Start here: what asthma is
Asthma is chronic inflammation of the airways, accompanied by airflow obstruction that varies — over hours, days and seasons — and by airways that overreact to things a normal airway ignores: cold air, exercise, smoke, allergens, a virus. The symptoms follow from that: wheeze, cough, chest tightness, breathlessness, characteristically worse at night and in the early morning.
The word that does the work is "variable." In an asthmatic airway the obstruction is largely reversible — it opens again, either spontaneously or within minutes of an inhaled bronchodilator. That single property is the diagnostic signature, and it is what separates asthma from its sibling in this series. In COPD the obstruction is largely fixed: it does not fully reverse, and the disease is defined by a persistent airflow limitation that accumulates. The two conditions can look similar at the bedside and can overlap in the same patient, and the airway inflammation underlying fixed obstruction genuinely differs between asthma and COPD [1].
The scale. Asthma affects roughly 5–10% of the population [2], and it is the most common chronic disease of childhood. The International Study of Asthma and Allergies in Childhood tracked worldwide trends in symptom prevalence across three phases and dozens of countries [3], and the global burden was summarised for the GINA dissemination committee [4]. The European Community Respiratory Health Survey did the equivalent for adults [5].
Why it matters. Asthma is usually controllable and not usually curable, and those two facts together define the whole management problem. People feel well between attacks, so treatment that prevents attacks feels unnecessary — while treatment that relieves an attack feels essential. That intuition is backwards, and correcting it is the single most consequential thing this review can do. The consequences of getting it wrong are not trivial: asthma still kills, and low-dose inhaled corticosteroids are associated with a reduced risk of death from asthma [6].
Why it is in the news. Two shifts. Severe asthma acquired biologics — monoclonal antibodies against specific inflammatory pathways — which changed the outlook for the minority of patients who remained uncontrolled on inhalers. And the field reversed a century-old habit by concluding that the reliever inhaler should itself contain an anti-inflammatory [7] [8].
Three pillars follow — measurements, medicines, and progress — with a dose-response model between the first two that explains why more steroid is usually not the answer.
Pillar 1: measurements and diagnosis
Spirometry and bronchodilator reversibility: the hallmark
Spirometry measures how much air someone can force out and how fast. In obstruction, the ratio of the volume exhaled in the first second (FEV₁) to the total (FVC) falls. The asthma-specific step is to repeat it: give an inhaled bronchodilator, wait, and measure again. A substantial improvement is bronchodilator responsiveness, and it is the closest thing asthma has to a positive test. Interpretation of spirometry, including the criteria for bronchodilator responsiveness, is set out in joint ERS/ATS technical standards [9] [10].
The European Respiratory Society's diagnostic guideline puts these in an explicit order: spirometry first, followed by bronchodilator reversibility testing if obstruction is present. If initial spirometry shows no obstruction — common, because asthma is variable and the patient may be well that morning — the next tests are exhaled nitric oxide, peak expiratory flow variability, or, in secondary care, bronchial challenge [2].
Two honest caveats belong here. Reversibility is a continuum, not a switch: measured in COPD patients it was normally distributed, the absolute change in FEV₁ was independent of the pre-bronchodilator value while the percentage change correlated with it, and response classification was not reliably reproducible between visits [11]. Responsiveness also declines with age [12], and it is partly genetic — β₂-adrenoceptor polymorphisms are associated with the albuterol response in children [13]. A negative test does not exclude asthma.
Peak flow, and variability over time
A peak flow meter is a cheap handheld device measuring the fastest exhalation someone can produce. One reading means little; a diary of readings captures the thing that defines asthma, which is variation — typically worse overnight, better after a bronchodilator, worse during an exacerbation. Peak flow variability is one of the ERS algorithm's confirmatory tests when spirometry is normal [2], and it appears as an outcome measure throughout the trial literature [14].
Bronchial challenge
If neither reversibility nor variability settles it, the airway can be provoked. Inhaling increasing doses of methacholine (or histamine, exercise, or mannitol) and measuring the dose that drops FEV₁ by a set amount quantifies airway hyperresponsiveness directly. Standardised challenge testing protocols exist for exactly this [15]. This is the most sensitive route to a diagnosis and the least convenient, which is why it sits last in the algorithm [2].
