Calle Rubio et al. Respir Res (2021) 22:36 https://doi.org/10.1186/s12931-021-01633-y RESEARCH Open Access COPD Clinical Control: predictors and long‑term follow‑up of the CHAIN cohort Myriam Calle Rubio1,2, Juan Luis Rodriguez Hermosa1,2* , Juan P. de Torres3, José María Marín4, Cristina Martínez‑González5, Antonia Fuster6, Borja G. Cosío7, Germán Peces‑Barba8, Ingrid Solanes9, Nuria Feu‑Collado10, Jose Luis Lopez‑Campos11 and Ciro Casanova12 on behalf of the CHAIN Study Investigators Abstract Background: Control in COPD is a dynamic concept that can reflect changes in patients’ clinical status that may have prognostic implications, but there is no information about changes in control status and its long-term consequences. Methods: We classified 798 patients with COPD from the CHAIN cohort as controlled/uncontrolled at baseline and over 5 years. We describe the changes in control status in patients over long-term follow-up and analyze the factors that were associated with longitudinal control patterns and related survival using the Cox hazard analysis. Results: 134 patients (16.8%) were considered persistently controlled, 248 (31.1%) persistently uncontrolled and 416 (52.1%) changed control status during follow-up. The variables significantly associated with persistent control were not requiring triple therapy at baseline and having a better quality of life. Annual changes in outcomes (health status, psychological status, airflow limitation) did not differ in patients, regardless of clinical control status. All-cause mortal‑ ity was lower in persistently controlled patients (5.5% versus 19.1%, p = 0.001). The hazard ratio for all-cause mortality was 2.274 (95% CI 1.394–3.708; p = 0.001). Regarding pharmacological treatment, triple inhaled therapy was the most common option in persistently uncontrolled patients (72.2%). Patients with persistent disease control more frequently used bronchodilators for monotherapy (53%) at recruitment, although by the end of the follow-up period, 20% had scaled up their treatment, with triple therapy being the most frequent therapeutic pattern. Conclusions: The evaluation of COPD control status provides relevant prognostic information on survival. There is important variability in clinical control status and only a small proportion of the patients had persistently good control. Changes in the treatment pattern may be relevant in the longitudinal pattern of COPD clinical control. Further studies in other populations should validate our results. Trial registration: Clinical Trials.gov: identifier NCT01122758. Keywords: Chronic obstructive pulmonary disease, Control, Management Background multidimensional component and the concept of phe- Over the last decade, we have seen new evidence that notype, which has meant a step forward on the road to has led to a new vision of chronic obstructive pul- personalized medicine and individualization of treatment monary disease (COPD) with the recognition of the [1–3]. Clinical practice guidelines in COPD establish the reduction of symptoms and minimization of risk as the *Correspondence: jlrhermosa@yahoo.es 1 Pulmonology Department, Hospital Clínico San Carlos, C/ Martin Lagos main therapeutic objectives [4, 5]. These objectives make S/N, 28040 Madrid, Spain it necessary to adapt actions to the changes experienced Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​ mmons​.org/publi​cdoma​in/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Calle Rubio et al. Respir Res (2021) 22:36 Page 2 of 12 by patients throughout their evolution, considering ther- Tenerife; Spain; IRB No. 258/2009). COPD was defined apeutic success to mean achieving disease control. as a smoking history of at least 10 pack-years and an The concept of COPD control is a new dimension FEV1/FVC ratio less than 0.70 after inhaling 400 mg of that is proposed as a tool to help make therapeutic deci- albuterol. Patients were stable for at least 6 weeks and sions and to modulate treatment [6, 7]. According to received optimal medical therapy. Exclusion criteria were this proposal, control is defined as a state of low clinical uncontrolled comorbidities such as malignancy or other impact and an absence of exacerbations maintained over confounding diseases that could interfere with the study. time. The prespecified criteria for clinical control were The follow-up of the subjects included annual office vis- described by Soler Cataluña [6] and have subsequently its and a telephone call was scheduled every 6 months to been evaluated in several studies [8–10]. compile data about the number of exacerbations, clinical Control in COPD is a dynamic concept that can reflect impact (health-related quality of life, subjective percep- changes in patients’ clinical status that may have prog- tion) and to verify the subject’s vital status. COPD treat- nostic implications. Some studies have observed a poten- ment followed national [5] and international guidelines. tial predictive value for poor outcomes and previous Data analyzed in the present study was obtained from studies have shown that improvement in control status the recruitment date through September 2018. Data was in the short term was associated with better outcomes, anonymized with hierarchical access control in order to improvement in health status, less frequent exacerba- guarantee that information was secure. tions [11] and a longer delay until hospitalization [8]. However, this new concept requires validation in terms of Clinical and physiological measurements its ability to predict outcomes and to provide additional Trained staff obtained information on age, sex, body clinical management insight. Given the limited infor- mass index (BMI) and smoking status at baseline and mation about the changes in clinical control in patients subsequent visits. Comorbidities were scored using the with COPD and the relationship with outcomes in those Charlson index [13]. Pulmonary function tests were per- patients, we assessed clinical control at baseline and lon- formed according to international criteria [14, 15]. Dysp- gitudinally (annually over 5 years) in patients participat- nea was evaluated using the modified Medical Research ing in the CHAIN (COPD History Assessment in Spain) Council (mMRC) scale [16]. To evaluate health-related cohort, aiming to use CHAIN data to explore the changes quality of life, the Spanish validated version of the COPD and consequences of clinical control in a large cohort of Assessment Test was used, which was self-administered patients with COPD. by each patient under the supervision of the interviewer We hypothesized that worse persistent control would [17]. Anxiety and depression were evaluated using the relate to worse clinical outcomes. We followed longitu- Hospital Anxiety and Depression Scale (HAD) question- dinal changes in physiological outcomes and patient- naire [18]. Exacerbations were defined as a worsening of reported outcomes for health status, dyspnea and respiratory symptoms (dyspnea, cough or sputum) that psychological status over 5 years in patients with COPD. required the use of antibiotics, systemic corticosteroids, The objectives of the present study were as follows: (1) to or both, or symptoms that necessitated an emergency evaluate the degree of control in patients with COPD; (2) room visit or hospital admission. All-cause mortality was to provide information on the longitudinal evolution of recorded using information obtained from the family and clinical control and to determine the factors associated then confirmed by reviewing the medical record. with worse control; (3) to validate the concept of control as a predictor of the risk of poor outcomes. Clinical control status assessment Control status was evaluated based on low clinical impact Methods and stability, according to clinical criteria. A patient was The CHAIN methodology has been extensively reported considered controlled when disease was clinically sta- previously [12]. Briefly, CHAIN is a Spanish multicenter ble and had low clinical impact, adjusted for the level of study carried out at pulmonary clinics. The main goal of disease severity. Stability was defined as the absence of this prospective observational study was to multidimen- exacerbations in the previous 6 months plus no change sionally evaluate the progression of patients with COPD or improvement in subjective perception referred to by to better define the natural history and phenotypes of the patient. Clinical impact was classified as low accord- the disease. The recruitment period began on January ing to the information collected on the dyspnea (mMRC) 15, 2010, and is ongoing (Clinical Trials.gov: identifier scale (0–1 if FEV1 ≥ 50% and 0–2 if FEV1 < 50%) and res- NCT01122758). All participants signed the informed cue medication usage (not needing to use rescue inhalers consent approved by the ethics committees of the par- regularly). The level of control was evaluated longitudi- ticipating centers (Hospital Universitario la Candelaria, nally during visits every 6 months. All participants had Calle Rubio et al. Respir Res (2021) 22:36 Page 3 of 12 a minimum of 12 months of follow-up with clinical con- versus uncontrolled patient characteristics is presented trol measurements. Based on the clinical control status in Table 1. Uncontrolled patients were older and had a evaluated at each visit during follow-up, the cohort was higher body mass index and greater degree of airflow lim- divided into three subgroups: persistently controlled, itation, with more chronic bronchitis and the presence of intermittently controlled and persistently uncontrolled dark sputum, more comorbidities and a poor quality of patients. life. Regarding pharmacological treatment, uncontrolled patients more frequently used inhaled triple therapy. Statistical analysis Data is summarized as frequencies for categorical vari- Control status according to degree of airflow limitation ables, median (5th–95th percentile) for ordinal or non- at recruitment normal scale variables and mean ± SD for normally Of a total of 300 patients with severe/very severe airflow distributed scale variables. Comparisons were made limitation, 228 patients (76%) were defined as having low- between groups using Pearson’s chi-squared test, the impact disease and 100 patients (33.3%) had stable dis- Kruskal–Wallis H test or the Mann–Whitney U test and ease; therefore, 26.7% were defined as controlled patients. one-way ANOVA or the t-test as appropriate. In mild/moderate COPD, there was a greater proportion Logistic regression was used to investigate factors con- of patients with stable disease: 262 patients (52.6%). Of tributing to clinical control in patients with COPD. A these, 36.9% patients were defined as controlled (Table 2). multivariate analysis considered variables with a statisti- cally significant association (p < 0.05). In the multivariate Prevalence and longitudinal follow‑up of clinical control model, we considered the following independent vari- Over a period of 5 years, the proportion of persistently ables: age, pack-years, chronic bronchitis, dark sputum, controlled patients with COPD was 16.8%, persistently eosinophils, Charlson index, FEV1, KCO, triple therapy, uncontrolled patients accounted for 31.1% and inter- CAT score and HDAS depression. mittently controlled patients represented 52.1% (Fig. 1). We chose the best predictive model, which only had There were significant differences in baseline clinical and the variables CAT score and triple therapy because the physiological characteristics between the persistently others weren’t as relevant to provide a good model. To controlled patients with COPD compared to those who select the model, we used the Akaike and Bayesian infor- were persistently uncontrolled or intermittently con- mation criteria. The final set of variables was selected trolled (Table 3). using a backward stepwise selection algorithm (p < 0.10 During this follow-up over 5 years, the median follow- to remain in the model). The discrimination capacity up time in the persistently controlled patient group was of the predictive model was analyzed by calculating the 2.4 (1.7) years, 4.2 (1.2) years in the intermittently con- area under the Receiver Operating Characteristics (ROC) trolled group and 1.8 (1.3) years for persistently uncon- curve along with a confidence interval at 95%. trolled patients. The loss of patients during follow-up was An unpaired t -test was used to compare baseline data 