The Forced Oscillation Technique in Children with Respiratory Disease

L. Solymar, MD, P-H. Aronsson, MD, and R. Sixt, MD

Summary. The single-frequency forced oscillation technique was used to estimate the total respiratory impedance, resistance, and reactance at 2, 4, and I2 Hz of children who had cricoid stenosis (CS; n = II), provoked bronchoconstriction (PB; n = 6), or cystic fibrosis (CF; n = 13). The selection criteria for patients who had CS and CF were abnormal forced expiratory volume in 1 sec (FEV!) and/or maximal expiratory flow at 50% vital ca- pacity (Vmax 50%). Sixteen of the 17 patients who had CS or PB were found to have resis- tance (Rrs) values outside the normal range at 4 Hz; at 2 Hz, 14 were found to have abnor- mal values and, at I 2 Hz, only 9. The reactance in the CS and PB groups was somewhat less discriminative than RrS at all frequencies. Of the 13 patients who had CF, only 4 had abnormal Rrs values; in this clinical condition, the reactance at 2 Hz was the most dis- criminative variable, being abnormal in 6 of the 13 patients. Irrespective of the clinical group, the RrS was, in absolute terms, highest at 2 Hz, and it decreased with increasing frequency. This pattern of negative frequency dependence was, however, not obviously specific for any of the tested clinical conditions and offered no information in addition to a single low-frequency determination. (Key words: airway resistance measurement; discriminative value of variables; frequency dependence; site of obstruction.) Pediatr Pul- mono/ 1985; 1:256-261

The forced oscillation technique is of special interest in pediatric pulmonary physiology be- cause it requires only passive cooperation from the subject and, furthermore, does not influence the characteristics of the system to be measured. Although the effective resistance (Rrs) and the effective reactance (XrJ are not equivalent to or readily interpretable in terms of established pul- monary parameters, several reports have indi- cated the usefulness of the technique in monitor- ing short-term changes in the bronchial status. 1-4

Experience with the technique in other patho- logic conditions, however, is limited.5 Grimby et a1.6 reported a marked negative frequency de- pendence of R, in obstructive lung diseases, and this has been confirmed by others.’.* Mndser et al.’ in a study of adults who smoked, and Clement et aLL0 in a study ofadults who had respiratory symptoms, both found the mean of the first derivative of R,, (R’rs) to be superior to the mean Rrs. It is interesting that in the find- ings of Clement et al. the discriminative capac- ity of R’r, was improved in subjects where the frequency range could be extended down to 4

From the Departments of Pediatrics I and Pediatric Clinical Physi- ology of the University of Goteborg and the Research Laboratory of Medical Electronics, Chalmers University of Technology, Goteborg, Sweden.

Received December 27, 1984; revision accepted for publication June 19, 1985.

Address correspondence and reprint requests to Dr. Solymar: Department of Pediatrics Ostra sjukhuset, S-416 85 Goteborg, Sweden.

256

Hz.’O We therefore sought to investigate, firstly, whether there is relatively more information to be derived from the low-frequency range, and, secondly, whether any information is to be de- rived from the pattern of frequency dependence in addition to that obtained from a single low- frequency measurement. These issues were ex- amined in clinical situations in which there were different sites of airway obstruction.

Materials and Methods

Methods

The principle and technical details of the forced oscillation technique used have been de- scribed elsewhere. l l Further improvements have recently been presented. l2 Our method implies separate single-frequency determinations of R,, Xrs, and the magnitude of the impedance [Z,] of the respiratory system at 2,4, and 12 Hz, respectively. The frequency dependence has been characterized by the mean first derivative of the R, (R’rs),’ which has been calculated as the difference between Rrs at 12 and 2 Hz divided by the frequency range (see discussion). For the sake of comparison with other investi- gations we have determined all variables for the whole breathing cycle. Each determination was based on at least 80 oscillations (after rejection of markedly distorted oscillations), which, for the 2-Hz measurement, required a sampling time of approximately 60 seconds.

pediatric pulmonology Vol. 1, No. 5-Sept.-Oct. 1985 257

table I-Results of Routine Lung Function Testing for the Different Clinical Subgroups*

Height vc FEVi Vmaxsos. Nz (m) (1) (1) (4 (%Nn/l)

Cricoid stenosis 1.25 1.6 1 .o 0.9 3.0 (1 .I 8-1.40) (1.2-1.8) (0.7-1.3) (0.3-1.3) (1.3-6.7)

