Real-world characterisation of severe asthma in Greece: results from the Greek Severe Asthma Registry

Abstract

Background

Severe asthma (SA) imposes a disproportionate burden on patients and healthcare systems worldwide, yet region-specific data from Southern Europe remain scarce. The Greek Severe Asthma Registry (GSAR), an initiative of the Hellenic Thoracic Society, provides the first nationwide characterisation of severe asthma in Greece.

Methods

We analysed data cross-sectionally from 619 adult patients (median age: 56 years; 61.1% female) with severe asthma, recruited across 13 specialised hospital centres and 11 private respiratory practices between March 2018 and November 2021. Patients were stratified by age of asthma onset, biologic therapy use and maintenance oral corticosteroid (mOCS) dependence. Clinical, inflammatory and treatment characteristics were assessed, alongside exacerbation predictors.

Results

The study population exhibited substantial disease burden, with 434 patients (70.1%) experiencing ≥1 exacerbation and 149 (24.1%) requiring hospitalisation in the prior year. Late-onset asthma (60.7%) was associated with a faster progression to severe disease, higher exacerbation risk (79.8% versus 69.5%, p=0.004), and more frequent cardiometabolic comorbidities versus early-onset asthma. Despite better asthma control (Asthma Control Test ≥20: 61.8% versus 50.5%, p=0.012), biologically treated patients (57%) had more severe disease, including lower forced expiratory volume in 1 s (FEV₁) % predicted (73% versus 81%, p<0.001). Patients receiving mOCS (18.9%) exhibited poorer asthma control, higher exacerbation rates and more frequent corticosteroid-related comorbidities. In multivariable models, elevated blood eosinophil count (incidence rate ratio: 1.044 per 100 cells per μL) and reduced FEV₁% (odds ratio: 0.97 per 1% increase) were independent predictors of exacerbations.

Conclusions

This study highlights the phenotypic heterogeneity of severe asthma in Greece and reveals region-specific treatment patterns and unmet clinical needs despite access to advanced therapies. These real-world findings may guide national asthma care strategies and contribute to European initiatives aimed at optimising asthma management.

Shareable abstract

Real-world stratified analysis of severe asthma in Greece https://bit.ly/4b7zqAe

Introduction

Severe asthma affects 3–5% of the global asthma population, yet contributes disproportionally to disease burden, quality of life and healthcare resource utilisation [1]. Despite therapeutic advances, including biologics targeting specific inflammatory pathways, many patients still experience frequent exacerbations, poor symptom control, systemic corticosteroid dependence and treatment-related adverse effects [2, 3].

Severe asthma is clinically and biologically heterogeneous, encompassing distinct phenotypes and endotypes that differ in presentation, immunopathology and treatment response [4, 5]. Factors such as age of onset, biomarker profiles and comorbidities vary widely, influencing both disease trajectory and therapeutic outcomes [6, 7]. While inflammatory phenotypes are fundamental markers of disease course in the biologic era, age of asthma onset provides complimentary clinical stratification, linked to disease progression, exacerbation risk and clinical outcomes [8, 9]. Nevertheless, the interplay between clinical phenotypes and treatment outcomes, particularly in biologic-treated patients, remains inadequately understood in routine care [10].

Clinical registries provide valuable real-world data across diverse populations, capturing severe asthma complexity beyond clinical trials [11]. Several European countries have established national registries, yielding important insights [12–14]. However, data from Greece remain scarce, despite broad access to biologics and efforts to implement precision asthma care [15–17].

To address this gap, the Hellenic Thoracic Society launched the Greek Severe Asthma Registry (GSAR), aiming to collect nationwide data and characterise severe asthma. This study presents cross-sectional data regarding clinical, inflammatory and therapeutic characteristics across key patient subgroups within the GSAR, emphasising variations in disease expression and treatment patterns. As a practice-based registry, GSAR does not aim to provide epidemiological estimates but rather captures the real-world clinical spectrum of severe asthma in Greece, complementing other national registries. By addressing the underrepresentation of Southern European populations in current asthma literature, this work aims to offer region-specific insights to inform personalised severe asthma management.

Methods

This cross-sectional, multicentre, observational study analysed data provided by 13 hospital outpatient severe asthma clinics and 11 private respiratory practices across Greece between March 2018 and November 2021. Eligibility criteria included physician-confirmed severe asthma diagnosis according to European Respiratory Society (ERS)/American Thoracic Society (ATS) criteria [18]: age ≥18 years; availability of core clinical and functional data; and provision of informed consent. Patients were excluded if asthma was not clearly differentiated from other chronic respiratory conditions (e.g., COPD, bronchiectasis) or if core data were incomplete. Data were prospectively recorded using a standardised electronic form including demographics, clinical history, atopy, asthma control, biomarkers (blood eosinophil count (BEC); fractional exhaled nitric oxide; serum immunoglobulin E (IgE)), treatment patterns, healthcare utilisation and comorbidities (supplementary table S1). Variables were selected based on clinical relevance and previous studies. For patients on biologics enrolment, outcomes reflect their status while on treatment. Medical/surgical history and comorbidities were coded per Medical Dictionary for Regulatory Activities (MedDRA) version 26.0. The study was exempt from ethics approval, using only anonymised clinical data, according to national regulations.

