Multimorbidity clusters in patients with chronic obstructive airway diseases in the EpiChron Cohort

Abstract

Chronic obstructive airway diseases such as chronic obstructive pulmonary disease (COPD), asthma, rhinitis, and obstructive sleep apnea (OSA) are amongst the most common treatable and preventable chronic conditions with high morbidity burden and mortality risk. We aimed to explore the existence of multimorbidity clusters in patients with such diseases and to estimate their prevalence and impact on mortality. We conducted an observational retrospective study in the EpiChron Cohort (Aragon, Spain), selecting all patients with a diagnosis of allergic rhinitis, asthma, COPD, and/or OSA. The study population was stratified by age (i.e., 15-44, 45-64, and ≥ 65 years) and gender. We performed cluster analysis, including all chronic conditions recorded in primary care electronic health records and hospital discharge reports. More than 75% of the patients had multimorbidity (co-existence of two or more chronic conditions). We identified associations of dermatologic diseases with musculoskeletal disorders and anxiety, cardiometabolic diseases with mental health problems, and substance use disorders with neurologic diseases and neoplasms, amongst others. The number and complexity of the multimorbidity clusters increased with age in both genders. The cluster with the highest likelihood of mortality was identified in men aged 45 to 64 years and included associations between substance use disorder, neurologic conditions, and cancer. Large-scale epidemiological studies like ours could be useful when planning healthcare interventions targeting patients with chronic obstructive airway diseases and multimorbidity.

Figure 1

Prevalence of multimorbidity clusters in the study population based on gender and age, and odds ratios of association with mortality. COPD chronic obstructive pulmonary disease, OSA obstructive sleep apnea.

Figure 2

Disease composition of the multimorbidity clusters identified in the study population based on sex and age, according to prevalence ((P); proportional to node size) and prevalence ratios (PRs, we used the lowest PR for each disease). Each cluster has a color, if a node (disease) has more than one color, that disease is included in other clusters.