to - thoraxthromboxane in asthma is less clear and has been probed with non-steroidal...

Thorax 1996;51:799-804

Comparison of three inhaled non-steroidalanti-inflammatory drugs on the airway response tosodium metabisulphite and adenosine5'-monophosphate challenge in asthma

Millie Wang, Antoni Wisniewski, Ian Pavord, Alan Knox, Anne Tattersfield

Respiratory MedicineUnit,City Hospital,Hucknall Road,Nottingham NG5 IPB,UKM WangA WisniewskiI PavordA KnoxA Tattersfield

Correspondence to:Dr A Knox.

Received 23 October 1995Returned to authors22 December 1995Revised version received7 March 1996Accepted for publication7 March 1996

AbstractBackground - Non-steroidal anti-inflammatory drugs (NSAIDs) are used toassess the role of prostaglandins inasthma but their effects on bronchocon-strictor challenges have been inconsistent.The effects of three nebulised non-steroidal anti-inflammatory drugs on theairway response to inhaled sodium meta-bisulphite (MBS) and adenosine 5'-monophosphate (AMP) were compared inthe same asthmatic subjects to seewhether contractile prostaglandins wereinvolved in MBS or AMP induced bron-choconstriction. A possible protectiveeffect of the osmolarity or pH of theinhaled solutions was also assessed.Methods - Two double blind placebo con-trolled studies were carried out. In study1, 15 non-aspirin sensitive patients withmild asthma attended on four occasionsand inhaled 5 ml of lysine aspirin (L-aspirin) 900 mg, indomethacin 50 mg,sodium salicylate 800 mg, or saline 20minutes before an inhaledMBS challenge.On four further occasions 14 of thepatients inhaled the same solutions fol-lowed by an inhaled AMP challenge. Instudy 2, 10 ofthe patients attended on fouradditional occasions and inhaled 5 ml of0.9%, 3%, 10%, or 9.5% saline withindomethacin 50 mg 20 minutes before aninhaled MBS challenge.Results - In study 1 inhaled lysine aspirinhad a similar effect on MBS and AMPinduced bronchoconstriction, increasingthe provocative dose causing a 20% fall inFEV1 (PD20) by 1.29 (95% CI 0.54 to 2.03)and 1.23 (95% CI 0.53 to 1.93) doublingdoses, respectively. Indomethacin in-creased the MBS PD20 and AMP PD20 by0.64 (95% CI -0.1 to 1.38) and 0.99 (95% CI0.29 to 1.69) doubling doses, respectively.Sodium salicylate had no significant effecton either challenge. The two solutionscausing most inhibition were the mostacidic and the most alkaline. In study 2inhaled 9.5% saline with indomethacin(osmolarity 3005 mOsm/kg) increased theMBS PD20 by 1.1 doubling doses (95% CI0.2 to 2.0) compared with only 0.09 (95%CI -0.83 to 1.0) and 0.04 (95% CI -0.88 to

0.95) doubling doses with 3% saline (918mOsm/kg) and 10% saline (2994 mOsmlkg), respectively.Conclusions - Inhaled L-aspirin andindo-methacin have broadly similar protectiveeffects against MBS and AMP inducedbronchoconstriction in the doses given,although the effect of indomethacin onMBS was not quite statistically signifi-cant. The osmolarity and pH of thesolutions did not appear to be importantdeterminants of the response. The effectof L-aspirin and indomethacin is likely tobe the result of cyclooxygenase inhibitionreducing the production of contractileprostaglandins during MBS and AMPchallenge.(Thorax 1996;51:799-804)

Keywords: asthma, non-steroidal anti-inflammatorydrugs, prostaglandins.

The cyclooxygenase products of arachidonicacidmetabolism include severalpotentbroncho-active prostaglandins (PGs) and thromboxane.Inhaled PGF2Q and PGD2 cause bronchocon-striction in patients with asthma' whereasPGE2 protects against challenges that act indi-rectly on the airways including sodium metabi-sulphite,' exercise, ultrasonically nebulised dis-tilled water,3 4and both the early and lateresponses to allergen.5The role of endogenous prostaglandins and

thromboxane in asthma is less clear and hasbeen probed with non-steroidal anti-inflammatory drugs (NSAIDs) which inhibittheir production by the cyclooxygenase path-way and with thromboxane receptor antago-nists.6 Ingestion and inhalation of NSAIDscause bronchoconstriction in a smallproportion of patients with asthma and, veryoccasionally, bronchodilatation. In most pa-tients, however, they cause little change in rest-ing lung function, suggesting that endogenousprostaglandins may not be important. Otherstudies, however, suggest that both constrictorand protective prostaglandins are active inasthmatic airways. Inhaled lysine aspirin, indo-methacin, and sodium salicylate inhibit the lateresponse to allergen7-' and lysine aspirin and

