Lung Function at 8 and 16 Years After Moderate-to-Late Preterm Birth: A Prospective Cohort Study
BACKGROUND AND OBJECTIVE: Knowledge regarding lung function after moderately preterm birth is limited. We therefore investigated lung function at early school age and adolescence among children born moderately preterm.
METHODS: Data were used from the Swedish prospective birth cohort BAMSE (Swedish abbreviation for Children, Allergy, Milieu, Stockholm, Epidemiology study; N = 4089), with a 4.8% prevalence of moderate to late preterm birth defined as a gestational age of 32 to 36 weeks. Participants underwent spirometry at ages 8 and 16 years, and impulse oscillometry additionally at age 16 years. In total, 2621 children (149 preterm and 2472 term) provided lung function data.
RESULTS: At age 8 years, adjusted forced expiratory volume in 1 second was lower in preterm female subjects (–64 mL [95% confidence interval (CI): –118 to –10]) compared with term female subjects but not in preterm male subjects. At age 16 years, both genders in the preterm group demonstrated lower forced expiratory volume in 1 second (female subjects: –116 mL [95% CI: –212 to –20]; male subjects: –177 mL [95% CI: –329 to –25]) compared with the term group. For the preterm group, impulse oscillometry demonstrated higher adjusted resistance at 5 Hz (female subjects: 31.3 Pa·L–1·s−1 [95% CI: 6.3 to 56.3]; male subjects: 34.9 Pa·L–1·s−1 [95% CI: 12.0 to 57.7]) and frequency dependence of resistance (resistance at 5 and 20 Hz) for male subjects (20.9 Pa·L–1·s−1 [95% CI: 9.8 to 31.9]) compared with the term group.
CONCLUSIONS: Measures of airway function assessed in adolescence were reduced in children born moderate to late preterm, and no catch-up in lung function between ages 8 and 16 years was observed.
- AX —
- reactance area
- AX0.5 —
- square root of the reactance area
- CI —
- confidence interval
- FEV1 —
- forced expiratory volume in 1 second
- FVC —
- forced vital capacity
- IOS —
- impulse oscillometry
- R5 —
- resistance at 5 Hz
- R20 —
- resistance at 20 Hz
What’s Known on This Subject:
Moderate to late preterm birth (ie, at 32–36 weeks’ gestation) is associated with increased risk of reduced lung function during childhood. Follow-up reports in adolescence and adulthood are sparse but have indicated normalization of airway function.
What This Study Adds:
This study found that children born moderate to late preterm have on average 3% to 4% lower forced expiratory volume in 1 second in adolescence compared with term control subjects. There was no catch-up of lung function between 8 and 16 years after moderate to late preterm birth.
Infants born preterm (ie, before 37 weeks’ gestation) will exhibit different stages of respiratory immaturity at birth, including incomplete deposition of the parenchymal elastic network of the lungs, as well as acute respiratory morbidity.1,2 Following increased survival after preterm birth, long-term lung recovery and health have become important issues. Most follow-up studies on lung function have focused on subjects born very preterm (ie, born at <32 weeks’ gestation). Far less is known about lung function in children born moderate to late preterm (ie, at 32–36 weeks’ gestation). This group constitutes the vast majority of preterm deliveries, with birth rates ∼10% worldwide, ranging from 5% to 15%; only 1% to 2% are born very preterm.3,4
Available studies on moderate to late prematurity in relation to airway obstruction in infants and toddlers report increased respiratory morbidity during early childhood.5–11 Few studies are available on lung function at school age and adolescence in this group, however, and the results are somewhat conflicting.12,13 Furthermore, although spirometry is a well-established method to assess lung function by measuring expiratory flow rates and volumes, other lung function techniques aiming to assess peripheral airway function and mechanics, such as forced oscillation, have been shown to be more sensitive than spirometry for detection of airway obstruction in children with lung disease, including follow-up studies of children born very preterm.14–18
We hypothesized that being born moderate to late preterm would have a negative effect on lung function through childhood and adolescence. Therefore, the aim of the present study was to determine associations between moderate to late preterm birth and various aspects of later lung function (measured by using spirometry at ages 8 and 16 years, and by using impulse oscillometry [IOS] at 16 years) with data from a large, prospective birth cohort.
