Fetal Origins of Asthma: A Longitudinal Study from Birth to Age 36 Years

Abstract

Rationale: Deficits in infant lung function—including the ratio of the time to reach peak tidal expiratory flow to the total expiratory time (tptef/te) and maximal expiratory flow at FRC (V̇maxFRC)—have been linked to increased risk for childhood asthma.

Objectives: To examine the individual and combined effects of tptef/te and V̇maxFRC in infancy on risk for asthma and abnormalities of airway structure into mid-adult life.

Methods: One hundred eighty participants in the Tucson Children’s Respiratory Study birth cohort had lung function measured by the chest-compression technique in infancy (mean age ± SD: 2.0 ± 1.2 mo). Active asthma was assessed in up to 12 questionnaires between ages 6 and 36 years. Spirometry and chest high-resolution computed tomographic (HRCT) imaging were completed in a subset of participants at age 26. The relations of infant tptef/te and V̇maxFRC to active asthma and airway structural abnormalities into adult life were tested in multivariable mixed models.

Measurements and Main Results: After adjustment for covariates, a 1-SD decrease in infant tptef/te and V̇maxFRC was associated with a 70% (P = 0.001) and 55% (P = 0.005) increased risk of active asthma, respectively. These effects were partly independent, and two out of three infants who were in the lowest tertile for both tptef/te and V̇maxFRC developed active asthma by mid-adult life. Infant V̇maxFRC predicted reduced airflow and infant tptef/te reduced HRCT airway caliber at age 26.

Conclusions: These findings underscore the long-lasting effects of the fetal origins of asthma, support independent contributions by infant tptef/te and V̇maxFRC to development of asthma, and link deficits at birth in tptef/te with HRCT-assessed structural airway abnormalities in adult life.

At a Glance Commentary

Scientific Knowledge on the Subject

Deficits in infant lung function—including the ratio of the time to reach peak tidal expiratory flow to the total expiratory time (tptef/te) and maximal expiratory flow at FRC (V̇maxFRC)—have been linked to increased risk for subsequent asthma, mainly in childhood.

What This Study Adds to the Field

We found partly independent effects of infant tptef/te and V̇maxFRC on risk of active asthma into the fourth decade of life and observed differential effects of low infant tptef/te on computed tomography–assessed airway measures and of low V̇maxFRC on spirometry-defined airflow limitation in adult life.

In the last few decades, evidence has been accumulating on the long-lasting influence that early-life exposures, events, and developmental processes can have on the prevalence of adult chronic conditions, including asthma. However, to what extent the early origins of asthma are established in utero and whether risk factors that are already present at birth can affect disease outcomes into mid-adult life are largely unknown (1).

In this context, reduced lung function measured shortly after birth or in early infancy has been shown to increase the risk for wheezing and asthma in childhood. In the birth cohort of the TCRS (Tucson Children’s Respiratory Study), we first reported that participants with low maximal expiratory flow at FRC (V̇maxFRC) in infancy had an increased risk for wheezing in the first 3 years of life (2–4). These findings were replicated in several other birth cohorts, including the Manchester Asthma and Allergy Study (5) and the Southampton Women’s Survey (6, 7), and the same trend was observed in another population-based study of 97 infants in Boston (8). A longitudinal cohort from the University of Western Australia has provided long-term follow-up on the relation of infant V̇maxFRC to asthma risk to date and reported significant associations of deficits in V̇maxFRC measured at 1 month of age to recurrent wheezing in infancy (9) and asthma at age 11 (10). This cohort has now completed an assessment in adult life and found the effects of infant V̇maxFRC on active asthma to persist up to age 24 years (11).

Of note, in the original TCRS report (2), we observed that an index of infant lung function that is measured during tidal breathing, the ratio of the time to reach peak tidal expiratory flow to the total expiratory time (tptef/te, also called Tme/Te), had a stronger relation to subsequent risk of early wheezing than V̇maxFRC. tptef/te measured a few days after birth was later shown to predict wheezing in infancy and asthma in childhood in two other studies (12, 13). In the Norwegian Environment and Childhood Asthma birth cohort (13), participants who had low tptef/te shortly after birth had a twofold greater risk of having active asthma at age 10 years as compared with participants with high tptef/te. However, in that cohort, the use of low tptef/te (defined as levels below the median) to predict childhood asthma had a positive predictive value below 15%, suggesting that this test by itself is unlikely to provide valuable information for screening asthma risk at birth.