1.59 3.2 2.4 2.2 1.2 (1.55-1.70) (1.9-4.5) (1 .O-2.6) (1 .O-2.3) (0.6-14.2)

Bronchial provocation - - Positive reaction 1.56 2.9 2.5

(1.37-1.63) (2.5-3.9) (1.8-2.9) - - Negative reaction 1.58 2.9 2.6

(1.35-1.61) (2.8-4.1) (2.3-3.2)

Cystic fibrosis 1.24 1.6 1.2 1.4 5.7

1.61 2.8 2.0 1.8 3.8

(1 .I 8-1.41) (1.1-2.0) (0.6-1.4) (0.7-1.7) (4.1 -1 3.1)

(2.1 -1 2.1) (1.51-1.65) (2.3-3.1) (0.8-2.1) (0.4-2.2)

* Results are reported as median values with ranges in parentheses

Materials

Cricoid Stenosis

For this study we selected patients who showed persistent signs of airway obstruction in terms of FEV, and/or Vm, 50% below - 2 SD of predicted normal values (n = 11). On exercise, all, and, at rest, some had stridorous breathing. Due to the size-related frequency dependence of R, in normal subjects,I2 we selected the pa- tients so they would fit into two groups-one with a median height of 1.25 m (n = 5) and an- other with a median height of 1.59 m (n = 6).

Asthma

We studied the bronchial response of patients who had perennial bronchial asthma using a routine allergen challenge procedure with step- wide increments of inhaled birch or timothy pol- len concentration. Six patients showed positive response in that they developed rhonchi and/or dyspnea together with a drop in peak expiratory flow (PEF) of more than 20%. Four patients were considered to have a negative response. None were on medication at the time of the investi- gation.

Normals

Sixty-one healthy children aged 2 to 18 years who were without recurrent respiratory com- plaints or acute respiratory infection within three weeks prior to the testing were used to es- tablish the reference equations (Table 2). From these subjects we selected size-matched controls for the subjects in the study group.

Results

R rs

The mean total Rrs at 2 Hz was significantly higher than at 4 Hz (P < 0.001), which, in turn, was higher than at 12 Hz (P < 0.001) in all pa- tients in the study group irrespective of disease or height. In the control subjects Rrs at 2 Hz also significantly exceeded that at 4 Hz, but a significant difference between Rrs at 4 and 12

table !&Prediction Equations for Total Respiratory Resistance (Rrs), Total Respiratory Reactance (Xrs), and the Magnitude of Total Respiratory Impedance ([zrs]) in crn H2O.l-'.s at 2, 4, and 12 Hz and the First Derivative of Resistance between

Frequencies of 2-12 Hz (R'rs) in crn H20.1-1~s.Hz-1.

Variable = y n Regression Equation RSD

Cystic Fibrosis

We evaluated 25 patients who had CF using routine lung function tests and selected those who had FEV, and/or Vm, 50% below -2 SD of predicted normal values. The selected patients were subdivided into the two groups described above (median height 1.24 m, n = 7; median height 1.61 m, n = 6).

The results of routine lung function testing of these patient groups are given in 'hble 1.

R n 2 HZ 4 Hz

12 Hz

XrS 2 Hz 4 Hz

12 Hz

4 Hz 12 Hz

[ Z r s l 9 Hz

- R'iS 2-12 HZ

61 61 61

61 61 61

61 61 61

61

103 y=2.057-0.849~ Ht* 103 y = 1.845 - 0.795 x Ht 109 y= 1.631 - 0.649 x Ht

103 ( - y)= 1.690 - 0.781 x Ht 103 (1 - y)= 1.337 - 0.573 x Ht 103 (2 - y) = 1.327 - 0.622 x Ht

109 y = 2.088 - 0.826 x Ht 103 y= 1.914 - 0.816 x Ht 103 y = 1.695 - 0.684 x Ht

103 (1 - y)= 0.435 - 0.237 x Ht

0.1155 0.0989 0.0853

0.0980 0.0837 0.1 402

0.0994 0.0996 0.0887

0.649

* Ht = height in meters

Forced Oscillation Technique in Children-Solymar el a/. 258

2.0

- YI - L k

1 p 10

LL

01

Ht=1.3’0.12 A patlents ln.5) 4 * controls ln=12)

2 4 12 Frequency I Hz I

T

Ht=1.620.10 A patients In-6) . controls In=lL)

i 12

)L. 1 . . . 2 L

Frequency 1 Hz I

figure 1 -Values for total respiratory resistance (RrS) and reactance (X,) determined a t oscillation frequencies of 2, 4, and 12 Hz in patients who had cricoid stenosis and controls. Values are reoorted as mean and SEM.

