Article Text

Perinatal factors associated with long-term respiratory sequelae in extremely low birthweight infants
  1. Katsuya Hirata1,
  2. Masahiro Nishihara1,
  3. Jun Shiraishi1,
  4. Shinya Hirano1,
  5. Katsura Matsunami1,
  6. Kiyoaki Sumi1,
  7. Norihisa Wada1,
  8. Yutaka Kawamoto1,
  9. Masanori Nishikawa2,
  10. Masahiro Nakayama3,
  11. Tadahiro Kanazawa4,
  12. Hiroyuki Kitajima1,
  13. Masanori Fujimura1
  1. 1Department of Neonatal Medicine, Osaka Medical Center and Research Institute for Maternal and Child Health, Izumi, Osaka, Japan
  2. 2Department of Radiology, Osaka Medical Center and Research Institute for Maternal and Child Health, Izumi, Osaka, Japan
  3. 3Clinical Laboratory Medicine and Anatomic Pathology, Osaka Medical Center and Research Institute for Maternal and Child Health, Izumi, Osaka, Japan
  4. 4Department of Comparative and Developmental Psychology, Graduate School of Human Sciences, Osaka University, Suita, Osaka, Japan
  1. Correspondence to Dr Katsuya Hirata, Department of Neonatal Medicine, Osaka Medical Center and Research Institute for Maternal and Child Health, 840 Murodo-cho, Izumi, Osaka 594-1101, Japan; khirata0513{at}


Objective To assess lung function at 8 years old in extremely low birthweight (ELBW) survivors and to identify perinatal determinants associated with impaired lung function.

Design Retrospective cohort study.

Setting Level III neonatal intensive care unit.

Patients ELBW survivors born in 1990–2004 with available spirometry at 8 years old were studied. Children were excluded if they had a Wechsler Intelligence Scale for Children Third Edition full IQ <70.

Main outcome measures Multivariate logistic regression analysis was used to identify perinatal determinants associated with airway obstruction (forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio <80%) at school age and the predictive power of potential determinants. Potential risk factors and predictors assessed in this study were gestational age, birth weight, small for gestational age, sex, chorioamnionitis, premature rupture of membranes, antenatal steroids, surfactant administration, respiratory distress syndrome, postnatal steroids, severe bronchopulmonary dysplasia and bubbly/cystic appearances of the lungs by X-ray during the neonatal period.

Results Of 656 ELBW survivors, 301 (45.9%) had attended a school-age follow-up at 8 years old. A total of 201 eligible children completed the lung function test. Bubbly/cystic appearance of the lungs (OR 4.84, 95% CI 1.26 to 18.70) was associated with a low FEV1/FVC ratio. Children with bubbly/cystic appearance had characteristics of immaturity and intrauterine inflammation.

Conclusions Within a cohort of ELBW infants, a bubbly/cystic appearance of the lungs in the neonatal period was the strongest determinant of a low FEV1/FVC ratio at school age.

  • Neonatology
  • Respiratory

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic?

  • Lung function at school age can be reduced in extremely low birthweight (ELBW) survivors, especially those who suffer from bronchopulmonary dysplasia. However, the importance of other perinatal characteristics on lung function remains uncertain.

What this study adds?

  • In school-age children who were born ELBW, a bubbly/cystic appearance in the lungs in the neonatal period was the strongest determinant of a low forced expiratory volume in 1 s/forced vital capacity ratio. This radiological finding is associated with immaturity and intrauterine inflammation in ELBW infants.


Because of advances in perinatal management,1–3 there has been a remarkable improvement in the rate of survival and short-term prognosis of extremely low birthweight (ELBW) infants with various complications. Evaluation of the quality of life in survivors of these infants has increased in importance. Despite widespread use of antenatal steroids and postnatal surfactant replacement therapy, reports have shown that lung function is impaired at school age in ELBW survivors, especially those who suffer from bronchopulmonary dysplasia (BPD).4–11 Additionally, these ELBW survivors have a lower forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio, consistent with airway obstruction,4–11 which may greatly decrease their quality of life. However, the importance of other perinatal factors on lung function of ELBW survivors remains uncertain.

