1 INTRODUCTION

The incidence of prematurity has increased and the survival of prematurely born individuals improved.¹ There has, however, been no reduction in respiratory morbidity with up to 60% of infants born at less than 26 weeks of gestational age being diagnosed with bronchopulmonary dysplasia (BPD).¹ As a consequence, more infants who had BPD are surviving to adulthood. Studies have shown that BPD is a predictor of respiratory morbidity in infancy² and childhood.³ The European Respiratory Society (ERS) has recently produced guidelines regarding the follow-up of those who had BPD suggesting they require dedicated care following discharge from the neonatal unit.⁴

A large population study has suggested that poor lung function at 18 years predicts subsequent chronic obstructive pulmonary disease in later adult life.⁵ It is, therefore, important to understand whether BPD within a prematurely born cohort can predict poor lung function after puberty, the latter being the last positive effect on lung development.⁶ We, therefore, aimed to assess the lung function of 16 to 19-year-old born before 29 weeks of gestation who had or had not developed BPD, in a cohort who were routinely exposed to antenatal corticosteroids and postnatal surfactant (more than 90% received both treatments). A study of 123 19-year-old adults born at less than 25 weeks of gestation showed that those who had BPD had poorer airway function (FEV₁) compared to those without BPD.⁷ Seventy-three percent of that study population, however, had been exposed to postnatal corticosteroids, an intervention we have shown to be associated with a dose dependent negative effect on airway function.^{8, 9} Thus, it is important to determine if a diagnosis of BPD per se rather than administration of postnatal corticosteroids results in poorer lung function at follow-up. Additionally, therefore, we will compare the lung function of those who did or did not develop BPD, but did not receive postnatal corticosteroids.

2 METHODS

Infants born at less than 29 weeks of gestational age born between 1998 and 2001, and recruited into the United Kingdom Oscillation Study¹⁰ were invited to attend King's College Hospital NHS Foundation Trust when they were between 16 and 19 years of age. There, they underwent lung function and exercise testing and completed a health questionnaire. The original UKOS trial was granted ethical approval by the South Thames Multicentre Research Ethics Committee. The follow-up study when the participants were 16 to 19 years of age was approved by The North East-Tyne and Wear South Research Ethics Committee respectively. Participants were invited to attend by letter, email and phone call, and gave written consent to take part in this study.

A detailed neonatal history was available from records taken during the original study.¹⁰ Any postnatal steroid exposure was recorded and subsequent correspondence with participating units had clarified the type of postnatal corticosteroid used. Two patients had received hydrocortisone, all the others dexamethasone. All were classified as having been exposed to postnatal steroids. Details regarding length of course have previously been described.⁸ For the purpose of these analyses, BPD was defined as an oxygen requirement at 36 weeks corrected gestational age.

Lung function assessments were done in accordance with ERS/ATS guidelines^11-19 as described previously.²⁰ Spirometry (FEV₁, FVC, FEV₁/FVC, PEF, FEF₂₅, FEF₅₀, FEF_75, and FEF_25-75), plethysmography (TLC_pleth, FRC_pleth, RV_pleth), airways resistance by impulse oscillometry (R at 5 and 20 Hz), diffusing capacity for carbon monoxide of the lungs (D_LCO, D_LCO/VA) and the lung clearance index using sulfur hexafluoride (LCI) were measured. The results were further examined by a respiratory physiologist and deemed acceptable if they adhered to ERS/ATS standards. Lung function results were converted to z scores.^{12, 21, 22} As described previously²³ abnormal lung function was defined as lung function below the 5th centile.²¹ Participants were asked not to take inhalers on the day of the study until the assessments had been completed. If a participant was unwell on the day of assessment, they were asked to rebook their appointment.

Exercise capacity was assessed by a modified shuttle sprint test where participants were asked to travel between cones 10 m apart in time to prerecorded tones where the time between tones became incrementally shorter each minute.²³ Participants were also asked how many hours of exercise they performed each week. Asthma and wheeze were assessed by self-report using a questionnaire.

Puberty was previously assessed by questionnaire in this cohort when they were between 11 and 14 years of age.²⁴ At that assessment, 90% had entered puberty indicated by a Tanner score of two or more.

3 RESULTS

One hundred and fifty young people completed the study. BPD was associated with lower mean gestational age and birthweight and increased use of postnatal steroids (Table 1).

The young people who had had BPD had significantly poorer mean airway function (FEF₇₅, FEF₅₀, FEF₂₅, FEF_25–75, FEV₁, FVC, FEV₁/FVC, PEF, DLCO, DLCO/VA) with mean differences ranging from 0.38 to 0.97 z scores. Differences remained statistically significant for FEV₁, FVC, and FRC_He after adjustment (Table 2). There was a high percentage of young people with abnormal lung function for several assessments, even after adjustment, with most being more than 5% points and 6% over 10% points (Table 2). There was no significant relationship between exercise distance and BPD before or after adjustment and similarly there was no evidence of a significant relationship with self-reported exercise. The proportion reporting wheeze in the past 12 months was significantly greater in those who had had BPD (22% vs. 9.9%) (Table 3).