Hz was found only in small children (height <1.4 m). The most pronounced increase of resistance compared with the controls was found at 2 and 4 Hz in small children who had CS and in chil- dren who had PB, while children who had CF had only a very moderate increase at all frequen- cies (figures 1-3, upperpanels). Among the sub- jects who had CS and PB 16 out of 17 were clas- sified as abnormal (Rrs > + 2 SD) at oscillation frequencies of 4 Hz and 14 at 2 Hz, while Rrs at 12 Hz was abnormal in only 9 (figure 4, upper panels). In the CF group only a minority of the children had abnormal resistance irrespective of frequency.

R’,

As shown in figures 1,2, and 3 (upperpanels) there was, on the average, a negative frequency dependence in both study and control groups. The frequency dependence was most pro- nounced in small children who had CS or PB (fig- ure 1, upper left panel; figure 2, upper right panel). Children who had CF, who had only slightly increased resistance, showed a fre- quency pattern rather similar to that of the con- trol subjects. The number of subjects classified as abnormal by values for R’rs that were 2 SD or more outside the mean was less than for all the other tested forced oscillatory variables except Rrs at 12 Hz (figure 4). All those who had abnor- mal R’rS also had Rrs values which were 2 SD outside the norm at 2 and 4 Hz.

X , In the study group the reactance shifted to-

ward more negative values at all frequencies, particularly in patients who had CS or PB. The negative shift seemed to be relatively more pro- nounced at 12 Hz. The number of CS and PB pa- tients classified as abnormal by values for XrS that were 2 SD or more outside the mean was somewhat less than that for R,. In subjects who had CF, Xrs at 2 Hz was the most dis- criminative variable, being abnormal in 6 out of 13 patients (figure 4, middle panels).

2,

The magnitude of the total respiratory im- pedance was as discriminative as Rrs and Xrs at low frequencies but was superior to Rrs at 12 Hz (figure 4, lower panels).

Discussion

In our evaluation of the forced oscillation tech- nique we chose to study three conditions in which the main pathologic involvement was at different levels of the respiratory system. CS is considered to represent a proximal and CF mainly a peripheral type of obstruction, while in PB the site of constriction is less well defined and may vary from subject to subject.

In the process of understanding the forced os- cillatory variables it is often useful to think in terms of a model of the respiratory system with parallel compartments, each with a resistance (R), a compliance (C), and an inertance (I) in se-

o pnprov In=W . cmtrok ln.17) b posrevt (n.61 o neg.react.(n=LI . CMroLs ln.17)

2 L 12 Frcqwncy rHll

2 L 12 Freqwncy (Hr I

figure 2-Values for total respiratory resistance (R(s) and reactance (Xrs) determined at oscillation frequencies of 2, 4, and 12 Hz in patients performing bronchial provocation tests and in controls. Values are reported as mean and SEM.

pediatric pulmonology Vol. 1, No. 5--Sept.-Oct. 1985 259

ries. The flow-response of such a single com- partment to an applied oscillatory pressure is, at each frequency, described by its impedance (2). The magnitude of this impedance ( [ Z ] ) is the composit effect of the compartmental resistance (R) and reactance (X), the latter being the net ef- fect of the opposing compliance- and inertia- related components. Even though R in a single compartment is constant, [Z] changes with fre- quency and is dominated by the compliant prop- erties at low frequencies, but, as frequency in- creases, the inertial properties become more important. In a parallel model with unequal compartments the oscillatory flow will be dis- tributed between the compartments in an in- verse relationship to their respective [Z] at that frequency. A consequence of this frequency- dependent redistribution of flow is that at each frequency the forced oscillatory variables will predominantly reflect the Compartment having the lowest [Z] , and the estimated overall resis- tance will change with frequency.

