Specific airway resistance, interrupter resistance, and ... - copsac

Pediatric Pulmonology 25:322–331 (1998)Specific Airway Resistance, Interrupter Resistance, andRespiratory Impedance in Healthy Children Aged2–7 YearsBent Klug, MD,* and Hans Bisgaard, MD, DR MED SCISummary. We report data on respiratory function in healthy children aged 2–7 years in whomwe measured respiratory resistance by the interrupter technique (Rint); total respiratory impedance(Zrs), respiratory resistance (Rrs), and reactance (Xrs) by the impulse oscillation technique;and specific airway resistance (sRaw) by a modified procedure method in the whole bodyplethysmograph. Measurements were attempted in 151 children and were successfully obtainedin 121 children with a mean (SD) age of 5.3 (1.5) years; no measurements were possible in 30children (mean age 3 (0.9) years).The repeatability of measurements was independent of the age of the subjects, and thewithin-subject coefficient of variation was 11.1%, 8.1%, 10.8%, and 10.2% for sRaw, Rint, Zrs,and Rrs at 5 Hz (Rrs5), respectively. All lung function indices were linearly related to age, height,and weight. A significant negative correlation with age, height, and weight was found for Rint,Zrs, and Rrs5. Xrs5 was positively correlated to age and body size. The mean values of Rint,Rrs5, Xrs5, and Zrs in children younger and older than 5 years were 1.04, 1.38, −0.5, and 1.48kPa L −1 s and 0.9, 1.18, −0.37, and 1.23 kPa L −1 s, respectively. sRaw showed no significantcorrelation with body size or age and the mean sRaw in children younger and older than 5years was 1.09 and 1.13 kPa s, respectively. None of the indices of respiratory function differedbetween boys and girls. Xrs and Rrs exhibited a significant frequency dependence in the rangeof 5–35 Hz. The techniques applied in this study require minimal cooperation and allow measurementof lung function in 80% of our population of awake young children. Further studies areneeded to evaluate the potentials of the presently established reference values for clinical andepidemiological purposes. Pediatr Pulmonol. 1998; 25:322–331. © 1998 Wiley-Liss, Inc.Key words: reference values; specific airway resistance; interrupter technique;impulse oscillation technique; respiratory impedance; preschool children.INTRODUCTIONMeasurement of respiratory function is important indiagnosing and monitoring asthma and other respiratorydiseases in children. 1 However, in preschool children reliableresults can rarely be obtained by standard lungfunction techniques. 2 We recently evaluated a number ofmethods: the interrupter technique for measurement ofrespiratory resistance (Rint), the impulse oscillation technique(IOS) for measurement of respiratory impedance,and a modified procedure for measurement of specificairway resistance (sRaw) by whole body plethysmography.These techniques require passive cooperation onlyand allow measurements during tidal breathing in unsedatedyoung children. 3,4 Reference values have not yetbeen established for Rint, sRaw, and respiratory impedancemeasured by IOS in young children. The purpose ofthe present study was, therefore, to establish referencevalues for sRaw, Rint, and indices derived from measurementof respiratory impedance by IOS, and to evaluatethe applicability of these measurements to healthychildren 2–7 years of age.Department of Pediatrics, National University Hospital, Copenhagen,Denmark.Contract grant sponsors: Astra-Draco, Sweden; the Danish Ministry ofHealth.*Correspondence to: Bent Klug, Dept. Pediatrics 5003, National UniversityHospital, Rigshospitalet, DK-2100 Copenhagen, Denmark.Received 6 June 1997; accepted 22 February 1998.© 1998 Wiley-Liss, Inc.

