Asthma is a chronic inflammatory disorder of the airways characterised by variable expiratory airflow limitation and a constellation of symptoms — wheeze, breathlessness, chest tightness and cough — that fluctuate in intensity over time (1). It affects an estimated 262 million people worldwide and accounts for over 450,000 deaths annually, the overwhelming majority of them in low- and middle-income countries where access to controller therapy is limited (2). India carries a disproportionate share of this burden, contributing approximately 13% of global asthma cases but a considerably higher fraction of asthma deaths (3).

Despite the availability of effective inhaled corticosteroid therapy, a substantial proportion of patients remain inadequately controlled. Poor control is clinically consequential in its own right, but it also predicts exacerbations — acute deteriorations requiring systemic corticosteroids, emergency attendance or hospitalisation — which drive most of the morbidity, mortality and cost attributable to the disease and are associated with accelerated decline in lung function (4). Identifying modifiable determinants of control and exacerbation risk therefore remains a priority, particularly where such determinants are inexpensive to measure and potentially inexpensive to correct.

Vitamin D has attracted sustained attention in this context. Once regarded as a hormone concerned solely with calcium and skeletal homeostasis, 1,25-dihydroxyvitamin D is now recognised to act through vitamin D receptors expressed on airway smooth muscle, bronchial epithelium, dendritic cells and both T and B lymphocytes (5). Several of its documented actions are directly relevant to asthma pathobiology. It attenuates airway smooth muscle proliferation and the remodelling response; it induces cathelicidin and other antimicrobial peptides, potentially modifying susceptibility to the viral and bacterial respiratory infections that precipitate the majority of exacerbations; it promotes regulatory T-cell function and suppresses Th2 and Th17 cytokine production; and, in corticosteroid-resistant asthma, it has been shown to restore the responsiveness of T cells to dexamethasone through enhanced interleukin-10 production (6,7).

Hypovitaminosis D is widespread in India, with reported prevalence exceeding 70% across diverse age groups and geographic regions — a paradox given the country's latitude and insolation, attributed to cutaneous melanin content, sun-avoidant behaviour, predominantly vegetarian diets, atmospheric pollution and the near-absence of food fortification (8). The coexistence of a high asthma burden and near-universal vitamin D insufficiency makes the relationship between the two of particular local interest.

The observational literature, however, is inconsistent. Brehm and colleagues, studying Costa Rican children, found lower 25(OH)D levels associated with higher IgE, greater eosinophil counts and increased hospitalisation risk (9), and Korn et al. reported a comparable association with poor control in adults (10). Other cohorts have found weak or null associations, and interventional evidence has been similarly mixed: the VIDA trial found no effect of supplementation on treatment failure in adults with symptomatic asthma and low vitamin D (11), whereas a Cochrane meta-analysis concluded that supplementation reduced the rate of exacerbations requiring systemic corticosteroids (12). Subsequent analyses have suggested that any benefit is concentrated in those who are deficient at baseline (13).

Much of this heterogeneity may reflect methodological variation — differences in the cut-off used to define deficiency, in the instrument used to assess control, in assay platform, and in the degree of adjustment for confounders such as obesity, which independently lowers 25(OH)D and worsens asthma control. Indian data, in particular, remain sparse. The present study was therefore undertaken to determine the prevalence of vitamin D deficiency in an adult asthma cohort and to examine its association with asthma control, exacerbation frequency and spirometric function.