Study Design and Study Subjects
The BAMSE (Swedish abbreviation for Children, Allergy, Milieu, Stockholm, Epidemiology) study is a prospective birth cohort that included 4089 Swedish children.19 Parents of all infants born between 1994 and 1996 in predefined areas of Stockholm (including inner city, urban, and suburban districts) were asked to participate in the study. The original cohort consists of 75% of eligible children. Exclusion criteria were: the family planned to move within 1 year of the study start; insufficient knowledge of the Swedish language; the family had a seriously ill child; or an older sister or brother was already included in the study. Data on detailed residential characteristics, environmental factors, and allergic heredity were collected from the parental questionnaires when the children were ∼2 months of age (time of inclusion). At age 8 and 16 years, all subjects were invited to attend a follow-up visit, including lung function measurements. The Supplemental Information provide additional details.
Registry Data and Definition of Gestational Groups
Gestational age (GA) in complete weeks was obtained from the Swedish Medical Birth Registry or, when registry data were not available, by parental reporting. GA was categorized into 2 groups: moderate to late preterm (32–36 weeks) and term (37–41 weeks).
Small for GA was defined as birth weight ≥2 SDs below the mean according to a Swedish gender- and GA-specific reference for normal fetal growth.20
Definition of Respiratory Symptoms
Asthma was defined as fulfilling at least 2 of the following 3 criteria: (1) symptoms of wheeze during the 12-month period before the date of questionnaire; (2) physician-diagnosed asthma (from birth to the date of the questionnaire); and (3) asthma medication taken occasionally or regularly during the past 12 months before the date of the questionnaire.21 Wheeze was defined as at least 1 episode of wheeze during the 12 months before the date of the questionnaire.
Measurements of Lung Function
At 8 years of age, patients underwent lung function testing, performed by using the 2200 Pulmonary Function Laboratory (SensorMedics, Anaheim, CA), and at 16 years of age, using the Jaeger MasterScreen–IOS system (Carefusion Technologies, San Diego, CA).22 At both occasions, spirometry was performed according to American Thoracic Society/European Respiratory Society criteria.23 The highest values of forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1) were used for analysis. FEV1/FVC ratios were expressed as percentages. Mid-expiratory flow rate was extracted from the curve with the highest sum of FEV1 and FVC.
At 16 years, lung function was further assessed according to IOS.24 The IOS system has been described in detail elsewhere,25–27 and a detailed description can be found in the Supplemental Information. Measurement was performed during tidal breathing with the lips tightly sealed around the mouthpiece and cheek support applied. Signals free from artifacts and lasting for at least 20 seconds were saved for analysis. At least 2 recordings were performed per subject. Quality control was performed at the time of examination by visual inspection of the waveforms. Given that coherence, which is a measure of testing reliability, was >0.80 at 10 Hz, the mean value of resistance at 5 Hz (R5), resistance at 20 Hz (R20), frequency dependence of resistance (ie, R5–20), and reactance area (AX) were used for analysis. AX may boost the assessed response inappropriately compared with the R5–20 response because it multiplies 2 partly independent reactions of the airways. We therefore “linearized” AX by reporting its square root as well (ie, AX0.5). The spirometer was calibrated each day by using a 3-L precision syringe. The IOS system was checked for accuracy daily using a reference resistance (0.20 kPa·L−1·s−1).
Demographic data were compared between groups of children by using Student’s t tests or the Pearson χ2 test. The associations between term/preterm groups and spirometry variables were assessed by using linear regression analysis for male and female subjects separately, adjusting for age and height at the time of measurement. The mother’s smoking habits during pregnancy and the first months of the child’s life, small for GA, socioeconomic status, allergic heredity, air pollution exposure (particulate matter with aerodynamic diameter <10 μm and nitrogen oxides)28 in the first year of life, BMI, pubertal staging, and personal smoking at 16 years were evaluated as confounders. The associations between preterm birth and z scores of FEV1 according to the Global Lung Initiative reference values were also evaluated.29
The distribution of the IOS values was skewed to the right. Quantile regression,30,31 also called regression on the median, was used to estimate associations between prematurity and IOS variables while adjusting for height, age, and maternal smoking in male and female subjects separately. This regression method is more robust to outliers or skewness of the distribution of the dependent variable than ordinary regression on the mean. The interpretation of the quantile regression coefficients is similar to that of linear regression, except that the parameter describes the difference in the medians, rather than means.