Lung Function in Healthy Young Children 323MATERIALS AND METHODSSubjectsThe subjects were recruited for this study through ahealth screening questionnaire mailed to a randomsample of children aged 2–7 years living in Copenhagen.The subjects were selected from a random sample of1,200 children, with 200 children (100 boys and 100girls) from each birth cohort aged 2–7 years; their namesand addresses were retrieved from the municipal populationdatabase. Children were considered eligible for thestudy if they were of Caucasian origin and had no chronicdiseases; no history of recurrent cough, wheeze, or severepneumonia; no history of eczema and no atopic firstdegreerelatives; and were not significantly exposed totobacco smoke (3 cigarettes/day). The children had tobe free of respiratory symptoms for one month prior tomeasurement of respiratory function. The study was approvedby the local ethics committee, and written informedconsent was obtained from the parents of all children.Techniques for Measurement of RespiratoryFunctionThe technical features of the equipment and measurementprocedures have recently been described in detail 3,4and are therefore only briefly summarized. Measurementswere carried out using commercially availableequipment (Master Screen, E. Jaeger GmbH, Würzburg,Germany). Flow and volume were measured with aAbbreviationsBTPS Body temperature and ambient pressure saturatedwith water vaporCV w Within-subject coefficient of variationfn Resonant frequencyICC Intraclass correlation coefficientIOS Impulse oscillation techniqueP mo Pressure at the mouthRint Interrupter resistanceRrs Respiratory resistance measured by IOSSD w Within-subject standard deviationsRaw Specific airway resistance measured by whole bodyplethysmographysRaw 0.5 Measurement of sRaw at a flow of 0.5L s −1 duringin- and expirationsRaw Vmax Measurement of sRaw using the maximum flowduring in- and expirationsRaw Vmax Measurement of sRaw using the flow at the point ofmaximum change of plethysmographic volume duringin- and expirationTGV Thoracic gas volumeV pleth Plethysmographic volumeV Respiratory airflowXrs Respiratory reactance measured by IOSZrs Respiratory impedanceheated differential pressure pneumotachograph for determinationof sRaw and Rint. A separate pneumotachographwith an additional built-in pressure transducer wasused for IOS measurements.All measurements were carried out during tidal breathingusing a face mask (Astratech No 2, ASTRA, Denmark)fitted with a flexible, noncompressible mouthpiece(internal cross-section of area of 2 cm 2 ). The face masksupported the cheeks and prevented nose breathing andprovided stable access to the airways. 3,4Measurements with the interrupter technique were performedby assessing mouth pressure (P mo ) at the end ofa brief (80 msec) interruption of airflow (V) during inspirationafter 50 ml of air had been inspired and subsequentlymeasuring airflow 70 msec after the interrupterwas reopened. Rint was calculated asRint P mo /V.The mean value of five sequentially obtained technicallysatisfactory measurements was retained.Measurements of respiratory impedance were obtainedby applying pressure oscillations at the mouth and measuringthe relationship between pressure oscillations andthe resulting air flow. 5,6 Brief square wave pressurepulses generated by means of a loudspeaker at 0.3 secintervals were superimposed on spontaneous breathingcycles at the mouth, while simultaneously and continuouslymeasuring the pressure and flow fluctuations at themouth. The sampling rate was 200 Hz. From each impulse,32 data samples were used to calculate the totalrespiratory impedance (Zrs), respiratory resistance (Rrs),and respiratory reactance (Xrs). Each measurementlasted 30 sec. The mean value of all impulses was used.Values of Rrs and Xrs were calculated at frequencies 5,10, 15, 20, 25, and 35 Hz. The resonant frequency (fn) atwhich Xrs equals zero was also calculated. The technicalquality of the measurements was assessed from the pressure,flow, and volume vs. time traces. During data acquisitionwe required uninterrupted breathing with noabrupt changes of pressure and flow, as described previously.3,4 The primary data from each subject was storedand reanalyzed at a later stage to calculate a coherencefunction from all single impulses recorded.A constant-volume whole body plethysmograph withelectronic BTPS-compensation (body temperature, barometricpressure, and saturated with water vapor) wasused for the measurement of sRaw. We measured therelationship between simultaneous variations of respiratoryflow and variations in plethysmographic volume(V/V pleth ), omitting the measurement of thoracic gasvolume (TGV). 7 The median value of five sequentialbreaths was used for estimating sRaw, which was measuredby three methods (Fig. 1):