Statistical analysis

Descriptive statistics summarised patient characteristics without imputation, categorised by age of asthma onset (early-onset asthma (EOA); late-onset asthma (LOA); UOA (unknown-onset asthma)), biologic use and maintenance oral corticosteroids (mOCS) use. Continuous variables were summarised as mean±sd, median (IQR) and range, categorical variables as counts and percentages. Annual event rates were compared using Wald test, with 95% confidence intervals (CI). Chi-squared or Fisher's exact tests provided p-values for categorical variables, as appropriate. One-way ANOVA calculated overall p-values across groups, with Tuckey's honestly significant difference for pairwise comparisons, and Pearson's coefficients for correlations. Analyses were restricted to EOA and LOA when UOA data were missing. A multivariable logistic regression model was used to evaluate sex, severe asthma duration, body mass index (BMI) and forced expiratory volume in 1 s (FEV₁%) impact on exacerbations likelihood the previous year. Statistical analyses were two-sided at 0.05, using SAS (v.9.4; SAS Institute, Cary, NC, USA) and GraphPad Prism software (Prism v9; GraphPad, San Diego, CA, USA).

Results

Study population

Overall, the analysis included 619 patients, with a median age of 56.0 (IQR 46.0–65.0) years and 61% being female. Median (interquartile range (IQR)) BMI was 28 kg·m⁻² (24.1–30.8), with 228 patients (36.8%) classified as overweight and 163 (26.3%) as obese. 46 patients (7.4%) were current, and 142 (22.9%) were former smokers, with a median smoking history of 20 pack-years (table 1).

TABLE 1.

Key demographic and clinical characteristics at baseline, overall and per timing of asthma onset

	Overall	EOA	LOA	UOA	Overall p-value	EOA versus LOA p-value	EOA versus UOA p-value	LOA versus UOA p-value
Patients, n	619	174	376	69
Sex, female	378 (61.1%)	99 (56.9%)	237 (63.0%)	42 (60.9%)	0.390	0.201	0.673	0.837
Age years	55.5±13.2	47.7±13.2	59.5±11.4	53.2±13.5	<0.001	<0.001	0.892	<0.001
BMI kg·m−2	28.0±5.8	27.2±5.3	28.5±6.0	26.6±5.8	<0.001	<0.001	0.798	0.039
Smoking status
Ever-smoker	188 (30.3%)	57 (32.8%)	121 (32.1%)	10 (14.5%)	0.010	0.971	0.007	0.005
Current smoker	46 (7.4%)	16 (9.2%)	28 (7.4%)	2 (2.9%)
Former smoker	142 (22.9%)	41 (23.6%)	93 (24.7%)	8 (11.6%)
Never-smoker	425 (68.7%)	115 (66.1%)	253 (67.3%)	57 (82.6%)
Unknown	6 (1.0%)	2 (1.1%)	2 (0.5%)	2 (2.9%)
Asthma duration years	24.7±14.5	35.5±13.8	19.7±11.8	na	<0.001	<0.001	na	na
Age (years) at first diagnosis of asthma	31.1±16.5	12.2±6.0	39.8±11.9	na	<0.001	<0.001	na	na
Age (years) at first diagnosis of severe asthma	45.1±14.1	35.5±12.6	49.5±12.5	55.3±12.0	<0.001	<0.001	<0.001	0.077
ACT score
Mean±sd	19.6±4.5	19.8±4.4	19.7±4.5	17.8±5.0	0.001	0.923	0.002	0.002
Well-controlled asthma (≥20 ACT score)	295 (47.7%)	90 (51.7%)	194 (51.6%)	11 (15.9%)	<0.001	1.000	<0.001	<0.001
Asthma exacerbations in the last year
Yes	434 (70.1%)	121 (69.5%)	300 (79.8%)	13 (18.8%)	<0.001	0.011	<0.001	<0.001
Mean±sd	1.9±1.8	1.9±1.9	2.1±1.8	0.5±1.1	<0.001	0.997	<0.001	<0.001
Number of asthma-related ED visits in the last year among all patients	0.5±1.0	0.6±1.1	0.5±1.1	0.1±0.5	<0.001	0.989	<0.001	<0.001
Number of asthma-related hospitalisations in the last year among all patients	0.5±1.2	0.4±0.7	0.6±1.4	0.1±0.5	<0.001	<0.001	0.073	<0.001
FEV1 % predicted, pre-BD	74.6±22.1	73.8±23.2	72.3±21.1	89.0±19.5	<0.001	0.537	<0.001	<0.001
FEV1/FVC (pre-BD)	71.2±12.8	71.2±13.9	70.2±11.9	76.8±13.2	0.005	0.984	0.013	0.004
Drug classes of asthma medications
ICS	592 (95.6%)	169 (97.1%)	368 (97.9%)	55 (79.7%)	<0.001	0.561	<0.001	<0.001
LABA	577 (93.2%)	168 (96.6%)	359 (95.5%)	50 (72.5%)	<0.001	0.722	<0.001	<0.001
SABA	456 (73.7%)	144 (82.8%)	272 (72.3%)	40 (58.0%)	<0.001	0.011	<0.001	0.024
Biologics and/or mOCS	391 (63.2%)	111 (63.8%)	263 (69.9%)	17 (24.6%)
Biologics	353 (57.0%)	103 (59.2%)	235 (62.5%)	15 (21.7%)	<0.001	0.518	<0.001	<0.001
Anti-IL-5 (mepolizumab)	198 (32.0%)	54 (31.0%)	139 (37.0%)	5 (7.2%)	<0.001	0.208	<0.001	<0.001
Anti-IgE (omalizumab)	119 (19.2%)	47 (27.0%)	62 (16.5%)	10 (14.5%)	<0.001	0.518	<0.001	<0.001
Anti-IL-5Ra (benralizumab)	36 (5.8%)	2 (1.1%)	34 (9.0%)	0 (0.0%)	<0.001	<0.001	na	na
LAMA	330 (53.3%)	83 (47.7%)	228 (60.6%)	19 (27.5%)	<0.001	0.021	<0.001	<0.001
Intranasal steroids	297 (48.0%)	93 (53.4%)	186 (49.5%)	18 (26.1%)	<0.001	0.255	<0.001	<0.001
Antileukotrienes	270 (43.6%)	83 (47.7%)	178 (47.3%)	9 (13.0%)	<0.001	1.000	<0.001	<0.001
Antihistamines	161 (26.0%)	56 (32.2%)	103 (27.4%)	2 (2.9%)	<0.001	0.293	<0.001	<0.001
mOCS	117 (18.9%)	26 (14.9%)	80 (21.3%)	11 (15.9%)	0.270	0.180	0.642	0.098
Theophylline	15 (2.4%)	2 (1.1%)	12 (3.2%)	1 (1.4%)	0.300	0.244	1.000	0.702

