Published ahead of print on August 6, 2003, doi:10.1164/rccm.200303-460OC
© 2003 American Thoracic Society Progressive Decline in Plethysmographic Lung Volumes in InfantsPhysiology or Technology?Portex Anaesthesia, Intensive Therapy and Respiratory Medicine Unit, Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust; Neonatal Units, Homerton University Hospital and University College London Hospitals, London, United Kingdom; and Department of Pediatrics, University of Münster, Münster, Germany Correspondence and requests for reprints should be addressed to Georg Hülskamp, M.D., Portex Anaesthesia, Intensive Therapy and Respiratory Medicine Unit, Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust, 30 Guilford Street, London WC1N 1EH, UK. E-mail: g.hulskamp{at}ich.ucl.ac.uk
During the last 30 years, there has been an unexplained trend toward declining values for plethysmographic assessments of lung volume at functional residual capacity (FRC) in infants. The aim of this study was to compare data collected from healthy infants using contemporary equipment with published reference data and to explore reasons for discrepancies. Lung volumes were measured in 32 healthy infants (age, 4–93 weeks; weight, 3.9–12.4 kg) using a new, commercially available infant plethysmograph. Mean (SD) FRC was 19.6 (3.4) ml/kg (within subject coefficient of variation 3.4 [2.3%]), which was on average 7.0 [3.5] ml/kg and 2.3 [1.2] SD (Z) scores lower than the recently collated reference data from an American Thoracic Society task force. A total of 66% of these healthy infants had a FRC that was below the predicted normal range. Comparison of equipment, software, and protocols with those from previous reports revealed the importance of minimization of dead space and of adequate subtraction of all compressible occluded volume when calculating FRC in infants. These findings emphasize the need to establish reference data for lung function tests in infants that are appropriate for the equipment and protocols in current use.
Key Words: functional residual capacity • plethysmography • infant • respiratory function tests Estimates of lung volume are an important part of infant lung function testing, both when assessing normal growth and development of the lungs and when interpreting changes in respiratory mechanics in response to disease or therapeutic interventions. Functional residual capacity (FRC), the only lung volume that can readily be assessed in infants, is measured either by whole-body plethysmography (FRCp) (1) or by gas dilution techniques (2). FRCp measurements in infants, as in adults, are rapid and reproducible but have generally been restricted to specialized centers as the technique involved was sophisticated and required considerable operator training. In 2001, a joint American Thoracic Society–European Respiratory Society (ATS–ERS) task force published recommendations on the methodology and the equipment for infant plethysmography (1, 3), and a new generation of infant body plethysmographs was developed that met most of these recommendations. Although these new developments could make infant plethysmography more applicable for routine physiologic and clinically orientated measurements, meaningful interpretation of results will still depend on the availability of appropriate reference data (4). A prediction equation with confidence limits for FRCp in infants was proposed in 1995 by Stocks and Quanjer (5) as part of an ATS task force on lung volumes (Clausen and Wagner, for the workshop participants. Consensus document on measurements of lung volume in humans. Developed from workshops sponsored by the American Thoracic Society and the National Heart, Lung, and Blood Institute. Manuscript submitted for posting on the American Thoracic Society website). When published data were collated as a basis for developing that equation, a seemingly "secular" trend toward lower values of FRCp became apparent in the more recent studies. Subsequent reports on FRCp from healthy infants have confirmed this trend (6–8). It is well recognized that prediction equations need to be constantly tested and validated and that investigators using such reference data should test a number of healthy subjects to determine whether selected prediction equations are appropriate for the specific populations, equipment, and protocols used (5, 9; Clausen and Wagner, submitted). This has rarely been done, and it remains unclear whether the observed decline in FRCp with time has a physiologic or epidemiologic basis or can be attributed to advances in measurement protocols and technology. The new Jaeger (J) MasterScreen BabyBodyplethysmograph (VIASYS Healthcare GmbH, Höchberg, Germany), which incorporates solid-state transducers with a much lower apparatus dead space than previously available, has been validated extensively in vitro (10) but, to our knowledge, values of FRC obtained in healthy infants using this equipment (FRCp-J) have not been compared with previously published normative data. The low absolute dead space of this system, together with a standardized and precise approach to correct for such apparatus dead space, might be expected to result in lower values of FRC than those previously reported. The aim of this study was to assess whether published reference data could be used to interpret results of FRCp obtained using current equipment and, if not, to investigate potential causes for discrepancies including those arising from recent changes in equipment, techniques, and protocols. We hypothesized that FRCp-J from a cohort of healthy infants tested in adherence to recent ERS–ATS recommendations (1) using contemporary equipment is significantly lower than previously published reference data (5; Clausen and Wagner, submitted). Some of the results of this study have been previously reported in the form of abstracts (11, 12).
