Reduced Exercise Capacity in Children Born Very Preterm

METHODS

Participants were identified from the New South Wales NICU database and were born in the years 1992, 1993, or 1994. This database contains records pertaining to details of all infants who are admitted to NICUs in New South Wales and the Australian Capital Territory during the neonatal period.¹²

All children who were in the NICU database and fulfilled the selection criteria (<32 weeks' gestation; weight <1000 g at birth; and birth between January 1, 1992, and December 31, 1994) were sent an invitation to participate. Responses were received from 165 families. This represented 28% of the total cohort available. Assessments were conducted on 126 (76%) of these because some (24%) families who initially expressed interest declined to participate further. In addition to the children who were born preterm, a cohort of 34 children who were born at term in the years 1992 to 1994 (there were no exclusion criteria) were recruited from 3 local primary schools. Responses were received from 50 families, and assessments were conducted on 34 children who were born at term.

Written informed consent was obtained from the parent or guardian (in 1 case) of the participant before access to medical charts occurred or any study procedures were performed. The study was approved by the University of Sydney Research Office, Children's Hospital at Westmead Ethical Review Committee, and the human research ethics committees of 6 Sydney hospitals (Royal Prince Alfred, Royal North Shore, Liverpool, Nepean District, Westmead, and Royal Hospital for Women) and 1 regional hospital (John Hunter Hospital, Newcastle).

All participants attended the respiratory function unit at Children's Hospital at Westmead for assessment, which involved a single 2-hour appointment during which the child underwent lung function (spirometry, lung volumes, and gas exchange) measurements in accordance with previously validated techniques.¹³ They also performed fitness assessments with the 6-minute walk¹⁴ and 20-m shuttle run test.¹⁵

Lung function assessments were conducted in a Sensormedics Vmax V62J Autobox (Sensormedics Corp, Yorba Linda, CA). Standard reference equations (variables included gender, height, and weight) were used for all lung function parameters to obtain percentage of predicted of normal values. Results were expressed as percentage predicted with generally accepted reference ranges being 80% to 120%, except for forced expiratory flow between 25% and 75% of forced vital capacity (FEF_25%–75%), which was 67% to 133%. Reference equations were obtained from the 1 source¹⁶ for all parameters except for peak expiratory flow¹⁷ and end reserve volume.¹⁸ Baseline lung function measurements began with spirometry, which was followed by measurement of lung volumes (using body plethysmography) and transfer factor according to American Thoracic Society criteria.¹³ Despite practice, results of tests that were considered to have unacceptable technique according to American Thoracic Society criteria (eg, not reproducible attempts, unable to perform the maneuvers, poor effort) were not included.

The fitness assessments followed lung function and began with the 6-minute walk test. Participants were instructed that the aim of the test was to measure how far they were able to walk in 6 minutes. They were informed that they were to be accompanied for the entire duration of the test; however, they were instructed to walk at their own normal walking pace and the tester (Dr Smith) would keep pace with them. The test was conducted on an indoor flat circuit that was 455 m long. The distance traveled during the test was measured with a meter trundle wheel. The test ceased when 6 minutes had elapsed on a handheld stopwatch, which was started as the first step was taken.

The 20-m shuttle run was conducted on an indoor flat straight track. Two markers were placed 20 m apart. The child was instructed to run to each marker and wait at the marker for a beep to be heard from a recorded cassette.¹⁹ The beep occurs at the beginning of each shuttle and signals when to run to the marker at the other end. Each stage takes 1 minute to complete and consists of at least 7 shuttles. As the test progresses, with the number of shuttles per stage increasing, the time allowed for completion of 1 shuttle decreases (time between successive beeps), so the child needs to run faster to keep in time. The child must be within 0.5 m of the marker at the time of the beep to be considered to have completed that shuttle. The tester (Dr Smith) encouraged all children in a standard manner with the phrases, “You're doing really well,” and, “Keep it up.” This process of running between markers in time to the beeps continued until the child did not reach the marker within the designated time frame on 2 consecutive shuttles.