Study design and setting. This hospital-based cross-sectional analytical study was conducted in the Department of Biochemistry and Respiratory Medicine, Basweshwar teaching tertiary care hospital, over a period of 6 Months. Patients were recruited from the outpatient chest clinic. Participants. Consecutive adults aged 18–65 years with physician-diagnosed asthma were screened. Diagnosis required a history of variable respiratory symptoms together with documented variable expiratory airflow limitation, defined as an increase in FEV₁ of more than 12% and 200 mL following inhaled bronchodilator, in accordance with GINA criteria (1). Patients were required to have been on a stable controller regimen for at least three months prior to enrolment. Exclusion criteria were: chronic obstructive pulmonary disease, bronchiectasis, interstitial lung disease or other structural lung disease; current smoking or a smoking history exceeding 10 pack-years; pregnancy or lactation; chronic kidney disease (eGFR <60 mL/min/1.73 m²) or chronic liver disease, both of which perturb vitamin D metabolism; primary hyperparathyroidism, sarcoidosis or malabsorptive disorders; intake of vitamin D or calcium supplements, anticonvulsants, rifampicin or antiretroviral therapy within the preceding six months; and an asthma exacerbation within four weeks of enrolment, which would confound both spirometry and control assessment. Sample size. Sample size was computed for the detection of a correlation coefficient using n = [(Z_α + Z_β)/C]² + 3, where C = 0.5 × ln[(1+r)/(1−r)]. Assuming a minimum clinically relevant correlation of r = 0.30 between serum 25(OH)D and ACT score, with α = 0.05 (two-sided) and power of 80%, a minimum of 85 subjects was required. Allowing for approximately 20% incomplete data and to permit stratified analysis across three vitamin D categories, 120 patients were enrolled. Data collection. A structured proforma recorded demographic details, duration of asthma, controller regimen and GINA treatment step, comorbidities, smoking and biomass exposure, occupational and environmental exposures, dietary history, estimated daily sun exposure, and anthropometry. Height and weight were measured with the participant in light clothing and body mass index calculated as weight in kilograms divided by height in metres squared. Exposure assessment. Venous blood was collected after an overnight fast. Serum 25-hydroxyvitamin D was measured by chemiluminescent immunoassay on the [platform] analyser, with intra-assay and inter-assay coefficients of variation of [x]% and [y]% respectively. Vitamin D status was classified according to Endocrine Society thresholds as deficient (<20 ng/mL), insufficient (20–29 ng/mL) or sufficient (≥30 ng/mL) (14). Serum calcium, phosphate and alkaline phosphatase were measured concurrently. Outcome assessment. Asthma control was assessed using the Asthma Control Test, a validated five-item instrument scored from 5 to 25, administered in the participant's preferred language by an investigator blinded to the vitamin D result. Scores of ≥20, 16–19 and ≤15 were classified as well controlled, partly controlled and poorly controlled respectively (15). Exacerbations during the preceding twelve months were defined as episodes requiring systemic corticosteroids for at least three days, an unscheduled emergency visit, or hospitalisation, and were ascertained from hospital records supplemented by structured patient recall; patients with two or more such events were designated frequent exacerbators. Spirometry was performed to ATS/ERS standards and post-bronchodilator FEV₁ expressed as a percentage of predicted using appropriate reference equations. Statistical analysis. Analyses were performed in [SPSS version / R version]. Continuous variables are presented as mean ± standard deviation and categorical variables as frequencies with percentages. Normality was assessed by the Shapiro–Wilk test. Continuous outcomes were compared across the three vitamin D strata by one-way ANOVA with Bonferroni-corrected post hoc comparisons, and exacerbation counts — which were non-normally distributed — by the Kruskal–Wallis test with Mann–Whitney post hoc testing. Categorical associations were examined by Pearson's chi-square test, with odds ratios and 95% confidence intervals calculated for dichotomised comparisons. Associations between serum 25(OH)D and continuous outcomes were quantified by Pearson correlation, with Spearman's rank coefficient as a distribution-free confirmation, and by multiple linear regression adjusted for age, sex and body mass index. A two-sided p value below 0.05 was considered statistically significant. Ethics. The protocol was approved by the Institutional Ethics Committee, [Institution] (Approval No. [ ], dated [ ]). Written informed consent was obtained from every participant.

One hundred and twenty patients completed all assessments.

