before enrolling in PR, participants presented low levels of PA, engaging only 15 min in MVPA/day and being mostly categorised as physically inactive (BPAAT of 0 [0–4] points) (Table A5 from appendix).

Physical activity and sedentary behaviour outcomes

ActiGraph valid days and wear-time data is provided in the appendix. Between group changes, at three- and six-month assessments, in the PA and sedentary behaviour outcomes, for the ITT and per-protocol analysis are presented in Table 2. The ITT and per-protocol analysis yielded similar PA and sedentary behaviour results. There was a significant time-group interaction for all PA but not for the sedentary behaviour outcomes. Between group changes in all PA outcomes were significantly different at the six-month follow-up (greater decline in the CG). According to the unadjusted BPAAT analysis, differences between group changes were also present at the three-month follow-up. Regarding within group changes, both in the ITT and per-protocol analyses all PA outcomes declined significantly in the CG from baseline to the 6-month, but no significant within-differences were observed in the EG (Fig. 2 and Table A6). In the ITT unadjusted analysis, in the CG, MVPA/day decreased from 15[8; 29] minutes (geometric mean [95 % confidence interval]) at the baseline to 6[3; 12] minutes at the six-month follow-up, and from 17[9; 33] to 15[8; 28] minutes in the EG. Similarly, in the CG, MVPA/week decreased from 106[53; 209] at the baseline to 37[19; 76] minutes at the six-month follow-up, and from 121[62; 234] to 101[5; 198] in the EG (Fig. 2 and Table A6). The decline observed on MVPA/day/week was 2.1 to 2.3 times greater in the CG compared to the EG (Table 2). For steps/day, the CG decreased from 4280[2980; 6147] at the baseline to 2952[2040; 4272] at the six-month follow-up, and from 5154[3651; 7275] to 4915[3452; 6997] in the EG (Fig. 2 and Table A6). The decline observed in steps/day was 1.4 times greater in the CG compared to the EG (Table 2). Unadjusted and adjusted means used in the LMM/GLMM models for PA and sedentary behaviour outcomes are presented in the appendix (Tables A9–12).

Table 2.

Between group changes for the physical activity and sedentary behaviour outcomes of the PICk UP trial.

		Per-protocol (EG: N = 14; CG: N = 19)			Intention-to-treat (EG: N = 32; CG: N = 29)
		Baseline-3months	Baseline-6months	P-value group*time	Baseline-3months	Baseline-6months	P-value group* time
MVPA/day	Unadjusted	1[0.5; 1.9]	2[1.1; 3.6]*	0.004	1.5[0.9; 2.6]	2.1[1.3; 3.5]*	0.002
	Adjusted	1[0.5; 1.9]	2[1.1; 3.5]*	0.005	1.5[0.9; 2.6]	2.1[1.3; 3.5]*	0.002
MVPA/week	Unadjusted	1.1[0.5; 2.3]	2.3[1.1; 4.5]*	0.009	1.6[0.9; 3]	2.3[1.3; 4.2]*	0.002
	Adjusted	1.1[0.5; 2.3]	2.2[1.1; 4.4]*	0.011	1.6[0.9; 3]	2.3[1.3; 4.2]*	0.003
Steps/day	Unadjusted	1.1[0.8; 1.5]	1.5[1.1; 2.1]*	0.002	1.2[0.9; 1.5]	1.4[1.1; 1.8]*	0.017
	Adjusted	1.1[0.8; 1.5]	1.5[1.1; 2.1]*	0.003	1.2[0.9; 1.5]	1.4[1.1; 1.8]*	0.018
BPAAT	Unadjusted	2.1[1.2; 3.8]*	2.5[1.3; 4.7]*	0.001	1.8[1.1; 3]*	2.3[1.3; 4.1]*	0.001
	Adjusted	1.8[0.9; 3.3]	2.6[1.2; 5.4]*	0.006	1.6[0.9; 2.8]	2.5[1.3; 4.8]*	0.004
SED/day, min	Unadjusted	−11.5[−78; 55]	25[−36.7; 86.7]	0.367	−17[−72.4; 38.4]	7.4[−46.9; 61.7]	0.582
	Adjusted	−11.9[−78.4; 54.6]	25.3[−36.4; 87]	0.354	−15.8[−71.2; 39.5]	8.7[−45.5; 62.9]	0.584
SED/week, min	Unadjusted	−80.4[−546; 385]	175.2[−257; 607]	0.367	−118.9[−507; 269]	51.8[−328; 432]	0.582
	Adjusted	−83.2[−549; 382]	177.1[−255; 609]	0.354	−111[−498; 276]	61[−318; 440]	0.584