To assess the importance of preterm birth on the change in lung function over time (ie, from 8 to 16 years of age), lung function was assessed longitudinally by using a general estimating equations model. Time-dependent covariates included in the model were height and age. Term/preterm birth group, gender, and maternal smoking were fixed covariates. To assess the change over time, a preterm/term group–by–time interaction term was used. A significant interaction term would indicate that the change over time was different in the preterm group compared with the term group. Analyses were performed by using Stata version 12.1 (Stata Corp LP, College Station, TX).
The BAMSE study was approved by the ethical committee of Karolinska Institutet (Stockholm, Sweden). All parents provided consent for their child to participate in the study.
Overall, 222 subjects (5.4%) in the original BAMSE cohort (N = 4089) were reportedly born preterm, and 198 subjects (4.8%) were born moderate to late preterm (32–36 weeks’ gestation). The term (control) group consisted of 3494 (85% of total cohort) children born in gestational weeks 37 through 41. An overview of the participants and number of successful lung function test results in each group tests is presented in Fig 1 and the Supplemental Information. With respect to population characteristics, the group of children who provided lung function data did not differ to any major extent from the original BAMSE cohort (Supplemental Table 4).
Characteristics of term and moderate to late preterm children are presented in Table 1. The moderate to late preterm groups required more respiratory support in the neonatal period than those born at term. Small for GA was more common among the moderate to late preterm male subjects. Wheeze at 8 and 16 years, and asthma at 8 years, was reported more often among children born moderate to late preterm compared with term female children.
Lung Function After Moderate to Late Preterm Birth
Unadjusted lung function values and z scores are presented in Supplemental Table 5. The z scores obtained with the Global Lung Initiative reference equations were outside the expected range at 8 years (mean ± SD z score in the term group: 0.50 ± 0.94 for female subjects, 0.36 ± 0.93 for male subjects) but showed a good fit at 16 years (mean: –0.01 ± 0.88 for female subjects and –0.01 ± 0.95 for male subjects).
Lung function differences between moderate to late preterm and term children adjusted for height, age, the mother’s smoking habits during pregnancy, and the first months of the child’s life are presented according to gender as differences in absolute values in Table 2. Small for GA, socioeconomic status, allergic heredity, air pollution exposure (particulate matter with aerodynamic diameter <10 μm and nitrogen oxides) in the first year of life, BMI, pubertal staging, and personal smoking at 16 years were evaluated as potential confounders. These variables showed no such effects, however, and were therefore not included in the final models. At 8 years of age, moderate to late preterm birth was associated with lower FEV1 (–64 mL [95% CI: –118 to –10] or –3.8%), mid-expiratory flow rate, and FEV1/FVC (–0.9% [95% CI: –2.5 to 0.7]) in female subjects compared with the term group. For male subjects, the corresponding value for FEV1/FVC was –1.8% (95% CI: –3.5 to 0.2). At 16 years, negative associations between preterm birth and all spirometric indices except FVC were observed for both genders. In male subjects, FEV1 demonstrated a reduction of –177 mL (95% CI: –329 to –25), or –4.0%, for the moderate to late preterm group compared with term control subjects. In female subjects, the corresponding number was –116 mL (95% CI: –212 to –20) or –3.4%. Expressed as z scores using the Global Lung Initiative reference values for FEV1, the corresponding differences between moderate to late and term groups were, at 8 years, as follows: z scores of –0.36 (95% CI: –0.64 to –0.08) for female subjects and –0.06 (95% CI: –0.32 to 0.20) for male subjects. At 16 years of age, the corresponding numbers were –0.37 (95% CI: –0.63 to –0.10) for male subjects and –0.31 (95% CI: –0.54 to –0.08) for female subjects. However, no significant interactions between gender and preterm group were found for any of the lung function variables (Supplemental Table 6).
The proportion of subjects with an FEV1 below the lower limit of normal (z score: less than –1.64) was significantly larger in male subjects in the moderate to late preterm group compared with the term group (14.9% vs 3.7% [χ2 test for difference, P < .001]), whereas the corresponding difference was smaller for female subjects (5.0% vs 2.3%; P = .19). Given the unsatisfactory fit of data observed at 8 years, proportions were not assessed for this time point.
The IOS results at age 16 years revealed significantly higher estimated medians for R5, R5–20, and AX0.5 in moderate to late preterm male subjects compared with term male subjects (Table 2). Although a similar trend was seen for female subjects, a significant difference was recorded only for R5.