324 Klug and Bisgaardno more than five attempts were allowed and if no technicallysatisfactory measurements were obtained, the reasonfor failure was noted.During plethysmographic measurements the childrenwere accompanied by the investigator. In children whowere reluctant to enter the plethysmographic when accompaniedby the investigator, measurements were attemptedwith the children accompanied by their guardian.When the child was breathing slowly during plethysmographicmeasurements, the child was coached toachieve a respiratory rate of 30 breaths/min.Fig. 1. Plots showing the relationship between respiratory flow(V) and variations of plethysmographic volume (V pleth ) duringthe respiratory cycle. The curves A, B, and C are identical copiesof a single curve from a healthy 4-year-old and D shows thecurve from a 4-year-old with acute asthma. sRaw is calculatedas: [sRaw = (V pleth / V) (P B −P H2 0)], where P B is the barometricpressure and P H2 0 is the pressure of water vapor at bodytemperature. V pleth / V is the tangens to the angle . Theestimate of sRaw may differ, depending on the inspiratory andexpiratory flow points () chosen for determining , which isseen clearly in the presence of airway obstruction (D). (A):sRaw Vmax is calculated using the flow at the points of maximumchange of plethysmographic volume. (B): sRaw Vmax iscalculated using the points of maximum flow. (C): sRaw 0.5 iscalculated using the points at which the flow is 0.5 L s −1 .1. From the V/V pleth loops, using a line connectingthe flows at the maximum changes in plethysmographicvolume during inspiration and expiration(sRaw Vmax );2. From a line connecting the points of maximum flowduring inspiration and expiration (sRaw Vmax ); and3. From a line connecting the intercepts of the V/V pleth loop at inspiratory and expiratory flows of 0.5L/s (sRaw 0.5 ). 8During the plethysmographic measurements the childrenwere accompanied by an adult who performed a constantslow expiration during the measurements, and the measuredvalues of sRaw were corrected for the volume ofthe accompanying adult, as described previously. 9ProceduresThe face mask with a built-in mouthpiece was appliedto the child prior to the measurements, to ensure that thechild was able to use the face mask correctly. Thereafter,measurements were carried out in a fixed sequence: Rint,IOS, and whole body plethysmography. This sequencewas repeated once within 15–20 min to obtain duplicatemeasurements by each technique. At each measurement,Statistical AnalysisThe reproducibility of the measurements was estimatedby the within-subject standard deviation (SD w ),calculated as the SD of differences of paired measurementsfrom all subjects divided by √2. The withinsubjectcoefficient of variation (CV w ) was calculatedfrom the SD w divided by the mean. For Xrs, the CV w wasnot calculated, owing to values close to 0. The intraclasscorrelation coefficient (ICC) was calculated as the between-subjectvariance divided by the total variance. 10The difference between paired measurements was plottedagainst their mean to examine whether variability wasindependent of the magnitude of the measurements.Calculations were made on the pooled data and on thedata from boys and girls, respectively, using the individualmean values. The dependence of the indices ofrespiratory function on age, weight, and height were determinedby simple linear regression. Prediction intervals(95% confidence interval for a predicted value) werecomputed assuming a normal distribution of the measurements.The interrelation between indices of respiratoryfunction was assessed for age, weight, and height bycoefficients of correlation. The F-test was used to testwhether the steepness of the regression curves was significantlydifferent from 0. The Wilcoxon matched-pairssigned rank sum test was used for comparison of thethree methods used to estimate sRaw, and for comparingthe coherence values of IOS measurements at 5 and 10Hz. P values < 0.05 were considered statistically significant.RESULTSA health screening questionnaire was sent to the parentsof 748 children; 460 completed questionnaires werereturned. 103 children did not wish to participate in thestudy and 206 were excluded because they did not fulfillone or more of the inclusion criteria (147 children werepassively exposed to tobacco smoke of more than 3 cigarettes/day;59 children were reported to have diseases notcompatible with the inclusion criteria). 151 children enteredthe study and measurements were successfully ob-

TABLE 1—Success of Measurements in Different Age GroupsAge (years) 2 3 4 5 6 7Number of subjects tested 28 31 34 21 22 15Number of subjects whocompleted measurements(%) 16 (57) 20 (65) 28 (82) 21 (100) 21 (95) 15 (100)Reasons for failure:Face mask not accepted 10 10 6 0 1 0Unacceptable quality ofmeasurements 2 a 1 b 0 0 0 0Total number of failures (%) 12 (43) 11 (35) 6 (18) 0 (0) 1 (5) 0 (0)a Failure of Plethysmographic measurements only.b Failure to achieve measurements by the interrupter technique and IOS.Lung Function in Healthy Young Children 325tained in 121 children; 13 of these subjects were occasionallyexposed to passive smoking (3 cigarettes/day).The reasons for failure to obtain measurements in 30children are presented in Table 1; it shows that failurewas mainly due to nonacceptance of the face mask in theyoungest children. Duplicate measurements were obtainedin 121 children (61 boys and 60 girls). Demographicdata on the children who