forced oscillation technique and impulse oscillometry

34
CHAPTER 5 Forced oscillation technique and impulse oscillometry H.J. Smith*, P. Reinhold # , M.D. Goldman } *Research in Respiratory Diagnostics, Berlin, Germany. # Friedrich-Loeffler-Institute, Jena, Germany. } David Geffen School of Medicine, University of California, Los Angeles, USA. Correspondence: H.J. Smith, Research in Respiratory Diagnostics, Bahrendorfer Str. 3, 12555 Berlin, Germany. Conventional methods of lung function testing provide measurements obtained during specific respiratory actions of the subject. In contrast, the forced oscillation technique (FOT) determines breathing mechanics by superimposing small external pressure signals on the spontaneous breathing of the subject [1]. FOT is indicated as a diagnostic method to obtain reliable differentiated tidal breathing analysis. Because FOT is performed without closure of a valve connected to the mouthpiece, and without maximal or forced respiratory manoeuvres, it is unlikely that FOT itself will alter airways smooth muscle tone [2]. FOT utilises the external applied pressure signals and their resultant flows to determine lung mechanical parameters. These pressure–flow relationships are largely distinct from the natural pattern of individual respiratory flows, so that measured FOT results are, for the most part, independent of the underlying respiratory pattern. Therefore, oscillometry minimises demands on the patient and requires only passive cooperation of the subject: maintenance of an airtight seal of the lips around a mouthpiece and breathing normally through the measuring system with a nose-clip occluding the nares. Potential applications of oscillometry include paediatric, adult and geriatric populations, comprising diagnostic clinical testing, monitoring of therapeutic regimens, and epidemiological evaluations, independent of severity of lung disease. Oscillometry is also applicable to veterinary medicine. The last two main sections Relevance of oscillometry in clinical practice and Oscillometry in the clinical pulmonary laboratory emphasise clinical aspects of application and interpretation of FOT rather than methodological details and technological solutions, which are discussed in the two sections that follow immediately below. Clinical application of FOT does not require mastery of the mathematical infrastructure of the technical methodology, and readers interested in the clinical use of FOT may find it more useful to begin with these clinical sections and refer subsequently to methodological and technological details. Methodology of impulse oscillation technique The mechanical basis of oscillometry concerns use of external forcing signals, which may be mono- or multifrequency, and applied either continuously or in a time-discrete manner. The impulse oscillation technique is characterised by use of an impulse-shaped, time-discrete external forcing signal. Eur Respir Mon, 2005, 31, 72–105. Printed in UK - all rights reserved. Copyright ERS Journals Ltd 2005. 72

Upload: lydiep

Post on 13-Feb-2017

254 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Forced oscillation technique and impulse oscillometry

CHAPTER 5

Forced oscillation technique and impulseoscillometry

H.J. Smith*, P. Reinhold#, M.D. Goldman}

*Research in Respiratory Diagnostics, Berlin, Germany. #Friedrich-Loeffler-Institute, Jena, Germany.}David Geffen School of Medicine, University of California, Los Angeles, USA.

Correspondence: H.J. Smith, Research in Respiratory Diagnostics, Bahrendorfer Str. 3, 12555 Berlin,Germany.

Conventional methods of lung function testing provide measurements obtained duringspecific respiratory actions of the subject. In contrast, the forced oscillation technique(FOT) determines breathing mechanics by superimposing small external pressure signalson the spontaneous breathing of the subject [1]. FOT is indicated as a diagnostic methodto obtain reliable differentiated tidal breathing analysis. Because FOT is performedwithout closure of a valve connected to the mouthpiece, and without maximal or forcedrespiratory manoeuvres, it is unlikely that FOT itself will alter airways smooth muscletone [2].

FOT utilises the external applied pressure signals and their resultant flows to determinelung mechanical parameters. These pressure–flow relationships are largely distinct fromthe natural pattern of individual respiratory flows, so that measured FOT results are, forthe most part, independent of the underlying respiratory pattern. Therefore, oscillometryminimises demands on the patient and requires only passive cooperation of the subject:maintenance of an airtight seal of the lips around a mouthpiece and breathing normallythrough the measuring system with a nose-clip occluding the nares. Potentialapplications of oscillometry include paediatric, adult and geriatric populations,comprising diagnostic clinical testing, monitoring of therapeutic regimens, andepidemiological evaluations, independent of severity of lung disease. Oscillometry isalso applicable to veterinary medicine.

The last two main sections Relevance of oscillometry in clinical practice andOscillometry in the clinical pulmonary laboratory emphasise clinical aspects ofapplication and interpretation of FOT rather than methodological details andtechnological solutions, which are discussed in the two sections that follow immediatelybelow. Clinical application of FOT does not require mastery of the mathematicalinfrastructure of the technical methodology, and readers interested in the clinical use ofFOT may find it more useful to begin with these clinical sections and refer subsequentlyto methodological and technological details.

Methodology of impulse oscillation technique

The mechanical basis of oscillometry concerns use of external forcing signals, whichmay be mono- or multifrequency, and applied either continuously or in a time-discretemanner. The impulse oscillation technique is characterised by use of an impulse-shaped,time-discrete external forcing signal.

Eur Respir Mon, 2005, 31, 72–105. Printed in UK - all rights reserved. Copyright ERS Journals Ltd 2005.

72

Page 2: Forced oscillation technique and impulse oscillometry

The most useful aspect of applying FOT pressure pulses rather than pseudo-randomnoise (PRN) is improved time resolution of the measurement. The impulse oscillationtechnique allows measurement of up to 10 impedance spectra per second. This permits auseful analysis of intra-breath variation in impedance, comparable to that obtained withmono-frequency applications. However, a disadvantage of such high impulse rates is theinability to record longer respiratory time constants that may be more informativeconcerning respiratory abnormalities. For this reason, the common application ofimpulse oscillation utilises recordings of 5 impedance spectra (5 impulses) per second.

An additional benefit of impulse oscillation is the simplicity of the hardware needed togenerate the forced signals, allowing smaller, more efficient electronic and mechanicalstructures with minimal power loss.

A unique aspect of applying pulses of pressure to the respiratory system is the fact thatthe entire energy of all applied pressure harmonics is applied within a very short timeperiod. This causes a higher impact to the respiratory system compared with sinusoidalor PRN applied pressures, and may be perceived by some patients as a slightlyunpleasant respiratory sensation during measurement.

Peculiarities of aperiodic waveforms

The impulse oscillation method applies aperiodic waveforms using an impulsegenerator that produces pressure pulses of limited magnitude and 30–40 ms duration.These pulses define specific amplitudes and phases of the inherent sinusoidalcomponents. The time-course of such practical pressure pulses applied to the respiratorysystem is not a true Dirac-impulse, defined as having virtually infinite magnitude andinfinitesimal time duration, which would provide a continuous spectrum of frequencieswith the same amplitude. Thus, the terms "impulse-shaped" oscillations and "impulsepressures" are used to indicate realistic practical pressure pulses rather thanmathematically defined impulses.

The short duration of the impulse-shaped waveform itself provides linearity ofpressure and flow signals in the face of within-breath dynamic changes in the respiratorysystem. In contrast, the longer time needed with PRN to embed a range of periodicfunctions decreases time resolution, resulting in increased noise of calculated impedancerelated to any time-dependence of dynamic changes in the respiratory system.

The characteristic feature of any aperiodic waveform is the resulting continuousspectrum after transformation of its time course into the frequency domain, using aFourier integral rather than a Fourier series, in the fast Fourier transform (FFT).

Thus, the advantage of continuous spectra is particularly important in abnormalrespiratory systems with regional nonhomogeneities (fig. 1), where resistance, reactanceand coherence spectra may manifest deviations from the normally smooth and uniformlycontinuous spectral courses.

In contrast, spectra resulting from FFT analyses of periodic multifrequency forcingsuch as PRN [3] are discontinuous. Discrete values of impedance are obtained with afrequency resolution determined by the frequencies of the included sinusoidalcomponents. As a result, the course of such discrete spectra often may require post-processing to smooth the PRN spectra [3]. To improve interpretation of discrete PRNspectra, it is common to approximate the overall spectral range by smoothing with linear,quadratic or logarithmic functions. However, such smoothing inherently diminishesinformation contained in characteristic peaks and plateaus of impedance that mayotherwise provide insight into superimposed parallel resonance phenomena, e.g. relatedto upper airway constriction.

Impulse power spectra of pressure and flow generated by the impulse oscillometry

OSCILLOMETRY: FOT AND IOS

73

Page 3: Forced oscillation technique and impulse oscillometry

system (IOS) are shown in figure 2 over a frequency range of 0.1–35 Hz. The energydistribution provides practical assessment of low (ƒ5 Hz), as well as high (w20 Hz)frequency ranges, with decreased amplitude at higher frequencies to minimise non-linearities due to acceleration of the moving air column [4]. Enhanced amplitudes at the

���

���

���

���

���

���

���

���

����

����

� �

� � �� �� �� �� �� ��

���

���

���

���

��

���

��

���������������

���������������������

�� ��������������

������������������

��������������

!�"#��� ���$%

&'�(()�

Fig. 1. – Representative data for spectra of respiratory resistance (Rrs(f); ––), reactance (Xrs(f); - - -) andcoherence (c2(f); - – - – ) are plotted, between 3 and 35 Hz, for a normal adult, during forced oscillation usingpulse-shaped forcing generated by the impulse oscillometry system. Resonant frequency (fres) is shown where theXrs(f) tracing crosses zero. The shaded area, below zero Xrs and above Xrs(f) tracing between 5 Hz and fres, isthe integral of Xrs(f) from 5 Hz to fres, and is designated the reactance area (AX). Regional nonhomogeneitiesmay manifest deviations from the normally smooth and uniformly continuous spectral courses.

��

���

���

���

���� � �� �� �� �� �� ��

!�"#��� ���$%

*����#

������+,

Fig. 2. – Power spectra of flow (––), and pressure (----), for the discrete pulse-shaped forcing signal generated bythe impulse oscillometry system. Spectra plotted at frequencies 0.1–35 Hz. Pressure and flow power are highestat 3–20 Hz. Less power is needed at frequencies w20 Hz, because "competing" higher harmonics of the patient’srespiratory flow are very small at these frequencies.

H.J. SMITH ET AL.

74

Page 4: Forced oscillation technique and impulse oscillometry

lower frequencies limit the influence of higher harmonics of spontaneous breathingfrequencies.

Impulse oscillometry system

The IOS measuring-head (fig. 3) is functionally similar to PRN systems designed forthe determination of input impedance [1, 5–12]. The characteristic feature of the IOS [13]is the generation of recurrent aperiodic impulse-shaped forcing signals of alternatingdirection.

Flow is measured by a Lilly-type heated screen pneumotachograph with a resistance of36 Pa?s?L-1, providing a common-mode rejection ratio ofw60 dB up to 50 Hz [12, 14, 15]for the combination of pneumotachograph and flow transducer system. At flow ratesv15 L?s-1 the heated pneumotachograph is linear within 2%. The proximal side of thepneumotachograph is connected to a pressure transducer. To guarantee suppression oftechnical influences and to avoid phase differences, both pressure and flow channels usematched transducers of the same type, SensorTechnics SLP 004D [16]. Pressure and flowsignals are sampled at a frequency of 200 Hz and digitally converted by a 12-bitanalogue-to-digital converter. Analogue anti-aliasing low-pass filtering is realised by afourth-order Bessel filter at a border frequency of 50 Hz, providing a damping at Nyquistfrequency ofy75 dB.

Tuning of the IOS impulse generator involves both volume displacement of theloudspeaker membrane and magnitude of the terminating resistor. The terminal resistor

)�#+-��&�./'

0����������

�/�����������-#��

!��1���+#��

'�#�����+#��

2�#��-����

2���������

3��+�-��

'��#���������-�������+�

4�-#����������

4�-#���5��1

��-���� �5��1

Fig. 3. – Schematic diagram of the impulse oscillometry system measuring-head and connectors: loudspeakerenclosure at top, connected to a y-adaptor at one upper arm, an exit for flow with terminal resistor at thesecond upper arm and a lower arm connected to the pneumotachograph. A mouthpiece is connected to theopen side of the pneumotachograph. Pulsatile flows generated by the loudspeaker are shown as a lightly shadedthick line, part of which exits through the terminal resistor and part of which flows through thepneumotachograph and mouthpiece. The patient’s normal respiratory flow is shown as a shaded thick line fromthe mouthpiece though the Y-connector, exiting through the terminal resistor. The resistance of the terminalresistor is 0.1 kPa?s?L-1 and the deadspace of the Y-connector/pneumotachograph/mouthpiece is y70 mL.

OSCILLOMETRY: FOT AND IOS

75

Page 5: Forced oscillation technique and impulse oscillometry

provides a low-impedance pathway for normal respiratory flow, which, at the same time,is high enough to minimise loss of energy of superimposed impulses so that sufficientimpulse pressure is transmitted into the respiratory tract. Both components determinethe linear working range of the unit and range of input impedances, which can bemeasured to maintain international recommendations [16, 17]. A terminal resistor ofy0.1 Pa?s?L-1, in combination with a volume displacement of 40 mL, which isaccelerated by the loudspeaker membrane in v40 ms, results in maximal peak-to-peakimpulse pressures of 0.3 kPa and minimises interference of underlying higher harmonicsof respiratory frequency that contribute "noise" to the oscillatory pressure and flowsignal [10].

International recommendations for electromechanical performance are maintained inthe IOS by use of advanced transducer technology and global mean spectral data derivedfrom IOS are generally comparable to those obtained by the pseudo-random noisemethod of Landser [6], Delecourt et al. [8] and Skloot et al. [18].