Data are presented as mean±sd, n (%) or n. EOA: early-onset asthma; LOA: late-onset asthma; UOA: unknown-onset asthma; BMI: body mass index; ACT: Asthma Control Test; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; pre-BD: pre-bronchodilation; IgE: immunoglobulin E; ICS: inhaled corticosteroids; LABA: long-acting β₂-agonists; SABA: short-acting β₂-agonists; mOCS: maintenance oral corticosteroids; LAMA: long-acting muscarinic antagonists; IL-5: interleukin-5; IL-5Ra: interleukin-5 receptor a.

Regarding medical history and comorbidities, 416 patients (67.2%) had atopic conditions, 344 (55.6%) allergic rhinitis and 150 (24.2%) nasal polyps, 84 of whom having undergone prior nasal surgery. 300 patients tested positive for ≥1 allergen (via skin-prick testing (SPT) and/or radioallergosorbent test), most commonly house dust mites (69%), Parietaria (46%), olive tree (32%), mixed grasses (28%) and cat dander (20%). Gastro-oesophegeal reflux disease (GORD) was reported by 174 patients (28.1%), anxiety or depression by 154 (24.9%) and sleep apnoea syndrome by 57 (9.2%). Other frequent comorbidities included hypertension (27.1%), dyslipidaemia (15.5%), thyroid disorders (14.9%) and diabetes mellitus (8.9%) (table 2).

TABLE 2.

Medical history/comorbidities of interest, overall and per timing of asthma onset

