Gastrointestinal safety of amtolmetin guacyl in comparison with celecoxib in patients with rheumatoid arthritis Z. Jajic´, M. Malaise, K. Nekam, É. Koó, K. Dankó, M. Kovacs, C. Scarpignato1

Department for Rheumatology, Physical Medicine and Rehabilitation, Medical Faculty University of Zagreb, Referral Centre for Inflammatory Rheumatic Disease, Zagreb, Croatia; 1Laboratory of Clinical Pharmacology, School of Medicine and Dentistry, University of Parma, Parma, Italy. Multicentre Study from the Department of Rheumatology, Physical Medicine and Rehabilitation, Medical Faculty University, Zagreb (Croatia); Department of Rheumatology, University of Liege, Liege (Belgium); Department of Clinical Medicine III, Carl Gustav University, Dresden (Germany); Department of Rheumatology, Clinical Medi- cine IV, Carl-Thiem Clinic, Cotbus (Germany); Department of Rheumatology, Hospital of Horovice, Horovice (Czech Republic); Department of Rheumatology, University Hospital, Klimentskà Polyclinic, Prague (Czech Republic); Department of Allergology and Immunology and Department of Rheumatology II, Polyclinic of the Hospitaller Brothers of St. John of God in Budapest, Budapest (Hungary); Department of Internal Medicine 1, County Hospital of Zala, Zalaegerszeg (Hungary); Department of Rehabilitation, “Markusovsky” Hospital and outpatient Clinic of Vas County, Szombathely (Hungary); University of Debrecen, Medical and Health Science Centre, Institute for Internal Medicine, III Department, Division of Immunology, Debrecen (Hungary)

Abstract Objective Selective inhibitors of cyclo-oxygenase-2 (COX-2) appear to be safer than conventional NSAIDs on the gastrointestinal (GI) tract. Amtolmetin guacyl (AMG), a NSAID that inhibits both COX-1 and COX-2, has an anti-inflammatory effect compara- ble to that of traditional NSAIDs, with a better GI safety profile. The primary end-point of this study was to evaluate the gas- trointestinal safety of amtolmetin guacyl in comparison with celecoxib in patients affected with rheumatoid arthritis. The assessment of efficacy was the secondary end-point.

Methods This study was a 24-week, randomized, parallel group, double-blind, double dummy, multicentre trial; 235 patients were enrolled and 180 patients (85 in the AMG group and 95 in the celecoxib group) completed the study. Each patient received twice daily amtolmetin guacyl 600 mg or celecoxib 200 mg. Assessment of safety was performed by upper GI endoscopy, gastrointestinal symptoms evaluation, electrocardiography, blood and urine laboratory tests, adverse events recording. Assessment of efficacy was performed by using the American College of Rheumatology (ACR-20) responder index.

Results Neither amtolmetin guacyl nor celecoxib determined a worsening of baseline gastro-duodenal endoscopy findings. The per- centage of patients with normal findings did not significantly change after treatment with both drugs, being virtually identi- cal with AMG (i.e. 75.29%) and increasing from 75.79% to 77.66% with celecoxib. Moreover an evaluation of the other safety parameters did not reveal any difference between the two treatment groups. Therapeutic efficacy was equivalent in both groups, with no statistical difference between the two drugs at all time intervals.

Conclusions In patients affected with rheumatoid arthritis, AMG and celecoxib proved to be equivalent, showing comparable gastrointestinal safety and therapeutic efficacy of treatment.

Key words Amtolmetin guacyl (AMG), celecoxib, NSAIDs, gastrointestinal safety, rheumatoid arthritis.

Clinical and Experimental Rheumatology 2005; 23: 809-818. Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

Zrinka Jajic´, MD, PhD; Michel Malaise, Introduction essentially associated with similar dam- MD, PhD; Kristof Nekam, MD, PhD; Rheumatoid arthritis is an autoimmune age. The strength of this idea is that it is Éva Koó, MD, PhD; Katalin Dankó, MD, disease characterized by chronic joint simple; there are reasonable associa- PhD, Monika Kovacs, MD; Carmelo inflammation with a fluctuating course tions between NSAID-induced reduc- Scarpignato, MD, DSc, PharmD. that can lead to progressive and des- tions in mucosal prostaglandin levels Please address correspondence and re- tructive arthropathy, followed by defor- and the damage ,and one report (13) print requests to: Prof. Zrinka Jajic´, MD, mities and disability (1, 2). Its manage- suggests that antibodies directed against Department for Rheumatology, Physical Medicine and Rehabilitation, Medical ment involves the use of new and tradi- COX cause in rabbits damage to the Faculty University of Zagreb, Referral tional disease-modifying anti-rheumat- gastrointestinal tract that is identical to Centre for Inflammatory Rheumatic ic drugs (DMARDs) (3), which can that induced by NSAIDs. Disease, Vinogradska c. 29, 10000 affect the course of the disease, and After the discovery of the two COX iso- Zagreb, Croatia. E-mail: [email protected] non-steroidal anti-inflammatory drugs forms (that is COX-1 and COX-2) com- Received in February 11, 2005; accepted (NSAIDs), which provide symptomatic pounds specifically designed to inhibit in revised form on July 14, 2005. relief by controlling pain and inflamma- COX-2, sparing COX-1 at therapeutic © Copyright CLINICAL AND EXPERIMEN- tion. Although these latter compounds doses, have been developed. Although TAL RHEUMATOLOGY 2005. represent a very effective class of drugs, increasing evidence indicates that also their use is associated with a broad COX-2 is normally expressed in the spectrum of untoward reactions in the mucosa of the digestive tract where it liver, kidney, skin and gut (4, 5); gas- plays a physiological role, assisting the trointestinal (GI) side effects are, how- housekeeping action of COX-1 in gas- ever, the most common adverse event troprotection (14, 15), the so-called encountered with this class of drugs (6- “selective COX-2 inhibitors” display a 10). GI problems encompass a wide significantly improved risk-benefit ratio range of different clinical pictures, compared with non-selective NSAIDs. spanning from mild symptoms such as Indeed, COX-1 expression represents dyspepsia, heartburn and abdominal the dominant isoform in healthy mucosa discomfort to more serious events, like (14) and PGs generated by COX-2 be- peptic ulcer and its life-threatening come important only during ulcer heal- complications, bleeding and perfora- ing (14, 15) and in the presence of Heli- tion. Indeed, gastroduodenal mucosa cobacter pylori infection (14,16). In ad- possesses an array of defensive mecha- dition, it is worth mentioning that selec- nisms and NSAIDs have a deleterious tive COX-2 inhibitors, like celecoxib effect on most of them (11). and rofecoxib, besides sparing PG syn- These drugs appear to cause gastroduo- thesis in the GI mucosa, do not display denal damage by two main mecha- any topical irritancy as they are non- nisms: a physiochemical disruption of acidic, nor do they increase gastric or the gastric mucosal barrier and a sys- intestinal permeability or cause mucosal temic inhibition of gastric mucosal pro- inflammation (17). Their failure to sig- tection, through inhibition of cyclo- nificantly affect the two basic biochem- oxygenase (prostaglandin endoperoxide ical mechanisms of NSAID-induced G/H synthase, COX) activity of the GI damage may account for their remark- mucosa. A reduced synthesis of mucus able GI tolerability. Large clinical trials and bicarbonate, an impairment of mu- have indeed demonstrated that they are cosal blood flow and an increase in acid as therapeutically effective as conven- secretion represent the main conse- tional NSAIDs (18,19), while being saf- quences of NSAID-induced prosta- er at gastroduodenal level (20,21). Not glandin (PG) deficiency (12). Addition- only have endoscopic studies shown a al mechanisms which may add to the significant reduction in the incidence of damage have been demonstrated. These gastric and duodenal ulcers (22, 23), but include the uncoupling of oxidative clinical trials have pointed out a signifi- phosphorylation, reduced mucosal cell cant reduction in peptic ulcer complica- proliferation and DNA synthesis as well tions . as neutrophil activation (12). The out- Amtolmetin guacyl (2-methoxyphenyl- standing candidate mechanism for initi- 1-methyl-5-p-methylbenzoyl-pyrrole- ating NSAID damage is their action to 2-acetamido acetate, AMG) is an inhibit cyclo-oxygenase-1 (COX-1) as NSAID (24) recently introduced into all conventional NSAIDs have this ac- the Italian market and approved for the tion (and not other agents) and all are treatment of rheumatoid arthritis, os-

