This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bar-Yishay, E. Articles by Ranganathan, S. Search for Related Content PubMed PubMed Citation Articles by Bar-Yishay, E. Articles by Ranganathan, S.

Prediction of maximum flow at functional residual capacity in infants

Normal range

The measurement of maximal flow at functional residual capacity (maxFRC) has been employed by many ever since the landmark publication coming from our laboratory (1). Since then, a few attempts have been made to define the normal range of maxFRC and these are clearly summarized in a recent task force statement (2), but even this committee was reluctant to recommend a single set of appropriate reference data. A few months ago, the first multicenter study of maxFRC in normal infants appeared in this journal (3), but this was a retrospective collation of previously published data.

The implications of using problematic data to define the normal range are far and many. If, for example, reference data demonstrate very wide variability, the lower limit of normality approaches zero, which, in practical terms, means that very few sick infants will have diminished maxFRC values. This point is exemplified by a report by the London group of this multicenter study in the November 2002 issue of this journal (4). In that study, 35 of 45 infants with cystic fibrosis (CF) had clinical evidence of prior lower respiratory tract infection and 8 had moderate/severe respiratory status. Only 1 infant demonstrated a maxFRC below the normal range, whereas 13 had significantly diminished FEV in 0.5 seconds. The authors thus concluded that "the raised volume technique identified diminished airway function more frequently than the tidal technique."

But is this really true? Over the years, we have collated 82 measurements of maxFRC in infants (42 boys) who were all judged to be free of cardiopulmonary disease after full clinical investigation, and were free of any symptoms at the time of testing. There were 32 healthy infants, 25 of whom appeared in publications coming from our laboratory (1, 5). Seventeen were siblings of infants with sudden infant death syndrome and two were siblings of infants with CF. Twenty-six were investigated after questionable episodes of suspected apnea, cough, or wheezing, two for suspected upper airway anomalies, and three for post viral pneumonia. Median age was 11.1 weeks (range 1 day to 79 months, 95% confidence interval [CI] 12.2, 19.5), weight 5.2 kg (5.0, 5.9), and length 58.5 cm (58.0, 62.1). Age distribution, and hence length, are skewed but no heteroscedasticity is indicated graphically.

Individual results are presented in Figure 1 along with a best nonlinear regression described by the following relationship: maxFRC (ml · sec^-1) = -6.139 + 0.00657 x Length^2.58 (cm); Sy,x = 101.86, r² = 0.918, p < 0.0001; Sy,x-standard deviation of the residuals. The regression presented here for our group of infants and young children, over the widest age range from birth to 6 years (length up to 124 cm), has an impressive power that is significantly more robust than that of Hoo and colleagues (3), i.e., r² = 0.92 versus r² = 0.48, 0.49 for boys and girls, respectively (p < 0.01). It is noted that almost all intersubject variability is explained by body size, and that the above nonlinear regression has a similar power function previously suggested by us (5).

View larger version (17K): [in this window] [in a new window]   Figure 1. Individual maxFRC values as well as the nonlinear regression and ± 95% confidence interval are plotted as a function of body length, n = 82. The lower panel compares the nonlinear regression of the present study (solid line) and lower 95% confidence interval range (upper dashed line) to the suggested regressions of Hoo and colleagues (3) for boys (lower dashed line) and girls (dotted line). For clarity, the graph was truncated at maxFRC of 1000 ml/second.

It is hard to believe that these substantial differences in maxFRC are due to patient selection. We cannot think of any reason to suggest why Israeli infants should yield much higher values than their British or American counterparts. It is recognized that our group does not constitute a random sample and cannot be construed as a true representative of the healthy population. Also, the suggested standards for the technique used to obtain maxFRC (2) were published only recently, after most of our investigations were carried out, but they are not critically different from our practice as four members of this committee ran all of our studies. Finally, reanalyzing only data of infants within the same body length range yielded essentially the same best-fit nonlinear regression, with a regression coefficient (r² = 0.741, n = 71) still significantly higher than that of Hoo and colleagues (3) (p < 0.01).

The fact that our results are higher makes us believe that the difference lies in the methodology of performing the test. The most likely explanation for the significant differences may be quite simple. We have been reporting the highest attainable maxFRC value ever since introducing this test and in agreement with recent published standards (1, 2), whereas the studies contributing to the multicenter publication (3) have opted to report the mean of the highest three values. This practice resulted in lower mean maxFRC values and very low lower threshold of normality.

In summary, the present data suggest that reference maxFRC values may be substantially higher and normal range narrower than those presented in the recent multicenter study. Using the suggested limits of normality (3) may result in an unnecessarily large number of false negatives. It is further suggested that each center should determine its own limits of normality, closely adhering to published recommendations, especially when investigating sick infants.

