Multidimensional Phenotyping of Severe Asthma Patients and Its Impact on Disease Control and Therapeutic Response - Analysis From the Portuguese Severe Asthma Registry
Multidimensional Phenotyping of Severe Asthma Patients and Its Impact on Disease Control and Therapeutic Response - Analysis From the Portuguese Severe Asthma Registry
Asthma currently affects 358 million individuals worldwide, posing a substantial burden on health care systems. In particular patients with severe asthma have higher morbidity, mortality and asthma-related costs than non-severe patients. The management of severe asthma is still an unmet need and improving the disease-related knowledge is important to optimize care pathways. Registries provide an opportunity to phenotypically describe a cohort of patients in real-world settings. We hypothesize whether patient profiling based on data in the Portuguese Severe Asthma Registry (RAG - Registo de Asma Grave) may contribute to identify predictors of disease control and therapeutic response.
This study aims to (Coprimary Objectives): 1) Identify multidimensional phenotypes associated with health outcomes and therapeutic responses, based on demographic characteristics, clinical features and biomarkers; 2) Explore novel composite endpoint measures of disease control and evaluate its association with the different severe asthma profiles.
This is a cross-sectional, observational, multicenter, real-world study. The study population are the patients of all ages with severe asthma included in the RAG, until Dec 2021. It is estimated that 150 patients will be enrolled, in approximately 12 sites throughout Portugal, which is expected to be a representative sample of Portuguese patients with severe asthma. Eligible patients will be invited to integrate RAG by clinicians at scheduled clinic appointments. The criteria for patients' inclusion in the RAG is based on the definition of Severe Asthma by GINA guidelines, based on step of treatment, adherence and comorbidities management. An additional inclusion criterion is the patient's signed consent to have his/her data included in the registry.
The main data source of this project is the data collected by RAG, an electronic Case Report Form. Descriptive and inferential statistics will be used to characterize and compare the characteristics across different sub-groups. Advanced data-driven statistical methods, such clustering analysis and latent class analysis, will be used for phenotype classification. Multivariate logistic regression modelling and Classification and Regression Tree analysis will be considered.
To address the potential limitations, the RAG has database specifications concerning data definitions and parameters and data validation rules enabling collection of data in the same manner for every patient, with specific and consistent data definitions. To minimize errors related to data completeness and consistency, several validation rules have been implemented and periodic data audits are planned. To avoid unnecessary burden within the clinical workflow, data will be collected at the time of routine medical appointments by the clinician and data entry personnel will assist on this task.
BACKGROUND Asthma currently affects 358 million individuals worldwide, posing a substantial burden on health care systems. Patients with severe asthma have higher morbidity, mortality and asthma-related costs than non-severe patients. Asthma phenotype definition is particularly important in patients with severe disease and who are not controlled with usual therapy. The management of severe asthma is challenging, involving treatment of comorbidities, optimization of medication adherence, and the choice of the best medical treatment for each patient, which remains one of the greatest difficulties. In the last decades, efforts have focused on the classification of different subsets of asthma patients according to its epidemiology, immunology, biomarkers, response to specific pharmacotherapies, and long-term prognosis. These are broadly called phenotypes: a set of clinical features of a specific genetic pattern in a specific environment. The main goal of the phenotype philosophy is the development of targeted and personalized pharmacological approaches. Allergic and nonallergic asthma are very similar in their clinical presentation and are distinguished by the presence or absence of clinical allergic reaction and in vitro Immunoglobulin E (IgE) response to specific aeroallergens. Although they both show several distinct clinical features and different biomarkers, there are no ideal biomarkers to stratify asthma phenotypes and guide therapy in clinical practice. A detailed and systematic clinical history, including comorbidities, spirometry with bronchodilator test, a skin or blood test panel for specific IgE (sIgE) to common regional airborne allergens, and a peripheral blood eosinophil count are very useful for establishing phenotypes. This distinction has prognostic and therapeutic implication, including the use of biologic therapies. It is not easy to choose between the biologics to be the first-choice treatment, and head-to-head comparison studies between them do not exist. Hence, clinical observational studies of real-world large patient populations should contribute to the knowledge on how to select the best biologic treatment for an individual patient.
