Key Takeaway

Medium- and high-dose inhaled corticosteroids (ICS) in adults with asthma were associated with higher 12-month risks of major adverse cardiac events (MACE), arrhythmia, pulmonary embolism (PE), and pneumonia, whereas low-dose ICS showed no increased risk; absolute event rates were low, supporting use of the lowest effective ICS dose.

Background

ICS are foundational for asthma control, preventing exacerbations and reducing mortality. Despite known systemic effects of oral corticosteroids, less is known about short-term harms from ICS, which can be systemically absorbed. This study assessed dose-related adverse events after initiating ICS in asthma.

Patients

• Adults with asthma (>18 years), ICS-naïve at cohort entry.
• United Kingdom primary care data (CPRD Aurum; validation in CPRD GOLD), 2004–2019.
• Required ≥1 year of prior data; asthma defined by ≥2 Read codes within 2 years.
• Main cohort size: 162,202.
• Exclusions (main cohort and SCCS): oral corticosteroid use in the prior year or during follow-up.

Intervention

Initiation of inhaled corticosteroids, analyzed by beclomethasone-equivalent average daily dose:

• Low: <200 μg/day
• Medium: 201–599 μg/day
• High: >600 μg/day

ICS molecules included beclomethasone, budesonide, and fluticasone.

Control

Asthma patients not yet initiated on ICS during the observation window, compared using stabilized inverse probability treatment weighting (IPTW) to balance confounders.

Outcome

• Primary outcomes: MACE (myocardial infarction, stroke, cardiovascular death), arrhythmia, pulmonary embolism, hospitalized pneumonia (12-month risk).
• Negative controls: urinary tract infection, forearm fracture, abdominal pain/diverticular disease, cellulitis (to assess residual confounding).
• Secondary endpoints/analyses: composite cardiovascular disease (MACE + arrhythmia), time since last ICS, self-controlled case series (SCCS) for cardiovascular disease.

Study Design

• Main: New-user IPTW cohort (CPRD Aurum) with dose-response modeling (restricted cubic spline) and weighted flexible parametric models for absolute risk.
• Secondary: Nested case–control studies (Aurum and GOLD) and SCCS (cardiovascular disease only).
• Linked to Hospital Episode Statistics and national mortality data; 7-day pre-exposure window excluded to reduce protopathic bias.

Level of Evidence

Oxford CEBM Level 2b: Large, weighted observational cohort with corroboration from nested case–control and SCCS analyses.

Follow up period

12 months after ICS initiation (time windows also examined in secondary analyses).

Results

Primary outcomes (dose-stratified relative risk vs. no ICS)

• No increased risk at low dose (<200 μg/day) for all outcomes.
• Medium dose (201–599 μg/day):
  • MACE: HR 2.63 (95% CI, 1.66–4.15); NNH 473 over 12 months.
  • Arrhythmia: HR 2.21 (1.60–3.04); NNH 567 over 12 months.
  • PE: HR 2.10 (1.37–3.22); NNH 1,221 over 12 months.
  • Pneumonia: HR 2.25 (1.77–2.85); NNH 230 over 12 months.
• High dose (>600 μg/day):
  • MACE: HR 4.63 (2.62–8.17); NNH 224 over 12 months.
  • Arrhythmia: HR 2.91 (1.72–4.91); NNH 396 over 12 months.
  • PE: HR 3.32 (1.69–6.50); NNH 577 over 12 months.
  • Pneumonia: HR 4.09 (2.98–5.60); NNH 93 over 12 months.
• Absolute event rates (overall cohort, per 1,000 person-years): MACE 1.5; arrhythmia 1.64; PE 0.84; pneumonia 3.40.
• Dose–response: Continuous analyses showed increasing risk with higher average daily ICS dose.

Secondary outcomes/analyses

• Nested case–control (Aurum and GOLD): Increased odds of composite CVD and pneumonia beginning ≥30–90 days after last ICS prescription; PE elevated in Aurum (power limited in GOLD).
• SCCS (cardiovascular disease): Risk increased within 60 days of starting ICS and remained elevated during exposure; returned to null after stopping.
• Negative controls: No association between ICS and urinary tract infection, cellulitis, abdominal pain, or forearm fracture; no association between citalopram (negative control exposure) and outcomes.
• By ICS molecule: No significant between-drug differences (fluticasone vs. beclomethasone; budesonide vs. beclomethasone), though pneumonia risk appeared numerically higher with fluticasone (underpowered).
• Effect modification: No modification by age or eosinophils; greater MACE risk among those without documented baseline CVD than those with known CVD.

Limitations

• Observational design; residual confounding possible despite IPTW, negative controls, and e-value analysis.
• Potential misclassification of asthma and outcomes (hospital-based diagnoses used to minimize this).
• No direct measures of medication adherence; average daily dose inferred from prescriptions.
• First 7 days post-ICS excluded to reduce protopathic bias, which may bias toward the null.
• Limited power for PE in smaller datasets and for between-molecule comparisons.
• No spirometry data available.

Funding

Supported by the National Institute of Health and Care Research (NIHR) award NIHR301726 and the NIHR Imperial Biomedical Research Centre. A.M. supported by the NIHR Applied Research Collaboration NW London. Funders had no role in study design, data collection, analysis, interpretation, writing, or publication decisions.

Citation

Bloom CI, Yang F, Hubbard R, Majeed A, Wedzicha JA. Association of Dose of Inhaled Corticosteroids and Frequency of Adverse Events. American Journal of Respiratory and Critical Care Medicine. 2025;211(1):54–63. Originally published online August 1, 2024. DOI: 10.1164/rccm.202402-0368OC. Open Access under CC BY 4.0.