The measurement is performed as follows: while the subject spontaneously breathesambient air via the tubing between mouthpiece and terminating resistor, the loudspeakergenerator transmits brief pressure impulses via Y-adapter, pneumotachograph andmouthpiece into the respiratory tract; pneumotachograph and pressure transducerregister the composite signals containing breathing activities and the forcing impulsesignal for further processing.

Further processing of digitised impulse data. Flow and pressure channels contain boththe underlying respiratory system flow and pressure, and the superimposed forcedoscillation signals.

By definition, respiratory input impedance is the transfer function or ratio of effectivepressure (Prs) and flow (V9rs), derived from the superimposed forced oscillations, afterbeing discriminated from underlying respiratory pressure and flow and their harmonics.In the mathematical sense, all components are considered "complex", characterised bymodulus and phase.

Multifrequency input impedance|Zrs(f )~ |

Prs (f )

V 0rs (f )0 < f ƒf maxf g ð1Þ

Discrimination of superimposed forced oscillations from the underlying respiratorypressure and flow in the IOS is focused on individual impulses, based on pressure andflow sampling intervals that include both the impulse stimulus and the respiratorysystem reaction to the impulse. Figure 4a gives an example of such a samplinginterval for flow.

A "baseline" straight line segment is inserted between the start- and end-points ofseparate sample segments of both pressure and flow. The baseline is a simple linearapproximation of the underlying respiratory flow and pressure throughout the singleimpulse excitation. Baseline approximation has proved to be a useful and reliabletechnique to decrease the respiratory component of the composite signals for pressureand flow. Spline reconstruction, sinusoidal approximation of underlying respiratorypressure and flow or high-pass digital filtering were not as useful. Baseline correction andoffset elimination, as can be seen in figure 4b, allow rectangular windowing prior to theFFT to effectively reduce spectral leakage and improve the signal-to-noise ratio.

Resolution of calculated pressure and flow spectra is increased by adding numericalzero values to the real sampling points of the corrected primary data segment. Thisprocedure allows formation of exactly 2n samples, compatible with FFT requirements.Choice of sampling interval as well as addition of zero values to this interval allow

H.J. SMITH ET AL.

76

Page 6: Forced oscillation technique and impulse oscillometry

adjustment of spectra concerning low-end border frequency as well as numericalresolution.

To further improve quality of calculated impedance data, impulses that do not fulfilldefined reliability criteria are rejected. The critical segments of respiration are the phasetransitions between inspiration and expiration. At these zero-crossings of pressure andflow, gradients of pressure and flow versus time are maximal, and the following principlesare implemented to establish reliability. Slopes of baseline corrections for pressure or floww0.7 indicate dominance of the underlying respiratory pressure–flow pattern, and the

����

*�-���#

+���5�-

��� �5��

1�)(��

�����

�����

����

����

����

����

��������� ��� ��� ��� ��� ���

/������������

*--�6�����+7������

'#�������(��

�����

����+�������������

���

���

����

���

���+

���-#

���5��1

�)(��

���

��

����������������

/��-�����������

���

���+

���-#

���-�#����

.#������5����������-#��

4�-#���5��1

,������

��#��������-#���-�#�

������8���-�����5��-���� � ���

��

7�

Fig. 4. – a) Primary flow recording. Note that the primary flow includes the patient’s expiratory respiratory flowwith pulses of "loudspeaker flow" superimposed. The dash-dot line shows the linearly approximated respiratoryflow if there were no superimposed pulsatile flow from the loudspeaker included. By referring all pulsatile flowand with a similar process utilised for the continuous pressure tracing to the approximated baseline, the flowand pressure related purely to the pulse generated by the loudspeaker, with the patient’s respiratory flow andpressure subtracted from the total pressure and flow signals, may be derived. b) Corrected impulse tracings offlow (––) and pressure (----), derived from baseline correction shown in figure 4a, as prepared for input into thefast Fourier transform. Duration of impulse corresponds to actual movement time of the loudspeaker diaphragm(y40 ms in duration). Initial flow response of the respiratory system begins almost immediately. Afterloudspeaker movement has ended, the respiratory system continues to respond in a "reactive" manner, withpressure reaching its peak and then declining, while flow falls below baseline, and then gradually returns tobaseline.

OSCILLOMETRY: FOT AND IOS

77

Page 7: Forced oscillation technique and impulse oscillometry

impulse is rejected because separation of the superimposed impulse pressure and flow isnot reliable. Absolute peak flow within the impulse segment must exceed 0.02 L?s-1.Pulses with less flow are rejected because of very small flow values in the denominator ofthe input impedance Equation (1) [19] and the resulting mathematical errors. Finally,impulses that yield negative resistance values at any frequency after FFT are rejected.

Impulse rate and sampling interval. Impulse rate and selected sampling interval effectson calculated impedance have been evaluated in vitro as well as in calves of comparablesize to adult humans [20, 21]. No significant effect of impulse rate was observed between 1and 5 impulses per second.

In contrast, different sampling durations led to significantly different low-frequencyrespiratory system reactance (Xrs) results. Using 16 sampling points (80 ms) for dataanalysis, no useful information was obtained for Xrs ƒ5 Hz. Use of 32 sampling points(160 ms) per impulse for data analysis provided useful Xrs5 data. However, samplingdurations of 320 ms did not improve impedance results and coherence c2

ƒ5 Hz wassignificantly decreased, consistent with deterioration of calculated impedance quality dueto interactions with spontaneous breathing signals. These findings underlie the currentrecommendation to use sampling intervals of 32 sampling points per impulse,corresponding to an optimised impulse rate of either 3 or 5 impulses per second.

Coherence. The coherence function is defined as the square of the cross-power spectrumdivided by the product of the auto-spectra of pressure and flow at any forcing frequency.Ranging between 0 and 1, it is a measure of the available linearity [22].

c2(f )~GV 0P(f )j j2

GV 0V 0(f ):GPP(f )0 < f ƒf maxf g ð2Þ

Landser et al. [6] found that when the coherence c2 i0.95, PRN impedance datashow a coefficient of variation CV%v10%. Subsequently, coherence value thresholdsof 0.9 or 0.95 have been widely used to accept FOT data. However, if methods otherthan PRN forcing input signals or data acquisition are used, this original thresholdvalue of coherence is not applicable. Miller and Pimmel [23] showed that estimatedvariance of calculated impedance is a function of coherence and the number ofestimates averaged. Use of pseudo-random noise techniques commonly includes threemeasures of impedance of 16 s each [5]. Three replicate measures require coherence c2

w0.9 to yield an estimated standard error of 10%. In contrast, IOS commonlyincludes the average of w100 separate FFT analyses and, accordingly, requires lessperfect coherence for the average data to provide an estimated standard error ofv10%. For clinical purposes it is recommended to use a coherence value of c2 i0.6 at5 Hz for the acceptance threshold, provided that at least 100 FFT analyses areaveraged. Coherence improves as a function of oscillatory frequency to c2 i0.9 at10 Hz and higher frequencies (fig. 1).

Interpretation of oscillation mechanics

Monofrequency oscillations provide a measure of total respiratory impedance (Zrs)that includes airway resistance, and elastic and inertive behaviour of lungs and chest wallat one oscillation frequency. In contrast, multifrequency oscillation methods, such aspseudo-random noise or impulse oscillation provide measures of respiratory mechanicalproperties in terms of Zrs, as a function of frequency (f), allowing the recognition ofcharacteristic respiratory responses at different oscillation frequencies.

E

H.J. SMITH ET AL.

78

Page 8: Forced oscillation technique and impulse oscillometry

Zrs has been described in engineering terms by so-called "real" and "imaginary"components. In medical use, preferred terms are resistance (Rrs) and reactance (Xrs),respectively.

|Zrs(f )~Rrs(f )zjXrs(f ) 0 < f ƒf maxf g ð3Þ

In contrast to Rrs, concepts of Xrs are not yet widely appreciated, largely because ofgreater complexity of reactive parameters, as well as their numerical characteristics.The present discussion emphasises the importance of consideration of both Xrs aswell as Rrs for interpretation of respiratory mechanical properties.

Most commonly, the oscillation frequency scale utilised for multifrequency oscillationmethods includes about 5–30 Hz [7, 17, 24]. Oscillation frequenciesv5 Hz are affected byhigher harmonics of underlying natural respiratory frequency [25]. Higher oscillationfrequencies are increasingly affected by "shunt" properties of the upper airways [6, 13, 26]and capacitive impact of face masks [27] when used.

Respiratory resistance

The resistive component of respiratory impedance, Rrs, includes proximal and distalairways (central and peripheral), lung tissue and chest wall resistance. Normally, centralresistance dominates, depending on airway calibre and the surface of the airway walls,while lung tissue and chest wall resistance are usually negligible [7, 28].

Rrs may be considered within normal limits if Rrs at 5 Hz (Rrs5) is within ¡1.64 sd ofthe predicted value. Rrs5 values between 1.64 and 2 sd above predicted may be consideredminor, w2 sd moderate and w4 sd above predicted severe obstruction.

In previous reports the calculation of per cent predicted has been used. Rrs5 values thatdid not exceed 150% predicted, defined in bronchial challenge comparing changes in Rrs5

to a 20% decrease in forced expiratory volume in one second (FEV1) and 50% increase inairway resistance (Raw), were considered within normal limits [29–31]. However, it is nowrecognised that a 20% decrease in FEV1 is a substantial abnormality, and normal limitsfor Rrs might be more profitably defined in their own right, without requiring a specificrelationship to arbitrary spirometric criteria.

In healthy subjects, Rrs is almost independent of oscillation frequency [7, 32, 33], butmay increase slightly at higher frequencies due to the upper airways shunt effect [6, 26].

When proximal (central) or distal (peripheral) airway obstruction occurs, Rrs5 isincreased above normal values. The site of airway obstruction is inferred from thepattern of Rrs, as a function of oscillation frequency, adjusting as necessary for subjectage [2, 25, 34–40]. Proximal airways obstruction elevates Rrs evenly independent ofoscillation frequency [32].

In distal airways obstruction, Rrs is highest at low oscillation frequencies and falls withincreasing frequency. This negative frequency-dependence of Rrs has been explained interms of intrapulmonary gas flow redistribution, due either to peripheral pulmonarynonhomogeneities or to changes in peripheral elastic reactive properties [6, 8, 34, 35]. Asperipheral resistance increases, Rrs becomes more frequency dependent [26, 36–38].Frequency dependence of Rrs may be a normal finding in small children [39, 40] and maybe greater than in adults in the presence of peripheral airflow obstruction [8].

Respiratory reactance

The reactive component of respiratory impedance, Xrs incorporates the mass-inertiveforces of the moving air column in the conducting airways, expressed in the terminertance (I) and the elastic properties of lung periphery, expressed in the term

OSCILLOMETRY: FOT AND IOS

79

Page 9: Forced oscillation technique and impulse oscillometry

capacitance (Ca).

Xrs(f )~v:I{1

v:Cav~2:p:f 0 < f ƒf maxf g ð4Þ

Most importantly, in medicine, it is emphasised that respiratory Ca is not identical tocompliance. The component of Xrs associated with Ca is defined to be negative insign. It is most prominent at low frequencies. In contrast, the component of Xrs

associated with inertance is always positive in sign and dominates at higherfrequencies. Thus, interpretation of Xrs is primarily influenced by the oscillationfrequency range under consideration.

Low frequency capacitive Xrs essentially expresses the ability of the respiratory tract tostore capacitive energy, primarily resident in the lung periphery. In both fibrosis andemphysema this ability is reduced: in fibrosis because of the stiffness of the lung; inemphysema because of hyperinflation and loss of lung elastic recoil. Distal capacitivereactance at 5 Hz (Xrs5) manifests increasingly negative values either in restriction or inhyperinflation. Thus, Xrs5 characterises the lung periphery, but is nonspecific as to thetype of limitation. Additional information is needed to differentiate peripheralobstruction from peripheral restriction. This is not usually problematic in clinicalpractice.

Capacitive and inertive elements have been modelled by a number of authors [2]. Manysimplifications have been attempted that include serial and parallel circuit elements, todefine an approximation of a normal Xrs spectrum with a zero-crossing exactly atresonant frequency and positive slope throughout. The frequency range utilised inmultifrequency oscillometric forcing signals should allow for determination of the serialresonance of the respiratory system under investigation [5].

Resonant frequency. Resonant frequency (fres) is defined as the point at which themagnitudes of capacitive and inertive reactance are equal.

v0:I~

1

v0:Ca

v0~2:p:fres ð5Þ

fres~1

2:p:ffiffiffiffiffiffiffiffiffiffi

I :Cap ð6Þ

Because fres can vary over a considerable range and, thereby, appear in closeproximity to oscillation frequencies dominated by either capacitative or inertiveproperties, this parameter should not be directly interpreted in terms of a particularmechanical property of the respiratory system. However, it is a convenient marker toseparate low-frequency from high-frequency Xrs. Thus, low-frequency capacitive Xrs

is dominant at oscillation frequencies below fres, while high-frequency inertive Xrs isdominant at frequencies above fres. In normal adults, fres is usually 7–12 Hz [33]. Inhealthy young children, fres is larger than in adults and increases with decreasing age.

In respiratory disease, both obstructive and restrictive impairments of the distalrespiratory tract cause fres to be increased above normal [7, 26, 36]. The relevance of fres isnormally revealed in within-individual trends over time during bronchial or therapeuticchallenge.

Capacitive spectral range of reactance. Thoraco-pulmonary elasticity is commonlyviewed as a static property, and normally is investigated in the absence of resistive orinertive mechanical losses. However, in oscillometry, capacitive Xrs is believed tocomprise useful information concerning peripheral airways mechanical properties. In

H.J. SMITH ET AL.

