Thorax 1993;48:936-946

Pulmonary rehabilitation in chronic respiratory insufficiency * 2Series editors: J7-F Muir and D 7 Pierson

Exercise in patients with chronic obstructivepulmonary disease

Michael J Belman

Exercise is widely promoted as a means ofimproving physical endurance. It is recom-mended, not only for the healthy, but also forindividuals with various disabilities and dis-ease. In respiratory medicine we have wit-nessed several decades of investigationdirected not only at the pathophysiology ofexercise in patients with chronic obstructivepulmonary disease (COPD), but also at theeffects of exercise training in improving func-tion. As was initially the case with coronaryartery disease, many physicians of the mid20th century adopted a very conservativeapproach and generally discouraged exercisein patients with significant COPD. Despitethe pleas for greater physical exercise forpatients with chronic lung disease by Barach,a pioneer of pulmonary medicine,' it was onlyin the late 1960s and early 1970s that hisideas were aggressively pursued. In the USAthere is now widespread support for pul-monary rehabilitation programmes which,almost without exception, include a liberaldose of exercise training. The transfer of thestandard recommendations for exercise train-ing to healthy subjects and even cardiacpatients has not been easy. The pattern ofexercise response in patients with COPD pre-sents some unusual and, in some cases,

Normal

FEV,NC 3-0/4-0 1400 r

300

E

0

c

0

._

(U

x

0

Co.a

C.

200

100 F

0

100

200 _

300

400i

Full expiration

Airways obstruction

FEV,NC 1 1/3-0 1

r,-;."\ Rest /

Max. expiration

0-5 Full inspiration

Figure 1 Spontaneous flow-volume curves at rest (dotted lines) and maximum exercise

(dashed lines) as well as maximum flow-volume curves at rest (outer solid line) in a

normal subject and a patient with chronic airways obstruction. Reproduced with permissionfrom ref 2.

unique features that require radical rethinkingof the traditional advice given to the normalsubject and those with heart disease. Thepurpose of this review is to highlight key fea-tures of the pathophysiology of this exercisepattern in patients with COPD and to analysethe evidence which supports exercise training.

Exercise limitation in COPDAbnormalities of ventilatory mechanics, respi-ratory muscles, alveolar gas exchange, andcardiac function are present to varyingdegrees in patients with COPD. Delineationof the major mechanisms underlying exerciselimitation has obvious value in that treatmentaimed at reducing the severity of a major lim-iting factor would be beneficial in improvingexercise function. This process has not alwaysbeen easy and it is likely that the importanceof limiting factors is not the same in everypatient. The discussion below deals with eachof these factors separately, although they areprobably interrelated in most patients.

VENTILATION AND PULMONARY MECHANICSThis is one of the most important factors thatlimits exercise performance. Expiratory airflow obstruction is the main pathophysiologi-cal result of the alveolar wall destruction andbronchiolar narrowing which characterisesthis disease. In moderate to severe obstructivelung disease resting expiratory airflowsapproach or are equal to maximal airflow.2'In contrast to normal subjects in whom expi-ratory flow limitation may only occur duringexpiration at the highest work rates, patientswith COPD show flow limitation over mostor all of expiration at low exercise levels(fig 1). In patients with severe disease flowlimitation is present at rest.4 The prolonga-tion of expiration together with a higher thannormal exercise breathing frequency leadsinexorably to dynamic hyperinflation with anincrease in end expiratory lung volume.5 Thedynamic hyperinflation causes an increase ininspiratory loading and work through (1) adecrease in static compliance as patients nowbreathe along a shallower portion of the pres-sure-volume curve; (2) a high inspiratorythreshold load caused by the need to generateadditional pressure required to overcomeelastic recoil pressure before inspiratory flowcan begin (this increased threshold pressurehas been referred to as intrinsic positive end

Pulmonary PhysiologyLaboratory, Cedars-Sinai Medical Center,8700 Beverly Blvd.,Room 6732,Los Angeles,California 90048, USAMichael J BelmanReprints requests to:Dr M J Belman

936

on 17 July 2018 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.9.936 on 1 Septem

ber 1993. Dow

nloaded from

Exercise in patients with chronic obstructive pulmonary disease

expiratory pressure); and (3) an exaggerationof dependence of compliance on frequency.As Younes5 has stated: "While the mechani-cal defect is primarily resistive in nature inexpiration, the mechanical consequences areencountered in inspiration and are primarilyrestrictive in nature." The more severe thereduction in the forced expiratory volume inone second (FEVy), the greater the increasein end expiratory lung volume.6 The dynamichyperinflation brings with it an increase ininspiratory load, but it is a necessary evil forwithout it the patients with COPD would notbe able to increase ventilation to meet thedemands of exercise. As end expiratory lungvolume rises the patient is able to increasemaximum expiratory airflow by breathingalong a higher portion of the expiratory flow-volume curve.47The importance of end expiratory lung vol-

ume in patients with mild COPD (FEV1/FVCratios of approximately 60%) has recentlybeen emphasised.8 In these patients it wasthought that ventilatory limitation plays onlya minor part in contrast to patients with moresevere disease. In a recent study, however, itwas shown that, although the ratio of maxi-mum exercise ventilation (VEmax) to themaximum voluntary ventilation at peak exer-cise was considerably less than 70%-a valuetraditionally used to rule out a ventilatorylimitation-these patients demonstrated a risein end expiratory lung volume and flow limi-tation during exercise. In contrast, agematched control subjects maintained orreduced their resting end expiratory lung vol-ume and achieved a maximum oxygen con-sumption (Vo,max) which was 30% higherthan the patients. These investigators con-cluded that, despite the mild degree ofCOPD, there was a significant impact on pul-monary mechanics during exercise.8

Because the respiratory system in the exer-cising patient with COPD fails to reach itsrelaxation volume, inspiration can only occurafter respiratory muscles develop sufficient

Thoracic cage a~elastic recoildirected inwards 4 Horizontal ribs

;clefibres Decreased zone

4""4of apposition?lmpaired Decreased diaphragmatic Medial orientation ofblood supply curvature diaphragmatic fibres

Figure 2 Detrimental effect of hyperinflation on respiratory muscle function. Reproducedwith permission from ref 16.

force to overcome the recoil pressure of thehyperinflated chest. Preliminary studies havenow examined the effect of applying continu-ous positive airway pressure as a means ofproviding inspiratory assistance.910 Theresults of this work showed that continuouspositive airway pressure reduced the work ofbreathing and dyspnoea. In the study byO'Donnell and colleagues9 the exerciseendurance was prolonged. This responseemphasises the importance of negating theloading effect of the intrinsic positive endexpiratory pressure.Dodd and colleagues'1 have shown that

patients with airflow obstruction attempt tocompensate for the increase in end expiratorylung volume by actively recruiting abdominaland expiratory rib cage muscles during expi-ration. At the onset of inspiration the patientsrapidly relaxed these muscles, an effect whichallows them to exploit the outward recoil ofthe chest wall and gravitational descent of thediaphragm at the onset of inspiration. Thisreaction functions as a form of inspiratoryassistance. On the other hand, excessive useof expiratory musculature during expirationincreases oxygen utilisation of the expiratorymuscles, further reducing the overall effi-ciency of breathing in these patients.12

It is well recognised that in moderate tosevere COPD maximal exercise ventilationreaches a high percentage of the maximumventilatory ventilation (MVV) at rest. ThisVEmax/MVV ratio may in fact even exceed100% in patients with severe airflow obstruc-tion.'3 There are significant correlationsbetween measures of expiratory airflow suchas the FEV1 and MVV on the one hand, andVEmax and Vo2max on the other.'4 15 Becauseof the relatively large scatter of the data, how-ever, the confidence intervals of individualpredictions are large and it is not possible topredict peak minute ventilation with greataccuracy in an individual patient. For exam-ple, the 95% confidence interval of one equa-tion is ±18 1/min despite a correlationcoefficient of 0 97.'4 Factors other thanmechanical ventilatory limitation also have arole (see below).