810 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al. teoarthritis, extra-articular rheumatism gastroprotective activity. percentage of gastrointestinal events and post-surgical pain. It affects both Taking all the above considerations in- induced by both treatments and fore- COX-1 and COX-2, the COX-2/COX- to account, this study was designed to seen dropouts. Since the number of pa- 1 selectivity ratio, evaluated in studies compare the GI safety and the clinical tients lost during the study was less using bovine aortic endothelial cells efficacy of AMG and celecoxib, a than that expected, the sample size was and LPS-stimulated macrophages, be- widely used and GI safe COX-2 selec- recalculated and the trial was stopped ing 4.4 (Vane et al., unpublished re- tive inhibitor, in patients affected by after the randomization of 235 patients sults). Like COX-2 selective com- rheumatoid arthritis. (118 to the AMG group and 117 to the pounds, the drug displays an effective celecoxib group). anti-inflammatory action with improv- Patients and methods ed GI tolerability, but, conversely from Patients Study design them, maintains an antiplatelet activity Out-patients of either sex aged over 18 This was a multicentre, double-blind, (25, 26). years who fulfilled the American Col- double dummy, active controlled, ran- Clinical trials have shown that its anti- lege of Rheumatology (ACR) criteria domized, parallel group trial. It consist- inflammatory, analgesic and antipyretic for clinical diagnosis of rheumatoid ed of 2 phases: a single-blind placebo effects are comparable to those of ref- arthritis in acute phase with Functional run-in period (one week) and a double- erence NSAIDs (27-29). Two meta- Capacity Classification of I-III, requir- blind active treatment period (24 analyses (30, 31) provided evidence ing continuous, prolonged NSAID weeks). Using a computer-generated that AMG displays a substantially low- therapy were eligible for the study. 1:1 randomization list by the coordinat- er incidence (7.2% and 28.3% respec- Patients were excluded from the trial if ing centre, the patients were assigned tively) of gastric adverse events in they had other rheumatic diseases, pso- to receive one of the following treat- comparison with traditional NSAIDs. riasis, uncontrolled diabetes, untreated ments: AMG 600 mg tablets or cele- The frequency and severity of gastric hyperthyroidism, significant renal im- coxib 200 mg capsules, both given mucosal lesions at endoscopic evalua- pairment (serum creatinine > 1.5 mg/ twice a day. AMG tablets (kindly pro- tion was lower for AMG in comparison dl), significant hepatic impairment vided by Medosan Ricerca, Rome, It- with other NSAIDs, the odds ratio (AST or ALT values over twice the aly) were administered 2-3 hours from being 0.3 (95% CI 0.1 to 0.7) for severe upper limit of the normal range), neu- meals, while celecoxib capsules were lesions and 0.1 (95% CI 0.1 to 0.4) for rological, haematological or autoim- administered regardless of meal con- mild and severe lesions, respectively mune (other than rheumatoid arthritis) sumption (Fig. 1). (30). diseases, cancer or any chronic disease. Prior to enrolment, the patients receiv- Pharmacological studies performed in Patients were also excluded if they had ed a full explanation of the nature and various animal models have shown that a history of ulcer or gastric bleeding or purpose of the study, provided written AMG is not only devoid of any damag- any clinically significant upper gastro- informed consent and underwent a ing effect on gastric mucosa, but actu- intestinal mucosal damage (i.e. > 10 physical examination and laboratory ally displays a gastroprotective effect erosions in the stomach and/or duode- testing. Follow-up clinic visits took (26), which extends down to the bowel num; oesophageal, gastric or duodenal place at weeks 4, 12 and 24 after the (32). This peculiar pharmacological ulcers). Other exclusion criteria were: start of both treatments. activity has been confirmed in humans treatment with steroids at doses greater This study, conducted in accordance where AMG protects the gastric mu- than the equivalent of 7.5 mg of pred- with the current ICH-GCP Guidelines cosa from ethanol damage with an effi- nisone per day; treatment initiation or and the Declaration of Helsinki as cacy comparable to that of misoprostol dosage alteration of any DMARD and/ amended, was approved by the Nation- (33). Gastric mucosal protection is like- or Disease–Controlling Anti-Rheumat- al Authorities of each Country involved ly to be due to the presence of a vanillic ic Drug (DCARD) during the 90 days and the Local Ethical and Drug Com- moiety in the molecule of AMG which, preceding the study and throughout the mittees of each participating Centre. through stimulation of capsaicin recep- study; intake of anti-ulcer drugs; aller- tors, causes Calcitonin Gene Related gy, sensitivity or intolerance to study Concomitant treatments Peptide (CGRP) release and a conse- drugs and/or study drugs formulation Concomitant medications (except for quent increase in nitric oxide (NO) pro- ingredients; use of antineoplastics (oth- the drugs reported in the exclusion cri- duction, which both counterbalance the er than methotrexate as antiarthritic teria), considered necessary for patient deleterious effects of prostaglandin therapy) during the 30 days preceding health and wellbeing, were allowed at depletion due to COX inhibition (34). the study; history of alcohol or drug the discretion of the Investigator. Activation of capsaicin receptors takes abuse; pregnancy or lactation. place through direct contact of the drug Assessment of safety with the gastric mucosa and therefore it Number of subjects Primary safety criteria were: gastric is maximal when the product is given A total of 304 patients (152 per each and/or duodenal damage (erosions, ul- on an empty stomach. This suggested treatment group) had to be randomized cers or both), assessed by an upper gas- the fasting administration of AMG in in the study in order to reach the 90% trointestinal endoscopy performed at clinical practice to better exploit its power of the test on the basis of the the beginning and at the end of the