Table 1. Baseline characteristics of the study population (n = 120)

The cohort was middle-aged with an even sex distribution and a mean BMI within the overweight range by Asian-Indian criteria. Mean serum 25(OH)D of 20.08 ng/mL sits at the boundary between deficiency and insufficiency, and the median of 19.4 ng/mL indicates that more than half the cohort fell below the deficiency threshold. Mean ACT score of 17.22 places the average patient in the partly controlled band, and a mean of 1.70 exacerbations per year indicates a symptomatic population.

Table 2. Distribution of vitamin D status and asthma control categories

Only one patient in ten had a sufficient vitamin D level, and nine in ten fell below 30 ng/mL. The correspondingly small size of the sufficient group (n = 12) is the principal statistical constraint on the stratified analyses that follow, and it should be borne in mind when interpreting the wide confidence intervals in Tables 4 and 5. Just over three-quarters of patients were not well controlled.

Table 3. Clinical and biochemical parameters across vitamin D strata

Values are mean ± SD. F, one-way ANOVA; H, Kruskal–Wallis.

A consistent gradient runs across all three outcome rows: as vitamin D status improves, ACT scores rise, exacerbation frequency falls and FEV₁ increases. The gradient is monotonic in every case, which argues against chance. Statistical significance, however, is achieved for exacerbations and FEV₁ but not quite for ACT score (p = 0.070), and this is attributable to the small sufficient group rather than to any reversal of direction — the deficient-to-sufficient difference in ACT is 2.58 points, which exceeds the accepted minimal clinically important difference of 3 points only marginally but is directionally consistent. Age and BMI did not differ across strata, indicating that the observed gradients are not explained by confounding on these variables.

Bonferroni-corrected post hoc testing localised the differences to the extremes of the vitamin D range. For ACT score, only the deficient-versus-sufficient contrast reached significance (p = 0.047); for exacerbations, the same contrast was significant (p = 0.017), while deficient-versus-insufficient was not (p = 0.837). The clinically meaningful distinction, in other words, is between frank deficiency and repletion, not between adjacent categories.

Table 4. Cross-tabulation of vitamin D status against asthma control category

χ² = 5.88, df = 4, p = 0.208

This table repays careful reading, because it is the one that does not reach significance. No patient in the sufficient group was poorly controlled, and the proportion well controlled was roughly double that in the other two strata — both observations consistent with the gradient in Table 3. Yet the chi-square test returns p = 0.208. The reason is that collapsing a 5-to-25 point continuous score into three ordered bins, and a continuous biomarker into three categories, discards most of the available information and leaves expected cell counts too small to support inference. The null result here is a statement about statistical power and measurement scale, not evidence of absent association.

Table 5. Dichotomised associations of vitamin D deficiency

Both odds ratios point in the hypothesised direction but neither excludes unity, and both confidence intervals are wide. The pattern mirrors Table 4: dichotomising both exposure and outcome collapses a real gradient into a comparison of two heterogeneous groups, the "non-deficient" category being dominated by patients who are merely insufficient rather than truly replete.

Table 6. Correlation of serum 25(OH)D with continuous outcome measures

Analysed as continuous variables, all three associations are statistically robust and of moderate strength, with confidence intervals that exclude the null comfortably. Spearman coefficients closely track the Pearson values, confirming that the relationships are monotonic rather than artefacts of linear modelling or of outlying observations. The contrast between this table and Tables 4 and 5 — the same underlying data yielding clear associations when treated continuously and null results when dichotomised — is the central methodological observation of this study.

Table 7. Multiple linear regression, outcome: ACT score

Model R² = 0.166; adjusted R² = 0.137

After adjustment for age, sex and body mass index, serum 25(OH)D remained independently associated with asthma control. The coefficient of 0.162 implies that a 10 ng/mL increment in 25(OH)D corresponds to a 1.62-point increase in ACT score — approximately half the minimal clinically important difference of 3 points, and therefore a modest but not negligible effect. The model explains only 16.6% of the variance in ACT score, a reminder that vitamin D status is one contributor among many to asthma control, alongside adherence, inhaler technique, allergen exposure, rhinitis and comorbidity, none of which were captured here.

