Lung Function of Adults Born at Very Low Birth Weight

The New Zealand 1986 VLBW cohort consists of survivors of the 413 VLBW infants who were live born and admitted to a neonatal unit in New Zealand that year, all of whom were included in a prospective audit of retinopathy of prematurity.¹⁰  Of these infants, 338 (82%) survived to be discharged from the hospital. Between March 2013 and October 2016, 250 cohort members (77% of the 323 known survivors), now aged 26 to 30 years, participated in a follow-up study, and 229 VLBW adults, together with 100 term controls, underwent assessment in Christchurch over 2 days, covering physical and mental health, physiologic assessments, and cognitive, neuropsychological, and social functioning (Supplemental Fig 2). Available perinatal data for the VLBW cohort included receipt of antenatal corticosteroids (ANS), days of assisted ventilation via an endotracheal tube, and a diagnosis of BPD (defined as receiving supplementary oxygen at 36 weeks’ postmenstrual age). Exogenous surfactant was not available in 1986. Participants in the control group, first recruited at 22 to 23 years, were born at term in New Zealand in 1986 and not admitted to a neonatal unit.¹⁰  Ethics approval was granted by the Upper South B Regional Ethics Committee, New Zealand. Written informed consent was given by all participants.

The assessments included anthropometric measurements (height was measured by using a Harpenden stadiometer; weight was measured by using electronic scales) and a comprehensive questionnaire of health and welfare, which included smoking history and other items that were based on the European Community Respiratory Health Survey II.¹¹  Lung function tests were undertaken at the Respiratory Physiology Laboratory, Christchurch Hospital, which is accredited by the Thoracic Society of Australia and New Zealand. All tests were performed on 1 day in the following order: spirometry, plethysmographic lung volumes, single-breath diffusing capacity of the lung for carbon monoxide (D_LCO) (all with Autobox V62J; CareFusion, Yorba Linda, CA), and single-breath nitrogen washout (SBN₂) (SensorMedics Vmax Spectra exercise test station; CareFusion). The tests were performed with the subject sitting and with a nose clip in place; respiratory physiologists were blinded as to the participants’ group. Testing protocols adhered to the American Thoracic Society and European Respiratory Society pulmonary function standards^12–15  and interpretation.¹⁶  D_LCO gas mixture contained 0.3% carbon monoxide and 0.3% methane with balance air. D_LCO values were corrected for venous blood hemoglobin. Only technically acceptable and repeatable data were included in data analyses. Values for spirometry, lung volumes, and D_LCO were presented as z scores and percentages of the predicted (% pred) on the basis of the reference values from the Global Lung Function Initiative (GLI) (2012),¹⁷  Quanjer et al (1993)¹⁸  and GLI (2017),¹⁹  respectively. The SBN₂ results were reported as the mean of the measured phase III slopes.

Between-group comparisons were conducted by using the independent samples t test for differences in means and the χ² test for comparison of percentages. Mean between-group differences in lung function were adjusted for covariate factors by using multiple linear regression methods. With a nominal P value of .05, the study had 80% power to detect a mean difference in lung function between VLBW participants and controls in excess of 0.34 SD. For subgroup comparisons (eg, by BPD status within the VLBW cohort), the corresponding effect size estimates ranged between 0.40 and 0.45 SD. A Bonferroni correction was applied to account for multiple significance testing across the lung function measures, taking into account the average correlation between outcomes. The Bonferroni adjusted P value was P < .009. Statistical analyses were conducted by using SAS 9.4 (SAS Institute, Inc, Cary, NC), and graphing of outcomes were conducted by using GraphPad Prism version 7.

Of the 323 known VLBW survivors, there were few differences between those assessed and not assessed in demographic and perinatal variables, except those assessed had a lower mean birth weight but were less likely to have moderate or severe disability at 7 to 8 years (Supplemental Table 6). A total of 228 VLBW survivors and 100 controls underwent lung function testing with valid data obtained from 226 VLBW survivors and all controls. The demographic and anthropomorphic data for VLBW adults and controls are shown in Table 1. Compared with controls, both men and women in the VLBW group were shorter (P < .001), but only women had lower weight (P < .001) and BMI (P = .022). Participants in the VLBW group were more likely to have ever smoked (P = .034). Among the VLBW group, 128 (57%) had received ANS, 72 (32%) were small for gestational age (SGA) (birth weight <10% percentile), and 46 (20%) were diagnosed with BPD. Self-reported respiratory symptoms are shown in Table 2, with the only statistically significant difference between the groups being a higher incidence of any wheeze in the past 12 months in the VLBW cohort.