In agreement with the work of other^^,^,^^ we found in both our study and control groups a varying degree of negative frequency depen- dence of Rrs consistent with a frequency- dependent redistribution of the oscillatory flow as discussed above. It has been suggested that in the upper airways both the oral cavity and the larger airway walls act as shunting compart- ments to applied ~ s c i l l a t i o n s . ~ ~ ' ~ ~ ' ~ At low fre- quencies the low compliance of the upper air- ways constitutes a very high parallel impedance (low shunting effect) which, however, progres- sively diminishes as the frequency increases (in- creasing shunting effect). This means that the properties of the shunting compartment will be increasingly reflected in the forced oscillatory variables at higher frequencies. Accordingly, we found in the CS group a considerable reduction of Rrs from 2 to 12 Hz. Rrs at lower frequencies was also the most discriminative forced oscilla- tory variable in separating CS from normal sub- jects. We also observed markedly negative Xrs in this group, a finding that can hardly be ex- plained by altered elastic properties distal to the cricoid obstruction but rather by an increased influence of the upper airways having a very low compliance. In the PB group the magnitudes and frequency patterns of Rrs were similar to that of the CS group, which indicates that the airway obstruction was mainly centrally located. The less discriminative capacity of Rrs at 12 Hz as compared with 2 to 4 Hz in the PB group may, as in CS, be attributed to the shunting effect of the upper airways with a possible contribution of the most proximal bronchi.

0,

H k 1 . 3 0.12 . patientslnd) . controls(n-12)

Ht=1,6t0.10 patientsln-61 - controls (n=lLI

;4 4 2 L 12 2 L 12

Frequency ( H z ) Frequency ( H z l

figure 3-Resistance (RG) and reactance (Xrs) determined at 2, 4, and 12 Hz in patients who had cystic fibrosis and con- trols. Values are reported as mean and SEM.

In CF the obstruction is predominantly lo- cated in the peripheral airways while the cen- tral airways may be unaffected. Normally the pe- ripheral airway resistance constitutes a minor part of the overall re~istance'~ and may be sub- stantially increased with only a moderate in- crease of the overall resistance. This could be a partial explanation of the very moderate in- crease and low discriminative power of Rrs in our CF group. Moreover, the peripheral obstruc- tion is often unevenly distributed: a redistribu- tion of flow between peripheral compartments with time constant inequalities, and, conse- quently, a frequency dependence of peripheral resistance and compliance, could therefore be expected. This peripheral redistribution is, how- ever, most pronounced at frequencies below l Hz. '~ TO that effect we found only a low degree of frequency dependence of Rrs between 2 and 12 Hz comparable to that of the control subjects. The frequency-dependent decrease of lung com- pliance, which would be apparent even at 2 Hz, may contribute to the relatively higher discrim- inative ability of Xrs in CF.

The frequency dependence of resistance ex- pressed as its mean derivative, R'rs, has been suggested to be more sensitive than the mean Rrs in separating smokers from nonsmokers9 and patients who have respiratory symptoms from normal subjects.'o The mean Rfrs is, as can be shown mathematically, derived from the difference in Rrs at the two extreme frequencies and is quite independent of the shape of the Rrs curve in between. An increased negative fre- quency dependence implies that Rrs at the low-

Forced Oscillation Technique in Children-Solymar et a/.

X,' 2nz

-151' A .

Iml

frequency endpoint is more elevated than is Rrs at the high-frequency endpoint, which is also the case in obstructive disease.6-8,10 However, be- cause in obstructive conditions R, is increased, although disproportionately, at both endpoints, one would expect the mean R'rs to be less dis- criminative than R,, at the low-frequency end- point. Our results support this view. As shown in figure 4, the number of patients classified as abnormal by Rfrs was less than, and besides, in- cluded in, those detected by Rrs at 2 Hz alone. Thus, no additional information was gained from the frequency-dependent change of the re- sistance when compared with single low- frequency determinations of Rrs.

Conclusion

Clinical conditions with central airway ob- struction were adequately detected by the forced oscillation technique. Resistance determina- tions at 2 and 4 Hz were superior to those at 12 Hz. The unfavorable signal-to-noise ratio, espe- cially at 2 Hz, may be overcome by refined filter- ing and prolonged data aquisition time; we be- lieve these efforts to be rewarded by the increased sensitivity of low-frequency determinations. Al- though the frequency dependence of the resis- tance, expressed as the mean derivative of the resistance between 2 and 12 Hz, correlated with the degree of airway obstruction, it provided no information in addition to that provided by sin- gle low-frequency determination of Rrs. The pat-

figure 4-Individual data on resistance (Rrs), reactance (Xrs), and the magnitude of im- pedance (prs]) at frequencies 2, 4, and I 2 Hz and the mean first derivative of the resistance (R'rs) in patients who had cricoid ste- nosis (A), provoked bron- choconstriction (0) and cystic fibrosis (x). The lines indicate mean f 2 SD.

tern of frequency dependence of R, was not specific for any of the clinical groups studied. Reactance, especially at 2 Hz, was the most sen- sitive variable for the detection of cystic fibrosis.