To evaluate if the observed differences were driven by individuals with asthma or very low lung function in the preterm group, an analysis was performed excluding individuals who fulfilled the asthma criteria or had z scores of FEV1 less than –1.96 at age 16 years. The negative association for FEV1 between preterm birth and lung function at 16 years remained present for the nonasthmatic individuals: –185 mL (95% CI: –346 to –24) for male subjects, and –120 mL (95% CI: –226 to –14) for female subjects. Few subjects (moderate to late preterm, n = 5 [4.75%]; term, n = 269 [1.5%]; χ2 test for difference, P = .012) presented with z scores less than –1.96. After excluding these subjects, however, the mean group z scores remained lower for the preterm groups compared with the term groups, although statistical significance was only reached for female subjects (z score for female subjects, –0.24 [95% CI: –0.47 to –0.01]; z score for male subjects, –0.22 [95% CI: –0.5 to 0.06]).
Lung Function From Age 8 to 16 Years
To compare how the levels of lung function changed over time in the moderate to late preterm and term groups, a general estimating equations model with an interaction term between the moderate to late preterm/term group and period (ie, 8- or 16-year follow-up) was used. There was an increasingly negative trend over time for FEV1 in preterm male subjects compared with term male subjects (–185 mL [95% CI: –320 to –51]) or FEV1 z score (–0.3 [95% CI: –0.58 to –0.03]). In female subjects, there was a further negative effect seen on FEV1/FVC for female subjects born moderate to late compared with those born term (–1.74% [95% CI: –3.3 to –0.2]) (Table 3).
In this large prospective longitudinal birth cohort study, we found that subjects born moderate to late preterm had lower lung function as measured by spirometry at 8 and 16 years of age compared with children born term. We found no evidence of catch-up in spirometric indices for either gender from school age until adolescence; the negative effects were instead augmented between age 8 and 16 years in male subjects. IOS measures at age 16 years confirmed the negative effects on airway function and further indicated possible small airway involvement, particularly in male subjects born moderate to late preterm. As a clinical measure, we found that female subjects born moderate to late preterm report respiratory wheeze and asthma symptoms and tend to use asthma medications more often compared with term female subjects.
Although reports of lung function at school age and adolescence are sparse in subjects born moderate to late preterm, a study by Kotecha et al12 reported similar results as in our study, with significantly lower lung function at 8 to 9 years of age for children born after 33 to 34 weeks of gestation, compared with term control subjects. However, they found no significant negative effects among those born after 35 to 36 weeks of gestation. When assessing lung function over time, the investigators reported significant increases in FEV1 z scores from age 8–9 years to 14–17 years in the 33- to 34-week GA subgroup, and no significant changes in those born after 35 to 36 weeks of gestation. Nevertheless, the FEV1/FVC ratios were consistently lower in the 33- to 34-week subgroup compared with the control subjects, which is in line with our findings of a persisting negative effect of late premature birth on airway function. Narang et al13 reported reduced expiratory flow rates (measured as FEV at 0.75 second) at 7 to 9 years of age in a study of children born from 27 to 37 weeks’ gestation, but no differences between the study group and control subjects for FEV1 in adulthood were reported. It should be noted that this study group primarily included children with low birth weight (<2000 g) who were born ∼15 years before our cohort was initiated, which makes direct comparisons difficult due to population differences. Hence, although all studies reported reduced lung function at school age in individuals born moderate to late preterm, the potential for catch-up growth seems less certain.
The results from the IOS recordings performed in the present study at age 16 years confirm the negative effect of moderate to late prematurity on airway function observed using spirometry findings. Increases in frequency dependency of resistance (ie, R5–20) and AX have been suggested as indicating increased peripheral airway involvement. This outcome was observed only among preterm male subjects in the present study. Although no significant interactions between preterm birth and gender were found, these results indicate that the preterm male and female groups may differ in the type of impairment, severity of bronchial obstruction, or other developmental effects on lung tissue. Associations between preterm birth and lung function measured by using IOS, although not gender specific, have previously been reported in follow-up studies on bronchopulmonary dysplasia after very preterm or extremely preterm birth.18,32,33 To our knowledge, no other studies have used IOS or methods other than spirometry to characterize lung function after moderate to late preterm birth.