Research networks are playing a critical role in the advances in severe asthma. They provide both multicenter severe asthma studies and real-life severe asthma patient registers. REAG, Rede de Especialistas em Asma Grave, is an open collaborative network of asthma specialists (allergists, pediatricians, and pulmonologists) who manage severe asthma patients in Portuguese hospitals. The foundational principle of REAG is the informal peer collaboration among colleagues with different medical specialties and backgrounds, maintaining a nonhierarchical organization and consensual decision processes to improve sharing of medical experience, data, and knowledge. Since 2011, this network of experts has been working towards a better care of severe asthma patients by (1) promoting a better coordination between medical specialties for early diagnosis and referral of severe asthma patients; (2) describing and implementing harmonized procedures to adopt in severe asthma healthcare; and (3) improving scientific knowledge on severe asthma in Portugal. In 2013, REAG conducted a 12-month study on patients with severe asthma under biological treatment. This was the first national effort of standardizing outcome assessment. In this study, REAG showed that in spite of being treated with omalizumab, 1/3 of the patients needed unscheduled medical care because of asthma and 29% had to start using or had to increase the dosage of oral corticosteroid (OCS). We could speculate whether before starting omalizumab the asthma control in these patients was even worse or if the treatment with omalizumab was ineffective and should be discontinued, but due to data constraints, we could not confirm these hypotheses. In face of the need to gather evidence on severe asthma in Portugal, in 2018, REAG developed and implemented the Portuguese Severe Asthma Registry (Registo de Asma Grave Portugal - RAG).
Disease registries are recognized as powerful tools to improve disease-related knowledge. They consist of organized systems that use observational study methods to collect uniform data aiming at evaluating specific outcomes for a heterogeneous population defined by a particular disease. Registries promote the harmonization of the healthcare delivery by physicians, the improvement of clinical processes, and are useful in severe asthma systematic evaluation, enabling the assessment of the effect of different therapies in the context of a single disease. They also provide an opportunity to improve the understanding of different phenotypes of asthma and their clinical implications, reported in a study based on the British Thoracic Society Severe Asthma Registry, which demonstrated the differences in severity of asthma in adult patients and the differences in inflammatory markers in late-onset asthma (Newby 2014). Another study on the Australian Severe Asthma Web-Based Database demonstrated that treatable traits in severe asthma patients can be identified using registry data and that they can be related with severe outcomes. The Portuguese Severe Asthma Registry - RAG is a national web-based disease registry of adult and pediatric patients with severe asthma. It was created in line with the International Severe Asthma Registry (ISAR) and with Severe Heterogenous Asthma Research collaboration, Patient-centered collaboration (SHARP). Specifically, RAG collects data with potential to improve the healthcare delivery by identifying the best treatment practices and standardizing the management of the disease control.
Patients with severe asthma included in RAG will be considered to 1) identify multidimensional phenotypes associated with health outcomes and therapeutic responses, based on demographic characteristics, clinical features and biomarkers and 2) Explore novel composite endpoint measures of disease control and evaluate its association with the different severe asthma profiles. Additionally, to explore predictors of future risk in different asthma profiles
METHODS This is a multicenter, cross-sectional, observational real-world study; based on data collected from the Portuguese Severe Asthma Registry, the RAG.
The main data source of this project is the data collected by RAG, an electronic Case Report Form (eCRF) specially designed for the study. Data collection will follow the Documentation of procedures (see section "Activities description" - Activity 1 Setup and Management). Only authorized physicians/investigators will have access to RAG through a unique login and password. Data entry will be performed by each physician, assisted by data entry personnel, during scheduled medical appointments. Additionally, to complete missing data, medical records may be consulted. (see section "Activities description" - Activity 3 Data collection). Data collectors will be trained in a structured manner. The investigator will allow study-related monitoring, audits, providing direct access to source data documents. The data collected by the RAG includes demographic data; asthma care information; comorbidities; atopy and inflammation biomarkers; diagnostic tests and study assessments; asthma control; therapy (see section "Activities description" - Activity 3 Data collection).
Methods are described in detail in following paragraphs (Activities 1-6).
Activity 1. Setup and Management (18 months) This activity will be responsible for the management of the entire project, including the setup and coordination, and financial support.
The first task comprises setup activities and coordination, including:
Development and management of the documentation of procedures, containing all the registry policies, protocols, procedures, and quality requirements, as well as a complete data dictionary listing all the data elements to be collected and their definitions
Publication of the summary protocol in ClinicalTrials.gov and, and at the end of the study, publication of the results summary in the same public registry
Recruitment of the hospitals and physicians:
Definition of the strategies to retain hospitals and physicians, by regular standardized training meetings, site audit and retraining visits, feedback, telephone helpline
Preparation of the data sources
The main data source for this project is the RAG, an eCRF. The preparation of the RAG for this project includes the:
Establishment of contracts and collaborative agreements with the involved stakeholders: Contract Research Organization (CRO), Investigators, software developers
Writing and production of 3 Internal Progress Summaries (each 3 months) containing summaries on the registry growth, trends, onsite visits, data completeness and cleaning
Writing and production of a Final Report on patients' enrollment, data collection and cleaning.