811 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

of frequencies equivalence was tested by means of Fisher’s exact test for con- tingency tables; comparisons of safety parameters within and between groups was tested by Fisher’s exact test and, Week -1 0 4 12 24 where appropriate, by McNemar’s test Visit 1 2 3 4 5 and t-test for paired samples. Efficacy Fig. 1. parameters were evaluated by Fisher’s exact test. study; the assessment included the ex- and after 4, 12 and 24 weeks of treat- The statistical tests were two-sided at amination of the oesophagus, stomach, ment. The primary outcome measure the 5% level of significance. Statistical pyloric channel and duodenum. The was ACR-20 Responder Index (35, 36), analysis was performed by using SAS endoscopic findings were scored from representing the number of patients 8.2 statistical modules running under 0 to 7 as follows: 0 = normal mucosa; achieving 20% improvement in tender Windows OS. 1 = 1 – 10 petechiae; 2 = > 10 petechi- and swollen joint counts and 20% ae; 3 = 1 – 3 erosions; 4 = 4 – 10 ero- improvement in 3 of the 5 following Results sions; 5 = > 10 erosions; 6 = oozing or core set measures: Patient’s global as- Patients intraluminal blood; 7 = ulcer or visible sessment, Physician’s global assess- Data analysis was performed on the vessels. A break in the mucosa without ment, pain, functional disability, acute 180 patients that completed the study a fibrous base was considered as an phase reactants. (85 in the amtolmetin guacyl group and erosion whereas, when a fibrous base The Physician’s global assessment and 95 in the celecoxib group). Patients not was present, the mucosal lesion was the Patient’s global assessment were included in the analysis were consid- defined as an ulcer. Score 0-2 was con- scored from 1 (very good) to 5 (very ered withdrawals. Demographic and sidered normal findings, while score 3- poor) with higher scores indicating clinical characteristics of the two groups 7 was considered abnormal. greater disease activity. Pain was of patients (Table I) show that they Each patient was given a gastric score scored on a 4-point scale (mild, moder- were quite homogeneous. (based on gastric mucosal findings) ate, severe, incapacitating). Functional and a duodenal score (based on the py- Capacity was evaluated in accordance Safety loric channel and duodenal mucosal with the ACR classification (36). Endoscopy. Gastric and duodenal scores findings). Moreover, each patient was are considered separately and are given a “maximal” gastro-duodenal Statistical analysis reported in Tables IIa and IIb, while in score defined as the higher of the gas- Statistical analysis was performed on Table III the number of patients with tric or duodenal ones. the Per Protocol (PP) cohort, defined as normal or abnormal gastroduodenal Secondary safety criteria were: gas- all treated patients that had completed findings is reported. No differences be- trointestinal symptoms, ECG abnor- the study. When the GI safety data were tween the two treatments were observ- malities, blood and urine laboratory considered, also the Intent-to-Treat ed: in fact both groups, homogeneous tests and adverse events. Gastrointesti- (ITT) analysis was done, i.e. all ran- at baseline, remained comparable at the nal symptoms (including heartburn, domized patients who received at least end of treatment; both AMG and cele- epigastric and/or abdominal pain, nau- one dose of study medication were coxib failed to worsen gastro-duodenal sea, vomiting, dyspepsia, flatulence, taken into account. Indeed, both GI endoscopic findings. In actual fact, diarrhoea) were evaluated by using a symptoms and mucosal lesions can with both drugs, the number of patients specific questionnaire at the beginning occur even after one single administra- with normal gastric score increased, and after 4, 12 and 24 weeks of treat- tion of an NSAID (8, 10). For this kind albeit not significantly. At baseline, all ment. of analysis, the LOCF (Last Observa- patients with abnormal scores present- A 12-lead ECG and laboratory tests, tion Carried Forward) method was ap- ed only gastric erosions (i.e. score 3 or including haematological and bio- plied in the presence of missing or not 4); at the end of treatment, 1 patient in chemical analysis, performed on blood available observations. the group treated with AMG and 4 and urine samples collected under fast- Descriptive analysis was performed for patients treated with celecoxib had a ing conditions, were assessed at study all demographic variables. Analysis to score 7, corresponding to the presence entry and at study end or withdrawal. assess the normal distribution was per- of ulcers or visible vessels. All adverse events observed by the in- formed by Fisher’s exact test or, where As far as the duodenal score is con- vestigator or reported by the patients appropriate, by Wilcoxon two-sample cerned, the number of subjects with were recorded during the entire study test, applied in order to verify the normal score increased after treatment period. homogeneity of means at baseline. The with AMG, while it decreased after analysis of safety parameters was treatment with celecoxib. At the end of Assessment of efficacy based on the comparison of the propor- treatment, the patients with abnormal The assessment of efficacy was per- tions of patients with or without gas- scores were 3 and 6 in the AMG group formed at the beginning of the study trointestinal diseases. The hypothesis and celecoxib group, respectively.

812 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

Table I. Baseline characteristics of the patients included in the trial. coxib group, among 71 patients with normal endoscopy at baseline, 61 re- Characteristic AMG Celecoxib (n = 85) (n = 95) mained normal also at the final visit (1 patient was excluded due to lack of Age (yrs) 57.2 ± 11.4 55.3 ± 11.9 final endoscopy) and out of 23 who had Males/females (n) 18 / 67 21 / 74 abnormal endoscopy at baseline, 12 Weight (Kg) 73.3 ± 16.0 71.9 ± 13.6 Height (cm) 164.1 ± 8.7 164.6 ± 7.5 improved at the final visit, showing normal findings. Between groups, the Concomitant treatment for rheumatoid arthritis (n) Glucocorticoids 39 (44.7%) 48 (50.5%) comparison did not highlight statisti- Metotrexate 42 (49.4%) 44 (46.3%) cally significant differences in gastro- Other disease-modifying drugs 34 (40.0%) 42 (44.2%) duodenal endoscopy findings at base- Prior use of NSAIDs (n) 42 (49.4%) 48 (50.5%) line, nor at the final examination, as User of low-dose aspirin (n) 4 (4.7%) 4 (4.2%) well as the within-groups comparison Clinical evaluation of RA: (McNemar’s test, p = 0.6698 for cele- Tender joints (n) 21.9 ± 15.7 22.2 ± 15.9 coxib and p = 1.0000 for AMG). When Swollen joints (n) 11.0 ± 7.4 11.8 ± 7.4 Physician’s global assessment (score) 3.3 ± 0.8 3.0 ± 0.8 the number of patients with normal and Patient’s global assessment (score) 3.3 ± 0.7 3.2 ± 0.9 abnormal gastro-duodenal findings are Pain (score) 2.5 ± 0.6 2.4 ± 0.7 considered (Table III), both PP and ITT CRP (µg/ml) 16.5 ± 17.8 16.8 ± 23.0 analyses did not reveal any significant American College of Rheumatology functional class (n) change from baseline values after treat- I 6 (7.1%) 12 (12.6%) ment with either AMG or celecoxib. II 43 (50.6%) 53 (55.8%) III 36 (42.3%) 30 (31.6%) Gastrointestinal symptoms. Table IV lists gastrointestinal symptoms report- Values are expresses as means ± SD. ed by patients at the beginning and at Baseline characteristics did not differ significantly between groups. the end of the study. Statistical analysis showed no significant difference in any When maximal gastroduodenal scores also at the final visit and out of 21 who of them between the two treatments. are considered, among 64 patients of had abnormal endoscopy at baseline, However, the symptom evaluation per- the AMG group with normal endo- 10 improved at the final examination, formed after 4 weeks of therapy reveal- scopy at baseline, 54 remained normal showing normal findings. In the cele- ed for AMG a significant lower inci-

