INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In 1967 Northway and colleagues first described bronchopulmonary dysplasia (BPD), a chronic lung disease acquired by some premature infants exposed to hyperoxia and positive-pressure assisted ventilation (1). Infants who die from BPD have extensive lung injury, characterized by alveolar loss, interstitial fibrosis, and vascular remodeling (2). Despite compensatory postnatal lung growth and repair, children who survive BPD may be left with a permanent reduction in the number of functional alveolar-capillary units. In a 3-yr-old who died from complications of BPD, both the alveolar number and alveolar surface area were reduced compared with these same variables in age-matched control subjects (3). Whether similar findings are present in infants with relatively milder forms of BPD is not known. Furthermore, there is little known as to the extent of lung structural abnormalities in long-term survivors.

Lung function in survivors of BPD is typically severely compromised in the first year of life, but then improves during the second year and into late childhood (4). However, even older adolescent survivors can still have reduced lung function (5). Abnormalities in pulmonary function described in long-term survivors of BPD include a low FEV₁, high ratio of residual volume to total lung capacity, and bronchial hyperreactivity (6). Pulmonary artery hypertension has been described in older infant survivors (7, 8), and may impair right ventricular function later in childhood (6). With exercise, survivors of BPD perform relatively poorly compared with peers of similar age born full term (9). Exercise performance in school-age survivors of BPD is impaired most commonly by an abnormal ventilatory response. Ventilatory derangements found in exercise challenge studies of BPD survivors and not age-matched control subjects include exercise-induced bronchospasm, low minute ventilation, decreased CO₂ elimination, low oxygen consumption, and hypoxemia (9, 11).

The structural basis for impaired cardiopulmonary function in survivors of BPD is not recognized, but could be due in part to a reduced alveolar capillary surface area. In an experimental animal preparation, primates with BPD from exposure to hyperoxia and positive-pressure ventilation had permanent alveolar loss (12). If a similar structural derangement were to occur in human survivors of BPD, it might significantly affect gas exchange. The transfer of soluble gases from the air spaces is influenced by several variables, including body mass, alveolar surface area, and pulmonary perfusion (13). During vigorous exercise, pulmonary perfusion acutely increases and thereby expands the alveolar capillary surface area (14). Based on the idea that the number of functional alveolar capillary units is relatively low, we hypothesized that gas transfer during exercise is reduced in survivors of BPD relative to age-matched control subjects. To study this, we used an intrabreath method to measure the transfer of two soluble gases at rest, during, and following exercise in school-age survivors of BPD. We also studied two control groups of similar age, children born prematurely but who did not acquire BPD, and children born full term.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Selection

School-age survivors of BPD were selected from a master data set of infants discharged during 1985, 1986, and 1987 from the Neonatal Intensive Care Unit of Grady Memorial Hospital, a university-affiliated hospital serving primarily an indigent urban population. The BPD survivors subgroup included children with a history of neonatal respiratory distress, exposure to mechanical ventilation, radiographic features of BPD (1), and dependence on supplemental oxygen at 4 wk of postnatal age or older. Children were excluded from the study if they had a history of lung disease other than BPD, cognitive impairment, or exercise limitation for any nonpulmonary reason. Episodic wheezing was not a criterion for exclusion. Twenty-six children met the study criteria. Of these, four could not be contacted and six had significant cognitive impairment (Bayley or Stanford Binet IQ < 70) prohibiting participation in the study protocol. One child was excluded because of symptomatic cor pulmonale. Fifteen children were initially enrolled in the study. We excluded four subjects who could not follow the protocol instructions adequately, and one with a nodal rhythm during exercise. The final ex-BPD subgroup included 10 children, all between 6 and 9 yr of age. None was being treated with supplemental oxygen at the time of study.

To select a subgroup of school-age children born prematurely but who did not develop BPD, we reviewed the data set of 200 premature infants discharged from Grady Memorial Hospital over the same time period as the BPD group. A total of 150 of these had outdated records and could not be contacted. The families of 30 children were contacted, 10 children had conditions which excluded them from study, including severe asthma, sickle cell anemia, or cognitive impairment. Ten did not keep their appointments, and 10 participated in the study protocol.