RESPIRATORY MUSCLE DYSFUNCTIONPatients with COPD exhibit respiratorymuscle weakness (see Tobin for review'6).Intrinsic factors such as hypoxia, hypercap-nia, acidaemia, and malnutrition impair respi-ratory muscle contractility. Superimposed onthis are the mechanical derangements whichfurther weaken diaphragmatic function.Hyperinflation shortens the diaphragm, mov-ing it to a disadvantageous portion of itslength-tension curve. Moreover, the zone ofapposition is reduced and this impairs theoptimal inspiratory action of the muscle(fig 2).16 Although patients with COPD showcompensatory changes in the diaphragmwhich allow for relative preservation of func-tion even at the limits of hyperinflation,'7these inspiratory pressures are still well belowthose of normal subjects at functional residualcapacity (FRC).18 Activity of the upper

Shortenedmus

937 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Belman

Figure 3 Tracings ofabdominal and thoracicexcursions. PandA showsexcursions in a patientwith synchronousthoracoabdominalmovements at rest duringleg exercise (LE) and armexercise (AE). PandBshows that the synchronouspattern observed at restand during LE changes todyssynchronous duringAE. Ful inward retractionduring inspection is seen inthe last two breaths of thearm exercise tracing.Reproduced withpermission from ref 19.

A

coc0

x

0

0)

E

3

.20

.0

B

C.. l

limbs'920 is an additional aggravating factorwhich hampers diaphragmatic function.During arm work the stabilising effect of theshoulder girdle on the thorax is lost and theinspiratory load is shifted onto the diaphragmand muscles of expiration. In these circum-stances the diaphgram is required to assume a

greater load and, as noted above, is ill pre-

pared to do so (fig 3).1920 The net result is a

greater limitation of arm than of leg exerciseassociated with the earlier onset of dyspnoeain many patients with severe airflow obstruc-tion. As performance of most activities ofdaily living require repetitive upper extremitymovement, this phenomenon has importantimplications for patients with COPD (see sec-

tion on upper extremity training below).

RESPIRATORY MUSCIE FATIGUEWhether or not respiratory muscle fatigueoccurs during exercise in patients with COPDis not clear. Preliminary evidence from Pardyand coworkers2' showed that in some patientsa decrease in the high to low ratio-an elec-tromyographic index of fatigue-occurred inonly some patients during exercise. The factthat the high to low ratio has proved useful incarefully controlled laboratory experimentsdoes not imply that it can be transferred touse in individuals during exercise. As Younes

has stated,22 a change in the ratio may occur

between rest and exercise because of changesin breathing pattern and changes in the spa-tial relationship of the electromyographicelectrode and the muscle. Moreover, othermuscles recruited at higher exercise intensi-tites may contaminate the signal. The pres-ence of thoracoabdominal asynchrony duringbreathing has also been cited as support forthe presence of inspiratory muscle fatigue.More recent evidence23 24 indicates, however,that asynchrony of the thorax and abdomenduring inspiration is not pathognomonic offatigue, but may be seen in circumstances inwhich an individual breathes against a highinspiratory load. With cessation of the loadingthe breathing pattern returns to normal, even

though presumably the low frequency fatigueinduced by the loading persists for severaladditional hours. Definitive proof of fatiguewould require documentation of decreasedmuscle contractility after performance ofwork. Rochester has emphasised that inspira-tory muscle weakness is more important thanfatigue.'825 He emphasises the ratio of thepressure required per breath to the maximuminspiratory pressure (Pbreath/Pmax) as anindex of the weakness. During exercisePbreath rises as inspiratory work increaseswhile the rise in end expiratory lung volumeand configurational changes in the diaphragmreduce the Pmax. The net effect is a reduc-tion in functional diaphragmatic strength dur-ing exercise.

IMPAIRED GAS EXCHANGEHypoxaemia, a common feature of COPD,frequently shows further reductions duringexercise. A low diffusing capacity (<55% ofthe predicted value) has been used as a pre-dictor of those patients in whom exercisedesaturation will occur.26 The hypoxaemia ofexercise is largely due to the effects of areduction in mixed venous Po2 on low venti-lation diffusion lung units27 aggravated insome cases by hypoventilation. On the otherhand, some patients do show an improve-ment in Pao2 with exercise which must reflectan improvement in intrapulmonary ventila-tion perfusion matching.28 There is little evi-dence for diffusion limitation. The absence ofthe normal exercise decrease in the physiolog-ical dead space to tidal volume ratio (VD/VT)further aggravates the ventilatory limitation inCOPD. In order to maintain efficient carbondioxide output in the presence of a reducedalveolar ventilation, greater than normalincreases in total minute ventilation arerequired as exercise intensity increases29 (seesection on Lactic acidosis and exercise train-ing below).

CARDIOVASCULAR FUNCTIONRemodelling. of the muscular arteries andarterioles is the main cause of the increase inpulmonary vascular resistance.30 Thesechanges lead to thickening of the intima andnarrowing of the arterial and arteriolarlumens and are more extensive than theincrease in muscle seen in the media ofmedium and small arteries in people exposedto high altitude hypoxia. Other factors thatplay a part in the elevated pulmonary vascularpressures include emphysematous destructionof the vascular bed, alveolar hypoxia,increased alveolar pressure, increased haema-tocrit and acidosis."3Commonly observed abnormalities of car-

diovascular function during exercise are anincreased heart rate/Vo2 ratio related to a shiftupwards and to the left of the heart rate/Vo2 slope which itself, however, may be nor-mal.232 In other words, at a comparable V02the heart rate in a patient with COPD isincreased with a corresponding decrease inthe oxygen pulse. This means that estimationof exercise intensity in these patients by

MM

Abdomen-m-i !Al -.l

Insp iratio n

Thorax

Abdomen

InspirationThorax

-44t

RE T L. E. A. E.