Osaka Medical Center and Research Institute for Maternal and Child Health (OMCRMCH) has provided school-age follow-up, including lung function tests for ELBW survivors since the early 1990s. Using this large amount of accumulated follow-up data linked with precise perinatal data, we conducted a retrospective study to evaluate lung function in ELBW survivors of school age at a postsurfactant timepoint. We hypothesised that there are important perinatal factors other than BPD that predict impaired lung function at school age. This study aimed to assess lung function and to identify perinatal determinants associated with impaired lung function at school age in ELBW survivors.

Patients and methods

Subjects and setting

Children eligible for the study were ELBW survivors who were born and admitted to the neonatal intensive care unit (NICU) at OMCRMCH between 1 June 1990 and 31 May 2004 and attended biyearly school-age follow-up from 1999 to 2011 at 8 years old. ELBW was defined as a birth weight (BW) <1001 g. Our school-age follow-up involved various fields of specialists, including anthropometry, the Wechsler Intelligence Scale for Children Third Edition (WISC-III) IQ test and a lung function test. Children were excluded if they had a WISC-III full IQ <70 at the time of the lung function test to reduce variance due to poor cooperation.

School-age factors

Height and weight were measured with heavy clothing and shoes removed at the school-age follow-up. The WISC-III IQ test was performed by clinical psychologists and full-scale IQ was recorded. Spirometry was measured by trained neonatologists who were blinded to the perinatal outcomes including X-ray findings using portable spirometers (1999–2010: AS-502, Minato Medical Science, Osaka, Japan; 2011: ARFEL III, AiVision, Tokyo, Japan). FVC, FEV1 and forced expiratory flow at 50% and 75% were obtained from maximum expiratory flow–volume curves. During spirometry, attempts were made to achieve at least three acceptable and two repeatable forced expiratory manoeuvres. Results of spirometric data are expressed as per cent predicted for age, height and sex relative to Japanese children.12 A decrease in spirometric parameters below their relevant 5th centiles is thought to be consistent with pulmonary defects according to the American Thoracic Society and European Respiratory Society consensus.13 According to reference equations for spirometry in Japanese children aged 6–18 years, cut-off values equivalent to the 5th centile were 80% for FVC predicted and the FEV1/FVC ratio.12 Therefore, we defined a restrictive abnormality as FVC <80% predicted with an FEV1/FVC ratio ≥80%, obstructive abnormality as an FEV1/FVC ratio <80% with FVC ≥80% predicted and mixed abnormality as FVC <80% predicted with an FEV1/FVC ratio <80%.

Perinatal factors

Perinatal and neonatal data were collected from the discharge summary report of children and mothers, and was confirmed by a detailed chart review. Information collected included demographic data, maternal, prenatal and perinatal history and delivery, and neonatal outcomes, such as gestational age (GA), BW, small for gestational age (SGA), sex, maternal age, mode of delivery, gravidity, multiple birth, antenatal steroids, premature rupture of the membranes (PROM), pregnancy-induced hypertension, histological chorioamnionitis (CAM), cord blood immunoglobulin M (IgM), Apgar scores, respiratory distress syndrome (RDS), surfactant administration, BPD, air leak syndrome/pulmonary interstitial emphysema, radiologically bubbly/cystic appearance in the lungs (hereafter ‘bubbly/cystic appearance’), necrotising enterocolitis, patent ductus arteriosus, retinopathy of prematurity, sepsis, severe intraventricular haemorrhage, cystic periventricular leukomalacia, inhaled nitric oxide, postnatal steroids, duration of oxygen therapy and ventilatory support, postmenstrual age (PMA) when oxygen therapy and ventilatory support were discontinued and oxygen at discharge. BPD was defined and classified according to the National Institute of Child Health and Human Development (NICHD) consensus definition (see online supplementary data 1).14 Other detailed definitions of perinatal factors are listed in online supplementary data 1.

