The sample will be composed by at least 20 patients with moderate-severe COPD from each institution involved (Catholic University of Leuven, Belgium and State University of Londrina, Brazil), adding up to 40 patients (see sample size calculation below). However, the sample size could be increased if necessary to achieve intermediate calculation of the sample power performed during the study. Both groups should be similar concerning anthropometric characteristics, gender distribution and severity feature of the disease (especially FEV1). Therefore, an eventual increase in the final number of patients may also be necessary in case this similarity is not achieved in the final data collection of the planned sample.
Data collection will be performed in the Laboratory of Research in Respiratory Physiotherapy from the State University of Londrina (Londrina, Brazil) and in the University Hospital Gasthuisberg from the Catholic University of Leuven (Leuven, Belgium).
The study has a longitudinal design. Two groups of patients, one group in Londrina and another group in Leuven, will be simultaneously and objectively evaluated concerning their physical activity in daily life during one week in two different moments: in January 2013 (winter in Leuven and summer in Londrina), and then the same two groups of patients will be once again simultaneously re-evaluated during another week in July 2013 (summer in Leuven and winter in Londrina). Changes in the level of physical activity in daily life between summer and winter in the two groups will then be compared. Monitoring of physical activity in daily life: two activity monitors (the SenseWear armband and the Fitbit) will be simultaneously worn during 7 consecutive days. The SenseWear armband (BodyMedia, Pittsburg, PA, USA) is a multisensor composed by a biaxial accelerometer and physiologic sensors. It is a small (8.8 x 5.6 x 2 cm) and lightweight (82g) monitor that is worn at the upper right arm and was already validated for energy expenditure estimation in patients with COPD \[12,13\]. The Fitbit (Fitbit, Inc., San Francisco, CA, USA) is a recently released triaxial accelerometer that is smaller (3.6 x 2.9 x 1 cm) and more lightweight (8g) than the SenseWear. According to the manufacturer, the Fitbit can be used in different positions, including at the waist. This device provides the number of steps, the distance walked and the energy expenditure in calories, among other variables. In the 7 days of assessment, the patient will remain with the device during 24 hours/day. The patients will be strictly instructed to avoid changing their daily physical activity habits during the monitoring period.
To characterize the sample, all patients will be submitted to spirometry, 6 minute walking test (6MWT) and quality of life assessment (Chronic Respiratory Disease Questionnaire - CRQ), also described in detail in the subsection "Methods".
For each patient, further information will be also registered: types of drugs used by the patient, their dosage and how often these drugs are used in the last month before inclusion in the study; characterization of the history of acute exacerbations occurred in the last year (if it is the case); presence of respiratory insufficiency on admission or during the study period (PaO2 \< 60 mmHg and/or SpO2 \< 90% with or without PaCo2 \> 50 mmHg, breathing on room air); need for long-term oxygen supplementation previously or during the study period; practice or not of regular physical activity in the last year.
LOGISTICS INVOLVED IN DATA COLECTION IN BOTH CENTERS
In the first period of data collection (January 2013), the Brazilian coordinator of the project (Prof. Fabio Pitta) will travel to Leuven to participate in data collection and organization on site. The Brazilian project coordinator will arrive in Leuven two weeks before the data collection, and in conjunction with the Belgian coordinator of the project (Prof. Thierry Troosters) and the European researchers will prepare all necessary logistic actions to ensure the adequate performance of the study. After the preparation period, data collection itself will be performed in three weeks, with 15 patients being assessed in the first week and 14 in the third week. Spirometry, 6MWT and CRQ will be performed on Friday, and data collection of physical activity will be performed from Saturday to next week's Saturday. The 1-week interval between the first and second weeks of assessment will be used to download the data and discuss eventual issues raised in the first week of assessment, and thus this justifies the need of 15 activity monitor devices (SenseWear and Fitbit) in each center adding up to 60 necessary devices (as data collection will simultaneously occur in both centers).
Previously, the Brazilian team involved in the project will be prepared by the Brazilian coordinator with the same instruction, since the first period of data collection in Londrina will occur simultaneously with Leuven. Therefore, the first period of data collection in Londrina will be performed by the Brazilian researchers involved in the study, coordinated from Leuven by Prof. Fabio Pitta. The research team will make sure that the Brazilian researchers are prepared and trained to perform all assessments, including the objective quantification of physical activity in daily life, as well as the identical application of the 6MWT and CRQ in both centers.
In the second phase of data collection (July 2013), the European and the Brazilian researchers will be responsible for the assessments with the same protocol trained and performed in the first phase of data collection. Therefore, Prof. Fabio Pitta will coordinate the data collection in Londrina while Prof. Thierry Troosters will coordinate the data collection in Leuven to ensure the reproducibility of all the methods in the second data collection moment..
After the second phase of data collection, further project activities will have the coordination centralized in Londrina, with the Brazilian coordinator of the project.
STATISTICAL ANALYSIS AND SAMPLE SIZE CALCULATION
Statistical analysis will be performed using the SPSS v. 17 Statistical Package (SPSS, Inc., Chicago IL, EUA) and the GraphPad Prism 5.0 (GraphPad Software Inc., San Diego, CA, EUA). The Shapiro-Wilk test will be initially used to analyze the normality in data distribution. In case of variables normally distributed, data will be described as mean ± standard deviation, and those variables with non-normal distribution will be described as median \[25%-75% interquartile range\]. Summer-winter changes in the same group of patients will be analyzed by the paired t test (parametric analysis) or the Wilcoxon test (non-parametric analysis). Summer-winter differences between the two groups of patients (Londrina and Leuven) will be analyzed by the unpaired t test (parametric analysis) or by the Mann-Whitney (non-parametric analysis). Variables of spirometry, 6MWT and CRQ of both groups will be analyzed by the unpaired t test in the two assessment moments. Correlations will be analyzed by the Pearson or Spearman coefficient, depending on the normality on data distribution. For all analysis, statistical significance will be set at p less than 0.05.
The sample size of the study was calculated according to the assumption of a mean (and standard deviation) difference of counts/day between two groups of patients with COPD, being one in Europe and one in Brazil. The first group was assessed in the summer and winter, as found in the study by Sewell et al. performed in England \[10\]. This study presented a mean difference of 5656 counts (8857 in summer and 3201 in winter), with a standard deviation of 7497 and 2629 counts, respectively. The second group was based on our hypothesis that there will be no (or little) change in physical activity in daily life due to seasonal variation in Londrina (Brasil) based on the study by Probst et al. (Respiratory Care 2011; 56\[11\]: 1799-87), taking into account a similar standard deviation of 5063 counts/day. With a power of 90%, alfa coefficient of 0.05, at least 12 patients in each group will be necessary to detect significant difference in the level of physical activity in daily life between summer and winter within the groups. Taking into account an eventual difficult setting with drop-out rate of 40% due to difficulties in the data collection protocol involving the motion sensors, a minimum of 20 patients will be included to compose either group (Londrina and Leuven), adding up to 40 patients in total.