Power calculations to determine the ideal sample size using 80% power and P < .05 were conducted with the primary outcome measure for the lung function data as percentage of predicted forced expiratory volume in 1 second (FEV₁). For demonstration of a 0.5-SD difference in FEV₁ between groups, 64 children per group were required; for demonstration of a 0.75-SD effect, 30 children per group were needed.²⁰ Because 126 children who were born very preterm and 34 term-born control subjects participated, the larger sample size allowed for clinically important differences to become statistically significant when the variation around outcome variables in the preterm group was larger than the variation in the control group. The large sample size decreased the risk for type II errors. Results were analyzed using SPSS (Gradpack) 13.0 (SPSS Inc, Chicago, IL). Descriptive statistics were used for continuous data. Initial comparisons between groups involved analyzing the differences in outcomes, which was undertaken by using a 2-sample (independent sample) t test and 1-way analysis of variance for data with normal distribution; results are expressed as mean (SD). When the data were not normally distributed, a nonparametric Mann-Whitney U test was used for making between-group comparisons, and results are expressed as median (range). A difference between groups was regarded as statistically significant at P < .05.

RESULTS

A total of 126 children (53% female) who were born very preterm and 34 control subjects (41% female) who were born at term in 1992, 1993, and 1994 participated in the study. There was no significant difference in gender proportion between the 2 groups (P = .214). The mean (SD) gestational age and birth weight for the preterm group was 26.9 (1.7) weeks and 862.4 (160.9) g, respectively, significantly less than the gestational age and birth weight for the control group of 39.4 (1.2) weeks and 3400.5 (512.5) g. A total of 104 children who were born very preterm consented to medical chart access to obtain detailed information regarding their neonatal course. Antenatal steroids were administered to the mothers of 73 children who were born very preterm. One hundred children required intubation, and 102 received supplemental oxygen after birth. BPD was diagnosed in 37 children and retinopathy of prematurity was diagnosed in 53 children who were born very preterm. At discharge, 23 children who were born very preterm had an abnormal head ultrasound finding.

The mean (SD) age of the preterm group [10.1 (1.1) years] was 18 months younger than that of the control group [11.6 (0.8) years] at the time of participation in the study. This difference was statistically significant (P < .0001). Because there was a difference in age between the 2 groups, z scores were calculated and used to compare the groups. The height z score for the preterm group (−0.25) was significantly lower than the height z score for the control group (0.17; P = .029). No significant difference between the preterm and control groups' weight (−0.26 vs 0.02) or BMI (−0.57 vs −0.41) z score was evident. A significant difference was found in the number of children in the preterm group compared with the control group who had spastic diplegia (19 [15.4%] vs 0 [0.0%]; P = .016); cognitive delay (34 [28.1%] vs 0 [0.0%]; P = .001), and visual impairment that required visual aids (28 [26.4%] vs 3 [8.8%]; P = .032). No significant difference was found in the number of children from the preterm group (n = 57) or the control group (n = 12) who reported ever having experienced asthma (46% vs 35%; P = .251). All preterm and control children had a clear chest on auscultation immediately before the lung function and fitness assessments.

Acceptable spirometry measurements¹³ were obtained from 123 children who were born preterm and 34 control children. Three children who were born preterm were unable to produce acceptable spirometry as a result of an inability to understand the technique required despite repeated attempts. Mean baseline spirometry for the preterm and control groups were within reference limits (80%–120% predicted), although the preterm group mean for FEV₁ was at the lower limit of reference, as shown by the mean values presented in Table 1. The preterm group had significantly lower values for all measured spirometric parameters compared with the control group. A total of 35 (28.5%) of 123 children in the preterm group had an FEV₁ <80% predicted, whereas only 4 (11.8%) of 34 children in the control group had an FEV₁ <80% predicted normal (P = .026). In the measurement of FEF_25%–75%, no children in the control group had a value of <67% predicted normal, but in the preterm group, 51 (41.5%) of 123 had a measured value of <67% predicted normal (P < .0001).

Acceptable lung-volume measurements¹³ were obtained from 94 children who were born preterm and 30 control children. Twenty-two children in the preterm group and 4 children in the control group were unable to perform the technique required for acceptable measurements despite repeated attempts. Lung volumes for the preterm and control groups were generally within reference limits, and the mean values are presented in Table 2. As was the case with spirometry, there were significant differences in the mean values for the control and preterm groups; however, in contrast to airflow, the preterm group had significantly higher percentage of predicted values in all standard lung-volume parameters than the control group, except for the vital capacity, which was reduced. In particular, the mean percentage of predicted residual volume (RV) was above the accepted upper limit of reference and 40% higher in the preterm group than in the control group (140.8% vs 98.5%; P = .001).

