Methods
This quasi-experimental study included a sample of consecutively recruited patients enrolled in a four-week, in-patient PR-program at LHL-Clinics, Glittre, Norway. This site were chosen for practical reasons (i.e. likelihood of reaching recruitment goals). The recruitment period was January to June 2015. Forty-five patients were included in the final analysis (Fig. 1). Inclusion criteria were a diagnosis of COPD, cognitive ability to provide informed written consent, and ability to understand and complete questionnaires. Exclusion criteria were ongoing exacerbation of COPD, inability to exercise, and co-morbidities limiting the patient’s physical performance more than the COPD alone (e.g., neurological disorder, severe angina). To minimize the ceiling effect, patients considered likely to achieve the maximum score of 12 on the SPPB were excluded prior to baseline testing. Exclusion on this ground, was determined by which test protocol that was used on the progressive treadmill test the patients was tested with in the beginning of PR at LHL-clinics Glittre. All patients tested with protocol 4 were excluded, because the initial walking speed (4.8 km/h) suggested that the patients would have scored 12 points (maximum) on the SPPB. Patients who obtained the maximum SPPB score at baseline and patients who did not exercise for five consecutive days or more prior to the post-test were excluded.
Outcome measures
Background data, including GOLD stages, were obtained from patients and medical records at baseline. All tests were administered by qualified health-care professionals. Exercise habits during PR were reported in a training log.
Primary outcome measure
The SPPB is comprised of three subtests: a hierarchical standing balance test (feet placed side-by-side, semi-tandem, and tandem, for 10 s each); a 4MGS test (timed four-meter walk test at habitual gait speed); and a 5STS test (timed five-repetition chair stands test performed as fast as possible). For scoring, see Background. The SPPB is reliable [11] and valid [7]. A one-point change is considered clinically meaningful [7]. Standardized instructions were followed, but in light of a possible learning effect [12], the SPPB was performed twice at baseline and twice at the end of PR. The patients rested for a minimum of 5 min before taking the test again. The best score was used in the data analysis.
Secondary outcome measures
The secondary outcome measures were administered at baseline and at the end of PR except for pulmonary function, which was measured at baseline only.
Exercise capacity
Exercise capacity was measured as distance walked (6MWD) during the six-minute walk test (6MWT). It is frequently used for measuring response to therapeutic interventions in COPD [13]. The 6MWT was performed according to the standardized protocol [6].
Dyspnea
The impact of dyspnea on daily activities was measured with the modified Medical Research Council dyspnea scale (referred to as the mMRC) [14]. The mMRC consists of five grades (range 0–4), where a higher number represents a higher degree of breathlessness on daily activities [15]. The mMRC is valid for assessing dyspnea [16].
Disease-specific quality of life
Disease-specific quality of life was measured using the COPD Assessment Test (CAT) [17]. The CAT consists of eight items designed to assess the impact of COPD on health status (range 0–40) and is a valid and reliable measure [17].
Pulmonary function
Pulmonary function (i.e. predicted forced expiratory volume in one second (FEV1%) was measured by spirometry (MasterScreen PTF, Jaeger GmbH, Würtzburg) in accordance with ATS/ERS guidelines [18].
Pulmonary rehabilitation
The PR-program consisted of individual and group-based strength and endurance training, education, and individual sessions with a multi-professional health-care team.
Statistics
Data analyses were performed using IBM SPSS Statistics version 22.0 for Windows (SPSS Inc., Chicago, IL, USA). Descriptive data were presented as mean and SD, or number and percentage out of total sample. Paired Student’s t test were used to evaluate pre- and post-test changes at group level. Results were confirmed with non-parametric statistics when the criterion of a normal distribution was not met. Analyses were performed per-protocol and confirmed with intention-to-treat. Cohen’s d was calculated to estimate the effect size of changes in SPPB scores with 0.2 indicating a small effect size, 0.5 a medium effect size, and 0.8 a large effect size [19]. Effect size was calculated using an online calculator [20, 21]. Sample size calculation for the main outcome measure was based on 80% power to detect a one-point change in SPPB score (SD 2.5), with an alpha level of 0.05. The estimated sample size was forty. Relationships between the SPPB and 6MWD, mMRC, CAT, and FEV1% were assessed with Pearson’s correlation coefficient. To explore relationships between changes in SPPB scores and changes in 6MWD and mMRC scores, two separate multiple regression analyses were performed. The significance level was set at p < 0.05.