Study Population and Background Data Healthy white infants (> 36 gestational weeks) without congenital abnormalities or neonatal respiratory compromise were recruited as control subjects into ongoing epidemiologic studies. The local research Ethics Committees approved these studies, and written informed parental consent was obtained before lung function testing. Background information on clinical, medical, and social factors was obtained from obstetric and neonatal records and from interviewing parents on the day of the test. Body weight and crown–heel length were measured as described previously (13).
Equipment
Protocol FRC was measured according to the manufacturer's instructions and ERS–ATS recommendations (1) during quiet sleep, after light sedation with chloral hydrate. A minimum of three end-inspiratory occlusions was performed from which the mean (SD) FRC was calculated. Rigorous quality control features were implemented for data acquisition and analysis (see online supplement for details).
Statistical Analysis
Based on collated reference data, Stocks and Quanjer (5) had proposed the following prediction equation:
(residual SD: 0.14; 95% limits of agreement: 76–132%), where FRCp-pred is the predicted FRCp; L crown–heel length (cm), and W body weight (kg).
This equation can be transformed to:
The Z score is calculated as the difference between the natural logarithm of the predicted and the actual measured FRC (FRCp-J) divided by the residual SD (abbreviated as RSD in Equation 2):
Technically satisfactory measurements of FRCp-J were obtained in 32 healthy infants on 35 occasions. Two studies in the youngest infants (< 6 weeks postnatal age) were performed in natural sleep without sedation. Table 1 summarizes subject details and results, with individual data documented in Table 2E in the online supplement. Reported results for each subject were based on a mean (SD; range) of 3.7 (1.1; 2–7) occlusions and were highly reproducible within each individual, the mean (SD) coefficient of variation for FRCp-J being 3.4% (2.3). Individual values of FRCp-J are plotted against length in Figure 2 , together with the predicted regression line (5). Mean (SD) FRCp-J was 172 (54) ml, which was on average (SD; 95% confidence interval), 59 (30; 49, 69) ml lower than predicted. The mean (SD) Z score for FRCp-J from this group of healthy infants was -2.3 (1.2), which was significantly lower than predicted (p < 0.001).
Plethysmographic measurements of FRC using new commercially available equipment were found to be significantly lower than predicted when based on the most recent and extensive compilation of reference data (5; Clausen and Wagner, submitted). When adjusted for weight, the mean (SD) FRCp-J was 19.6 (3.4) ml/kg, with a range from 13.8 to 27.9 ml/kg. The mean (SD) within-subject difference between weight-adjusted FRCp-J and predicted values was -7.0 (3.5) ml/kg, which is physiologically and clinically highly relevant and would preclude use of the ATS (1995) reference data when interpreting these results. Results from 23 of 35 (i.e., 66%) of these studies fell outside the 95% limits of agreement (i.e., less than -2 Z scores) of the prediction equation. The presented data need to be interpreted in the broader context of the declining plethysmographic lung volumes in healthy infants that have been reported over the last 30 years. During this period, the reported mean weight-corrected FRCp diminished from 35 to 23 ml/kg (see Table 2 for overview). This trend and the observed discrepancies between the presented data and those predicted could be related to:
Infant Characteristics Mean birth weight and gestational age of the present cohort were close to the predicted norms for British full-term infants, as was weight for postnatal age at the time of test (mean [SD] Z scores -0.21[1.16]) (24). In contrast, crown–heel length in our cohort was somewhat higher than predicted (mean [SD] + 0.7 [1.0] Z scores) (24). This may reflect differences in measurement protocols compared with those used by Freeman and coworkers (24). A similar elevation of length Z score in healthy infants has been reported in a recent epidemiologic study (19) that used an identical protocol when measuring infant length (including calibrated precision stadiometer and two trained operators) (13). Such details have not always been reported in previous studies (18, 25, 26), and it is possible that underestimation of infant length in at least some of the studies contributing to the collated reference data (5; Clausen and Wagner, submitted) could account for some of the observed discrepancies in this study. The magnitude of such errors would however need to be extremely large to make a substantial group difference. In this study, we estimated that a mean change in length Z score of 0.7 (on average equivalent to 1.7 cm in actual length for this cohort) would have resulted in an increase in mean Z score for FRCp-J of 0.13 only (from -2.36 to -2.23), with the predicted "weight-corrected" FRC falling by 0.4 ml/kg only (from 26.6 to 26.2 ml/kg). Environmental tobacco smoke exposure is known to be relatively high for white infants living in London (22, 27), which could potentially have influenced our findings. However, in contrast to the adverse effect of nicotine exposure on airway function (8, 27–29), previous large epidemiologic studies have not shown any significant impact of smoke exposure on resting lung volume in newborns and young infants (8, 30, 31). Although the present study was not powered to verify any effect of nicotine exposure on FRC, mean [SD] Z scores for FRCp-J in infants whose mothers smoked (-2.2 [1.3]) was not different from those in nonexposed infants (-2.3 [1.2]). Ten of the infants in this study had experienced their first lower respiratory infection before the test but were symptom free for at least 3 weeks before testing. Because wheezing illnesses occur in up to 40% healthy infants during the first year of life, exclusion of such infants from a cohort with a mean (SD) age at test of 43 (21) weeks is probably inappropriate. Previous studies have shown that, after recovery, there is no difference in FRC between infants with and without earlier uncomplicated episodes of wheezing (30, 32). Thus, it seems unlikely that there were any substantial differences in background characteristics of the infants that we studied to explain the magnitude of observed differences in FRCp-J from predicted values.