Compared with controls, VLBW adults showed significantly lower z scores for forced expiratory volume in 1 second (FEV₁), FEV₁/forced vital capacity (FVC) ratio (FEV₁/FVC), and forced expiratory flow at 25% to 75% of forced vital capacity (FEF_25–75), whereas the difference in FVC z scores was not statistically significant (Table 3, Fig 1). The spirometric pattern is shown in Table 4; 63% of VLBW participants had normal spirometry versus 82% of controls (P = .0004). The predominant abnormality in VLBW participants was obstruction (35% vs 14% of controls), which was mild in 66 (30%) and more severe in 13 (5.5%) compared with 2% of control participants.

For lung volumes, there were no significant differences in the z scores for total lung capacity (TLC) or functional residual capacity (FRC) between the groups. However, higher z scores in residual volume (RV) and RV/TLC ratio were found in the VLBW group compared with controls (Table 3, Fig 1).

The VLBW group had significantly lower z scores for D_LCO and transfer coefficient for carbon monoxide (K_CO) than controls. The z scores for alveolar volume (V_A) were not significantly different between the groups. The phase III slope in SBN₂ in the VLBW group was significantly higher than controls. All of these lung function differences persisted when adjusted for sex and smoking status (Table 3, Fig 1). Between-group differences in lung function revealed the same statistical significance when variables were analyzed as percent predicted values (Supplemental Table 7).

Within the VLBW group, z scores for FEV₁, FEV₁/FVC, and FEF_25–75 were all significantly lower in subjects with BPD versus subjects without BPD (Table 5). There was no between-group difference in z score for TLC, FRC, or FRC/TLC after correction for multiple comparisons, but RV and RV/TLC were higher in subjects with BPD than in subjects without BPD. D_LCO was lower in subjects with BPD but not statistically significant on Bonferroni correction; V_A and K_CO were not significantly different. The phase III slope was significantly higher in subjects with BPD than in subjects without BPD. These effects also persisted when adjusted for sex, smoking status, birth weight (<1000 g vs ≥1000 g), gestation (<28 weeks vs ≥28 weeks), SGA status, and maternal smoking during pregnancy. The patterns remained the same when data were analyzed as percent predicted values (Supplemental Table 8). Duration of ventilation was significantly correlated with BPD (r = 0.51). Comparing ventilated and nonventilated subjects revealed broadly similar results to the BPD and non-BPD comparisons (Supplemental Table 9).

Pulmonary function comparisons between SGA and non-SGA VLBW adults revealed nominally significant differences of lower FEV₁/FVC and FEF_25–75 z scores and higher phase III slope in the SGA group, but none remained significant on Bonferroni correction (Supplemental Table 10). Restricting the analysis to those <30 weeks’ gestation to reduce any sampling bias did not alter these results.

Differences in lung function between extremely low birth weight (ELBW) (birth weight <1000 g) subjects and other VLBW participants were small and did not reach statistical significance when adjusted for multiple comparisons (Supplemental Table 11).

There were no differences in spirometry and lung volumes between extremely preterm (EP) (gestation <28 weeks) and other VLBW participants (Supplemental Table 11). D_LCO and K_CO were significantly lower in EP survivors, but there was no difference in V_A or phase III slope.

There was no statistically significant difference in any lung function parameter between those VLBW survivors who received ANSs and those who did not (Supplemental Table 9).

In this comprehensive assessment of static lung function in a national cohort of VLBW survivors at age of 26 to 30 years, we found that compared with term-born controls, VLBW adults showed a higher incidence of airflow obstruction, gas trapping, reduced gas exchange, and higher ventilatory inhomogeneity. The majority of VLBW survivors did however have normal spirometry. Within the VLBW group, participants with a history of BPD had further reduced airflow, increased gas trapping, and ventilatory inhomogeneity compared with those without BPD. These findings extend and enhance our understanding of lung function in adult VLBW survivors by incorporating lung volumes and diffusing capacity, which are rarely reported, and measurement of ventilatory inhomogeneity by SBN₂, which is novel.