Definition of bubbly/cystic appearance in lungs

A bubbly/cystic appearance in the lungs was defined as diffuse, streaky infiltrates with small cystic areas (1–10 mm in diameter) on day 28 (±3 days) from chest radiographs based on a previously published description.15–17 A typical radiographic bubbly/cystic appearance is shown in figure 1. In our experience of treating premature babies, we observed that many who develop a bubbly/cystic appearance within a few weeks eventually develop severe chronic lung disease.17 ,18 Based on this preliminary observation, we included a bubbly/cystic appearance on day 28 chest X-ray, in addition to the NICHD consensus definition for registering and classifying chronic lung disease,16 to establish a national database of very-low-birthweight infants.3 Thus, all infants were X-rayed on day 28 (±3 days) in our study cohort and a bubbly/cystic appearance was included as an important perinatal factor to be analysed. Films of chest radiographs were reviewed and a diagnosis was made by a paediatric radiologist who was unaware of the patient data. Radiological findings other than a bubbly/cystic appearance such as pneumonia were not found on day 28 X-rays in our cohort.

Figure 1

Day 28 chest radiograph of a patient with bubbly/cystic appearance in lung.

Statistical analysis

A comparison of the different groups with respect to baseline characteristics and outcomes was performed. The χ2 test or Fisher's exact test for categorical variables, Student t test, Mann–Whitney non-parametric U-test or Kruskal–Wallis test followed by the Steel–Dwass multiple comparison test for continuous variables were used as appropriate. The Jonckheere–Terpstra test was used to assess trends in lung function variables across the four groups with varying BPD classification.

Logistic regression analysis was used to identify perinatal determinants associated with airway obstruction at school age and the predictive power of potential determinants. The primary outcome was an FEV1/FVC ratio <80% at 8 years old. Theoretically possible potential risk factors and predictors were GA, BW, SGA, sex, CAM, PROM, antenatal steroids, surfactant administration, RDS, postnatal steroids, severe BPD and bubbly/cystic appearance by X-ray.4–11 ,19–21 Findings are presented as unadjusted results and results adjusted for all possible predictors, age and birth cohort. Detailed methods of sensitivity power calculation to estimate the minimal detectable OR (MDOR) and multiple imputations are described in online supplementary data 1.

Data management and statistical analyses were performed with PASW statistics V.18 (SPSS, Chicago, Illinois, USA), R statistical software V.2.13.0 (R Foundation for Statistical Computing, Vienna, Austria) and G*Power V. ( All reported p values are two-sided; p<0.05 was considered statistically significant.


Population characteristics

A total of 765 ELBW infants were admitted to the NICU over the 14-year study period. Of these, 656 (85.8%) were discharged alive and 301 (45.9%/656 subjects) attended the school-age follow-up at 8 years old. Of the 278 ELBW children who performed the WISC-III test and the lung function test, 201 had a full IQ ≥70 with a valid lung function test (see online supplementary data 2). Therefore, we analysed these 201 children in detail as the ELBW study cohort. All of these 201 children were ethnic Japanese, and the median age at the time of the follow-up was 8.02 years (range, 7.02–9.25 years). With the exception of a lower follow-up rate in men (40.8% vs 54.3%) in the study cohort, there were no significant differences in descriptors or neonatal pulmonary morbidities between the ELBW study cohort (n=201) and ELBW survivors who were not included in the study (n=455, table 1).

Table 1

Descriptors and neonatal pulmonary morbidity between the ELBW study cohort and ELBW survivors not included in the study

Lung function variables in ELBW children with respective BPD classification

In this study cohort, 40.8% (82/201) of ELBW survivors required supplemental oxygen at 36 weeks' PMA. Lung function results with respective BPD classification are shown in table 2. After adjustment for age, sex and height using a reference for Japanese children,12 children who were born ELBW had a lower baseline spirometry, as shown by a reduction in the per cent predicted compared with the reference for Japanese children. There was a trend towards a decreased level of all lung function variables, except for FEV1/FVC ratio, as the grade of BPD increased.