938

on 17 July 2018 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.9.936 on 1 Septem

ber 1993. Dow

nloaded from

Exercise in patients with chronic obstructive pulmonary disease

160

42) 140co

a,120------

t: 100 A -Normal60 year man

80 FEV1 -0 ICOPD

60 05 1 1 5 2

V02 (I/min)

Figure 4 Normal heart rate response to exercise isillustrated by the parallel solid lines. In the patient withCOPD, in whom the heart response is generally at theupper limit of normal, the Vo, achieved at a heart rate of120 beats/min (1 0 I/min) would be less than a normalsubject (1 33 1/min).

means of heart rate can be erroneous (fig 4)."A heart rate of 120-130 beats/min in a nor-mal individual reflects a Vo2 of greater than 1litre/min while in a patient with COPD thismay represent a significantly lower V2. Thisphenomenon has implications for the inten-sity of exercise training which will be dis-cussed later. The reduction in heart ratesachieved at peak exercise are proportionatelyless than the reduction in peak Vo2, and themaximal oxygen pulse at peak exercise istherefore smaller. The rise in the cardiac out-put/Vo2 relationship is considered generally tobe normal2 despite the fact that pulmonaryvascular resistance is increased and there is ahigher than normal rise in pulmonary arterypressure. The increased afterload on the rightventricle can cause right ventricular dysfunc--tion, but whether or not this actually limitsexercise is unclear."3

LACTIC ACIDOSISLactic acid is produced during incrementalexercise although the time at which it appearsin arterial blood varies and is dependent oncirculatory function and level of fitness. Thepoint at which blood lactate rises has beentermed the lactic acid threshold and precedes,by approximately 150 ml of Vo2, the increasein minute ventilation related to the increasedcarbon dioxide output.34 35 This increase inventilation can be detected by one of manyindices as described by Wasserman andcoworkers but, most recently, they haveemphasised the use of the "V slopecriterion."36 In this index the rate of rise incarbon dioxide output is plotted againstoxygen uptake. While oxygen uptake remainslinear at the onset of lactic acid production,carbon dioxide output increases and so abreak point can be discerned. This inflectionpoint has been termed the "anaerobic thresh-old" by Wasserman and colleagues, whileother investigators have used the term "venti-lation threshold." This difference in terminol-ogy symbolises a heated controversy.Wasserman et al feel that the appearance oflactic acid truly indicates a transition toanaerobic glycolysis because of tissue

hypoxia.37 Their critics disagree and considerthat, while lactic acid certainly does rise dur-ing exercise, it does not necessarily implyanaerobiasis but merely an imbalancebetween lactate production on the one handand its utilisation on the other.'8 In patientswith COPD lactic acid and anaerobic thresh-olds can be determined even in those withmoderately severe disease39 although, clearly,peak lactate levels will be considerablyreduced in these patients because of theiroverall reduction in exercise capacity.'940 Itshould be noted that the lactic acid in thesepatients probably arises from working limbmuscles since it is those patients who reachthe highest work rates who show the highestlactate levels.40 Conversely, patients with verysevere obstructive disease in whom respira-tory muscle work is high have low lactatelevels and, moreover, lactate levels during iso-capnic hyperpnoea are only marginallyincreased.4'

Sue et al 39 feel that the V slope criterion isuseful in detecting metabolic acidosis in thesepatients (fig 5) but recognition of V slopemay not always be easy, as was shown in arecent study42 in which not only was thereconsiderable interobserver variability in Vslope detection but a significant number ofpatients with exercise induced metabolic aci-dosis did not develop inflection points.Conversely, inflection points were found inpatients without metabolic acidosis. Thisfinding detracts from the value of the V slopein detecting metabolic acidosis in patientswith COPD (see section on Exercise trainingintensity below).

PERIPHERAL MUSCLE FATIGUEDespite the emphasis on impaired ventilatorymechanics and dyspnoea it is now well docu-mented that a significant number of patientswith COPD will stop exercising because ofperipheral muscle fatigue.4' In a recent studyabout one third of patients stopped for thisreason. In addition, both limb and respiratory

1.5

0- 1.0 X

.>0-5 45

1*1

V02 (I/min) (STPD)

Figure 5 Carbon dioxide output (17CO2) plotted againstoxygen uptake (102). When the anaerobic threshold isreached, 1co2 accelerates compared with P'2. Theinflection point marking this acceleration is shown by thearrow. Reproduced with permission from ref 39.

939 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Belman

muscle show parallel decrements in strengthand contribute independently to reducedexercise capacity.

These are the major pathophysiologicalabnormalities seen in COPD, but otherfactors do play a part in limiting exercise.Their recognition is important as treatmentaimed at improving functional capacity musttake them into account. Additional factorsinclude nutritional status through its effect onboth limb and respiratory muscle strengthand endurance," perception of and responseto breathlessness which varies amongstpatients,4546 and psychological factors such asdepression, anxiety, and fear of exercise.47Furthermore, the role of deconditioning-a common problem in these patients becauseof their chronic inactivity-can aggravatethe impaired exercise tolerance.40 Althoughbreathlessness is clearly related to the severityof abnormalities in expiratory air flow this isnot the only factor, as recently emphasised byO'Donnell et al 48 who showed that patientswith comparable levels of airway obstructionmay have varying degrees of breathlessness.The major differences between mildly andseverely breathless patients were the presenceof hypoxaemia during exercise and an abnor-mally low diffusing capacity in the lattergroup. Other investigators have shown anadditional effect of psychological and psycho-social factors on functional capacity over andabove that of lung function. Their analysesshowed that dyspnoea, respiratory musclestrength, and spirometry each contributedindependently to functional limitation andemphasised that each of them should beassessed separately.45

Exercise training in COPDPulmonary rehabilitation programmes vary intheir complexity and may include severaltherapeutic components4950 including (1)patient and family education; (2) treatment ofbronchospasm by means of bronchodilatorsor reduction in bronchial secretions; (3) treat-ment of bronchial infections; (4) treatment ofcongestive heart failure; (5) oxygen therapy;(6) chest physical therapy including breathingtechnique training; (7) exercise recondition-ing; and (8) psychosocial therapy and voca-tional rehabilitation.

CONTROLLED EXERCISE STUDIESAlthough exercise reconditioning has longbeen considered an essential component ofthe rehabilitation process it is only veryrecently that a randomised study has con-firmed this belief.5' In this eight week study119 patients with COPD were randomisedeither to a comprehensive rehabilitation pro-gramme including exercise reconditioning or,alternatively, to an education control pro-gramme. The investigators provided educa-tion, physical and respiratory therapy,psychosocial support, and supervised exercisetraining to the treated group while the controlgroup received twice weekly classroominstruction in respiratory therapy, lung

disease, pharmacology, and diet but did notexercise. Before and after the treatment andafter an additional six months both groupsunderwent extensive physiological andpsychosocial tests. The major finding of thisstudy was that at eight weeks the improve-ment in exercise endurance as measured bytreadmill walking showed a mean increase intreadmill time from 12-5 minutes to 23 min-utes compared with an insignificant changefrom 12 to 13 minutes in the control group.At six months the treated group still main-tained a comparable advantage with a tread-mill endurance of approximately 21 minutescompared with 12 minutes in the controlgroup. No difference in the quality of well-being scale-a measure of health related qual-ity of life-was noted. This well designedrandomised controlled study definitivelyestablished exercise therapy as an essentialcomponent of the pulmonary rehabilitationprocess.

Relatively few other studies have comparedtreated and control groups. In the study byCockcroft and colleagues52 a treated group of19 patients was compared with a controlgroup of 20 patients. During training thepatients used cycle exercise, rowingmachines, and swimming and, in addition,free range walking was performed. This treat-ment was carried out for six weeks in a reha-bilitation centre; patients were subsequentlydischarged and encouraged to continue walk-ing and stair climbing. The control group wasgiven no special instructions to exercise. Thefindings showed an increase in 12 minutewalking distance and peak exercise V02 andVE in the treated group at two months andthese differences were significantly greaterthan those in the control group. The treatedgroup also showed improvement in generalwellbeing and dyspnoea. In a study byMcGavin and coworkers53 training was car-ried out by stair climbing at home, but thepatients were tested with a 12 minute walk.In this study of 24 patients (12 in the exercisegroup and 12 in a control group) a signifi-cant, albeit small, improvement in the 12minute walking distance was noted. Othernotable findings were an increase in stridelength in the exercise group but no change inpeak Vo2, heart rate, or minute ventilation asmeasured during an incremental cycleergometer test. Additional studies comparingtreated and control groups are summarisedelsewhere.49

UNCONTROLLED EXERCISE STUDIESNumerous uncontrolled studies of exercisetraining have been performed during the pastthree decades, the results of which have beensummarised in recent publications.333495054 Inthis review several more recent studies will bedealt with in detail. Apart from the study ofCasaburi and coworkers40 the findings aresimilar to those in previous work. These stud-ies do, however, effectively highlight themethods of testing and training and discussunresolved issues, including mode and inten-sity of exercise training.

940 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Exercise in patients with chronic obstructive pulmonary disease

LACTIC ACIDOSIS AND EXERCISE TRAININGIn an editorial published in 1986 Casaburiand Wasserman55 emphasised the role of car-bon dioxide output as the major drive to ven-tilation during exercise. Recognising the wellknown relationships between VE on the onehand and Vco,, arterial Pco2, and VD/VTratio on the other, they suggested that aerobictraining in patients with COPD would reducecarbon dioxide output and the ventilatorystimulus. The interrelationship of these vari-ables is expressed in the equation

VE k x Vco2Paco2 (1 -VDNVT)

where VE is expired minute ventilation, Vco2is carbon dioxide output, Paco, is partialpressure of arterial carbon dioxide, VDNVT isthe physiological dead space to tidal volumeratio, and k is a constant.The lactic acid produced during exercise

is buffered mainly by bicarbonate with thegeneration of carbonic acid which dissociatesto carbon dioxide and water. The carbondioxide produced by the buffering of lacticacid must be excreted by the lungs in addi-tion to carbon dioxide produced by musclemetabolism during exercise. Exercise trainingdelays the rise in blood lactate levels so anydelay in lactic acid production will, by reduc-ing the carbon dioxide load, decrease the ven-tilatory requirements during exercise. Theeffect of aerobic training and reduction in VEduring exercise has been well documented innormal subjects by these investigators. Athigh levels of work near peak V02, largereductions of 30-40 I/min in VE can beachieved in normal individuals.40With this rationale in mind, Casaburi and

Wasserman from the USA, in conjunctionwith a group of Italian investigators,40 per-formed a study in which high and low inten-sity training was performed in patients with

High work ratetraining group

Low work ratetraining group

X1) a)

co CN .2 t (,( oUJ *

N*>

Z!*> I

C14 0 'el-, a)cU W 0 C. wU a) LU 0 0 wU (U

-201

-30F

Figure 6 Changes in physiological responses to identical exercise tasks in high and lowwork rate training groups. Reproduced with permission from ref 40.

COPD and the effects on lactate productionwere examined in detail. Exercise testing wasperformed on a cycle ergometer with breathby breath measurements of gas exchangebefore and after the training. Arterial bloodgas measurements and arterial lactate mea-surements were also made. The anaerobicthreshold was determined by means of themodified V slope technique.36 Training wasperformed on a calibrated cycle ergometerfive days a week for eight weeks. The highintensity group performed exercise at 45min/day at an intensity 60% of the differencebetween the anaerobic threshold and theVo,max. The low intensity group exercised at90% of this level, but the duration wasincreased so that total work performed in thetwo groups was similar.The major results of Casaburi's study were

a reduction in the peak Vco, and the maximalventilatory equivalent for oxygen (VE/VO,) inthe high intensity group. In a high work rate,constant load test, the high intensity trainedgroup showed significant reductions in bloodlactate, VE, VCo2, Vo,, and the VE/V02 ratio.Heart rate at comparable work rates wasreduced. All these findings confirm the devel-opment of a true aerobic training effect(fig 6). On the other hand, the group whotrained at the low intensity, even though thetotal work performed was similar, showedsmaller changes in these variables. In thisgroup, although the lactate decrease was sig-nificant (10%), the decreases in VE, VCo,0and Vo2 were not significantly different.Furthermore, a significant increase inendurance of exercise at the higher work rateseen in the high intensity trained group(6-6-11-4 min) was not seen in the low inten-sity trained group (6-9-7-5 min).

There was a significant relationshipbetween the decrease in minute ventilationduring exercise and the decrease in blood lac-tate (r = 0 73, AVE = 2-46 Vlmin/mEq lac-tate). The slope of the relationshipAVE/Alactate in these patients (fig 7) was con-siderably lower than that recorded in a previ-ous study in normal subjects in whom the VEdecreased by 7-2 Vlmin/mEq lactate. Thisstudy clearly shows that (1) significant lacticacidaemia occurs in patients with mild tomoderate chronic airways obstruction, and insome cases this may develop at low workrates (pedalling at 0 W); (2) both high andlow intensity training reduce the rise in lac-tate but the effect with high intensity trainingis considerably greater; (3) although lactatelevels and VE are lower after training inpatients with COPD, the reduction in ventila-tion in patients is only about a third as largeas that seen in normal subjects. The explana-tion for this difference is related to the factthat these patients show a reduced ventilatoryresponse to the lactic acidosis of exercise andtherefore a decrease in lactic acid after train-ing produces a comparably smaller decreasein VE.

Although this study clearly shows the gen-eration of a true aerobic training response,this was accomplished in a group of relatively

0

-10

-20

(U0)

co

o-0

941 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Belman

Figure 7 Relationbetween the decrease inventilation and thedecrease in arterial lactatein response to a highconstant work rate test as aresult ofa programme ofexercise training: A, highwork rate trained group;A, low work rate trainedgroup. Solid line isobtained by linearregression. APE= 2-84,A lactate = 1 19.Reproduced withpermission from ref 40.

1E

-E

a1)

co,C1)L,>L

Lactate decrease (mEq/1)

young patients (mean age 49) with rathermild disease (FEV, percentage predicted 56%and FEV,IFVC ratio 58%). These are not thetype of patients commonly found in mostrehabilitation programmes. In the past most

studies examined patients in whom the FEV,was considerably lower. In the USA it is not

unusual to find patients participating in exer-

cise programmes with FEV, values < 1.01.49Moreover, before the training these patientswere relatively unfit as shown by the lactatethreshold which was found to be at oxygen

consumptions below 1 /min. It is not surpris-ing, therefore, that they responded dramati-cally to the exercise programmes. The factthat the ventilatory response to exercise aci-dosis is blunted in patients with COPD fur-ther detracts from the practical benefit oflactate reduction. Casaburi et a140 document a

AVE/Alactate change of only one third of thatin normal subjects. This, however, was inmildly affected patients. In severely affectedpatients one would anticipate an even smallerreduction in lactate levels and consequently a

smaller decrease in exercise in ventilation.Most studies other than that of Casaburi et alhave treated patients with moderate to severe

disease.49 In many patients the average FEV,is lower, generally about 1 litre, and in some

cases less than that. Similar results in a more

severely affected group of patients would behelpful before these exercise recommenda-tions could be generalised.56

EXERCISE TRAINING TO TOLERANCE

Exercise testing and training directed at

manipulation of blood lactate levels involvescomplex measurements and should be con-

trasted with the more unstructured approachof exercising to tolerance, an approach whichhas been used in a large number of studies.49These studies, despite the fact that they havenot necessarily documented reductions in lac-tate levels, have shown that even severelyobstructed patients (some with extreme

hypercapnia57) can be exercised safely andshow impressive gains in submaximal exerciseendurance. This is particularly striking in thestudy of Niederman et al58 who showed thegreatest percentage improvement in those

patients with the lowest FEV, values and low

pretraining exercise endurances. In that studythe training was done without emphasis on

intensity, patients being allowed to choosetheir own exercise level. Exercise sessionswere conducted three times a week for twohours for a total of nine weeks. During eachsession the time was divided among cycling,treadmill walking, and lifting weights. Themost impressive gains were in cycleendurance which increased from 129-5 to726 1 W/min. Similar results were obtainedby Holle et al59 who showed large increases in-treadmill endurance.

In two recent studies large gains inendurance were achieved, although highintensity training was used and patients wereencouraged to reach maximal levels of venti-lation during training.606' In the former studypatients were initially separated into twogroups based on whether an anaerobicthreshold was reached. In those unable toreach an anaerobic threshold training wasperformed at the maximal work load achievedon the treadmill. In the patients who passedthe anaerobic threshold training intensity wasinitially aimed at the threshold level itself. Inboth groups intensity and duration wereincreased as tolerated. Of interest was thefinding that these patients could train at exer-cise ventilations close to or even exceedingthe maximum level reached on initial testing.In contrast to the work of Casaburi et al40both groups showed significant and compar-able improvements in endurance on thetreadmill. The investigators were quick topoint out that this does not mean that a highintensity training regimen is therefore desir-able for patients with COPD. In fact, in com-parison with the previously described study,5'gains in endurance were similar. Clearly bothapproaches are successful in the moderate toseverely affected patient; there does notappear to be an intrinsic benefit in demand-ing that training be performed at almost max-imal ventilatory capacity. High intensityexercise may also be disadvantageous becauseof the higher risk of injury and because thediscomfort of extreme exercise may reducecompliance with exercise programmes.62These findings are summarised in the table.

UPPER LIMB EXERCISE TRAININGThe impact of upper extremity exercise hasbeen discussed (see section on Respiratorymuscle dysfunction). Ries and coworkers63performed a randomised study which com-pared a control group with two groups whoused two different forms of upper arm train-ing. Testing was done by means of cycleergometry and unsupported arm exercise. Inaddition, three tests of activities of daily livingwere used, namely, dishwashing, dusting ablackboard, and placing grocery items onshelves. Training was performed for at leastsix weeks and showed that, although thepatients who underwent the upper extremitytraining improved their performance on anarm cycle ergometer, they did not improveperformance in arm activities of daily living.The specificity of limb training is emphasisedby the findings of Lake et al 64 who ran-

domised patients to one of three groups. The

942

AA

A& A

on 17 July 2018 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.9.936 on 1 Septem

ber 1993. Dow

nloaded from

Exercise in patients with chronic obstructive pulmonary disease

first group was a control group who receivedno training; the second group received upperlimb training only; and the third group per-formed combined upper and lower limbtraining. Upper limb training included cycleergometry with varying resistances, throwinga ball against a wall with the arms above thehorizontal, passing a bean bag over the head,and arm exercise with ropes and pulleys.Lower limb exercise was tested by cycleergometry and a six minute walk distance.Training was continued for one hour, threetimes a week, for eight weeks. The resultsshowed that limb training was limb specific.Thus it was only in the group that trainedwith the upper extremity that upper extremityendurance increased, while walk distanceimproved in the lower limb trained group.The combination trained group showedimprovements in both upper and lower limbendurance. A modified quality of life ques-tionnaire was also used but only producedsignificant changes in the group that receivedcombined upper and lower limb training.Two recent studies6566 have confirmed the

value of specific arm training. In both studiesspecific arm training resulted in increasedarm endurance and a reduction in the meta-bolic cost for arm exercise. In the latter studythe improvements in unsupported arm activ-ity were seen only in the group who per-formed unsupported exercise and not in thegroup who did supported arm exercise.66 Theimproved endurance in conjunction with thereduced metabolic cost is indicative ofimproved mechanical efficiency of movementof the arms and possibly breathing musclesduring arm activity.An interesting finding of previous studies63 64

was the lack of change in ventilatory musclefunction. Before and after the upper extrem-ity training these investigators examined ven-tilatory muscle endurance and neither foundan improvement. Early work by Keens et al67in patients with cystic fibrosis suggested thatupper extremity exercise may have a

Exercise training in chronic obstructive pulmonary disease

Casaburi et al40 Punzal et al60 Niedermnan et a158

No. of patients 9 57 24

FEVI/FVC 58% 44% 50%

Intensity High (60% of (1) VEmax Unstructured,difference between laissez faireanaerobic threshold (2) At anaerobic

and Vo,max) threshold

Frequency 5/week Daily treadmill 3/week

Duration 8 weeks inpatient Supervised 2/week x 4, 9 weeks45 min cycle then 1/week x 4 free 20 min on cycle,

daily unsupervised treadmill, upperwalking extremity

Test Cycle endurance Treadmill endurance Cycle endurance66-11-4 min 12-1-22-0 min 5-0-12-0 minAnaerobic thresholdT 12 min walk T

PeakVo, 10%T 10%T 19%TPsychosocial Not measured Breathlessness 4 Depression 4

Fatigue 4 Disability 4

crossover effect and improve respiratory mus-cle endurance but this was not confirmed in astudy of patients with COPD by Belman andKendregan,68 in which a group of patientswho performed upper extremity cycle ergo-metry did not improve their ventilatorymuscle endurance. Similarly, in both thesestudies of upper extremity training no signifi-cant change in ventilatory muscle functionwas found.6364

Mechanisms ofimprovmentImprovements in exercise tolerance may beascribed to one or more of the following fac-tors: improved aerobic capacity, or musclestrength, or both; increased motivation;desensitisation to the sensation of dyspnoea;improved ventilatory muscle function; andimproved technique of performance. Despitethe multiplicity of studies performed there is,as yet, no clear consensus on the predomi-nant mechanism of improvement.

IMPROVED AEROBIC CAPACITYIn normal subjects increased endurance haslargely been ascribed to changes in thetrained muscles.69 These changes, which con-sist mainly of increased capillary and mito-chondrial density together with increasedconcentrations of oxidative enzymes, occurconcomitantly with training induceddecreases in the exercise heart rate and con-stitute the major components of the aerobictraining response in normal subjects.70 Apartfrom the study by Casburi et al 40 this patternhas not been observed in patients with COPDso it is not possible to ascribe improved exer-cise endurance to improved aerobic perfor-mance. A striking feature of the results ofexercise training in patients with COPD is thefact that, almost without exception, investiga-tors have claimed success for their respectiveprogrammes despite the fact that trainingmodes, intensity, and frequency have variedwidely.49 Moreover, in a study in which aero-bic training effects were specifically examinedby means of muscle biopsies from the trainedlimbs no significant improvement in oxidativeenzymes was found.7' These authors con-cluded that patients with COPD were unableto exercise at the threshold intensity neces-sary to elicit a true aerobic response.

Although the emphasis on training hasconcentrated on endurance activities, recentevidence supports an important role forperipheral muscle strength. A third ofpatients with COPD implicated musclefatigue as the limiting factor during exercise.A subsequent randomised study evaluated theeffect of a weightlifting programme in thesepatients.72 The patients performed weighttraining three times a week for eight weeks.Both arm and leg strengthening exerciseswere done. The results showed an increase incycle endurance and reduction in symptomsas assessed by a questionnaire. This studycertainly reinforces the need not to neglectstrength training as an important componentof the training regimen.

943 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Belman

INCREASED MOTIVATIONIncreased motivation might easily account forthe improvement seen in some studies. Thiscould be evaluated by noting an increase inthe maximal VE or heart rate. However, nei-ther of these variables has increased consis-tently in cases where there has been anincrease in endurance. In submaximal steadystate exercise tests, where exercise endurancetime is the measure of improvement, motiva-tion may be a factor.

REDUCTION IN DYSPNOEAResearch into the mechanisms of dyspnoea iscomplicated by the inherent problems withmeasurement of intensity of a symptom. Thistopic has been reviewed recently.73 Variousscales and questionnaires are in use includingthe Borg scale for perceived exertion, thebaseline and transitional dyspnoea indices,and the chronic respiratory disease question-naire.74 Moreover, techniques are availablewhich allow measurement of quality of life.Improved measurement in these areas isessential to gauge the impact of pulmonaryrehabilitation programmes in general, andexercise training in particular.

Dudley et al 47 have reviewed the psycho-social aspects of pulmonary rehabilitation andcited several studies which have found corre-lations between improved exercise enduranceand improved feeling of wellbeing. One studyfound that psychological improvementresulted from either pulmonary rehabilitationincluding exercises or psychotherapy alone. Ithas also been shown that there is a better cor-relation between mood and motivation andexercise endurance than between pulmonaryfunction and exercise endurance.75 Severalstudies of exercise training have shownimprovements in wellbeing and reduction inbreathlessness.4950 In the study by Agle andcoworkers76 many of the patients alsoreported an improved sense of wellbeing anddecreased sensation of breathlessness afterexercise training. These authors speculatedthat the process of graduated exercise trainingin the presence of trained medical personnel"inadvertently functioned as a desensitisingform of behaviour therapy." They felt, there-fore, that progressive exercise led to adecrease in the unrealistic fear of activity anddyspnoea. A recent study by Belman andcoworkers77 showed that four repetitiveepisodes of treadmill walking over 10 days ata relatively high intensity resulted in adecrease in the perceived level of breathless-ness over this short period of exercise andspeculated that "desensitisation" may haveplayed a part. In a study of ventilatory muscletraining78 a control group showed significantincreases in exercise after participating in thetesting sequence only. This evidence hasgiven rise to the speculation that, whenpatients with dyspnoea experience theirsymptoms in a medically controlled environ-ment while simultaneously receiving supportand encouragement, they learn to overcomethe anxiety and apprehension associated withtheir dyspnoea. This desensitisation to

dyspnoea may be a key component to im-proved endurance after exercise, but furtherinvestigation is necessary to prove this point.

VENTILATORY MUSCLE TRAININGVentilatory muscle training will be dealt within a separate article in this series. Its role inimproving endurance is as yet unclear. Arecent meta analysis of ventilatory muscletraining concluded that any effect, if present,is small and unlikely to contribute signifi-cantly to improved exercise tolerance in thesepatients.79

IMPROVED MECHANICAL SKILLImproved skill in performance has beenfound in several studies including the earlystudies by Paez and coworkers80 who showedthat skill in treadmill walking improved withrepeated attempts. Clearly, skillful perfor-mance of the task decreases both the oxygencost and the ventilatory requirements ofwork, although the actual work rate isunchanged.8' This effect constitutes trainingof technique and can be used to advantage inthat these patients can be trained to performspecific tasks more efficiently. Although thetechnique of treadmill walking has beenshown to improve in some studies, it is notknown if this is indeed a component ofimprovement seen in walking other than on atreadmill.From the large number of studies per-

formed to date it is striking that there is noappreciable benefit on pulmonary functionand gas exchange.334950 As noted above, withthe singular exception of the work ofCasaburi et al40 no true aerobic training effecthas been found. Even in the absence of atraining effect it is impressive that there isalmost universal success shown for studies ofexercise training when the outcome measureis increased exercise endurance. This includesstudies in which the training intensity is low.The precise mechanism responsible for theimprovement is not clear, but the absence ofobjective cardiopulmonary improvementsraises the possibility that a reduction in dys-pnoea perception is important. Furtherresearch to evaluate this mechanism isrequired. Moreover, additional researchwhich combines measurements of exercise aswell as valid measures of breathlessness andquality of life are indicated. The transfer ofimproved walking endurance to increasedendurance for carrying out activities of dailyliving also requires improved documentation.

SummarySporadic visits to the local doctor followedsometimes by changes in oral and inhaledbronchodilators and occasionally by theaddition of steroids frequently does little tosignificantly improve symptoms and functionin the disabled patient with COPD. As inother chronic diseases, the management ofthese patients is facilitated by a teamapproach in conjunction with general re-habilitation principles.50 The rationale and

944 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Exercise in patients with chronic obstructive pulmonary disease

practical implementation of such a pro-gramme has recently been outlined by theAmerican Association of CardiopulmonaryRehabilitation.50 These are multifaceted pro-grammes but a key component, as outlinedabove, is exercise training. In this brief reviewthe various approaches available have beendescribed. Controversy still reigns regardingthe optimal modes of training and there areimportant differences among the severalapproaches. Two main groups can be delin-eated. One emphasises the detailed definitionof the impaired physiology with therapeuticmeasures targeted to specific defects.40 Thereis good documentation that, conversely,unstructured programmes that use treadmilland free range walking and cycling alsoimprove endurance for walking.59 Upperextremity training is of additional benefit.Programmes with as little as three sessionsper week of 1-2 hours of low intensity activityhave achieved success so we know that simpleprogrammes can be helpful. Moreover, with-out the necessity for complex testing andtraining methods these programmes can beimplemented with relatively low costs. Futureinvestigations to examine the relationshipbetween improved exercise capacity for walk-ing and arm exercise on the one hand, andthe ease of performance of activities of dailyliving on the other, will help to reinforce theeffectiveness of exercise programmes.

1 Barach AL, Bickerman HA, Beck G. Advances in thetreatment of non tuberculosis pulmonary disease. BullNYAcad Med 1964;1134:28-36.

2 Gallagher CG. Exercise and chronic obstructive pul-monary disease. Med Clin North Am 1990;74:619-41.

3 Olopade CO, Beck KC, Viggiano RW, Staats BA.Exercise limitation and pulmonary rehabilitation inchronic obstructive pulmonary disease. Mayo Clin Proc1992;67:144-57.

4 Grimby G, Stiksa J. Flow-volume curves and breathingpatterns during exercise in patients with obstructed lungdisease. ScandJ7 Clin Lab Invest 1970;25:303-13.

5 Younes M. Load responses, dyspnea, and respiratory fail-ure. Chest 1990;97:59-68S.

6 Regnis JA, Alison JA, Henke KG, Donnelly PM, ByePTP. Changes in end-expiratory lung volume duringexercise in cystic fibrosis relate to severity of lung dis-ease. Am Rev RespirDis 1991;144:507-12.

7 Stubbing DG, Pengelly LD, Morse JLC, Jones NL.Pulmonary mechanics during exercise in subjects withchronic airflow obstruction. 7 Appl Physiol 1980;49:511-5.

8 Babb TG, Viggiano R, Hurley B, Staats B, Rodarte JR.Effect of mild-to-moderate airflow limitation on exercisecapacity.J Appl Physiol 1991;70:223-30.

9 O'Donnell DE, Sanii R, Younes M. Improvement in exer-cise endurance in patients with chronic airflow limita-tion using continuous positive airway pressure. Am RevRespirDis 1988;138:1510-4.

10 Petrof BJ, Calderini E, Gottfried SB. Effect of CPAP onrespiratory effort and dyspnea during exercise in severeCOPD.3ApplPhysiol 1990;69:179-88.

11 Dodd DS, Engel L. Chest wall mechanics during exercisein patients with severe chronic air-flow obstruction. AmRev RespirDis 1984;129:33-8.

12 Johnson BD, Reddan WG, Seow KC, Dempsey JA.Mechanical constraints on exercise hyperpnea in a fitaging population. Am Rev Respir Dis 1991;143:968-77.

13 Spiro SG. Exercise testing in clinical medicine. Br J DisChest 1977;71:145-72.

14 Dillard TA. Prediction of ventilation at maximal exercisein patients with chronic airflow limitation. Chest 1987;92:195-6.

15 Dillard TA, Piantadosi S, Rajagopal KR. Prediction ofventilation at maximal exercise in chronic air-flowobstruction.Am Rev Respir Dis 1985;132:230-5.

16 Tobin MJ. Respiratory muscles in disease. Clin Chest Med1988;9:263-86.

17 Similowski T, Sheng Y, Gauther AP, Macklem PT,Bellemare F. Contractile properties of the humandiaphragm during chronic hyperinflation. N Engl J Med1991;325:917-23.

18 Rochester DF. The Diaphragm in COPD: better thanexpected, but not good enough. N Engl J Med1991;325:961-2.

19 Celli BR, Rassulo J, Make BJ. Dyssynchronous breathingduring arm but not leg exercise in patients with chronicairflow obstruction. NEnglJ3Med 1986;314:1485-90.

20 Criner GJ, Celli BR. Effect of unsupported arm exerciseon ventilatory muscle recruitment in patients withsevere chronic airflow obstruction. Am Rev Respir Dis1988;138:856-61.

21 Pardy RL, Rivington RN, Despas PJ, Macklem PT. Theeffects of inspiratory muscle training on exercise perfor-mance in chronic airflow limitation. Am Rev Respir Dis1981 ;123:426-33.

22 Younes M. Determinants of thoracic excursions duringexercise. In: Whipp BJ, Wasserman K, eds. Exercise:pulmonary physiology and pathophysiology. New York:Marcel Dekker, 1992:1-65.

23 Tobin MJ, Perez W, Guenther SM. Does rib cage-abdom-inal paradox signify respiratory muscle fatigue? J ApplPhysiol 1987;63:851-60.

24 Johnson BD, Saupe KW, Dempsey JA. Mechanical con-straints on exercise hyperpnea in endurance athletes._J Appl Physiol 1992;73:874-86.

25 Rochester DF. Respiratory muscle weakness, pattern ofbreathing, and CO, retention in chronic obstructive pul-monary disease: editorial. Am Rev Respir Dis 1991;143:901-3.

26 Owens GR, Rogers RM, Pennock BE. The diffusingcapacity as a predictor of arterial oxygen desaturationduring exercise in patients with chronic obstructive pul-monary disease. NEnglJMed 1984;310:1218-21.

27 Dantzker DR, D'Alonzo GE. The effect of exercise onpulmonary gas exchange in patients with severe chronicobstructive pulmonary disease. Am Rev Respir Dis1986;134:1135-9.

28 Barbera JA, Roca J, Ramirez J, Wagner PD, Ussetti P,Rodriguez-Roisin R. Gas exchange during exercise inmild chronic obstructive pulmonary disease. Am RevRespirDis 1991;144:520-5.

29 Wasserman K, Whipp BJ. Principles of exercise testing andinterpretation. Philadelphia: Lea and Febiger, 1987.

30 Wilkinson M, Ranghom CA, Heath D, Barer G, HowardPA. A pathophysiological study of 10 cases of hypoxiccor pulmonale. QJTMed 1988;66:65-85.

31 Matthay RA, Wiedemann HP. Cardiovascular pulmonaryinteraction in chronic obstructive pulmonary diseasewith special reference to the pathogenesis and manage-ment of cor pulmonale. Med Clin North Am 1990;74:571-618.

32 Nery LE, Wasserman K, French W. Contrasting cardio-vascular and respiratory responses to exercise in mitralvalve and chronic obstructive pulmonary diseases. Chest1983;83:446-53.

33 Belman MJ. Exercise training in pulmonary rehabilitation.Clin Chest Med 1986;7:585-98.

34 Beaver WL, Wasserman K, Whipp BJ. Bicarbonate buffer-ing of lactic acid generated during exercise .J ApplPhysiol 1986;60:472-8.

35 Wasserman K, Beaver WL. Anaerobic threshold and res-piratory gas exchange during exercise. J Appl Physiol1973;35:236-43.

36 Beaver WL, Whipp BJ. A new method for detectinganaerobic threshold by gas exchange. J7 Appl Physiol1986;60:2020-7.

37 Davis JA. Anaerobic threshold: review of the concept anddirections for future research. Med Sci Sports Exerc1985;17:6-18.

38 Brooks GA. Anaerobic threshold: review of the conceptand directions for future research. Med Sci Sports Exerc1985;17:22-31.

39 Sue DY, Wasserman K, Moricca RB, Casaburi R.Metabolic acidosis during exercise in patients withchronic obstructive pulmonary disease. Chest 1988;94:931-8.

40 Casaburi R, Patessio A, Ioli F, Zanaboni S, Donner CF,Wasserman K. Reductions in exercise lactic acidosis andventilation as a result of exercise training in patientswith obstructive lung [see comments]. Am Rev RespirDis 1991;143:9-18.

41 Belman MJ, Mittman C. Ventilatory muscle trainingimproves exercise capacity in chronic obstructive pul-monary disease patients. Am Rev Respir Dis 1980;121:273-80.

42 Belman MJ, Epstein L, Doombos D, Elashoff J, KoernerSK, Mohsenifar Z. Reliability and validity of non-inva-sive detection of the anaerobic threshold in patients withchronic obstructive pulmonary disease. Chest 1992;102:1028-34.

43 Killian KJ, LeBlanc P,Martin DH, Summers E, JonesNL, Campbell EJM. Exercise capacity and ventilatory,circulatory, and symptom limitation in patients withchronic airflow limitation. Am Rev Respir Dis 1992;146:935-40.

945 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from

Belman

44 Lewis MI, Belman MJ. Nutritional supplementation inambulatory patients with chronic obstructive puhnonarydisease. Am Rev Respir Dis 1987;135:1062-7.