80

Page 10: Forced oscillation technique and impulse oscillometry

practice, determination of peripheral capacitive Xrs is determined at the lowest frequencythat is not highly interfered with by fundamental respiratory frequency and its harmonics[41]. In the IOS method, Xrs5 is commonly utilised. In children breathing at highrespiratory frequencies, reactance at 10 Hz may be useful [42] in following bronchial and/or therapeutic challenge.

Interpretation of Xrs5, is clearly different from that of conventional lung function testparameters and, in particular, lung compliance. One striking feature of Xrs5 is its negativevalue. The minus sign is derived from a general definition in natural sciences todifferentiate elastic properties from moments of inertia, the latter of which is alwayspositive. Therefore, the minus sign simply confirms a relationship with elastic properties.

Definition of abnormality has previously been based on increased negative readingsrelated to expected normal values of Xrs5. A differencew0.15 kPa?s?L-1 is most definitelyagreed to imply abnormal lung function, independent of Rrs, although abnormal lungfunction may be present with smaller differences.

Capacitance versus dynamic pulmonary compliance. Dynamic pulmonary complianceis derived from the relationship between oesophageal pressure and changes in lung volume[43]. In contrast, oscillometry assesses the elastic properties of the respiratory system fromthe out-of-phase relationship of simultaneously recorded transrespiratory pressure (Prs)and central flow signals (V9rs), measured from superimposed oscillations and transformedinto Xrs. Therefore, the term capacitance, Ca, an equivalent of capacitive phaseinformation between the primary signals Prs and V9rs should be used. The frequency rangefor such measures is always below fres.

In pulmonary fibrosis, dynamic lung compliance (CLdyn) is decreased below normal. Ina similar manner, oscillometry yields a decreased estimate of Ca, due to negativedisplacement of low frequency Xrs. Both dynamic lung compliance and Ca reflect elasticlimitation and they trend together. Previous oscillometry studies have reported lesssensitivity than dynamic lung compliance in the early stages of restrictive disease [7, 17].

In contrast, pulmonary hyperinflation is associated with loss of lung elastic recoil andincreased CLdyn. Because of the loss of lung elastic recoil, peripheral airways are notsupported externally by lung recoil and the resultant partial peripheral airwayobstruction prevents the applied oscillometric signals from reaching peripheralcompliant areas. In this way, loss of lung elastic recoil indirectly causes a decrease inCa, with associated increased negative magnitude of low frequency Xrs, [26, 36]. Indeed,low frequency Xrs is particularly sensitive to pulmonary hyperinflation, and while Rrs5

may be nearly normal or only moderately increased, Xrs5 is highly abnormal [44].

Reactance area. The index designated reactance area (AX) is a quantitative index of totalrespiratory reactance Xrs at all frequencies between 5 Hz and fres. The integration of thesenegative values of Xrs [45] creates an area between the reactance zero axis and Xrs,providing an integrative function to include changes in the magnitude of low-frequencyreactance Xrs, changes in fres and changes in the curvature of the Xrs(f)-tracing. It isrepresented graphically in figure 1 as the area under the zero axis of the reactance graphabove the Xrs(f) tracing.

AX~

ð

fres

5

Xrs(f ):df ð6Þ

This integrative index provides a single quantity that reflects changes in the degree ofperipheral airway obstruction and correlates closely with frequency dependence ofresistance [18].

OSCILLOMETRY: FOT AND IOS

81

Page 11: Forced oscillation technique and impulse oscillometry

Inertive spectral range of reactance. During resting breathing at normal respiratoryfrequencies, transrespiratory pressure is dissipated in resistive and elastic losses, whereasinertive pressure losses are negligible [46]. In contrast, during forced oscillation, whenoscillatory frequencies are more than 10-times greater than normal respiratory frequency,inertance (I) contributes significantly to dissipation of the externally applied pressures [16,46]. As noted above, the inertive spectral range of Xrs is at frequencies above fres. Thesefrequencies reflect mechanical properties of the proximal conducting airways. However,specific clinical interpretation of Xrs in this range is limited due to the wide variety ofinfluences that may appear in the upper respiratory tract and resultant resonant effectsthat may change Xrs unpredictably. Changes in central airway calibre correlate morestrongly with resistive parameters and are less well represented in the inertive spectralrange.

Finally, it should be noted that oscillometric reactance occasionally demonstrates alow-to-mid frequency plateau in the Xrs (f) tracing, which is suggestive of possible upperairway obstruction [47–50].

Variability of oscillometric parameters

The majority of oscillometric parameters can usefully be assessed using the coefficientof variation (CV%). Short-term variability should not exceed 10%, for magnitude of Zrs

and Rrs at frequencies i5 Hz [17]. Variability of Xrs is larger, because of physiologicaland numerical characteristics. Xrs can be positive or negative and is commonly close tozero. As a result, the calculation of CV% is not suitable to estimate the variability of mostof the Xrs values. Therefore, it is recommended to estimate variability of Rrs and Xrs

using standard deviation, 95% percentiles for normally distributed values or calculatingthe absolute difference (range) between minimum and maximum of the Xrs parameters[20, 51].

Methodological versus biological variability. Low methodological variability inmeasures of impedance (Zrs) has been shown using physical models with rather highreproducibility [20].

Biological variability is much more complex and incorporates intra-breath and intra-and inter-subject variability. Intra-subject variability of consecutive measurementswithin a specified time period, including circadian variability, day-to-day variability andvariability associated with changes in diseased airways has been reported previously [52–56].

Intra-breath variability is of specific physiological interest. Commonly, Zrs isdetermined as an average over a number of consecutive breathing cycles within 15–30 s. However, flow- and volume-dependence of Rrs and Xrs within each breathing cyclemay be apparent during both inspiration and expiration, and have been shown in anumber of investigations to reflect specific pathophysiological characteristics [57–61].

Special application of impulse oscillometry system to animals

Application of the impulse oscillation technique has been described in different animalspecies. The standard IOS device, originally developed for humans, has been validatedcarefully in calves aged up to 6 months and weighing 35–150 kg [20, 27, 62] and in pigs[63, 64]. For large animals, such as horses or adult cattle, a specially designed IOS unithas been developed, which is capable of analysing larger flow and volume characteristics.While no data are available in adult bovines, methodological validation and clinicalapplication of IOS in horses has been reported [65–67] and is still in progress.

H.J. SMITH ET AL.

82

Page 12: Forced oscillation technique and impulse oscillometry

Oscillometry in spontaneously breathing conscious animals requires the imposition ofapplied pressure signals via a rigid mask. Without correcting measured impedance for thefacemask, the limit of frequencies that can be clinically analysed is v15 Hz. Higherfrequencies are substantially influenced by the capacitance of the facemask itself [27].

The most useful frequency range for clinical evaluation of respiratory impedancediffers in different species, primarily dependent on animal size. The lower the specificfrequency is, the more sensitive the measurements to the periphery of the respiratorysystem are. For example, while the resonant frequency is between 5–12 Hz in calves,depending on body weight, it isv5 Hz in horses. Accordingly, frequencies that reflect thelung periphery in horses are lower than in calves.

In agreement with results in human medicine, peripheral airway obstruction ischaracterised by a marked increase in magnitude of low frequency respiratory reactance(|Xrs|) and in fres. In addition, the resistance spectrum of Rrs shows increased negativefrequency dependence and increase in low frequency Rrs (v5 Hz). Upper airwaynarrowing is characterised by a parallel increase in Rrs at all frequencies with no changein the frequency dependence of Rrs. No significant changes occur in reactance with upper(large) airway narrowing.

Relevance of oscillometry in clinical practice

This section offers a perspective on clinical use of the FOT method of determining Zrs

that may be summarised as follows:

1) FOT provides useful clinical information that prominently includes functionalassessment of small, peripheral airway behaviour beyond that available fromcommonly used pulmonary function tests (PFT).

2) Because of its sensitivity to peripheral airway function, FOT in its own right, apartfrom other PFT results, provides useful guidance in clinical patient management.

3) The prominence of peripheral airway functional assessment provided by FOTderives both from Xrs as well as Rrs.

4) The importance of Xrs is amplified by recognition of different Xrs-characteristics atlow, i.e. below resonant frequency (vfres), and high (wfres) oscillation frequencies.The last issue is considered in detail in the foregoing technological sections.

Briefly, it is noted here that original technical descriptions of FOT included calculationof the magnitude (|Zrs|) and phase (Q) of Zrs, i.e. polar coordinates [1, 5]. This engineeringdescription gave way to clinical research descriptions of Rrs and Xrs in Cartesiancoordinates. As noted in the technology section, Xrs relates to peripheral airwayproperties at oscillation frequencies vfres, and to central conducting airways atfrequencies wfres. The use of the magnitude of Xrs (|Xrs|) rather than Xrs itself inalgebraic manipulation of reactance data is emphasised, because this increases sensitivityto changes in peripheral airway mechanical properties, as demonstrated later in thissection in reference to previously reported studies.

The section that follows includes discussion of monofrequency, pseudo-random noiseand pulse-shaped pressure oscillations, as these three methods are currently in commonuse. The general principles apply to all methods of FOT for the most part where issuesconcern specific use of multifrequency rather than monofrequency or IOS rather thanPRN this difference is stated explicitly. Previous published reviews have discussedtheoretical and modeling aspects of FOT [2, 45, 68]. The present discussion does notinclude an exhaustive review of clinical research investigations from the Barcelona group[69–73], Leuven [6, 29, 32, 36, 39, 74–78], London [7, 79–81], Paris [8, 82–86] and

OSCILLOMETRY: FOT AND IOS

83

Page 13: Forced oscillation technique and impulse oscillometry

Vandoeuvre-les-Nancy [25, 87–91], which have all helped to provide the essentialinfrastructure for clinical application of oscillometry. Instead, the authors focus onpublished studies in relation to current work that permit practical establishment ofoscillometry in the routine clinical pulmonary function laboratory. Furthermore,because the authors have worked more intimately with IOS, illustrative examples ofcurrent work with this technology are included.

The relative advantages of each type of FOT may be summarised by noting that thesimplest form for clinical practitioners is monofrequency sinusoidal pressure application[2, 68]. Measurement of Rrs with this method is applicable to patients with sleep apnoeaand those using continuous positive airway pressure or mechanical ventilation.

Multifrequency FOT provides further characterisation of respiratory mechanics,including variation of Rrs and Xrs with oscillation frequency. PRN FOT has been appliedto the description of patients with asthma, bronchitis, emphysema, diffuse interstitiallung disease and thoracic wall deformities, and to assess bronchial or therapeuticchallenge [6, 29, 32, 36, 74–78]. Multifrequency FOT using PRN imposes a more gentleforcing signal perturbation than IOS, and has not been noted to provokebronchoconstriction. IOS differs from PRN by utilising brief pressure pulses of 30–40 ms duration. These pulses result in respiratory pressure responses that may beperceived as a slightly unpleasant respiratory sensation in some subjects. The briefpressure pulses provide convenient time-trend analyses and within-breath changes of Rrs

and Xrs not available with PRN. IOS is most familiar to the current authors, andtherefore, this and the following section relate current clinical work with IOS topreviously published reports of both PRN and IOS.

The clinical relevance of FOT may be assessed, as with any test of physiologicalfunction, in terms of its utility in diagnosis. Two general approaches are considered: first,the use of FOT as part of an initial complete diagnostic evaluation, including spirometry,body plethysmography, and gas distribution and exchange measures; secondly, the use ofFOT as a means of monitoring response to treatment can include both bronchial andtherapeutic challenge.

Summary information is presented here concerning the utility of FOT in assessingseverity of lung disease, degree of airway reactivity, reversibility of airflow obstruction,and stratification of breathing mechanics between central and peripheral airways.

Current clinical relevance of FOT relates significantly to the broad range of patientsthat may conveniently be evaluated. In contrast to standard PFT requiring maximalcoordinated efforts, FOT requires only normal quiet breathing with the lips tightly closedto avoid airflow leak, and the wearing of a nose-clip. For this reason, children can beeasily studied, often as early as 3 yrs [8, 19, 92–94]. Similarly, elderly subjects, those withsevere airflow obstruction or those with neuromuscular disease who find maximal forcedrespiratory efforts difficult to perform are able to breathe normally for FOT testing [95–98]. Portability of commercial FOT instruments permits lung function testing at thebedside or, for occupational lung disease studies, at the place of work [18].

FOT places minimal performance demands upon the patient, often described aspassive cooperation. However, the operator must take considerable care with the testprocedure. Because of the freedom offered to patients by simply breathing normally,moment-to-moment changes in respiratory resistance may be anticipated. Accordingly, aminimum of three technically acceptable FOT tests of 20–30 s duration or longer shouldbe performed. The mouthpiece of the FOT instrument must be supported at a position toensure maintenance of a neutral relaxed head and neck posture, avoiding body posturesthat might affect Zrs. Children should be seated in an appropriately-sized chair tocomfortably support their legs and adults should avoid crossing their legs, which requiresabdominal muscle contraction that may lead to end-expiratory lung volumes belowrelaxation volume. Patients should be comfortably relaxed to maintain a constant body

H.J. SMITH ET AL.

84

Page 14: Forced oscillation technique and impulse oscillometry

position without muscular effort. In contrast, firm contraction of the facial muscles isnecessary to support airtight closure of the lips about the mouthpiece. It is also desirablethat FOT testing be performed in a quiet examination room, sufficiently far from othersundergoing spirometry so as to be undisturbed by vigorous operator instructions forspirometry. The operator must review each FOT test immediately to ensure adequaterecording time free of artefacts; a minimum of 20 s of consecutive artefact-free recordingtime is advisable. Lack of attention to these fundamental principles may result in highlyvariable FOT tests in an individual that are not clinically interpretable. This is not atroublesome issue in experienced clinical investigators’ laboratories but it is an importantconsideration in clinical PFT laboratories who have not previously used FOT.