799

on April 22, 2020 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.51.8.799 on 1 A

ugust 1996. Dow

nloaded from

http://thorax.bmj.com/

Wang, Wisniewski, Pavord, Knox, Tattersfield

indomethacin protect against ultrasonicallynebulised distilled water induced bronchocon-striction,"' suggesting that bronchoconstrictortone in these circumstances is prostaglandindependent. Inhibitory prostaglandins such as

PGE2 appear to have a role in the developmentof refractoriness following indirectly actingchallenges since this is inhibited by NSAIDs,and there is some evidence to suggest that theymediate the effects of frusemide although thishas been less consistent. The protective effectof frusemide on bronchoconstriction followingexercise was attenuated by oral indomethacinin one study" but augmented by inhaled lysineaspirin'2 and oral flurbiprofen'3 in others.Thus, there is evidence that both constrictorand protective prostaglandins play a part indetermining airway calibre under certain cir-cumstances, although there are many unex-

plained inconsistencies in the literature.One problem when comparing studies on

the effects of NSAIDs in asthma is the consid-erable variation in the drugs given, doses,routes of administration, osmolarity and pH ofthe inhaled solutions, and the bronchialchallenge tests used to assess the response.

NSAIDs are a non-homogeneous group ofchemical compounds with varying effects on

the two cyclooxygenase isoenzymes currentlyidentified, COX-1 and COX-2." Discrepan-cies in the literature may therefore be due todifferences in the drugs, the methodologiesused, or to differences in osmolarity or pH ofthe inhaled solutions.The bronchoconstriction caused by sodium

metabisulphite is thought to act via release ofsulphur dioxide (SO2)"' and activation of neu-ral pathways.'6 7 In contrast, the broncho-constriction caused by inhalation of adenosineand its related nucleotide, adenosine 5'-monophosphate (AMP), appears to be largelydue to histamine release from primed airwaymast cells;'8 19 selective HI-receptor antago-nists such as terfenadine inhibit the response toAMP20 in patients with asthma but not theresponse to sodium metabisulphite.2' Theeffect of inhaled NSAIDs on the airwayresponse to inhaled MBS and AMP has notbeen studied. We have therefore compared theprotective effect of three nebulised NSAIDsagainst inhaled MBS and AMP induced bron-choconstriction in asthmatic subjects who were

not sensitive to aspirin to see whether contrac-tile prostaglandins were involved in these chal-lenges. We also assessed whether the osmolar-ity or pH of the inhaled solutions were relatedto any protective effect.

MethodsSUBJECTSFifteen non-smoking subjects (one woman)aged 21-51 years with mild asthma and no his-tory of aspirin induced asthma were studied.Twelve were atopic (two or more positive skinprick tests to common allergens). All hadasthma that was well controlled, a forcedexpiratory volume in one second (FEVy)greater than 70% predicted, and no evidence ofbronchoconstriction in response to inhaledlysine aspirin in the laboratory (see below). All

were taking an inhaled P2 agonist as required,six were taking inhaled beclomethasone (100-200 ztg per day), and one took oral loratadineas required for rhinitis. Inhaled f agonists werewithheld for at least 10 hours and oralloratadine for at least 48 hours before eachchallenge. Subjects were asked to use onlyparacetamol for pain relief if required duringthe study period. The study was approved bythe City Hospital ethics committee and allsubjects gave written informed consent.

MATERIALSThe drugs and chemicals used in the studyincluded lysine aspirin (Laboratoires Synthe-labo, Paris, France), indomethacin meglumine(Liometacen, Chiesi Farmaceutici, Parma,Italy), sodium salicylate (E Merck, UK), saline= 0.9% (wt/vol) sodium chloride (SteripakLtd, UK), 30% sodium chloride (MartindalePharmaceuticals Ltd, Essex, UK), sodiummetabisulphite (MBS) (Thornton Boss Ltd,UK), and adenosine 5'-monophosphate(AMP) (Sigma, Poole, Dorset, UK).Forced expiratory volume in one second

(FEVy) was measured with a dry wedgespirometer (Vitalograph, Buckingham, UK)and the higher of two successive readingswithin 100 ml was recorded.Sodium metabisulphite and adenosine 5'-

monophosphate were dissolved freshly in salineto produce a doubling concentration range of0.6-160 mg/ml for sodium metabisulphite and3.12-200 mg/ml for adenosine 5'-monophosphate. Solutions were delivered by anebuliser attached to a breath actuated dosi-meter (Mefar, Brescia, Italy) at a nebulisationtime of one second, pause time six seconds,pressure 22 lb/m (152 kPa), output 10.3 jl and12.8 ,ul per puff for MBS and AMP, respec-tively. Normal saline was inhaled initially withthe subjects inspiring slowly through themouthpiece from functional residual capacityto total lung capacity over three seconds afterthe nebuliser was triggered and then breathholding for three seconds. FEV, was measuredafter two minutes. If the response to saline wasless than a 10% fall from baseline FEV1,subjects inhaled doubling doses of MBS orAMP until a 20% or greater fall in FEV, wasrecorded or the maximum dose of MBS (135imol) or AMP (1 18 gmol) had been given.

PROTOCOLScreening studyAn initial inhaled aspirin challenge was per-formed using a disposable nebuliser (DeVil-biss, Somerset) according to the method ofPhillips et al.22 1800 mg lysine aspirin (L-aspirin) powder (equivalent to aspirin 1000mg) was made up freshly each day in 5 mlsaline to produce a 360 mg/ml L-aspirin solu-tion. This solution was diluted further in salineto produce doubling concentrations from 5.6to 360 mg/ml. Three ml of each solution wasplaced in the nebuliser which was driven bycompressed air at 8 1/min (nebuliser output0.41 ml/min). Subjects wore a nose clip andinhaled five breaths of normal saline, inspira-

800



j.com/


ugust 1996. Dow

nloaded from


Effect ofNSAIDs on airvay responses in asthma

tion slowing from functional residual capacityto total lung capacity followed by a threesecond breath hold. FEVy was measured fiveminutes later and followed by 10 breaths ofL-aspirin at 15 minute intervals in doublingconcentrations starting at 5.6 mg/ml. FEV,was measured five, 10, and 15 minutes aftereach concentration. The test was discontinuedif the FEVy fell more than 10% after saline orL-aspirin compared with the value after saline.

Subjects only entered the two main studies ifFEVy fell by less than 10% with the highestdose of L-aspirin (360 mg/ml). Both studieswere double blind and placebo controlled.

Study 1: Effect of three NSAIDs on the airwzayresponse to MBS andAMPSubjects attended for eight visits at the sametime of day, each visit separated by at least 72hours. After resting for 15 minutes they inhaledone of the following drugs dissolved in 5 mlsaline and made up freshly by a second investi-gator: 900 mg L-aspirin (equivalent to 500 mgaspirin), 77.2 mg indomethacin meglumine(equivalent to 50 mg indomethacin), 800 mgsodium salicylate, or placebo (saline). Thesolutions were administered using a jet neb-uliser (Unineb, Unimed (UK) Ltd, Dorset,UK) and mask driven by compressed air at 81/min and run by regular breathing until thenebuliser mist was no longer visible (output0.36 ml/min). The same nebuliser was usedthroughout the study. The pH and osmolarityof the solutions were measured before adminis-tration by a pH meter (M82 Radiometer A/S,Copenhagen, Denmark) and an osmometer(Osmomat 030, Gonotec GmbH, Berlin, Ger-many). FEV, was recorded before and at two,five, 10 and 20 minutes after treatment. At 20minutes patients underwent an inhaled MBSchallenge during the first four visits and aninhaled AMP challenge on the second fourvisits.

Study 2: Osmolarity studyTen patients from the first study attended onfour further occasions. 30% sodium chloridewas diluted in sterile water to produce concen-trations of 3%, 10%, and 9.5% sodiumchloride. Subjects inhaled 5 ml of each of thefollowing: 0.9% saline (292 mOsm/kg), 3%saline (918 mOsm/kg), 10% saline (2994mOsm/kg), or 9.5% saline with indomethacin50 mg (3005 mOsm/kg) 20 minutes before aninhaled MBS challenge as described above.

Based on our previous data, the study had95% power to detect a difference of 1.4doubling doses at the 5% level of significance.

DATA ANALYSISTwo way analysis of variance (ANOVA) wasused to compare baseline values of FEV, andthe change in FEV, at two, five, 10, and 20minutes after each drug. The dose of broncon-strictor agent causing a 20% fall in FEV, frombaseline (PD20 FEVy) was calculated by linearinterpolation of the log dose/response curve. Ifthe fall in FEV, was less than 20% after the

Table 1 pH and osmolarity of the solutions placed in thenebuliser in studies 1 and 2

OsmolaritypH (mOsmlkg)

Study 10.9% saline 5.80 295Lysine aspirin 5.18 1734Sodium salicylate 6.39 2100Indomethacin 7.92 463

Study 20.9% saline 5.80 2923% saline 6.50 91810% saline 6.80 29949.5% saline + 8.20 3005indomethacin

maximum dose of 135 jsmol MBS or 118 gmolAMP the curve was extrapolated by onedoubling dose to provide an estimated PD20. Ifthe estimate was greater than the next doublingdose this value (270 or 235 jmol) was used asa censored value. This occurred on 11 of 60occasions after MBS challenge and on 14 of 56occasions after AMP challenge. The distribu-tion of censored values after saline, L-aspirin,sodium salicylate, and indomethacin were two,three, two, and four occasions, respectively, forMBS and one, five, four, and four occasionsafter AMP (for MBS one subject had censoredvalues after all four solutions and one after allbut sodium salicylate; for AMP one subjecthad censored values after all four solutions andtwo after all three NSAIDs). A censored valuewas used on five of 40 occasions of the PD20MBS in study 2 (one, three, and one occasionafter 0.9% saline, indomethacin with 9.5%saline, and 3% saline, respectively; one subjecthad a censored value after all three solutions).PD20 values were log transformed and are

presented as geometric mean values. The logvalues were compared using ANOVA. Themean difference in PD20 values relative tosaline are expressed as doubling doses with95% confidence intervals (CI). The differencein doubling doses was calculated as: [log MBSor AMP PD20 (drug)] - [log MBS or AMPPD20 (saline)]/log 2.

Results were calculated for all values includ-ing censored values. For study 1 we repeatedthe analysis excluding the two subjects whohad a censored value on at least three of thefour occasions. The results given refer to theanalysis including all censored values unlessotherwise stated.

ResultsStudy 1: Effect of three NSAIDs on PD20 MBSand PD20AMPThe drugs were well tolerated apart fromsodium salicylate which caused short livedcough in all subjects and prevented twosubjects from providing an FEV1 measurementtwo minutes after treatment. Only 14 of the 15subjects were available for the AMP study. ThepH of the solutions in the nebuliser rangedfrom 5.18 (L-aspirin) to 7.92 (indomethacin)and the osmolarity from 295 (saline) to 2100mOsm/kg (sodium salicylate) (table 1).Mean baseline FEV, did not differ signifi-

cantly between the four MBS or four AMP

801



j.com/


ugust 1996. Dow

nloaded from



1.6

1.2A

0.8

ini0.4UL)0

N

CL

< 1.6

0.

0~

Fiur0 (SE)L-aspirin S-salicylate Indomethacin900 mg 800 mg 50 mg

Figu1re Mean (SE) difference in PD20 MBS (above)and PD20AMP (below) after inhalation of lysine (L)aspirin 900 mg, sodium (S) salicylate 800 mg, andindomethacin 50 mg relative to 0. 9% saline. *p < 0. 05.A represents the difference for all subjects included.

study days. There was a small fall in FEV, frombaseline following inhalation of all the NSAIDsand saline ranging from 1.2% to 5.5%. Thiswas maximum at two minutes and hadreturned to 0-2% of baseline by 20 minutes.There were no significant differences betweenthe four treatments at any time after inhalation.PD20 MBS was higher after all three NSAID

treatments relative to saline (fig 1) although theincrease was only statistically significant forL-aspirin. When compared with the PD20 val-ues following saline, PD20 MBS increased by1.29 doubling doses with L-aspirin (95% CI0.54 to 2.03), by 0.64 doubling doses withindomethacin (95% CI -0.1 to 1.38), and by0.63 doubling doses with sodium salicylate(950% CI -0. 1 1 to 1 .38). There was no signifi-cant difference in PD20 MBS values betweenthe three drugs.PD20 AMP was significantly higher after

L-aspirin and indomethacin (fig 1). When

1.6 T

1/)a)c" 1.20-0

CD

D 0.80-a

0

EL 0.4'1

09.5% saline + 3% salineindomethacin

Figure 2 Mean (SE) difference in PD20 Ainhalation of3% saline, 100% saline, and 9. <indomethacin relative to 0. 9% saline. *p <

with 0. 9% saline.

compared with PD20 values following saline,PD20 AMP increased by 1.23 doubling doseswith L-aspirin (95% CI 0.53 to 1.93), by 0.99doubling doses with indomethacin (95% CI0.29 to 1.69), and by 0.31 doubling doses withsodium salicylate (95% CI -0.39 to 1.01). Thedifference in PD20 AMP between the threedrugs was significant (p < 0.05).When the two subjects with three and four

censored values were excluded the PD20 MBSvalues increased to 1.49, 0.84, and 0.79doubling doses after L-aspirin, indomethacin,and sodium salicylate, respectively, the increasefor indomethacin being then statisticallysignificant (p = 0.05) and that for sodiumsalicylate being almost significant (p < 0.1).The changes to the PD20 AMP values after thesubjects with censored values were excludedwere minor and did not alter the statisticalsignificance of the findings.

Study 2: Osmolarity studyAll four solutions were well tolerated. The pHranged from 5.8 (0.9% saline) to 8.2 (indo-methacin with 9.5% saline) and the osmolarityfrom 292 (0.9% saline) to 3005 mOsm/kg(9.5% saline with indomethacin) (table 1). Themean baseline FEV, did not differ significantlybetween the four study days.There was a small increase from baseline in

FEV, following inhalation of normal saline(0.4%) and a small decrease following theother three solutions (range 1.8-5.7%) causinga difference between the four solutions whichwas statistically significant at two minutes afterinhalation (p = 0.05) but not at five, 10, or 20minutes.There was a significant difference in MBS

PD20 between the three hyperosmolar solu-tions (p < 0.05). When compared with thePD20 values following 0.9% saline, MBS PD20increased significantly by 1.1 doubling dosewith indomethacin dissolved in 9.5% saline(95% CI 0.2 to 2.0). The 3% and 10% salinesolutions had little effect, altering MBS PD20by 0.09 (95% CI -0.83 to 0.1) and 0.04 (95%CI -0.88 to 0.95) doubling doses, respectively(fig 2).No significant difference in mean PD20

MBS values after saline was seen in the 10patients who took part in studies 1 and 2 (geo-metric mean 4.9 [tmol and 7.5 jmol, respec-tively) or in the mean difference in PD20 valuesrelative to saline after indomethacin alone instudy 1 (0.93 doubling doses) and indometh-acin with 9.5% saline in study 2 (1.1 doublingdoses).

DiscussionWe have compared the effects of L-aspirin,indomethacin, and sodium salicylate againsttwo bronchoconstrictor challenges with differ-ent modes of action in patients with mild

10% saline asthma who were not aspirin sensitive. Mostpatients with asthma show bronchial hyper-

4BS after responsiveness to a large number of stimuli.5% saline with0.05 compared These may act directly on airways smooth

muscle or indirectly by stimulating neural

802



j.com/


ugust 1996. Dow

nloaded from


Effect ofNSAIDs on airway responses in asthma

reflexes (for example, metabisulphite) or bycausing inflammatory cell mediator release (forexample, exercise, AMP).Three NSAIDs (L-aspirin, indomethacin,

and sodium salicylate) were given by the same

route to the same patients. None caused a

change in FEV, compared with saline in study1, confirming the absence of aspirin sensitiveasthma in these patients. Pretreatment withinhaled L-aspirin (900 mg) protected againstbronchoconstriction induced by MBS andAMP when compared with saline, causing a

1.3 and 1.2 doubling dose increase in PD20values, respectively. Inhaled indomethacin (50mg) shifted the dose-response curves of MBSand AMP challenge to the right by 0.6 and 1.0doubling doses, respectively, although only thechange with the 1.0 doubling dose was statisti-cally significant. The small changes in PD20MBS and AMP after inhaled sodium salicylatewere not significant (0.6 and 0.3 doublingdoses, respectively). The protective effect ofinhaled indomethacin against MBS challengemay have been underestimated as four of thePD20 values after indomethacin were censored.When results from the two subjects with threeor more censored values were excluded fromthe analysis the increase in PD20 MBS (0.84doubling doses) was statistically significant.Sodium salicylate is a weak cyclooxygenase

inhibitor compared with both aspirin andindomethacin,21-25 and inhaled L-aspirin andindomethacin were active in our study whereassodium salicylate was not. We have consideredthat the protection is due to cyclooxygenaseinhibition.

In this study L-aspirin 900 mg and indo-methacin 50 mg had broadly similar effects on

the response to MBS and AMP - two broncho-constrictor agents with different mechanismsof action. The same doses of L-aspirin andindomethacin caused a similar increase inultrasonically nebulised water PD20 (2.1 and1.6 doubling doses) compared with placebo inpatients with asthma studied by Bianco.'0 If thetwo drugs are causing protection through COXinhibition, both MBS and AMP induced bron-choconstriction must be dependent on releaseof contractile prostaglandins (or potentiated byprostaglandins in the airways) to a roughlysimilar extent. Sodium metabisulphite isthought to activate sensory nerves.2' 26 PGD2can enhance cholinergic neurotransmission.27It has been shown that inhaled ipratropiumbromide has a protective effect on the airwayresponse to inhaled PGD228 which suggeststhat PGD2 augments the parasympathetic con-

tractile response prejunctionally and this islikely to involve the accelerated release ofacetylcholine at the neuromuscular junction.27It follows that PGD2 could potentiate the neu-

ral response to sodium metabisulphite. Inhibi-tion of COX would reduce PGD2 productionand thereby reduce the response to MBS. Ithas been shown that exogenous histamineinduces release of PGF2Q in the airways invitro.29 30 It is feasible that AMP induced hista-mine released from mast cells may producesome of its bronchoconstrictor effect by releas-ing contractile cyclooxygenase products. Thus,

COX inhibitors could attenuate AMP inducedbronchoconstriction by this mechanism.

If the effect of aspirin and indomethacin isdue to COX inhibition, their relative effects onMBS and AMP challenges should reflect theirrelative potencies as COX inhibitors. Cyclo-oxygenase exists as two isozymes, COX-1 andCOX-2, COX-1 being a constitutive enzymeinvolved in cellular "housekeeping" functions3'whereas COX-2 is an inducible enzyme whichis increased in the presence of inflammation32 33and is responsible for the production of proin-flammatory prostaglandins. When the relativepotency of aspirin and indomethacin in inhib-iting cyclooxygenase has been studied inbroken cell, purified enzymes and intact cellpreparations both drugs have been morepotent inhibitors of COX-1 than COX-2,although the ratio of COX-1 to COX-2 inhibi-tion has differed depending on the preparationstudied.34 It is difficult therefore to predict therelative potency of aspirin and indomethacin ininhibiting the two COX enzymes in the airwaysother than to say that both drugs are likely tobe more selective for COX-1.

If the protective effect of NSAIDs in asthmais due to COX-2 inhibition, larger doses wouldbe needed and this is more likely to be achievedby inhalation. High doses of inhaled NSAIDshave caused consistent inhibition of broncho-constrictor responses as did oral indomethacinand flurbiprofen when given before adenosineinduced bronchoconstriction.3 6 Oral indo-methacin (50 mg three times daily) for threedays did not protect against MBS inducedbronchoconstriction,37 however, in contrast tothe findings with inhaled indomethacin. Thelower concentration of indomethacin in theairways following oral administration may nothave been in the millimolar range needed toinhibit COX-2.'4 Induction of COX-2 has notyet been demonstrated in mast cells from asth-matic patients, however.We also considered whether production of

1 5-HETE (1 5-hydroxy eicosatetraenoic acid)might contribute to the protective effect ofaspirin since it increases in response to aspirinin cultured ovine tracheal epithelial cells'4 andcan inhibit the 5-lipoxygenase pathway inneutrophils and T lymphocytes.38 Since it isnot increased with indomethacin'4 it is unlikelyto play an important part in the protection seenin our study.

In study 2 all three hyperosmolar solutionscaused a small transient decrease in FEV, frombaseline compared with 0.9% saline whichreturned to placebo values after two minutes.Inhaled indomethacin (50 mg) with 9.5%saline protected against MBS challenge andthe magnitude of this effect (1. 1 doublingdoses) was similar to that seen in study 1 withthe same dose of indomethacin in 0.9% salinein the same subjects (0.9 doubling doses). The3% and 10% saline solutions had no effect.The osmolarity of the solutions does not there-fore appear to be an important determinant ofthe protective effect of indomethacin. Hyper-osmolar saline induced bronchoconstrictionhas been shown to cause refractoriness40 butwhether cross refractoriness occurs to affect

803



j.com/


ugust 1996. Dow

nloaded from



other bronchial provocation challenges has notbeen studied. The lack of effect of hyperosmo-lar saline on subsequent challenge by MBS inour study may be because the hyperosmolarsolution caused minor bronchoconstrictiononly.The two solutions causing most inhibition

were the most acidic (L-aspirin) and the mostalkaline (indomethacin), which suggests thatpH was not a major determinant of the protec-tion seen.Thus, inhaled L-aspirin and indomethacin

have broadly similar effects against MBS andAMP induced bronchoconstriction in thedoses given, although the effect of indometh-acin on MBS was not quite statistically signifi-cant. The osmolarity and pH of the solutionsdid not appear to be important determinants ofthe response. The effect of L-aspirin andindomethacin is likely to be due to cyclooxygen-ase inhibition which reduces the production ofcontractile prostaglandins during MBS andAMP challenge.

The authors thank S Pacey for coding and randomising thedrugs and J Williams for help with subject recruitment.

1 Hardy CC, Robinson C, Tattersfield AE, Holgate ST. Thebronchoconstrictor effect of inhaled prostaglandin D innormal and asthmatic men. N EnglJ Med 1984;311:2b9-13.

2 Pavord ID, Wisniewski A, Mathur R, Wahedna I, Knox AJ,Tattersfield AE. Effect of inhaled prostaglandin E2 onbronchial reactivity to sodium metabisulphite and metha-choline in patients with asthma. Thorax 199 1;46:633-7.

3 Pasargiklian M, Bianco S, Allegra L, Moavero NE, PetrigniG, Robuschi M, et al. Aspects of bronchial reactivity toprostaglandins and aspirin in asthmatic patients. Respira-tion 1977;34:78-91.

4 Pasargiklian M, Bianco S, Allegra L. Clinical, functional andpathogenetic aspects of bronchial reactivity to prostaglan-dins F2,,,, E1, and E2. Adv Prostaglandin Thromboxane Res1976;1:461-75.

5 Pavord ID, Wong CS, Williams J, Tattersfield AE. Effect ofinhaled prostaglandin E2 on allergen-induced asthma. AmRev RespirDis 1993;148:87-90.

6 Fujimura M, Sakamoto S, Saito M, Miyake Y, Matsuda T.Effect of a thromboxane A2 receptor antagonist (AA-2414) on bronchial hyperresponsiveness to methacholinein subjects with asthma. Jf Allergy Clin Immunol 1991;87:23-7.

7 Bianco S, Refini RM, Pieroni MG. Protective effect ofinhaled lysine acetylsalicylic acid on bronchial specificallergen challenge. Eur RespirJ 199 1;4:478s.

8 Sestini P, Pieroni MG, Refini RM. Protective effect ofinhaled sodium salicylate on the early and late asthmaticreaction to allergen challenge. Eur RespirJ' 1992;5:80s.

9 Pieroni M, Vaghi A, Refini RM. Inhaled indomethacin par-tially prevents the late but not the early bronchial allergicreaction. Am Rev Respir Dis 1993;147:25A.

10 Bianco S, Vaghi A, Pieroni MG, Robuschi M, Refini RM,Sestini P. Protective activity of inhaled nonsteroidal antiin-flammatory drugs on bronchial responsiveness to ultra-sonically nebulized water. J Allergy Clin Immunol 1992;90:833-9. [Erratum Jf Allergy Clin Immunol 1993;91:966].

11 Pavord ID, Wisniewski AF, Tattersfield AE. Inhaledfrusemide and exercise-induced asthma. Evidence of a rolefor inhibitory prostaglandins. Am Rev Respir Dis 1991;143:A2 10.

12 Robuschi M, Scuri M, Vaghi A. Inhaled acetylsalicylic acidenhances the protective activity of furosemide againstexercise-induced bronchoconstriction. Am Rev Respir Dis1992;145:A729.

13 O'Connor BJ, Barnes PJ, Chung KF. Inhibition of sodiummetabisulphite induced bronchoconstriction by frusemidein asthma: role of cyclooxygenase products. Thorax 1994;49:307-11.

14 Meade EA, Smith WL, DeWitt DL. Differential inhibitionof prostaglandin endoperoxide synthase (cyclooxygenase)isozymes by aspirin and other non-steroidal anti-inflammatory drugs. J Biol Chem 1993;268:6610-4.

15 Fine JM, Gordon T, Sheppard D. The role of pH and ionicspecies in sulfur dioxide and sulfite-induced bronchocon-striction. Am Rev Respir Dis 1987;136: 1122-6.

16 Nadel JA, Salem H, Tokiwa Y. Mechanism of bronchocon-striction during inhalation of sulfur dioxide. Jf Appl Physiol1965;20: 164-7.

17 Nichol GM, Nix A, Chung KF, Barnes PJ. Characterisationof bronchoconstrictor responses to sodium metabisulphiteaerosol in atopic subjects with and without asthma. Thorax1989;44:1009-14.

18 Mann JS, Holgate ST, Renwick AG, Cushley MJ. Airwayeffects of purine nucleosides and nucleotides and releasewith bronchial provocation in asthma. JAppl Physiol 1986;61:1667-76.

19 Cushley MJ, Tattersfield AE, Holgate ST. Inhaled adenosineand guanosine on airway resistance in normal andasthmatic subjects. Br GClin Pharmacol 1983;15:161-5.

20 Phillips GD, Rafferty P, Beasley R, Holgate ST. Effect oforal terfenadine on the bronchoconstrictor response toinhaled histamine and adenosine 5'-monophosphate innon-atopic asthma. Thorax 1987;42:939-45.