tptef/te and V̇maxFRC likely measure related but different components of lung function. V̇maxFRC approximates maximal airflow at a low lung volume theoretically reflecting the elastic recoil pressure of the lung opposed by the resistance of the airways upstream from multiple flow-limiting segments or choke points that are positioned peripherally, effectively excluding the greater resistance in the central and upper airways (14). In contrast, tptef/te likely reflects neuromuscular respiratory control through braking of expiratory flow by inspiratory muscles (15, 16) in response to the balance of lung recoil and chest wall compliance and the overall time constant of the respiratory system. In this case, the proportionally greater resistance of the central airways is contributing to this time constant of the system, which is not the case during flow limitation at FRC. It is therefore plausible that infant tptef/te and V̇maxFRC affect subsequent asthma risk through mechanisms that are partly independent. However, to date, no attempt has been made to determine whether tptef/te and V̇maxFRC in infancy may have independent and additive effects on the risk for subsequent asthma. Most importantly, the nature of these fetal-origin abnormalities remains unknown because no study has determined the relations of these two indices of early lung function to both airway function and airway structure achieved in adult life.

The TCRS birth cohort has now followed subjects to age 36 years. In the current study, we sought to use the extensive longitudinal data of this cohort to determine the relation of low tptef/te and low V̇maxFRC in early infancy to the risk of active asthma, airflow limitation, and imaging-based airway structural abnormalities into adult life.

Methods

The TCRS Birth Cohort

Between 1980 and 1984, 1,246 healthy infants were enrolled at birth in the Tucson Children’s Respiratory Study, a longitudinal nonselected birth cohort study of the early origins of respiratory disease. The chest compression technique for assessing pulmonary function in infancy was developed as the last 376 infants were enrolled in the study. Only these infants were eligible for infant pulmonary function testing, and 180 participants less than 6 months of age were tested (mean ± SD: 2.0 ± 1.2 mo, min–max: 0.7–5.7 mo). Details of the selection process and exclusion criteria have been previously reported (2).

Lung Function Measurement in Early Infancy

Infant lung function testing included measurement of tidal breathing indices, FRC, and partial expiratory flow–volume curves as previously reported (2–4, 17). The mean of the percentage of expiratory time necessary to reach peak tidal expiratory flow (tptef/te) was determined from the tidal breathing flow–time curves (2, 16). FRC was measured by the helium-dilution-gas-equilibration method (18) with a minimum of two measurements within 10%. Flow at the end-tidal point (V̇maxFRC) was determined from the partial expiratory flow–volume curves obtained by the chest-compression technique (19–21), and the greatest flow obtained after three to five maneuvers was used in the analysis (2).

Early-Life Factors

Infant race and ethnic background were determined from parental questionnaires collected at enrollment. Parental years of education, age, and smoking status were reported by questionnaire at enrollment. Parental asthma, defined as ever having received a physician diagnosis of asthma, was determined from a questionnaire completed by each parent shortly after enrollment. Gestational age, birthweight, and birth length were obtained from medical records. The ponderal index at birth was computed as the birth weight divided by the birth length cubed.

Asthma Outcomes at Ages 6–36 Years

For participants, active asthma was defined as a physician diagnosis of asthma received at any age with active symptoms during the past year based on responses to 12 questionnaires completed at ages 6, 8, 11, 13, 16, 18, 22, 24, 26, 29, 32, and 36 years of age.

Spirometry and High-Resolution Computed Tomographic Measurements at Age 26

At the year 26 survey, spirometry and, in a subset of participants, high-resolution computed tomographic (HRCT) imaging of the chest were completed. Spirometry was performed according to guidelines by the American Thoracic Society and European Respiratory Society (22), and percentage of predicted values were calculated based on reference equations from the Global Lung Initiative (23). Following baseline pulmonary function testing and the administration of 360 μg of albuterol through an Aerochamber (Monaghan Medical Corp.), participants were transferred to the imaging center where lung volume testing via N₂ washout in the supine position was performed to establish TLC and end-expiratory lung volume (referred to as FRC). Once TLC and FRC were established, and still within 30 minutes of albuterol administration, HRCT scanning commenced. As described previously (24), for each participant, airway generations 0 to 10, ranging in size from approximately 14 to 2 mm in diameter, were measured at both TLC and FRC. The Volumetric Image and Display Analysis Pulmonary Workstation 2 software (Department of Radiology, Division of Physiologic Imaging, University of Iowa), which has been validated and demonstrated to have low user bias (25, 26), was then used to analyze airway luminal areas and diameters. For statistical analyses, after removing data from generation 0 (trachea), we used data from generations 1 to 9 because the majority of participants (68%) had missing values for generation 10. However, all results were confirmed in sensitivity analyses that also included data from generation 10.