35.7%. and annual changes between persistently controlled and persistently uncontrolled status. P values less than 0.05 Factors accounting for persistently controlled patient status were considered to be statistically significant. A backward logistic multivariate model was developed A Kaplan–Meier analysis for survival due to all causes with persistent control as the independent variable and was performed in persistently uncontrolled patients. the dependent variables were clinical and demographic Finally, to predict the risk of death, we performed Cox variables, which were not related to the definition of proportional hazard regression analyses with the persis- control. The adjusted model showed that triple therapy tently controlled and uncontrolled subgroups. Signifi- (OR, 0.3026; 95% CI, 0.1776–0.51573; p < 0.001) and cance was established as two-tailed p < 0.05. CAT (OR, 0.9399; 95% CI 0.9032–0.9781; p < 0.001) were independently and significantly associated with persis- Results tently controlled status. The AUC was 0.7029 (95% CI, Participant characteristics 0.64209–0.76367). The population of this study was 798 patients with COPD from the CHAIN study who underwent a minimum of Changes in treatment patterns for COPD in persistently 12 months of follow-up with clinical control measure- controlled and uncontrolled patients ments. Stability was defined as the absence of exacerba- Regarding pharmacological treatment, persistently tions in the last 12 months during the recruitment visit. uncontrolled patients more frequently used inhaled A total of 264 (33%) patients met the criteria for con- corticosteroids, particularly as part of triple ther- trolled status at recruitment. A comparison of controlled apy (72.2%). Of these, 71.8% showed no changes in Calle Rubio et al. Respir Res (2021) 22:36 Page 4 of 12 Table 1 Characteristics of the study population according to control status at recruitment Total Controlled Uncontrolled n = 534 P-value n = 798 n = 264 (33%) (66.8%) Age (years), m (SD) 65.7 (10.5) 62.8 (11.6) 67.2 (9.7) < 0.001 Gender (male), n (%) 663 (82.9) 226 (85.6) 435 (81.5) 0.144 Active smoker, n (%) 229 (28.6) 92 (34.8) 136 (25.5) 0.006 Tobacco exposure, pack-years, m (SD) 56.3 (28.7) 52.6 (25.8) 58.2 (29.9) 0.018 BMI (kg/m2), m (SD) 28.0 (5.1) 27.2 (4.7) 28.4 (5.2) 0.001 Post-bronchodilator FEV1 (%), m (SD) 60.2 (25.9) 68.0 (20.8) 56.4 (27.3) < 0.001 Post-bronchodilator FEV1 (mL), m (SD) 1629.9 (690.4) 1944.0 (746.4) 1476.6 (602.7) < 0.001 KCO%, median (P25-P75) 73 (51–92.9) 76.05 (60–95) 70 (46–91) 0.002 Chronic bronchitis, n (%) 466 (58.2) 132 (50.0) 334 (62.5) 0.001 Dark sputum, n (%) 122 (15.2) 29 (11.0) 93 (17.4) 0.018 Bronchial asthma, n (%) 26 (3.3) 11 (4.2) 15 (2.8) 0.309 Eosinophils (%), median (P25-P75) 2.3 (1.5–3.6) 2.4 (1.6–3.6) 2.3 (1.5–3.6) 0.666 Charlson index, m (SD) 1.2 (1.5) 1.0 (1.5) 1.3 (1.5) 0.009 Treatment, n (%) Inhaled triple therapy 454 (56.8) 105 (39.8) 348 (65.2) < 0.001 Theophylline 73 (9.1) 8 (3.0) 65 (12.2) < 0.001 Influenza vaccine 430 (53.7) 102 (38.6) 328 (61.4) < 0.001 LTOT 104 (13.0) 9 (3.4) 95 (17.8) < 0.001 Home ventilation 41 (5.1) 9 (3.4) 32 (6.0) 0.120 CAT score, m (SD) 12.6 (7.2) 10.3 (6.5) 13.8 (7.3) < 0.001 Anxiety, HDAS, m (SD) 11.1 (4.8) 11.06 (4.8) 11.19 (4.9) 0.576 Depression, HDAS, m (SD) 8.6 (4.7) 8.2 (4.5) 8.9 (4.7) 0.088 BMI body mass index, FEV1 forced expiratory volume in 1 s, KCO carbon monoxide transfer coefficient, Inhaled triple therapy: long-acting beta-2 agonist (LAMA) with corticosteroids (ICS) with long-acting antimuscarinic agent (LAMA), LTOT long-term oxygen therapy, CAT​COPD Assessment Test, HDAS Hospital Anxiety and Depression Scale Table 2 Factors accounting for the control status with triple therapy being the most frequent therapeutic of patients with COPD by level of severity at recruitment pattern (Fig. 2). Total FEV1 ≥ 50% FEV1 < 50% P-value (n = 798) (n = 498) (n = 300) Outcomes in patients with COPD according to longitudinal Clinical impact < 0.001 control status pattern Low, n (%) 546 (68.4) 318 (63.9) 228 (76.0) Longitudinal changes in clinical outcomes (health status, High, n (%) 252 (31.6) 180 (36.1) 72 (24.0) psychological status and airflow limitation) according Stability to a persistently uncontrolled or controlled longitudinal Stable, n (%) 362 (45.4) 262 (52.6) 100 (33.3) control status pattern are shown in Table 4. Regarding the baseline data, persistently uncontrolled Not stable, n (%) 436 (54.6) 236 (47.4) 200 (66.7) < 0.001 patients were significantly worse as rated by CAT and Control status HDAS scores and FEV1 levels. However, there were Controlled, n (%) 264 (33.0) 184 (36.9) 80 (26.7) no significant differences in annual changes in out- Uncontrolled, n (%) 534 (67.0) 314 (63.1) 220 (73.3) 0.003 comes between persistently controlled and uncontrolled FEV1 forced expiratory volume in 1 s patients. Regarding survival, there were 94 (24.6%) deaths in 382 patients with a persistently uncontrolled or controlled treatment during follow-up, 13.3% underwent de- status pattern, of which 73 (19.1%) were persistently escalation and 14.9% escalation in treatment. Patients uncontrolled and 21 (5.5%) were persistently controlled who were persistently controlled more frequently used (p = 0.001). The Kaplan–Meier analysis for all-cause bronchodilators, particularly monotherapy (53%), fol- mortality showed that persistently uncontrolled status lowed by triple therapy (37%). Of these, 5.2% de-esca- was associated with a shorter survival time (3.58 years; lated treatment and 19.4% scaled up their treatment, 95% CI, 3.31–3.85) than persistently controlled status Calle Rubio et al. Respir Res (2021) 22:36 Page 5 of 12 Fig. 1 The evolution of the clinical control pattern every year (4.43 years; 95% CI, 4.19–4.67) (Fig. 3). The hazard ratio The current analysis describes the progression of clini- for all-cause mortality was 2.274 (95% CI, 1.394–3.708; cal control in a well-characterized COPD cohort over a p = 0.001). period of 5 years as monitored at pulmonary clinics. In our study, only 33% of patients with different degrees of Discussion COPD severity met the criteria required to be consid- This study provides novel information on the longitudinal ered controlled at recruitment. In the