PDB

KR*BI

BntaaflmCNv2vtmdpntCaoanTigrttst

Cat((e

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2006;4:196–202

revalence and Mechanism of Nonsteroidal Anti-Inflammatoryrug–Induced Clinical Relapse in Patients With Inflammatoryowel Disease

EN TAKEUCHI,* SIMON SMALE,* PURUSHOTHAMAN PREMCHAND,* LAURENCE MAIDEN,*OY SHERWOOD,‡ BJARNI THJODLEIFSSON,§ EINAR BJORNSSON,� and INGVAR BJARNASON*Department of Internal Medicine, Guy’s, King’s, St Thomas’ Medical School, London, United Kingdom; ‡Department of Clinicaliochemistry, King’s College Hospital, London, United Kingdom; §Department of Gastroenterology, University Hospital Iceland, Reykjavik,

celand; and �Department of Gastroenterology, Sahlgrenska University Hospital, Gothenburg, Sweden

tpnpiaCatimia

tfntcisliumtrtmecs

s

ackground & Aims: It has been variably suggested thatonselective NSAIDs and cyclooxygenase (COX)-2 selec-ive inhibitors aggravate or ameliorate clinical diseasectivity in patients with inflammatory bowel disease. Wessessed the effect of these drugs in patients with in-ammatory bowel disease (n � 209) and the possibleechanisms. Methods: First, patients with quiescent

rohn’s disease and ulcerative colitis received the non-SAID analgesic acetaminophen (n � 26) and the con-entional NSAIDs naproxen (n � 32), diclofenac (n �9), and indomethacin (n � 22) for 4 weeks. The Har-ey-Bradshaw index was used to define relapse. Second,o assess the mechanism of relapse, intestinal inflam-ation was quantitated (fecal calprotectin) before and

uring treatment (20 patients/group) with acetamino-hen, naproxen (topical effect, COX-1 and -2 inhibitor),abumetone (COX-1 and -2 inhibitor), nimesulide (selec-ive COX-2 inhibitor), and low-dose aspirin (selectiveOX-1 inhibition). Results: Nonselective NSAIDs weressociated with a 17%–28% relapse rate within 9 daysf ingestion. No patient had an early relapse on acet-minophen, nimesulide, or aspirin, whereas those onaproxen and nabumetone (20%) experienced relapse.hese clinical relapses were associated with escalatingntestinal inflammatory activity. Conclusions: NSAID in-estion is associated with frequent and early clinicalelapse of quiescent inflammatory bowel disease, andhe mechanism appears to be due to dual inhibition ofhe COX enzymes. Selective COX-2 inhibition with nime-ulide and COX-1 inhibition with low-dose aspirin appearo be well-tolerated in the short-term.

onventional NSAIDs are the most prescribed of theanti-rheumatic drugs, which attests to their efficacy

s anti-inflammatory analgesics. These therapeutic ac-ions are largely due to inhibition of cyclooxygenaseCOX)-2. The side effects of NSAIDs on the stomachNSAID-gastropathy)1 and the small bowel (NSAID-

nteropathy)2 are, however, of considerable concern, con-

ributing significantly to the morbidity and mortality ofatients receiving these drugs long-term.3 The mecha-isms underlying this damage are complex and incom-letely understood. In experimental animals the damages initiated by various combinations of the 3 biochemicalctions common to all conventional NSAIDs,2,4 ie,OX-1 inhibition, COX-2 inhibition, and the “topical”ction. The biochemistry of this topical action is thoughto involve an NSAID–surface membrane phospholipidnteraction5 and/or an effect on mitochondrial energyetabolism,2 both of which are consequent to the phys-

cochemical properties of conventional NSAIDs, namelycidity and their high lipid solubility (lipophilicity).

