Development of an automated pressure-threshold loading device for evaluation of inspiratory muscle performance

Development of an automated pressure-threshold loadingdevice for evaluation of inspiratory muscle performance

M. P. Caine, G. R. Sharpe* and A. K. McConnell

Wolfson School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough, UK*School of Life Sciences, The Nottingham Trent University, Clifton Campus, Nottingham, UK Department of Sport Sciences, Brunel University, Osterley Campus, Middlesex, UK

AbstractIn 1996 Johnson et al. (Medicine and Science in Sports and Exercise, 28, 1129±1137)concluded that respiratory muscle fatigue may limit human performance whereby thework done by the respiratory muscles to drive the ventilatory pump might signi®cantlyaffect exercise tolerance. As a result of these and similar ®ndings there is a growinginterest in respiratory muscle performance. However, methods for appraising inspir-atory muscle function which are relevant to exercise performance are limited. Thepurpose of the present paper is to describe the development and validation of anautomated pressure-threshold loading device for use in evaluating the maximumincremental performance of the inspiratory muscles and other endurance-basedmeasures of inspiratory muscle performance.The device utilizes a stepper motor to vary the resistance of a spring-loaded poppetvalve arrangement, thereby permitting automatic selection of various inspiratory muscleloading pro®les. The device is accurate to within 1.5% (0.1±1.5 cmH2O) of selectedload across its range and permits continuous incremental loading with a resolution of0.1 cmH2Oámin±1. This apparatus allows an endurance-based measure of inspiratorymuscle endurance to be obtained which is distinctly different from measures ofinspiratory muscle strength such as peak inspiratory mouth pressure. Furthermore,reproducible measures can be attained without the need to regulate breathing pattern orto maintain isocapnia. In summary, the current device is useful when a simple yetreproducible endurance-based measure of inspiratory muscle performance is required.

Keywords: incremental loading, lung function, respiratory muscle

Introduction

Historically, the pulmonary system was thought notto limit exercise capacity (Wasserman et al. 1981;Dempsey et al. 1982). This belief was supported bythe observation that pulmonary ventilation and gas

exchange remained adequate even during heavyexercise in most healthy populations. Recentlyhowever, it has emerged that work done by therespiratory muscles to drive the ventilatory pumpmight impact upon exercise tolerance. Johnson et al.(1996) concluded that respiratory muscle fatiguemay limit human performance. In addition, recentstudies have also shown that training the respiratorymuscle improves exercise tolerance (Caine &McConnell 1998; Spengler et al. 1999). With thisgrowing interest in respiratory muscle performance

Correspondence address:M. P. Caine, Wolfson School of Mechanical and Manufactur-ing Engineering, Loughborough University, LoughboroughLE11 3TU, UK. Tel.: 01509 227650. Fax: 01509 267648.E-mail: [email protected]

Ó 2001 Blackwell Science Ltd · Sports Engineering (2001) 4, 87±94 87

it has become apparent that methods for appraisinginspiratory muscle function which are relevant toexercise performance are somewhat lacking.

Due to their con®guration, appraising the con-tractile function of the diaphragm and otherinspiratory muscles is a unique challenge. Directmeasurement of force is not possible, hence res-piratory pressures are often substituted as anindirect measure of force production (strength).Inspiratory muscle strength can be assessed usingphrenic nerve stimulation allied to measures ofoesophageal and transdiaphragmatic pressures.However, this method is both invasive and resourceintensive. A more commonly adopted approach hasbeen to measure maximum pressures at the mouthduring voluntary inspiratory and expiratory efforts(Moxham 1996).

Although useful for assessing changes in respir-atory muscle function, measures of force alone failto describe the ®tness of the inspiratory muscles toaccommodate the demands placed upon them bythe increases in ventilation resulting from exercise.In light of this, several endurance-based measuresof inspiratory muscle function have been described,for a review refer to Clanton (1995). Inspiratorymuscle endurance is typically inferred from theability to sustain breathing tasks. These can bebroadly divided into two types: (i) hyperpnea ±elevated minute ventilation with no added externalresistance; and (ii) loaded breathing ± forcedbreathing against an external resistance.

Hyperpnea

Hyperpnea-based breathing tasks include suchmeasures as maximal voluntary ventilation, max-imal sustainable voluntary ventilation and incre-mental breathing tests. All these tasks requirerespiratory ¯ow or volume measurement, andwhilst useful functionally, are con®ned to special-ized laboratory-based facilities.