To study a subgroup of healthy children of similar age born full term, we interviewed healthy children who were children of the departmental or laboratory staff members. The first 10 volunteers who met selection criteria and assented participated in the study protocol.

Measurements

To measure the transfer of soluble gases across the alveolar capillary barrier, we used a noninvasive intrabreath technique (SensorMedics 2200; Yorba Linda, CA). Starting from residual volume, the subjects inspired to total lung capacity a test gas mixture containing 0.3% CH₄ (methane), 0.3% C₂H₂ (acetylene), 0.3% CO (carbon monoxide), and 21% oxygen balanced with nitrogen. After a brief breath hold, the subjects exhaled to residual volume at a constant flow rate maintained by a resistor in the expiratory circuit. The concentrations of the three gases were measured continuously by a rapid-response infrared analyzer.

The transfer of acetylene measured with this technique is perfusion-limited as a result of its high solubility in plasma. It therefore is a measurement of effective pulmonary blood flow (15) and can be normalized for body surface area and referred to as effective cardiac index (ECI). ECI does not include the extrapulmonary shunt fraction and is dependent upon the rate of change of lung volume during expiration at constant flow rate.

The transfer of carbon monoxide (DL_CO) measured with the intrabreath technique is diffusion-limited as a result of its high affinity to hemoglobin, and its transfer highly correlates with lung capillary blood volume. The original method was developed by Hallerborg and colleagues (19) and meets American Thoracic Society criteria for single breath measurements of diffusing capacity (12). Results for intrabreath diffusing capacity were corrected for alveolar volume (VA) measured by the uptake of the insoluble tracer gas methane and expressed as percent predicted from standard equations developed by Crapo and coworkers (20).

With coaching, all of the healthy and ex-premature children learned the technique quickly. However, some of the school-age survivors of BPD had difficulty with the technique, especially during exercise. In studies of children in our laboratory with diverse types of obstructive airways disease, we have found the mean intrasubject coefficient of variation for repeat intrabreath measurements of ECI on the same day to agree within 17%.

We measured pulmonary function with a calibrated spirometer (SensorMedics 2200). The participants performed the forced vital capacity maneuver in the standing position, wore nose clips, and repeated the maneuver according to standard criteria (21). Results were expressed as percent predicted based on height, gender, and race using the standard equations of Hsu and coworkers (22) and Crapo and coworkers (20).

A complete 12-lead ECG was obtained via standard leads, and the ECG tracing was displayed continuously on a monitor during the protocol for accurate measurement of heart rate. We measured systolic and diastolic blood pressures from the left arm with a sphygmomanometer and Sa_O2 by a pulse oximeter (Nellcor; Bennet-Puritan Corp., Minneapolis, MN) with the probe secured to the temporal region.

Study Protocol

Study participants suspended the use of ₂-adrenergic agonists or methylxanthines for 48 h and inhaled glucocorticoids or cromolyn sodium for 96 h before the exercise challenge. On the day of study, the subjects and their parents completed a written questionnaire concerning respiratory symptoms and exercise impairment. A co-author explained the study protocol and consent form in detail, which was approved by the Institutional Human Investigation Committee. Each subject was questioned as to recent respiratory illnesses, and confirmed that he or she had avoided use of the medications as specified above. The subjects then underwent a cursory physical examination that included height, weight, vital signs at rest, and room air Sa_O2.