Equipment and Protocol None of the studies quoted in Table 2 commented on the volume within the tubings connecting the pneumotachometer and mouth pressure port to the transducers, which, although not representing any dead space for the infant, is part of the total compressible dead space. It is therefore difficult to know how other centers have corrected for this potential source of error. However, we do know that the volume of such tubing (which was subsequently found to be 21 ml [i.e., 3 x 7 ml]) was not subtracted from our own plethysmographic studies at the Hammersmith Hospital (35) or at the Institute of Child Health in London, until after 1995 (30, 33). Appropriate adjustments for mask dead space were introduced shortly afterward, thereby explaining the apparently large total compressible dead space accounted for by Dezateux and coworkers (8, 32) and the relatively lower values of FRCp when compared with previous studies from the same group (30, 33). Similar, and potentially larger, problems could occur if a bias flow is delivered to minimize rebreathing through a relatively large volume circuitry, unless the occlusion valve is close to the mask. To our knowledge, the total compressible dead space of the current Jaeger device (using solid-state transducers, with no associated tubing; flowsensor + shutter + mask = 11.8 to 14.3 ml) is lower than previously reported equipment. As can be seen from Table 2, a similarly low volume has been subtracted from some other studies, but the details of these calculations are not mentioned. It is difficult to precisely account for a relatively large total compressible dead space as, in contrast to conditions within the lung, this volume will not be compressed under isothermal conditions during airway occlusion. Thus Boyle's law cannot be applied. It is unclear whether conditions within the total compressible dead space will be polytropic or adiabatic, but it can be assumed that the volume of the total compressible dead space, as assessed by water displacement, probably overestimates the "effective total compressible dead space" (3). Differences in data acquisition could also account for some of the observed discrepancies. Early plethysmographic studies routinely used airway occlusion at end expiration (18, 25, 26), whereas recent studies have used end-inspiratory occlusions (6–8, 30, 33). The latter protocol was recommended in the recent ERS–ATS statement (1) as this approach disturbs infants to a lesser degree and results in less glottic activity during respiratory efforts against the occlusion. In addition, there may be better equilibrium of pressures throughout the respiratory system when occlusions are performed at a higher lung volume. Although there is some controversy regarding the precise effect (36), most authors (37–40) agree that in both healthy infants and in those with airway disease, FRCp is consistently higher when measured at end expiration rather than at end inspiration. In the most recent comparison, McCoy and coworkers (6) estimated the magnitude of this effect on FRCp in healthy infants to be on average (SD; range) 1.3 (1.7; -1.6 to 4.8) ml/kg. These differences may be explained by a lack of equilibration between changes in alveolar pressure and those measured at the airway opening in the presence of airway narrowing at lower lung volumes. In this situation, a relative underestimation of pressure changes at the airway opening would result in an overestimation of FRCp during end-expiratory occlusions. When performing end-inspiratory occlusions, the occluded (VT) volume above FRC should be corrected to "body temperature, pressure, and saturation" conditions before subtracting it from the total occluded volume. Although this correction is incorporated into the current Jaeger software, no details are given in many papers (6, 7, 31), and to our knowledge, it has not been consistently used in the past (8, 30, 32, 33). Failure to apply this correction could result in an underestimation of the VT occluded above FRC by up to 10% and hence in an overestimation of FRC of about 1 ml/kg in an infant with a VT of 10 ml/kg. The very high values of thoracic gas volume reported during the 1970s will also have been biased by the fact that plethysmographic measurements of FRC were often preceded by assessments of airway resistance, which required the infant to breathe from a heated and humidified rebreathing bag (18). Although the significance of this was not appreciated at the time, this practice, together with the relatively large apparatus dead space common (though not always acknowledged) at the time, would inevitably have caused some hyperventilation and probably dynamic elevation of resting lung volume. During the last decade, most FRCp measurements (including those with the current Jaeger system) appear to have been made with the infants breathing either directly from the box or via a bias flow, so this should no longer be a problem. Nevertheless, a similar phenomenon could occur, albeit to a lesser extent, if the infant has to breathe through a relatively large dead space before FRC recordings and could account for some of the discrepancies between our current findings and previously reported FRCp values (8, 30, 32). With the exception of one report (31), the lowest published values of FRCp in infants have been reported from the most recent studies (6–8, 32). This may reflect gradual improvement in equipment and