FeNO and eosinophils: biomarkers of type-2 inflammation
The newer measurements do something different: rather than confirming asthma, they ask which kind of asthma, and therefore which drug.
Airway inflammation in asthma is heterogeneous, and the dominant axis is whether type-2 inflammation is present. Gene-expression analysis of airway epithelial brushings from 42 patients and 28 healthy controls separated asthma into "Th2-high" and "Th2-low" subphenotypes that differed in cytokine expression, markers of inflammation and remodelling, and — crucially — in responsiveness to inhaled corticosteroids [16]. Type-2 inflammation is now understood as a shared mechanism across asthma and other airway diseases [17].
Fractional exhaled nitric oxide (FeNO) is a breath test that rises with type-2 airway inflammation, and blood eosinophil count is a simple blood test that does the same. The ERS guideline recommends FeNO in its algorithm and separately recognises the value of measuring blood eosinophils and serum IgE to phenotype the patient [2]. Induced sputum cell counts, the older research method for measuring airway inflammation directly, established the underlying approach [18].
These matter because they select who will benefit from the biologics in Pillar 2 — none of which work well in patients without the pathway they block.
Control versus severity
One conceptual distinction is worth stating because it is constantly muddled. Control is how the patient is doing now: symptoms, night waking, reliever use, activity limitation. Severity is how much treatment it takes to achieve control. A patient on high-dose therapy with no symptoms has severe but well-controlled asthma; a patient on nothing with daily symptoms has uncontrolled asthma of unknown severity. Only the second needs more drug. Surveys of European patients found asthma control was substantially worse than patients themselves believed it to be [19].
Centerpiece: a simple simulatable model of the controller dose-response
The central practical question in asthma is what to do when someone is not controlled. The intuitive answer is more steroid. The evidence says the curve has already flattened.
A meta-analysis of eight placebo-controlled randomised trials in 2,324 adolescents and adults examined the dose-response relation of inhaled fluticasone across FEV₁, morning and evening peak flow, night awakenings, β-agonist use and major exacerbations. It reported the curve in an unusually usable form: the raw dose-response "began to reach a plateau at around 100–200 µg/day and peaked by 500 µg/day", and a negative exponential model indicated that 80% of the benefit at 1000 µg/day was achieved at 70–170 µg/day and 90% by 100–250 µg/day [14].
Take the source's own functional form:
f(d) = 1 − exp(−k·d)
and fit the single rate constant k to one of those statements — the 80% anchor. Everything else the paper says then becomes a test the model either passes or fails.
What the model explains. Four things.
First, why the model is worth trusting at all. It was given one number and independently reproduced three others. The 90%-of-benefit dose lands at 172 µg/day, inside the published 100–250 µg/day range that played no part in the fit. That is not a decorative curve; it is the source's own model reconstructed and checked.
Second, why escalating the steroid is usually the wrong move. By 200 µg/day you have 93% of the benefit available at 1000 µg/day. Going from 500 to 1000 buys under one percentage point, while systemic exposure keeps rising. The paper draws the conclusion itself: this "partially explains why adding a long acting β agonist to inhaled corticosteroids is more efficacious than increasing the dose of inhaled steroid beyond this dose range" [14].
Third, what the trial evidence says to do instead. FACET randomised 852 patients already on inhaled glucocorticoids to a 2×2 of low or high budesonide, each with or without formoterol, for a year. Adding formoterol to the low dose cut severe exacerbations by 26% and mild ones by 40%; quadrupling the budesonide instead cut them by 49% and 37%; doing both cut them by 63% and 62% [20]. Two things follow. The two levers are not redundant — the combination beat either alone. And on mild exacerbations, adding the bronchodilator (40%) actually beat quadrupling the steroid (37%). That is the empirical basis for combination inhalers, supported by the pharmacological rationale for combining the two classes [21].
Fourth, why none of this makes the reliever the treatment. The controller curve is about preventing obstruction. A bronchodilator relieves obstruction that has already happened and does nothing to the inflammation causing it — which is why the paradigm shift in Pillar 2 was to put a steroid inside the reliever rather than to use more bronchodilator.
