Evaluation of the effect of multimorbidity on difficult-to-treat asthma using a novel score (MiDAS): a multinational study of asthma cohorts

Background

Multimorbidity (ie, co-existence of two or more health conditions) is highly prevalent in patients with difficult-to-treat asthma. However, it remains unclear how multimorbidity correlates with disease severity and adverse health outcomes in these patients and which comorbidities are most important. We aimed to address this knowledge gap by developing a patient-centred, clinically descriptive multimorbidity score for difficult-to-treat asthma.

Methods

We used data from the UK-based Wessex Asthma Cohort of Difficult Asthma (WATCH; n=500, data collected between April 22, 2015, and April 1, 2020) to develop the Multimorbidity in Difficult Asthma Score (MiDAS). Initially, we created a modified Asthma Severity Scoring System (m-ASSESS) in WATCH. We then conducted univariate association analysis to test the association between the 13 commonest comorbidities and m-ASSESS in WATCH and used a branch-and-bound approach to select the most relevant comorbidities for inclusion in MiDAS. We calculated MiDAS values for all patients with complete information in WATCH (n=319) and assessed them for correlation with components of m-ASSESS, proinflammatory biomarkers, and St George's Respiratory Questionnaire (SGRQ) score, a quality-of-life measure. We also assessed the association of MiDAS with multiple clinical outcomes in four international cohorts: two from Australia (n=236, data collected between June 14, 2014, and April 1, 2022; and n=140, Aug 6, 2012, to Oct 18, 2016), one from southeast Asia (n=151, March 21, 2017, to Jan 16, 2024), and one from the USA (n=100, July 9, 2021, to Dec 14, 2023).

Findings

We selected seven common comorbidities (ie, rhinitis, gastro-oesophageal reflux disease, breathing pattern disorder, obesity, bronchiectasis, non-steroidal anti-inflammatory drug-exacerbated respiratory disease, and obstructive sleep apnoea) for inclusion in MiDAS on the basis of the branch-and-bound analysis and combined them using multivariate linear regression to derive a MiDAS model associated with m-ASSESS in WATCH. The range of MiDAS scores was 9·6–16·2. In WATCH members, mean MiDAS value was 11·97 (SD 1·21) and MiDAS was nominally correlated with m-ASSESS components of poor asthma control (τ=0·31 [95% CI 0·24–0·38]) and exacerbations (τ=0·16 [0·08–0·24]). MiDAS was also correlated with worse total SGRQ score (r=0·39 [95% 0·28–0·49], p<0·0001) and with the proinflammatory plasma cytokines interleukin (IL)-4 (r=0·19 [95% CI 0·06–0·31], p=0·0036), IL-5 (r=0·35 [0·24–0·46], p<0·0001), and leptin (r=0·29 [0·17–0·40], p<0·0001) in WATCH. MiDAS values across the four international cohorts were similar to those of WATCH (UK cohort), with mean values of 12·33 (SD 1·47) and 12·31 (1·37) in the Australian cohorts, 11·80 (1·20) in the USA cohort, and 11·55 (1·23) in the Singapore cohort. In these cohorts, MiDAS correlated with worse asthma control, worse quality of life, anxiety, depression, and increased inflammation.

Interpretation

MiDAS highlights the co-occurrence of multimorbidity with the worst outcomes in difficult-to-treat asthma. These findings strongly indicate that an airway-centric approach is inadequate and that holistic and multidisciplinary care is imperative. This clinical score could help clinicians to identify patients most at risk from their multimorbidity.

Funding

UK National Institute for Health and Care Research, Australian National Health and Medical Research Council, Hunter Medical Research Institute, University of Newcastle (Australia), and John Hunter Hospital Charitable Trust.