Numerous published spirometry data,⁴  including a recent individual patient data meta-analysis,⁷  reveal ongoing expiratory airflow obstruction in survivors born VLBW or VP in childhood through early adulthood. Airflow obstruction in VP children is associated with abnormal lung structure changes, including subpleural opacities, bronchial wall thickening, and hypoattenuated lung areas, as detected by computed tomography scan.²⁰  Our results are broadly consistent with the literature, with significantly lower z scores for parameters reflecting airflow (FEV₁, FEV₁/FVC, and FEF_25–75) in adult VLBW survivors; and these values were further reduced in those with a history of BPD. Similar results in those with SGA compared with non-SGA participants did not remain significant after Bonferroni correction. The finding of reduced airflow with stable FVC indicates reduced airway caliber,²⁰  and reduced FEF_25–75 suggests that obstruction presents in distal small airways. In our VLBW cohort, 63% had normal spirometry; airflow obstruction was generally mild with only a small proportion having more severe airflow obstruction. We found no association between either male sex or ANS and spirometry results, in contrast to a recently published meta-analysis of late adolescence and adult cohort studies.⁷  Male sex is associated with increased mortality in VLBW and VP infants²¹  and with increased BPD in survivors,²²  whereas ANS increase survival but have little impact on rates of BPD.²³  The effects on survival may well mask any longer-term consequences.

Studies assessing lung volumes in former VLBW and/or VP subjects in childhood through late adolescence have provided conflicting results.^20,24–27  Reports in adulthood, however, are rare. Our data revealed increased gas trapping, demonstrated by significantly elevated RV and RV/TLC ratio in VLBW adult survivors compared with controls, and in those with BPD compared with those without. These findings are physiologically consistent with our spirometry data, which reveal more airflow obstruction in these participants than in their counterparts, and with published lung volume data in VP children.^27,28 

Although there are several studies reporting on gas exchange (D_LCO) in former VLBW and/or VP infants in childhood, with conflicting results, there have been few studies in adults.²⁹  Narang et al³⁰  reported preterm subjects (<38 weeks) had significantly reduced D_LCO compared with term controls at 21 years of age. Landry et al³¹  reported preterm subjects (<37 weeks) with BPD had significantly reduced D_LCO than other preterm subjects or term controls at 22 years of age. Our data revealed reduced D_LCO and K_CO in VLBW adults compared with controls, although significance was no longer found in those with BPD compared with no BPD after Bonferroni correction. These findings support the hypothesis that damaged alveolar compartments and pulmonary vasculature resulting in reduced gas exchange, as previously reported in VLBW children, persist into adulthood, with more evidence required to assess the importance of BPD.

Using the lung clearance index, a few researchers have investigated ventilatory inhomogeneity in preterm infants and children with inconsistent results.²⁹  To our knowledge, this study is the first used to report SBN₂ in VLBW adults. We have demonstrated evidence of increased ventilatory inhomogeneity, as determined by significantly elevated phase III slope, in VLBW survivors and in those with BPD.

Taken together with previous studies, there are several implications from our results. By the end of the third decade of life, these young adults will have already passed their peak lung function,^4,8,17  and anything that may hasten respiratory function decline should be avoided. Chief among those is smoking, and it is disappointing that 31% of the VLBW cohort (and 21% of controls) were current smokers. Neonatologists and pediatricians engaged in follow-up clinics should emphasize the importance of VLBW and/or VP adults not smoking. Annual influenza immunization should be recommended⁸  and not just for individuals with a history of BPD. The arguments that everybody should be free of exposure to secondhand smoke and air pollution are particularly pertinent for VP adults.^8,9  There might also be a case for routine lung function testing of VLBW and/or VP survivors in middle childhood with targeted messages for those identified as already having abnormal lung function. With increasing evidence of long-term health consequences after VP birth, it is important that clinicians in all settings caring for adolescents and young adults ask about gestational age^8,32 ; yet a recent survey suggested that respiratory specialists did so infrequently.³³  Taking an appropriate history should alert practitioners, and VP adults with respiratory symptoms should be referred for detailed lung function tests.