We measured pulmonary function, heart rate, Sa_O2, and intrabreath gas transfer at 3-min intervals with the subjects at rest, during moderate exercise on a treadmill, and during recovery from exercise. In this way, each subject attempted the intrabreath maneuver once before exercise, twice during exercise at 1 to 3 min (acute) and at 6 to 12 min (steady state), and three times (at 3, 9, and 12 to 15 min) during recovery from exercise. Subjects exercised in an air-conditioned indoor laboratory with a mean ambient temperature of 20 ± 2 (SD)° C and barometric pressure of 737 ± 9 mm Hg. The subjects walked or ran on a treadmill via the modified Bruce protocol (23) in which the treadmill speed and incline were increased in four 3-min intervals, progressing to a 16-degree incline at 4.2 miles per hour in the final interval. We discontinued the exercise phase of the protocol if the subjects reached a 12-min endurance limit or sooner if they complained of exhaustion and were unable to continue the protocol despite verbal coaching. The endurance time was the elapsed time from the start to the stop of the exercise phase. During recovery from exercise, we continued pulmonary function measurements for up to 20 min before ending the study. Study participants with exercise-induced bronchospasm were treated with albuterol (0.15 mg/kg) by small-volume nebulizer. Follow-up spirometry showed an improvement in FEV₁ to at least 10% of baseline before the subjects were permitted to leave the laboratory.

Statistics

We analyzed the data with the SPSS for Windows (SPSS Inc., Chicago, IL) software. For continuous variables, results are expressed as the mean ± 1 standard deviation or standard error. The intrasubject coefficient of variation was calculated as the standard deviation/mean of repeated measurements under the same conditions. To test the null hypothesis that there was no difference in results for specific variables among the three study groups, we used one-way analysis of variance (ANOVA). The difference between results for specific subgroups was compared post hoc with the Student-Newman-Keuls test. For categorical variables, differences in outcome were tested by chi-square analysis. To test changes in pulmonary function from pre- to postexercise, we used Student's t test for paired variables for each study group. To test the association between two variables, we applied Pearson's correlation coefficient. For all comparisons, a value of p < 0.05 indicated statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The ex-BPD and ex-premature birth study groups were similar in respect to gender composition, birth weight, birth gestational age, and postnatal age (Table 1). At the time of study, survivors of BPD weighed significantly less and tended to be shorter than children born full term, and had a significantly lower body surface area (BSA). Although BSA tended to be lower in survivors of BPD than in children born prematurely, this was not statistically significant. To account for these anthropometric differences, effective pulmonary blood flow (EPBF) measurements were therefore normalized for BSA and expressed as corrected acetylene transfer or ECI.

In response to questions about past respiratory symptoms, parents of survivors of BPD were more likely to report past respiratory symptoms, including wheezing (9 of 10), recurrent cough (4 of 10), and chronic otitis media (4 of 10), than did parents of children born full term (² < 0.01 for all). However, the incidence of these symptoms was not significantly different in survivors of BPD than it was in children born prematurely.

In response to questions pertaining to exercise history, parents of children born full term were more likely (² < 0.05 for all) to report that their children participated regularly in organized sports (7 of 10) than parents of survivors of BPD (4 of 10) and children born prematurely (2 of 10). Parents of survivors of BPD reported more wheezing (4 of 10), shortness of breath (5 of 10), and cough (9 of 10) with exercise in their children than parents of children born prematurely and full term (² < 0.01 for all). However, ability to keep up with peers during exercise was reported no differently among the three study groups.

Five survivors of BPD had significant respiratory symptoms during the exercise phase of the study. The Sa_O2 fell abruptly in one to the high 80th percentile. Three of the ex-BPD subjects complained of chest tightness or pain during exercise and required treatment with inhaled albuterol at the completion of the protocol. One of the BPD survivors had recurrent cough during exercise, but continued to run. In contrast, none of the children in the ex-full term or ex-premature infant groups had symptoms with exercise, although one of the children born prematurely did receive treatment with albuterol for exercise-induced bronchospasm.

Mean corrected acetylene transfer was lower pre-exercise (p = 0.04), during the acute phase of exercise (p = 0.03), and during recovery from exercise (p = 0.01) in survivors of BPD than it was in children born prematurely and full term (Table 2 and Figure 1). There was no significant difference in mean corrected acetylene transfer between children born prematurely and full term during any phase of the study protocol.