An important perceived concern about the clinical relevance of FOT is the availabilityof a normal database from which to judge results in a particular patient. Publishedreports of normative FOT data in children and adults are available, but the number andsize of these studies represent a much smaller normal population than is available forspirometry [6, 32, 39, 74, 79, 80, 93]. Differences in techniques of FOT and more recentlyin mouthpiece design may allow some degree of uncertainty concerning ranges of normalexpected values for both Rrs and Xrs, in both adults and children. Nevertheless, thesimilarity of FOT data using monofrequency sine-wave oscillations, PRN or pulse-shaped multifrequency FOT over the past four decades supports the acceptance ofclinically useful guidelines at this time. As expected, Rrs and Xrs are dependent on bodysize, and recent data suggest the possibility of racial/ethnic differences [99].

It is suggested that concern over precise definition of "normality versus abnormality" inan individual should not preclude clinical implementation of FOT at this time, becausebronchial and therapeutic challenges are very helpful in assessing airway responsivity. Ifinitial baseline Rrs/Xrs data do not clearly identify abnormality relative to existingnormative data, retesting after b-agonist inhalation will immediately identify increasedairway responsivity, and sequential studies over time provide similarly useful guidelinesfor clinical patient management.

Diagnostic evaluations

The most obvious relationship with other PFT procedures concerns the widespread useof spirometry. At the outset, it must be emphasised that spirometry measures maximalforced respiratory efforts, while FOT measures quiet breathing. Accordingly, it is notappropriate to demand that FOT and spirometric parameters be closely correlated as amandatory requirement for FOT to be considered valid. For example, children withasthma most commonly manifest normal spirometry [100] with no spirometric responseto inhaled b-agonist [42], while they may manifest abnormal baseline FOT parametersthat are responsive to therapeutic challenge [101]. At the other end of the spectrum,patients with advanced chronic obstructive pulmonary disease (COPD) commonlymanifest marked dynamic airway compression during spirometry with little spirometricresponse to pharmacological treatment, but may often manifest significant FOTresponses [45]. For these reasons, use of spirometry to define severity of obstructive lungdisease or receiver-operating characteristic of true sensitivity and specificity ofoscillometry must no longer be considered the optimal standard.

In some patients, spirometry cannot provide optimal clinical information. Patientswith significant neuromuscular disease are unable to provide the motive force needed forclear interpretation of spirometric results [97, 98]. Patients with lung allografttransplantation have obvious thoracic wall limitations that preclude truly maximalrespiratory efforts until many months after surgery, during which time it is often notpossible to detect adverse events, such as infection or acute allograft rejection by

OSCILLOMETRY: FOT AND IOS

85

Page 15: Forced oscillation technique and impulse oscillometry

spirometry alone. It is common for forced vital capacity (FVC) and FEV1 to increaseover the first 18–26 weeks post-lung transplantation [102]. Despite this apparentimprovement in spirometric parameters, peripheral airway disease, if it occurs, will leadto substantial worsening of FOT parameters [103] or of gas distribution [104].

Spirometric determination of responses to bronchial or therapeutic challenge may belimited by the necessary deep inspiration immediately prior to maximal expiration,allowing for distinctly different FOT responses [105–108]. This disagreement betweenspirometric and FOT parameters during bronchial challenge may be readily documentedby IOS during quiet breathing immediately before and after a deep inspiration, whichcommonly reveals immediate but transient bronchodilation in asthmatic subjects duringbronchial challenge. Figure 5 illustrates a 70-s recording of tidal breathing and a deepinspiration with simultaneous display of calculated magnitude of impedance at 5 Hz(Zrs5) using impulse oscillometry.

Rrs5 in this patient had increased markedly from 0.3 kPa?s?L-1 at baseline to1.2 kPa?s?L-1 after cumulative inhalation of 0.25, 0.5 and 1.0 mg?mL-1 methacholine,when FEV1 had changed by only 260 mL from a baseline of 2.73 L. Zrs5 during normalbreathing and following a deep inspiration with relaxed expiration shows that, at theonset of the IOS test, Zrs5 is much increased, while after an inspiration to maximal lungvolume, there is a marked fall in Zrs5, which gradually "recovers" with time. Since deepinspiration to maximal lung volume must immediately precede the FEV1 measurement,patients like the one whose data are shown in figure 5 will manifest FEV1 responses tomethacholine that are altered by the immediate decrease in airway smooth muscle tonefollowing maximal inspiration [106, 107].

With these caveats it is suggested that FOT can provide useful supplementaryinformation to spirometry that may not be tightly correlated with spirometric results. Toput such FOT information into proper perspective, it is necessary to review commonlyobserved patterns of FOT results in lung disease (primarily airflow obstruction), and inresponse to bronchial and therapeutic challenge.

���

��

9��#���)

���

���

��

���

���

:��&'�

(()

��

� �� �� �� �� �� �� ������+���������

Fig. 5. – Tidal volume and magnitude of respiratory impedance at 5 Hz (Zrs5) plotted as a function of timeduring methacholine challenge. First 40 s are resting breathing. After 40 s, subject inspired to total lungcapacity, followed by relaxation back to normal resting breathing. Note that Zrs5 increases markedly duringeach exhalation. After the deep inspiration, Zrs5 decreases transiently, with gradual return towards initial levelsover the following 24 s. The moving average of Zrs5 is shown by the dash-dot line.

H.J. SMITH ET AL.

86

Page 16: Forced oscillation technique and impulse oscillometry

Oscillometry in relationship to other diagnostic pulmonary functional tests

An important body of work has related FOT to body plethysmography [29, 76–78]. Agroup led by Van Noord have reported high correlations between Raw and FOTparameters in patients with obstructive lung disease and between absolute total lungcapacity (TLC) and FOT parameters in patients with diffuse interstitial lung disease. Infurther studies comparing plethysmography, spirometry and FOT to assess reversibilityof airflow obstruction, Van Noord’s group reported the distinctly lower sensitivity ofFOT than plethysmography [77]. The importance of this and other work by the VanNoord group is discussed further in the following sections.

It is widely recognised that body plethysmographic resistance, Raw includes only theresistance of the extrathoracic and intrathoracic airways, while Rrs includes that of thechest wall and lung tissue in addition to airway resistance. Resistance of the chest wallhas been reported [75], but there has been limited clinical interest in this parameterbecause of the technical difficulty of the measurements. Another difference between Raw

and Rrs relates to the status of the glottic aperature: it is commonly assumed that theglottis is maximally open during panting, but Jackson et al. [109] have shown that thisoccurs only in totally unrestricted panting. During voluntary attempts to control pantingfrequency and tidal volume, there is significant adduction of the vocal cords. Similarly,during quiet breathing, normal subjects commonly manifest a small, but variable, degreeof vocal cord adduction during expiration. In patients with obstructive lung disease, thisphasic expiratory adduction, visualised during the course of bronchscopy, does notappear to be systematically different from that observed in normal subjects. Thus, it is tobe expected that average Rrs will differ systematically from panting plethysmographicRaw.

It is also well established that Raw is more prominently influenced by large airway thanby small airway resistance. Thus, Smith and Dubois [110] reported a comparable increasein deadspace when compared to the decrease in Raw in response to scopolamine innormal subjects. In addition, Hensley et al. [111] reported similar changes in Raw anddeadspace after inhaled atropine. These results are consistent with the idea that Raw isprimarily influenced by large airways. In contrast, Rrs is importantly influenced by smallairway resistance, and, accordingly, it may be expected that FOT responses tointerventions that improve peripheral airway obstruction will be more prominent thanRaw responses. Because of the prominent effect of peripheral airway obstruction on FOTmeasurements, it may be expected that FOT indices of peripheral airway obstructionmight correlate more closely with indices of gas distribution ("Closing volume" [104]) andareas of lung hyperinflation manifested by computerised tomography [112] than withplethysmographic or spirometric indices, although there are no published comparisons atthis time.

Oscillometry as a clinical monitor of response to treatment

By way of summarising the clinical relevance of FOT, it is worth considering thespecial value of FOT as a means of monitoring response to interventions. FOT has beenreported to show greater sensitivity to inhaled corticosteroid or to b-agonist inhalation[8, 113–115] than spirometry. Both inhaled corticosteroids and b-agonists improve smallairways function, and FOT responses manifest prominent changes in indices ofperipheral airway obstruction. In contrast, spirometric sensitivity to small airwaysfunction is less prominent. Accordingly, it is expected that FOT might provide usefulindices of peripheral airway change in response to therapeutic interventions. Such use ofFOT provides a clinically valuable monitoring tool to follow therapeutic changes in small

OSCILLOMETRY: FOT AND IOS

87

Page 17: Forced oscillation technique and impulse oscillometry

airways function over time. This use of FOT for therapeutic monitoring is not dependenton the use of FOT as an initial diagnostic evaluation.

Finally, as a matter of practical convenience, FOT is more readily utilised in theclinical pulmonary function laboratory than body plethysmography. This issue isrelevant to recent interest in the therapeutic value of anticholinergics in patients withCOPD. As noted above, anticholinergics result primarily in large airway bronchodila-tion, and changes in Raw and deadspace are considerable [110, 111]. Thus, effectiveairway cholinergic blockade decreases large airway bronchomotor tone and increasesdeadspace, with relatively little effect on small peripheral airways disease. Whereas bodyplethysmography may be considered a useful technique to monitor such treatmenteffects, it is substantially less convenient to use routinely than FOT.

Oscillometry in the clinical pulmonary laboratory

Clinical interpretations of FOT responses in patients have often been related directlyto the application of a particular mechanical or electrical model of the respiratorysystem. Van Noord et al. [78] discussed their results in diffuse interstitial lung diseasewith respect to an electrical analogue of the respiratory system. They further confirmedearlier work (vide infra) that ascribed negative frequency dependence of resistance toperipheral airway obstruction [116, 117]. Engineering models of the respiratory systemhave provided predictions of FOT characteristics in normal human subjects, and changesin FOT parameters in lung disease. However, the fact that many of these predictions havebeen observed empirically does not constitute proof of validity of one or otherengineering models. Rather, it provides evidence that under the particular conditions ofthe FOT measurements undertaken, the empirical results show patterns that would beintuitively expected in normal subjects and those with lung disease. Over the past 3decades, a body of empirical evidence has accumulated that relates FOT results toparticular lung diseases, indeed establishing patterns that are characteristic of lungdisease. The following sections draw heavily upon this clinical research and codify FOTresults with respect to obstructive lung disease, with very limited data in diffuseinterstitial lung disease. These FOT data are not intended as validations of engineeringmodels, but instead, to illustrate commonly observed patterns of FOT characteristicsassociated with lung disease.

Obstructive lung disease

The relationships of FOT to spirometry noted above have a common theme.Spirometry does not provide a clear indication of peripheral airway obstruction, despitethe general information contained within the shape of the flow–volume curve, and valuesof mid-flow rates (forced expiratory flow between 25 and 75% of the forced vitalcapacity). Thus, the most striking characteristic of FOT in relation to spirometry is therelatively greater sensitivity of FOT to peripheral airway disease [2, 18, 25, 29, 32, 42, 45,52, 68, 82].

Peripheral airway disease. The most well-known FOT result empirically observed inperipheral airway disease in adults is frequency dependence of resistance (fdr). Grimby

et al. [116], using multiple replicates of monofrequency FOT, were the first to demonstratethe pattern of frequency dependence, wherein calculated Rrs was greater at 3 Hz than at 5,7 or 9 Hz in patients with chronic airflow obstruction (CAO). As calculated Rrs decreasedas oscillation frequency increased, patients with CAO might manifest nearly normal

H.J. SMITH ET AL.

88

Page 18: Forced oscillation technique and impulse oscillometry

values of Rrs at sufficiently high frequencies. For this reason, Grimby et al. [116] focusedon low (3 Hz) monofrequency Rrs to avoid masking differences between patients withairflow obstruction and normal subjects [118]. Many subsequent reports [18, 32, 36, 74,82, 117, 119] confirmed that subjects with early peripheral airways disease, includingsmokers, certain industrial workers and normal subjects after histamine infusion,manifested frequency dependence, even with normal values of low-frequency Rrs insmokers. Importantly, the abnormal frequency dependence of resistance occurred in thepresence of normal spirometry in those subjects with early peripheral airways disease [18,32, 117]. This body of clinical research is largely empirical, although it had been shown onautopsy many years earlier that cigarette smokers who died early in life had manifestedperipheral airway inflammation on autopsy [120]. Similarly, there is now ample evidenceof peripheral airway inflammation in patients with asthma, and, as will be illustratedbelow, frequency dependence of resistance occurs prominently in asthma.

The sensitivity of frequency dependence to peripheral airway disease is the firstdiscriminant between methods of FOT in general: while monofrequency FOT isconvenient to dissect within-breath patterns of change in Rrs [61, 68] or changes in Rrs

during sleep-disordered breathing or in patients on mechanical ventilators [2],multifrequency FOT is most convenient to document frequency dependence of resistancein practical use in the clinical pulmonary function laboratory. Monofrequency FOT maybe used at two or more single frequencies; however, multifrequency FOT uses differentoscillation frequencies applied within a single burst to dissect patterns consistent withperipheral rather than central airway obstruction. This dissection is based uponestablished observations that pressure oscillations at frequencies w15 Hz are severelydamped out before reaching peripheral airways, while those at frequencies v10–15 Hzpenetrate much further to the lung periphery [25, 121].

The transition between "large central" airways and "small peripheral" airways isneither precisely fixed anatomically nor precisely defined in terms of airway lumendiameter. The illustrations in figures 6 and 7 reflect common patterns observed inchildren with asthma and in adults with COPD.

Figure 6 shows representative IOS tests pre- and post-salbutamol in a 6-yr-old patient

���

��

���

���

���

���� � �� �� �� �� �� ��

!�"#��� ���$%

���

����

���

��&'�

(()

��

Fig. 6. – Conventional plots of respiratory resistance (Rrs(f)), as a function of oscillation frequency, in a 6-yr-oldchild with asthma. Note that frequency axis is shown between zero and 35 Hz, while data are plotted between3–35 Hz. A vertical dotted line is shown at 5 Hz, the lower limit at which most impulse oscillometry systemdata are reported. ––: data prior to nebulised b-agonist bronchodilator; ----: data after bronchodilator.

OSCILLOMETRY: FOT AND IOS

89

Page 19: Forced oscillation technique and impulse oscillometry

with mild asthma. Rrs is plotted as a function of oscillation frequency (Rrs(f)-tracing).Note that prior to b-agonist, Rrs5 is 0.93 kPa?s?L-1. Rrs falls steeply with increasingoscillation frequency to a minimum at 18 Hz, after which it increases with furtherincreases in oscillation frequency. While Clement et al. [39] have shown that normalchildren manifest a mild degree of frequency dependence, the very large fall in Rrs

between 5 and 15 Hz in figure 6 is consistent with abnormal peripheral airways functionin a 6 yr old. This is confirmed by administration of nebulised b-agonist, after which Rrs5

decreased to 0.59 kPa?s?L-1 (37% change). Note also that the fall of Rrs between 5–15 Hzpost b-agonist is much less than pre b-agonist. Baseline and post b-agonist IOS data inthis child may be compared with data in normal children of this age, who manifestedv15% fall in Rrs from 5–15 Hz [122]. The response to b-agonist in figure 6 may also becompared with responses of normal nonatopic children who manifested an averagechange in Rrs5 after salbutamol of 19% [122]. Finally, it can be seen in figure 6 that Rrs(f)at frequencies w20 Hz decreased substantially after b-agonist. Adult patients withasthma show similar findings to those in figure 6. Rrs may be nearly independent ofoscillation frequency in adult asthmatics after beta agonist.

Figure 7 illustrates an adult patient with COPD, pre- and post-b-agonist. In contrastto asthma, Rrs(f) decreases continuously with oscillation frequency between 5–25 Hz atbaseline, and after b-agonist there is a decrease in Rrs(f) only at low frequencies,v12 Hz.

The findings in figures 6 and 7 are consistent with bronchodilation produced by b-agonist in both large and small airways in the asthmatic subjects, but only in smallairways in the patient with COPD. In some patients with COPD who also have reactiveairways, b-agonist results in bronchodilation of large airways as well.

Central airway obstruction. Because of the frequency-dependent distribution ofoscillatory pressures within the airway tree, FOT provides separate, although not entirelyindependent, indices of large and small airway responses. Thus, changes of resistance inthe larger airways are manifest by FOT as uniform changes in Rrs at all oscillationfrequencies, both low and high. An increase in resistance of the large airways was reportedin a recent study of rescue workers at the World Trade Center site in New York [18].Rescue workers with no history of cigarette smoking exposed to large-particle air

���

��

���

���

���

���� � �� �� �� �� �� ��

!�"#��� ���$%

���

����

���

��&'�

(()

��

Fig. 7. – Respiratory resistance (Rrs(f)) in an adult patient with chronic obstructive pre- and post-nebulised b-agonist bronchodilator. ––: pre-bronchodilator; ----: post-bronchodilator. Note that Rrs(f) is unchanged afterbronchodilator at frequencies w12 Hz.

H.J. SMITH ET AL.

90

Page 20: Forced oscillation technique and impulse oscillometry

pollution at the World Trade Center site manifested, at baseline, uniformly increased Rrs

at all oscillatory frequencies studied between 5–35 Hz, as shown in figure 8, pre- and post-nebulised b-agonist.

Ironworkers with a history of cigarette smoking showed greater baseline increases inlow-frequency Rrs5, and, accordingly, a frequency dependence of resistance that ischaracteristic of peripheral airway obstruction [7, 18, 29, 36, 82, 116, 117]. Responses tonebulised b-agonist showed uniform decreases in Rrs across all frequencies innonsmokers, while smokers manifested larger decreases in Rrs5 than in Rrs20 [18].

Nonresistive components of forced oscillation technique. A second discriminantbetween mono- and multifrequency FOT concerns reactive components of appliedpressure oscillations, reactance Xrs(f)-tracings, which appear prominently inmultifrequency FOT. Reactance can be assessed by computer-assisted analysesutilising FFT; however, monofrequency FOT has not been widely adapted toconveniently calculate Xrs.

Just as with the interpretation of Rrs, oscillation frequency provides a means ofexamining different regions of the airway tree using Xrs (vide infra): at low oscillationfrequencies, elastic elements in peripheral airways are the dominant reactant to appliedpressure, and reflect small airway mechanical properties. In airflow obstruction, smallairways are functionally obstructed, due to peripheral airway inflammation in bothasthma and COPD. This results in portions of the distal lung periphery that are "in theshadow" of effective obstruction of small airways. This pathological process leads to anincrease in the magnitude of Xrs at low frequencies. At high frequencies, accelerativeforces are the predominant reactant to applied pressures and occur virtually exclusivelyin large airways where they are related to inertial properties. It should be noted that Van

Noord et al. [78] reported qualitatively similar changes in Xrs(f) tracings in patients withdiffuse interstitial lung disease and obstructive lung disease. Accordingly, changes in low-frequency Xrs are not specific to obstructive lung disease, but rather reflect peripheralairway disease. Mechanical interpretations of changes in low-frequency Xrs in obstructiveand restrictive lung disease are considered in detail in the methodology section.

���

��

���

���

���

���� � �� �� �� �� �� ��

!�"#��� ���$%

���

����

���

��&'�

(()

��

Fig. 8. – Respiratory resistance (Rrs(f)), plotted as a function of oscillatory frequency pre- (––) and post (----)-bronchodilator in an ironworker exposed to large-particle air pollution at the World Trade Center site. Noteincreased Rrs at all frequencies, with no significant frequency dependence of resistance at baseline. Afternebulised b-agonist, Rrs(f) decreases in a parallel manner relative to baseline pre-bronchodilator.

OSCILLOMETRY: FOT AND IOS

91

Page 21: Forced oscillation technique and impulse oscillometry

Figure 9 illustrates IOS Xrs(f) tracings in the asthmatic child whose Rrs(f) tracings pre-and post-nebulised b-agonist are shown in figure 6. Figure 9 shows that after b-agonist,Xrs at 5 Hz (Xrs5) changes from -0.31 to -0.26 kPa?s?L-1 and the frequency at which Xrs iszero (resonant frequency = fres, vide infra) changes from 18 to 17 Hz (6%). Morestrikingly, the overall curvature of the Xrs(f) tracing changes from concave to the zero-Xaxis to being convex to the zero-X axis. This change in curvature is consistent with Xrs(f)tracings described by Clement et al. [36], who reported that patients with airflowobstruction manifested a loss of the downward concavity of Xrs(f) tracings that iscommonly seen in normal subjects. b-agonist produced only small changes in Xrs5 andfres in figure 9; however, the change in the overall Xrs(f) tracing curvature at frequenciesbelow fres was dramatic.

Previous investigations have emphasised that low-frequency Xrs and Rrs are mostsensitive to changes in peripheral airway function, and Xrs5 has been used as a primaryefficacy variable [113–115, 122, 123]. However, in small children, respiratory frequency iscommonlyw20–30 breaths?min-1, and, accordingly, the higher harmonics of fundamentalrespiratory frequency may encroach on the lowest FOT frequencies analysed [41, 68]. Asa result, Xrs5 manifests relatively greater measurement noise. Some IOS studies inchildren have reported failure of Xrs5 to manifest significant changes after inhaledcorticosteroids or b-agonists [42, 115, 124]. Marotta et al. [42] showed that the Xrs10

response to b-agonist, but not the Xrs5 response, manifested a significant differencebetween asthmatic and nonasthmatic atopic children. This was associated with lessvariability in Xrs10 responses compared with Xrs5.

The absolute value of Xrs (|Xrs|) changes differently as a function of oscillationfrequency below and above fres. At low frequencies of oscillation, below fres, |Xrs|decreases as oscillation frequency increases up to fres. At fres, |Xrs| is zero. As oscillationfrequency increases above fres, |Xrs| increases with further increases in oscillationfrequency.

Because of this difference in the relationship between magnitude of Xrs and oscillationfrequency below and above fres, a quantitative index of Xrs magnitude at frequenciesbelow fres was developed by integrating all negative values of Xrs [18, 45, 52]. This index,

����

����

����

����

����

���

���

���

���

���

����

����

���

�&'�

(()

��

��������������!�"#��� �� �$%

���-�

���-��

Fig. 9. – Respiratory reactance (Xrs(f)), plotted as a function of oscillatory frequency pre- (––) and post- (----)bronchodilator. Same subject as figure 6. The integrated low-frequency reactance area (AX), is shown by verticalhatching pre-bronchodilator, and by diagonal hatching post-bronchodilator. This area is reduced by y50% frompre- to post-bronchodilator, associated with marked change in curvature of the Xrs(f) tracing. In comparison,there are small changes in Xrs at 5 Hz and resonant frequency from pre- to post-bronchodilator.

H.J. SMITH ET AL.

92

Page 22: Forced oscillation technique and impulse oscillometry

designated AX, provides an integrative function to include changes in the magnitude oflow-frequency Xrs, changes in fres and changes in curvature of the Xrs(f) tracing. AXincludes Xrs magnitudes at 5 Hz and slightly higher oscillation frequencies whichmanifest improved signal-to-noise ratio, as noted above for atopic asthmatic children[42]. It is represented graphically in figure 9 as the area under the zero Xrs axis above theXrs(f) tracing. As discussed below, this integrative index provides a single quantity thatreflects changes in the degree of peripheral airway obstruction and correlates closely withfrequency dependence of resistance. In figure 9, AX decreases by 50% after b-agonist,closely comparable to the decrease in frequency dependence of resistance measuredbetween 5 and 15 Hz in the same child shown in figure 6, when R5-R15 decreased from0.36 kPa?s?L-1 to 0.16 kPa?s?L-1 after b-agonist.

The perspective presented here of AX in relation to peripheral airway function, resultsdirectly from the details presented in the methodology section, namely that low-frequency Xrs essentially expresses the ability of the respiratory tract to store capacitiveenergy, which is primarily resident in the lung periphery. In contrast, at frequenciesabove fres, I, which is primarily resident in proximal conducting airways, contributessignificantly to the dissipation of externally applied pressures [6, 46]. Thus, oscillationfrequencies above fres reflect mechanical properties of more proximal conductingairways. Because of these differences in mechanical properties reflected by low- and high-frequency oscillation, calculation of the arithmetic mean Xrs is not likely to be optimallyuseful to assess pulmonary mechanical responses. Thus, Van Noord et al. [29] reportedthat the mean value of Xrs change was less sensitive than FEV1 in assessing the effect ofhistamine. However, their graphic mean Xrs–frequency data reveal changes in theestimate of AX from y0.15 at baseline, to 1.7 or 2.5 kPa?L-1 when mean decrease inspecific airway conductance (sGaw) was 40% or 15% in FEV1. These increases in AX of1,000–1,500% are comparable to those measured using IOS in the current author’slaboratory during methacholine challenge: the patient shown in figure 6 manifested anincrease in AX from 0.34 at baseline to 7.5 kPa?L-1 (w2,000% increase) after cumulativeexposure to methacholine up to 1.0 mg?mL-1. Furthermore, the study of Van Noord

et al. [77] during assessment of reversibility of airflow obstruction by FOT, bodyplethysmography and spirometry reported that changes in mean Xrs did not contributesignificantly to discriminant function beyond spirometry, Raw and Rrs at 6 Hz. Incontrast, estimated AX, approximated from their graphic mean Xrs–frequency data,showed a 50% reduction after salbutamol, fromy3.1 to 1.5 kPa?L-1. Both the baselineAX in patients with airflow obstruction and the AX decreases in response to b-agonist arecomparable to those recently reported using IOS [101].

The empirical observations discussed above do not prove that any particularengineering model is a true representation of the lung, especially the diseased lung asemphasised by Van Noord et al. [78]. Nevertheless, FOT results predicted by modelsmay correlate usefully with independent clinical physiological and pathophysiologicalevidence. Quite apart from engineering models, an intuitive understanding of Xrs may beappreciated from the physical principles elucidated above and amplified in the foregoingdiscussion of methodology. The applicability of high frequencies to large central airwaysand low frequencies to peripheral airways is not a consequence of any particularengineering model, but is observed empirically both in Rrs and Xrs.

Clinical interpretation of forced oscillation technique

As also noted in the methodology discussion of FOT, abnormalities of Xrs are notspecific to obstructive lung disease, because these same patterns have been reported inlung fibrosis [78]. In clinical diagnostic lung function testing including FOT, spirometry,

OSCILLOMETRY: FOT AND IOS

93

Page 23: Forced oscillation technique and impulse oscillometry

gas diffusion and body plethysmography, the problem is not usually distinguishingbetween obstructive and restrictive disease. The more important issue is the relativeseverity of pathophysiological abnormality. Furthermore, there is evidence to suggestthat the early pathological lesion in lung fibrosis is inflammation of the small airways[125]. Thus, changes in Xrs magnitude in lung fibrosis and in peripheral airflowobstruction may both reflect peripheral airway inflammation. If these nonspecific Xrs

abnormalities in lung fibrosis represent small airway inflammation, they may respond toanti-inflammatory treatment, analogous to the way asthmatic peripheral airwayinflammation responds to corticosteroids. Such responses are more likely to be foundin FOT parameters than in spirometry. Clinical interpretation may then be considered inthe setting of response to treatment interventions.