We also found that female subjects born moderate to late preterm reported more respiratory symptoms at both 8 and 16 years of age than female subjects born term. Positive associations between respiratory symptoms and preterm birth have been previously reported, including 1 study reporting a higher prevalence in female subjects compared with male subjects in young adulthood.6,7,34 No significant interactions between gender and preterm group for any lung function indices were observed; however, the findings that the proportion of preterm adolescents with an FEV1 below the lower limit of normal was more pronounced in male subjects (and symptoms were more pronounced in preterm female subjects) indicate that there is a need for a broader investigation into the origins of the pattern of symptoms after moderate to late preterm birth.
A major strength of the present study is that we used prospective data from a large unselected birth cohort. GA was reported by the Swedish Medical Birth Registry, and in >90% of the pregnancies, was determined by using ultrasound. The proportion of prematurely born children in the cohort is comparable to other European reports,3 but it should be noted that the narrow border between 36+6 days and 37+1 poses a risk of nondifferential misclassification that would likely result in an underestimation of the true differences in lung function between the groups. Another important point of discussion is loss to follow-up, with only 34% of the original preterm group in the study having undergone spirometry at both occasions. However, follow-up rates were similar in the term group and were also comparable to other studies.12,13 Examination after bronchodilation was not performed, making it difficult to exclude increased airway smooth muscle tone disclosed as reversible airway obstruction as the cause of the differing airway function (as opposed to impaired airway growth and development). Nevertheless, because restricting the analysis to those not reporting asthma did not change the main results, reactive airways apparently do not explain the differences seen in lung function between the groups in the present study. In addition, excluding individuals with FEV1 z scores less than –1.96 at 16 years did not change the main results, indicating that the difference between the moderate to late preterm and term groups is not explained by a few subjects with very poor lung function.
At 8 years of age, mean FEV1 z scores were not centered at zero, whereas there was a good fit at 16 years. We speculate that this outcome could be due to a slightly different growth pattern during childhood in Swedish children compared with those included in the reference population. Although the effect is likely to be similar for term and preterm children and not affect comparisons between groups, we have refrained from estimating proportions of the population below the lower limit of normality at 8 years of age.
A recent study has suggested that low FEV1 in early adulthood is important in the genesis of chronic obstructive pulmonary disease.35 Concern has also been raised as to whether chronic lung disease after extremely premature birth is a precursor of a chronic obstructive pulmonary disease–like phenotype later in life, given the markedly reduced airway function in early adulthood shown in this group.36–38 Although our findings suggest that children born moderate to late preterm are at risk of reaching adulthood with a modest reduction in lung function, the future clinical importance of these findings remains to be evaluated. Nevertheless, a larger proportion than expected of the moderate to late preterm population had a FEV1 below the lower limit of normal, indicating that this relatively large subgroup of the general population could benefit from lung function follow-up in adulthood.
Measures of airway function assessed in adolescence by spirometry and IOS are reduced in females and males born moderate-to-late preterm. Further, there was no catch-up of lung function between 8 and 16 years of age.
- Accepted January 5, 2016.
- Address correspondence to Per Thunqvist, MD, Sachs' Children and Youth Hospital, Department of Pediatrics, Södersjukhuset, 118 83 Stockholm, Sweden. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Swedish Research Council, the Swedish Heart-Lung Foundation, the Freemason Child House Foundation in Stockholm, the Stockholm County Council, and the Strategic Research Programme in Epidemiology at the Karolinska Institutet.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Henschen M,
- Stocks J,
- Brookes I,
- Frey U
- Blencowe H,
- Cousens S,
- Oestergaard MZ, et al
- Kotecha SJ,
- Edwards MO,
- Watkins WJ, et al
- Kotecha SJ,
- Watkins WJ,
- Paranjothy S,
- Dunstan FD,
- Henderson AJ,
- Kotecha S
- Quanjer PH,
- Weiner DJ,
- Pretto JJ,
- Brazzale DJ,
- Boros PW
- Miller MR,
- Hankinson J,
- Brusasco V, et al; ATS/ERS Task Force
- Quanjer PH,
- Stanojevic S,
- Cole TJ, et al; ERS Global Lung Function Initiative
- Malmberg LP,
- Mieskonen S,
- Pelkonen A,
- Kari A,
- Sovijärvi AR,
- Turpeinen M
- Doyle LW,
- Faber B,
- Callanan C,
- Freezer N,
- Ford GW,
- Davis NM
- Vollsæter M,
- Røksund OD,
- Eide GE,
- Markestad T,
- Halvorsen T
- Copyright © 2016 by the American Academy of Pediatrics