Statistical Analyses

The relation of age, sex, and length to each of the infant lung function indices was assessed using backward selection in multivariate linear regression models. The independent predictors identified in these regression models were age for tptef/te, sex and length for V̇maxFRC, and age and length for FRC. The standardized residuals from these models were then saved and used as standardized, adjusted levels for each infant lung function index. Standardized, adjusted levels for each index were also divided into tertiles to evaluate asthma risk associated with low infant lung function (defined as being in the lowest tertile).

Participants were classified based on whether they had asthma at any point between ages 6–36 years (ever active asthma), ages 6–18 (active asthma in childhood), and ages 22–36 (active asthma in adult life), and mean levels of infant lung function indices were compared between groups with and without asthma. The positive and negative predictive value, sensitivity, and specificity of low tptef/te, low V̇maxFRC, and their combination were computed for the risk of having active asthma at any point between ages 6 and 36 years.

Generalized estimating equations (GEEs) were used to determine the relation of each infant lung function index to asthma risk longitudinally using all observations between ages 6 and 36 years, with an exchangeable correlation structure (best fitting by quasi-likelihood under the independence model criterion) and a logit link. GEE models were adjusted for other covariates measured at birth that were either selected a priori (i.e., sex and ethnicity) or found in our data to predict the risk of the offspring developing asthma during follow-up (i.e., maternal education and maternal asthma). The relation of infant lung function indices to asthma risk was also tested with Kaplan-Meier survival curves for incidence of active asthma from age 6 years onward.

The relation of each infant lung function index to % predicted values of FEV₁, FVC, FEV₁/FVC, and forced expiratory flow, midexpiratory phase (FEF_25–75) at age 26 was tested with Pearson’s correlation coefficients. Because HRCT parameters (Figure E1 in the online supplement, inner diameter, outer diameter, luminal area, wall area, total area, and wall thickness) were measured in 10 airway generations for each participant, we standardized HRCT parameters within each airway generation and tested their association with infant lung function indices in random effects models that control for the serial correlation of intrasubject measurements across airway generations. Models were also adjusted for sex and body surface area.

All analyses were limited to the 180 participants who had at least one index (tptef/te, V̇maxFRC, or FRC) for infant lung function available in the first 6 months of life. Four participants had a lower respiratory tract illness before completing the lung function measurement, and, in sensitivity analyses, all main results were confirmed after removing these four individuals. Informed consent was obtained from parents (and from participants starting from age 18), and the Institutional Review Board of the University of Arizona approved the study.

Results

Characteristics of the 180 TCRS participants included in this study are compared with those of the 1,066 participants that were not included in Table E1. Included and excluded participants did not differ in their risk of having active asthma in either childhood or adult life.

The infants included in this study had mean levels (SD) of 29.9% (10%, n = 180) for tptef/te, 121 ml/s (54 ml/s, n = 159) for V̇maxFRC, and 103 ml (27 ml, n = 134) for FRC. Among demographic and anthropometric factors, sex, age, and length at the time of the measurement were associated with infant lung function (Table 1). These factors (Table E2) were used to compute standardized adjusted levels for each infant lung function index as described in the Methods section. Among the adjusted lung function indices, only tptef/te and V̇maxFRC were found to correlate with each other (r = 0.28, n = 158, P = 0.0003).

Table 1.