mild or moderate evolution of clinical control in a large cohort of patients subgroup of patients, 36.9% were defined as controlled with COPD as well as factors associated with persistent whereas only 26.7% of severe patients were defined as clinical control and their clinical consequences. controlled. These results are similar to those obtained The main results of our study indicate three things. in an international multicenter study, obtaining an over- First, in the population with COPD, there were frequent all control value of 32% using the clinical evaluation of changes in clinical control status. Only a small percent- control criteria [10]. Another prospective study showed age of patients could be classified as persistently con- similar results, with only 27.5% [8] of patients being con- trolled over the following 5 years. Secondly, the main sidered controlled. However, it should be mentioned that variables associated with persistent clinical control are unlike these studies, almost 40% of the sample analyzed a better quality of life as evaluated by the CAT and not in our study had severe airflow obstruction. In addition, requiring inhaled triple therapy. Finally, the clinical con- the level of physical activity referred to by the patient and sequences of persistent clinical control are observed in the presence of sputum purulence were not included in the risk of death. the clinical impact assessment. In our study population, Calle Rubio et al. Respir Res (2021) 22:36 Page 6 of 12 Table 3 Baseline characteristics of longitudinal clinical control patterns Persistently controlled Intermittently controlled Persistently uncontrolled P-value (n = 134) (n = 416) (n = 248) Demographics and clinical data Male, n (%) 117 (87.3) 336 (80.6) 209 (84.3) 0.153 Age (years), m (SD) 63.2 (9.7)a,b 64.9 (10.9) 68.5 (9.9) < 0.001 Pack-years, m (SD) 53.5 (26.1) 54.6 (28.4) 60.8 (30.2)c 0.013 Active smoker, n (%) 48 (35.8) 115 (27.6) 66 (26.6) 0.128 BMI (kg/m2), m (SD) 27.1 (4.6) 28.0 (4.8) 28.5 (5.6) 0.060 Chronic bronchitis, n (%) 71 (53.0)d 233 (55.9) 162 (65.3) 0.022 Dark sputum, n (%) 15 (11.2)d 54 (12.9) 53 (21.4)c 0.005 Bronquial asthma, n (%) 5 (3.7) 9 (2.2) 12 (4.8) 0.164 Eosinophils (%), median (P25-P75) 2.6 (1.6–3.9)d 2.4 (1.6–3.6) 2.1 (1.3–3.3)c 0.002 Charlson index, m (SD) 1.2 (1.6) 1.0 (1.4) 1.5 (1.6)e 0.002 Physiology FEV1 (L), 2050 (1467–2505)b 1600 (1190–2050)a 1360 (940–1730)£ < 0.001 median (P25-P75) FEV1%pred, 72 (55–88)b 60 (46–74)a 51 (39–63)£ < 0.001 median (P25-P75) FVC (L), 3585 (2847–4380)b 3100 (2450- 3710)a 2745 (2197–3227)e < 0.001 median (P25-P75) FVC %pred, 94 (80–110)b 84 (71–101)a 75 (63–90)e < 0.001 median (P25-P75) FEV1/FVC, 58 (49–65)b 54 (44–63)f 51 (41–60)c < 0.001 median (P25-P75) KCO%, median (P25-P75) 79.5 (62.5–99.7)b 72.6 (52.2–92.8)f 66 (41–85.2)c < 0.001 Treatment Triple therapy, n (%) 50 (37.3)b 226 (54.2)f 178 (71.8)e < 0.001 b f Influenza vaccine, n (%) 50 (37.3) 213 (51.1) 167 (67.3)c < 0.001 LTOT, n (%) 3 (2.2)b 37 (8.9) 64 (25.8)e < 0.001 VMNI, n (%) 5 (3.7)b 12 (2.9) 24 (9.7)c < 0.001 CAT score, median (P25-P75) 8 (5–14.2)b,f 11 (7–16) 14 (9–21)e < 0.001 HDAS anxiety score, median (P25-P75) 11 (6–15) 12 (8–15) 12 (8–15) 0.576 HDAS depression score, median (P25-P75) 8.0 (4.6) 8.5 (4.5) 9.3 (4.9) 0.048 Follow-up time (years), m (SD) 2.4 (1.7) d 4.2 (1.2)a 1.8 (1.3)e < 0.001 BMI body mass index, FEV1 forced expiratory volume in 1 s, FVC forced vital capacity, KCO carbon monoxide transfer coefficient, Triple therapy long-acting beta-2 agonist (LAMA) with corticosteroids (ICS) with long-acting antimuscarinic agent (LAMA), LTOT long-term oxygen therapy, CAT​COPD Assessment Test, HDAS Hospital Anxiety and Depression Scale a p < 0.001 persistently controlled compared with intermittently controlled b p < 0.001 persistently controlled compared with persistently uncontrolled c P ≤ 0.05 persistently uncontrolled compared with intermittently controlled d P < 0.05 persistently controlled compared with persistently uncontrolled e P ≤ 0.001 persistently uncontrolled compared with intermittently controlled f p ≤ 0.05 persistently controlled compared with intermittently controlled the use of rescue medication as a high impact criterion controlled patients compared to 21.4% of persistently was present in 70% of patients classified as having a high uncontrolled patients (p = 0.015). clinical impact. This is a widely justified criterion if we Regarding the longitudinal clinical control patterns, we fear that the increased use of rescue medication has been found that there were frequent changes in clinical con- associated with an increased risk of future exacerbations trol status, with 42.1% of patients changing control status [19]. However, sputum color has shown the lowest dis- during the observation period. Few studies have shown criminative property for the level of impact [10]. In our data on the progression of clinical control in COPD. A study, dark sputum was present in 11.2% of persistently recent international study showed that 53.7% of patients Calle Rubio et al. Respir Res (2021) 22:36 Page 7 of 12 Fig. 2 Changes in treatment patterns for COPD at baseline and last visit in persistently controlled and uncontrolled patients Table 4 Comparisons of baseline data and annual changes between persistently controlled patients and persistently uncontrolled patients during 5 years of follow-up Baseline data Annual changes (/year) Characteristics Persistently controlled Persistently Persistently controlled Persistently uncontrolled uncontrolled CAT score 9.6 (5.9) 15.5 (7.8)* 0.0 (− 1.0–1.7) 0.2 (− 2.5–3.0) HDAS anxiety 10.7 (4.8) 11.1 (4.8) 0.0 (− 1.0–2.0) 0.3 (− 1.0–3.0) HDAS depression 8.0 (4.6) 9.3 (4.9)* 0.5 (− 0.3–2.3) 0.0 (− 1.6–3.0) FEV1, %pred 71.2 (20.8) 52.3 (19.9)* 0.3 (− 2.8–3.0) − 0.2 (− 3.4–1.5) CAT​COPD Assessment Test; HDAS Hospital Anxiety and Depression Scale; FEV1 forced expiratory volume in 1 s Data is presented as mean (SD) or median (5th–95th percentile) *Statistically significant differences between persistently uncontrolled and persistently controlled patients (p < 0.05) changed control status, 29.8% of patients remained con- control status could be used as a supplementary assess- trolled and 16% persistently uncontrolled during an ment tool for decision-making at each medical visit, 18-month follow-up [20]. These results are not compara- similar to the evaluation of asthma control. Table 5 sum- ble to our analysis, where follow-up is greater. Another marizes studies that examined the proportion of con- observational study analyzed changes in control over a trolled patients and changes in clinical control. 