techniques and the precision of recent calculations and corrections for the total compressible dead space. Nevertheless, all reported FRCp values remain significantly higher than those found in this study. The remaining discrepancies between these studies and the data from the present study may be explained by subtle differences between the equipment (particularly the amount of dead space and resistance), software, and protocol, which have not been accounted for or are difficult to quantify. For example, when using the Jaeger system, FRCp-J is calculated from the integrated slope of both inspiratory and expiratory efforts, according to recent recommendations (1), whereas data from the Columbus group has traditionally been calculated from the inspiratory portion (6, 7), a practice which could introduce subtle differences, especially in the presence of any substantial drift of the box signal due to lack of complete thermal equilibrium. In addition, caution is required when making comparisons of weight-corrected FRCp between different studies, as disproportionate growth and comparison of results from cohorts with different age ranges might influence the magnitude of differences. Ideally, when collating reference data from several centers (41), multiple regression analysis should be applied taking into account anthropometric and ethnic background details as well as measurement protocols. This Jaeger system was extensively tested in vitro (10) before this study, and we have subsequently undertaken regular checks of both equipment and software as described in the online supplement. In addition to the routine procedures recommended by the manufacturers, all sensors were intermittently checked for accuracy and precision, and results were always within the accepted limits (see Table 1E in the online supplement). The integrity of the algorithms used for calculating FRCp-J was cross checked by calculating FRCp manually from the slope of pressure at the airway opening versus box volume changes during respiratory efforts against the occlusion (see online supplement) and was found to be accurate within 1.5% (equivalent to a reading error of 0.5 mm from the printed plot). Rigorous criteria for quality control were implemented throughout this study (see online supplement), and measures of FRCp-J were highly reproducible within individuals. We are thus reasonably confident that FRCp results obtained using the Jaeger equipment represent correct and unbiased estimates of lung volume in healthy infants and that these results could be reproduced by other groups using the same equipment, protocols, and criteria for quality control. In summary, a combination of errors resulting from a failure to adequately subtract all relevant dead space, including that of the facemask and any compressible volume proximal to the occlusion, together with failure to correct the occluded VT signal to body temperature, pressure, and saturation before subtraction from the total occluded gas volume could easily account for the observed discrepancy between published reference data and those presented in this study.
Clinical and Physiologic Implications In addition, the lower values of FRCp-J now reported suggest that the relationship between FRC measurements from plethysmographic and gas dilution techniques in infants needs to be reevaluated. In contrast to findings in healthy adults and older children, in whom results are very similar, plethysmographic lung volumes in infants have been consistently higher than those obtained by gas dilution (6, 33, 43). Therefore, separate prediction equations for the two techniques were recommended by the 1995 ATS report (5; Clausen and Wagner, submitted). The reason for this discrepancy has traditionally been attributed to the increased tendency to airway closure and hence gas trapping during tidal breathing, even in healthy infants, as a result of developmental differences including the high compliance of the chest wall and airways in early life (44) but in light of the current findings may be more attributable to technologic factors.
Conclusions
G.H. has no declared conflict of interest; A-F.H. has no declared conflict of interest; H.L. has no declared conflict of interest; S.L. has no declared conflict of interest; J.S. has been reimbursed (travel and accommodation only) by Jaeger/Viasys Healthcare to attend the ATS and ERS international conferences during the past three years, and the Portex Respiratory Physiology Unit received a £30,000 Educational grant (2001–2003) from Viasys Healthcare; J.J.P. has no declared conflict of interest. The authors gratefully acknowledge the ongoing support of Jane Hawdon and Kate Costeloe in allowing measurements to be undertaken in infants delivered at University College London Hospital and Homerton University Hospital, London, UK.
Supported by the Innovative Medizinische Forschung, University of Münster, and the Gesellschaft für Pädiatrische Pneumologie, Germany (G.H.), by Portex Ltd. (A.-F.H. and J.S.), by a European Respiratory Society Long Term Research Fellowship (H.L.), and by a Neil Hamilton Fairley NHMRC Postdoctoral Fellowship (J.P.). Research at the Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust benefits from R&D funding received from the NHS Executive. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form March 31, 2003; accepted in final form August 4, 2003
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