Association of Baseline Demographic and Parental Factors with Unadjusted Infant Lung Function

Categorical Factors	Infant Lung Function Index
	tptef/te (%)			V̇maxFRC (ml/s)			FRC (ml)
	n	Mean (SE)		n	Mean (SE)		n	Mean (SE)
Sex
F	88	30.7 (1.1)		76	131.8 (6.5)		65	97.4 (3.0)
M	92	29.1 (0.9)		83	110.7 (5.6)		69	108.2 (3.3)
P value		0.276			0.014			0.019
Ethnicity/Race
Non-Hispanic white	121	29.7 (0.9)		104	122.8 (5.3)		87	104.4 (2.7)
Any Hispanic white	34	29.7 (1.6)		32	112.7 (9.3)		28	102.6 (6.0)
All other	25	30.8 (2.1)		23	122.7 (11.9)		19	96.3 (5.7)
P value		0.873			0.647			0.481
Delivery
Vaginal	141	30.2 (0.8)		127	118.1 (5.0)		104	99.0 (2.4)
Cesarean section	39	28.9 (1.4)		32	131.3 (8.4)		30	116.5 (5.2)
P value		0.488			0.219			0.001
Maternal education
>12 yr	140	29.9 (0.8)		123	123.5 (5.1)		102	103.8 (2.4)
≤12 yr	40	29.9 (1.5)		36	111.3 (7.4)		32	100.1 (5.9)
P value		0.992			0.236			0.489
Paternal education
>12 yr	130	29.7 (0.8)		114	124.9 (5.4)		96	104.5 (2.6)
≤12 yr	48	30.6 (1.5)		43	110.5 (6.9)		37	98.2 (5.0)
P value		0.567			0.140			0.224
Maternal smoking
No	146	30.0 (0.8)		128	124.3 (4.8)		109	102.3 (2.5)
Yes	34	29.5 (1.6)		31	106.3 (9.1)		25	105.4 (6.3)
P value		0.813			0.098			0.604
Paternal smoking
No	134	29.7 (0.8)		114	118.7 (5.2)		95	102.0 (2.6)
Yes	44	30.7 (1.8)		43	126.8 (8.0)		38	104.8 (4.9)
P value		0.536			0.408			0.591
Maternal asthma
No	156	30.3 (0.8)		141	121.1 (4.6)		117	102.1 (2.5)
Yes	22	27.4 (1.6)		17	120.6 (13.5)		16	107.8 (5.3)
P value		0.185			0.974			0.436
Paternal asthma
No	139	30.3 (0.8)		120	124.7 (4.7)		102	101.6 (2.4)
Yes	28	28.0 (2.0)		28	112.2 (12.3)		22	107.5 (7.6)
P value		0.254			0.273			0.339

Continuous Factors	tptef/te (%)			V̇maxFRC (ml/s)			FRC (ml)
	n	R	P Value	n	r	P Value	n	r	P Value
Age at time of test, d	180	−0.45	<0.001	159	0.19	0.014	134	0.68	<0.001
Length at time of test, cm	174	−0.35	<0.001	158	0.20	0.011	134	0.69	<0.001
Gestational age, wk	138	0.08	0.329	121	0.16	0.073	108	0.06	0.557
Ponderal index at birth, weight/length3	135	−0.10	0.251	119	0.03	0.767	106	0.09	0.381
Maternal age, yr	180	0.002	0.977	159	0.04	0.610	134	−0.01	0.898
Paternal age, yr	178	0.06	0.517	157	0.11	0.161	133	−0.03	0.701

Definition of abbreviations: tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time; V̇maxFRC = maximal expiratory flow at FRC.

The Relation of Infant Lung Function to Asthma Risk up to Age 36 Years

Prevalence of active asthma in participants in this analysis ranged between 8% at year 6 and 18% at year 36. Overall, levels of infant tptef/te and V̇maxFRC (but not FRC) were lower for participants who went on to develop active asthma during the study follow-up than for participants who never did (Figure 1). This held true for active asthma both in childhood and in adult life (Figure 1). When assessed longitudinally during follow-up, at each age, participants with active asthma had had lower levels of tptef/te and V̇maxFRC in infancy than participants with no active asthma (Figure E2). In GEE models that included observations from age 6 to 36 years, both indices were significantly associated with the risk for active asthma (P = 0.003 for tptef/te and P = 0.007 for V̇maxFRC). Similarly, as shown in Figure 2, participants in the lowest tertile of adjusted infant tptef/te and in the lowest tertile of adjusted infant V̇maxFRC had the highest risk for having active asthma at age 6–36 years (P = 0.009 and P = 0.01, respectively, for comparison of rates of active asthma across tertiles of infant lung function from GEE models). No significant effects of FRC tertiles on risk of asthma were found.