completed measurementsare presented in Table 2. During the plethysmographicmeasurements, 105 children were accompaniedby the investigator and 16 were accompanied by theirguardian. The mean respiratory rate was 41 (10)breaths min −1 .Plots of the mean value of paired measurementsagainst their difference for all techniques showed thatvariability was independent of the magnitude of the measuredvalues. Measures of repeatability are given inTable 3. Repeatability was independent of age, thoughfor sRaw Vmax the repeatability tended to be poorer inthe youngest children (less than 4 years of age) in comparisonto the oldest children (age 6 years or more) withCV w ’s of 9.8% and 6.9%, respectively. The ICC for Rintand sRaw Vmax was 0.86 and 0.92, respectively. ForRrs5 and Xrs5 the ICC was 0.85 and 0.79, respectively.For measurements of Rrs and Xrs in the range of 10–35Hz, the ICC’s ranged from 0.80–0.85 and 0.86–0.91,respectively. The coherence function of impedance measurementsat 5 and 10 Hz was 0.79 (0.09) and 0.91(0.04), respectively (P < 0.0001). The dependence of thecoherence function on frequency in different age groupsis shown in Figure 2.All indices of lung function showed a linear relationshipwith age, weight, and height, and no significantdifferences were found between boys and girls. For Rint,Rrs5, fn, and Zrs, the pooled measurements from boysand girls showed a poor but significant correlation, withdecreasing values with increasing age, weight, and height(P < 0.001), whereas sRaw measurements showed nosignificant correlation with either of the aforementionedparameters (Table 4). The degree of correlation with age,weight, and height was not markedly different for Rint,Rrs5, fn, and Zrs. For Xrs5, a significant (P < 0.001)TABLE 2—Demographic Data of Study PopulationAge (years)No.positive correlation was found with age, weight, andheight. In 37 children in whom sRaw 0.5 could be estimated,sRaw 0.5 was significantly lower than sRaw Vmaxand sRaw Vmax (P < 0.0001) (Table 3).Table 4 presents the regression equation and parametersfor calculating predicted values of lung functionindices according to age, weight, and height. Figure 3depicts the regression line (based on height) and predictionlimits for each technique with the individual measurementsplotted separately for boys and girls. Frequencydependence of Rrs and Xrs could be demonstratedin the range of 5–35 Hz with no differencebetween boys and girls. The magnitude of frequency dependenceof Rrs and Xrs was not significantly related toheight, age, or weight. In Figure 4 the mean values of Rrsand Xrs are plotted vs. frequency in the tallest and shortest50% of subjects, respectively. The results show practicallyidentical shapes of the curves depicting frequencydependence. Respiratory resistance as assessed by Rrs5was higher than Rint, with a mean difference of 0.16(0.1) kPa L −1 s.DISCUSSIONAge(months)Gender(boys/girls)Weight(kg)Height(cm)2 16 35 (0.9) 8/8 14.5 (1.3) 94 (3.7)3 20 42 (3.4) 12/8 15.9 (2.0) 102 (5.6)4 28 56 (3.8) 11/17 19.0 (1.7) 109 (4.7)5 21 66 (4.5) 9/12 22.1 (3.7) 117 (5.1)6 21 82 (3.6) 13/8 24.4 (2.6) 126 (5.9)7 15 87 (1.6) 8/7 25.9 (3.3) 127 (4.7)Total 121 60.4 (18) 61/60 20.0 (4.3) 112 (12)Values are given as mean (SD).Measurement of respiratory function is a valuable adjunctto the clinical assessment of respiratory function inchildren with respiratory diseases; however, it is difficultto assess lung function reliably by standard lung functiontests in preschool children. Until recently, few techniqueswere available for testing respiratory function in awake

326 Klug and Bisgaardyoung children. The present study demonstrates thatmeasurements by modified whole body plethysmography,the interrupter technique, and IOS can be achievedin more than two-thirds of children 2–5 years of age, andin nearly all children 5–7 years of age. Moreover, therepeatability of the measurements is practically independentof age.At present, none of the applied techniques have beenstandardized, implying that data from different laboratoriesmay not be readily comparable. For instance, measurementsof Rint may vary significantly depending onthe method used to measure mouth pressure 11 and timingof the interruption of airflow during the respiratorycycle. 12 Reference values for Rint in young children havenot been reported previously. We found that Rint wassignificantly and negatively correlated with age, height,and body weight, reflecting the increase of airway dimensionswith growth.Measurement of sRaw provides an estimate of airwayresistance, which, beyond infancy, is practically independentof body size. In keeping with other investigators, wefound no significant correlation between sRaw and eitherage, weight, or height. 8,13,14 We measured sRaw by asingle-step procedure from the relation between variationsof respiratory flow and variations of plethysmographicvolume during normal breathing. The single-stepprocedure obviates the need for measuring breathing effortsagainst a closed shutter. 7,15 The latter procedure isnecessary to measure TGV and calculates airway resistance(Raw) by the classic two-step plethysmographicprocedure. 