3-month period and showed that 29.2% changed their In our study, 31.1% of patients had persistently poor control status [11]. In this study, these changes were sig- disease control during follow-up and only a small pro- nificantly more frequent than changes in GOLD stage, portion (16.8%) of patients had persistently good con- risk level or in phenotype, which further suggests that trol. We found that persistently controlled patients were Calle Rubio et al. Respir Res (2021) 22:36 Page 8 of 12 The use of maintenance respiratory therapy is usually thought to reduce risk. However, data reported in the ECLIPSE [26] and SPIROMICS [27] cohorts reported that patients did shift from high-risk to low-risk groups over time, though the reasons for doing so were unclear. In any case, adequate therapy seems to improve the ratio of infrequent to frequent exacerbators over time [28–30]. In our study, triple therapy at baseline was less frequent in persistently controlled patients (37%) versus persis- tently uncontrolled patients (72.2%). At the end of the follow-up period, 20% of persistently controlled patients had scaled up their treatment, with triple therapy being the most frequent therapeutic pattern. On the con- trary, in persistently uncontrolled patients, 13.3% had increased their pharmacological treatment while 15% had decreased it, observing a decrease in triple therapy and an increase in double bronchodilator therapy. These results for the changes in treatment pattern according to longitudinal control status provide interesting informa- tion, showing an increase in triple therapy in persistently Fig. 3 Kaplan–Meier analysis for all-cause mortality. Persistently controlled patients. In our study, not requiring triple controlled patients were associated with a longer survival time than persistently uncontrolled patients therapy at baseline and having a better quality of life were identified as variables that were significantly associated with persistent disease control. A likely explanation why patients are given triple therapy to prevent exacerba- younger, had less frequent chronic bronchitis, a lower tions is because they are believed to be progressing more degree of airflow obstruction, lower involvement in the poorly and are thus more likely to relapse in the future, diffusion test, a better quality of life as evaluated by the irrespective of any positive effect of their therapy. CAT and a higher level of peripheral eosinophilia. In A previous publication described control status as a previous studies [9, 10, 20, 21], the presence of chronic marker of increased risk of poor outcomes in the short bronchitis, female sex, lower BMI and a history of prior term. According to data reported in the studies by Soler- exacerbations were identified as variables that were sig- Cataluña et al. [8] and Barrecheguren et al. [31], con- nificantly associated with poor control. In addition, poor trolled patients showed a lower risk of complications, lung function and worse health status were demonstrated with a longer delay until the first combined event, the to be the best predictors of the risk of future exacerba- first exacerbation and hospitalization, as well as better tions and were associated with a significant increase health status at 1 year of follow-up. However, they did in the risk of mortality [22]. However, our study found not report any significant difference in survival between that sex, tobacco history, BMI and comorbidities such as controlled and uncontrolled patients. In the Miravitlles bronchial asthma or anxiety and depression were simi- et al. [20] study, uncontrolled patient visits resulted in a lar in patients, irrespective of longitudinal clinical con- highly significant increased risk of poor outcomes over trol status. These results are similar to those reported the next 6 months, with an OR of 4.25 for hospitalization by Calverley et al. [23], who showed that tobacco history due to exacerbation compared to controlled patient vis- and BMI were similar in individuals with frequent exac- its. In addition, it has been reported that control status erbations and those who never experienced an exacerba- determined by clinical criteria was a better predictor of tion over the 2 years of follow-up. However, continued exacerbations compared to CAT criteria (AUC: 0.67 vs smoking in patients with COPD has been associated with 0.57) [32]. Our analysis showed that although a further higher disease impact and increased exacerbations [24]. worsening in CAT and HDAS scores and FEV1 levels was In addition, former smokers had a significantly reduced observed in persistently uncontrolled patients, there were risk of death and hospitalization compared to active no significant differences in annual changes between per- smokers [25]. In our study, the majority of the patients sistently controlled and uncontrolled subjects. However, maintained their tobacco use status. There were no dif- we found that persistently controlled patients had a sig- ferences in longitudinal clinical control patterns regard- nificantly lower risk of death than those who were persis- ing smoking cessation during follow-up. tently uncontrolled. In our study, there were 94 (24.6%) Calle Rubio et al. Respir Res Table 5 Summary of studies that examined the distribution of control status, changes in clinical control and evaluated the predictive value of control Number of patients Investigation objective Study findings Study design (2021) 22:36 Baloira A et al. [9] (2016) 481 patients Distribution of control status 36.8% of patients were controlled Spanish cross-sectional multicenter study (primary care vs respiratory care) Nibber A [9] et al. (2017) 2788 patients To validate the concept of control 4.5% of patients were controlled Retrospective observational cohort study Distribution of control status Time to first exacerbation