Figure 1.

Comparison of standardized, adjusted levels of infant lung function (tptef/te, V̇maxFRC, and FRC) between participants who had and did not have active asthma at any time between ages 6–36 years (entire follow-up), 6–18 years (childhood), and 22–36 years (adult life). tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time; V̇maxFRC = maximal expiratory flow at FRC.

Figure 2.

Prevalence of active asthma from age 6 to 36 years among participants who had low, medium, and high lung function in infancy (tptef/te in A; V̇maxFRC in B; FRC in C). For each lung function index, tertiles were computed using standardized, adjusted levels. P values refer to the comparison of rates of active asthma across the tertiles from generalized estimating equation models. tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time; V̇maxFRC = maximal expiratory flow at FRC.

Kaplan-Meier curves (Figure E3) indicated that, although the effects of having low tptef/te or low V̇maxFRC (and even more so, having low levels for both indices) on asthma risk were already present in childhood, the magnitude of these effects increased over time into adult life.

We then tested the effects of each infant lung function index on asthma risk between ages 6 and 36 years in adjusted multivariate longitudinal GEE models that included sex, ethnicity, maternal education, and maternal asthma as covariates. When tested as predictors in separate models (Table 2, upper part), we found a 1-SD decrease in infant tptef/te to be associated with 70% increased odds and a 1-SD decrease in infant V̇maxFRC with 55% increased odds for active asthma between ages 6 and 36 years (P = 0.001 and P = 0.005, respectively). In contrast, FRC was not associated with asthma risk. When the independent effects of tptef/te and V̇maxFRC were tested in mutually adjusted models (Table 2, lower part), we found both indices to be independently associated with asthma risk in GEE models including all observations ages 6–36 years, although V̇maxFRC appeared a stronger predictor for asthma risk in childhood and tptef/te for asthma in adult life. Consistent associations of low tptef/te and low V̇maxFRC with asthma risk were found when infant lung function data were analyzed in tertiles (Table E3).

Table 2.

Effects of Infant Lung Function Levels on Subsequent Risk for Active Asthma between Ages 6 and 36 Years

Effect for 1-SD Decrease in Standardized, Adjusted Index	Active Asthma Age 6–36 yr*			Active Asthma Age 6–18 yr†			Active Asthma Age 22–36 yr‡
	adjOR	95% CI	P Value	adjOR	95% CI	P Value	adjOR	95% CI	P Value
Individual GEE models§
(1) tptef/te	1.70	1.23–2.36	0.001	1.55	1.07–2.25	0.021	1.78	1.18–2.68	0.006
(2) V̇maxFRC	1.55	1.14–2.11	0.005	1.61	1.13–2.63	0.008	1.38	0.84–2.04	0.127
(3) FRC	0.93	0.68–1.27	0.666	1.16	0.77–1.74	0.474	0.85	0.57–1.27	0.436
Jointly adjusted GEE models||
tptef/te	1.53	1.07–2.18	0.020	1.32	0.89–1.96	0.164	1.74	1.10–2.76	0.018
V̇maxFRC	1.40	1.02–1.91	0.038	1.50	1.05–2.15	0.028	1.19	0.80–1.80	0.392

Definition of abbreviations: adjOR = adjusted odds ratio; CI = confidence interval; GEE = generalized estimating equation; tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time; V̇maxFRC = maximal expiratory flow at FRC.

All GEE models adjusted for sex, ethnicity, maternal education, and maternal asthma. In all models, standardized, adjusted tptef/te, V̇maxFRC, and FRC were included as inverse values (i.e., values multiplied by −1) to compute the effect of 1-SD decrease in their levels on asthma risk.

^*Model Ns (subjects, observations): 168, 1,459 for tptef/te; 149, 1,318 for V̇maxFRC; and 128, 1,135 for FRC.

^†Model Ns (subjects, observations): 168, 807 for tptef/te; 149, 729 for V̇maxFRC; and 128, 626 for FRC.

^‡Model Ns (subjects, observations): 137, 652 for tptef/te; 124, 589 for V̇maxFRC; and 108, 509 for FRC.

^§Three separate GEE models were run that included among independent variables 1) standardized adjusted tptef/te, 2) standardized adjusted V̇maxFRC, and 3) standardized adjusted FRC.