Clinical interpretation of FOT can be related both to effects of airway smooth muscletone and airway inflammation. Below, effects of anticholinergic, b-agonist orcorticosteroid medications are represented, because these agents are most commonlyutilised clinically. Commonly observed changes in FOT measures in patients withobstructive lung disease are illustrated. Rrs is considered first.

Rrs effects. While inflammation is a cellular process, it has mechanical consequences.These consequences may be considered in relation to bronchoconstriction, defined asincreased tone of airway smooth muscles, and the common perception of bronchodilationdefined as a decrease in smooth muscle tone.

When airway smooth muscle tone increases, Rrs increases because of decreased airwaylumen. Airway lumen is also decreased with inflammation or oedema in the walls of theairways. Therefore, Rrs increases as a result of inflammation and oedema. Peripheralairways have much smaller lumina than central (large) airways, and inflammation/oedema in the walls of peripheral airways can reasonably be expected to have aproportionately larger effect on lumen size than inflammation/oedema in larger airways.In the discussion that follows, "low-frequency Rrs" will be denoted as Rrs at frequenciesv15 Hz and "high-frequency Rrs" as Rrs at frequencies w20 Hz, with the latter termsynonymous with large airway resistance.

If an intervention, such as inhaled anticholinergic, achieves bronchodilation with noeffect on inflammation, it may be expected that large airway lumen will increase, withlittle effect on peripheral airway lumen. In this event, large airway Rrs will decrease. Low-frequency Rrs will decrease to a similar degree and little or no change in frequencydependence occurs. This is illustrated in a 55-yr-old patient with COPD in figure 10.

If an intervention such as inhaled b-agonist achieves bronchodilation with little or noeffect on inflammation, it may be expected that peripheral airway lumina will increase, inaddition to any release of bronchoconstriction in large airways. In this case, low-frequency Rrs will decrease out of proportion to high-frequency Rrs.

In asthmatic patients, both high-frequency and low-frequency Rrs may decrease, withrelatively greater decrease in low-frequency Rrs and associated decrease in frequencydependence of resistance as illustrated in figure 6.

In patients with COPD, a decrease in low-frequency Rrs after b-agonist with little orno change in high-frequency Rrs is commonly observed, as illustrated in figure 7. In COPDpatients with lung hyperinflation, little or no decrease in Rrs may occur afterb-agonist inhalation, particularly if there is an associated fall in resting end-expiratorylung volume. If Rrs remains the same after b-agonist while end-expiratory lung volumedecreases, this represents "functional bronchodilation", because the same resistancepertains at lower operating lung volumes. Accordingly, failure of Rrs to decrease in patientswith COPD need not be considered as "no response" to b-agonist bronchodilator.

If an intervention such as inhaled corticosteroids achieves a decrease in inflammation

H.J. SMITH ET AL.

94

Page 24: Forced oscillation technique and impulse oscillometry

with no effect on airway smooth muscle tone, FOT responses might be expected to reflecta relatively greater impact due to decrease in peripheral airway inflammation withresultant increase in peripheral airway lumina. Such an effect results in a significant"dilation" of peripheral airways due to decreased inflammatory encroachment onperipheral airway lumina. Thus, a decrease in peripheral airway resistance can beexpected, manifest as a greater decrease in low-frequency than in high-frequency Rrs, andconcomitant decrease in frequency dependence.

In asthmatic patients, a decrease in large airway Rrs (high-frequency Rrs) may alsooccur. In patients with COPD, there may be a decrease in low-frequency Rrs; however,little or no decrease in Rrs may be manifest, especially if lung hyperinflation is present.

Xrs effects. How then does decreased inflammation manifest itself in patients withCOPD? As noted in the preceding section, describing nonresistive components of FOT,the magnitude of low-frequency Xrs is increased in COPD due to functional peripheralairway obstruction, with resultant contraction of surface area of the lung peripheryexposed to low-frequency oscillations. Indeed, low-frequency Xrs is more sensitive toperipheral airway obstruction in COPD/emphysema than Rrs. In the presence ofperipheral airway obstruction in patients with COPD, relatively small increments toairways resistance may occur because of the large cumulative cross-sectional diameter ofall airways in the lower generations of airways, as manifest in the trumpet model ofWeibel [126]. Accordingly, body plethysmographic measurement of airway resistancemay be nearly normal, and only by measuring absolute thoracic gas volume is theabnormality manifest.

If peripheral airway inflammation is decreased by administration of inhaledcorticosteroids, peripheral airway lumina increase and the patency of small airwaysexpands in the direction of the lung periphery. As a result, a portion of the lung peripherycomes "out of the shadow" of small airway obstruction and a larger surface area ispresented to the low-frequency oscillations. This acts to decrease the magnitude of low-frequency Xrs. This is illustrated in figure 11a showing IOS tracings in a COPD patient at

���

���

���

��

���

���� � �� �� �� �� �� ��

!�"#��� ���$%

���

����

���

��&'�

(()

��

Fig. 10. – Respiratory resistance (Rrs), plotted as a function of oscillation frequency in a 55-yr-old male withchronic obstructive pulmonary disease pre- (––) and post- (----) anticholinergic bronchodilator. Note that Rrs at5 Hz is markedly elevated with marked frequency dependence of Rrs. After inhaled anticholinergic broncho-dilator, there is a significant decrease in Rrs that is nearly identical at all frequencies, indicating a decrease inproximal airway resistance, with little or no effect on peripheral airways

OSCILLOMETRY: FOT AND IOS

95

Page 25: Forced oscillation technique and impulse oscillometry

baseline (–––) and after 4 weeks of inhaled corticosteroids treatment (-----). AX decreasedfrom 3.0 to 1.3 kPa?L-1 after inhaled corticosteroid treatment. In this patient, Rrs andfrequency dependence of the Rrs(f) tracing show significant decreases, as illustrated infigure 11b; Rrs5 improved from 0.68 to 0.48 kPa?s?L-1, Rrs15 from 0.41 to 0.35 kPa?s?L-1.Furthermore, this patient with COPD manifests somewhat "responding" large airways,as his high-frequency Rrs (at 30–35 Hz) also decreased byy20%.

Figure 12 shows tracings in a COPD patient when stable at baseline and during theonset of an exacerbation. Figure 12a shows baseline reactance area AX = 1.2 kPa?L-1,with increases to 1.6, 2.4 and 4.3 kPa?L-1 over a duration of 9 days of exacerbation. Rrs5

increased significantly, but less dramatically from 0.51 at baseline to 0.54, 0.59 and0.65 kPa?s?L-1 during exacerbation, shown in figure 12b. Frequency dependence,calculated as the fall in Rrs from 5 to 15 Hz, was 0.15 kPa?s?L-1 when stable at baseline,and increased to 0.2, 0.23 and 0.31 kPa?s?L-1 during exacerbation.

����

����

����

����

����

���

���

���

���

���

���

����

����

���

�&'�

(()

��

���-�

���-����

�����

���

�&'�

(()

��

��������������!�"#��� �� �$%

��

7�

���

���

���

���

��

���

Fig. 11. – a) Respiratory reactance (Xrs), plotted as a function of oscillatory frequency in a 73-yr-old male withchronic obstructive pulmonary disease before and after 4 weeks inhaled corticosteroid (ICS) therapy. ––: pre-ICS; ----: post-ICS. Integrated low-frequency reactance area (AX) is vertically hatched pre-ICS and diagonallyhatched post-ICS. AX is decreased by y50% after 4 weeks of therapy, resonant frequency by 10% and Xrs at5 Hz decreased by 0.12 kPa?s?L-1. b) Respiratory resistance (Rrs) plotted as a function of oscillation frequency inthe same patient. ––: pre-ICS; ----: post-ICS. Note that Rrs at 20–22 Hz is relatively unchanged, while Rrs at 5and 10 Hz are substantially decreased. High-frequency resistance (Rrs at 25–35 Hz) is decreased after ICStherapy. See text for discussion.

H.J. SMITH ET AL.

96

Page 26: Forced oscillation technique and impulse oscillometry

The close correlation between changes in AX and changes in frequency dependence ofresistance in individual patients shown in figures 6–12 are confirmed across individuals inthe occupational study reported by Skloot et al. [18], who found a correlation of 0.92between frequency dependence of resistance and AX across a sample of ironworkers atthe World Trade Center site, with variable exposure to cigarette smoking and large-particle air pollution, and resultant variability in both large and small airwayobstruction. The close correlation between frequency dependence of resistance andAX is consistent with both indices reflecting small airway function.

Coherence. Coherence, first introduced by Michaelson et al. [5], is defined as the auto-and cross-correlations of phase and amplitude of oscillatory pressure and flowcomponents. It reflects, in an "engineering sense", the linearity of the respiratorysystem and, in a "biological sense", the variability of the respiratory system from time totime within the sample of data. Marked temporal variability of the respiratory systemwithin a breath occurs commonly in COPD, where even during quiet breathing, dynamic

����

����

����

����

����

���

���

���

���

���

���

����

����

���

�&'�

(()

��

��

���

����

���

�&'�

(()

��

7� ���

��

���

���

���

���� � �� �� �� �� �� ��

!�"#��� �� �$%

Fig. 12. – a) Respiratory reactance (Xrs(f)) in a 73-yr-old male chronic obstructive pulmonary disease patientwhen stable at baseline and during onset of exacerbation over a duration of 9 days. Note progressive increase inthe reactance area (AX) from trial 1 to trial 4. b) Respiratory resistance (Rrs(f)) in the same patient. Noteprogressive increase in Rrs at 5 Hz (Rrs5) from trial 1 to trial 4. Changes in Rrs5 were relatively smaller thanchanges in AX over a duration of 8 days. –––: trial 1, baseline; -----: trial 2, 7 days after trial 1; –– - ––: trial 3,8 days after trial 1; –– -- ––: trial 4, 9 days after trial 1.

OSCILLOMETRY: FOT AND IOS

97

Page 27: Forced oscillation technique and impulse oscillometry

compression of intrathoracic airways may occur. As a result, Rrs and Xrs may increasesubstantially during expiration.

Figure 13a illustrates a patient with severe COPD, with volume and Zrs5 as functionsof time. Figure 13b shows coherence plotted as a function of oscillation frequency, withseparate tracings for average combined inspiration/expiration as well as for inspiration-only and expiration-only.

Figure 13a shows marked changes in Zrs5 with respiratory phase, similar to thoseshown by Marchal and Loos [68]. There is a clear decrease in Zrs5 at the onset ofinspiration, keeping minimal values until end-inspiration. A marked abrupt rise in Zrs5occurs at the onset of expiration with elevated values including end-expiration. In figure13b, the value of averaged coherence at 5 Hz, 0.4, is distinctly lower than at frequenciesi10 Hz.

��

���

9��#���)

����

� �� �� ��0����

���

���

���

���&'�

(()

��

���

���

��

���

���

���

���

���

���

������

����

���

����

7�

� �������������!�"#��� �� �$%

��

Fig. 13. – a) Volume and magnitude of respiratory impedance at 5 Hz (Zrs5) plotted as a function of timeduring a 40-s impulse oscillometry system (IOS) test in a 53-yr-old patient with severe chronic obstructivepulmonary disease. Note marked increase in Zrs5 during the expiratory phase of every tidal volume. b)Coherence of all IOS data (global average (–––) and separated inspiratory (–– - ––) and expiratory (-----)respiratory phases) plotted as a function of oscillation frequency during the 40-s test. Note low coherence foraverage data (v0.5) at 5 Hz with prominent increase to 0.7 at 10 Hz. See text for discussion.

H.J. SMITH ET AL.

98

Page 28: Forced oscillation technique and impulse oscillometry

Numerical calculations of coherence during the separate respiratory phases reveal thatinspiratory-only and expiratory-only coherences are systematically greater than thecombined coherence over the entire spectrum of oscillatory frequencies. The tracings infigure 13 are consistent with more uniformity of respiratory mechanics within theseparate inspiratory and expiratory phases than for the combined total breath average.At 10 Hz, the coherence averaged across both respiratory phases is 0.7, while that forseparate inspiratory and expiratory data are both 0.9, reflecting differences in respiratorymechanical parameters that pertain to the separate respiratory phases. Despite the verylow coherence for combined inspiration/expiration, three consecutive 40-s impulseoscillometry system recordings obtained within 5 min manifested an average respiratoryresistance at 5 Hz of 1.08, 1.0 and 1.02 kPa?s?L-1 and reactance area of 5.9, 6.1 and6.2 kPa?L-1. Thus, the within-phase uniformity of mechanical parameters reflected byseparate inspiratory and expiratory coherences is borne out by standard deviations ofv5% for triplicate measures.