Table II. Gastric and duodenal scores at baseline and 24 weeks after drug treatment in patients given AMG or celecoxib. a

Treatment Visit Gastric score Normal Abnormal 01234567

AMG (n = 85) Baseline 48 (56.47%) 16 (18.82%) 0 (0%) 13 (15.29%) 8 (9.41%) Celecoxib (n = 95) Baseline 56 (59.95%) 16 (16.84%) 1 (1.05%) 12 (12.63%) 10 (10.53%) p = 0.9597 AMG (n = 85) 24 weeks 50 (58.82%) 12 (14.12%) 3 (3.53%) 11 (12.94%) 5 (5.88%) 3 (3.53%) 0 (0%) 1 (1.18%) Celecoxib (n = 94) 24 weeks 62 (65.96%) 8 (8.51%) 5 (5.32%) 8 (8.51%) 5 (5.32%) 1 (1.06%) 1 (1.06%) 4 (4.26%) p = 0.5532 b

Treatment Visit Duodenal score Normal Abnormal 012345 67

AMG (n = 85) Baseline 77 (90.59%) 4 (4.71%) 0 (0%) 2 (2.35%) 2 (2.35%) Celecoxib (n = 95) Baseline 86 (90.53%) 6 (6.32%) 0 (0%) 2 (2.11%) 1 (1.05%) p = 0.8950 AMG (n = 85) 24 weeks 78 (91.76%) 4 (4.71%) 0 (0%) 0 (0.%) 0 (0%) 0 (0%) 0 (0%) 3 (3.53%) Celecoxib (n = 95) 24 weeks 81 (85.26%) 6 (6.32%) 2 (2.11%) 2 (2.11%) 3 (3.16%) 0 (0%) 0 (0%) 1 (1.05%) p = 0.1818

Statistical analysis: Fisher’s exact test.

813 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

Table III. Number of patients given AMG or celecoxib with normal and abnormal findings at endoscopy at baseline and 24 weeks after drug treatment.

PP analysis ITT analysis Treatment Patients Time Patients with Patients with p Treatment Patients Time Patients with Patients with p normal abnormal normal abnormal findings findings findings findings (score 0-2) (score 3-7) (score 0-2) (score 3-7)

AMG 85 Baseline 64 (75.3%) 21 (24.7%) n.s. AMG 118 Baseline 94 (79.7%) 24 (20.3%) n.s. Celecoxib 95 Baseline 72 (75.8%) 23 (24.2%) Celecoxib 116 Baseline 90 (77.6%) 26 (22.4%)

AMG 85 24 weeks 64 (75.3%) 21 (24.7%) n.s. AMG 118 24 weeks 92 (78.0%) 26 (22.0%) n.s. Celecoxib 94 24 weeks 73 (77.7%) 21 (22.3%) Celecoxib 116 24 weeks 91 (78.4%) 25 (21.6%) dence of dyspepsia (Fisher’s exact test, increased, without exceeding in any Adverse events (AEs) p = 0.0301). case the upper reference value and pro- The incidence of adverse events (AEs), Electrocardiographic evaluation. At bably being of no clinical relevance. calculated on all randomized patients, screening, in the AMG group 92.9% of Blood haemoglobin values before and during the study period was similar for patients had normal tracings, while in after treatment were not significantly both drugs (Table VI). The percentage the celecoxib group 90.5% of patients different in patients of both groups, of patients with treatment-related gas- had normal tracings; at the final visit suggesting the absence of any gastroin- trointestinal AEs (reported in Table the normal tracings for AMG and cele- testinal bleeding during the study peri- VII) was 21.2% for the AMG group coxib groups were 94.1% and 93.6%, od. The mean values of systolic and and 26.5% for the celecoxib group. respectively. No statistically significant diastolic blood pressure, as well as Figure 2 shows the cumulative inci- differences in ECG findings were ob- heart rate, did not change during the dence of gastrointestinal events during served, neither in the between-groups treatment with both drugs. the study. comparison (at baseline Fisher’s exact Non-gastrointestinal adverse events test, p = 0.6004 and at final examina- Efficacy were 12 and 11 for AMG and celecoxib, tion Fisher’s exact test, p = 1.0000) nor Clinical improvement, assessed by respectively; the AMG-treated patients in the within-groups comparison (Mc- ACR-20 responder index, (Table V) did complained of headache (n = 3), aller- Nemar’s test: p=0.5637 for the AMG not show any differences between the gic skin reactions (n = 3), worsening of group and p= 0.1797 for the celecoxib two groups at the different time intervals pathology (n = 3), hypotension (n = 1) group). during treatment. In particular, both and abnormal laboratory parameters (n Laboratory tests. Laboratory tests on drugs showed a similar efficacy at 4, 12 = 2); the celecoxib treated patients blood and urine did not evidence any and 24 weeks. The observed response experienced tinnitus (n = 1), ECGraphic statistically significant alteration in the rates were comparable with those re- abonormalities (n = 1), allergic skin re- examined parameters except for uric ported in literature for conventional actions (n=3), abnormal laboratory par- acid, in the AMG group, which slightly NSAIDs in rheumatoid arthritis (37) . ameters (n = 4), hypertension (n = 1), haemorrhage (n = 1). Table IV. Gastrointestinal symptoms in patients given AMG or celecoxib at baseline and Serious adverse events. During the stu- 24 weeks after drug treatment. dy, 11 serious adverse events were re- ported in 7 patients (4 in the AMG Symptom AMG Celecoxib p AMG Celecoxib p group and 3 in the celecoxib group); T= 0 T = 0 24 weeks 24 weeks none of these events was considered Heartburn 3 (3.5%) 3 (3.2%) n.s. 5 (5.9%) 8 (8.4%) n.s. drug-related for AMG and only 1 (an- gioedema) was connected to celecoxib Epigastric pain 0 1 (1.0%) n.s. 4 (4.7%) 5 (5.3%) n.s. administration. Abdominal pain 0 0 n.s. 1 (1.2%) 4 (4.2%) n.s. Withdrawal. 54 patients over a total of Nausea 1 (1.2%) 0 n.s. 2 (2.3%) 2 (2.1%) n.s. 235 prematurely interrupted the study; Vomiting 0 1 (1.0%) n.s. 0 1 (1.0%) n.s. the most frequent reason for drop out Dyspepsia 0 2 (2.1%) n.s. 5 (5.9%) 4 (4.2%) n.s. was the lack of cooperation or consent Flatulence 0 1 (1.0%) n.s. 7 (8.2%) 6 (6.3%) n.s. withdrawal from patients (Table VIII). Only 11 withdrawals were due to treat- Diarrhoea 2 (2.3%) 0 n.s. 1 (1.2%) 2 (2.1%) n.s. ment-related adverse events: 7 for Other 0 0 n.s. 1 (1.2%) 0 n.s. AMG and 4 for celecoxib. These ad- Total 6 8 n.s. 26 32 n.s. verse events were gastrointestinal com- plaints (n = 4), allergic reaction (n = 1), Statistical analysis : Fisher’s exact test. hypotension (n = 1) and headache (n =