45 Mahler DA, Harver A. A factor analysis of dyspnea rat-ings, respiratory muscle strength, and lung fimction inpatients with chronic obstructive pulmonary disease.Am Rev Respir Dis 1992;145:467-70.

46 Mahler DA, O'Connor GT. Impact of dyspnea and physi-ologic function on general health status in patients withchronic obstructive pulmonary disease. Chest 1992;102:395-401.

47 Dudley DL, Glaser EM, Jorganson PN. Psychosocial con-comitants to rehabilitation in chronic obstructive pul-monary disease. Chest 1980;77:413-20.

48 O'Donnell DE, Webb KA. Breathlessness in pataents withsevere chronic airflow limitation: physiologic correla-tions. Chest 1992;102:824-31.

49 Ries AL. Position paper of the American Association ofCardiovascular and Pulmonary Rehabilitation. Scientificbasis of pulmonary rehabilitation. J CardiopldmonaryRehabil 1990;10:418-41.

50 Connors G, Hilling L. Guidelines for pulmonary rehabilita-tion programs. Champaign, Illinois: Human Kinetics,American Association of Cardiopulmonary Rehabilita-tion, 1992.

51 Toshima MT, Kaplan RM, Ries AL. Experimental evalu-ation of rehabilitation in chronic obstructive pulmonarydisease: short term effects on exercise endurance andhealth status. Health Psychol 1990;9:237-52.

52 Cockcroft AE, Berry G. Randomised controlled trial ofrehabilitation in chronic respiratory disability. Thorax1981;36:200-3.

53 McGavin CR, Gupta SP, ILloyd El, McHardy GJR.Physical rehabilitation for the chronic bronchitic: resultsof a controlled trial of exercises in the home. Thorax1977;32:307-1 1.

54 Carter R, Coast JR, Idell S. Exercise training in patientswith chronic obstructive pulmonary disease. Med SciSports Exerc 1992;24:281-91.

55 Casaburi R, Wasserman K. Exercise training inpulmonary rehabilitation. N Engl J Med 1986;314:1509-11.

56 Belman MJ, Mohsenifar Z. Reductions in exercise lacticacidosis and ventilation as a result of exercise training inpatients with obstructive lung disease. Am Rev RespirDis 1991;144:1220-1.

57 Foster S, Thomas HM. Pulmonary rehabilitation inCOPD patients with elevated' Pco,. Am Rev Respir Dis1988;138:1519-23.

58 Niederman MS, Clemente PH, Fein AM, Feinsilver SH,Robinson DA, flowite JS, et al. Benefits of a multi-disciplinary pulmonary rehabilitation program. hnprove-ments are independent of lung function. Chest 1991;99:798-804.

59 Holle RHO, Schoene RB. Increased muscle efficiency andsustained benefit in an outpatient community hospitalbased pulmonary rehabilitation program. Chest 1988;94:1161-8.

60 Punzal PA, Ries AL, Kaplan RM, Prewitt LM. Maximumintensity exercise training in patients with chronicobstructive pulmonary disease. Chest 1991;100:618-23.

61 Carter R, Nicotra B, Clark L, Zinkgraf S, Williams J,Peavler M, et al. Exercise conditioning in the rehabilita-tion of patients with chronic obstructive pulmonary dis-ease. Arch Phys Med Rehabil 1988;69: 118-22.

62 Beiman MJ, Gaesser GA. Exercise training below andabove the lactate threshold in the elderly. Med Sci SportsExerc 1991;23:562-8.

63 Ries AL, Ellis B, Hawkins RW. Upper extremity exercisetraining in chronic obstructive pulmonary disease. Chest1988;93:688-92.

64 Lake FR, Henderson K, Briffa T, Openshaw J, Musk AW.Upper-limb and lower-limb exercise training in patientswith chronic airflow obstruction. Chest 1990;97:1077-82.

65 Couser JI, Martinez FJ, Celli BR. Pulmonary rehabilita-tion that includes arm exercise reduced metabolic andventilatory requirements for simple arm elevation. Chest1993;103:37-41.

66 Martinez FJ, Vogel PD, Dupont DN, Stanopoulos I, GrayA, Beamis JF. Supported arm exercise vs unsupportedarm exercise in the rehabilitation of patients with severerchronic airflow obstruction. Chest 1993;103:1397-402.

67 Keens TG, Krastins IRB, Wanamaker EM, Levison H,Crozier DN, Bryan AC. Ventilatory muscle endurancetraining in normal subjects and patients with cysticfibrosis. Am Rev Respir Dis 1977;116:853-60.

68 Belman MJ, Kendregan BK. Physical training fails toimprove ventilatory muscle endurance in patients withchronic obstructive pulmonary. Chest 1982;81:440-3.

69 Holloszy JO, Coyle EF. Adaptations of skeletal muscle toendurance exercise and their metabolic consequences.IJAppi Physiol 1984;56:831-8.

70 Clausen JP. Effect of physical training on cardiovascularadjustments to exercise in man. Physiol Rev 1977;57:779-815.

71 Belman MJ, Kendregan BK. Exercise training fails toincrease skeletal muscle enzymes in subjects withchronic obstructive pulmonary disease. Am Rev RespirDis 1981;123:256-61.

72 Simpson K, Killian K, McCartney N, Stubbing DG,Jones NL. Randomised controlled trial of weightliftingexercise in patients with chronic airflow limitation.Thorax 1992;47:70-5.

73 Mahler DA, Harver A. Clinical measurement of dyspnea.In: Mahler DA, ed. Dyspnea. Mount Kisco, New YDrk:Futura, 1990, 75.

74 Guyatt GH, Chambers LW. A measure of quality of lifefor clinical trials in chronic lung disease. 7Torax1987;42:773-8.

75 Morgan AD, Peck DF, Buchanan DR. Effects of attitudesand beliefs on exercise tolerance in chronic bronchitis.BMJ 1983;286: 171-3.

76 Agle DP, Baum GL, Chester EH. Multi-discipline treat-ment of chronic pulmonary insufficiency: functional sta-tus at one year follow-up. In: Johnston RF, ed.Pulmonary medicine A Hahnemann Symposium. NewYork: Grune and Stratton 1973:355.

77 Belman MJ, Brooks LR, Ross DJ, Mohsenifar Z.Variability of breathlessness measurement in patientswith chronic obstructive pulmonary disease. Chest 1991;99:566-71.

78 Levine S, Weiser P, Gillen J. Evaluation of a ventilatorymuscle endurance training program in the rehabilitationof patients with COPD. Am Rev Respir Dis1986;133:400-6.

79 Smith K, Cook D, Guyatt GH, Madhavan J, Oxman AD.Respiratory muscle training in chronic airflow limita-tion: a meta-analysis. Am Rev Respir Dis 1992;145:533-9.

80 Paez PN, Phillipson EA, Masangkay M, Sproule BJ. Thebasis of training patients with emphysema. Am RevRespir Dis 1967;95:944-53.

81 Lustig FM, Haas A, Castillo R. Clinical and rehabilitationregime in patients with COPD. Arch Phys Med Rehabil1972;53:315-22.

946 on 17 July 2018 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.936 on 1 S

eptember 1993. D

ownloaded from