Summary

The aim of this chapter has been to describe the unique and clinically relevantinformation that forced oscillation technique (FOT) provides. This may be derivedwithout mathematical mastery of technological principles of the equipment and/or ofnumerical models. It is emphasised that recognition of the change in respiratorymechanical parameters as a function of oscillation frequency is necessary to appreciatethe outstanding value of FOT in its ability to assess peripheral airway function. Thishas been one of the major challenges in respiratory diagnostics up to the present time.The short duration of the FOT test, 20–30 s, makes it particularly useful as part of adiagnostic programme of lung function testing; it is not suggested that FOT be used inlieu of conventional pulmonary function testing, but rather in addition. FOT measuresresting breathing while spirometry assesses maximal respiratory performance of thepatient. The special value of FOT in terms of short-term response to bronchial andtherapeutic challenge has been emphasised as well as its value in monitoring long-termtrend responses to therapy.The simplicity of FOT measurements and its minimal requirements on subjectcooperation are in rather sharp contrast to its current limited clinical acceptance. Twoprimary reasons for the present limited application of FOT include the need forviewing respiratory mechanical parameters over a range of frequencies and theresultant central-peripheral specificity of oscillatory parameters, with specificemphasis on the reflection of peripheral airway function by low-frequency reactance.Indeed, lack of awareness of this ability of FOT to assess peripheral airway functionhas turned physicians to the use of multiple replicates of high-resolution computedtomography lung scans to assess small airway function. Other reasons for limited useof FOT currently may include the greater variability of FOT measures compared withspirometry. Despite such variability, use of at least three replicate FOT measurescombined with therapeutic challenge can provide sensitive evaluation of small airwayfunction.The freedom allowed to the subject to breathe "naturally" imposes increased demandsfor vigilance on the operator, who must maintain a quiet environment for forcedoscillation technique testing. Operators must also reassure subjects that theirrelaxation is needed, except for the facial musculature ensuring tight lip closure onthe mouthpiece. Posture must be supported to maintain subject comfort and the

OSCILLOMETRY: FOT AND IOS

99

Page 29: Forced oscillation technique and impulse oscillometry

instrument mouthpiece must be brought to the subject to avoid stretching of the neck.Finally, the availability of results from a brief test must not lead the operator to accepta single measurement, but rather, the usual clinical testing procedure of at least threereplicate measures is required.

Keywords: Forced oscillation technique, impulse oscillation technique, reactance,resistance, respiratory impedance.

References

1. DuBois AB, Brody AW, Lewis DH, Burgess BF. Oscillation mechanics of lungs and chest in man.

J Appl Physiol 1956; 8: 587–594.

2. Navajas D, Farre R. Oscillation mechanics. Eur Respir Mon 1999; 4 (12): 112–140.

3. Bedenice D, Mazan M, Reinhold P, Harrington K, Hoffman A. Comparison of two different

oscillatory techniques for the non-invasive measurement of respiratory system resistance (Rrs) in

the horse. Eur Respir J 2003; 22: Suppl. 45, 688.

4. Lorino A, Zerah F, Mariette C, Harf A, Lorino H. Respiratory resistive impedance in obstructive

patients: linear regression analysis vs viscoelastic modelling. Eur Respir J 1997; 10: 150–155.

5. Michaelson E, Grassman E, Peters W. Pulmonary mechanics by spectral analysis of forced

random noise. J Clin Invest 1975; 56: 1210–1230.

6. Landser F, Nagels J, Demedts M, Billiet L, Van De Woestijne K. A new method to determine

frequency characteristics of the respiratory system. J Appl Physiol 1976; 41: 101–106.

7. Pride N. Forced oscillation techniques for measuring mechanical properties of the respiratory

system. Thorax 1992; 47: 317–320.

8. Delacourt C, Lorino H, Herve-Guillot M, Reinert P, Harf A, Housset B. Use of the forced

oscillation technique to assess airway obstruction and reversibility in children. Am J Respir Crit

Care Med 2000; 161: 730–736.

9. Cauberghs M, Van de Woestijne K. Effect of upper airway shunt and series properties on

respiratory impedance measurements. J Appl Physiol 1989; 66: 2274–2279.

10. Daroczy B, Hantos Z. Generation of optimum pseudorandom signals for respiratory impedance

measurements. Int J Biomed Comput 1990; 25: 21–31.

11. Solymar L, Landser F, Duiverman E. Measurement of resistance with the forced oscillation

technique. Eur Respir J 1989; 2: Suppl. 22, 150–153s.

12. MacLeod D, Birch M. Respiratory input impedance measurement: forced oscillation methods.

Med Biol Eng Comput 2001; 39: 505–516.

13. Vogel J, Smidt U. Impulse Oscillometry – Analysis of Lung Mechanics in General Practice and

the Clinic, Epidemiological and Experimental Research. Frankfurt, Pmi Verlagsgruppe GmbH,

1994.

14. Peslin R, Jardin P, Duvivier C, Begin P. In-phase rejection requirements for measuring respiratory

input impedance. J Appl Physiol 1984; 56: 804–809.

15. Van de Woestijne K, Desager K, Duiverman E, Marchal F. Recommendations for measurement

of respiratory input impedance by means of the forced oscillation method. Eur Respir Rev 1994;

4: 235–237.

16. Lorino H, Mariette C, Karouia M, Lorino A. Influence of signal processing on estimation of

respiratory impedance. J Appl Physiol 1993; 74: 215–223.

17. Oostveen E, MacLeod D, Lorino H, et al. The forced oscillation technique in clinical practice:

methodology, recommendations and future developments. ERS Task Force. Eur Respir J 2003;

22: 1026–1041.

18. Skloot G, Goldman M, Fischler D, et al. Respiratory symptoms and physiologic assessment of

ironworkers at the World Trade Center disaster site. Chest 2004; 125: 1248–1255.

H.J. SMITH ET AL.

100

Page 30: Forced oscillation technique and impulse oscillometry

19. Bisgaard H, Klug B. Lung function measurement in awake young children. Eur Respir J 1995;

8: 2067–2075.

20. Reinhold P, Smith HJ, Langenberg A, Lekeux P. Measurement of respiratory impedance in

healthy calves using the impulse oscillation technique - physiological and methodological aspects.

Vet J 1998; 155: 27–38.

21. Muller E. Nichtinvasive Differenzierung der mechanischen eigenschaften des respiratorischen

systems - strukturermittlung und parameterbestimmung mit hilfe systemtheoretischer verfahren

[Noninvasive differentiation of mechanical properties of the respiratory system – determination of

structures and parameters using methods of system theory]. Dissertation B, Ilmenau, Technische

Hochschule Ilmenau, 1989.

22. Oostveen E, Zwart A. Reliability of the coherence function for rejecting respiratory impedance

data. Eur Respir Rev 1991; 1: 218–221.

23. Miller T, Pimmel R. Standard errors on respiratory mechanical parameters obtained by forced

random excitation. IEEE Trans Biomed Eng 1983; 30: 826–832.

24. Van Noord JA, Van de Woestijne KP, Demedts M. Clinical applications and modelling of forced

oscillation mechanics of the respiratory system. Eur Respir J 1991; 4: 247–248.

25. Peslin R, Fredberg J. Oscillation mechanics of the respiratory system. In: Macklem P, Mead J, eds.

Handbook of Physiology, Section 3: The respiratory system. Vol III, Mechanics of breathing.

Bethesda, American Physiological Society, 1986; pp. 145–166.

26. Cauberghs M, Van de Woestijne KP. Changes of respiratory input impedance during breathing in

humans. J Appl Physiol 1992; 73: 2355–2362.

27. Reinhold P, Smith HJ, Close R, Genicot B, Lekeux P. Validation of impulse oscillometry in

Friesian and Blue Belgian calves with respect to changes in extrathoracic upper airway resistance.

Res Vet Sci 1998; 65: 93–101.

28. Ritz T, Dahme B, Dubois AB, et al. Guidelines for mechanical lung function measurements in

psychophysiology. Psychophysiology 2002; 39: 546–567.

29. Van Noord J, Clement J, Van de Woestijne K, Demedts M. Total respiratory resistance and

reactance as a measurement of response to bronchial challenge with histamine. Am Rev Respir Dis

1989; 139: 921–926.

30. Bouaziz N, Beyaert C, Gauthier R, Monin P, Peslin R, Marchal F. Respiratory system reactance

as an indicator of the intrathoracic airway response to methacholine in children. Pediatr Pulmonol

1996; 22: 7–13.

31. Duiverman EJ, Neijens HJ, Van der Snee-van Smaalen, Kerrebijn KF. Comparison of forced

oscillometry and forced expirations for measuring does-related responses to inhaled methacholine

in asthmatic children. Bull Eur Physiopathol Respir 1986; 22: 433–436.

32. Landser F, Clement J, Van De Woestijne K. Normal values of total respiratory resis-

tance and reactance determined by forced oscillations. Influence of smoking. Chest 1982;

81: 586–591.

33. Wouters EFM, Polko AH, Schouten HJA, Visser BF. Contribution of impedance measurement of

the respiratory system to bronchial challenge tests. J Asthma 1988; 25: 259–267.

34. Lorino AM, Zerah F, Mariette C, Harf A, Lorino H. Respiratory resistive impedance

in obstructive patients: linear regression analysis vs viscoelastic modelling. Eur Respir J 1997;

10: 150–155.

35. Daroczy B, Hantos Z. Generation of optimum pseudorandom signals for respiratory impedance

measurements. Int J Biomed Comput 1990; 25: 21–31.

36. Clement J, Landser F, Van de Woestijne K. Total resistance and reactance in patients with

respiratory complaints with and without airways obstruction. Chest 1983; 2: 215–220.

37. Pasker HG, Schepers R, Clement J, Van de Woestijne KP. Total respiratory impedance measured

by means of the forced oscillation technique in subjects with and without respiratory complaints.

Eur Respir J 1996; 9: 131–139.

38. Cuijpers CEJ, Wesseling G, Swaen GMH, Wouters EFM. Frequency dependence of oscillatory

resistance in healthy primary school children. Respiration 1993; 60: 149–154.

OSCILLOMETRY: FOT AND IOS

101

Page 31: Forced oscillation technique and impulse oscillometry

39. Clement J, Dumoulin B, Gubbelmans R, Hendriks S, Van de Woestijne K. Reference values of

total respiratory resistance and reactance between 4 and 26 Hz in children and adolescents aged 4–

20 years. Bull Eur Physiolpathol Respir 1987; 23: 441–448.

40. Solymar L, Aronsson PH, Bake B, Bjure J. Respiratory resistance and impedance magnitude in

healthy children aged 2-18 years. Pediatr Pulmonol 1985; 1: 134–140.

41. Landser FJ, Nagels J, Clement J, Van de Woestijne KP. Errors in the measurement of total

respiratory resistance and reactance by forced oscillations. Respir Physiol 1976; 28: 289–301.

42. Marotta A, Klinnert M, Price M, Larsen G, Liu A. Impulse oscillometry provides an effective

measure of lung dysfunction in 4-year-old children at risk for persistent asthma. J Allergy Clin

Immunol 2003; 112: 317–322.

43. Mead J. Contribution of compliance of airways to frequency-dependent behaviour of lungs. J Appl

Physiol 1969; 26: 670–673.

44. Govaerts E, Demedts M, Van de Woestijne KP. Total respiratory impedance and early

emphysema. Eur Respir J 1993; 6: 1181–1185.

45. Goldman M. Clinical application of forced oscillation. Pulm Pharm & Therap 2001; 14: 341–350.

46. Imberger H. Elementary complex number analysis of lung models. Respiration 1985; 47: 57–69.

47. Mobeireek A, Alhamad A, Al-Subaei A, Alzeer A. Psychogenic vocal cord dysfunction simulating

bronchial asthma. Eur Respir J 1995; 8: 1978–1981.

48. Bucca C, Rolla G, Brussino L, De Rose V, Bugiani M. Are asthma-like symptoms due to bronchial

or extrathoracic airway dysfunction? Lancet 1995; 346: 791–795.

49. Beraldo PSS, Mateus SRM, Araujo LM, Horan TA. Forced oscillation technique to detect and

monitor tracheal stenosis in a tetraplegic patient. Spinal Cord 2000; 38: 445–447.

50. Horan T, Mateus S, Bernaldo P, et al. Forced oscillation technique to evaluate tracheostenosis in

patients with neurologic injury. Chest 2001; 120: 69–73.

51. Uystepruyst C, Reinhold P, Coghe J, Bureau F, Lekeux P. Mechanics of the respiratory system in

healthy newborn calves using impulse oscillometry. Res Vet Science 2000; 68: 47–55.

52. Goldman M, Carter R, Klein R, Fritz G, Carter B, Pachucki P. Within- and between-

day variability of respiratory impedance using impulse oscillometry in adolescent asthmatics.

Ped Pulmonol 2002; 34: 312–319.

53. Barnes PJ. Circadian variation in airway function. Am J Med 1985; 79: Suppl. 6a, 5–9.

54. Bonnet R, Jorres R, Heitman U, Magnussen H. Circadian rhythm in airway responsiveness and

airway tone in patients with mild asthma. J Applied Physiol 1991; 71: 1598–1605.

55. Reinhold P, Uystepruyst C. The intra-subject variability of respiratory impedance - a new

parameter of clinical relevance. Eur Respir J 2000; 16: Suppl. 31, 134s.

56. Smolensky M, Barnes P, Reinberg A, McGovern J. Chronobiology and asthma. I. Day-night

differences in bronchial patency and dyspnea and circadian rhythm dependencies. J Asthma 1986;

23: 321–343.

57. Barnikol W. Perspektiven einer innovativen funktionsdiagnostik des bronchialsystems

[Prospectives of innovative function diagnostics of the bronchial system]. Munchen-

Deisenhofen, Dustri-Verlag, Dr. Karl Feistle, 1997; pp. 22–40.

58. Barnikol W, Diether K. Zwei neue kenngroßen zur charakterisierung des bronchialen status in

vivo und bronchial wirksamer substanzen. Das diskordante verhalten des bronchialsystems [Two

new parameters to characterise the bronchial status in vivo and after administration of bronchial

active substances. The discordant behaviour of the bronchial system]. Respiration 1980; 39: 97–

104.

59. Hafer C, Strohl K, Fouke J. Phasic changes in upper airway impedance. Respir Physiol 1987;

70: 13–24.

60. Peslin R, Ying Y, Gallina C, Duvivier C. Within-breath variations of forced oscillation resistance

in healthy subjects. Eur Respir J 1992; 5: 86–92.