Loaded breathing

Martyn et al. (1987) described an incrementalthreshold-loading technique based on an earlier

design by Nickerson & Keens (1982) that mimics astep protocol used in whole-body exercise testing.Subjects begin inspiring at a pressure load of »30%peak inspiratory mouth pressure (pMIP); the loadis then increased every 2 min by »5±10% pMIPuntil the subject is no longer able to tolerate theload. This test is well tolerated (Morrison et al.1989), is sensitive to the effects of inspiratorymuscle training (Preusser et al. 19941 ) and has beenshown to be reproducible (McElvaney et al. 1989).

The major limitation presented by the device ofMartyn et al.'s (1987) and those based on the sameprinciple lie in the stepped nature of the test.Ideally a true continuous ramp would be employed,as this would yield greater resolution in testmeasures. However, Martyn et al.'s approach didnot permit this, as the loading was achievedmanually by adding weights to the central valveplunger arrangement. A shift away from discrete`step-wise' loading in favour of continuous incre-mentation would permit a ®xed percentage pMIPramp rate to be used, therefore standardizing theprotocol and facilitating more meaningful compar-isons between individuals.

The purpose of the present paper is to describethe development and validation of an automatedpressure-threshold loading device for use in eval-uating the maximum incremental performance ofthe inspiratory muscles and other endurance-basedinspiratory muscle performance measures. A briefoverview of currently employed methodologies willbe given, followed by a description of this noveldevice. Results from test procedures will also bepresented and discussed.

Implementation

The descriptions that follow provide an overview ofthe device's subsystems. However, reference toFig. 1 may be useful as it provides an illustration ofthe complete assembly.

The loading mechanism

Larson et al. (1988) and later Caine & McConnell(2000) successfully utilized a spring-loaded poppet

Automated inspiratory muscle loading device · M. P. Caine, G. R. Sharpe and A. K. McConnell

88 Sports Engineering (2001) 4, 87±94 · Ó 2001 Blackwell Science Ltd

valve to achieve inspiratory muscle loading. Theloading mechanism of the current device possesseda number of similarities to that described by Caine& McConnell (2000).

The following description is provided withreference to the assembly illustrated in Fig. 1.The area of the inlet valve was set at 250 mm2 toavoid increases in resistance as a result of air¯owlimitation. A stainless steel spring with closedground ends was used to provide a loading rangeof between 10 and 150 cmH2O. The poppet valvecomprised a milled aluminium head with a recessedlip housing a 17-mm diameter, 70 shaw nitrileo-ring. The valve seat comprised a stainless stealcone (wall angle 60°) housed within a polythene

outer sleeve. A locating pin on the poppet valvehead articulated with a cavity on the outer sleeve,thereby maintaining the alignment between valveand valve seat.

The external architecture and drive mechanics

A `T' shaped con®guration was used for theexternal architecture, similar in dimension to thatdescribed by Caine & McConnell (2000). Thevertical section of the `T' main body housed thespring, poppet valve and valve seat; this wasconstructed from polythene and was milled to sizeto achieve an airtight seal with the upper portionof the body. A prefabricated, nylon component

Figure 1 A cross-section of the automatedpressure-threshold loading device.

M. P. Caine, G. R. Sharpe and A. K. McConnell · Automated inspiratory muscle loading device


constituted the two horizontal arms of the `T'.A PVC respiratory mouthpiece was connected toone opening whilst the end distal to the mouth-piece, housed a self-contained one-way expiratory¯ap valve.

A 9-V stepper motor was attached to the uppersurface of the `T' section such that the drive shaftof the motor was housed internally. The driveshaft was connected to a length of brass studdingpositioned within the coils of the spring. A nylonbar containing a brass nut was located on thestudding; this was free to move in a vertical planebut was prevented from rotating by two brass rodslocated on the internal surface of the outersleeve. Rotation of the drive shaft resulted invertical motion of the nylon bar. Anti-clockwiserotation resulted in the spring being compressed,whilst clockwise rotation produced spring decom-pression.

The controller

Control of the stepper motor was achieved usinga remote unit; this comprised a 4 ´ 4 matrixkeyboard, a stepper motor driver, a 16 ´ 2 LCDdisplay module and electronic circuitry to interfacethese components. The unit was powered via a 9-VDC transformer. A 12-way ribbon (length 2 m)connected the stepper motor to the control unit.The device was software controlled, using codewritten on an independent platform and thendownloaded onto a 16-kB EPROM. This unit, inconjunction with a P8031-1 microcontroller and8-kB RAM microchip, permitted full control of thedevice.

Input to the controller was via the 4 ´ 4 keypadcoupled to the 16 ´ 2 LCD screen upon whichfeedback text and prompts were displayed. A drop-down menu format was used to access options.Scroll keys allowed the menus to be explored,whilst menu selections were made by entering thenumber corresponding to the desired selection.A choice of three menus was given:

(i) Manual. The user employs scroll keys duringthe test to adjust the load.

(ii) Set protocols. The user selects a pre-deter-mined ramp rate (choice of 8, range ±1 to±30 cmH2O á min±1).

(iii) New protocols. The user is provided withprompts to create a bespoke loading pro®le.

Having selected the appropriate test pro®le, thecontroller gives a `press any key to start' prompt.Upon initiation of the test, the controller's internalclock is started; the time (s) and calculated load(cmH2O) are then displayed throughout the test.The controller also displays a `press any key to®nish' prompt. Upon completion of the test anumber of options are available, these includere-running the test, displaying the protocol, orresetting the stepper motor. Whilst these optionsare self-explanatory it is worth expanding on thelatter option. When powering up the controller,and upon selection of the `reset' command, thestepper motor is automatically returned to its startposition. This is achieved using a microswitchoperation. The stepper motor's drive rod is rotatedfrom in a clockwise direction until the nylon bar,travelling upward (decompressing the spring) trig-gers the microswitch. Upon activation of the switchthe drive rod's direction is reversed for onecomplete revolution, thus moving the nylon bardownward. The bar subsequently makes contactwith the spring and comes to rest in a positionwhere compression would accompany furtherdownward movement. This process ensures thatthe correct start position is achieved prior to eachtrail and thus permits accurate load selectionthroughout the test.

Evaluation of the loading device

To appraise the accuracy of the loading mechanismthe selected load was compared to the actualload. This was achieved across the working rangeof the device by selecting an incremental rampof 20 cmH2O á min±1 with an initial load of10 cmH2O. A 3-L syringe was used to simulatebreathing through the device. The syringe wasconnected to the mouthpiece of the loading devicevia air-tight rubber tubing; pressure was recorded at



the mouthpiece using an electronic pressure mano-meter. Flow was also measured at the inspiratoryport using an ultrasonic ¯ow meter. Pressure and¯ow signals were logged using dedicated software.A `breathing frequency' of 10 breaths per minutewas used, with a constant `tidal volume' of 3 Lused throughout the evaluation.

Figure 2 plots measured load against selectedload, both indices require further explanation.Measured load was de®ned as the plateau pressuregenerated on each inspiratory manoeuvre. To fullycomprehend this concept it is necessary to consi-der the loading pro®le across the inspiratory phase.At end-expiration, there is neither pressure gen-eration nor ¯ow; this is followed by a sharp rise innegative pressure, generated as the inspiratorymuscles contract to work against the resistancegenerated by the loaded poppet valve. In this phasethere is still no ¯ow as the pressure beinggenerated by the contracting inspiratory musclesis insuf®cient to overcome the spring-generatedload, thus the valve remains closed. However, atthe point where pressure at the mouth exceedsthe resistance of the spring/valve arrangement thevalve opens, thereby permitting air¯ow. Thepressure generated in this phase was termed

the `plateau pressure' as it remained essentiallyconstant until such point where the inspiratorymuscles were no longer able to generate suf®cientpressure to keep the valve open. This point de®nedend-inspiration and thus the end of ¯ow genera-tion. Conceptually, `plateau pressure' describes themean load across the ¯ow generating phase of theinspiratory manoeuvre.

The selected pressure is that displayed andlogged by the loading device and is thus deter-mined by the extent of spring compression. Thus,the controller achieves a selected pressure based onthe relationship between spring length and load(for further explanation refer to Caine & McCon-nell 2000).

Figure 2 illustrates the agreement between theselected load and measured load; 99.5% of thevariance in the measured load was explained bythe variance in the selected load. The predictedload underestimated actual load slightly (predictedload � 0.986 ´ measured load).

Human implementation

The pressure-threshold loading device describedhas been used as a stand alone unit to appraise themaximum incremental performance of the inspira-tory muscles in several research studies (Caine1999). In these studies the device was hand-heldby subjects; a respiratory mouthpiece was used toprovide an interface with the device, whilst nose-clips were employed to occlude the nasal cavity,thereby preventing unloading breathing throughthe nose. The controller remained attached butremote from the loading device and was covered toprevent subject access. Subjects were asked toassume a comfortably standing position and wereinstructed to hold the loading device to their mouthin a manner that ensured the horizontal potion ofthe `T' shaped body was kept parallel to the ¯oor.

Prior to commencement of the test subjects werereminded that the trial was progressive and ulti-mately maximal. The sensations likely to beencountered were described and it was explainedthat they, not the investigator, were responsible fortermination of the test. Subjects performed 2 min

Figure 2 A plot of measured load against selected load obtainedusing the pressure-threshold loading device during an incre-mental protocol. Open circles denote actual data points, thesolid line depicts best ®t, whilst the bold, dashed line depictsthe line of identity.



of unloaded breathing (actual load 10 cmH2O) tofacilitate familiarization with the apparatus. In thisperiod, inappropriate breathing patterns weremodi®ed, by instructing subjects on technique.The test was initiated discretely with subjects beingleft largely undisturbed throughout the test. How-ever, towards the end of the trial, as the loadingbecame arduous, encouragement was given. How-ever, at no point in the test proper was coaching ofbreathing pattern undertaken. The end-point of thetest was usually well de®ned, being preceded by thesubjects failing to open the valve on a couple ofoccasions, followed by expedient removal of themouthpiece.

Choice of protocol

The present system was devised to permit continu-ous incremental loading with the primary measurebeing either test duration or peak pressure. Whendeciding upon the choice of protocol a number ofopposing issues needed to be considered. Shortertests are better tolerated, in addition, the end pointis less dependent upon subject motivation and istypically well de®ned. Unfortunately, the shorterthe test the greater the importance of the strengthcomponent. Shallow ramp rates yield longer teststhat, by de®nition, are more endurance dependent.The compromise is thus straightforward, longertests provide a more relevant measure of endurancewhilst shorter tests are more readily tolerated andreproducible.

The tests proposed by Nickerson & Keens(1982) and Martyn et al. (1987) both lasted»10 min. Indeed, sustainable inspiratory pressureis de®ned as the inspiratory pressure that can besustained for 10 min. Furthermore, given theexponential nature of the inspiratory muscleendurance curve, the strength component of anytest will be substantially reduced where loading istolerated for periods in excess of 2±3 min.

Breathing pattern can be controlled by usingpacing cues to determine breathing frequencycoupled with a biofeedback, incentive system toregulate tidal volume, although this detracts fromthe simplicity and practicability of the procedure.

Discussion

The loading device and protocol described providea means of appraising the maximum incrementalperformance of the inspiratory muscles. The deviceis accurate to within 1.5% (0.1±1.5 cmH2O) ofselected load across its range. Furthermore,because the ramp rate can be selected with aresolution of 0.1 cmH2O á min±1, and the loadupdated at 1-s intervals, the sensitivity of themeasure is good. This feature permits subject-speci®c loading pro®les to be employed whereabsolute loads can be selected accurately to ensurestandard test intensities between users. This facilitypermits meaningful comparisons to be madebetween subjects.

The loading pro®le is almost independent of¯ow; thus, the device may be used without the needto control breathing pattern. However, it should benoted that relatively small changes in duty cycleinspiratory ¯ow rate and tidal volume can substan-tially alter the performance measures obtained(Clanton et al. 1985; Clanton et al. 1990).

Although emphasis has been placed on incre-mental loading it should be noted that the system'sversatility permits a wide range of loading pro®lesto be achieved automatically. Practically, thismeans that breath-by-breath variable loading couldbe accomplished if desired.

Limitations

The major limitations of the device arise from thearti®cial nature of the loading. The measure isdif®cult to interpret functionally, as the relation-ship between breathing pattern and patterns ofmuscle activation during the test are not the sameas those employed during exercise. Whilst maximalincremental performance is not considered a puremeasure of inspiratory muscle endurance, evidencedoes exist that it is an endurance-based measure(Eastwood et al. 1994) and in this respect isdistinctly different from measures of inspiratorymuscle strength such as pMIP. In addition, becauseof the incremental nature of the test, improvementsin maximum tolerable load distort the true increase



in work performed. For example, a 50% improve-ment in peak load from ±50 to ±75 cmH2Orequires the work done to be more than doubled.In light of these factors a measure of external workwould be useful and could be achieved withappropriate instrumentation of the device. Thiswould compensate to some extent for differentbreathing strategies and provide a more robustmeasure of inspiratory muscle endurance than thatcurrently derived.

Conclusions

An automated pressure-threshold loading devicehas been described which is capable of appraisingthe maximum incremental performance of theinspiratory muscles. The device utilizes a steppermotor to vary the resistance of a spring-loadedpoppet valve arrangement, thereby permittingautomatic selection of various inspiratory muscleloading pro®les. The device is accurate to within1.5% (0.1±1.5 cmH2O) of selected load across itsrange and permits continuous incremental loadingwith a resolution of 0.1 cmH2Oámin±1. This designovercomes many of the limitations inherent in thesystem described by Nickerson & Keens (1982) andpermits more effective implementation of the load-ing protocol of Martyn et al. (1987). This apparatusallows an endurance-based measure of inspiratorymuscle endurance to be obtained which is distinctlydifferent from measures of inspiratory musclestrength such as pMIP. Furthermore, reproduciblemeasures can be attained without the need toregulate breathing pattern or maintain isocapnia. Insummary, the current device is useful when a simpleyet reproducible endurance-based measure ofinspiratory muscle performance is required.