^||A single GEE model was run that included simultaneously standardized adjusted tptef/te and standardized adjusted V̇maxFRC among independent variables.

Next, we assessed the positive and negative predictive value, sensitivity, and specificity of low tptef/te, low V̇maxFRC, and their combination on asthma risk (Table 3). The combination of low tptef/te and low V̇maxFRC in infancy had limited sensitivity (32%) but a strong positive predictive value (63%), with nearly two out of three participants in this group developing active asthma during follow-up. When analyses were restricted to asthma in adult life, the positive predictive value for this group remained virtually unchanged (65%, data not shown). Consistent with these results, infants who had both low tptef/te and low V̇maxFRC had a greater than threefold increased risk for asthma, as compared with all other participants (Table 3).

Table 3.

Standardized, Adjusted Infant Lung Function Indices as Predictors of Active Asthma up to Age 36 Years

	Positive Predictive Value	Negative Predictive Value	Sensitivity	Specificity	adjOR (95% CI)*
Active asthma ages 6–36 yr
Low tptef/te	48%	70%	46%	72%	1.94 (1.09–3.45)
Low V̇maxFRC	56%	75%	54%	76%	2.48 (1.33–4.61)
Low tptef/te + low V̇maxFRC	63%	70%	32%	90%	3.36 (1.67–6.76)

Definition of abbreviations: adjOR = adjusted odds ratio; CI = confidence interval; tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time; V̇maxFRC = maximal expiratory flow at FRC.

For each infant lung function index, the “low” category corresponds to the lowest tertile of the standardized adjusted levels. Among the 150 participants with available adjusted levels of both tptef/te and V̇maxFRC and available information on subsequent asthma, 52 (35%) had low tptef/te, 52 (35%) had low V̇maxFRC, and 27 (18%) had both low tptef/te and low V̇maxFRC.

^*Odds ratios from generalized estimating equation models adjusted for sex, ethnicity, maternal education, and maternal asthma comparing participants in the low lung function category versus all other participants. Age and length at time of infant lung function were not included as covariates because standardized adjusted levels of infant lung function (already adjusted for these factors, see Methods) were used for analyses.

The Relation of Infant Lung Function to Spirometric and HRCT Parameters in Adult Life

Last, we evaluated the relation of infant tptef/te and V̇maxFRC to spirometric and HRCT parameters measured at age 26 years. Table E4 shows the characteristics of the subsets of TCRS participants with spirometric and HRCT parameters at age 26 who also had available data on lung function in infancy. As shown in Table E5 and consistent with our previous report up to age 22 (17), we observed significant and direct correlations between standardized, adjusted V̇maxFRC levels in infancy with FEV₁, FEV₁/FVC, and FEF_25–75 levels measured at age 26 years. In contrast, infant tptef/te levels were not associated with any spirometric parameters at age 26.

When we analyzed HRCT data measured at FRC at age 26 years, we observed that low infant tptef/te was associated with significant reductions in inner diameter, outer diameter, total area, luminal area, and wall area (Table 4 and Figure E4), indicating smaller, thinner airways. Of note, these associations were not evident in HRCT measurements taken at TLC, and when we computed the ratio between FRC and TLC values for each HRCT parameter, we found participants with low infant tptef/te to have remarkably decreased ratios (i.e., to have the largest difference between TLC and FRC values) for inner diameter, outer diameter, total area, luminal area, and wall area (Table 4). Apart from a reduction in wall area at FRC, no consistent associations of low V̇maxFRC were observed with HRCT parameters measured at either FRC or TLC (Table E6).

Table 4.