Conventional NSAIDs are occasionally implicated inhe development of de novo colitis,2,6 but they are morerequently involved in aggravating pre-existing intesti-al diseases.2 This is most obviously illustrated in pa-ients with the inflammatory bowel disease ulcerativeolitis and Crohn’s disease, who frequently require anti-nflammatory analgesics because of peripheral arthritis,acroiliitis, ankylosing spondylitis, and osteoporosis-re-ated fractures.7 Although early studies suggested thatntestinal inflammation and symptoms in patients withlcerative colitis improved with NSAID treatment,8,9

ost subsequent studies, albeit not all,10,11 suggestedhat NSAIDs might be one of the causes of clinicalelapse of inflammatory bowel disease.12–17 In either casehere are substantial uncertainties as to the possibleechanisms (inhibition of COX-1 or COX-2, the topical

ffect, or a combination of these). Theoretically it isonceivable that selective COX-2 inhibition exerts aimilar anti-inflammatory effect in the inflamed intestine

Abbreviations used in this paper: COX, cyclooxygenase; NSAID, non-teroidal anti-inflammatory drug.© 2006 by the American Gastroenterological Association Institute

1542-3565/06/$32.00

PII: 10.1053/S1542-3565(05)00980-8

aaiCoirrtrfpnitmdewcd

cgiflwt

bpptib1rctdaPcswwNrctm

cdgssmdnsms

(wpcgpmw(p

cawt

February 2006 NSAID-INDUCED RELAPSE IN PATIENTS WITH IBD 197

s it does in inflamed joints. The data in human beingsre inconclusive. There are reports of exacerbation ofnflammatory bowel disease with the use of selectiveOX-2 inhibitors celecoxib and rofecoxib,18–20 whereasthers consider them safe.21 A significant problem fornterpretation of these studies is that many are caseeports, and some are retrospective in which the temporalelationship between the disease activity and drug inges-ion is not always clear. Furthermore, patients in clinicalemission with substantial inflammation, as reflected byecal calprotectin concentrations (a neutrophil selectiverotein) that are in 5-fold excess of the upper limit oformal,22 are at particular risk of a “spontaneous” clin-cal relapse. Finally, the clinical disease activity indiceshat are used to define the clinical relapse can be made toimic relapse of disease with NSAIDs, because these

rugs frequently cause abdominal pain, diarrhea, nausea,tc23 in rheumatic patients as well as healthy volunteersithout a corresponding inflammatory response that

haracterizes the clinical relapse of inflammatory bowelisease.24,25

The aims of this study were to assess the prevalence oflinical relapse of inflammatory bowel disease wheniven conventional NSAIDs, to assess whether the clin-cal relapse is associated with escalating intestinal in-ammation, and to assess the possible mechanisms, ie,hether it is due to inhibition of COX-1 or COX-2, the

opical effect, or a combination of these effects.

Subjects and MethodsPatients with ulcerative colitis and Crohn’s disease

etween the ages of 20–70 years were invited to participate,rovided that they were in clinical remission. These wereatients attending gastroenterology outpatient clinics wherehey are followed up every 4–12 months. The diagnoses ofnflammatory bowel disease were biopsy and/or radiologicallyased in all and had been made between 2–23 years (median,1 years) before this study. The diagnosis was verified byeview of case notes and interview according to acceptedriteria.26 The principal inclusion criterion for this study washat patients were in full clinical remission defined by clinicalisease activity score questionnaire, and that they had not hadclinical relapse during the last 6 weeks requiring treatment.atients on NSAIDs or aspirin, those with serious psychiatric,ardiovascular, respiratory, renal, hepatic, and central nervousystem disease, malignancy, and women at risk of pregnancyere specifically excluded as were those misusing alcohol orith a history of substance misuse. A history of significantSAID intolerance (hypersensitivity reactions, asthma, acute

hinitis, urticaria, and angioneurotic edema) was an exclusionriterion. However, patients who had experienced adverse gas-rointestinal side effects (dyspepsia or flare-up of their inflam-

atory bowel disease) of NSAIDs in the past were not ex-

luded, but it is possible that some of these would haveeclined participation in this study. Patients who had under-one extensive surgery (colectomy, colostomy, ileostomy,mall bowel resection of more that 40 cm, bypass or diversionurgery) were excluded. Also excluded were patients receivingore than 10 mg prednisolone/day, those requiring elemental

iets or nutritional support, and those who had received tumorecrosis factor antibodies (or other biologicals) or undergoneurgery during the preceding 6 months. No change in treat-ent for inflammatory bowel disease was made during the

tudy.Altogether 209 patients with inflammatory bowel disease

ulcerative colitis/Crohn’s disease) underwent study. Thereere 2 aspects to these studies. The first part involved 109atients with inflammatory bowel disease to assess whetheronventional NSAIDs caused clinical relapse of disease wheniven during a period of 4 weeks. Thirty-two of these werelaced on naproxen (500 mg twice a day), 29 on diclofenac (75g twice a day), and 22 on indomethacin (75 mg twice a day),hereas 26 received the non-NSAID analgesic acetaminophen