Three findings emerge from this cohort. First, vitamin D deficiency was near-ubiquitous: 53.3% of patients were deficient and only 10.0% sufficient. This is consistent with Indian population data reporting prevalence above 70% across age groups and regions, attributed to cutaneous melanin, sun-avoidant behaviour, low dietary intake and absent fortification (8). It should be noted that this high background prevalence limits what any single-country cross-sectional study can demonstrate, since the comparison group of truly replete patients is necessarily small. Second, serum 25(OH)D was associated with all three outcome domains — symptom control, exacerbation frequency and lung function — in a consistent direction, and the association with ACT score persisted after adjustment for age, sex and BMI (B = 0.162 per ng/mL; p < 0.001). The direction and magnitude accord with the wider literature. Korn and colleagues found insufficiency associated with poorer control and lower FEV₁ in adults (10), and Brehm et al. reported that lower levels predicted hospitalisation in children (9). Mechanistically, the finding is plausible: vitamin D modulates airway smooth muscle proliferation, induces cathelicidin with consequent effects on the respiratory infections that trigger most exacerbations, and enhances corticosteroid responsiveness through interleukin-10-dependent pathways (5–7). The inverse association with exacerbation count is the most therapeutically interesting, given that the Cochrane review of supplementation trials found a reduction specifically in exacerbations requiring systemic corticosteroids (12), with subsequent analyses suggesting benefit concentrated among those deficient at baseline (13). Third, and methodologically most instructive, the associations evident on continuous analysis largely disappeared when both exposure and outcome were dichotomised: the chi-square across control categories was non-significant (p = 0.208), and the odds ratio for frequent exacerbation among deficient patients was 1.38 (95% CI 0.67–2.84). This is not a contradiction but a consequence of information loss. Categorising a continuous biomarker at conventional thresholds, and collapsing a 21-point symptom score into three bins, reduces statistical efficiency substantially. It is plausible that some of the null findings in the published observational literature reflect the same artefact rather than genuine absence of association, and studies reporting only dichotomised analyses should be interpreted with this in mind. These observations must be set against the negative interventional evidence. The VIDA trial found that supplementation did not reduce treatment failure in adults with symptomatic asthma and low vitamin D (11), and Martineau's meta-analysis found no effect on FEV₁ or ACT score even where exacerbations were reduced (12). A cross-sectional association is entirely compatible with reverse causation — poorly controlled asthmatics may go outdoors less and thus synthesise less vitamin D — and with confounding by obesity, physical activity and socioeconomic position. Nothing in the present design distinguishes these possibilities. Limitations are substantial. The cross-sectional design cannot establish temporality. Exacerbations over twelve months were ascertained partly by recall, introducing misclassification. Single-timepoint 25(OH)D does not capture seasonal variation. The sufficient group comprised only 12 patients, so stratified estimates are imprecise. Adherence and inhaler technique — powerful determinants of control — were not measured, and residual confounding by these is likely. Single-centre recruitment from a tertiary clinic limits generalisability.

Vitamin D deficiency was present in more than half of this adult asthma cohort, and fewer than one patient in ten had a sufficient level. Serum 25(OH)D showed consistent moderate associations with asthma control, exacerbation frequency and post-bronchodilator FEV₁, and the association with ACT score persisted after adjustment for age, sex and body mass index. These associations were apparent when vitamin D and outcomes were analysed as continuous variables but attenuated substantially on dichotomisation, indicating that conventional categorical thresholds discard clinically relevant information.

The cross-sectional design precludes any causal inference, and the negative findings of major supplementation trials counsel against extrapolating these associations into a treatment recommendation. What the data do support is measurement of 25(OH)D in patients with poorly controlled or frequently exacerbating asthma as part of a broader assessment, and correction of deficiency where found on established skeletal and general-health grounds. Adequately powered randomised trials restricted to deficient patients, using exacerbation rate as the primary endpoint, remain the appropriate route to determining whether repletion improves asthma outcomes.