From the Directive Committee of the SEPAR COPD IRP, members of the COPD research area were contacted in order to propose the study to them. An initial meeting took place in April 2009, where the objectives of the project were specified and the scientific committee for the project was selected. During the following months, the scientific committee worked on the design of an initial draft of the study. Later, pulmonologists with research experience in COPD or those who had shown interest in participating were invited to a meeting before the start of the recruitment period in order to evaluate their participation in the project and to reach a consensus on the study protocol. This meeting took place in November 2009. At these meetings, it was decided that there would be a main researcher for every participating unit. After an initial contact with 52 hospitals, 36 centers from 13 Spanish regions decided to participate in the study (Fig. 1). Later, once the recruitment of cases had been started, a third meeting was held in April 2010 to evaluate the project start-up and to assess possible problems after patient recruitment.

Figure 1.

Number of participating centers.

(0.24MB).

Due to the multidimensional nature of the project, the BODE index8 was established as the main variable of the study and basis for calculating the sample size. Contemplating an alpha error of 5% and a power of 80%, for the comparison of the progression on the BODE index between 2 samples of COPD participants (severe versus very severe COPD, or mild versus moderate COPD), since in the reference group the BODE is 0 by definition, and considering a mean BODE of 3 and standard deviation of 2,10 at least 300 participants are required per group in order to detect an increase of 0.6 units or higher in the BODE index between groups. Due to the number of participating centers, the reference was determined to be 50 cases and 10 controls per center.

The inclusion and exclusion criteria for each study group are defined in Table 2. The patients were recruited consecutively from the outpatient consultations during scheduled appointments. The patients were in stable clinical situation and had been free from exacerbations for at least 8 weeks. It was decided that the selected cases and controls should be paired for age and sex.

The variables collected during the follow-up were agreed upon by the researchers. Due to the fact that the objective was to carry out a multidimensional evaluation of COPD, the number of variables had to be high. During the development of the method, some variables were excluded to make the protocol simpler and more viable, but the researchers were given the opportunity to register other variables in case a group of researchers wished to carry out a sub-study including these data. After several consensus sessions, the selected variables were grouped according to the following aspects: socioeconomic data, comorbidities, smoking, clinical data, treatments, and complementary studies.

The socioeconomic data that were registered included: marital state, level of education, occupation, employment, and social class. Employment status was collected in an open field. In addition to the comorbidities included in the Charlson index,14 an extensive questionnaire was established for the systematic compilation of other comorbidities. In the case of depression and anxiety, these were evaluated with the Hospital Anxiety and Depression Scale (HADS).15

As for smoking, patients are registered as either active smoker or ex-smoker, as well as data for age of initiation and calculated accumulated consumption in pack-years. In addition, at each visit, co-oximetry is done in the exhaled air to verify if there have been any changes in tobacco habit. The co-oximetries are all done with the same device (PiCO Smokerlyzer, Bedfont Scientific, Kent, United Kingdom) acquired specifically for this project.

The clinical data obtained cover the following aspects: the date of diagnosis, the presence of respiratory symptoms (dyspnea, cough, expectoration, and color of sputum), symptoms related with sleep (frequent snoring, insomnia), and cardiologic symptoms (presence of malleolar and orthopneic edemas). Dyspnea is quantified by means of the modified Medical Research Council (MRC) scale.16 The impact of the symptoms is assessed by means of the Clinical COPD Questionnaire (CCQ).17 In addition, activities of daily living are evaluated with the London Chest Activity of Daily Living (LCADL) scale.18 The variables for exacerbations in the previous year are registered as part of the protocol (Table 3), identifying if the information were revealed by the patient or instead obtained from the patient files. Quality of life is evaluated by mean of the COPD Assessment Test (CAT) questionnaire.19

As for treatment, the drugs compiled are shown in Table 4. It also registers anti-flu and anti-pneumonia vaccinations, home therapies such as oxygen therapy and home mechanical ventilation, as well as any type of respiratory rehabilitation program. The patients of the COPD cohort are treated according to the current national and international recommendations for the treatment of COPD, as decided by the head researcher. All the included subjects who are active smokers received minimal anti-smoking intervention, trying to advance in the process of quitting according to the Prochaska and DiClemente method20 and offering complete anti-smoking treatment according to national guidelines.21

The complementary studies that were ordered are listed in Table 5. All the analyses are done in the reference laboratories of each center. Blood samples were taken for hemogram and general biochemistry (with special interest in cardiovascular risk parameters, total IgE), and the first analysis included determination of α1-antitripsin levels. Finally, 35-ml EDTA tubes were extracted to centrifuge (3000rpm for 10min at room temperature) and extract serum, which was stored at −80°C for later biomarker studies. For storage, at least 6 aliquots of 500μl were separated before freezing. Each center was provided with labels identifying each aliquot by center code, case number, and aliquot number.