1 g 3 times a day), and these served as a control group forossible spontaneous clinical relapse of disease.The second part assessed the possible mechanism of the

linical relapse by administering drugs that have differentialctions on the COX enzymes and the topical effect. Patientsere assigned to 1 of 5 groups (20 patients/group) and received

he following:

1. Acetaminophen (1 g 3 times a day) as a non-NSAIDcontrol (acetaminophen does not inhibit COX-1 orCOX-227 or have the topical effect28)

2. Naproxen (500 mg twice a day), a conventional NSAID(naproxen inhibits COX-1 and COX-227 and exerts thetopical effect28)

3. Nabumetone (1 g twice a day), a non-acidic pro-NSAID, which undergoes hepatic biotransformation tothe active component, 6-methoxy-2-naphthylaceticacid, which is not excreted in bile (nabumetone does nothave a topical effect,28 whereas 6-methoxy-2-naphthyl-acetic acid inhibits both COX-1 and COX-2,27 but theintestine is not exposed to its topical action)

4. Nimesulide (100 mg twice a day), a preferential orselective COX-2 inhibitor29 (nimesulide inhibitsCOX-2 without affecting COX-1 significantly at thesedoses,27,29 and it does not have a topical effect30)

5. Aspirin (75 mg once a day). At this low dose, aspirinappears to be a relatively selective COX-1 inhibitor.Although aspirin is acidic and therefore exerts a topicaleffect, most of the dose is absorbed from the stomachand duodenum (aspirin is not excreted in bile), and asignificant topical effect on the rest of the gastrointes-tinal tract is unlikely31 and has not been evident withhigher (2.4 g) doses in human beings.32 It is, however,more difficult to assess whether aspirin has an effect onCOX-2 at these doses. Ex vivo studies indicate that it

does not inhibit COX-2,33,34 it is not associated with

mr1s(wds

T

Ka

cifwdbcHi2

ngcwtwtr

opnptasdpl(

httaoTtqcgom

le4

gtcD

T

MAU

C

T

198 TAKEUCHI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 2

any significant anti-inflammatory action (a COX-2–dependent action),35 and judging from its COX-1 se-lectivity in vitro,27 it seems to be a relatively selectiveCOX-1 inhibitor at these doses.

All drug treatments in both parts of the study were for 1onth unless the activity score met the criteria for clinical

elapse, in which case drug treatment was discontinued. The00 patients in the second part of the study also underwenttudies that allowed quantitation of intestinal inflammationfecal calprotectin) before and after 1 and 4 weeks of treatmentith the drugs. When a clinical relapse occurred or if theyiscontinued the study, these patients collected the next stoolample for calprotectin measurement.

The demographic details of the patients are shown inable 1.These studies were approved by the ethical committees of

ing’s College Hospital and Sahlgrenska University Hospital,nd all patients gave informed consent.

Clinical Disease Activity

Clinical remission was defined as a Harvey-Bradshawlinical disease activity index36 of 4 or less during the preced-ng week. Although the Harvey-Bradshaw index was designedor use in patients with Crohn’s disease, we used it for patientsith both Crohn’s disease and ulcerative colitis as previouslyone.22,37 All patients completed diary cards during the weekefore drug treatment and then a daily diary if there was ahange in clinical symptoms. A clinical relapse was defined asarvey-Bradshaw score of equal to or greater than 5 and an

ncrease of 4 or more from the pretreatment baseline values onconsecutive days while receiving the drugs.Some of the side effects of NSAIDs, such as diarrhea and