16 Hence, sRaw and TGV are the measuredvariables from which Raw is derived through the equationRaw sRaw/TGV. By obviating the need ofbreathing against a closed shutter, the single-step procedureis a particularly useful measurement in awakeyoung children, because breathing against a closed shutteris either not accepted or inadequately performed bymost preschool children. 7,17 Moreover, the estimate ofsRaw is similar when using the single-step and the twostepprocedure, provided that a minor correction is employedto the single-step measurements to allow for theresistance of the apparatus. 18,19 A potential limitation ofthe single-step procedure is that it provides no data onlung volume; however, this does not seem to restrict theclinical usefulness of this method. 3,4,15 This is furthersupported by the findings of Buhr et al., 17 who comparedmeasurements of Raw, TGV, and sRaw by the singlestepmethod in healthy children and in asthmatic children.Measurements were categorized as being normal orabnormal depending on whether they were within orabove the 95% confidence limit of measurements inhealthy children. The proportion of asthmatic children inwhom abnormal values were observed was higher whenmeasuring sRaw by the single-step procedure than whenmeasuring Raw and TGV separately. When the two lattermeasurements were combined, the proportion of childrenwith abnormal values was similar to that found whenmeasuring sRaw by the single-step procedure. 17The single-step procedure for measuring sRaw onlypartly solves the practical problems associated with plethysmographicmeasurements in young children, who oftenare unwilling to enter the plethysmograph alone orbecome uncooperative when left alone inside the closedplethysmograph. We, therefore, performed the measurementsof sRaw with an adult accompanying the child. Wehave found that this method greatly improves the acceptabilityof the plethysmographic procedure in young childrenwithout introducing bias to the estimate of sRaw orloss of reproducibility. 9The mean sRaw Vmax (1.1 kPa s) observed in thepresent study is higher than in other studies, which havereported values ranging from 0.45 to 1.0 kPa s. 8,13,14,20This discrepancy can probably be ascribed to the fact thatwe, in contrast to most previous workers, used a plethysmographwith electronic BTPS compensation, which hasintroduced a positive frequency dependence and has resultedin significantly higher estimates of sRaw in comparisonwith measurements obtained with a heated rebreathingsystem to ensure BTPS conditions. 9 One of theadvantages of sRaw Vmax and sRaw Vmax is that thesetwo measures can be estimated irrespective of the magnitudeof the flow, and they are therefore applicable tosubjects of all ages. Conversely, this may not be the casewhen estimating sRaw from some fraction of the approximatelylinear part of the curve. In the present study,sRaw 0.5 could be estimated in a third of the children only,due to flows lower than 0.5 L/s in the youngest children.This problem can be eliminated by choosing flow sufficientlylow to ensure that sRaw can be estimated even invery young children. The clinical usefulness of such ameasure remains to be established. Also, the comparativeusefulness of the various estimates of sRaw is unknown,and studies in children with respiratory disorders areneeded to address this issue.Rrs, Zrs, and fn were found to be negatively correlatedwith age, weight, and height, and this is in accord withprevious studies; 21–28 Xrs was positively correlated withage, weight, and height. 21–23,29 For all IOS indices, wefound a linear relationship with age, weight, and height,and no significant differences were observed betweenboys and girls. A linear relationship has been reported inprevious studies between growth and Xrs and Rrs inwhich Xrs and Rrs were measured at lower frequencies inyoung children. 21,23,26–28,30When comparing measurements, it must be taken intoaccount that measurements of respiratory impedance inyoung children may vary considerably depending on thefrequency at which they are measured, as illustrated bythe present data (Fig. 4). In adults, the finding of frequencydependence of Rrs is considered an indication of

TABLE 3—Measures of Repeatability Estimated From Paired MeasurementssRaw VmaxkPa ssRaw VmaxkPa ssRaw 0.5kPa sRintkPaL −1 sLung Function in Healthy Young Children 327ZrskPal −1 sRrs5kPaL −1 sXrs5kPaL −1 sNo. of subjects 119 119 37 a 120 120 120 120 120Mean (SD) 1.31 (0.20) 1.11 (0.19) 0.85 (0.18) 0.97 (0.19) 1.36 (0.24) 1.29 (0.22) −.44 (0.14) 21.2 (3.17)SD w 0.11 0.12 0.09 0.08 0.15 0.13 0.10 1.49CV w (%) 8.3 11.1 10 8.1 10.8 10.2 — b 7.0ICC 0.86 0.84 0.91 0.92 0.84 0.85 0.79 0.90sRaw Vmax : specific airway resistance estimated from flow at the maximum change of plethysmographic volume during in- and expiration;sRaw Vmax : specific airway resistance estimated at the maximum flow during in- and expiration; sRaw 0.5 : specific airway resistance estimatedat the flow 0.5 L s −1 during in- and expiration. Rint: interrupter resistance; Zrs5: total respiratory impedance at 5 Hz; Rrs5: respiratory resistanceat 5 Hz; Xrs5: respiratory reactance at 5 Hz; fn: resonant frequency.SD w : within-subject standard deviation; CV w : within-subject coefficient of variation; ICC: intraclass correlation coefficient.a sRaw 0.5 could be estimated in 37 children with mean age 5.7 (1.6) years; in 82 children aged 4.7 (1.4) years a flow of 0.5 L s −1 was notachieved.b CV w not calculated for Xrs5 owing to values close to zero.fnHzFig. 2. The mean value of the coherence function in the frequencyrange of 5–35 Hz in three age groups. The frequencydependence of the coherence function shows a similar patternin all age groups, but in the youngest children the absolutevalues are consistently lower. Vertical bars indicate 1 SD. Age