Hadassah University Hospital Jerusalem, Israel

REFERENCES

1. Taussig LM, Landau LI, Godfrey S, Arad I. Determinants of forced expiratory flows in newborn infants. J Appl Physiol 1982;53:1220–1227.[Abstract/Free Full Text]
2. Sly PD, Tepper RS, Henschen M, Gappa M, Stocks J. Standards for infant respiratory function testing: tidal forced expirations. Eur Respir J 2000;16:741–748.[Abstract]
3. Hoo AF, Dezateux C, Hanrahan JP, Cole TJ, Tepper RS, Stocks J. Sex-specific prediction equations for maxFRC in infancy: a multicenter collaborative study. Am J Respir Crit Care Med 2002;165:1084–1092.[Abstract/Free Full Text]
4. Ranganathan SC, Bush A, Dezateux C, Carr SB, Hoo A-F, Lum S, Madge S, Price J, Stroobant J, Wade A, et al. Relative ability of full and partial forced expiratory maneuvers to identify diminished airway function in infants with cystic fibrosis. Am J Respir Crit Care Med 2002;166:1350–1357.[Abstract/Free Full Text]
5. Shulman DL, Bar-Yishay E, Beardsmore CS, Beilin B, Godfrey S. Partial forced expiratory flow-volume curves in young children during ketamine anesthesia. J Appl Physiol 1987;63:44–50.[Abstract/Free Full Text]

From the Authors:

We thank Dr. Bar-Yishay for his interest in our recent paper, and agree that caution is needed whenever interpreting results using reference data (1). However, his arguments regarding the appropriateness of our data are seriously flawed (see details in online supplement). Numerous potential reasons could explain differences between the equations presented by Hoo and colleagues (1) and those by Bar-Yishay, not least because the latter violate most of the underlying assumptions for creation of reference data (2). The population described by Bar-Yishay comprised a small heterogeneous group of children up to age 6 years, many of whom were "hospital controls" studied under ketamine anesthesia, which may contribute to higher maxFRC values (3). Inclusion of some "outliers" up to age 6 years will not only distort the calculated r² value but skew the entire regression equation. Both equipment and techniques have advanced considerably since early descriptions in the 1980s, and the type of curves presented in papers quoted by Dr. Bar-Yishay as being the source for his current "reference data" (4, 5) would not meet current recommendations for quality control (6).

Even had we used the ‘best’ value, predicted maxFRC values would only be 7% higher than those based on the mean (1). The suggestion that the collated reference equations are inappropriate since they failed to identify abnormal airway function in so few individual infants with cystic fibrosis (7) not only overlooks the potential limitations of maxFRC if there is substantial gas trapping as may occur in cystic fibrosis, but the fact that in this study, results from infants with cystic fibrosis were compared directly with those from a large prospective control group rather than published reference data. The potential value of using the maxFRC SD scores in other applications has been illustrated in recent publications (8, 9).

We believe that the discrepancies noted by Dr. Bar-Yishay reflect the limitations of his own data set rather than the way we chose to express maxFRC. Furthermore, the 95% confidence intervals he presented in Table 1 appear to be calculated for the mean, rather than population confidence intervals, as presented by Hoo and colleagues (1).

Thus, while recognizing the limitations, we would continue to advocate use of these SD scores, but stress that users should ascertain whether such equations and the derived scores are applicable in their own laboratory. With recent advances in equipment and standards, it may be necessary to revise and upgrade these equations in the future.

Ah-Fong Hoo^a, Janet Stocks^a, Tim Cole^a, Carol Dezateux^a and Rob Tepper^b

^a Institute of Child Health London, United Kingdom
^b James Whitcomb Riley Hospital for Children Indianapolis, Indiana

FOOTNOTES

This letter has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org

REFERENCES

1. Hoo A-F, Dezateux C, Hanrahan JP, Cole TJ, Tepper RS, Stocks J. Sex-specific prediction equations for VmaxFRC in infancy: a multicenter collaborative study. Am J Respir Crit Care Med 2002;165:1084–1092.
2. American Thoracic Society. Lung function testing: selection of reference values and interpretative strategies. Am Rev Resp Dis 1991;144:1202–1218.[Medline]
3. Hatch DJ, Fletcher M. Anaesthesia and the respiratory system in infants and young children. Br J Anaesth 1992;68:398–410.[Free Full Text]
4. Taussig LM, Landau LI, Godfrey S, Arad I. Determinants of forced expiratory flows in newborn infants. J Appl Physiol 1982;53:1220–1227.
5. Shulman DL, Bar-Yishay E, Beardsmore CS, Beilin B, Godfrey S. Partial forced expiratory flow–volume curves in young children during ketamine anesthesia. J Appl Physiol 1987;63:44–50.
6. Sly P, Tepper R, Henschen M, Gappa M, Stocks J. Standards for infant respiratory function testing: tidal forced expirations. Eur Respir J 2000;16:741–748.
7. Ranganathan SC, Bush A, Dezateux C, Carr SB, Hoo A-F, Lum S, Madge S, Price J, Stroobant J, Wade A, et al. Relative ability of full and partial forced expiratory maneuvers to identify diminished airway function in infants with cystic fibrosis. Am J Respir Crit Care Med 2002;166:1350–1357.
8. Hoo AF, Dezateux C, Henschen M, Costeloe K, Stocks J. Development of airway function in infancy after preterm delivery. J Pediatr 2002;141:652–658.[CrossRef][Medline]
9. Hofhuis W, Huysman MW, van der Wiel EC, Holland WP, Hop WC, Brinkhorst G, de Jongste JC, Merkus PJ. Worsening of VmaxFRC in infants with chronic lung disease in the first year of life: a more favorable outcome after high-frequency oscillation ventilation. Am J Respir Crit Care Med 2002;166:1539–1543.[Abstract/Free Full Text]

The letter from Dr. Bar-Yishay contains several errors. Since its introduction in 1978 (1) the tidal compression technique has been employed by many investigators to assess airway function in infancy. Standardized procedures have been introduced (2) and reference data collated from two centers in the United States and from our institution in London were published recently (3). The London data, contrary to Dr. Bar-Yishay's statement, had not been published previously.