What the model deliberately does not do. Its y-axis is a pooled abstraction across outcomes with different units, so it cannot predict any specific measurement. It is fluticasone-specific and does not transfer across molecules or devices, where deposition differs substantially [22]. It describes an average when steroid responsiveness is precisely what differs between the Th2-high and Th2-low phenotypes [16], and some patients respond poorly for identifiable reasons [23] — smoking among them [24]. And a flat average dose-response is entirely compatible with individual patients who genuinely need more.
Pillar 2: medicines — the inhaler logic
Two drugs, two receptors, two jobs
Almost all asthma inhalers contain one or both of two drug classes, and the public confusion about asthma comes from not knowing which does what.
Bronchodilators are β₂-agonists. They bind the beta-2 adrenergic receptor on airway smooth muscle, raise cyclic AMP and relax the muscle, opening the airway within minutes [25] [26]. They do nothing to inflammation. Short-acting agents (salbutamol/albuterol) work for hours; long-acting agents (salmeterol, formoterol) for around twelve, with formoterol distinctive in having a fast onset as well as a long duration, which is what makes the single-inhaler strategy below possible [27].
Inhaled corticosteroids (ICS) act on the glucocorticoid receptor to suppress the airway inflammation that produces the hyperresponsiveness in the first place. They do not relieve an attack in progress. They prevent the next one, and low-dose ICS is associated with reduced asthma death [6].
The names given to these in practice — "reliever" and "preventer" — capture the pharmacology but have historically encouraged exactly the wrong behaviour: patients feel the bronchodilator work and do not feel the steroid work, so adherence to the drug that matters is poor and reliance on the one that does not is high.
The shift: an anti-inflammatory reliever
The correction was to stop offering a bronchodilator alone as the reliever. Because formoterol acts fast enough to relieve symptoms and is a long-acting agent that can also be used for maintenance, budesonide–formoterol in a single inhaler can serve as both controller and reliever — the maintenance-and-reliever (MART/SMART) approach [28].
The evidence in mild asthma is striking. SYGMA 1 randomised 3,849 patients to as-needed terbutaline, as-needed budesonide–formoterol, or maintenance budesonide plus as-needed terbutaline for 52 weeks. Annual severe exacerbation rates were 0.20 with terbutaline, 0.07 with as-needed budesonide–formoterol, and 0.09 with maintenance budesonide — a rate ratio of 0.36 (95% CI 0.27–0.49) for the combination against terbutaline. As-needed budesonide–formoterol was superior to terbutaline for symptom control but inferior to maintenance budesonide; and it achieved its exacerbation benefit on a median daily steroid dose of 57 µg, 17% of the maintenance group's 340 µg [7]. SYGMA 2 examined the same comparison against maintenance therapy [8], and a single inhaler combining rescue beclomethasone and albuterol had tested the concept earlier [29].
Read honestly: as-needed anti-inflammatory reliever is dramatically better than a bronchodilator alone and gives up some day-to-day symptom control relative to daily maintenance, at a fraction of the steroid exposure. That is a trade-off, not a free win, and which side of it suits a given patient depends on their adherence and their exacerbation history.
Severe asthma: the biologics
A minority of patients remain uncontrolled on high-dose inhaled therapy. For those with type-2 inflammation, four antibody classes now exist, each blocking a different point in the pathway. What they share is that they reduce exacerbations more reliably than they improve lung function — and that they work in patients selected by the biomarkers in Pillar 1.
Anti-IgE. Omalizumab binds IgE and prevents it engaging its receptors, and was the first biologic in severe allergic asthma [30] [31], with anti-inflammatory effects confirming IgE's central role in allergic inflammation [32].
Anti-IL-5 and anti-IL-5 receptor. IL-5 drives eosinophil maturation and survival. Mepolizumab, in 61 patients with refractory eosinophilic asthma and recurrent severe exacerbations, reduced severe exacerbations from 3.4 to 2.0 per subject over 50 weeks (relative risk 0.57, 95% CI 0.32–0.92) and improved quality of life — while producing no significant difference in post-bronchodilator FEV₁, symptoms or airway hyperresponsiveness [33]. That dissociation is the signature of the class and worth stating plainly. Benralizumab targets the IL-5 receptor α and depletes eosinophils directly [34] [35] [36]; the anti-IL-5 therapies have been reviewed together [37], as has anti-IL-5's effect on airway remodelling [38].