814 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

Table V. Evaluation of ACR-20 responders amongst patients with rheumatoid arthritis Table VII. Number of gastrointestinal given AMG or celecoxib for 24 weeks. adverse events in patients with rheumatoid arthritis given AMG or celecoxib for 24 Treatment Week ACR ACR p weeks. responders (%) non responders (%) Adverse event AMG Celecoxib AMG 0 1.2 98.8 nn Celecoxib 0 2.1 97.9 1.0000 Heartburn 7 13 AMG 4 24.7 75.3 Epigastric pain 7 7 Celecoxib 4 26.3 73.7 0.8649 Abdominal pain 4 5 AMG 12 34.1 65.9 Nausea 6 3 Celecoxib 12 37.9 62.1 0.6428 Vomiting 0 2 Dyspepsia 3 6 AMG 24 43.5 56.5 Flatulence 7 6 Celecoxib 24 51.6 47.4 0.2950 Diarrhoea 1 3 Fisher’s exact test. Lost of appetite 1 1 Esophagitis 0 1 Esophageal ulcer 1 2 Table VI. Drug-related adverse events in patients with rheumatoid arthritis given AMG or Gastric petechiae 3 1 celecoxib for 24 weeks. Gatric erosions 1 1 Parameter AMG Celecoxib Gastric ulcer 1 1 Duodenal petechiae 0 2 N° of patients with adverse events 32 34 Duodenal erosions 0 1 N° of patients with gastrointestinal adverse events 25 31 Duodenal ulcer 2 1 N° of total adverse events 57 68 Other 1 1 N° of gastrointestinal adverse events 45 57 Total 45 57

arthritis who were taking a NSAID Fig. 2. Cumulative inci- dence of GI adverse alone had upper gastrointestinal com- events in patients with plications over a period of 6 months; rheumatoid arthritis given the combined treatment determined a AMG or celecoxib for 24 40% reduction in the risk. weeks. Along the same lines, studies carried out with selective COX-2 inhibitors, like celecoxib (22) and rofecoxib (23), showed on average a 50% reduction in the incidence of clinically significant upper gastrointestinal events and ulcer complications, in comparison with non-selective NSAIDs. Amongst the different COX-2 inhibitors so far dev- eloped and available on the market, the selectivity ratio towards the two isoen- zymes varies greatly depending on the methodology employed (in vivo, ex- 1) for AMG; those reported for cele- the addition of a gastroprotective agent, vivo or in vitro assays, intact cells or coxib were gastrointestinal complaints such as proton pump inhibitor or miso- purified enzyme as drug target, varia-

(n = 2) and allergic reactions (n = 2). prostol, a synthetic PGE1 derivative, or tions in substrate concentration and the use of cyclooxygenase-2 selective incubation time) (41, 42). As a conse- Discussion inhibitors which, while maintaining the quence, classification of these agents Gastrointestinal toxicity induced by same activity of traditional NSAIDs in clinically is difficult because there are NSAIDs represent one of the most reducing arthritis pain and inflamma- insufficient data to predict the correla- common serious adverse drug events in tion, markedly improves the disease- tions between biochemical and phar- the industrialized world, with a signifi- related quality of life (38,39) as a result macological properties and the clinical cant impact both on public health and of their better gastrointestinal tolerabil- effect of a given agent. In any case, costs related to the management of side ity. A prospective study (40) that com- from a clinical standpoint selectivity, effects. Different strategies have been pared conventional NSAIDs alone with COX-2 selectivity should translate in adopted in order to reduce the NSAID- NSAIDs plus misoprostol reported that effective anti-inflammatory and anal- associated gastro-duodenal damage: 0.95% of patients with rheumatoid gesic activities while sparing COX-1

815 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

Table VIII. Reasons for withdrawal among patients with rheumatoid arthritis given AMG adverse renal events (e.g. patients with or celecoxib. congestive heart failure, renal or hepatic disease, as well as those of advanced Reason AMG Celecoxib (n = 33) (n = 21) age receiving therapy with diuretics or angiotensin-converting enzyme (ACE) Did not cooperate/withdrew consent 39.4% 52.4% inhibitors) should be monitored with the Adverse events 21.2% 19.0% same caution when receiving selective Intercurrent illness 15.2% 4.8% COX-2 inhibitors as when receiving Loss of follow-up 9.1% 9.5% treatment with conventional NSAIDs. It Lack of compliance/poor compliance 3.0% 0 is therefore evident that COX-2 selec- Lack of safety 3.0% 0 tive NSAIDs do not offer any advantage Other protocol violations 3.0% 0 Other reasons 6.1% 14.3% over non-selective drugs in terms of renal safety (54). On the contrary, AMG does not impair renal function, evaluat- associated physiological functions like, gastroduodenal mucosa but rather pro- ed with clinical and biochemical para- for instance, mucosal PG synthesis and tects it from noxious stimuli. Mucosal meters after repeated administration platelet aggregation (42, 43). Although protection is achieved via the inhibition (55). Here again, maintenance of renal in some assays, celecoxib displayed a of acid secretion, increase of bicarbon- blood flow by the released NO is likely low selectivity ratio (44, 45), it fulfills ate production and stimulation of mu- to be the underlying mechanism of the “clinical definition” of selective cosal blood flow, mechanisms all dri- AMG renal safety. COX-2 inhibitor and large clinical tri- ven by the increase of CGRP and NO Although the use of selective COX-2 als (19, 20) and meta-analyses (22) production (26,49-51). inhibitors in place of conventional have provided strong evidence for its It is now clearly established that mi- NSAIDs for the treatment of arthritis gastrosparing activity. crovascular damage represents the ear- appears to reduce the risk of serious The present study shows that both liest event following NSAID adminis- gastrointestinal toxicity, the role played AMG and celecoxib are similarly ef- tration (52), preceding and contributing by these inhibitors in the generation or fective in controlling the symptoms of to the development of epithelial le- exacerbation of ischaemic cardiovascu- rheumatoid arthritis. Like the therapeu- sions. In a recent experimental study lar disease is less clear. Clinical studies tic efficacy, the GI safety and tolerabil- (53), electron microscopic examination demonstrate that hypertension can be ity of both drugs are also similar, con- of gastric mucosa did show that even induced or aggravated by COX-2 inhi- firming the results of previous endo- the selective COX-2 inhibitor celecox- bitors to a degree similar or higher than scopic studies where AMG was com- ib induced, like conventional NSAIDs, that which occurs with non-selective pared to traditional NSAIDs (27, 28, marked endothelial damage. However, drugs (56). Endothelial dysfunction, an 46). under the same experimental condi- indicator of cardiac ischaemia, may al- The lack of correlation between symp- tions, both acute and chronic AMG ad- so be exacerbated by COX-2 inhibition toms and mucosal damage after ministration produced very limited da- (57) and there is much debate as to NSAID administration has been repeat- mage, most likely thanks to the release whether these changes lead to an abso- edly shown in clinical trials and the of the vasodilator and endothelial pro- lute increase in ischaemic cardiac ev- lack of symptoms cannot therefore be tective NO. ents (58). These effects on cardiovascu- considered a mirror of the lack of mu- Both COX isoforms (COX-1 and lar risk factors appear all more impor- cosal damage; in fact, a large propor- COX-2) are constitutively expressed in tant in patients with rheumatoid arthri- tion of serious NSAID induced gas- the adult mammalian kidney and con- tis where there is an increase in the in- trointestinal complications can be tribute to the biosynthesis of prosta- cidence of ischaemic heart disease asymptomatic (47, 48). In this study, noids. Inhibition of COX activity in the (59). both symptoms and mucosal damage kidney by NSAIDs has relatively mild Soon after the introduction of celecox- have been investigated by specific consequences in healthy individuals, ib and rofecoxib into the market, it was questionnaires and GI endoscopy, re- but can lead to serious adverse events reported that both drugs suppressed the spectively; the subjective and objective in patients whose renal function is PG formation of prostacyclin I2 (PGI2) in evaluations highlight the remarkable dependent. Most studies have reported healthy volunteers (60). PGI2 had pre- gastroduodenal tolerability of AMG transient decreases in sodium excretion viously been shown to be the predomi- and celecoxib, with a similar incidence upon initiation of therapy with either nant COX product in endothelium, of adverse events possibly or probably traditional NSAIDs or coxibs that, in inhibiting platelet aggregation, causing related to treatment; in particular, gas- patients whose renal function is depen- vasodilatation and preventing the pro- trointestinal events resulted for both dent on prostanoids, can affect the liferation of vascular smooth muscle drugs comparable in frequency, intensi- glomerular filtration rate (GFR). cells in vitro (61). However, it was as- ty and type. Changes in renal function may result in sumed that PGI2 was derived mainly Despite COX inhibition and prostanoid hypertension and edema; as a conse- from COX-1, the only isoform ex- depletion, AMG does not damage the quence, patients who are at risk for pressed constitutively in endothelial