21 Dixon CSM, Ind PW. Metabisulfite induced bronchocon-striction: mechanisms. Am Rev Respir Dis 1988;137:238.

22 Phillips GD, Foord R, Holgate ST. Inhaled lysine-aspirin asa bronchoprovocation procedure in aspirin-sensitiveasthma: its repeatability, absence of a late-phase reaction,and the role of histamine. J Allergy Clin Immunol1989;84:232-41.

23 Vane JR. Inhibition of prostaglandin synthesis as amechanism of action for aspirin-like drugs. Nature197 1;231:232-5.

24 Cockbill SR, Heptinstall S, Taylor PM. A comparison of theabilities of acetylsalicylic acid, flurbiprofen and indometh-acin to inhibit the release reaction and prostaglandin syn-thesis in human blood platelets. Br J Pharmacol 1979;67:73-8.

25 Smith JB, Willis AL. Aspirin selectively inhibits prostaglan-din production in human platelets. Nature 1971;231:235-7.

26 Wright W, Zhang YG, Salome CM, Woolcock AJ. Effect ofinhaled preservatives on asthmatic subjects. I. Sodiummetabisulfite. Am Rev Respir Dis 1990;141: 1400-4.

27 Tamaoki J, Sekizawa K, Graf PD, Nadel JA. Cholinergicneuromodulation by prostaglandin D2 in canine airwaysmooth muscle. _Appl Physiol 1987;63: 1396-400.

28 Beasley R, Varley J, Robinson C, Holgate ST. Cholinergic-mediated bronchoconstriction induced by prostaglandinD2, its initial metabolite 9a, 11P-PGF2 and PGF2,, inasthma. Am Rev Respir Dis 1987;136:1140-4.

29 Platshon LF, Kaliner M. The effect of the immunologicrelease of histamine on human lung cyclic nucleotidelevels and prostaglandin generation. J Clin Invest 1978;632:1113-21.

30 Adkinson NF, Newball HH, Findlay S, Adams K, Lichten-stein LM. Anaphylactic release of prostaglandins fromhuman lung in vitro. Am Rev Respir Dis 1980;121:911-20.

31 Smith WL. The eicosanoids and their biochemical mecha-nisms of action. Biochemj_ 1989;259:315-24.

32 Kujubu DA, Herschman HR. Dexamethasone inhibitsmitogen induction of the TIS10 prostaglandin synthase/cyclooxygenase gene. J Biol Chem 1992;267:7991-4.

33 Masferrer JL, Seibert K, Zweifel B, Needleman P.Endogenous glucocorticoids regulate an inducible cyclo-oxygenase enzyme. Proc Natl Acad Sci USA 1992;89:3917-21.

34 Mitchell JA, Akarasereenont P, Thiemermann C, Flower RJ,Vane JR. Selectivity of nonsteroidal antiinflammatorydrugs as inhibitors of constitutive and inducible cyclo-oxygenase. Proc Nad Acad Sci USA 1994;90:11693-7.

35 Crimi N, Palermo F, Polosa R, Oliveri R, Maccarrone C,Palermo B, et al. Effect of indomethacin on adenosine-induced bronchoconstriction. J Allergy Clin Immunol1989;83:921-5.

36 Phillips GD, Holgate ST. The effect of oral terfenadinealone and in combination with flurbiprofen on thebronchoconstrictor response to inhaled adenosine 5'-monophosphate in nonatopic asthma. Am Rev Respir Dis1989;139:463-9.

37 Pavord ID, Wisniewski A, Tattersfield AE. Refractoriness toinhaled sodium metabisulphite in subjects with mildasthma. EurRespirj 1994;7:50-4.

38 Vanderhoek JY, Bryant RW, Bailey JM. Inhibition of leuko-triene biosynthesis by the leukocyte product 15-hydroxy-5,8,11,13-eicosatetraenoic acid. J Biol Chem 1980;255:10064-6.

39 Goetzl EJ. Selective feed-back inhibition of the5-lipoxygenation of arachidonic acid in humanT-lymphocytes. Biochem Biophys Res Commun 1981;101:344-50.

40 Belcher NG, Rees PJ, Clark TJ, Lee TH. A comparison ofthe refractory periods induced by hypertonic airway chal-lenge and exercise in bronchial asthma. Am Rev Respir Dis1987;135:822-5.

804



j.com/


ugust 1996. Dow

nloaded from


to - thoraxthromboxane in asthma is less clear and has been probed with non-steroidal...

Documents