For the nutritional assessment, the body mass index was calculated and a body bioimpedance study was done. All the bioimpedance studies were carried out with the same device, an Omron BF500 (Omron Corporation, Kyoto, Japan) acquired specifically for this project. With this technique, body and visceral fat mass can be estimated. Lung function tests were done with the equipment available in each participating hospital, registering the absolute values of each determination to be able to apply the same theoretical values for the analysis of the results. The peripheral muscle strength study was done with a GRIP-A hand dynamometer (Takei Scientific Instruments, Niigata, Japan), acquired specifically for this project. The 6-min walk test22 was done in accordance with international recommendations.23 If the patient did not know the test, a second test was done (separated by at least 30min of rest) to make amends for the possible training effect.

At each annual and semestral office visit, the vital state of the subject will be registered and, in case of death, the cause of death will also be reported (coded as: respiratory, cardiovascular, cerebrovascular, lung neoplasm, other neoplasms, sudden death, other or unknown) as well as the exacerbation variables (Table 3). Specifying the cause of death in each case will be the responsibility of the local researcher, taking into account the information provided by family members and the patient files.

Notebooks will be provided to collect the following exacerbation variables: the time of the exacerbation in days, the main location of care (home, outpatient, hospital emergency department, hospitalization or ICU), the Anthonisen criteria24 and the treatment received.

For the control cohort, the same data were collected as in the case cohort, but the respiratory function tests are simplified, requiring at least spirometry and an oxygen saturation measurement by pulse-oximetry.

The case recruitment period lasted 26 months, from January 15, 2010 to March 31, 2012. After the end of the inclusion period, there were a total of 991 cases and 132 controls. The follow-up of the subjects includes annual office visits, where all the information will be collected for each case. In addition, every 6 months a telephone call is programmed to compile data about the number of exacerbations and to verify the vital status of the subject. Finally, the patients were instructed that they could come to the consultation any time if they were having an exacerbation with the aim to collect data about the exacerbations. The project includes a scientific notebook (SNB) to collect data from the annual scheduled visits and the biannual phone calls.

A company was selected with experience in the design of web medical applications (Claveweb, Seville, Spain) to create the project webpage (www.chain.org.es). The web page offers information about the project and provides the researchers and scientific committee access to the data. The web has a hierarchical structure, so that the scientific committee has access to all the information, but the local researchers can only access the data from their respective centers. The web application also has an option to download data for the scientific committee in order to be able to download data and do quality controls.

A data manager was hired to communicate with the researchers and to resolve incidences, maintain the database, and control the quality of the data. The quality control entails confirming the veracity of the registered data. The quality controls of the procedures carried out at each center are done by each participating hospital, following the internal regulations at their centers. The project manager reviews the database weekly to identify lost, inconsistent or extreme values. In these cases, the project manager contacts the researcher to confirm or correct the final value. In addition, the project manager periodically sends the researchers an e-mail with the recruitment information of cases and the state of the project.

The present protocol was approved by the Healthcare Research Ethics Committees of the participating centers, and the principles of the Declaration of Helsinki are observed for research with human subjects. All the participants were informed about the nature of the study and its objectives, and they signed their consent to participate, while being able to withdraw this consent at any time.

All the procedures and variables related with this paper are compiled in a SNB. There is an SNB for each included subject, regardless of the group to which he/she belonged, and in said SNB all the variables related with the study are annotated.

During the study, no personal data of the participating patients are taken which would enable the patient to be identified. These data were made anonymous in the web application. The data obtained will be maintained under strict confidentiality (Law 15/1999, protection of personal data) and only the researcher responsible for the project in each center will have access to them.

The biological samples related with the study, as well as the SNB, are coded in order to guarantee the confidentiality of the sample and the data. During the study period, the SNB are stored under lock and key, which is kept exclusively by the researcher in charge of the project at the center and, after their transcription to the database, will be destroyed.

The results for each variable will be presented by means, standard deviation, minimum, median, maximum, and number of cases validated in the case of the continuous variables, and using the case count for each category and the frequency related to the total number of responses in of the categorical variables will be done. For the statistical comparisons between groups, in the continuous variables an initial ANOVA will be done and then later a bilateral Student's t test or a χ2 for the categorical variables. The Spearman's test will be used to determine correlation between non-parametric variables, and Pearson's for the parametrical ones. In all the statistical tests done, a level of statistical significance of 0.05 will be used.