Acceptable transfer factor measurements¹³ were obtained from 81 children who were born preterm and 34 control children. A total of 45 children in the preterm group were unable to perform the technique required for acceptable measurements despite repeated attempts. The mean baseline transfer factors for the preterm and control groups were within reference limits as presented in Table 2.

All children in the preterm group (n = 126) and the control group (n = 34) participated in the 6-minute walk test. One child from the preterm group (n = 125) did not participate in the 20-m shuttle run because that child was unable to stand unassisted as a consequence of cerebral palsy.

Exercise outcomes for the preterm and control groups are presented in Table 3 and Fig 1. There was no significant difference between the preterm and control groups in the distance walked in the 6-minute walk test. The mean completed shuttle stage reached by the preterm group was 48% lower than the shuttle stage reached by the control group (2.9 vs 5.6; P < .0001), and the median distance traveled during the 20-m shuttle run by the preterm group was only 44% of that reached by the control group (300 vs 680 m; P < .0001). The mean predicted peak oxygen consumption (V̇o_2peak) in the preterm group was 9% lower than that calculated for the control group (41.6 vs 45.5 mL/kg per min; P < .0001). The values obtained from the children in the preterm group were also well below published reference data,¹⁵ and this is highlighted in Fig 1A.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

A, Preterm cohort: completed 20-m shuttle stage. B, Control cohort: completed 20-m shuttle stage.

DISCUSSION

The most striking finding in this study was that the exercise capacity of children who were born very preterm was half that of term-born control subjects despite that the mean values of lung function of these children were within the reference range. This study confirms previous reports of normal lung function in childhood, although airflow measurements were significantly reduced when compared with a term-born control group.^4–7,11,21 In particular, the preterm group had a significantly lower FEV₁ (85% vs 95%) and FEF_25%–75% (71.8% vs 91.4%) and a significantly higher RV (141% vs 99%) than the control group. The RV in the preterm group (141%) does not meet the formal definition of gas trapping (>150%), although mild gas trapping may be present. Reduced spirometry in addition to an elevated RV in the preterm group is evidence of airway obstruction. This may be reflecting the size of their airways. They were deprived of 3 months of in utero development and were born when the airways end in thin-walled terminal saccules (which will eventually lead to respiratory bronchioles, alveolar ducts, and saccules²²). Growth of the airways, which would have occurred during the third trimester, has taken place postnatally. It is possible that the airways may not have grown to their full potential and are likely to be narrower, limiting expiratory airflow as measured by spirometry.

Similar to this cohort, other studies reported children who were born very preterm to have lung function in the reference range, although significantly lower than a term-born control group.⁶ Reduced FEF_25%–75% has consistently been reported in the literature, even when there was no difference in other spirometric parameters (forced vital capacity, FEV₁, and peak expiratory flow rate).^1,4,5 Although the findings of this study suggest that airway obstruction is present, there is recent evidence to suggest that children who are born very preterm may have a different pathophysiology of airflow limitation than children who are born at term when airway inflammation has been assessed by measurement of fractional exhaled nitric oxide (FE_NO).²³ Although preterm children had significantly lower lung function than healthy children, there was no significant difference in the measured levels of FE_NO; however a significant difference in the measured levels of FE_NO was found between children who were born very preterm and children who were born at term and had a history of asthma, suggesting that there are distinct pathophysiologic pathways that can lead to airflow limitation.²³

The finding of raised lung volumes (in particular RV and RV/total lung capacity) in the preterm group of this study is evidence of mild gas trapping and consistent with the premise of airway obstruction. Airflow during expiration is limited as a result of narrow airways; therefore, at end expiration, there is residual air trapping that leads to an increase in the measured lung volumes. There is limited information regarding lung-volume assessment in children who were born preterm. The findings of this study, in which volumes were assessed by plethysmography, are in agreement with the literature showing increased lung volumes that have been determined by either the nitrogen washout technique² or using the body plethysmograph.⁶