816 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al. cells. This assumption later proved to fit/risk of therapies. Clin Exp Rheumatol 19. EMERY P, ZEIDLER H, KWIEN T et al.: Ce- be incorrect, since studies in mice and 2004; 22 (Suppl. 35): S2-11. lecoxib versus diclofenac in long term man- 4. HENRY DA: Side effects of non steroidal agement of rheumatoid arthritis: randomized, humans showed that COX-2 was the anti-inflammatory drugs. Balliere’s Clin double blind comparison. Lancet 1999; 354: dominant source (62, 63). The individ- Rheumatol 1988; 2: 425-54. 2106-11. 5. BRATER DC: Anti-inflammatory agents and 20. SILVERSTEIN FE, FAICH G, GOLDSTEIN JL ual cardiovascular effects of PGI2 in vitro contrast with those of thrombox- renal function. Semin Arthritis Rheum 2002; et al.: Gastrointestinal toxicity with celecox- 32 (Suppl. 1): 33-42. ib vs non steroidal anti-inflammatory drugs ane A2 (TXA2), the major COX-1 prod- 6. WOLFE MM, LICHTENSTEIN DR, SINGH G: for osteoarthritis and rheumatoid arthritis. uct of platelets, which causes platelet Gastrointestinal toxicity of non-steroidal The CLASS study: a randomized, controlled aggregation, vasoconstriction and vas- anti-inflammatory drugs. N Eng J Med 1999; trial. JAMA 2000; 284 (10):1247-55. cular proliferation (64). Whereas tradi- 340: 1888-99. 21. BOMBARDIER C, LAINE L, REICIN A et al.: tional NSAIDs inhibit both TXA and 7. MACDONALD TM, MORANT SV, ROBINSON Comparison of upper gastrointestinal toxicity 2 GC et al.: Association of upper gastrointesti- of rofecoxib and naproxen in patients with PGI2, the coxibs leave TXA2 genera- nal toxicity of non-steroidal anti-inflammato- rheumatoid arthritis. N Engl J Med 2000; tion unaffected, reflecting the absence ry drugs with continued exposure: cohort 343: 1520-8. of COX-2 in platelets, and thus in- study. BMJ 1997; 315: 1333-7. 22. DEEKS JJ, SMITH LA, BRADLEY MD: Effica- crease thrombotic risk, predisposing 8. LANGMAN MJ, WEIL J, WAINWRIGHT P et cy, tolerability, and upper gastrointestinal al.: Risks of bleeding peptic ulcer associated safety of celecoxib for treatment of osteo- patients to heart attack and stroke (64). with individual non-steroidal anti-inflamma- arthritis and rheumatoid arthritis: systematic As a matter of fact, a series of epidemi- tory drugs. Lancet 1994; 343: 1075-8. review of randomised controlled trials. BMJ ologic analyses (65, 66) have also rais- 9. HASLOCK I: Review of induced upper gas- 2002 ; 325: 619-24. ed questions about the cardiovascular trointestinal morbidity and mortality. In 23. WATSON DJ, YU Q, BOLOGNESE JA, REICIN CHELI R (Ed.): Treatment and Prevention of AS, SIMON TJ: The upper gastrointestinal safety of the coxibs and both the US NSAID-Induced Gastropathy. Royal Society safety of rofecoxib vs. NSAIDs: an updated FDA and EMEA(67, 68), followed by of Medicine Services International Congress combined analysis. Curr Med Res Opin the Italian Drug Agency [AIFA (69)], and Symposium Series 1989; 147: 3-10. 2004; 20: 1539-48. have issued warnings about the cardio- 10. GARCIA-RODRIGUEZ LA, JICK H: Risk of 24. TUBARO E, BELOGI L, MEZZADRI CM, RUCO upper gastrointestinal bleeding and perfora- L, STOPACCIARO A: Studies on the gastric vascular risk of COX-2 selective inhi- tion associated with individual non steroidal tolerability of the new non steroidal anti-in- bitors as well as non-selective NSAIDs anti-inflammatory drugs. Lancet 1994; 343: flammatory drug amtolmetin guacyl. Arz- and given guidance on their use in pa- 769-72. neim Forsch/Drug Res 1995; 45: 1298-302. tients with cardiovascular risk factors. 11. SCARPIGNATO C, PELOSINI I: Prevention 25. TUBARO E, BELOGI L, MEZZADRI CM: and treatment of non-steroidal anti-inflam- Anti-inflammatory and antiplatelet effect of Conversely from COX-2 selective matory drug-induced gastro-duodenal dam- Amtolmetin Guacyl, a new gastroprotective NSAIDs which do not affect platelet age: rationale for the use of antisecretory non-steroidal anti-inflammatory drug. Arz- aggregation (70-72), AMG displays compounds. Ital J Gastroenterol Hepatol neim Forsch/Drug Res 2001, 51: 737-42. both in vitro and ex vivo an antiplatelet 1999; 31 (Suppl. 