was longer for controlled patients (p < 0.001) Miravitlles M [10] et al. (2018) 314 patients To validate concept of control 32% of patients were controlled Multicenter prospective observational study Distribution of control status Soler-Cataluña JJ [8] et al. (2018) 265 patients To validate “modified” control criteria 61.5% of patients were controlled Spanish multicenter prospective observational study To evaluate predictive value of control The time to the first combined event (emergency room visit, hospitalization, or death) was significantly greater in controlled patients (p < 0.001) Barrecheguren [31] et al. (2020) 2044 patients To validate the prospective value of control 20% of patients were controlled Multicenter double-blind SPARK study The rate of exacerbations was lower in controlled patients (OR 0.56, p < 0.0001) and time to first exacer‑ bation was significantly delayed Miravitlles [32] et al. (2020) 307 patients To validate the concept of control in COPD 65% of patients were controlled International, multicenter study Time to first exacerbation was significantly delayed for controlled patients Soler-Cataluña JJ [11] et al. (2020) 354 patients To compare changes in control over a 3-month 50.3% of patients were controlled Prospective multicenter observational study period with changes in risk level and GOLD stage Changes in control over a 3-month period was 29.3% Miravitlles M [20] et al. (2020) 267 patients To describe the changes in control status during During 18 months of follow-up, 29.8% of patients International multicenter study follow-up (18 months) and the predictive value of remained controlled, 16% persistently uncontrolled Follow-up for 18 months control (6 months) and the remaining 53.7% changed control status during follow-up Page 9 of 12 Calle Rubio et al. Respir Res (2021) 22:36 Page 10 of 12 deaths in 5 years of follow-up, a mortality rate similar to prognostic implications such as mortality. In this sense, that of the Spanish PAC-EPOC cohort (3.6 fatal events/ it is worth mentioning that in our analysis, there was a year/100 patients) [33]. Specifically, there were 73 deaths higher number of exitus in the first year of follow-up: 48 in persistently uncontrolled patients and 21 in controlled patients defined as persistently uncontrolled and 13 as patients. Our analysis further confirmed that subjects persistently controlled. In addition, if we establish a mini- who died were older, had a greater degree of airway mum of 3 years of follow-up as a criterion, the majority obstruction, and had worse health status than those who of patients (76.8%) would be classified as intermittently survived. These results are similar to those reported by controlled. Therefore, we defined the longitudinal pattern Oga et al [34]. Changes in mortality occur after the first with a minimum of 1 year of follow-up, also keeping in year and tend to increase in the second year, which could mind that this criterion perhaps most closely resembles explain why this was not observed in previous studies [8]. ordinary clinical practice. Another limitation is that this Our study extends our understanding of the concept of was not an interventional study, we could not investigate control in COPD and its possible application in clinical whether a change in treatment could modify control sta- practice. Previous studies have found that improvement tus and influence the outcomes. This has to be demon- in control status in the short term was associated with strated in future interventional studies. better outcomes, with a reduced frequency of exacerba- tions and improved health status. Our results show that patient control status frequently changes in subsequent Conclusions clinical visits and we observed that there are long-term This is the first study to show the impact of control sta- consequences: persistently uncontrolled patients have tus on long-term mortality. There is important variability higher mortality. This is the first study to show the impact in clinical control status and only a small proportion of of control status on long-term mortality. This increased patients had persistently good control. The study high- risk justifies the use of control evaluation as a warning lights the significantly increased risk of death in uncon- sign to foster more careful evaluation of the patients and trolled patients. Consequently, control criteria should the adoption of therapeutic measures. be incorporated into clinical practice as a simple tool to This study has several strengths. It included a large help reassess patients with COPD at each follow-up visit. number of well-characterized patients being treated for Further studies in other populations should validate our COPD in “real life” with a long follow-up time, provid- results. ing invaluable information on outcomes which is not usually available in most pharmacological trials. How- Abbreviations ever, it is necessary to keep in mind some characteristics AECOPD: Acute exacerbations of COPD; BMI: Body mass index; CATTM: COPD of the cohort in order to correctly interpret our results. assessment test; COPD: Chronic obstructive pulmonary disease; ICS: Inhaled The CHAIN cohort was obtained from an observa- corticosteroids; IQR: Interquartile range; LABA: Long-acting beta-2 agonists; LAMA: Long-acting antimuscarinic agents; mHealth apps: Health-related tional study of patients visiting pulmonary clinics and mobile applications; mMRC: Modified Medical Research Council. not from general medical practice. In fact, patients with COPD treated in a specialized clinic have been found Acknowledgements The authors thank Irene Serrano Garcia, from Health Research Institute at to have better clinical control [35]. In the Baloira et al. Hospital Clínico San Carlos, for statistical support in the present analysis. The [36] study, patients at the primary care level were more authors are also grateful to AstraZeneca for funding the project. poorly controlled. However, our cohort included a large Authors’ contributions population of patients with different degrees of severity MCR, JLRH: conception and design, recruitment of patients, analysis and (16.4% mild, 46% moderate, 26.8% severe and 10.8% very interpretation and drafting the manuscript