328 Klug and BisgaardTABLE 4—Linear Regression: y = + (x−x), Based on Height (H), Weight (W) and Age (A)Method (y) x x SDCoefficientof correlation P-valuesRaw Vmax A (months) 60.8 1.31 −1.4 10 −3 0.187 −0.13 0.15W (kg) 20.14 1.31 −6.4 10 −3 0.186 −0.16 0.08H (cm) 112.3 1.31 −1.8 10 −3 0.188 −0.12 0.21sRaw Vmax A (months) 60.8 1.11 1.0 10 −3 0.196 0.09 0.31W (kg) 20.14 1.11 0.5 10 −3 0.197 0.01 0.89H (cm) 112.3 1.11 1.4 10 −3 0.196 0.09 0.32Rint A (months) 60.8 0.97 −3.9 10 −3 0.172 −0.39 *W (kg) 20.14 0.97 −1.9 10 −3 0.163 −0.48 *H (cm) 112.3 0.97 −6.7 10 −3 0.166 −0.46 *Rrs5 A (months) 60.8 1.29 −6.0 10 −3 0.193 −0.49 *W (kg) 20.14 1.29 −26 10 −3 0.185 −0.55 *H (cm) 112.3 1.29 −9.1 10 −3 0.189 −0.52 *Rrs10 A (months) 60.8 1.07 −4.8 10 −3 0.151 −0.50 *W (kg) 20.14 1.07 −20 10 −3 0.148 −0.53 *H (cm) 112.3 1.07 −6.9 10 −3 0.152 −0.50 *Xrs5 A (months) 60.8 −0.44 3.5 10 −3 0.129 0.44 *W (kg) 20.14 −0.44 14 10 −3 0.128 0.46 *H (cm) 112.3 −0.44 5.2 10 −3 0.129 0.45 *Xrs10 A (months) 60.8 −0.25 2.7 10 −3 0.099 0.45 *W (kg) 20.14 −0.25 11 10 −3 0.098 0.47 *H (cm) 112.3 −0.25 4.1 10 −3 0.098 0.47 *Zrs5 A (months) 60.8 1.36 −7.3 10 −3 0.202 −0.55 *W (kg) 20.14 1.36 −30.4 10 −3 0.195 −0.60 *H (cm) 112.3 1.36 −10.9 10 −3 0.200 −0.57 *fn A (months) 60.8 21.2 −73 10 −3 2.89 −0.42 *W (kg) 20.14 21.2 −300 10 −3 2.83 −0.45 *H (cm) 112.3 21.2 −110 10 −3 2.88 −0.43 **P < 0.001.For abbreviations, see legend for Table 3.later stage by reanalyzing the IOS primary data. Thecoherence function varies with frequency and with ageand not surprisingly the values were low in the youngestchildren (Fig. 2). The mean value of the coherence functionof the impedance measurements at 5 Hz was low(0.71); if a coherence function of at least 0.95 had beenused as the lower limit for acceptability, practically allmeasurements at 5 Hz would have been rejected. However,for a number of reasons we did not use the coherencefunction as a means for accepting or rejecting measurements.First, in recent studies in young children wefound that IOS measurements at 5 Hz were superior tomeasurements at higher frequencies in terms of providingclinically useful information when assessing bronchoconstrictioninduced by methacholine, bronchodilationfrom inhaled terbutaline, or impairment of lungfunction in acute asthma. 3,4 Also, measurements at 5 Hzcan provide repeatable estimates of bronchial responsivenessin 2 to 4-year-old children. 34 Second, the coherencefunction of impedance measurements in the aforementionedstudies and in the present study may be assumedto be comparable, due to the fact that identical criteriahave been employed for accepting the IOS. These observationssuggest that despite a low coherence function,suggestive of poor reliability in technical terms, our measurementsand those of others apparently provide informationthat is clinically useful. Moreover, the diagnosticvalue of measurements at low frequencies has been documentedpreviously. 35 Measurements at frequencies below5 Hz seem promising; 36 however, this is currentlynot possible with the IOS. In obtaining IOS measurements,we used 0.3-sec intervals between the impulsesrather than 0.2 sec, as used in our previous studies. Thereason for prolonging the interval between the impulseswas that we wished to avoid impulses that were superimposedon the low-frequency oscillations, since suchinterference would probably impede future analysis ofthe IOS primary data at low frequencies. The interindividualvariation of the present measurements is comparableto that reported in some studies, 21,25 but greaterthan that reported in others. 23,37 It remains to be seenwhether the relatively large dispersion of measurementswill limit their usefulness when it comes to interpretingsingle measurements in individual patients.The study population conformed with the recommendationsfor establishing reference values for lung functionin children. 38 We excluded children if they weresignificantly exposed to tobacco smoke or had a historyof eczema or doctor-diagnosed atopy in first-degree relatives,due to the increased risk of subclinical airway diseasein these children. 39 We arbitrarily defined a significantexposure to tobacco smoke as exposure to smoke

Fig. 3. Measurements of respiratory function as a function of height in 121 healthy children ( Girls; Boys). For abbreviations,see legend for Table 3. Solid lines indicate the regression line, and broken lines indicate the 95% prediction limits.