onspecific crampy abdominal pain,23 cannot be easily distin-uished on clinical grounds from the diarrhea associated withlinical disease activity of inflammatory bowel disease. Patientsere warned about the possible occurrence of upper gastroin-

estinal dyspepsia, nausea, and nonspecific abdominal pain andere instructed not to include these in the questionnaire unless

his had been a prominent feature of their previous clinical

able 1. Demographic and Clinical Details

Acetaminophen Naproxen Diclofe

ale/female 12/14 14/18 19/1ge (y) median (range) 37 (24–62) 40 (20–70) 33 (20–lcerative colitisTotal 5 4 5Left-sided 5 5 7Proctitis 6 3 5

rohn’s diseaseSmall bowel 4 11 7Colonic 6 9 5

reatmentMesalamine 17 19 16Azathioprine 7 6 3Corticosteroids 2 1 0

elapse. t

Intestinal Inflammation

Intestinal inflammation was assessed by measurementf calprotectin in feces. Calprotectin is a neutrophil selectiverotein (also present in small quantities in other polymorpho-uclear white cells), which resists bacterial degradation.38 Itsresence in feces relates quantitatively to the neutrophil flux tohe gastrointestinal tract, ie, it is proportional to the degree ofcute inflammation.24,39 Patients in the second part of thetudy provided a stool sample within 3 days of commencingrug treatment and on days 7 and 28 during treatment. In caseatients fulfilled the criteria for a clinical relapse, they col-ected the next stool sample for calprotectin measurementsthis was in all cases within 24 hours of the clinical relapse).

Stool samples were delivered to the laboratory within 48ours and were frozen at �20°C on arrival. Samples werehawed at room temperature, and the calprotectin concentra-ion was determined by an enzyme-linked immunosorbentssay after an extraction procedure as previously described24,38

r with the use of a commercial kit (Calprest; Eurospital,rieste, Italy).40 The 2 methods are essentially comparable in

heir performance, sensitivity, and specificity,40 but they giveuantitatively different data. The commercial kit providesalprotectin values and upper normal limits that are 5 timesreater than the older assay. All the results obtained with theld method during these studies were therefore presented afterultiplying the values by a factor of 5.The normal range of fecal calprotectin concentration (upper

imit of normal is 52 �g/g stool), with the commercial kit, wasstablished by the laboratory in 88 healthy volunteers (42 men,6 women; median age, 47 years; range, 20–70 years).

Statistics

The statistical differences between the treatmentroups (clinical relapses) were tested with 2-tailed Fisher exactest with Yates correction. Sequential changes in calprotectinoncentrations were assessed by the paired Student t test.emographic data were assessed with the Wilcoxon rank sum

Indomethacin Acetaminophen Naproxen Nabumetone

9/13 13/7 9/11 8/1238 (24–70) 36 (21–66) 40 (20–58) 41 (26–69)

7 3 1 46 6 9 72 1 2 3

5 7 4 43 3 4 2

15 15 13 174 3 6 30 1 2 0

nac

068)

est. Values of P �.05 were regarded as significant.

NswadB(d[aipc(maN

acdtweCa(nmatwrmcd

rt2tp(w

mtmfrNatitpimCaRa

dboiIrfl

ptaTirk

mpWnaa

T

ANDI

Nta

p

February 2006 NSAID-INDUCED RELAPSE IN PATIENTS WITH IBD 199

ResultsTable 2 shows that 17%–28% of the patients on

SAIDs experienced relapse during the first part of thetudy. The relapse rates as compared with acetaminophenere statistically significant for naproxen and indometh-

cin. Those who experienced relapse did so within 2–9ays of receiving the NSAID. Their median Harvey-radshaw score increased from 1 (range, 0–4) to 9

range, 6–12). The relapse rate in patients with Crohn’sisease (8/40 [20%]) and ulcerative colitis (n � 10/4423%]) was comparable (P � .6), and there was nopparent association with disease location, treatment fornflammatory bowel disease, age, or sex. Six of the 18atients who experienced relapse on NSAIDs were suc-essfully treated as outpatients with oral corticosteroids10–30 mg/day for 4–7 days with dose reductions of 5g/day every 3–5 days), and the remaining 12 became