for important intellectual content. severe). Another consideration is that few women were JPT, JMM, CMG, AF, BC, GPB, IS, NFC, CC: conception and design, recruitment of patients, analysis and interpretation. JLL-C: conception and design, analysis included in the cohort and the findings reported in rela- and interpretation and drafting the manuscript for important intellectual tion to this must be interpreted with caution. There was content. All authors read and approved the final manuscript. also a loss of patients during follow-up that could result Funding in measurement bias. Regarding the limitations of the This study has been funded by AstraZeneca. The funder had no role in study present study, it is important to consider that the prob- design, data collection and analysis, decision to publish, or preparation of the ability of change in clinical control status will be greater manuscript. for a longer follow-up period. In our analysis, a minimum Availability of data and materials of 1 year of follow-up was established as a criterion to Data can be shared upon request by writing to mcallerubio@gmail.com. define the longitudinal pattern since our objective was to Ethics approval and consent for participation explore the differences between persistently controlled All participants signed the informed consent approved by the ethics commit‑ and uncontrolled classifications and to analyze their tees of the participating centers (Hospital Universitario la Candelaria, Tenerife; Calle Rubio et al. Respir Res (2021) 22:36 Page 11 of 12 Spain; IRB No. 258/2009). This manuscript does not contain any individual 2. Miravitlles M, Soler-Cataluña JJ, Calle M, Soriano JB. Treatment of COPD by person’s data. clinical phenotypes: putting old evidence into clinical practice. Eur Respir J. 2013;41(6):1252–6. Consent for publication 3. McDonald V, Higgins I, Wood LG, Gibson PG. Multidimensional assess‑ Not applicable. ment and tailored interventions for COPD: respiratory utopia or common sense? Thorax. 2013;68(7):691–4. Competing interests 4. Vogelmeier CF, Criner GJ, Martínez FJ, et al. Global strategy for the MCR has received speaking fees from Boehringer Ingelheim, AstraZen‑ diagnosis, management, and prevention of chronic obstructive lung eca, GlaxoSmithKline, Menarini and Novartis, and consulting fees from disease 2017 report: GOLD executive summary. Arch Bronconeumol. GlaxoSmith¬Kline, Gebro Pharma and Novartis. There is no real or perceived 2017;53(3):128–49. conflict of interest between these sources and the present paper. JLRH has 5. Miravitlles M, Soler-Cataluña JJ, Calle M, et al. Spanish COPD guidelines received speaking fees from Boehringer Ingelheim and Gebro Pharma. There (GesEPOC). 2017. Pharmacological treatment of stable chronic obstruc‑ is no real or perceived conflict of interest between these sources and the tive pulmonary disease. Arch Bronconeumol. 2017;53(6):324–335. present paper. JPT does not have a real or perceived conflict of interest. JMM 6. Soler-Cataluña JJ, Alcázar B, Miravitlles M. The concept of control of COPD does not have a real or perceived conflict of interest. CMG does not have a in clinical practice. Int J Chron Obstruct Pulmon Dis. 2014;9:1397–405. real or perceived conflict of interest. AF does not have a real or perceived 7. Soler-Cataluña JJ, Alcazar-Navarrete B, Miravitlles M. The concept of conflict of interest. BC reports grants, personal fees and non-financial support control in COPD: a new proposal for optimising therapy. Eur Respir J. from GSK; grants, personal fees and non-financial support from Chiesi; grants, 2014;44(4):1072–5. personal fees and non-financial support from Astrazeneca; grants from Menar‑ 8. Soler-Cataluña JJ, Marzo M, Catalán P, Miralles C, Alcazar B, Miravitlles M. ini and Boehringer-Ingheilm; non-financial support from Novartis; personal Validation of clinical control in COPD as a new tool for optimizing treat‑ fees and non-financial support from Sanofi, outside the submitted work. GPB ment. Int J Chron Obstruct Pulmon Dis. 2018;13:3719–31. reports grants, personal fees and non-financial support from GSK, Boehringer 9. Nibber A, Chisholm A, Soler-Cataluña JJ, Alcazar B, Price D, Miravitlles M. Ingelheim, Chiesi and Orion Pharma. There is no real or perceived conflict of Validating the concept of COPD control: a real-world cohort study from interest between these sources and the present paper. IS does not have a the United Kingdom. COPD. 2017;14(5):504–12. real or perceived conflict of interest. NFC does not have a real or perceived 10. Miravitlles M, Sliwinski P, Rhee CK, et al. Evaluation criteria for clinical conflict of interest. JLLC reports personal fees and non-financial support from control in a prospective, international, multicenter study of patients with AstraZeneca; grants, personal fees and non-financial support from Boehringer COPD. Respir Med. 2018;136:8–14. Ingelheim; grants, personal fees and non-financial support from Chiesi; per‑ 11. Soler-Cataluña JJ, Alcázar B, Marzo M, Pérez J, Miravitlles M. Evaluation of sonal fees and non-financial support from CSL Behring; grants, personal fees changes in control status in COPD: an opportunity for early intervention. and non-financial support from Esteve; personal fees and non-financial sup‑ Chest. 2020;157:1138–46. port from Ferrer; grants, personal fees and non-financial support from GebroP‑ 12. Lopez-Campos JL, Peces-Barba G, Soler-Cataluna JJ, et al. Chronic harma; grants, personal fees and non-financial support from GlaxoSmithKline; obstructive pulmonary disease history assessment in Spain: a multi‑ grants, personal fees and non-financial support from Grifols; grants, personal dimensional chronic obstructive pulmonary disease evaluation: study fees and non-financial support from Menarini; grants, personal fees and methods and organization. Arch Bronconeumol. 2012;48(12):453–9. non-financial support from Novartis; grants, personal fees and non-financial 13. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prog‑ support from Rovi; and grants, personal fees and non-financial support from nostic comorbidity in longitudinal studies: development and validation. J Teva, outside the submitted work. CC has received speaker fees from Novartis, Chronic Dis. 1987;40:373–83. Menarini, Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline and Teva, and 14. American Thoracic Society. Lung function testing: selection of reference consulting fees from AstraZeneca, Esteve, GlaxoSmithKline and Novartis. values and interpretative strategies. Am Rev Respir Dis. 1991;144:1202–18. 15. Macintyre N, Crapo RO, Viegi G, et al. Standardisation of the single-breath Author details determination of carbon monoxide uptake in the lung. Eur Respir J. 1 Pulmonology Department, Hospital Clínico San Carlos, C/ Martin Lagos S/N, 2005;26:720–35. 28040 Madrid, Spain. 2 Medical Department, School of Medicine, Universidad 16. Mahler DA, Weels CK. Evaluation of clinical methods for rating dyspnea. Complutense de Madrid, Madrid, Spain. 3 Respirology and Sleep Division, Chest. 1988;93:580–6. Queen’s University, Kingston, ON, Canada. 4 Respiratory Department. Hospital, 17. Jones PW, Harding G, Berry P, Wiklund I, Chen WH, Kline LN. Develop‑ Universitario Miguel Servet and IISAragón, Ciber Enfermedades Respiratorias, ment and first validation of the COPD Assessment Test. Eur Respir J. Madrid, Spain. 5 Pulmonology Department, Hospital Universitario Central 2009;34(3):648–54. de Asturias, Universidad de Oviedo, Oviedo, Spain. 6 Pulmonology Depart‑ 18. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta ment, Hospital Universitario Son Llàtzer, Palma de Mallorca, Spain. 7 Depart‑ Psychiatr Scand. 1983;67:361–70. ment of Respiratory Medicine, Hospital Universitario Son Espases-IdISBa 19. Jenkins CR, Postma DS, Anzueto AR, et al. Reliever salbutamol use as a and CIBERES, Palma de Mallorca, Spain. 8 Pulmonology Department, IIS- measure of exacerbation risk in chronic obstructive pulmonary disease. Fundación Jiménez Díaz-CIBERES, Madrid, Spain. 9 Pulmonology Department, BMC PulmMed. 2015;15:97. Hospital de La Santa Creu Y San Pau, Universidad Autónoma de Barcelona, Bar‑ 20 Miravitlles M, Sliwinski P, Rhee CK, et al. Changes in control status of celona, Spain. 10 Pulmonology Department, Hospital Universitario Reina Sofía, COPD over time and their consequences: a prospective international, Instituto Maimónides de Investigación Biomédica de Córdoba, Universidad study. Arch Bronconeumol. 2020a. https​://doi.org/10.1016/j.arbre​ de Córdoba, Córdoba, Spain. 11 Unidad Médico-Quirúrgica de Enfermedades s.2020.06.003. Respiratorias, Instituto de Biomedicina de Sevilla (IBIS), Hospital Universitario 21. Busch R, Han MK, Bowler RP, et al. Risk factors for COPD exacerbations in Virgen del Rocio, Universidad de Sevilla, CIBERES, Seville, Spain. 12 Pulmonol‑ inhaled medication users: the COPDGene study biannual longitudinal ogy Department, Hospital Universitario Nuestra Señora de Candelaria, Univer‑ follow-up prospective cohort. BMC Pulm Med. 2016;16:28. sidad de La Laguna, Tenerife, Spain. 22. Hurst JR, Vestbo J, Anzueto A, et al. Susceptibility to exacerba‑ tion in chronic obstructive pulmonary disease. N Engl J Med. Received: 21 October 2020 Accepted: 21 January 2021 2010;363(12):1128–38. 23. Calverley PM, Tetzlaff K, Dusser D, et al. Determinants of exacerbation risk in patients with COPD in the TIOSPIR study. Int J Chron Obstruct Pulmon Dis. 2017;12:3391–405. 24. Riesco JA, Alcazar B, Trigueros JA, Campuzano A, Perez J, Lorenzo JL. References Active smoking and COPD phenotype: distribution and impact on prog‑ 1. Agusti A, MacNee W. The COPD control panel: towards personalised nostic factors. Int J Chron Obstruct Pulmon Dis. 2017;12:1989–99. medicine in COPD. Thorax. 2013;68(1):687–90. 25. Josephs L, Culliford D, Johnson M, Thomas M. Improved outcomes in exsmokers with COPD: a UK primary care observational cohort study. Eur Respir J. 2017;49(5):1602114. Calle Rubio et al. Respir Res (2021) 22:36 Page 12 of 12 26. Donaldson GC, Müllerova H, Locantore N, et al. Factors associated with la EPOC en el estudio PAC-COPD: diseño y metodología [Phenotypic change in exacerbation frequency in COPD. Respir Res. 2013;14:79. characterization and course of chronic obstructive pulmonary disease 27. Han MK, Quibrera PM, Carretta EE, et al. Frequency of exacerbations in in the PAC-COPD Study: design and methods]. Arch Bronconeumol. patients with chronic obstructive pulmonary disease: an analysis of the 2009;45:4–11. SPIROMICS cohort. Lancet Respir Med. 2017;5(8):619–26. 34. Oga T, Nishimura K, Tsukino M, Sato S, Hajiro T. Analysis of the fac‑ 28. Vestbo J, Fabbri L, Papi A, et al. Inhaled corticosteroid containing combi‑ tors related to mortality in chronic obstructive pulmonary disease: nations and mortality in COPD. Eur Respir J. 2018;52(6):1801230. role of exercise capacity and health status. Am J Respir Crit Care Med. 29. Singh D, Papi A, Corradi M, et al. Single inhaler triple therapy versus 2003;167:544–9. inhaled corticosteroid plus long-acting β2-agonist therapy for chronic 35. Calle Rubio M, Soler-Cataluña JJ, López-Campos JL, et al. Assessing the obstructive pulmonary disease (TRILOGY): a double-blind, parallel group, clinical practice in specialized outpatient clinics for chronic obstructive randomised controlled trial. Lancet. 2016;388:963–73. pulmonary disease: analysis of the EPOCONSUL clinical audit. PLoS ONE. 30. Lipson DA, Barnhart F, Brealey N, et al; IMPACT Investigators. Once-daily 2019;14:e0211732. single-inhaler triple versus dual therapy in patients with COPD. N Engl J 36. Baloira A, Gonzalez-Moro JMR, Sanjuán E, Trigueros JA, Casamor R. Degree Med. 2018;378:1671–80. of control of patients with chronic obstructive pulmonary disease in 31. Barrecheguren M, Kostikas K, Mezzi K, et al. COPD clinical control as a Spain: SINCON study. BMC Pulm Med. 2018;18:183. predictor of future exacerbations: concept validation in the SPARK study population. Thorax. 2020;75:351–3. 32. Miravitlles M, Sliwinski P, Rhee CK, et al. Predictive value of control of Publisher’s Note COPD for risk of exacerbations: an international, prospective study. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Respirology. 2020b;25(11):1136–43. lished maps and institutional affiliations. 33. Garcia-Aymerich J, Gómez FP, Antó JM, en nombre del Grupo Investi‑ gador del Estudio PAC-COPD. Caracterización fenotípica y evolución de Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions