Captopril-associated cough is not due to increased airway

BACKGROUND: Angiotensin-converting enzyme (ACE) inhibitors commonly cause cough. There is controversy as to whether preexisting airway hyperresponsivencss predisposes patients to ACE inhibitor-induced cough and whether ACE inhibitor therapy can cause airway hypcrresponsiveness.

BACKGROUND : Angiotensin-converting en,yme (ACE ) inhibitors commonly cause cough.There is controversy as to whether preexisting airway hyperresponsi vencss predisposes patients to ACE inhibitor-induced cough and whether ACE inhibitor therapy can cause airway hypc rresponsiveness.OBJECTIVES : To lest the predictive value of base line airway res pon siveness fur the development of cough induced by AC E inhibitors , and whether patients who developed cough during ACE inhibitor thn:.1ryshowed an increase in airway res ponsiveness.'.\,IETIIODS: Baseline spiromctry and airway res pons iveness tu methacholinc (expressed as the conce ntrat ion of mcthacholine required to cause a rocro fall in forced expi re d rnlumc in I s I FEY I J or PC 10) were measured in 23 hypertt'nsive patients before and afte r IO to l 2 wee ks of treatment 11ith caplopril at an initi:.1!daily dose of 50 mg .
RESULTS : During caplopril therapy se ve n (307a ) paticnh Jcn:loped new coug h, which rcsol vcd after d iscontinuatio n of captopril.The base line PC 10 was the same in coug he rs and noncoughers and did not rrcdicl the dnelop ment of cough .During caplopril therapy the re was a nun sign ificant decrease in the geometric mean PC 10 fo r a ll patients from 15 rng/rn L (ra nge 0.8 to 62) to 13 mg/ml (r:.1ngL' 0.7 to 62) (P=0.08) .These changes were not different in cougher, and noncoughers.FEY 1, forced vital capacity and total lung capacity di d not change during the study.CONCLUSIONS : Cough caused by captopril is not predicted hy baseli ne airway respo nsiveness and is not associated with increased airway res ponsiveness .Kr)' \>Vords: A11gio1e11 si11-com •ati11g e11~y111c.As1h11w.Cough.
Brond 1ial hypcrrcspu11sivencss has been thl' most 'xtensivcly invesl igatcJ potentially prcdispo ... ing fac 1or.but the results of these stud ies arc contl icting.Dixon cl al ( 14) fo und no change in lhe airway r 'sponsivcncss to hislaminc or mcthacholinc of asthmalic ... ubjccts gi \ Cn a single dose of ACE inhi bilor.In another sludy.only two paticnls wi th cough caused hy ACE inhibitors had evidence o f increased bronch ial reactivity ( 10).However, Buck nall et a l ( 15) observed a small but ... ignificant increase in bronchial responsiveness when symptornal ic raticnts were studied during and artcr d iscontin uation o f ACE inhibitors and suggeslcd thal lhcsc pal icnts had increased bronch ial responsiveness al base line.Kaufman et a l ( 16) reported that cough that developed on CE inhibil ion seemed lo occur in pat ients with increased airway responsive ness bolh during and without lreatmcnt and called for a prospect ive study of cough during ACE inhi bitor therapy.Ogihara ct al ( 17) reporled incrcas 'd cough in rcsronsc to c itric ac id and increased brnnchoconslric tion in response to histami ne i 11 normal volunteers trealcd wit h ACE inhibilors.On !he olher hand Sala ct al ( 18) reported no change in mcthacholinc resronsivcness after chron ic ACE inhibitor lhcrapy in asthmatic and nonal-,thmatic subjects.Since severe bronchoconstriction can develop in a subset n r pat ients receiving ACE inhibitor treatment, the abi lity to predict adverse effects would be advantageous.
We have r rospcctively studied th airway resr onsi cness to mc1hacholinc o f _3 hypcrtcnsiv • patients who wl'rc <,tarting therapy with captopri l.We focused on the rdationshir> of cough to baseline a irway respon'>ivencss as well as the rcla-tionshir> of changes in a irway responsiveness during captopri l thcrary to the deve lopment or cough .

PATIENTS AND METHODS
Patient population : Patient:,, who were tal,.ingra11 in a singkblind crossover slud or eaptorril wen.' asl,.cdt\l participalc in the sludy.Patients gave informed consent before parlicipat ing in !he study which was approved by 1hc eth ical re, icw c mmittees ol' St Paul 's Hospita l and the llniversi t 1 of Hri tish o lumbia.All previous antihypcrtcnsi,e 111cdic;1li\1ns had bee n stoprcd al least three weci,.sbefore the initial rcsriratory asscssmcnl.No subject was taking captorril or an _' other ACE inhi bi1 or before be ing recruited, although three or the 236 subjects had been treated in the past with ACE inh ibitor\.
None or them had a hisiory or cough or any other resrirator) d isease.Study protocol: Each subject co mpleted a modified mcrican T horac ic Society questionnaire designed tu e lic it any history or rcspira1ory or atopic disease.Basel ine ru lmon:1ry function, including spiromclry and lung vol umes.were mea,urecl.Spiromctry (forced expired vo lume in I s IFEV1I.fo rced vital capacity iFVC I) was done in a volu me Jisplac-:mcnt.pressure comrcnsatcJ body plethysmograph.Volume was measured using a Krogh sp iromelcr courlcd to a linear d isplacemcnl transducer (Type 300 HR, Shacvit1.Engineering, New krsey) and airtlow was measured using a Flcisch o 3 pncumolacho mctcr coupled to a r ress ure transducer (Sanborn 270, Sanborn ).Partial ~tnd comp lete max imal cxpiralory manoeuvres were r crformcd, T he partial expirator), !low-volume curves were initiated fro m end inspiration during tidal breathi ng and were fo llowed by immcd ialc inhalation to total lung capac ity (TLC) and a complete expiraLOr), !low-volume cu rve.Flow and vo lume signals were •Jigititcd' and fed to an Apple lie computer progra mmed to calcu late FEY I and FYC and to dis play cxpiralory l'lcrnvolumc loops.After three technically adequate trac ings had been obtained, the mean aluc or the th ree measuremen ts wa, calculated and used as the baseli ne FEY 1 and FYC.Mc~bLircmcnts of llow al 50q, and 25CA ur measured vital capacil) (Y111a, 50, Yma, 25 l on the complete !low-volu me curve wcr~ obtained from the d ig it i1cd !'low and volu me data as well a, !lows at an isovo lumc point between 30 and 351k TLC on the partial and complete c.\ piratory llow vol ume curves.Thoracic gas volume was also determ ined at baseline using Boy le's La,v 1ech11iquc foll owed by an i1 bpiratury capaL•ity manocuH c lo calculate TL ' .Mouth pressure was measured usi ng a different ial rrcssure transducer (Val idync 45 MP ±100 cm H:i.O.Va lidync Co, Cali forn ia ) whi le the subjccl was ocdudcd al lhc moulh and panting al I Ht.. Arte r these basl'linc mcasurcm 'nls, patients performed a modi l'ical ion or I he inha lat ion challenge lest described b: Coch-roft ct ,ti ( 19).crosols were ge nerated by a Be nnc11 twin jct ne bu li1cr fHcnnctt) at an a irtlow rate of 6 Umin.llnJc r these u111d itio11s, the ncb ul iLer prod uced an outpul of 0 .13 ml. of .tL'n>sul/m in with a mean part icle site o f 3.6 µ111 and a geometric slanda rd dev iation o f 3.47 µ 111.The aerosol \\as de livered tl inn1gh a thrcc-wa stopcock and rlastic lubing wilh ,t deadspacc of approx imalc ly 200 mL while the patienl inhaled lhrough the mouth d uring quiet tidal breal hing with the nose closed by a clip.Aeroso ls were del i crc<l fur 2 mins each lime.Patients first inhaled an aerosol of isoionic ... al ine fo llowl'd at 5 min interva ls by aerosols of acctyl methy !chol ine ch loride (mclhacholine ) in doubling concc11 t1• a1 ion\ from 0 .5 mg/mL to a maximu m o f 64 mg/ml.T he st ar! ing dose w,ts low red lo 0. 125 mg/mL if tile subject had a hi..,tu1-y uf atopic rhinit i,.Doses were sl,.ippcd if a fall o f lc ... s than 3'.f in FEY I occ urred over I wo \Uccessive dose,.Nn dusc.., \\'L'rc sl,.ippcd arter the 4 mg/mL Jose was admini,tc rcd .Ma.\ ima l fo rced expiratory l'lows were measured al 30, and 2 min\ a fter the eomr lc tion o f each inha lation .
Can Respir J Vol 1 No 4 Winter 1994 rile test was stopped when there was a fall in FEY I of 20"/r or more, or whe n the 64 mg/m l dose of mcth acho li ne had hccn given.
The percentage fal l in FEY I was calcul ated from the hiwe~t FEY 1 obtained after the saline aerosol (FEY 1-A) and alk r methachol im: ( f<EY 1-B ), as follows: II the postsali ne tEY I was less than :w, cli lle ren t from the tia,eline FEY I then the mea n of all four efforts was used as f EV1-A.The results were ex pressed as the concen tra tion o r rnl'thaeho line requi red to provoke a I 0% rai l in FEY 1 (PC I o) . 1 1 hich was ca lculated by interpo lation be tween the two poi nts on ei ther .iclc of a I 0% Lill in FE Y 1. Dose response rel at ion-,hi p~ using maximal !lows at 50 and 25% of FYC from the ,,1rn plctc flo w vo lume c urve and !l ows at the isovo lumc point 1)1 1 part ial and com pl ete cur\'cs were also constructed.The patie nts were lreated with cap topril only (in itia l dail y J1h C 50 mg.max im um daily close I 00 mg) fo r 12 weeks .The patients had weekly assessme nts o r blood pressu re thro ugh-Put the study.The dose o r captopri l was adjusted accord ing 1t1 the pati ent's blood pressure.Ir hypertens ion was uncontrp llcJ or significan t si de e ffect s of treatme nt developed.11ccC)~itati ng withclraw :tl or captopril.the protocol was immediately term inated ( captopril was discontinued at I 0 11 ccks in three pat ie nt ~. one because of poor cont ro l and two hcc;1u~e or coughing ).Ju st hcfore the encl or the 12-weck period.patients returned for anothe r assessrne nt.They were 4uc,1ioned at !hat ti me regarding the occ urre nce of any new rc, pi ra tory sym ptoms.Spi ro metry, measure ment or lu ng volume, and inha lat ion challenge testing we re repeated in an 1JL ' 11tical manner to their li rst visi t. ~uug h threshold fo r inhaled citric ac id was al so measured in :1 subgroup of 13 pati ents (six coughcrs and seven noncnughers) after completi on ol thc initi al pro tocol.All patie nts h.iJ been wit hdra wn rro m captopril the ra py for at lea\t one 111011th at the ti me or the citric acid in halat ion test.Cough thrc, hold w:1 s determ ined as previously described by Tay lor ,'l :ii (20) .In a sing le-bli nd fashion each subject inhaled a ,untrol so lut ion or O.'.'i M sal ine rolluwcd by increas ing cun -,cnli ations (OS I. 2. 4. 6. 8. 12 .16. 24 and 12 gl I 00 ml) ol ,ry,tall ine citric acid mo nohydra te (Fisher Sc ie nti fic.New k r,cy) dissol ved in O. ' .' i M saline.So lut ions were inhaled ,•1cry S mi ns fro m a Puritan Bennett twin jct nebu li7cr, with ,ubjccts carryi ng out a slow inspiratory capac ity manoeuvre ,,1~r S s.The cough thres ho ld was defined as the lowest ,tincentrat ion of citric aci d that consiste ntl y elici ted an invo lunta ry cough during three separate inha lations.lh ta analysis: Data arc prese nted as mean ± SD or as the geometric mea n and ra nge.T he anth ropometric data of the ,nughcrs and noncoughcrs were compared using a no n pai red rt~, t.Lu ng function and log PC 10 be f"ore and during captopril th,Tap were compared using a paired r test.The PC 10 and ,•, ,ugh thres hold values of the two gro ups we re compared u,ing the Wi lcoxon signed ra nks test.Pd l.05 was cons idered

RESULTS
Re\ults from then pat ients who completed the study are su mmari;:cd 111 Tab le I. Seven (30% ) patienh compla ined or new cough during captopril therapy that resolved when captopril was discont inued .On ly one subjec t was an act ive smoker and he did not develop a cough .A much higher inc idence of cough occu rred among l"cma les (s ix of 11 .5571) ve rsus males (one o f 12, 8Ck ).P<CW5.The pe rcen tage of subjec ts who gave a history consiste nt with atof it.: rhinitis was simil ar between coughers and no ncoughcrs.Basel ine FEY 1. FYC.Yrna, 50.Ynrn 25 and TLC were no rma l in bot h groups.one nf these mea'>ure ments we re different during therapy.