The BPAAT per-protocol data analysis included 43 participants (EG=20; CG=23). MVPA/day, MVPA/week, steps/day and BPAAT are presented as mean ratio [95 %CI]; mean ratios were considered significant if the 95 %CI excluded the value 1; ratios were calculated as: (CG baseline ÷ CG 3 months) ÷ (EG baseline ÷ EG 3 months); and (CG baseline ÷ CG 6 months) ÷ (EG baseline ÷ EG 6 months). SED/day and SED/week are presented as mean difference [95 %CI]; mean differences were considered significant if the 95 %CI excluded the value zero; differences were calculated as: (CG baseline – CG 3 months) – (EG baseline – EG 3 months); and (CG baseline – CG 6 months) – (EG baseline – EG 6 months).

⁎

P-value for the multiple comparison <0.05 (Sidak adjustment). Adjusted models included pack-years and body mass-index as factors. Additionally, adjusted models for the per-protocol analysis using accelerometer outcomes included pulmonary rehabilitation attendance rate as factor. A sensitivity analysis was conducted for the BPAAT, adding site location as a covariate in the intention-to-treat adjusted model, with no change observed in the results.

Legend: BPAAT: brief physical activity assessment tool; CG: control group; EG: experimental group; MVPA: time spent in moderate to vigorous physical activity; SED: time spent in sedentary behaviour per day.

Fig. 2.

Unadjusted geometric means and 95 % confidence intervals used in the linear-mixed models of the intention-to-treat analysis in A) time spent in moderate to vigorous physical activity per day; B) time spent in moderate to vigorous physical activity per week; C) steps/day; and D) Brief physical activity assessment tool of the PICk UP trial.

(0.19MB).

Legend: BPAAT: Brief physical activity assessment tool; CG: control group; EG: experimental group; MVPA: time spent in moderate to vigorous physical activities.

This study showed that, six-months after completing PR, the PICk UP intervention was effective and efficacious in preventing PA decline in people with COPD. Sit-to-stand performance was also preserved in the PICk UP group. PR benefits on the remaining health-related outcome measures were sustained in both groups.

The PICk UP intervention was successful in maintaining the PA levels after PR independently of the outcome measure (objective/subjective) or analysis (ITT/per-protocol, adjusted/unadjusted) used. Our study differs from previous post-PR maintenance studies6 by allowing participants to choose and enrol in diverse PA modalities, aligning with the principle that individual preferences/needs vary and “one size does not fit all”.10,11,23,26 A good attendance rate was registered (67 %), surpassing attendance previously reported in PA maintenance studies14,18,36,39 and the typical 50 % rate seen in long-term treatments for chronic diseases.37 The social connections, supervision and peer-support,19,22,23,26 health professional support, home proximity and infrastructure convenience are well-known PA facilitators23,25,26 and might have contributed to our PA results.

Over twenty maintenance studies using PA-based interventions have been published,6-10 yet only a few reported PA as an outcome.14,20,39,40 Integrating PA as an endpoint in maintenance trials is crucial given its prognostic value in COPD.34 The only study measuring PA objectively14 observed similar findings to our study, i.e., in the EG the PA levels were preserved for the whole six-month period and in the CG a significant PA decline was present at six- but not at three-months.14 In fact, the decline noticed in steps/day in our CG exceeded the minimal clinically important difference established for the likelihood of medical events (350–1110 steps/day) in COPD.41 Therefore, our findings highlight the need for policy makers to support effective PA maintenance strategies.

The PICk UP and the PR programmes used in this study targeted PA but did not focus on reducing time spent in sedentary behaviour, which might explain the absence of improvements in sedentary lifestyle after both interventions.42,43 How to successfully reduce sedentary behaviour in COPD is still an enigma,44 however, it is worth noting, that our population spent an average of 400 min in sedentary behaviour, below the critical 510-min threshold tied to COPD mortality.45

Overall, six-months after PR completion, its benefits were still present in both groups, except in sit-to-stand performance, which declined in the CG. Our findings suggest that PA is likely to be the first benefit to decline after PR. A reduction in PA while exercise capacity (6MWT) remained stable has been previously observed in COPD.46 Several reviews have reported inconsistent findings regarding the effect of maintenance strategies on the 6MWT,6,7,9,10 with some reporting positive significant results at six-months,7,9 while others highlight the existent uncertainty in this field.6,10 Unlike the 6MWT, the 1-minSTS was not included as an outcome measure in previous maintenance reviews.6,7,9,10 Although an association between these instruments has been shown,47-49 the fact that daily mobility primarily involves walking-based activities whereas sit-to-stand or squat activities tend to occur less frequently throughout the day and are typically not repeated (unless for exercise purposes), may contribute to a faster decline measured by the 1-minSTS.