	Overall	EOA	LOA	UOA
Patients, n	619	174	376	69
With any atopic condition	416 (67.2%)	95 (54.6%)	171 (45.5%)	22 (31.9%)
Allergic rhinitis	344 (55.6%)	109 (62.6%)	216 (57.4%)	19 (27.5%)
Eczema	68 (11.0%)	27 (15.5%)	37 (9.8%)	4 (5.8%)
Nasal polyps	150 (24.2%)	34 (19.5%)	106 (28.2%)	10 (14.5%)
Prior nasal surgery	84 (13.6%)	17 (9.8%)	62 (16.5%)	5 (7.2%)
History of GORD	174 (28.1%)	46 (26.4%)	125 (33.2%)	3 (4.3%)
Screened for osteoporosis	206 (33.3%)	51 (29.3%)	152 (40.4%)	3 (4.3%)
Osteopenia (% of screened)	71 (34.5%)	13 (25.5%)	58 (38.2%)	–
Osteoporosis (% of screened)	70 (34.0%)	17 (33.3%)	50 (32.9%)	3 (100.0%)
Anxiety/depression syndrome	154 (24.9%)	46 (26.4%)	104 (27.7%)	4 (5.8%)
Sleep apnoea syndrome	57 (9.2%)	13 (7.5%)	44 (11.7%)	–
Aspirin hypersensitivity	37 (6.0%)	14 (8.0%)	21 (5.6%)	2 (2.9%)
Other medical conditions	297 (48.0%)	66 (37.9%)	219 (58.2%)	12 (17.4%)
Vascular disorders	170 (27.5%)	30 (17.2%)	132 (35.1%)	8 (11.6%)
Hypertension	168 (27.1%)	29 (16.7%)	131 (34.8%)	8 (11.6%)
Metabolism and nutrition disorders	121 (19.5%)	22 (12.6%)	96 (25.5%)	3 (4.3%)
Dyslipidaemia	96 (15.5%)	17 (9.8%)	77 (20.5%)	2 (2.9%)
Diabetes mellitus	55 (8.9%)	10 (5.7%)	43 (11.4%)	2 (2.9%)
Endocrine disorders	94 (15.2%)	27 (15.5%)	62 (16.5%)	5 (7.2%)
Thyroid disorder	92 (14.9%)	26 (14.9%)	62 (16.5%)	4 (5.8%)
Cardiac disorders	49 (7.9%)	4 (2.3%)	44 (11.7%)	1 (1.4%)
Coronary artery disease	36 (5.8%)	3 (1.7%)	32 (8.5%)	1 (1.4%)
Atrial fibrillation	13 (2.1%)	2 (1.1%)	10 (2.7%)	1 (1.4%)
Respiratory, thoracic and mediastinal disorders	40 (6.5%)	16 (9.2%)	24 (6.4%)	–
Bronchiectasis	30 (4.8%)	13 (7.5%)	17 (4.5%)	–
COPD	5 (0.8%)	1 (0.6%)	4 (1.1%)	–
Musculoskeletal and connective tissue disorders	18 (2.9%)	5 (2.9%)	12 (3.2%)	1 (1.4%)
Neoplasms (benign/malignant/unspecified)	16 (2.6%)	4 (2.3%)	11 (2.9%)	1 (1.4%)
Infections and infestations	11 (1.8%)	3 (1.7%)	8 (2.1%)	–
Sinusitis	3 (0.5%)	2 (1.1%)	1 (0.3%)	–
Immune system disorders	8 (1.3%)	1 (0.6%)	7 (1.9%)	–
Nervous system disorders	8 (1.3%)	3 (1.7%)	5 (1.3%)	–
Eye disorders	6 (1.0%)	2 (1.1%)	4 (1.1%)	–
Cataract	4 (0.6%)	2 (1.1%)	2 (0.5%)	–
Glaucoma	4 (0.6%)	1 (0.6%)	3 (0.8%)	–
Gastrointestinal disorders	6 (1.0%)	1 (0.6%)	5 (1.3%)	–
Reproductive system/breast disorders	6 (1.0%)	3 (1.7%)	3 (0.8%)	–

Data are presented as n (%). EOA: early-onset asthma; LOA: late-onset asthma; UOA: unknown-onset asthma; GORD: gastro-oesophageal reflux disease.

Most patients (60.7%) had LOA, defined as onset at ≥18 years of age, with median (IQR) disease duration of 25 years (13–33) and severe asthma duration of 8 years (3–17). Despite broad access to biologics (57% of patients), 19% remained mOCS-dependent. Fewer than half (47.7%) had an Asthma Control Test (ACT) score ≥20, and overall disease burden was high: 70.1% experienced at least one exacerbation and 31.3% had ≥3 exacerbations the preceding year. Asthma-related emergency visits occurred in 30.4%, hospitalisations in 24.1% and intensive care unit admissions in 3.4% of patients (tables 1 and 2).

Clinical and treatment differences by age of asthma onset

Table 1 summarises demographic and clinical characteristics stratified by age of asthma onset, and figure 1 by onset, biologic use, and OCS dependence. Compared with EOA, patients with LOA were diagnosed with severe disease at an older median age (50 versus 36 years, p<0.001), with a shorter interval from initial asthma diagnosis (median time: 10 versus 23.0 years, p<0.001). LOA was associated with a higher prevalence of heavy smoking (≥20 pack-years: 19.2% versus 11.0%, p=0.017) and a slightly higher mean BMI (27.3 kg·m⁻² versus 26.5 kg·m⁻², p=0.014). Despite comparable ACT scores and lung function parameters, patients with LOA experienced more frequent exacerbations (79.8% versus 69.5%, p=0.004) and higher asthma-related hospitalisation rates (0.60 versus 0.40, p<0.001). They also required higher median doses of inhaled corticosteroids (ICS) (1600 µg (1000–2000) versus 1500 µg (800–2000) beclomethasone-equivalent, p=0.036) and more frequent use of long-acting muscarinic antagonists (60.6% versus 47.7%, p=0.004).

FIGURE 1.

Severe asthma profiles by onset, biologic use and OCS dependence in the Greek Severe Asthma Registry (GSAR). Distribution of patients with severe asthma (n=619) according to age of asthma onset (early-onset asthma (EOA), late-onset asthma (LOA) and unknown-onset asthma (UOA)), stratified by biologic treatment status and maintenance oral corticosteroid (OCS) use. Bars represent total numbers within each onset category. Each bar is subdivided according to treatment category: biologic treatment, OCS treatment and no biologic treatment.