330 Klug and BisgaardACKOWLEDGMENTThe authors thank T. Bengtsson for the statisticalevaluation.REFERENCESFig. 4. Measurements of respiratory resistance (Rrs) and respiratoryreactance (Xrs) in frequency range 5–35 Hz in 120 healthychildren showing a similar pattern of the frequency dependencyin small and in tall children. Mean value of measurements inthe smallest 50% of the children [mean height 102 (6.6) cm]. Mean value of measurements in the tallest 50% of the children[mean height 123 (6.4) cm]. Vertical bars indicate 1 SD.from more than 3 cigarettes/day; exposure to doses lowerthan this was reported in 10% of the children in thisstudy. The detrimental effect of environmental tobaccosmoke on respiratory function in children increases withthe magnitude of the exposure; 40 however, the effectfrom exposure to very small doses is probably negligibleand not likely to influence the reference values establishedin this study. Maternal smoking during pregnancymay adversely affect respiratory function in their offspringeven beyond infancy; 40 we did not collect any dataon maternal smoking during pregnancy. The children enrolledin this study were living in the city and suburbs ofCopenhagen, which may have implications for the use ofthe reference values in other populations.Furthermore, use of the present reference valuesshould be considered appropriate only if the specificationsof the equipment and measurement procedures conformwith those described in this study. The accuracy ofthe applied techniques may be questioned, but presentlyit is not possible to address the issue of accuracy properly,due to the lack of data comparing the various techniqueswith what may be considered ‘‘gold standards.’’However, maximum accuracy should not be considered aprerequisite for these measurements to be clinically useful,which is supported by our recent findings in studiesin preschool children. 3,4,34 It remains to be demonstratedwhether the reference values established in the presentstudy will further expand the clinical relevance of applyingthese techniques for measuring respiratory functionin young children.1. Warner JO, Götz M, Landau LI, Levison H, Milner AD, PedersenS, Silverman M. Management of asthma: A consensus statement.Arch Dis Child. 1989; 64:1065–1079.2. Kanengiser S, Dozor AJ. Forced expiratory maneuvers in childrenaged 3 to 5 years. Pediatr Pulmonol. 1994; 18:144–149.3. Bisgaard H, Klug B. Lung function measurement in awake youngchildren. Eur Respir J. 1995; 8:2067–2075.4. Klug B, Bisgaard H. Measurement of lung function in awake2–4-year-old asthmatic children during methacholine challengeand acute asthma: A comparison of the impulse oscillation technique,the interrupter technique, and transcutaneous measurementof oxygen versus whole-body plethysmography. Pediatr Pulmonol.1996; 21:290–300.5. Vogel J, Smidt U. Impulse oscillometry. pmi Verlagsgruppe,Frankfurt am Main; 1994; 1–106.6. Landser FJ, Nagles J, Demedts M, Billiet L, van de Woestijne KP.A new method to determine frequency characteristics of the respiratorysystem. J Appl Physiol. 1976; 41:101–106.7. Dab I, Alexander F. A simplified approach to the measurement ofspecific airway resistance. Pediatr Res. 1976; 10:998–999.8. Leben M, von der Hardt H. Airway resistance, airway conductance,specific airway resistance, and specific airway conductancein children. Pediatr Res. 1983; 17:508–513.9. Klug B, Bisgaard H. Measurement of the specific airway resistanceby plethysmography in young children accompanied by anadult. Eur Respir J. 1997; 10:1599–1605.10. Chinn S. Statistics in respiratory medicine. 2. Repeatability andmethod comparison. Thorax. 1991; 46:454–456.11. Phagoo SB, Watson RA, Pride NB, Silverman M. Accuracy andsensitivity of the interrupter technique for measuring the responseto bronchial challenge in normal subjects. Eur Respir J. 1993;6:996–1003.12. Carter ER, Stecenko AA, Pollock BH, Jaeger MJ. Evaluation ofthe interrupter technique for the use of assessing airway obstructionin children. Pediatr Pulmonol. 1994; 17:211–217.13. Doershuk CF, Fisher BJ, Matthews LW. Specific airway resistancefrom the perinatal period into adulthood. Alterations inchildhood pulmonary disease. Am Rev Respir Dis. 1974; 109:452–457.14. Dab I, Alexander F. Lung function measured with a whole bodyplethysmograph: Standard values for children and young adults.Acta Paediatr Belg. 1979; 32:259–267.15. Dab I, Alexander F. On the advantages of specific airway resistance.Pediatr Res. 1978; 12:878–881.16. Stocks J, Beardsmore C. Flow resistance in infancy. Eur Respir J.