Figu re I shows both se ts of measure me nt\ of PC 10 for the two groups.The base li ne geometric mean PC II) 1 •or the 2.\ su bjects was 15 mg/ml ( ra nge 0.8 to 62 mg/ml ) and :tfk r treatme nt with cap topril was 13 mg/ml (ra nge 0 .7 to 62 mg/ml) ( P<! U)8 ).The re was no di llercncc betwee n the baseline PC to or the coughcrs and no ncoughc rs: there was nu differe nce in the change~ in bronc hial responsiveness between the coughers and no ncoughe rs.
The re was no differe nce in the shape or position of the dose response c ur ves construc ted using met hacho linc concen tration and pa rti al or com pl ete max imal !low s at low lu ng volumes between coughers and no ncoug hcrs or in coughers before and a rtcr captopril therapy .Th irtee n pat ients (seven noncoughers.six coughers ) comp leted the ci tric :icid inhalation test.As shown in Figure 2. the geometric mean cough

DISCUSSION
The results of this study demonstrate tlwt captopril-in .. duced cough is 1101 predicted by increased bas1::line airway responsiveness.The d,1ta also show that caplopril-induced cough is not associated with an increase in airway responsi veness during 12 weeks of trt'atment.We found a high incidcnCL' of captopril-induced cnugh (30'k ) with a marked preponderance in womL•n.
The cough reported in association with ACE inhibitors is unique among respiratory side effects of drugs .The lung is a common site for drug-re lated side effects (2 1 ).but coug h apparently without changes in the lung parenchy ma or a irway obstruction is not assoc:iated with other classes of drugs .ACE inhibitors have been implicated in the worse ning of asthmatic symptoms (il.22.23) and i11 parcnehynnl lung disease (24 L but the.e reports are rare compared w ith the m11nber or reports on the occurrence of isolated cough.The mechanism of' the cough rema ins unk.no wn .W e chose to study the relationship of cough induced by ACE inhibitors to airway hype1Tesponsiveness because of the implications of this potent ial rel:.itionship on the trc:.itment of hypertension and/or congesti ve heart failure in asthmatic patients. W e used PC to as the marker of airway responsiveness rathe r than the more commonly used PC 20 in order to obtain a measured value for bronchial responsiveness in a large r percentage or patients.We expected to he able to measure PC20 in no more than 20% or patients \Vithout a previous diag nosis of asthm:.ior other respiratory disease: this may ha ve prevented inte rpretation of the results in a large number of subjects and excluded patients with normal bronchial responsiveness .PC 10 is a convenient method or expressing the responsivene ss of the airways, although it may not have the same implicatio ns for clinical disease as does PC20.The  The inc ide nce of cough (3()1/r ) we observed is at the upper end of the range previously reported (7,25 .26).O ne possible expl ,u1ation for this is that we were specifically asking about the occu rrence of couoh, and the patients were aware that cough was a potential side effect o f captopril.In addition , a rel a tivel y high proportion of our pa tients was female , a group of patienb previously demonstrated to have a hi g her incidence of ACE inhibitor-induced cough (26.27).Our findin g~ support these earlier reports.Finally, not all patients who report coug h initially while on ACE inhibitor therapy repon cou gh again when rcchallengcd with the drug ( 10).
Results from prc\'ious studies of ACE inhibitors, in whi ch investigators measured airway respo nsiveness.are confl ic ting.T he results of two studies ( 15.16) showed increascu airway responsiveness in patients who had a hi story of AC E inhibitor-induced cough hut were no longer taking A E inhibitors.One or thcsL'.studies reported a small but signilicant increase in airway respo ns ive ness when their patient, were rcc:hallcngcd with ACE inhibitors ( 15). while the other study did not show consistent ch anges ( I (i).Both stud ic~ were retrospective investigation s o f patient s pre vio usly identified as ~usc:eptiblc to ACE inhibitor-induced coug h.J\ thiru retrospective study (28) demonstrated increased histamine responsiveness during treatment with cnalapril which wa~ not prc~ent in a placebo-treated g roup.Thi~ study lacked a control group or noncoughcrs.In a study of normal volunteers.Ogihara ct al ( 17) reported increased cough in res po n~e to citric acid and increased bronchoconstriction in response to histamine following captopril or enalapril the rapy.