Table 2

Lung function variables in ELBW survivors with respective NICHD BPD classification14

Determinants of impaired lung function in children born ELBW

In total, 27 (13.4%) of the 201 children had restrictive abnormalities, 36 (17.9%) had obstructive abnormalities and 11 (5.5%) had mixed abnormalities at the age of 8 years (see online supplementary data 3). Among the four lung function categories assessed, mixed abnormalities, characterised by the coexistence of obstruction and restriction,13 were associated with various perinatal characteristics, such as PROM (p=0.005), severe BPD (p=0.002), bubbly/cystic appearance (p<0.001), postnatal steroids (p=0.04), duration of oxygen therapy (p=0.002) or mechanical ventilation (p<0.001), PMA when oxygen therapy (p<0.001) or mechanical ventilation (p<0.001) were discontinued, and oxygen at discharge (p<0.001), with an inverse correlation with RDS (p=0.03) and surfactant administration (p=0.007).

For logistic regression analysis, nine children with missing data (4.5%/201 ELBW cohort) were excluded to perform a completed case analysis. GW (<26 weeks), CAM, PROM, surfactant administration, RDS, severe BPD and bubbly/cystic appearance showed a modest association with a low FEV1/FVC ratio in univariate analysis (table 3). After adjustment for all other factors in table 3, bubbly/cystic appearance was the only statistically significant predictor (OR 4.84, 95% CI 1.26 to 18.70). Surfactant administration, PROM, RDS, GW (<26 weeks) and severe BPD were not independent predictors in multivariate analyses; however, ORs of these predictors suggested moderate effects on a low FEV1/FVC ratio and were below an MDOR of 3.6. The concordance index in the multivariate model was 0.81, suggesting suitability of the model for discrimination.

Table 3

Multivariate logistic regression analysis to estimate the predictive power of independent variables with a low FEV1/FVC ratio

The analysis using multiple imputations provided similar associations to those by complete case analysis where a bubbly/cystic appearance was the only independent predictor of a low FEV1/FVC ratio (OR 5.70, 95% CI 1.75 to 18.59, see online supplementary data 4).

Perinatal characteristics of children with bubbly/cystic appearance in the lungs during the neonatal period

We then focused on bubbly/cystic appearances in the lungs as an important factor to predict lung function disorders and examined the perinatal characteristics associated with this condition. Bubbly/cystic appearance was associated with low GA (p<0.001), low BW (p=0.01), PROM (p<0.001), CAM (p<0.001), elevation of cord blood IgM (p=0.01), lower occurrence of SGA (p=0.002) and absence of RDS (p=0.001, table 4). Use of these parameters indicated immaturity and intrauterine inflammation.

Table 4

Perinatal risk factors for developing bubbly/cystic appearances in the neonatal period


Within a cohort of ELBW children born in the era after the introduction of surfactant, lung function at 8 years of age was significantly impaired compared with a reference for Japanese children (table 2).12 Consistent with previous reports,4–11 ,14 ,23 ,24 we confirmed that the NICHD classification of BPD14 was useful for predicting the impairment of long-term lung function (table 2).

The strongest perinatal factor related to an FEV1/FVC ratio <80% (OR 4.84, 95% CI 1.26 to 18.70, table 3) in the present study was a bubbly/cystic appearance on day 28. The association of an FEV1/FVC ratio <80% with surfactant administration, PROM, RDS, low GW, severe BPD or others was not statistically significant after adjustment. However, these results should be interpreted cautiously given the small sample size of the current study. Future studies using a larger sample size might identify other variables as independent predictors. The utility of chest radiograph criteria compared with the duration of supplemental oxygen criteria for predicting pulmonary outcomes is controversial.25–29 In contrast to previous studies, we focused on an apparent bubbly appearance for the diagnosis of bubbly/cystic appearance, which might have enhanced the significance of the study findings.