61. Reinhold P, Barnikol W. Breath-by-breath evaluation of the oscillatory resistance as assessed with

monofrequency oscillation method in calves - a non-invasive procedure to characterize the motility

of the bronchial system. Eur Respir Rev 1994; 4: 178–181.

H.J. SMITH ET AL.

102

Page 32: Forced oscillation technique and impulse oscillometry

62. Reinhold P, MacLeod D, Lekeux P. Comparative evaluation of impulse oscillometry and a

monofrequency forced oscillation technique in clinically healthy calves undergoing

bronchochallenges. Res Vet Sci 1996; 61: 206–213.

63. Klein C, Reinhold P. Analysis of respiratory mechanics by impulse oscillometry in non-sedated

and diazepam-sedated swine. Res Vet Sci 2001; 70: 181–189.

64. Klein C, Smith HJ, Reinhold P. Respiratory mechanics in conscious swine: effects of face

mask, head position and bronchoconstriction evaluated by impulse oscillometry. Res Vet Sci 2003;

75: 71–81.

65. Klein C, Smith H-J. Analysis of respiratory mechanics in sedated and non-sedated horses. Eur

Respir J 2003; 22: Suppl. 45, 689.

66. Van Erck E, Votion D, Art T, Lekeux P. Measurement of respiratory function by impulse

oscillometry in horses. Equine Vet J 2004; 36: 21–28.

67. Van Erck E, Votion D, Kirschvink N, Art T, Lekeux P. Use of the impulse oscillometry system

for testing pulmonary function during methacholine bronchoprovocation in horses. Am J Vet Res

2003; 64: 1414–1420.

68. Marchal F, Loos N. Respiratory oscillation mechanics in infants and preschool children. Eur

Respir Mon 1997; 5: 58–87.

69. Badia J, Farre R, John Kimoff R, et al. Clinical application of the forced oscillation technique

for CPAP titration in the sleep apnea/hypopnea syndrome. Am J Respir Crit Care Med 1999;

160: 1550–1554.

70. Badia J, Farre R, Montserrat J, et al. Forced oscillation technique for the evaluation of severe

sleep apnoea/hypopnoea syndrome: a pilot study. Eur Respir J 1998; 11: 1128–1134.

71. Farre R, Ferrer M, Rotger M, Torres A, Navajas D. Respiratory mechanics in ventilated COPD

patients: forced oscillation versus occlusion techniques. Eur Respir J 1998; 12: 170–176.

72. Farre R, Gavela E, Rotger M, Ferrer M, Roca J, Navajas D. Noninvasive assessment of

respiratory resistance in severe chronic respiratory patients with nasal CPAP. Eur Respir J 2000;

15: 314–319.

73. Farre R, Peslin R, Rotger M, Barbera J, Navajas D. Forced oscillation total respiratory

resistance and spontaneous breathing lung resistance in COPD patients. Eur Respir J 1999; 14:

172–178.

74. Decramer M, Demedts M, van de Woestijne KP. Isocapnic hyperventilation with cold air in

healthy non-smokers, smokers and asthmatic subjects. Bull Eur Physiopathol Respir 1984; 20: 237–

243.

75. Nagels J, Landser F, van der Linden L, Clement J, van de Woestijne K. Mechanical properties of

lungs and chest wall during spontaneous breathing. J Appl Physiol 1980; 49: 408–416.

76. Van Noord J, Clement J, Van de Woestijne K, Demedts M. Total respiratory resistance and

reactance in patients with asthma, chronic bronchitis and emphysema. Am Rev Respir Dis 1991;

143: 922–927.

77. Van Noord J, Smeets J, Clement J, Van de Woestijne K, Demedts M. Assessment of reversibility of

airflow obstruction. Am J Respir Crit Care Med 1994; 150: 551–554.

78. Van Noord JA, Clement J, Cauberghs M, Mertens I, Van de Woestijne KP, Demedts M. Total

respiratory resistance and reactance in patients with diffuse interstitial lung disease. Eur Respir J

1989; 2: 846–852.

79. Coe C, Watson A, Joyce H, Pride N. Effects of smoking on changes in respiratory resistance with

increasing age. Clin Sci 1989; 76: 487–494.

80. Phagoo S, Watson R, Silverman M, Pride N. Comparison of four methods of assessing airflow

resistance before and after induced airway narrowing in normal subjects. J Appl Physiol 1995;

79: 518–525.

81. Yap J, Watson R, Gilbey S, Pride N. Effects of posture on respiratory mechanics in obesity. J Appl

Physiol 1995; 79: 1199–1205.

82. Brochard L, Pelle G, de Palmas J, et al. Density and frequency-dependence of resistance in early

airway obstruction. Am Rev Resp Dis 1987; 135: 579–584.

OSCILLOMETRY: FOT AND IOS

103

Page 33: Forced oscillation technique and impulse oscillometry

83. Beydon L, Malassine P, Lorino A, et al. Respiratory resistance by end-inspiratory occlusion and

forced oscillations in intubated patients. J Appl Phyiol 1996; 80: 1105–1111.

84. Lorino AM, Hamoudi K, Lofaso F, et al. Effects of continuous negative airway pressure on lung

volume and respiratory resistance. J Appl Physiol 1999; 87: 605–610.

85. Zerah-Lancner F, Lofaso F, Coste A, Ricolfi F, Goldenberg F, Harf A. Pulmonary function in

obese snorers with or without sleep apnea. Am J Respir Crit Care Med 1997; 156: 522–527.

86. Zerah F, Lorino A, Lorino H, Harf A, Macquin-Mavier I. Forced oscillation technique vs.

spirometry to assess bronchodilatation in patients with asthma and COPD. Chest 1995; 108: 41–47.

87. Peslin R. Methods for measuring total respiratory impedance by forced oscillations. Bull Eur

Physiopath Resp 1986; 22: 621–631.

88. Peslin R, Duvivier C. Partitioning of airway and respiratory tissue mechanical impedances by body

plethysmography. J Appl Physiol 1998; 84: 553–561.

89. Peslin R, Duvivier C. Removal of thermal artifact in alveolar pressure measurement during forced

oscillation. Respir Physiol 1999; 117: 141–150.

90. Peslin R, Duvivier C, Didelon J, Gallina C. Respiratory impedance measured with head generator

to minimize upper airway shunt. J Appl Physiol 1985; 59: 1790–1795.

91. Peslin R, Duvuvier C, Gallina C, Cervantes P. Upper airway artifact in respiratory impedance

measurements. Am Rev Respir Dis 1985; 132: 712–714.

92. Klug B, Bisgaard H. Measurement of lung function in awake 2–4-year-old asthmatic children

during methacholine challenge and acute asthma: a comparison of the impulse oscillation

technique, the interrupter technique, and transcutaneous measurement of oxygen versus whole-

body plethysmography. Pediatr Pulmonol 1996; 21: 290–300.

93. Hellinckx J, De Boeck K, Bande-Knops J, van der Poel M, Demedts M. Bronchodilator response

in 3-6.5 years old healthy and stable asthmatic children. Eur Respir J 1998; 12: 438–443.

94. Buhr W, Jorres R, Berdel D, Landser F. Correspondence between forced oscillation and body

plethysmography during bronchoprovocation with carbachol in children. Pediatr Pulmonol 1990;

8: 280–288.

95. Carvalhaes-Neto N, Lorino H, Gallinari C, et al. Cognitive function and assessment of lung

function in the elderly. Am J Respir Crit Care Med 1995; 152: 1611–1615.

96. Wesseling GJ. Respiratory impedance measurements in clinical lung function testing. Proefschrift,

Maastricht, April 1993.

97. Sharma G, Arora H, Lester L, Goldman M. Forced oscillatory (IOS) assessment of pulmonary

function and bronchodilator response in children with neuromuscular disorders. Am J Respir Crit

Care Med 2003; 167: A512.

98. Radulovic M, Spungen A, Wecht J, et al. Effects of neostigmine and glycopyrrolate on pulmonary

resistance in spinal cord injury. J Rehabil Res Dev 2004; 41: 53–58.

99. Niven A, Backenson T, Goldman M, Weisman I. Resistance values measured by forced oscillation

(IOS) are race dependent and increased at low frequencies in non-caucasian subjects. Am J Respir

Crit Care Med 2004; 169: A103.

100. Spahn J, Cherniack R, Paull K, Gelfand E. Is forced expiratory volume in one second the best

measure of severity in childhood asthma? Am J Respir Crit Care Med 2004; 169: 784–786.

101. Saadeh C, Goldman M, Gaylor P, et al. Forced oscillation using impulse oscillometry (IOS)

detects false negative spirometry in symptomatic patients with reactive airways. J Allergy and Clin

Immun 2003; 111: S136.

102. Hachem R, Chakinala M, Yusen R, et al. The predictive value of bronchiolotis obliterans

syndrome stage 0-p. Am J Respir Crit Care Med 2004; 169: 468–472.

103. Ross D, Goldman M, Strieter R, Belperio J, Ardehali A. "Multi-frequency forced oscillation

technique (FOT) for assessment of lung allograft function: a pilot study". J Heart Lung Transplant

2004; 23: S131.

104. Estenne M, Van Muylem A, Knoop C, Antoine M. Detection of obliterative bronchiolitis

after lung transplantation by indexes of ventilation distribution. Am J Respir Crit Care Med 2000;

162: 1047–1051.

H.J. SMITH ET AL.

104

Page 34: Forced oscillation technique and impulse oscillometry

105. Quanjer PhH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volume and

forced ventilatory flows. Report Working Party, Standardization of lung function tests. European

Coal and Steel Community. Official statement of the European Respiratory Society. Eur Respir J

1993; 6: Suppl. 16, 5–40.

106. Orehek J, Nicoli MM, Delpierre NS, Beaupre A. Influence of the previous deep inspiration in the

spirometric measurements of provoked bronchoconstriction in asthma. Am Rev Respir Dis 1981;

123: 569–572.

107. Salome CM, Thorpe CW, Diba C, Brown NJ, Berend N, King GG. Airway re-narrowing

following deep inspiration in asthmatic and nonasthmatic subjects. Eur Respir J 2003; 22: 62–68.

108. Schmekel B, Smith HJ. The diagnostic capacity of forced oscillation and forced expiration

techniques in identifying asthma by isocapnic hyperpnoea of cold air. Eur Respir J 1997; 10: 2243–

2249.

109. Jackson AC, Gulesian PJ Jr, Mead J. Glottal aperture during panting with voluntary limitation of

tidal volume. J Appl Physiol 1975; 39: 834–836.

110. Smith T, DuBois A. The effects of scopolamine on the airways of man. Anesthesiology 1969;

30: 12–18.

111. Hensley MJ, O’Cain CF, McFadden ER Jr, Ingram RH Jr. Distribution of bronchodilatation in

normal subjects: beta agonist versus atropine. J Appl Physiol 1978; 45: 778–782.

112. Bankier AA, Van Muylem A, Knoop C, Estenne M, Gevenois PA. Bronchiolitis obliterans

syndrome in heart-lung transplant recipients: diagnosis with expiratory CT. Radiology 2001;

218: 533–539.

113. Nielsen K, Bisgaard H. The effect of inhaled budesonide on symptoms, lung function, and cold air

and methacholine responsiveness in 2- to 5-year-old asthmatic children. Am J Respir Crit Care

Med 2000; 162: 1500–1506.

114. Nielsen K, Bisgaard H. Lung function response to cold air challenge in asthmatic and healthy

children of 2-5 years of age. Am J Respir Crit Care Med 2000; 161: 1805–1809.

115. Nielsen K, Bisgaard H. Discriminative capacity of bronchodilator response measured with three

different lung function techniques in asthmatic and healthy children aged 2 to 5 years. Am J Respir

Crit Care Med 2001; 164: 554–559.

116. Grimby G, Takishima T, Graham W, Macklem P, Mead J. Frequency dependence of flow

resistance in patients with obstructive lung disease. J Clin Invest 1968; 47: 1455–1465.

117. Kjeldgaard J, Hyde R, Speers D, Reichert W. Frequency dependence of total respiratory resistance

in early airway disease. Am Rev Respir Dis 1976; 144: 501–508.

118. Goldman M, Knudson R, Mead J, Peterson N, Schwaber J, Wohl M. A simplified measurement of

respiratory resistance by forced oscillations. J Appl Physiol 1970; 28: 113–116.

119. Bhansali P, Irvin C, Dempsey J, Bush R, Webster J. Human pulmonary resistance: effect of

frequency and gas physical properties. J Appl Physiol 1979; 47: 161–168.

120. Niewoehner DE, Kleinerman J, Rice DB. Pathologic changes in the peripheral airways of young

cigarette smokers. N Engl J Med 1974; 291: 755–758.

121. Frantz I, Close R. Alveolar pressure swings during high frequency ventilation in rabbits. Pediatr

Res 1985; 19: 162–166.

122. Malmberg LP, Pelkonen A, Poussa T, Pohianpalo A, Haahtela T, Turpeinen M. Determinants of

respiratory system input impedance and bronchodilator response in healthy Finnish preschool

children. Clin Physiol Funct Imaging 2002; 22: 64–71.

123. Ortiz G, Menendez R. The effects of inhaled albuterol and salmeterol in 2- to 5-year-old asthmatic

children as measured by impulse oscillometry. J Asthma 2002; 39: 531–536.

124. Nielsen K, Bisgaard H. Lung function response to cold air challenge in asthmatic and healthy

children of 2-5 years of age. Am J Respir Crit Care Med 2000; 161: 1805–1809.

125. Ward P, Hunninghake G. Lung inflammation and fibrosis. Am J Respir Crit Care Med 1998;

157: S123–S129.

126. Weibel ER. Morphometry of the Human Lung. New York, Academic Press 1963.

OSCILLOMETRY: FOT AND IOS

105