In contrast, other studies have reported no effect of ACE inhibitor~ on airway responsi veness.J\ single dose of ramipril did not alter the ai rway responsiveness of asthmati c ~ubjccts ( 14 ).Sala ct al ( 18) l' L' portcd 1h l change in methacholine responsiveness after chronic captopril treatment in 16 hypertensive asthmatic subjects and 15 hypcrten-,iYe nonast hmatics.In another study.only two of seven patients with ACE inhibitor-induced cuugh had increased airnay responsiveness at basel ine and nu chan ges in airway re~ponsiveness were observed during therapy ( 10).Boulet et al (29) reported no significant change in lung function or airway responsiveness during eight weeks or treatment with ~aptopril in 15 patients, four or whom developed cough.However, there was a trend towards increased responsiveness in patients on captopril.and this study may have been weakened by the inability tu qu.intify the degree or airway respon-,il'eness (as judged by ho) at either visit in six of their 15 ,uhjccts.These authors raised the possibility that murc sensi-1i1e tests of airway responsiveness might detect ch anges that their study had missed.Our results, using f1uws at low lung 1olurnes and bdure a dee p inhalation.do not support this h~ pothesis.We did not find a significant change in the airway respon-1i1eness of our patients treated with captopri l for 12 weeks.We considered that a change l)f one doubling dl>sc in the airway res ponsiveness or our patients would be clinically ,igniticant.The puwer of our study to exclude such a clifferenc' I? in the entire patient group (n=23) exceeds 95 % and in l>llr grou p of coughcrs (n= 7) exceeds 90%.It is poss ible that a small change in airway responsiveness occurs during ther-ap~ with ACE inhibitors.Previous reports in the literature 115.17.28,29) have suggested this possibility.and the change ,w1 in th is study approached but did nut reach stat istical ,ignificance (P=0.08).Such a small change in airway rcspun-1i1enes, is unli ke ly to be clinically important.
Furthermore, it is clear rrum our findings that there is no relat ionship between the development of cuugh and inffeased bronchial responsiveness, e ither at baseline or during 12 weeks therapy with captopril.Figure I demon strates the range of PC 10 values measured.Clearly the range of PC 10 \CKNOWLEDGEMENTS: SupportcJ by the Medical Resea rch Cl>uncil of Canada.Dr Black ie was a Fellow of the Canauian Lung A"ociation.Dr Rangno was supported hy the BC Hearl Fou ndat ion val ues of the coughers is not lower than that of the noncoughers.Measurement of a patient's methacholine responsiveness before receiving the rapy with captopril would not he lp a cli nician to predict the de velopment of cough.Because captopril treatment did not increase non specific airway responsiveness, these data suggest that sym ptoms of asthma should not worsen during captopril therapy.However.asthmatic symptoms do develop in a small proportion of pati ents who receive ACE inhibitor the rapy, and preexisting airway disease is a predictor of these rare adverse responses (8).
Prev ious studies have examined the relationship between cough threshold and ACE inhibitor-induced cough.Fuller and Choudry (30) demonstrated an increased cough response tu capsaiein inhalation in five patients who had developed ACE inhibitor-induced cough.Testing after stopping the drug showed that the cough sensitivity had returned to normal levels.The res ults of a subsequent study from this group (JI) showed that baseli ne sensitivity to eapsaicin is normal in pati ents who develop cough while tak ing ACE inhibitors.Morice et al (32) al so showed that captopril significantly shifts the cough Jose response rnrvc to capsaicin inhalation in normal subjects.In this study.there was no change in cough frequency afte r inhalation of citric acid, but dose respon se curves for nei the r citric acid nor thresholds for cough were assessed.Fro m our small study population (n= 13) it appears that.citric acid sens itivity is not d iffere nt in coughcrs from nUJtcoughers.Whether the re is any change in cough threshold to citric ac id during ACE inh ibitor therapy is unknown.

CONCLUSIONS
This prospective study showed that baseline airway responsive ness does not pred ict the development of captopril-inducL'd cough.As we ll , captopril-induccd cough is not associated with an increase in airway responsiveness .Cough is a comnwn sidl' effect of ACE inhibitors and women are at increased risk or this cUJnplication.

Can Respir J Vol 1
No 4 Wi nter 1994 Captopri l an d cough

TABLE 1 Data at baseline and on captopril Noncoughers (n =16) Coughers (n=7)
' The only significant difference between the coughers and noncoughers is the female .maleratio (P<0.05) .Th ere were no differences in any of the measurements between the two visits.t Reference 33; t Reference 34.FEV1 Forced expired volume in 1 s; FVC Forced vital capacity; TLC Total lung capacity; PC1 0 The concentration of methacholine required to cause a 10° o fall in FEV1 .All va lues except PC,o are expressed as mean ± SD.PC,o is expressed as the geome tric mean and the numbers within brackets represent the range significant.All calcu lations, inc ludi ng power ana lysi~.were do ne using Systat ( Illino is).