View larger version (11K): [in this window] [in a new window]   Figure 1.   Corrected acetylene transfer (effective cardiac index) and heart rate measured at rest (circles), during acute exercise (triangles), and at 12-15 min during recovery from exercise (squares) in 10 school-age survivors of BPD, 10 children born prematurely without BPD, and 10 children born full term. Each point represents the mean ± 1 standard error. Even though heart rate at the time of the measurements was not significantly different during the three phases of the protocol, effective cardiac index at rest, during exercise, and during recovery from exercise was significantly lower in survivors of BPD than in children born prematurely and full term.

In all three groups exercise led to a nearly twofold acute increase in mean corrected acetylene transfer above baseline (Table 2 and Figure 1), but its magnitude in survivors of BPD was never as great as it was in children born prematurely or full term. During the steady-state phase of exercise, mean corrected acetylene tranfer fell in all three groups, but this was not significant and its level remained above pre-exercise values. During the recovery phase, mean corrected acetylene transfer decreased to near its resting value by 20 min in all three groups.

Mean DL_CO was lower pre-exercise in survivors of BPD than in children born full term (Table 3). However, mean VA pre-exercise was also lower in the BPD survivors. As a result, mean DL_CO corrected for VA was not significantly different among the three groups. During the acute exercise and recovery phases of the protocol, mean DL_CO/VA was significantly lower in childhood survivors of BPD than in children full term. This difference was due to a reduction in mean DL_CO uptake. Mean VA was not significantly different among the three groups during these two phases of the protocol.

During acute exercise, mean DL_CO/VA increased significantly above its pre-exercise magnitude in children born prematurely and full term, but not in survivors of BPD (Table 3). The mean increase in DL_CO/VA during exercise was only 4% in survivors of BPD, 29% in children born prematurely, and 14% in children born full term. During recovery from exercise, mean DL_CO/VA fell below its pre-exercise level in survivors of BPD, a change that did not occur in either of the two control groups.

Mean heart rate was similar among the three groups at the time of the gas transfer measurements (Figure 1). Although school-age survivors of BPD did tend to tire earlier during the exercise phase than children born prematurely and full term, heart rate during the acute phase of the exercise protocol, maximum observed heart rate, and endurance time were not significantly different (Table 4).

Pulmonary function variables pre-exercise showed mild airways obstruction in the school-age survivors of BPD (Table 5). The mean FEF_25-75 was significantly lower in the ex-BPD group than in children born prematurely and full term. The mean FEV₁ also tended to be lower in survivors of BPD, but this was not statistically significant. The FEV₁/FVC ratio was significantly lower than in children born full term but no different from children born prematurely. Pulmonary function pre-exercise was also abnormal in children born prematurely. Whereas the mean FVC was within the normal range, the mean FEV₁ by comparison tended to be below the normal range for age. As a result, the mean FEV₁/FVC ratio pre-exercise in this group was significantly lower than it was in children born at term.

Although pulmonary function did not change significantly during exercise, recovery from exercise was accompanied by a significant reduction in mean FEV₁ from pre-exercise both in survivors of BPD and children born prematurely (Table 5). The magnitude of exercise-induced bronchospasm, though statistically significant, was mild with a mean decrease in FEV₁ from pre-exercise of 9 ± 7% (p = 0.003) for survivors of BPD and 8 ± 7% (p = 0.006) in children born prematurely. There was no significant correlation between corrected acetylene transfer and FEV₁ pre-, during, or postexercise for any of the study subgroups. However, uncorrected DL_CO transfer did correlate significantly with both FVC and FEV₁ when measurements from all three groups were combined. However, the correlation between pulmonary function variables and DL_CO/VA did not hold.