symptomatic within 2–11 days of discontinuation of theSAID.The second part of the study was carried out in an

ttempt to elucidate the possible mechanisms of thelinical relapse. Baseline fecal calprotectin levels andemographic factors did not differ significantly betweenhe 5 groups. Figure 1 shows the overall results. Thereere a total of 10 patients who had a clinical relapse, 1

ach receiving acetaminophen and nimesulide (selectiveOX-2 inhibition) and 4 each in the naproxen (COX-1nd COX-2 inhibition � topical effect) and nabumetoneCOX-1 and COX-2 inhibition) treated groups. Theumber of patients with relapse with naproxen and nabu-etone differed significantly (P � .01) from that of

cetaminophen, nimesulide, and aspirin, regardless of theiming of relapse. The timing of the clinical relapsesas day 22 and 29 for acetaminophen and nimesulide,

espectively, and on days 2–7 with naproxen and nabu-etone. These relapses were in all cases (Figure 1) asso-

iated with an increase in fecal calprotectin levels (me-

able 2. Clinical Relapse Rates in Patients WithInflammatory Bowel Disease ReceivingAcetaminophen and 3 Different NSAIDs

DrugNumber ofpatients

Number ofrelapses (%)

Pvaluesa

cetaminophen 36 0 (0)aproxen 32 9 (28) .01iclofenac 29 5 (17) .08

ndomethacin 22 5 (24) .04

OTE. Statistical differences were assessed with the Fisher exactest.P values refer to the differences between the relapse rates of theatients on acetaminophen and the NSAID.

ian, 418 and 3537 �g/g before and after relapse values, r

espectively; P � .03), whereas the calprotectin levels inhe rest of the patients did not alter significantly (243,60, and 278 �g/g before and after 1 and 4 weeks ofreatment; P � .6) during this study. Three relapsedatients were treated with corticosteroids, including onetaking nabumetone) who required hospitalization,hereas the rest improved on stopping the drugs.

DiscussionThese studies show that conventional NSAIDs

ight cause clinical relapse within a few days of inges-ion in 17%–28% of asymptomatic patients with inflam-atory bowel disease. Those who tolerated the NSAID

or a week did not seem to be at serious risk for clinicalelapse. Furthermore, the clinical relapse caused bySAIDs is not simply due to nonspecific abdominal pain

nd dyspepsia, which is seen in 40% of rheumatic pa-ients taking these drugs. On the contrary, the NSAID-nduced clinical relapse is associated with escalating in-estinal inflammatory activity, similar to that seen inatients with active inflammatory bowel disease not tak-ng NSAIDs.24,39 Within the constraints of the experi-

ental design it seems unlikely that either selectiveOX-1 or COX-2 inhibition by itself is the sole mech-nism by which NSAIDs lead to the clinical relapse.ather, it is suggested that the relapse might be brought

bout by dual inhibition of the enzymes.The cause of clinical relapse of inflammatory bowel

isease is unknown in most cases, but it is characterizedy greater inflammatory activity than that seen in mostther intestinal diseases.25,41 When a cause is suspected,t is usually ascribed to stress, infection, or NSAID use.n the case of NSAIDs the data on the prevalence ofelapse are sparse,8–17 but many case reports suggest thatair of quiescent disease occurs within a few days.The first part of this study prospectively assessed the

revalence of symptomatic relapse of disease when pa-ients were placed on diclofenac, naproxen, or indometh-cin as compared with patients receiving acetaminophen.he NSAIDs were associated with a symptomatic relapse

n 17%–28% of patients receiving the drugs, and theseelapses occurred within days of receiving the drugs, ineeping with published case reports.2,15,16

The second part of the study assessed the possibleechanism of relapse. In this context it might be im-

ortant to distinguish between the early and late relapse.ith regards to the early relapse, it is clear that the

onselective NSAIDs naproxen and nabumetone weressociated with clinical relapse. Furthermore, this wasssociated with increased inflammatory activity compa-

able to that seen in relapsed patients not on NSAIDs.24

Fpilwh

200 TAKEUCHI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 2

igure 1. The effect of acetaminophen, naproxen, nabumetone, nimesulide, and aspirin on clinical relapse and intestinal inflammation inatients with inflammatory bowel disease. Four patients each had a clinical relapse (identified by *) within a week of naproxen and nabumetone

ngestion, which is significantly greater (P � .009) than the no relapses in patients taking acetaminophen at 1 week, and 1 patient each had aate relapse of disease on paracetamol and nimesulide. All of the clinical relapses were associated with increased calprotectin levels. Patientsho discontinued the study prematurely (for various personal reasons) without an increase in clinical disease activity (identified by �) did not

ave increased fecal calprotectins.

Bhfrc(CcigCobiqcmteiC

nwjsa

(bwpdspabtispit

1

1

1

1

1

1

1

1

1

1

2

2

2

2

February 2006 NSAID-INDUCED RELAPSE IN PATIENTS WITH IBD 201

ecause these relapses occurred equally in subjects atigh and low risk of spontaneous relapse (as reflected byecal calprotectin levels above and below 250 �g/g,espectively),22 it seems unlikely that these relapses oc-urred by chance. By comparison, non-NSAID analgesiaacetaminophen) and selective COX-1 (aspirin) andOX-2 (nimesulide) inhibition did not lead to earlylinical relapse of disease, and there was not a significantncrease in intestinal inflammation. Collectively this sug-ests that it is the dual inhibition of both isoforms of theOX enzymes that is responsible for the clinical relapsef inflammatory bowel disease. The precise mechanismy which this occurs is speculative, but there is increas-ng evidence that the main pathophysiologic conse-uence of COX-1 inhibition is impaired mucosal micro-irculatory blood flow,4 whereas the COX-2 enzymeight have an immunomodulatory role in the gastroin-

estinal tract.42 Interestingly, the gastrointestinal sideffects of conventional NSAIDs are also caused by dualnhibition of the COX enzymes4,30 and not by selectiveOX-1 inhibition, as often suggested.Two patients, one on acetaminophen and the other on

imesulide, had a “late” clinical relapse. Both of theseere at significant risk of spontaneous clinical relapse as

udged by high fecal calprotectin levels at the start of thetudy.22 Whether this is the case or whether this truly isdrug effect is uncertain and will require further study.The clinical implications of these studies are that most

70%–80%) patients with quiescent inflammatoryowel disease tolerate conventional nonselective NSAIDshen administered for 4 weeks. However, 20%–30% ofatients are intolerant and experience a clinical relapse ofisease within a week of ingesting these drugs. Thesetudies were done on patients without arthritic com-laints, and it is conceivable that the drugs might havedifferent effect in patients with active inflammatory

owel disease and arthritis as an extraintestinal manifes-ation. The mechanism of relapse seems to involve dualnhibition of the COX enzymes. The tolerability to theelective COX-2 inhibitor nimesulide and low-dose as-irin might be important when contemplating anti-nflammatory analgesic and cardiovascular prophylacticreatment, respectively, in these patients.

References1. Scarpignato C, Bjarnason I, Bretagne J-F, et al. Towards a GI safer

antiinflammatory therapy. Gastroenterol Int 1999;12:180–215.2. Bjarnason I, Hayllar J, Macpherson AJ, et al. Side effects of

nonsteroidal anti-inflammatory drugs on the small and large in-testine. Gastroenterology 1993;104:1832–1847.

3. Wolfe MM, Lichtenstein DR, Singh G. Gastro intestinal safety ofnonsteroidal antiinflammatory drugs. N Eng J Med 1999;340:1888–1899.

4. Wallace JL, McKnight W, Reuter BK, et al. NSAID-induced gastric 2

damage in rats: requirement for inhibition of both cyclooxygenase1 and 2. Gastroenterology 2000;119:706–714.

5. Darling RL, Romero JJ, Dial EJ, et al. The effects of aspirin ongastric mucosal integrity, surface hydrophobicity, and prostaglan-din metabolism in cyclooxygenase knockout mice. Gastroenterol-ogy 2004;127:94–104.

6. Puspok A, Kiener HP, Oberhuber G. Clinical, endoscopic, histo-logic spectrum of nonsteroidal anti-inflammatory drug inducedlesions of the colon. Dis Colon Rectum 2000;43:685–691.

7. Smale S, Natt RS, Orchard T, et al. Spondylarthropathy andinflammatory bowel disease. Arthritis Rheum 2001;44:2728–2736.

8. Rask-Madsen J, Bukhave K, Laursen LS, et al. Eicosanoids ininflammatory bowel disease - physiology and pathology. In: PetersTJ, ed. The cell biology of inflammation in the gastrointestinaltract. Corners Publication, Hull, UK: 1990:255–271.

9. Campieri M, Franchi LGA, Bazzocchi G, et al. Prostaglandins,indomethacin and ulcerative colitis. Gastroenterology 1980;78:193.

0. Bonner GF, Walczak M, Kitchen L, et al. Tolerance of nonsteroidalantiinflammatory drugs in patients with inflammatory bowel dis-ease. Am J Gastroenterol 2000;95:1946–1948.

1. Bonner G, Fakhri A, Vennamaneni SR. A long-term cohort study ofnonsteroidal anti-inflammatory drug use and disease activity inoutpatients with inflammatory bowel disease. Inflamm Bowel Dis2004;10:751–757.

2. Rampton DS, Sladen GE. Relapse of ulcerative proctocolitis dur-ing treatment with NSAID. Postgrad Med J 1981;57:297–299.

3. Rampton DS, Sladen GE. Prostaglandin synthesis inhibitors inulcerative colitis: flurbiprofen compared with conventional treat-ment. Prostaglandins 1981;21:417–425.

4. Rampton DS, McNeil MI, Sarner M. Analgesic ingestion and otherfactors preceding relapse in ulcerative colitis. Gut 1983;24:187–189.

5. Kaufmann HJ, Taubin HL. Nonsteroidal anti-inflammatory drugsactivate quiescent inflammatory bowel disease. Ann Intern Med1987;107:513–516.

6. Felder JB, Korelitz BI, Rajapakse R, et al. Effects of nonsteroidalantiinflammatory drugs on inflammatory bowel disease: a case-control study. Am J Gastroenterol 2000;95:1949–1954.

7. Evans JM, McMahon AD, Murray FE, et al. Non-steroidal anti-inflammatory drugs are associated with emergency admission tohospital for colitis due to inflammatory bowel disease. Gut 1997;40:619–622.

8. Bonner GF. Exacerbation of inflammatory bowel disease associ-ated with the use of celecoxib. Am J Gastroenterol 2001;96:1306–1308.

9. Gornet JM, Hassani Z, Modiglian R, et al. Exacerbation of Crohn’scolitis with severe colonic hemorrhage in a patient on rofecoxib.Am J Gastroenterol 2002;97:3209–3210.

0. Matuk R, Crawford J, Abreu MT, et al. The spectrum of gastroin-testinal toxicity and effect on disease activity of selective cyclo-oxygenase-2 inhibitors in patients with inflammatory bowel dis-ease. Inflamm Bowel Dis 2004;10:352–356.

1. Mahadevan U, Loftus EV, Tremaine WJ, et al. Safety of selectivecyclooxygenase-2 inhibitors in inflammatory bowel disease. Am JGastroenterol 2003;97:910–914.

2. Tibble J, Sigthorsson G, Fagerhol M, et al. Surrogate markers ofintestinal inflammation are predictive for relapse in patients withinflammatory bowel disease. Gastroenterology 2000;119:15–22.

3. Kivitz AJ, Nayiager S, Schimansky T, et al. Reduced incidence ofgastroduodenal ulcers associated with lumiracoxib comparedwith ibuprofen in patients with rheumatoid arthritis. Aliment Phar-macol Ther 2004;19:1189–1198.

4. Tibble J, Teahon K, Thjodleifsson B, et al. A simple method for

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

PSe

202 TAKEUCHI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 2

assessing intestinal inflammation in Crohn’s disease. Gut2000;47:506–513.

5. Bjarnason I, Takeuchi K, Bjarnason A, et al. The GUT of gut.Scand J Gastroenterol 2004;39:807–815.

6. Malchow H, Ewe K, Brandes JW, et al. European cooperativeCrohn’s disease study (ECCDS): results of drug treatment. Gas-troenterology 1984;86:249–266.