1): S63-72. 26. TUBARO E, BELOGI L, MEZZADRI CM: The 12. SCARPIGNATO C: Non steroidal anti-inflam- mechanism of action of amtolmetin guacyl, a activity comparable to that of aspirin matory drugs: how do they damage gastro- new gastroprotective non-steroidal anti- (25). This pharmacological action duodenal mucosa? Dig Dis 1995; 13 (Suppl. inflammatory drug. Eur J Pharmacol 2000; should render the drug more suitable 1): 9-39. 387: 233-44. than selective COX-2 inhibitors for 13. REDFERN JS, FELDMAN M: Role of prosta- 27. BIANCHI PORRO G, MONTRONE F, LAZZA- glandins in preventing gastrointestinal ulcer- RONI M, MANZIONNA G, CARUSO I: Clini- patients with one or more CV risk fac- ation: Induction of ulcers by antibodies to cal and gastroscopic evaluation of amtol- tors. prostaglandins. Gastroenterology 1989; 96: metin guacyl versus diclofenac in patients Taken together, all these data allow the 596-605. with rheumatoid arthritis. Ital J Gastroen- conclusion that AMG is an effective 14. HALTER F, TARNAWSKI AS, SCHMASS- terol Hepatol 1999; 3: 378-85. anti-inflammatory drug, with a safety MANN A, PESKAR BM: Cyclooxygenase 2- 28. TAVELLA G, URSINI G: Studio clinico sul- implications on maintenance of gastric l’attività antinfiammatoria e sulla tollerabili- profile encompassing the GI tract, the mucosa integrity and ulcer healing: contro- tà gastroenterica di amtolmetin guacil, un kidney and the CV system. Due to versial issues and perspectives. Gut 2001; 49: nuovo FANS, in confronto a diclofenac, su these properties and its peculiar mecha- 443-53. pazienti anziani con patologie osteoarticolari. nism of action, this compound seems to 15. WALLACE JL, DEVCHAND PR: Emerging Clin Ter 1997; 148: 543-8. roles for cyclooxygenase-2 in gastrointesti- 29. MONTRONE F, SANTANDREA S, CARUSO I et be particularly suitable for long-term nal mucosal defence. Br J Pharmacol 2005; al.: Amtolmetin guacyl versus piroxicam in use, especially in older subjects and 145: 275-82. patients with osteoarthritis. J Int Med Res patients with renal or CV disease. 16. WU CY, WU MS, CHEN CJ, LI MC, LIN JT, 2000; 28: 91-100. CHEN GH: The interaction of H. pylori infec- 30. MARCOLONGO R, FREDIANI B, BIASI G, tion and NSAIDs in cyclooxygenase-2 MINARI C, BARRECA C: Metanalysis of the References mRNA expression in gastric antral, corpus tolerability of amtolmetin guacyl, a new, effi- 1. SOKKA T: Work disability in yearly rheuma- mucosa, and gastric ulcer. J Clin Gastroen- cacious, non-steroidal anti-inflammatory toid arthritis. Clin Exp Rheumatol 2003; 21 terol 2005; 39: 50-5. drug, compared with traditional NSAIDs. (Suppl. 31): 571-4. 17. SCARPIGNATO C, BJARNASON I, BRE- Clin Drug Invest 1999; 17: 89-96. 2. SCOTT DL, SMITH C, KINGSLEY G: Joint TAGNE J-F et al.: Working Team Report: 31. DE PRETIS G, TASINI E: Tollerabilità gastri- damage and disability in rheumatoid arthri- Towards a GI Safer Antiinflammatory Thera- ca dell’antiinfiammatorio Amtolmetina Gua- tis: an updated systematic review Clin Exp py. Gastroenterology Int 1999; 12: 186-215. cil: meta-analisi dei trials terapeutici. Medici- Rheumatol 2003; 21 (Suppl. 31): 520-7. 18. BENSEN WG, FIECHTNER JJ, MCMILLEN JI na Terapia 2002. 3. PINCUS T, KAVANAUGH A, SOKKA T: Ben- et al.: Treatment of osteoarthritis with cele- 32. TUBARO E, BELOGI L, MEZZADRI CM, BET- efit/risk of therapies for rheumatoid arthritis: coxib, a cyclooxygenase 2 inhibitor: a ran- TELLI E: Impact on the bowel of amtolmetin underestimation of the “side effects” or risks domized, controlled trial. Mayo Clin Proc guacyl, a new gastroprotective non-steroidal of RA leads to underestimation of the bene- 1999; 74: 1095-105. anti-inflammatory drug. Eur J Pharmacol