The mean room air Sa_O2 was in the normal range and not significantly different pre-exercise among the three study groups (Table 6). However, exercise led to a significant decrease from baseline (p = 0.04) in mean Sa_O2 in survivors of BPD but not in children born prematurely or full term. Thus Sa_O2 was significantly lower (p = 0.02) during exercise in survivors of BPD than in children born prematurely or full term. During the recovery phase of the protocol, Sa_O2 returned to baseline in survivors of BPD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In school-age children who survived BPD, the transfer of both acetylene and carbon monoxide is lower both at rest and during acute exercise than in children of similar age range born prematurely who did not acquire BPD and children born full term. With exercise, corrected acetylene transfer increased significantly in all three groups, but did not increase as much and remained relatively lower in survivors of BPD. DL_CO/VA increased during exercise in both control groups, but did not change significantly in survivors of BPD. We did not find differences among the three study groups in body size, heart rate, exercise time, VA, or the degree of airways obstruction to account for these findings. Although children born prematurely had a similar magnitude of exercise-induced bronchospasm as survivors of BPD, this was not accompanied by a reduction in corrected acetylene transfer or DL_CO/VA. Whereas total alveolar capillary surface area is a fundamental variable regulating the magnitude of soluble gas transfer (13), these results suggest that lung structural abnormalities persist long-term in survivors of BPD. Alternatively, residual right ventricular dysfunction and a corresponding reduction in cardiac output in response to exercise could produce similar results.

In this study we used exercise as a means to examine gas transfer in the presence of increased pulmonary perfusion and ventilation. Weibel advocated this method to measure the true pulmonary diffusing capacity for oxygen (13). Previous work has shown that measurements of carbon monoxide and acetylene transfer reach steady state within 1 min following the onset of vigorous exercise in adult volunteers (18). Thus, the acute intrabreath measurements reflect gas transfer under conditions of increased pulmonary blood flow and an expanded pulmonary vascular bed.

The magnitude of oxygen consumption and cardiac output likely were not similar in survivors of BPD and control subjects at the time of the gas transfer measurements. As a result, we cannot exclude differences in cardiac output at the time of the measurements as an explanation of our results. Whereas intrabreath acetylene transfer is dependent in part on lung perfusion, carbon monoxide transfer is relatively more dependent on variables related to lung membrane diffusion (17). During exercise both measures of gas transfer were relatively lower and accompanied by a reduction in Sa_O2 in survivors of BPD than in control subjects. In previous studies in healthy adults, mean intrabreath acetylene transfer during exercise increased by 100% and mean carbon monoxide transfer increased by 30% above their values at rest (24). By comparison, mean acetylene transfer increased by only 62% and mean carbon monoxide transfer by 4% in children who survived BPD. In that carbon monoxide transfer barely increased with exercise in survivors of BPD, we think that impaired gas transfer is an important mechanism for exercise limitation in survivors of BPD.

What is the likely explanation for impaired gas transfer at rest and during exercise in survivors of BPD? Comparative studies in several mammals show a strong relationship between body mass, alveolar surface area, and the magnitude of oxygen transfer (13). In the developing ovine lung, the transfer of carbon monoxide correlated directly with estimates of alveolar surface area (25), and lung volume (26). Thus impaired gas transfer in survivors of BPD could at least in part be explained by a reduction in alveolar surface area. Such an explanation is supported by the results of morphometric studies in neonatal primates with BPD induced by hyperoxia and assisted ventilation (12) and human infants who die with BPD (3). Both of these studies showed a substantial reduction in the number of alveoli.

We cannot exclude right ventricular dysfunction and an associated reduction in cardiac output as factors contributing to reduced gas transfer in survivors of BPD. The calculated effective right ventricular stroke volume, the ratio of effective pulmonary blood flow/heart rate, was significantly lower in survivors of BPD than it was in children born prematurely and healthy control subjects at rest, during exercise, and during recovery from exercise. Mean resting effective stroke volume in survivors of BPD was 20.4 ± 7.1 ml versus 37.8 ± 9.2 ml in children born prematurely and 31.6 ± 12.3 ml in children born full term (p = 0.001). A similar abnormality has been found in children with cystic fibrosis, in which cardiac stroke volume during incremental exercise was significantly lower than in age-matched control subjects (27). Due to motion artifact, we did not attempt to assess right ventricular function during exercise with echocardiography in the current studies. In a previous report of school-age survivors of BPD, right ventricular enlargement was present by echocardiography (6). However, exercise studies in adult patients with chronic obstructive airways disease suggest that pulmonary blood flow does not fall despite significant pulmonary artery hypertension until there is cor pulmonale (28).