It is interesting that the TL_CO at rest was not impaired. This suggests that the alveolar-capillary membrane is not limiting gas exchange and the vascular network is sufficient for gas exchange to occur effectively at rest. Information regarding gas exchange in childhood is extremely limited, and the findings of this study are in contrast to other reports³ in which a significantly lower TL_CO was demonstrated in children who were born very preterm, predominantly from cohorts born in the 1980s. In addition, the lungs of long-term ventilated infants have been shown to have immature microvasculature with a dual capillary pattern and simplified, nonbranching vessels.²⁴ The authors suggested that although the capillaries were frequently seen to be aligned with the alveolar lining, the paucity of branching could lower the efficiency of gas exchange in the expanding parenchyma. This study assessed TL_CO in 81 children who were born preterm and is the first to assess gas exchange in children who were born in the early 1990s. The results indicate that there was no impairment in TL_CO at 10 years in this cohort of children. It may be that derangements in lung structure in this cohort are minimal so as not to affect gas exchange at rest,²⁵ or the single-breath technique of measuring gas exchange may not have the required sensitivity. In addition, during exercise, limitations may become evident. This may be the result of limitations in alveolarization (alveolar hypoplasia²⁶), the alveolar-capillary membrane (thick membrane as a result of immaturity²²), or the vascular network (as a result of larger and less dense alveoli, less surface area for gas exchange and vascular connections²⁷) that are not obvious at rest; however, because the measurements in this study were done at rest, no additional information is available in relation to the latter potential mechanism.

The lung function results of this study indicated no significant abnormalities in respiratory function that would have been expected to result in significantly reduced exercise capacity in the preterm group. It is possible that the children in the preterm group are deconditioned and a lack of fitness plays a part that could be improved with a training program. This study used the 20-m shuttle run test to assess exercise capacity, whereas other investigators used a treadmill^2,5 or bicycle ergometer.^28,29 Only 1 study reported significantly lower V̇o_2peak,⁵ although the others reported marginally reduced V̇o_2peak.^2,28,29

There are a number of possible reasons for the inconsistent results from this study and those previously published in terms of exercise capability and preterm birth. This study represents the largest cohort of children who were born very preterm to assess exercise capability. The V̇o_2peak measured in this study was calculated (using previously validated prediction equations¹⁵) from the level reached in the 20-m shuttle and not directly measured as in the treadmill or bicycle exercise test. Nonetheless, it has been shown that a good correlation exists between the 20-m shuttle level reached and V̇o_2peak in children.¹⁵

There are a number of limitations to this study. It is conceivable that those who consented represented the “better” survivors (because parents took into account the lung function and fitness assessments that were required), although all survivors who were born in 1992 to 1994 were invited to participate. Unfortunately, it is not possible to identify potential bias because the identity of nonresponders and their characteristics is not known because of privacy constraints stipulated by the participating neonatal centers. It could be argued that z scores allow more direct comparisons between various measurements; however, percentage predicted is more widely reported and allows comparisons between various studies notwithstanding the limitations involved when these studies have used different prediction equations. The prediction equations used in this study do not have a correction for ethnicity, which could also be seen as a limitation. It is also difficult to compare the results of studies that recruited children who were born in different eras of neonatal treatment strategies or had different birth weights or different gestational ages and that assessed lung function at different ages. Cross-sectional studies provide important information, but longitudinal follow-up is required for a more clear understanding of whether reduced lung function in infancy continues throughout childhood. Some longitudinal data exist,²¹ although, at present, much is limited to children who were born before 1990. Opportunities are now becoming available, with technologic advances in equipment to assess early lung function, to follow a cohort of children from infancy through childhood and adolescence and into adulthood to determine how lung function develops. Clearly, additional information is required to determine the extent of reduced exercise capacity in survivors of very preterm birth, whether cardiac abnormalities are contributing to exercise limitations or whether it is “fitness” related, and an exercise training program may be of benefit.

CONCLUSIONS

Our clinical assessments of lung function and fitness in 126 children who were born very preterm in the early 1990s demonstrated evidence of mild small-airway obstruction, gas trapping, and impaired exercise capacity, the last seeming to be out of proportion to the degree of airflow limitation. Additional studies are required to evaluate the cause of this exercise limitation and whether it can be improved with a training program.