Comorbidity patterns also differed (table 2). EOA was more often associated with atopic conditions, including allergic rhinitis, eczema, anaphylaxis and food allergy; LOA showed higher prevalence of nasal polyps and systemic comorbidities, particularly cardiometabolic disorders, including arterial hypertension, cardiovascular disease, diabetes mellitus, and dyslipidaemia. Notably, LOA patients had a higher multimorbidity burden, with 60.2% having ≥1 nonrespiratory comorbidity, compared to 39.3% in EOA.

Biologic therapy use mirrored these patterns: omalizumab was more frequently prescribed in EOA (27.0% versus 16.5%, p=0.004), consistent with predominant allergic sensitisation, whereas anti-interleukin-5(IL-5)/IL-5Receptor(R) antagonists were more commonly used in LOA (9.0% versus 1.1%, p<0.001). Interestingly, BEC and serum total IgE levels showed no significant differences between groups (table 3).

TABLE 3.

Blood eosinophil count (BEC), fractional exhaled nitric oxide (F_ENO) and IgE, overall and per timing of asthma onset

	Overall	EOA	LOA	UOA
Patients, n	619	174	376	69
BEC %	4.9±4.2	4.1±3.3	5.3±4.6	5.3±4.4
Absolute BEC cells·μL−1
Mean±sd	377.7±361.7	365.5±404.0	383.7±341.0	374.5±369.0
<150 cells·μL−1	134 (29.8%)	39 (29.5%)	89 (31.1%)	6 (19.4%)
150 to <300 cells·μL−1	101 (22.5%)	37 (28.0%)	55 (19.2%)	9 (29.0%)
≥300 cells·μL−1	214 (47.7%)	56 (42.4%)	142 (49.7%)	16 (51.7%)
FENO ppb
Mean±sd	37.0±47.3	39.3±74.5	37.8±32.0	23.2±19.0
<25 ppb	82 (47.4%)	26 (53.1%)	46 (42.2%)	10 (66.7%)
25 to <50 ppb	50 (28.9%)	14 (28.6%)	34 (31.2%)	2 (13.3%)
≥50 ppb	41 (23.7%)	9 (18.4%)	29 (26.6%)	3 (20.0%)
IgE IU·mL−1
Mean±sd	298.0±471.1	346.7±558.3	276.7±432.1	252.1±280.9
<100 IU·mL−1	167 (37.9%)	49 (34.5%)	110 (39.7%)	8 (36.4%)
≥100 IU·mL−1	274 (62.1%)	93 (65.5%)	167 (60.3%)	14 (63.6%)

Data are presented as mean±sd or n (%). EOA: early-onset asthma; LOA: late-onset asthma; UOA: unknown-onset asthma; IgE: Immunoglobulin E.

Impact of biologic therapy

Stratification by biologic use at registry entry revealed clinically meaningful differences, with outcomes reflecting the preceding 12 months (table 4). Patients receiving biologics were slightly older (median age 57.0 versus 54.0 years, p=0.002), with a longer interval from asthma onset to severe asthma diagnosis (median time 11.0 versus 9.0 years, p=0.006).

TABLE 4.

Patient characteristics per biologic use

	Biologic-treated	Non-biologic-treated	p-value of comparison
Sex (female)	215 (61.1%)	163 (61.3%)	0.960
Age at study visit years	57.0 (48.0–67.0)	54.0 (43.0–63.0)	0.002
BMI kg·m−2	27.0 (24.5–31.7)	26.8 (24.0–29.3)	0.096
BMI <30 kg·m−2	232 (68.4%)	180 (76.3%)	0.040
BMI ≥30 kg·m−2	107 (31.6%)	56 (23.7%)
Age (years) at first diagnosis of severe asthma	50.0 (39.0–59.0)	40.0 (30.0–50.0)	<0.001
Asthma duration years	20.5 (10.0–36.0)	27.0 (20.0–30.0)	0.011
Time (years) from first diagnosis of asthma to severe asthma diagnosis	11.0 (5.0–25.0)	9.0 (4.0–19.0)	0.006
ACT score	21.0 (17.0–23.0)	20.0 (17.0–25.0)	0.491
Well controlled (≥20 ACT score)	196 (61.8%)	99 (50.5%)	0.012
Occurrence of asthma exacerbations in the last year	289 (83.0%)	145 (55.8%)	<0.001
Rate of exacerbations in last year per patient year (95% CI)	2.24 (2.09–2.41)	1.43 (1.29–1.58)	<0.001
Rate of medical appointments in last year per patient year (95% CI)	1.77 (1.63–1.92)	1.07 (0.95–1.20)	<0.001
Asthma-related ER visits	103 (29.9%)	85 (32.3%)	0.515
Sick leave due to asthma in the last year among employed patients	69 (40.8%)	31 (25.8%)	0.008
FEV1 % predicted (pre-BD)	73.0 (57.0–85.0)	81.0 (61.0–99.0)	<0.001
<80% predicted	190 (65.5%)	111 (48.5%)	<0.001
≥80% predicted	100 (34.5%)	118 (51.5%)
FEV1/FVC (pre-BD)	71.0 (62.0–79.0)	73.5 (62.0–82.0)	0.048
Absolute BEC cells·μL−1	220.0 (100.0–515.0)	300.0 (180.0–510.0)	0.094
IgE IU·mL−1	180.0 (76.0–417.0)	105.0 (32.5–197.0)	<0.001
ICS dose μg BDP equivalent per day	1280 (920.0–2000)	1600 (1000–2000)	0.035
≤1000 μg BDP equivalent per day	146 (43.3%)	62 (29.0%)	<0.001
>1000 μg BDP equivalent per day	191 (56.7%)	152 (71.0%)
Current treatment with mOCS	79 (22.6%)	38 (18.3%)	0.221
Presence of any atopic medical condition (atopy, allergic rhinitis, eczema, anaphylaxis, food allergy)	263 (75.6%)	153 (60.2%)	<0.001
History of atopy	202 (58.9%)	86 (33.6%)	<0.001
Allergic rhinitis	223 (64.1%)	121 (46.9%)	<0.001
Eczema	50 (14.4%)	18 (7.3%)	0.007
Nasal polyps	108 (31.5%)	42 (16.4%)	<0.001
Prior nasal surgery	62 (18.0%)	22 (8.6%)	<0.001
History of GORD	116 (33.4%)	58 (22.4%)	0.003
Cardiac disorders	38 (11.2%)	11 (4.7%)	0.006
Vascular disorders	105 (30.9%)	65 (27.7%)	0.405
Hypertension	104 (30.6%)	64 (27.2%)	0.385
Diabetes	34 (10.0%)	21 (8.9%)	0.670
Dyslipidaemia/hypercholesterolaemia	68 (20.0%)	29 (12.3%)	0.016
Osteoporosis	43 (32.1%)	27 (38.6%)	0.355