The outputs of this activity will be 3 progress summaries, 1 final report
Activity 2. Patients' enrolment (12 months) Patients' recruitment is the essential element in the operation of the registry. It is estimated that 150 severe asthma patients will be enrolled, in approximately 12 sites throughout Portugal. Enrollment will be competitive, with a minimum of 10 patients and a maximum of 30 patients per site. Eligible patients will be invited to integrate RAG by physicians at scheduled medical appointments. Patients will be informed that participation on RAG is free and voluntary and that they may, in any moment and without penalty, withdraw the registry or verify and/or delete their data. Patients will also be informed on the purposes of RAG, the data collected, and the implications of participating in this registry. The informed consent form will be automatically generated at the time of inclusion. Only patients that agree, by a clear affirmative consent given by a written statement will be included in RAG.
Recruitment and retention strategies will be applied according with the definition established in the previous tasks.
The output of this activity will be the inclusion of at least 150 new patients with severe asthma in RAG.
Activity 3. Data collection (12 months) This activity aims to collect data in a uniform manner for every patient. Data collection will follow the Documentation of procedures established in the setup task. Data entry will be performed by each physician, assisted by data entry personnel, during scheduled medical appointments. Additionally, to complete missing data, medical records may be consulted.
The data collected though the eCRF of the RAG includes 7 domains: 1) Demographic data: gender, month and year of birth, birthplace, place of residence, body mass index, education years, smoking habits, occupation, family history of asthma and of asthma-related death, personal history of respiratory infections during early childhood, environmental exposures; 2) Asthma care information: age at asthma diagnosis, age at severe asthma classification, first year of specialized asthma follow-up, medical specialty of the attending physician; 3) Comorbidities; 4) Atopy and Inflammation biomarkers: Atopy (serum IgE, allergic sensitization, type(s) of diagnostic test used to confirm allergic sensitization); Inflammation biomarkers (Fractional exhaled Nitric Oxide - FeNO, blood eosinophils, sputum eosinophils, sputum neutrophils); 5) Diagnostic tests and study assessments: Lung function tests (FEV1, FVC, MEF, residual volume, specific airway resistance, carbon monoxide diffusion capacity, bronchial challenge test); Imaging (X-ray, thorax CT scan, sinus CT scan, bronchial endoscopy, bone densitometry); Arterial blood gases; 6) Control: Asthma-related healthcare utilization in previous 12 months or since the last appointment, when at follow-up visit (number of routine primary care medical appointments, routine hospital care medical appointments, nonscheduled medical appointments, emergency department visits, hospital admissions, intensive care unit admissions, need for mechanical ventilation, school or labor absenteeism); Asthma control assessment according to GINA recommendations (frequency of daytime symptoms, activity limitations due to asthma, any night awakening due to asthma, frequency of use of reliever medications for asthma, respiratory function, number of exacerbations in last year/week); Asthma control self-questionnaires (CARAT - Control of Allergic Rhinitis and Asthma Test and ACT - Asthma Control Test); 7) Therapy: Asthma medication OCS, ICs, LTRAs, LABAs, SABAs, LAMAs, SAMAs, xanthines, immunosuppressors, immunotherapy, monoclonal antibodies, antibiotics, therapy adherence, inhalation technique); Other medication (proton pump inhibitor, anti-depressive/anxiolytics, intranasal steroids, antihistamines, long-term oxygen therapy, non-invasive ventilation).
To ensure the quality of the data collection, data collectors will be trained in a structured manner. Furthermore, a data monitor will review samples of data to identify potential sites at higher suspicion for inaccuracy or intentional errors, such as discrepancies between enrolment and logs, missing data, and overly high or low enrolment and onsite periodic audits will be conducted for the identified sites.
Security features compliant with the European GDPR and required procedures according to this legislation are incorporated into the RAG. RAG does not record any identifiable personal data from patients (e.g., date of birth is replaced by the month and year of birth, no ID numbers are registered, and patients' names are pseudo-anonymized by replacing with a code number).