(Suppl) 1989; 4:144S–149S.17. Buhr W, Jorres R, Knapp M, Berdel D. Diagnostic value of bodyplethysmographic parameters in healthy and asthmatic young childrenis not influenced by breathing frequency. Pediatr Pulmonol.1990; 8:23–28.18. Haluszka J, Willim G, Cutrera R, Ronchetti R, Filtcher S. Acorrection formula for computing specific airway resistance froma single-step measurement. Pediatr Pulmonol. 1989; 6:118–121.19. Stanescu DC, Rodenstein DO. Is simpler better? New approachesfor computing airway resistance. Bull Eur Physiopathol Respir.1986; 22:323–328.20. Haluszka J. Application of the whole body plethysmography inexamination of respiratory system in children. Predicted values,

Lung Function in Healthy Young Children 331interrelations, methodological suggestions. National Institute forMother and Child, Branc in Rabka, Poland. 1976.21. Duiverman EJ, Clement J, van de Woestijne KP, Neijens HJ, vanden Bergh AC, Kerrebijn KF. Forced oscillation technique. Referencevalues for resistance and reactance over a frequency spectrumof 2–26 Hz in healthy children aged 2.3–12.5 years. Bull EurPhysiopathol Respir. 1985; 21:171–178.22. Cuijpers CEJ, Wesseling G, Swaen GMH, Wouters EFM. Heightand gender related impedence values in primary school children.Eur Respir Rev. 1994; 4:150–154.23. Clement J, Dumoulin B, Gubbelmans R, Hendriks S, van deWoestijne KP. Reference values of total respiratory resistance andreactance between 4 and 26 Hz in children and adolescents aged4–20 years. Bull Eur Physiopathol Respir. 1987; 23:441–448.24. Hordvik NL, Konig P, Morris DA, Kreutz C, Pimmel RL. Normalvalues for forced oscillatory respiratory resistance in children.Pediatr Pulmonol. 1985; 1:145–148.25. Solymar L, Aronsson PH, Bake B, Bjure J. Respiratory resistanceand impedance magnitude in healthy children aged 2–18 years.Pediatr Pulmonol. 1985; 1:134–140.26. Cogswell JJ. Forced oscillation technique for determination ofresistance to breathing in children. Arch Dis Child. 1973; 48:259–266.27. Mansell A, Levison H, Kruger K, Tripp TL. Measurement ofrespiratory resistance in children by forced oscillations. Am RevRespir Dis. 1972; 106:710–714.28. Stanescu D, Moavero NE, Veriter C, Brasseur L. Frequency dependenceof respiratory resistance in healthy children. J ApplPhysiol. 1979; 47:268–272.29. Lierl MB, Hilman BC. Normal values for forced oscillation parametersin small children. J Asthma. 1986; 23:179–185.30. Hantos Z, Daroczy B, Gyurkovits K. Total respiratory impedancein healthy children. Pediatr Pulmonol. 1985; 1:91–98.31. Clement J, Landser FJ, van de Woestijne KP. Total resistance andreactance in patients with respiratory complaints with and withoutairways obstruction. Chest. 1983; 83:215–220.32. Marchal F, Haouzi P, Peslin R, Duvivier C, Gallina C. Mechanicalproperties of the upper airway wall in children and their influenceon respiratory impedance measurements. Pediatr Pulmonol. 1992;13:28–33.33. Solymar L, Landser FJ, Duiverman E. Measurement of resistancewith the forced oscillation technique. Eur Respir J. (Suppl) 1989;4:150S–153S.34. Klug B, Bisgaard H. Repeatability of methacholine challenges in2- to 4-year-old children with asthma, using a new technique forquantitative delivery of aerosol. Pediatr Pulmonol. 1997; 23:278–286.35. Sixt R, Aronsson PH, Solymar L. The forced oscillation techniquein children—with special reference to the use of low frequencies.Eur Respir Rev. 1991; 1:222–225.36. Sly PD, Hayden MJ, Petak F, Hantos Z. Measurement of lowfrequencyrespiratory impedance in infants. Am J Respir Crit CareMed. 1996; 154:161–166.37. Lebecque P, Desmond K, Swartebroeckx Y, Dubois P, Lulling J,Coates A. Measurement of respiratory system resistance by forcedoscillation in normal children: A comparison with spirometricvalues. Pediatr Pulmonol. 1991; 10:117–112.38. Taussig LM, Chernick V, Wood R, Farrell P, Mellins RB. Standardizationof lung function testing in children. Proceedings andRecommendations of the GAP Conference Committee, Cystic FibrosisFoundation. J Pediatr. 1980; 97:668–676.39. Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M,Morgan WJ. Asthma and wheezing in the first six years of life. NEngl J Med. 1995; 332:133–138.40. Wang X, Wypij D, Gold DR, Speizer FE, Ware JH, Ferris BG,Dockery DW. A longitudinal study of the effects of parentalsmoking on pulmonary function in children 6–18 years. Am JRespir Crit Care Med. 1994; 149:1420–1425.