Although neonatal care has continued to advance with improved survival rates for VP infants compared to 1986, BPD is still a common outcome, albeit with perhaps a slightly different pathologic basis and milder manifestation in the neonatal period³⁴  and beyond.³⁵  Lung function tests in midchildhood in survivors with this “new BPD” revealed increased airway obstruction compared with EP survivors without BPD and with term controls, similar to follow-up studies from the presurfactant era.³⁶  Although the longer-term outcome of these individuals is unclear, it seems probable that the results from our cohort will have relevance for contemporary neonatal care survivors.³⁷ 

There is a need for ongoing monitoring of lung function and functional status as this and other similar cohorts age to assess the pattern of physiologic and functional decline. Of interest will be whether lung function decline is uniform or whether 2 distinct groups emerge. If the rate of decline in lung function is accelerated in the group identified as having abnormality in childhood, this allows targeted interventions in this group and reassurance to others. Currently, only a minority of VLBW adults in this cohort have any (generally mild) respiratory symptoms. But the literature suggests that in addition to obstructive airways disease, which is not always reversible, structural lung changes of emphysema may lead to future symptoms.⁸  Our finding of increased ventilatory inhomogeneity is consistent with this scenario, and there may be more information available from cardiopulmonary exercise testing. Additionally, the lung changes associated with VP birth coupled with the effects on cardiac and pulmonary vascular development could increase the risks of pulmonary hypertension in adult survivors.^9,38,39 

Strengths of our study include the fact that this is a national population-based cohort prospectively enrolled at birth in a single year and assessed in 1 center with a high participation rate, and is thus able to give a more complete picture of the respiratory phenotype of VP adults than hospital-based studies or those focused only on graduates with BPD. We performed a range of lung function tests using the latest GLI reference equations. Although born before the availability of surfactant, more than half of the cohort had been exposed to ANS, and both assisted ventilation and parenteral nutrition were routinely available. Weaknesses include the fact that we did not assess bronchial responsiveness to direct stimulation (eg, methacholine) or the response to bronchodilators. High-resolution computed tomography scans would be desirable to further delineate structural lung abnormalities.⁸  The cohort was recruited on the basis of birth weight, which was usual in the 1980s, but the inclusion of more mature infants of 32 weeks’ gestation and greater, many of whom were SGA, could have affected the overall results. Although we assessed the cohort at 7 to 8 years of age, when questionnaires included evidence of wheeze and a diagnosis of asthma (52% VLBW versus 27% contemporary controls),⁴⁰  we did not conduct lung function testing at that time and therefore cannot report longitudinal data for these outcomes. Lastly, although this is a national cohort and the largest data set reported so far,⁷  our numbers are still relatively small.

The current study reveals a higher incidence of airflow obstruction, gas trapping, reduced gas exchange, and increased ventilatory inhomogeneity in adult VLBW survivors versus age-matched controls; subjects with BPD showed further worsened airflow obstruction compared with those without BPD. The findings suggest the pulmonary effects due to VLBW, although mild, persist into adulthood, and BPD is a further insult on small airway function. Our data highlight the importance of avoiding factors that may accelerate lung function decline, particularly smoking, in VLBW survivors. Routine lung function testing and ongoing monitoring are imperative in VLBW and/or VP survivors, with targeted intervention for those with abnormal lung function.

We are grateful to all participants in this study for their enthusiastic support of the assessment days. In addition, we acknowledge the support of scientific and medical staff in the Respiratory Physiology Laboratory at Christchurch Hospital. We thank the study project manager, Julia Martin, for all her hard work and organizational skills.

This article is dedicated to the memories of Dr Maureen Swanney and Josh Stanton, senior members of the research team and dearly loved respiratory physiologist colleagues, both of whom died during the course of the study.

ANS

antenatal corticosteroids

BPD

bronchopulmonary dysplasia

D_LCO

single-breath diffusing capacity of the lung for carbon monoxide

ELBW

extremely low birth weight

EP

extremely preterm

FEF_25–75

forced expiratory flow at 25% to 75% of forced vital capacity

FEV₁

forced expiratory volume in 1 second

FRC

functional residual capacity

FVC

forced vital capacity

GLI

Global Lung Function Initiative

K_CO

transfer coefficient for carbon monoxide

RV

residual volume

SBN₂

single-breath nitrogen washout

SGA

small for gestational age

TLC

total lung capacity

V_A

alveolar volume

VLBW

very low birth weight

VP

very preterm

% pred

percentage of the predicted