Determinants of Adolescent Lung Function in Indians: Race, Nutrition, and Systemic Inflammation

To the Editor:

Spirometry, one of the most commonly used pulmonary function tests, is additionally informative of general health, with reductions in spirometry volumes corresponding to respiratory disease as well as increased cardiometabolic morbidity and mortality (1–3). However, to know whether spirometry volumes are reduced, we must first define the normal. Owing to large geoethnic variations, such normal values are usually derived from regional data. Others and we have previously postulated that the use of regional data of apparently healthy subjects to derive a local normal may mask widely prevalent adverse exposures and subclinical abnormalities (4). For example, Indians have the lowest “normal” lung function globally, about 30% lower than Caucasians, but Indian-origin U.S.-born children have better lung function than their immigrant parents (5, 6). Furthermore, India has the greatest genetic diversity after Africa, with regions of relatively distinct genetic ancestries referred to as Indo-European (IE), Dravidian (DR), and Tibeto-Burman (TB), as well as areas of admixture (7). Here, we examined the correlates of spirometric lung function in 2,609 (1,210 girls, 1,399 boys) adolescent children and young adults, aged 9–20 years, from a government residential school system (Jawahar Navodaya Vidyalaya, run by Navodaya Vidyalaya Samiti, India) to determine whether nature or nurture is behind the lower lung function of Indians and whether it represents a true normal. A portion of the work contained in this manuscript has been previously reported in the form of a preprint (https://doi.org/10.1101/2021.03.01.21252646).

Fourteen districts were carefully selected to provide adequate genetic and geographical representation within India. Briefly, height, weight, shoulder width, and waist circumference were measured, and spirometry was done using portable spirometers (EasyOne air, NDD) as per guidelines. Data were analyzed for 2,338 subjects without asthma with acceptable spirometry. Both parental and subject questionnaires were administered with questions related to background, habits, and symptoms. Genetic ancestry was mapped at the degree of geographic regions that have previously been mapped by the Indian Genome Variation Database to correspond to IE, DR, TB, or mixed. In a subset of participants (n = 1,582), 8 ml of blood was collected, and cytokine concentrations (IL-1, IL-4, IL-6, IL-8, IL-13, and IL-17; tumor necrosis factor [TNF]-α; and IFN-γ) were measured in serum using a custom MILLIPLEX MAP Human Cytokine/Chemokine Magnetic Bead Panel (Merck Millipore). Associations of lung functions were investigated using multiple (linear) regression models adjusted for covariates (age, sex, height, altitude, and ethnicity). A model-based clustering method (latent profile analysis) was used to identify hidden subphenotypes correlated with lung function among 903 participants with complete anthropometric and inflammometric data.

Adjusted associations are shown in Table 1. The TB group had the highest lung volumes (FVC and FEV₁) followed by the IE group, after adjusting for all remaining covariates. The DR group was characterized by the poorest lung function despite better or comparable anthropometry. As the marker of central adiposity, waist to height ratio (WHtR) showed a positive association with lung volumes. Subjects with broader shoulders proportional to height (shoulder to height ratio) and a higher z-score of body mass index for age (World Health Organization growth curves) showed higher lung volumes. Hemoglobin concentrations were also positively associated with the lung volumes, with moderately severe anemia being associated with significantly reduced lung function. Type of diet showed no association with lung function; however, skipping meals and chronic digestive health issues (gastrointestinal symptoms) were associated with a reduction in lung volumes. IL-8 and IL-17 showed a negative association with FVC but not with FEV₁ and FEV₁/FVC ratio. Subjects with a family history of asthma and allergic diseases had a lower FEV₁/FVC ratio but similar lung volume. There was no association with birth weight declared on the parental questionnaire.

Table 1.