Data are presented as median (IQR) or n (%) unless indicated otherwise. Bold type for p-values indicates statistical significance. BMI: body mass index; ACT: Asthma Control Test; ER: emergency room; FEV1: forced expiratory volume in 1 s, pre-BD: post bronchodilation; FVC: forced vital capacity; BEC: blood eosinophil count; IgE: Immunoglobulin E; ICS: inhaled corticosteroids; mOCS: maintenance oral corticosteroids; GORD: gastro-oesophageal reflux disease.

Although asthma control was better in the biologic group (ACT ≥20: 61.8% versus 50.5%, p=0.012), overall disease burden remained high, with patients reporting more frequent exacerbations in the previous year (83.0% versus 55.8%, p<0.001), greater healthcare utilisation (1.77 versus 1.07 asthma-related medical appointments/year, p<0.001), and increased rates of asthma-related sick leave (40.8% versus 25.8%, p=0.008). Lung function was also more impaired despite advanced therapy, with lower FEV₁% (73% versus 81%, p<0.001), lower FEV₁/forced vital capacity (FVC) ratios (71.0 versus 73.5, p=0.048), and more patients with FEV₁<80% (65.5% versus 48.5%, p<0.001). Biologic-treated patients also showed higher rates of Type 2(T2)-related comorbidities, including allergic rhinitis (64.1% versus 46.9%, p<0.001), eczema (14.4% versus 7.3%, p=0.007) and nasal polyps (31.5% versus 16.4%, p<0.001), as well as systemic disorders, such as GORD (33.4% versus 22.4%, p=0.003), cardiac disease (11.2% versus 4.7%, p=0.006) and dyslipidaemia (20.0% versus 12.3%, p=0.016).

A larger proportion of biologic-treated patients were receiving lower ICS doses (≤1000 μg·day⁻¹: 43.3% versus 29.0%, p<0.001), possibly reflecting steroid-sparing strategies or self-de-escalation after improvement; however, mOCS use remained similar between groups.

Characteristics of biologic switchers

Among the 353 patients receiving biologic therapy, 21.8% had switched biologics at least once (supplementary table S2), with notable differences between those who switched and those who did not. Switchers were diagnosed with severe asthma at a younger median age (45.0 versus 50.0 years, p=0.029), progressed faster to severe disease (median time: 9.0 versus 11.0 years, p=0.039) and had a higher prevalence of allergic comorbidities, particularly allergic rhinitis (74.0% versus 61.3%, p=0.039) and elevated serum IgE levels (>100 IU·mL⁻¹: 74.0% versus 61.3%, p=0.039). They also demonstrated more frequent exacerbations (94.8% versus 79.7%, p<0.001) and mOCS use (37.7% versus 18.4%, p<0.001), indicating a severe, treatment-refractory phenotype (supplementary table S3).