Health-related quality of life benefits were sustained independently of the PICk UP intervention. These findings are aligned with previous maintenance strategies reviews.6,7,9,10 In accordance with previous conclusions,6 the PICk UP did not provide any extra advantage in terms of health care utilization or prevention of exacerbations. Both groups experienced few exacerbations and used the healthcare system occasionally. Two factors could explain this: the six-month follow-up period may have been insufficient to assess these outcomes; most of our sample consisted of non-exacerbators (94 %), and it is known that exacerbation history is the most important predictor for future occurrences.50 Moreover, the decrease in PA in the CG, even with few exacerbations occurring during the follow-up period, reinforces the inconsistency noted in the connection between PA decline and exacerbation frequency.51

This study has several strengths. Firstly, it introduced a novel maintenance strategy that was personalised according to participants’ preferences and needs, embedded in their local context, and settled in an intersectoral partnership. To the best of the authors knowledge, this was the first study where participants could select and engage in various PA options, according to their preferences. Secondly, aligning with other PA interventions,21 the PICk UP intervention was safe, with only minor adverse events being reported. An additional strength of this study is its multicentre design with a rigorous methodology, including blinded assessors, and closely reflecting real-world conditions. Furthermore, the agreement between the ITT, per-protocol, adjusted and unadjusted analysis enhances confidence in our results.

Nevertheless, some limitations should be acknowledged. Firstly, effects of the PICk UP were only assessed over a medium-term period of six-months. A longer follow-up, ideally at least one year, is necessary to determine if the PICk UP programme was able to preserve PA levels in the EG, and whether other PR benefits were still maintained in both groups, in the long-term. Secondly, the possibility for participants of integrating different PA modalities added heterogeneity. Nevertheless, all PA modalities were properly validated and were mostly categorized as of moderate intensity32 (appendix). In addition, this study used the community real-world resources, therefore, the availability of PA options varied across city councils, resulting in some participants having a broader range of options compared to others. A sensitivity analysis based on type of PA chosen, city council or participants’ severity would have been relevant, however, the sample size was not large enough to perform it. The sample size may also have been insufficient to detect changes in the secondary outcomes (no power calculations were performed), therefore, caution is advised when interpreting these results. It is worth noting that the organisation, enrolment capacity, predisposition to accept and adherence to our study was affected by the COVID-19 pandemic. Although changes in organisation (e.g., physical distance among people, different entrances and exits) were similar across all site locations and adherence to the PICk UP intervention was considered good, the fear of COVID-19 may have precluded some participants from integrating community-based PAs. Another shortcoming of this study is related to the cutoff used to estimate MVPA,52 developed with healthy individuals and never validated in people with COPD. Our findings may not be generalizable to people with more severe COPD (GOLD grades 4 and E, and oxygen-dependent), or females, since these groups were underrepresented in our sample. Similarly, results might not apply to those recruited from inpatient or hospital-based settings, or failing to adhere to PR as only participants who completed an outpatient community-based PR programme were recruited. Finally, due to the nature of the intervention, blinding of participants was not possible. Currently, the Portuguese Health System does not cover the expenses related to community-based PAs. In this project, participants had free access to the entire PICk UP intervention, so the potential impact of financial constraints was not considered. Future studies performing a cost-analysis of the PICk UP intervention and assessing the efficacy/ effectiveness of this programme based on the type of PA are required to support its broader implementation and influence policy makers.

The personalised community-based PA programme implemented was effective and efficacious in preventing a PA decline over a six-month period, following PR in people with COPD. Although there seems to have been a decline in sit-to-stand performance, this was counteracted with the PICk UP intervention. The remaining PR benefits seem to have lasted after six-months of PR, regardless of the maintenance strategy employed. Our findings highlight the importance of policymakers supporting sustainable PA maintenance strategies which bring the healthcare systems closer to patients’ communities and foster multisectoral partnerships to prevent the PA reduction after PR in people with COPD.

This work was funded by the project “CENTR(AR): pulmões em andamento” by Programa de Parcerias para o Impacto, Portugal Inovação Social through Programa Operacional Inclusão Social e Emprego (POISE-03-4639-FSE-000597); by Fundação para a Ciência e a Tecnologia (SFRH/BD/148738/2019 and COVID/BD/153476/2023) by Fundo Social Europeu through Programa Operacional Regional Centro, and by Programa Operacional Competitividade e Internacionalização (COMPETE 2020 - POCI-01-0145-FEDER-007628; UIDB/04501/2020 and UIDP/04501/2020).