Ventilation inhomogeneity, exacerbated by exercise-induced bronchospasm, could reduce the transport of soluble gases to regions of gas exchange. The intrabreath method in a previous study of carbon monoxide transfer was especially sensitive to ventilation inhomogeneity (29). The effect of abnormal pulmonary function on the accuracy of the intrabreath method as a means to measure effective pulmonary blood flow has been tested against standard methods by previous investigators (15, 17). Mild ventilation inhomogeneity did not affect the accuracy of the intrabreath method in adults during exercise (15). In adults with cardiac disease, pulmonary blood flow measurements with the intrabreath method were accurate as long as the FVC was 60% predicted or higher (17). Thus the intrabreath method provides a reasonably accurate estimate of pulmonary perfusion when the degree of airflow obstruction is mild.

While it is possible that ventilation inhomogeneity can account for some of the reduction in acetylene transfer in survivors of BPD, this is not the most likely reason. Although survivors of BPD had mild airflow obstruction at rest and exercise-induced bronchospasm, we found no correlation between FVC, FEV₁, or FEF_25-75 and the magnitude of acetylene transfer corrected for alveolar volume during any phase of the study protocol. Whereas children born prematurely had a similar degree of exercise-induced bronchospasm as children who survived BPD, acetylene transfer was significantly higher in children born prematurely during recovery from exercise (Table 2 and Figure 1). We would not expect that if ventilation inhomogeneity alone accounted for reduced acetylene transfer in survivors of BPD.

In adults with obstructive airways disease, tachypnea with or without exercise was associated with gas trapping and an increase in mean intrathoracic pressure (30). The intrathoracic gas pressure, if sufficiently high, might impede pulmonary capillary perfusion and account in part for our results. However, we found a reduction in ECI in BPD survivors both with low and fast respiratory rates. Although lung overexpansion has been described in survivors of BPD (5, 31), the subjects in the current study had a relatively mild degree of airflow obstruction. While hyperinflation and raised intrathoracic gas pressure could certainly contribute to our results, this is not likely the principal mechanism involved.

Mean corrected acetylene transfer at rest in healthy children born at term in the present study was 3.4 ± 0.3 L/min/m². This result is comparable to that reported by Bowyer and associates using a single breath method in children 5 to 14 yr of age, 3.9 ± 0.7 L/min/m² (32). In their study, younger children had higher single-breath measurements of cardiac index than older children, a difference attributed to technique in performing the single-breath maneuver in the younger subjects. However, we studied children in a relatively narrow age range and repeated the measurements during recovery from exercise. In all three study groups, corrected acetylene transfer returned to near baseline during recovery. This finding supports the reproducibility of the intrabreath method as a means of measuring gas transfer in school-age children.

The results of the present study bear on the approach to clinical problems in school-age survivors of BPD. Foremost, they support a previous study which showed reduced carbon monoxide transfer at rest in Scandinavian school children who survived BPD (33). In the Scandinavian study children born prematurely who did not acquire BPD also had reduced carbon monoxide transfer than age-matched children born full term. In the present study survivors of BPD reported reduced exercise participation and had mild exercise-induced bronchospasm with hypoxemia. Pretreatment with albuterol before exercise might improve lung function and pulmonary perfusion in survivors of BPD as it does in children with exacerbations of asthma (34). Despite the impairment in gas transfer reported herein, exercise may still impart significant benefit to children who survive BPD. In previous studies of children with cystic fibrosis, a structured program of aerobic conditioning despite a compromised ventilatory capacity still led to significant cardiovascular benefits (35).