There are two main differences between the Israeli prediction equations reported by Dr. Bar-Yishay and those of the collated reference data of Hoo and colleagues. First, his data predict much higher values for maxFRC. For example, a boy of 65 cm is predicted to have a value 55% greater on average by Bar-Yishay than by Hoo (267 vs. 172 ml/second, respectively). As the coefficient of variation for maxFRC within our healthy population was only 6% (4), this cannot be explained by the policy of reporting best values in his laboratory and mean values in ours. A more likely explanation for obtaining such higher values for maxFRC are late timing of the thoraco-abdominal squeeze resulting in delayed attainment of peak expired flow, as was evident in previous partial flow–volume curves published by their group (5).

The second difference is the report of a significantly lower intersubject variability. Dr. Bar-Yishay reports a coefficient of determination (r²) of 0.92 but does not question the biological plausibility of this value. In no other data obtained in any age group has length explained such a large proportion of the variability of parameters of forced expiratory flow. His model also fails to indicate any association between gender and exposure to maternal smoking and airway function. The number of measurements is far fewer than the collated data, and he reports a median age at test lower than the lower 95% CI, which, if true, indicates an extremely skewed population. The reason that heteroscedasticity is not detected graphically from his data is that too few measurements have been obtained in the older subjects. When restricting the analysis to infants of similar body length, the r² decreased to 0.741, which was still greater than that reported by Hoo and colleagues. However, as for all regression models, it is important to know whether the residuals follow a normal distribution, are centred around zero, and are statistically independent.

Importantly, Dr. Bar-Yishay makes an error in comparing the lower 95% CI of the mean predicted maxFRC for his data with the lower limit of normality defined as -2 SD as reported by Hoo and colleagues. The table shows the limits of normality for the Jerusalem data recalculated using the SD of the residuals he provides. It can be seen that the lower limits of normality are actually quite similar to those from the multicenter study of Hoo and colleagues, and for smaller infants are in fact lower.

Finally, we agree with Dr. Bar-Yishay when he suggests that each center should "determine its own limits of normality closely adhering to published recommendations, especially when investigating sick infants." If he reads our paper again he will discover that this is exactly what we did when we compared maxFRC measured in infants with CF directly with that obtained prospectively in 187 healthy infants, which in contrast to the data Dr. Bar-Yishay reports, did indeed constitute a random sample truly representative of the healthy population (6). Our conclusion that the raised volume technique identified diminished airway function more frequently than the tidal technique in infants with CF remains valid.

Institute of Child Health and Great Ormond Street Hospital London, United Kingdom

REFERENCES

1. Adler SM, Wohl MEB. Flow-volume relationship at low lung volumes in newborn infants. Pediatrics 1978;61:636–640.[Abstract/Free Full Text]
2. Sly P, Tepper R, Henschen M, Gappa M, Stocks J. Standards for infant respiratory function testing: tidal forced expirations. Eur Respir J 2000;16:741–748.
3. Hoo AF, Dezateux C, Hanrahan J, Cole TJ, Tepper R, Stocks J. Sex-specific prediction equations for maxFRC in infancy: a multicenter collaborative study. Am J Respir Crit Care Med 2002;165:1084–1092.
4. Ranganathan SC, Hoo A-F, Lum SY, Goetz I, Castle RA, Stocks J. Exploring the relationship between forced maximal flow at functional residual capacity and parameters of forced expiration from raised lung volume in infants. Pediatr Pulmonol 2002;33:419–428.[CrossRef][Medline]
5. Taussig LM, Landau LI, Godfrey S, Arad I. Determinants of forced expiratory flows in newborn infants. J Appl Physiol 1982;53:1220–1227.
6. Ranganathan SC, Bush A, Dezateux C, Carr SB, Hoo A, Lum S, Madge S, Price J, Stroobant J, Wade A, et al. Relative ability of full and partial forced expiratory maneuvers to identify diminished airway function in infants with cystic fibrosis. Am J Respir Crit Care Med 2002;166:1350–1357.

This article has been cited by other articles:

				M. J. Tobin The Role of a Journal in a Scientific Controversy Am. J. Respir. Crit. Care Med., September 1, 2003; 168(5): 511 - 511. [Full Text] [PDF]

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bar-Yishay, E. Articles by Ranganathan, S. Search for Related Content PubMed PubMed Citation Articles by Bar-Yishay, E. Articles by Ranganathan, S.