Profile of patients on mOCS

Patients requiring mOCS exhibited significantly poorer asthma control (median ACT scores 18.0 versus 21.0, p<0.001), higher annual exacerbation rates (rate ratio (RR): 2.63 (2.35–2.95) versus 1.89 (1.77–2.03), p<0.001), more asthma-related hospitalisations (RR: 1.12 (0.94–1.34) versus 0.38 (0.33–0.44), p<0.001) and higher work absenteeism (30.2% versus 13.2%, p=0.003), despite more intensive ICS therapy (supplementary table S4). Lung function was also more impaired, with lower mean FEV₁% (66.1% versus 74.1%, p<0.001) and FEV₁/FVC ratios (64.5 versus 72.0, p<0.001). Notably, they demonstrated lower median serum IgE levels, fewer atopic comorbidities and more corticosteroid-related conditions (e.g. GORD and osteoporosis) (supplementary table S5), suggesting a severe, nonatopic phenotype, characterised by persistent airflow limitation and systemic morbidity.

Predictors of exacerbation risk

Multivariable regression analysis identified BEC as an independent predictor of increased annual exacerbation rates (incidence rate ratio (IRR) per 100 cells/μL increase: 1.044, p<0.001) (supplementary table S6). In contrast, better lung function was protective, with each 1% increase in FEV₁% reducing exacerbation risk by 0.8% (IRR: 0.992, p<0.001). In a separate multivariable logistic regression model, higher FEV₁% remained significantly protective against experiencing ≥1 exacerbation (odds ratio (OR): 0.97; 95% CI: 0.95–0.99, p<0.001), while male sex was independently associated with reduced risk (OR: 0.44; 95% CI: 0.22–0.89, p=0.022) (supplementary table S7). In contrast, higher BEC showed a borderline association with increased exacerbation risk (OR: 1.14; 95% CI: 1.00–1.30, p=0.053) (figure 2).

FIGURE 2.

Predictors of experiencing ≥1 severe asthma exacerbation in the prior 12 months. Odds ratios (ORs) with 95% confidence intervals are shown from a multivariable logistic regression model (n=359). The model included duration of severe asthma (years), sex (male versus female), body mass index (BMI, kg·m⁻²), absolute blood eosinophil count (BEC, per 100 cells/μL) and pre-bronchodilator forced expiratory volume in 1 s (FEV₁) % predicted. An OR >1 indicates increased odds of exacerbation, while an OR <1 indicates reduced odds. In this cohort, lower FEV₁ and higher blood eosinophil counts were associated with higher exacerbation risk, whereas male sex was associated with reduced risk.

Discussion

This national, registry-based, cross-sectional study provides the most comprehensive real-world overview of severe asthma in Greece to date. By capturing detailed clinical, functional and treatment-related data across a real-world population, it contributes to the growing body of evidence on severe asthma heterogeneity in routine care. Our findings broadly align with patterns reported in other national registries, while offering novel insights into disease expression and management practices within the Greek healthcare setting.

Consistent with observations from other large-scale registries, we demonstrated a predominance of LOA and female sex among severe asthma patients [12–14, 19–22]. Distinct clinical and comorbidity profiles between EOA and LOA highlight the clinical value of age at asthma onset for phenotypic differentiation [23]. Patients with LOA experienced faster progression from initial diagnosis to severe asthma, indicating an intrinsically more aggressive disease, delayed recognition or the contribution of comorbidities, aligning with recent evidence of compressed disease trajectories in this phenotype [24]. Despite comparable lung function and patient-reported control, LOA was associated with higher exacerbation and hospitalisation, rates indicating that conventional clinical assessments may underestimate disease burden in this group [25]. The higher prevalence of obesity and cardiometabolic comorbidities in LOA aligns with prior reports [26, 27], potentially reflecting reduced physical activity, shared systemic inflammatory pathways, ageing or the cumulative impact of corticosteroid exposure. Interestingly, the lack of significant differences in BEC or total serum IgE levels between EOA and LOA contradicts traditional endotype-based expectations [28]. This finding may reflect overlapping endotypes, the impact of corticosteroid exposure or both, highlighting the limitations of relying solely on inflammatory biomarkers for phenotypic differentiation [29].

Patterns of biologic use reflect the complexity of managing severe asthma in real-world practice. Although biologic-treated patients demonstrated better asthma control, they exhibited greater overall disease burden, with more frequent exacerbations, persistent airflow limitation and higher healthcare utilisation. This discrepancy likely arises from the use of biologics in patients with more advanced or treatment-refractory disease and established airway remodeling, which may attenuate therapeutic benefits [30]. However, beyond baseline severity, poorer outcomes in biologic-treated patients may also relate to longer disease duration before biologic initiation, prolonged corticosteroid exposure, comorbidities, delayed specialist referral, and potential healthcare access or socioeconomic disparities, all of which can independently sustain disease burden despite advanced therapy [30–32].

These observations support emerging evidence advocating for earlier biologic initiation to prevent irreversible structural changes and lung function decline, particularly in patients with frequent exacerbations or T2-high inflammation [30, 33, 34]. The continued use of mOCS in nearly one-fifth of biologic-treated patients underscores the limitations of current treatment strategies and the ongoing need for more effective steroid-sparing interventions [35, 36]. Furthermore, the clustering of T2-related and systemic comorbidities within this population supports the concept of a complex, multimorbid asthma phenotype that warrants multidisciplinary management [37].