7. Warner TD, Giuliano F, Vojnovic I, et al. Nonsteroid drug selectiv-ities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are as-sociated with human gastrointestinal toxicity: a full in vitro anal-ysis. Proc Natl Acad Sci U S A 1999;96:7563–7568.

8. Mahmud T, Rafi SS, Scott DL, et al. Nonsteroidal antiinflamma-tory drugs and uncoupling of mitochondrial oxidative phosphory-lation. Arthritis Rheum 1996;39:1998–2003.

9. Shah AA, Thjodleifsson B, Murray FE, et al. A randomised, doubleblind, double dummy, crossover study of the effects of nimesu-lide and naproxen on the gastrointestinal tract and an in vivoassesment of their selectivity for cyclooxygenase 1 and 2. Gut2001;48:339–348.

0. Sigthorsson G, Jacob M, Wrigglesworth JM, et al. A comparison ofindomethacin and nimesulide, a selective cyclooxygenase-2 in-hibitor, on key pathophysiological steps in the pathogenesis ofNSAID enteropathy in the rat. Scand J Gastroenterol 1998;33:728–735.

1. Whittle BJR. Unwanted effects of aspirin and related agents onthe gastrointestinal tract. In: Vane JR, Botting RM, eds. Aspirinand other salicylates. London: Chapman & Hall Medical, 1992:465–509.

2. Bjarnason I, Williams P, Smethurst P, et al. The effect of NSAIDsand prostaglandins on the permeability of the human small in-testine. Gut 1986;27:1292–1297.

3. Capone ML, Tacconelli S, Sciulli MG, et al. Clinical pharmacologyof platelet, monocyte, and vascular cyclooxygenase inhibition bynaproxen and low-dose aspirin in healthy subjects. Circulation2004;109:1468–1471.

4. Fiorucci S, Santucci L, Wallace JL, et al. Interaction of a selective H

cyclooxygenase-2 inhibitor with aspirin and NO-releasing aspirinin the human gastric mucosa. Proc Natl Acad Sci U S A2003;100:10937–10941.

5. Botting R. COX-1 and COX-3 inhibitors. Thromb Res 2003;110:269–272.

6. Harvey RF, Bradshaw JM. A simple index of Crohn’s diseaseactivity. Lancet 1980;1:514.

7. Mahmud NMG, Kelleher D, Weir DG. Microalbuminuria correlateswith intestinal histopathological grading in patients with inflam-matory bowel disease. Gut 1996;38:99–103.

8. Roseth AG, Fagerhol MK, Aadland E, et al. Assessment of theneutrophil dominating calprotectin in feces: a methodologicstudy. Scand J Gastroenterol 1992;27:793–798.

9. Roseth AG, Schmidt PN, Fagerhol MK. Correlation between faecalexcretion of Indium-111-labelled granulocytes and calprotectin, agranulocyte marker protein, in patients with inflammatory boweldisease. Scand J Gastroenterol 1999;34:50–54.

0. Johne B, Kronborg O, Ton HI, et al. A new fecal calprotectin testfor colorectal neoplasia: clinical results and comparison withprevious method. Scand J Gastroenterol 2001;36:291–296.

1. Tibble J, Sigthorsson G, Foster R, et al. The use of surrogatemarkers of intestinal inflammation and symptom based criteria todistinguish organic from non-organic disease. Gastroenterology2002;123:450–460.

2. Newberry RD, Stenson WF, Lorenz RG. Cyclooxygenase-2-depen-dent arachidonic acid metabolites are essential modulators ofthe intestinal immune response to dietary antigen. Nature Med1999;5:900–906.

Address requests for reprints to: Ingvar Bjarnason, MD, MSc, FRC-ath, FRCP(Glasg), DSc, Professor of Digestive Diseases, Guy’s, King’s,t Thomas’ Medical School, Bessemer Road, London SE5 9PJ, UK.-mail: ingvar.bjarnason@kcl.ac.uk; fax: 00-44-20-7346-6474.The authors have received grants from Merck Sharp & Dohme and

elsinn Pharmaceuticals, the manufacturers of nimesulide.