The considerable negativity of Xrs in cases with marked central airway obstruction may have been a secondary phenomenon to the in- creased resistance, reflecting the increased con- tribution of the low compliant properties of the shunting upper airways.

References

1. Lenney W, Wilner AD. Recurrent wheezing in the preschool child. Arch Dis Child 1978; 53:468-473.

2. Baur X, Bergstermann F, Fruhmann H, Polke H, Pram1 G. Oszil- latorische und ganzkorperplethysmographische Messung der Atemwegwiderstandes bei allergeninduzierten Bronchialob- struktionen. Atemwegs- und Lungenkr 1978; 4:262-264.

3. Klein C, Vrbanek R, Kohler D, Matthys H. lnhalative bronchiale Provokations bei Kinder Vergleichende Messungen des Oscillations-Verschlussdruck- und Plethysmographischen Resis- tance. Klin Padiatr 1983; 195:33-37.

4. Solymar L, Aronsson P-H, Engstrom I, Bake B, Bjure J. Forced os- cillation technique and maximum expiratory flows in bronchial provocation tests in children. Eur J Respir Dis (in press).

5. Cogswell J. Forced oscillation technique for determination of re- sistance to breathing in children. Arch Dis Child 1973;

6. Grimby G, Takishima T, Graham W, Macklem PT, Mead J. Frequency dependence of flow resistance in patients with obstructive lung disease. J Clin Invest 1968; 47:1455-1465.

7. Holle Je Undser FJ, Schuller B, Hartmann V, Magnusson H. Mea- surement of respiratory mechanics with forced oscillations. Respiration 1981; 56:1210-I 230.

8. Michaelson ED, Grassman ED, Peters WR. Pulmonary mechanics

4 a : w - w

pediatric pulmonology Vol. 1, No. 5-Sept.-Oct. 1985 26 1

by spectral analysis of forced random noise. J Clin Invest 1975;

9. Wndser FJ, Clement J, Van de Woestijne KI? Normal values of to- tal respiratory resistance and reactance determined by forced oscillations. Influence of smoking. Chest 1982; 81 :586-591.

10. Clement J, Bndstr FJ, Van de Woestijne KP Total resistance and reactance in patients with respiratory complaints with and with- out airway obstruction. Chest 1983; 2:215-220.

11. Aronsson P-H, Solymar L, Dempsey J, Bjure J, Olsson T, Bake B. A modified forced oscillation technique for measurements of respiratory resistance. J Appl Physiol 1977; 42:650-655.

12. Solymar L, Aronsson P-H, Bake B, Bjure J. Pulmonary resistance and impedance magnitude in healthy children aged 2-18 years. Pediatr Pulmonol 19j5; 1:134-140.

13. Zapletal A, HouItEkJ, SambnekM, Vavrova V, Srajer J. Lungfunc-

56:1210-1230. tion abnormalities in cystic fibrosis and changes during growth. Bull Eur Physiopathol Respir 1979; 15:575-592.

14. Stanescu D, Moavero NE, Veriter C, Brasseur L. Frequency depen- dence of respiratory resistance in healthy children. J Appl Phys- iol 1979; 47:268-272.

15. Mead J. Contribution of compliance of airways to frequency- dependent behaviour of lungs. J Appl Physioll969; 26:670-673.

16. Cauberghs M, Van de Woestijne Kf? Mechanical properties of the upper airway. J Appl Physiol: Respirat Environ Exercise Physiol

17. Ferris BG, Mead J, Opie LH. Partitioning of respiratory flow resis- tance in man. J Appl Physiol 1964; 19:653-658.

18. Otis AB, McKerrow CB, Bartlett RA, Mead J, Mcllroy MC, Selver- stone NJ, Redford EP Jr. Mechanical factors in distribution of pulmonary ventilation. J Appl Physiol 1956; 83427-443.

1983; 55:335-342.