Patients requiring biologic switching represent a particularly complex, treatment-refractory phenotype. In our cohort, switchers were diagnosed with asthma at a younger age and had higher serum IgE levels and more frequent allergic comorbidities, indicating a predominantly allergic profile. They also experienced more frequent exacerbations and required higher mOCS doses, potentially reflecting suboptimal initial phenotyping, persistent disease activity or steroid dependence [38, 39]. However, these findings should be interpreted within the study period context, when treatment options were limited mainly to omalizumab and mepolizumab, with benralizumab only recently introduced. Therefore, the need for switching may reflect, at least in part, limited initial therapeutic options rather than intrinsic nonresponsiveness to biologic treatment. Although this registry did not capture reasons for switching, initial biologic treatment duration or specific switching patterns, the data reflect real-world practice in Greece.

Patients on long-term mOCS represent a particularly vulnerable subgroup, experiencing significantly worse clinical outcomes, including poorer symptom control, higher exacerbation rates and lower lung function. Their comorbidity profile included a higher prevalence of GORD, osteoporosis and other systemic conditions, along with fewer atopic features. The UK Severe Asthma Registry (UKSAR) [13] and Severe Asthma Network in Italy (SANI) registries [14] have reported similar findings, identifying persistently high mOCS use despite broad access to biologics. This reliance on systemic corticosteroids likely reflects both persistent disease activity and care gaps, including suboptimal efficacy of current therapies in certain non-T2 endotypes, under-recognition of noninflammatory treatable traits and limited options for patients with fixed airflow obstruction [40]. These findings also highlight the importance of accurate phenotypic characterisation and timely response assessment in biologic selection and adjustment. As the therapeutic landscape expands, more individualised treatment algorithms will be critical to optimise long-term outcomes [41]. Addressing these unmet needs requires holistic management strategies that consider comorbidities, psychological burden and quality-of-life impairment.

In addition to phenotypic characterisation, this study identified elevated BEC, reduced FEV₁ and female sex as independent predictors of exacerbation risk. These findings align with previous reports demonstrating the prognostic importance of inflammatory, functional and demographic factors in severe asthma [33, 42, 43]. The observed protective effect of male sex corroborates earlier findings of lower exacerbation rates and reduced healthcare utilisation among men compared to women [43–45]. Taken together, these findings support incorporating routine clinical parameters into risk assessment models for personalised disease management.

This study has several strengths. GSAR registry employes a standardised nationwide platform that enhances internal validity and ensures consistency, supporting generalisability across diverse clinical settings. By investigating clinically relevant subgroups and treatment patterns, this analysis provides better understanding of phenotypic heterogeneity, treatment response and care gaps. Together these features position GSAR as a valuable complement to other European registries, providing real-world insights that can inform both clinical practice and health policies.

Beyond mirroring patterns observed in other national registries, our findings highlight distinct characteristics within the Greek severe asthma population. Multimorbidity, particularly cardiometabolic diseases among patients with LOA, indicates a substantial, additional burden. Despite the widespread availability of biologics, nearly one-fifth of patients continued mOCS, similar to other Southern European cohorts [14]. These findings underscore the importance of national registries in capturing the local context of severe asthma, complementing pan-European initiatives such as SHARP CRC (Severe Heterogeneous Asthma Registry, Patient-centred Clinical Research Collaboration) [21].

Limitations should also be acknowledged. The cross-sectional design limits causal inference and precludes longitudinal assessment. Variability in local clinical practices and potentially missing or inaccurately recorded data may have introduced bias. Age of asthma onset was patient-reported and, although confirmed where possible through medical history, recall bias may exist. However, consistency with previous studies supports this variable's robustness. Furthermore, data reflect an earlier treatment landscape, with limited representation of newer biologics such as dupilumab and tezepelumab. It should also be noted that this registry does not capture longitudinal outcomes before and after biologic initiation; inadequate response was inferred from cross-sectional markers such as persistent exacerbations, mOCS dependence and biologic switching. Moreover, the absence of data on biologic treatment duration limits interpretation of treatment response in relation to exposure time. Future registry updates with longitudinal follow-up will allow better evaluation of biologic response. Finally, selection bias may also exist, as the registry predominantly includes patients referred to specialised centres, limiting generalisability to the entire Greek severe asthma population. However, our findings still provide a real-world picture of clinical practice, in line with the scope of other European registries.

In the current era of biologics, severe asthma registries such as GSAR continue to play an important dual role. They provide real-world evidence regarding treatment patterns, OCS dependence and biologic switching, which remain insufficiently captured in clinical trials. Additionally, by aggregating clinical and biological data, they contribute to a better understanding of disease heterogeneity, comorbidity burden and unmet needs, thereby supporting the development of targeted therapeutic strategies and informing healthcare policies.

Conclusion

This study offers a comprehensive overview of severe asthma in Greece, contributing data largely consistent with findings from larger European cohorts. At the same time, it elucidates region-specific patterns and healthcare dynamics that reflect local practices and potential unmet needs, thereby assisting both clinical decision-making and health policies at a national level. These insights contribute to European real-world evidence and emphasise the importance of context-specific data in shaping optimal asthma care.