In the present ELBW study cohort, 10.9% (22/201 subjects) had bubbly/cystic appearance. Immaturity and absence of RDS, which are a background characteristic in the bubbly/cystic appearance group (table 4), are similar to the symptoms observed in Wilson–Mikity syndrome (WMS), consistent with previous studies.15 ,18 ,30–34 Relatively large case series reports of WMS with bubbly/cystic appearance were published in North America in the 1960s: 22 cases by Swyer et al33 and 34 cases by Hodgman et al.15 Recently, nine cases of WMS were reported by Hoepker et al from the Canadian NICU.34 Exact percentages of WMS in the above case series reports are unknown because of a limited description of their prevalence. Thus, to what extent our findings that the bubbly/cystic appearance may be a strong predictor of long-term lung sequelae in ELBW survivors can be generalised to ELBW populations in other countries remains uncertain. However, because similar symptoms were also observed in countries other than Japan, focusing on a bubbly/cystic appearance in ELBW infants and evaluating their long-term outcome might be of benefit.

Other background characteristics in the bubbly/cystic appearance group in the present study were intrauterine inflammation such as CAM, PROM and elevation of IgM levels. These results are consistent with our previous reports.17 ,18 A recent report from the UK35 also described a bubbly/cystic appearance in extremely premature infants similar to our cases, and suggested a close association with colonisation by Ureaplasma urealyticum. There is increasing concern that the diagnosis of BPD, simply defined by a need for oxygen supplementation, provides limited information about the pathophysiology, disease progression or variability of lung pathology.36–38 Further studies focused on the bubbly/cystic appearance and intrauterine inflammation may help reveal the aetiology of lung pathology in extremely premature infants.

There were several limitations to this study. First, it was a retrospective study. Our school-age follow-up without a predefined protocol resulted in a large loss to follow-up (69%). The relationship between perinatal factors and respiratory outcomes at school age is potentially biased by differential dropout. The male sex is reported to be associated with short-term respiratory morbidity in premature infants;39 therefore, because of a lower follow-up rate in men in the study cohort (table 1), we might have underestimated the true prevalence of lung dysfunction. To minimise this attrition bias, we used multiple imputations by chained equations to account for missing data. Results of the imputed case analysis were consistent with the complete case analysis results (see online supplementary data 4). Second, the exclusion of patients with severe neurological disorders at the time of lung function tests might have skewed the analysis, although this was required to evaluate exact lung functions. Third, the reproducibility of the radiological diagnosis may be of concern because chest X-rays were only reviewed by one blinded radiologist.

The strengths of the study are the relatively large number of patients with precise perinatal data and valid lung function tests at school age. Notwithstanding the limitations described, this study indicates the importance of some perinatal parameters that are associated with long-term respiratory sequelae, especially a bubbly/cystic appearance in the lungs during the neonatal period.


In conclusion, within a cohort of ELBW infants assessed at 8 years of age, a bubbly/cystic appearance in the lungs during the neonatal period was the strongest determinant of a low FEV1/FVC ratio.


We thank Dr Michael Dunn, Dr Tetsuya Isayama, Dr Naohiro Yonemoto, Dr Hisaya Hasegawa and Dr Hirotaka Minami for their helpful advice, and Dr Kei Tamai, Dr Kazue Moon and Ms Ai Miyazaki for data collection and discussion. We also thank all of the children and parents who participated in this study and the medical staff working in the NICU at OMCRMCH for their support during this study.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors KH, MNi, HK and MF conceptualised and designed the study, contributed to data analysis and drafted the manuscript. JS, SH, KM, KS, NW and YK contributed to data collection and analysis, and reviewed results. MNi supervised data interpretation of the radiological findings. MNa supervised data interpretation of the histological findings. TK evaluated the developmental status of ELBW children at school age. All authors reviewed the draft manuscript and approved the final manuscript.

  • Competing interests None.

  • Ethics approval This study was approved by the Ethical Committee of Osaka Medical Center and Research Institute for Maternal and Child Health (no. 589).

  • Provenance and peer review Not commissioned; externally peer reviewed.