References

Caine, M.P. (1999) Inspiratory muscle training in a non-clinical population: development, implementation andevaluation of a new technology. PhD Thesis, University

of Birmingham, Birmingham, UK.

Caine, M.P. & McConnell, A.K. (1998) Pressure thresholdinspiratory muscle training improves submaximal cycling

performance. In: Proceedings of the 3rd Annual Congress ofthe European College of Sports Science (eds A.J. Sargant &H. Siddons), pp. 101. The Centre for Health CareDevelopment, Liverpool, UK.

Caine, M.P. & McConnell, A.K. (2000) Development andevaluation of a pressure threshold inspiratory muscletrainer for use in the context of sports performance.Journal of Sports Engineering, 3, 149±159.

Clanton, T.L. (1995) Respiratory muscle endurance inhumans. In: The Thorax, 2nd edn (ed. C. Roussos),pp. 1199±1230. Marcel Dekker, Inc., New York.

Clanton, T.L., Ameredes, B.T., Thomson, D.B. & Julian,M.W. (1990) Sustainable inspiratory pressures overvarying ¯ows, volumes, and duty cycles. Journal of AppliedPhysiology, 69, 1875±1882.

Clanton, T.L., Dixon, G.F., Drake, J. & Gadek, J.E. (1985)Inspiratory muscle conditioning using a threshold load-ing device. Chest, 87, 62±66.

Dempsey, J.A., Hanson, P., Pegelow, D., Claremont, A. &Rankin, J. (1982) Limitations to exercise capacity andendurance: pulmonary system. Canadian Journal of AppliedSport Science, 7, 4±13.

Eastwood, P.R., Hillman, D.R. & Finucane, K.E. (1994)Ventilatory responses to inspiratory threshold loadingand role of muscle fatigue in task failure. Journal ofApplied Physiology, 75, 185±195.

Johnson, B.D., Aaron, E.A., Babcock, M.A. & Dempsey,J.A. (1996) Respiratory muscle fatigue during exercise:implications for performance. Medicine and Science inSports and Exercise, 28, 1129±1137.

Larson, J.L., Kim, M.J., Sharp, J.T. & Larson, D.A. (1988)Inspiratory muscle training with a pressure thresholdbreathing device in patients with chronic obstructivepulmonary disease. American Review of Respiratory Disease,138, 689±696.

Martyn, J.B., Moreno, R.H., Pare, P.D. & Pardy, R.L.(1987) Measurement of inspiratory muscle performancewith incremental threshold loading. American Review ofRespiratory Disease, 135, 919±923.

McElvaney, G., Fairbarn, M.S., Wilcox, P.G. & Pardy, R.L.(1989) Comparison of two-minute incremental thresholdloading and maximal loading as measures of respiratorymuscle endurance. Chest, 96, 557±563.

Morrison, N.J., Richardson, D.P.T., Dunn, L. & Pardy,R.L. (1989) Respiratory muscle performance in normalelderly subjects and patients with COPD. Chest, 95,90±94.

Moxham, J. (1996) Respiratory muscle testing. MonaldiArchives for Chest Disease, 51, 483±488.

Nickerson, B.G. & Keens, T.G. (1982) Measuring venti-latory muscle endurance in humans as sustainable



inspiratory pressure. Journal of Applied Physiology, 52,768±772.

Preusser, B.A., Winningham, M.L. & Clanton, T.L. (1994)High- vs low-intensity inspiratory muscle interval train-ing in patients with COPD. Chest, 106, 110±117.

Spengler, C.M., Roos, M., Laube, S.M. & Boutellier, U.(1999) Decreased exercise blood lactate concentrations

after respiratory endurance training in humans. EuropeanJournal of Applied Physiology, 79, 299±305.

Wasserman, K., Whipp, B.J. & Davies, J.A. (1981) Respir-atory physiology of exercise: metabolism, gas exchange,and ventilatory control. In: Respiratory Physiology III, Vol. 23,International Review of Physiology (ed. J.G. Widdicombe),pp. 149±211. University Park Press, Baltimore, MD, USA.



Documents

Development of an automated pressure-threshold loading device for evaluation of inspiratory muscle performance