Relation of Infant tptef/te Tertiles to HRCT Parameters at Age 26 Years

Standardized HRCT Parameters	tptef/te Tertile	HRCT Measured at						FRC/TLC Ratio
		FRC			TLC
		Coef*	95% CI	P Value	Coef*	95% CI	P Value	Coef*	95% CI	P Value
Inner diameter	High	ref			ref			ref
	Med	−0.64	−1.1 to −0.2	0.004	0.19	−0.3 to 0.7	0.431	−0.18	−0.3 to −0.1	0.008
	Low	−0.65	−1.1 to −0.2	0.007	0.20	−0.3 to 0.7	0.452	−0.24	−0.4 to −0.1	0.001
Outer diameter	High	ref			ref			ref
	Med	−0.77	−1.2 to −0.4	<0.001	0.08	−0.4 to 0.6	0.729	−0.17	−0.3 to −0.1	0.005
	Low	−0.75	−1.2 to −0.3	0.001	0.14	−0.4 to 0.7	0.586	−0.22	−0.4 to −0.1	0.001
Luminal area	High	ref			ref			ref
	Med	−0.58	−1.0 to −0.1	0.011	0.33	−0.1 to 0.8	0.168	−0.35	−0.6 to −0.1	0.007
	Low	−0.55	−1.0 to −0.1	0.026	0.36	−0.2 to 0.9	0.171	−0.46	−0.7 to −0.2	0.001
Wall area	High	ref			ref			ref
	Med	−0.89	−1.3 to −0.5	<0.001	−0.03	−0.5 to 0.4	0.904	−0.32	−0.6 to −0.1	0.016
	Low	−0.80	−1.3 to −0.3	0.001	0.03	−0.5 to 0.5	0.899	−0.37	−0.6 to −0.1	0.011
Total area	High	ref			ref			ref
	Med	−0.76	−1.2 to −0.3	0.001	0.19	−0.3 to 0.7	0.428	−0.36	−0.6 to −0.1	0.006
	Low	−0.70	−1.2 to −0.2	0.003	0.25	−0.3 to 0.8	0.343	−0.44	−0.7 to −0.2	0.002
Wall thickness	High	ref			ref			ref
	Med	−0.54	−1.0 to −0.1	0.019	−0.28	−0.7 to 0.1	0.149	−0.08	−0.2 to 0.1	0.270
	Low	−0.39	−0.9 to 0.1	0.121	−0.22	−0.6 to 0.2	0.310	−0.06	−0.2 to 0.1	0.404

Definition of abbreviations: CI = confidence interval; Coef = coefficient; HRCT = high-resolution computed tomographic; ref = reference; tptef/te = ratio of the time to reach peak tidal expiratory flow to the total expiratory time.

^*HRCT parameters were standardized within each airway generation. Therefore, coefficients represent effects of tptef/te tertiles on HRCT parameters across airway generations in SDs.

Discussion

In a long-term birth cohort followed into the fourth decade of life, we found that reduced levels of tptef/te and V̇maxFRC measured in early infancy predicted an increased risk for active asthma up to age 36 years and that the effects of tptef/te and V̇maxFRC on subsequent asthma risk were additive and partly independent from each other. Infant V̇maxFRC (but not tptef/te) was found to predict reduced forced expiratory flow in young adult life. However, infant tptef/te (but not V̇maxFRC) was found to predict reduced airway caliber at FRC by age 26 years. Despite the relatively small number of TCRS infants who had both low tptef/te and low V̇maxFRC, two out of three participants in this group had active asthma in their adult life.

tptef/te is a complex measure of infant lung function that is assessed during tidal breathing and largely dependent on the braking of expiratory flow by inspiratory muscles (15, 16). Thus, reductions of infant tptef/te could also potentially originate from alterations in the neural control of breathing. However, in our study, not only did low infant tptef/te predict subsequent asthma risk, it was also strongly associated with smaller airway diameter and reduced luminal area at FRC in young adult life. This was most evident in the proximal bronchial generations that account for the largest proportion of airway resistance. Although we do not know whether these airway structural characteristics were already present at birth, it is tempting to speculate that central airways may have been congenitally narrower in these participants leading to reduced airway conductance in early infancy. This would have resulted in a lower rate constant of tidal exhalation leading to an accelerated decay of inspiratory muscle tone during expiration (i.e., reduced braking) and hence a lower tptef/te. Interestingly, individuals with low tptef/te had narrower airways at FRC but not TLC, demonstrating increased airway compliance. Reduced conductance with increased airway compliance would be expected to increase susceptibility to wheezing and the development of subsequent asthma (12, 27–30).

Previous studies on CT imaging of airway morphometry in asthma have demonstrated increased wall thickness or percent wall area, often with reduced luminal area in longstanding, severe, or uncontrolled asthma, likely owing to airway inflammation and remodeling (31–33). In our sample, 8 out of 38 (21.1%; Table E4) participants in the HRCT group reported asthma at age 26, but only 1 had symptoms and medication history compatible with moderate to severe, persistent asthma. In a study of the change in luminal area between TLC and FRC, Shim and colleagues (34) demonstrated that, as compared with normal participants and participants with severe asthma, those with milder asthma had increased change in luminal area from TLC to FRC, suggesting that the central airways in mild asthma may actually be more compliant than normal airways. This apparent increase in airway compliance in milder asthma corresponds to our finding of thinner walls and increased compliance in participants with low tptef/te and greater asthma risk.