Association of Lung Function with Ethnicity, Growth, and Other Health Parameters

		FVC (L)			FEV1 (L)			FEV1/FVC Ratio
	n or n (%)	Estimate	95% CI	P Value	Estimate	95% CI	P Value	Estimate	95% CI	P Value
Ethnicity
Indo-Aryan	869 (37.2%)	—	—	—	—	—	—	—	—	—
Dravidian	507 (21.7%)	−0.13	−0.17 to −0.09	<0.001	−0.12	−0.16 to −0.09	<0.001	−0.002	−0.008 to 0.004	0.56
Mongoloid	440 (18.8%)	0.17	0.11 to 0.22	<0.001	0.18	0.13 to 0.23	<0.001	0.006	−0.002 to 0.015	0.12
Mixed	522 (22.3%)	0.01	−0.03 to 0.05	0.52	0.02	−0.01 to 0.06	0.21	0.004	−0.002 to 0.01	0.2
Skip meals
Never/rarely	1,996 (85.9%)	—	—	—	—	—	—	—	—	—
Often/sometimes	327 (14.1%)	−0.08	−0.12 to −0.03	<0.001	−0.06	−0.1 to −0.02	0.004	0.004	−0.003 to 0.011	0.23
Digestive issues
Never/rarely	2,141 (92.2%)	—	—	—	—	—	—	—	—	—
Often/sometimes	182 (7.8%)	−0.08	−0.13 to −0.02	0.008	−0.09	−0.14 to −0.03	0.001	−0.003	−0.012 to 0.006	0.49
Family history of  asthma and allergic diseases
No	2,046 (93.1%)	—	—	—	—	—	—	—	—	—
Yes	151 (6.9%)	0.008	−0.053 to 0.069	0.8	−0.04	−0.092 to 0.019	0.2	−0.013	−0.022 to −0.004	0.004
Anthropometry
WHtR	2,315	1.6	1.3 to 1.9	<0.001	0.6	0.31 to 0.9	<0.001	−0.27	−0.32 to −0.22	<0.001
SHtR	2,335	2.3	1.7 to 2.9	<0.001	1.4	0.83 to 1.94	<0.001	−0.23	−0.32 to −0.14	<0.001
zBMI (WHO)	2,310	0.12	0.1 to 0.13	<0.001	0.07	0.055 to 0.08	<0.001	−0.013	−0.015 to −0.011	<0.001
Inflammation
ln (IL-8)	1,478	−0.02	−0.039 to −0.001	0.039	−0.015	−0.032 to 0.002	0.092	0.00034	−0.0023 to 0.003	0.8
ln (TNF-α)	1,492	−0.017	−0.049 to 0.015	0.3	−0.024	−0.053 to 0.005	0.1	−0.0031	−0.0076 to 0.0014	0.18
ln (IFN-γ)	1,106	−0.009	−0.022 to 0.004	0.2	−0.009	−0.021 to 0.003	0.16	−0.0007	−0.0026 to 0.0012	0.45
ln (IL-17)	1,254	−0.017	−0.03 to −0.004	0.012	−0.011	−0.023 to 0.0007	0.066	0.0006	−0.0012 to 0.0024	0.51
Hemoglobin	2,014	0.04	0.023 to 0.05	<0.001	0.04	0.026 to 0.05	<0.001	0.002	−0.0004 to 0.0037	0.12

Definition of abbreviations: CI = confidence interval; ln = natural logarithm; SHtR = shoulder to height ratio; TNF = tumor necrosis factor; WHO = World Health Organization; WHtR = waist to height ratio; zBMI = z-score of body mass index.

Regression parameters for statistically significant associations are shown in bold.

Given the associations of adiposity, digestive health, nutritional markers, and inflammation with lung function (particularly FVC), we further determined whether there was a profile associated with poorer lung function in India that operates across ethnicities. Global z-scores of FVC (zFVC) were calculated for all subjects, independent of ethnicity, from The Global Lung Function Initiative equations for the Caucasian ethnic group (8). zFVC_GLI was positively associated with WHtR and negatively associated with IL-8, TNF-α, IFN-γ, and IL-17. Latent profile analysis was performed to find subclasses within the zFVC_GLI spectrum. Using central obesity and inflammatory parameters, a relatively well-defined endotype of poor lung function, thinness and elevated inflammation was seen in one-third of the subjects, with comparable membership across all ethnicities (Figure 1). Compared with the remaining subjects, subjects with this endotype had a reduction in zFVC_GLI and WHtR and an increase in IFN-γ, IL-17, and TNF-α (fold increase: 2.7, 1.8, and 1.5, respectively). A more modest elevation in IL-8 (1.2-fold) was also seen. It is noted that IL-8 and TNF-α concentrations were above Western clinical reference values in the majority of subjects studied.

Figure 1.

Mean comparison of latent profiles. Data are represented as mean (SD) of model parameters for latent profiles (Cluster 1 and 2); the 95% confidence interval was calculated for the mean difference between the two clusters. P values were calculated using the Welch two-sample t test. CI = confidence interval; GLI = The Global Lung Function Initiative; ln = natural logarithm; TNF = tumor necrosis factor; WHtR = waist to height ratio; zFVC = z-score of FVC (using GLI [8]).

This is the first study from India that looks at adolescent to young adult spirometric lung function through a multiprismatic lens combining anthropometry, ethnogenetics, life-history questionnaire, and general clinical markers of nutritional deficiency and inflammatory state. We show that genetic and environmental factors are both operational in determining lung function in Indians. Although we did not see significant associations with available ambient air quality indices in our study (not shown), more detailed studies looking at life-course exposures may reveal such effects. Gastrointestinal symptoms, meal skipping, thinness, low body mass index, vitamin deficiency symptoms, moderate to severe anemia, and elevated inflammation markers were all associated with lower FEV₁ or FVC. The endotype of low FVC, thinness, and inflammation is notable because it suggests that a large fraction of Indians may not be reaching their genetic potential because of nutritional or environmental adversity. Because the study subjects are unlikely to be calorically limited, other possibilities to be investigated are low nutritional quality or gut dysbiosis leading to nutritional steal and enteropathy. We speculate that this represents prevalent environmental enteropathy, a poorly recognized condition that is seen in low- to middle-income tropical countries and is thought to be driven by poor sanitation and recurring fecooral exposures (9). This fits previous reports that children from rural and semiurban India have poorer lung function than those in cities, and first-generation Indian-origin Americans have better lung function than their immigrant parents (6, 10). The route to better lung function and health for Indians may therefore lie through better sanitation and nutrition.

Supplemental material containing detailed methods, data, and analysis is available at https://doi.org/10.6084/m9.figshare.14986182.