Forty two percent of those with a neonatal diagnosis of BPD (33/78) and 11% of those without (8/72) were exposed to postnatal corticosteroids. When excluding those who had postnatal corticosteroids, 109 remained and the same pattern of confounders was seen among demographic variables as for the whole sample (Supporting Information: E-Table 1). The sensitivity analysis in those without exposure to postnatal corticosteroids showed that the differences in mean lung function were similar to those observed in the whole sample (Table 4). The sizes of differences were very similar in the unexposed subset to the whole sample. The main difference was the wider CIs due to smaller numbers when those exposed to corticosteroids were removed. The relationships between BPD and exercise were similar in those unexposed to postnatal corticosteroids and not statistically significant. Wheeze and asthma remained more common among those who had had BPD, but this was not statistically significant (Supporting Information: E-Table 2).

4 DISCUSSION

We have demonstrated that a diagnosis of BPD was associated with poorer lung function in adolescence, including in those who had not received postnatal corticosteroids. Furthermore, those with BPD were more likely to have abnormal lung function deemed clinically significant, for example, 38% of participants had an abnormal value for FEV₁ compared to 22% of those who did not have BPD. Approximately in one half of the lung function results there were differences between those with BPD and no BPD greater than 10% points (Table 5).

Our data showed deficits in airway function (FEV₁ and FEF₇₅) and vital capacity (FVC) were associated with a diagnosis of BPD which remained significant after correcting for possible confounders. A study of 164 25-year-old born before 28 weeks of gestation assessing airway function by spirometry demonstrated significantly worse lung function in those who had had BPD (mean difference in FEV₁ −0.66, 95% CI −0.99 to −0.33, p<0.001).²⁸ There was also more evidence of airways obstruction (FEV₁:FVC) (−0.55 95% CI −0.92 to −0.19, p= 0.003), but with preserved FVC. In that study, only 74% were exposed to antenatal steroids and 40% received postnatal surfactant. In our study, over 90% of those with or without BPD were exposed to antenatal steroids and received postnatal surfactant. The routine use of antenatal steroids and postnatal surfactant may have altered the subsequent course of lung development²⁹ and explain the relative lack of obstructive lung disease seen in our study. In another study of adults aged 19 years born prematurely in the post surfactant era, there was significantly poorer airway function (FEV₁) and obstruction (FEV₁/FVC) in those with BPD compared to those without.⁷ In that cohort 76% of those studied had received postnatal corticosteroids compared to less than 30% in our cohort. We have previously shown that postnatal corticosteroid exposure was associated with poorer FEV₁/FVC and this may explain the differences in the results seen between the two cohorts.^{8, 9} Approximately 10% in both groups had abnormal RV/TLC ratios, this could be suggestive of early signs of physiology (gas trapping) consistent with COPD.

We found that 16% of participants without BPD and 37% with BPD were found to have clinically abnormal lung function. The higher rates of abnormal lung function in those with BPD compared with those without were observed alongside more than a two-fold significant increase in the proportions reporting wheeze, but there were no significant differences in the proportions diagnosed with asthma. Those data are in keeping with the results of 24 young adults who had a diagnosis of BPD compared to 63 without when studied at 26 years of age.³⁰ In that study, there were significant differences in FEV₁, FVC and FEV₁:FVC between the two groups, but there was no significant difference in the proportion who had asthma (30% vs. 25%, p = 0.62).

Our study has strengths and some limitations. We did not include a term controls, but compared comprehensive lung function test results in those born prematurely who had or had not had BPD. We felt our study could be underpowered to undertake a subanalysis in only those who received postnatal corticosteroids, but in the cohort overall there were significant differences in lung function between the BPD and the no BPD groups after adjustment for confounders including postnatal corticosteroids. Unfortunately, we did not have data on possible post NICU confounders, such as RSV infection, but we did adjust for a number of other factors including antenatal steroid use, birthweight, gestation, sex of infant, postnatal corticosteroids, maternal smoking in pregnancy and the participant's age at assessment. After that adjustment, there were still significant differences in lung function, being poorer in those who had BPD.

We demonstrated that 16% of participants without BPD had abnormal lung function indicating that prematurity alone impacts on lung function. This is in keeping with another study which showed that in adulthood FEV₁ and FVC z scores were lower in those born prematurely without BPD compared to term controls.⁷

In conclusion, we have demonstrated that those who had had BPD compared to those who had not were more likely to have abnormal lung function at 16 to 19 years of age. Those results were independent of postnatal corticosteroid exposure.

AUTHOR CONTRIBUTIONS

Anne Greenough and Christopher Harris designed the study. Christopher Harris, Alan Lunt, and Samuel Morris collected the data. Christopher Harris, Anne Greenough, and Janet Peacock undertook the analysis. Christopher Harris wrote the initial draft. All authors (except Alan Lunt who has sadly died) approved the final version of the manuscript.

CONFLICT OF INTEREST

The authors declare no conflict of interest.