The outputs of this activity will be a complete database from a representative sample of at least 150 patients.
Activity 4. Data cleaning (12 months) This activity aims to correct data problems and prepare the database for analysis.
A data monitor will be responsible for data cleaning to assure validity and quality of the data in the registry. Data checks to assess sample size and compare data entered into the registry against predefined rules will be conducted, by checking and correcting incorrect or out-of-range values, responses that are logically inconsistent with other responses in the database, and duplicate patient records. When missing values are detected a reminder will be sent to the data collector to check and correct them.
To improve data quality, automated simple data checks are specified in the database of RAG for presentation at the time of data entry. To detected unexpected discrepancies and abnormal trends a periodic manual data cleaning will be performed.
Tracking of the data entered and data cleaned will be reported. The outputs will be a database ready for data analysis integrating all patients registered in RAG until December 2020.
Activity 5. Data analysis (6 months) The aim of data analysis is the identification of patients' profiles and its association with disease control measures.
Descriptive statistics will be used to characterize all collected variables. The normality of each variable will be investigated with Kolmogorov-Smirnoff and visual analysis of histograms. Relative and absolute frequencies will be used for categorical and mean, median, standard deviation, maximum, minimum, interquartile range for continuous variables. Inferential statistics will also be conducted to compare the characteristics across different sub-groups using chi-square tests, one-way analysis of variance or Kruskal-Wallis tests. Correlations with Pearson's coefficient or Spearman's rho will also be used to measure the association between characteristics. Advanced data-driven statistical methods, such as distance-based (clustering analysis) and model-based (latent class analysis), will be used for phenotype classification. The importance of the variables that best distinguish between the obtained phenotypes will then be explored. Multivariate logistic regression modelling and Classification and Regression Tree (CART) analysis will be considered. Other appropriate statistical methods can also be applied.
Activity 6. Manuscript writing and Dissemination (10 months) The aim of this activity is to disseminate the project, including the scientific writing.
The results from this project will be disseminated through the participation on a scientific international meeting and the publication of a full-text original article in an international journal indexed in the Web of Science.
The outputs will be the publication of a full-text original article and of a conference abstract.
LIMITATIONS Real-world studies, presented as a pragmatic approach to everyday clinical practice, are fundamental to assess the therapeutic responses and clinical outcomes in patients with severe asthma. However, this type of studies may have some limitations.
The first is the risk of divergent approaches adopted by each recruiting site. To address this limitation the RAG includes database specifications concerning data definitions and parameters and data validation rules that were determined through consensus by the physicians of the REAG. As so data will be collected in the same manner for every patient, with specific and consistent data definitions. To minimize errors related to data completeness and consistency, several logical and validation rules have been implemented and periodic data audits are planned.
This study relies on physicians' active involvement, as they are responsible for making the decision about the inclusion of a patient in RAG and are responsible for inserting patients' data. To avoid unnecessary burden within the clinical workflow, data will be collected at the time of routine medical appointments, several features to support healthcare providers on severe asthma care are implemented in RAG and data entry personnel will assist on this task. Importantly, as suggested by the members of REAG, RAG includes the automatic generation of clinical notes based on the inputted data that can be pasted into the institutional electronic clinical record of the patient, avoiding duplication of effort.
An additional challenge will be the retention of sites/physicians. The participation in the RAG should not be an overload on the usual care that physicians provide to their patients and features have been implemented to reduce the burden of participation. Additionally, the participation in the RAG may be seen as an opportunity to increase the communication of the clinician with their patient, and a to contribute to the knowledge on severe asthma. Retention strategies such as regular standardized training meetings, site visits, feedback and reminders, will be applied.
Informed consent and data protection Informed consent forms were prepared according to age - under 6, between 6 and 15, between 16 and 17 and above 17 years old - in compliance to the European General Data Protection Regulation (GDPR). The informed consent form must be downloaded, dated and signed at the time of inclusion. The informed consent forms are automatically generated in triplicate, in order to provide one copy to the patient, one to the investigator and another to be included in the institutional clinical process of the patient.
Security features compliant with the European GDPR and required procedures according to this legislation are incorporated into the RAG. RAG does not record any identifiable personal data from patients (e.g., date of birth is replaced by the year of birth, no ID numbers are registered, and patients' names are pseudo-anonymized by replacing with a code number).
Inclusion Criteria:
Exclusion Criteria:
cl_loureiro@hotmail.com919795255
anasasousa@gmail.com