We also found infant V̇maxFRC to be inversely related to the subsequent risk of active asthma, although these effects were stronger for childhood than for adult asthma. These results are in line with reports from other birth cohorts showing associations between infant V̇maxFRC or forced expiratory flow at 50% of VC (FEF₅₀) and risk of recurrent wheezing, cough, and asthma in childhood (5, 7, 35). In the Perth study, the only other birth cohort with available data on infant V̇maxFRC and clinical outcomes into adult life, levels of V̇maxFRC below the median at 1 month of age were associated with risk of active asthma and deficits in FEV₁/FVC and FEF_25–75 at age 24 years (11). Consistent with that report, we had previously shown that TCRS participants in the lowest quartile of infant V̇maxFRC had reduced levels of FEV₁, FEV₁/FVC, and FEF_25–75 by age 22 years (17). Other studies (36) have demonstrated that only a minority of infants with low V̇maxFRC will go on to have a persistently low trajectory of lung function throughout childhood, indicating the existence of other postnatal factors that influence long-term trajectories of lung function into adult life.

The lack of relationship between FRC and later asthma risk or HRCT findings in our study is not surprising. Studies of infants with chronic lung disease of prematurity have often found only minor alterations in FRC despite significantly reduced V̇maxFRC values (37, 38). Furthermore, Hanrahan and colleagues demonstrated striking reductions in V̇maxFRC in infants born to mothers who smoked during pregnancy, but no difference in FRC (39).

In the present study, for the first time, we tested for and found independent effects of infant tptef/te and V̇maxFRC on risk of active asthma into the fourth decade of life and observed differential effects of low infant tptef/te on CT-assessed airway measures and of low V̇maxFRC on spirometry-defined airflow limitation in adult life. Taken together, these observations suggest that these two indices of infant lung function are associated with asthma risk through partly different mechanisms. It is thus not surprising that in our study, the group of participants with low levels for both infant tptef/te and infant V̇maxFRC—a group that accounted for nearly 20% of our study population—had the highest risk and the best positive predictive value for subsequent asthma. Indeed, two out of three participants who were in the lowest tertile for both indices developed asthma by their adult life. These results provide novel, cogent evidence in support of the fetal origins of asthma, the long-lasting influence of early-life factors on disease risk well into adult life, and the need for early interventions to modify the trajectory of this disease. In this context, interventions aimed to reduce the impact of noxious exposures (e.g., maternal exposure to smoking [40], nicotine [41], and air pollution [42]) on fetal lung development in utero may have significant effects on subsequent asthma risk of the offspring from childhood into adult life. Our findings have also potential implications for risk stratification because they could contribute to the early identification of patients with persistent asthma who are at high risk for long-term sequelae of the disease (43, 44), although the feasibility of programs to measure infant lung function indices derived from both tidal breathing and maximal expiratory flows remains unclear.

Among the limitations of our study is the relatively small number of TCRS participants with available infant lung function. However, this subset was overall representative of the original cohort, and the impact of this limitation was tempered by the magnitude of the associations we observed between infant lung function and adult outcomes. Among the TCRS strengths are the measurement of multiple infant lung function indices, the longest follow-up (36 yr) ever reported for outcomes of infant lung function, the assessment of asthma outcomes in up to 12 follow-up surveys, and the integration of questionnaire, spirometry, and HRCT data in adult life.

In conclusion, low levels of tptef/te and V̇maxFRC in early infancy are associated with an increased risk for active asthma into mid-adult life, and their effects on asthma are partly independent. Infant tptef/te and V̇maxFRC have different relations to airway function and airway structure achieved in adult life, with the former being associated with CT-assessed narrower airways and the latter with spirometry-defined airflow limitation. These observations underscore the long-lasting effects of the early-life origins of asthma, support the value of both tptef/te and V̇maxFRC at birth for early risk stratification of infants at risk of future asthma, and provide novel evidence of the structural airway abnormalities underlying the fetal origins of this disease.