Anti-IL-4 receptor. Dupilumab blocks the shared IL-4Rα subunit and therefore both IL-4 and IL-13 signalling, with efficacy in moderate-to-severe uncontrolled asthma [39] [40] and — importantly for patients on long-term oral steroids — in glucocorticoid-dependent severe asthma [41].
Anti-TSLP. Tezepelumab blocks thymic stromal lymphopoietin, an epithelial-cell-derived cytokine acting upstream of the type-2 cascade. In NAVIGATOR it reduced exacerbations and improved lung function (FEV₁ difference 0.13 L, 95% CI 0.08–0.18), asthma control and quality of life, with adverse events not differing meaningfully from placebo [42] [43]. Its upstream position is the reason it is the one biologic with meaningful effects in patients who are not eosinophil-high.
A network comparison of benralizumab, dupilumab, mepolizumab, omalizumab and reslizumab in severe eosinophilic asthma has been published [44]. The uncomfortable context is what these drugs replaced: long-term oral corticosteroids, whose consequences in severe asthma have been catalogued in their own right [45].
Not every candidate succeeded — lebrikizumab, an anti-IL-13 antibody, gave equivocal pooled results in moderate-to-severe asthma [46], a reminder that blocking one cytokine in a redundant network often is not enough.
Pillar 3: progress
Phenotype-guided care
The strategic change is that "asthma" is being resolved into treatable traits. The Th2-high/Th2-low split [16], cluster analyses of clinical phenotypes, and the recognition of type-2 inflammation as a mechanism shared across airway diseases [17] together mean that a severe-asthma patient is now characterised — FeNO, blood eosinophils, IgE, exacerbation history — before a biologic is chosen [2]. This is unusually mature precision medicine for a common chronic disease, and it exists because the biomarkers are cheap.
The reliever paradigm, and what it asks of guidelines
The SYGMA results and their successors overturned the assumption that a short-acting bronchodilator alone is an acceptable reliever [7] [8], and global strategy documents have had to absorb that [4]. It is a rare instance of a guideline change that reduces both exacerbations and total steroid exposure at once.
Delivery, technique and adherence
An inhaler that is not used correctly delivers a fraction of its dose, and device errors are common. Airway deposition differs substantially between drugs and devices [22], and nebuliser and delivery-system characteristics change both airway and systemic effects [47] [48]. Aerobic training reduces bronchial hyperresponsiveness and systemic inflammation in moderate-to-severe asthma [49], and comorbidities matter: upper and lower airways share characteristics in rhinitis and asthma [50], and cigarette smoking both worsens asthma and impairs the steroid response [24].
What has not been solved
Asthma remains controllable rather than curable. Airway remodelling — the structural change that accompanies chronic inflammation — is only partly reversible, and the extent to which any current therapy prevents it is unresolved [38]. Steroid resistance has identifiable mechanisms in some patients, including Th17-mediated inflammation [51] [23], and biologics only help those with the pathway they block, which leaves non-type-2 severe asthma as the field's clearest unmet need.
Dig deeper in lmmol
Asthma is best read against the other obstructive airway diseases in this collection:
- COPD — the direct comparison. Same organ, same spirometry, overlapping symptoms, but obstruction that is largely fixed rather than reversible, and airway inflammation that genuinely differs even when the obstruction looks the same [1].
- Obesity — a risk factor for and a comorbidity of asthma, and one that alters both its phenotype and its response to treatment.
- Silicosis and black lung — occupational lung diseases where the exposure is known and the obstruction is not reversible, a useful contrast to a disease defined by variability.
- Tuberculosis — the other chronic airway review here, and a reminder that inhaled corticosteroids are not harmless in a lung that may harbour infection.
- The health reviews index collects the rest of the series.
Then move down into lmmol's graph, to the two receptors inside almost every asthma inhaler:
- Beta-2 adrenergic receptor — what salbutamol, salmeterol and formoterol bind to relax airway smooth muscle, and where β₂-adrenoceptor polymorphisms modify the response [25] [13] [26].
- Glucocorticoid receptor — where the inhaled corticosteroid acts, the target of the dose-response curve above, and the receptor whose impaired signalling explains some steroid-resistant asthma [21] [23].
- For entities without a linked static page here, use the graph index, all proteins, or all diseases rather than guessing an entity URL.