817 Safety and efficacy of amtolmetin guacyl and celecoxib / Z. Jajic´ et al.

2003; 467: 173-83. cal selectivity of novel COX-2 inhibitors in ALD GA: Cardiovascular hazard and non- 33. RIEZZO G, CHILOIRO M, MONTANARO S: whole blood assays of COX-isozyme activi- steroidal anti-inflammatory drugs. Curr Opin Protective effect of amtolmetin guacyl versus ty. Curr Med Res Opin 2002; 18: 503-11. Pharmacol 2005; 5: 204-10. placebo, a typical NSAID and misoprostol in 46. LAZZARONI M, ANDERLONI A, BIANCHI 59. WALLBERG-JONSSON S, CEDERFELT M, healthy volunteers evaluated as gastric elec- PORRO G: The effects on gastroduodenal RANTAPAA DAHLQVISP S: Hemostatic fac- trical activity in alcohol induced stomach mucosa of a new non-steroidal anti-inflam- tors and cardiovascular disease in active damage. Dig Dis Sci 2001; 46: 1797-804. matory drug, amtolmetin guacyl, versus piro- rheumatoid arthritis: an 8-year follow-up 34. PISANO C, GRANDI D, MORINI G et al.: xicam in healthy volunteers: a short term, study. J Rheumatol 2000; 27: 71-5. Gastrosparing effect of new anti-inflammato- double blind, endoscopically controlled 60. FITZGERALD GA: COX-2 and beyond: ap- ry drug amtolmetin guacyl in the rat: involve- study. Eur J Gastroenterol Pathol 2001; 13: proaches to prostaglandin inhibition in hu- ment of nitric oxide. Dig Dis Sci 1999; 44: 833-9. man disease. Nat Rev Drug Discov 2003; 2: 713-24. 47. HAWKEY CJ, LANGSTROM G, NAESDAL J 879-99. 35. FELSON DT, ANDERSON JJ, BOERS M et al.: et al.: Significance of dyspeptic symptoms 61. SCHROR K, WEBER AA: Roles of vasodila- The American College of Rheumatology pre- during healing and maintenance of NSAID tatory prostaglandins in mitogenesis of vas- liminary core Set of disease activity mea- associated gastroduodenal lesions with cular smooth muscle cells. Agents Actions sures for rheumatoid arthritis clinical trials. omeprazole, misoprostol and ranitidine. Gas- 1997; 48 (Suppl.): 63-91. Arthritis Rheum 1993, 36: 729-40. troenterology 1997; 112: A144. 62. EGAN KM, LAWSON JA, FRIES S et al.: 36. FELSON DT et al.: American College of 48. SINGH G, RAMEY DR, MORFELD D, SHI H, Cyclooxygenase-2-derived prostacyclin con- Rheumatology. Preliminary definition of im- HATOUM HT, FRIES JF: Gastrointestinal fers atheroprotection on female mice. Obstet provement in rheumatoid arthritis. Arthritis tract complications of non-steroidal anti- Gynecol Surv 2005; 60: 309-10. Rheum 1996, 39: 535-7. inflammatory drug treatment in rheumatoid 63. MCADAM BF, CATELLA-LAWSON F, MARDI- 37. SIMON LS, WEAWER AL, GRAHAM DY et arthritis. Arch Intern Med 1996; 156: 1530-6. NI IA, KAPOOR S, LAWSON JA, FITZGER- al.: The anti-inflammatory and upper gas- 49. CORUZZI G, COPPELLI G, SPAGGIARI S et ALD GA: Systemic biosynthesis of prostacy- trointestinal effects of celecoxib in rheuma- al:. Gastroprotective effects of amtolmetin clin by cyclooxygenase (COX)-2: the human toid arthritis: a randomized, controlled trial. guacyl: a new non-steroidal anti-inflammato- pharmacology of a selective inhibitor of COX. JAMA 2000; 82: 1921-8. ry drug that activates inducible gastric nitric Proc Natl Acad Sci USA 1999; 96: 272-7. 38. MITCHELL JA, WARNER TD: Cyclooxyge- oxide synthase. Dig Liv Dis 2002;34:403-10. 64. FITZGERALD GA: Coxibs and cardiovascu- nase 2: pharmacology, physiology, biochem- 50. BRZOZOWSKI T, KONTUREK PC, KONTUR- lar disease. N Eng J Med 2004; 351: 1709-11. istry and relevance to NSAID therapy. Br J EK SJ et al.: Gastroprotective and ulcer heal- 65. MUKHERJEE D, NISSEN SE, TOPOL EJ: Risk Pharmacol 1999; 128: 1121-32. ing effects of nitric oxide releasing non- of cardiovascular events associated with 39. ZHAO SS, MCMILLEN JI, MARKENSON JA steroidal anti-inflammatory drugs. Dig Liv selective Cox-2 inhibitors. JAMA 2001; 286, et al.: Evaluation of functional status aspects Dis 2000; 32: 583-94. 954-9. of health related quality of life of patients 51. TEPPERMANN BL, SOPER BD: Nitric oxide 66. RAY WA, STEIN CM, DAUGHERTY JR, HALL with osteoarthritis treated with celecoxib. synthase induction and cytoprotection of rat K, ARBOGAST PG, GRIFFIN MR: COX-2 Pharmacotherapy 1999; 19: 1269-78. gastric mucosa from injury by ethanol. Can J selective non-steroidal anti-inflammatory 40. SILVERSTEIN FE, GRAHAM DY, SENIOR JR Physiol Pharmacol 1994; 72: 1308-12. drugs and risk of serious coronary heart dis- et al.: Misoprostol reduces serious gastroin- 52. LOPEZ-BELMONTE J, WHITTLE BJR, MON- ease. Lancet 2002; 360: 1071-3. testinal complications in patients with rheu- CADA S: The actions of nitric oxide donors 67. EMEA/62838/2005 Public statement: Euro- matoid arthritis receiving non-steroidal anti- in the prevention or induction of injury to the pean Medicines Agency announces regulato- inflammatory drugs: a randomized, double rat gastric mucosa. Br J Pharmacol 1993; ry action on COX-2 inhibitors [http://www. blind, placebo controlled trial. Ann Intern 108: 73-8. emea.eu.int/htms/hotpress/ d6275705.htm]. Med 1995; 123: 241-9. 53. MORINI G, GUAIATA E, LAZZARETTI M, 68. FDA Public Health Advisory: FDA announ- 41. SCARPIGNATO C, LAZZARONI M, PELOSINI GRANDI D, CORUZZI G: Morphological fea- ces important changes and additional warn- I, BIANCHI PORRO G: Selective inhibitors of tures of rat gastric mucosa after acute and ings for COX-2 selective and non-selective cyclooxyganase-2: myth or reality ? [in Ital- chronic treatment with amtolmetin guacyl: non-steroidal anti-inflammatory drugs ian]. Arg Gastroenterol Clin 1997; 10: 153- comparison with non-selective and COX2 (NSAIDs) [http://www.fda.gov/cder/drug/ad- 75. selective NSAIDs. Digestion 2003; 68: 124- visory/ COX2.htm]. 42. BROOKS P, EMERY P, EVANS JF et al.: In- 32. 69. AIFA: Comunicato n. 5 del 17 febbraio 2005. terpreting the clinical significance of the dif- 54. SCHNITZER TJ: Cyclooxygenase-2-specific COX 2, AIFA adotta misure a tutela salute ferential inhibition of cyclooxygenase-1 and inhibitors: are they safe ? Am J Med 2001; cittadini [http://www.agenziafarmaco.it/com cyclooxygenase-2. Rheumatology 1999; 38: 110 (1A): 46S-49S. _n5_17_02_2005.html]. 779-88. 55. NICCOLI L, BELLINO S, CANTINI F: Renal 70. LEESE PT, HUBBARD RC, KARIM A, ISAK- 43. LIPSKY LP, ABRAMSON SB, CROFFORD L, tolerability of three commonly employed SON PC, YU SS, GEIS GS: Effects of celecox- DUBOIS RN, SIMON LS, VAN DE PUTTE LB: non-steroidal anti-inflammatory drugs in ib, a novel cyclooxygenase-2 inhibitor, on The classification of cyclooxygenase inhibi- elderly patients with osteoarthritis. Clin Exp platelet function in healthy adults: a random- tors. J Rheumatol 1998; 25: 2298-303. Rheumatol 2002; 20: 201-7. ized, controlled trial. J Clin Pharmacol 2000; 44. WARNER TD, GIULIANO F, VOJNOVIC I, 56. AW TJ, HAAS SJ, LIEW D, KRUM H: Meta- 40: 124-32. BUKASA A, MITCHELL JA, VANE JR: Nons- analysis of cyclooxygenase-2 inhibitors and 71. WILNER KD, RUSHING M, WALDEN C et al.: teroid drug selectivities for cyclo-oxygenase- their effects on blood pressure. Arch Intern Celecoxib does not affect the antiplatelet 1 rather than cyclo-oxygenase-2 are associat- Med 2005; 165: 490-6. activity of aspirin in healthy volunteers. J ed with human gastrointestinal toxicity: a full 57. JUSTICE E, CARRUTHERS DM: Cardiovas- Clin Pharmacol 2002; 42: 1027-30. in vitro analysis. Proc Natl Acad Sci USA cular risk and COX-2 inhibition in rheuma- 72. CATELLA-LAWSON F, REILLY MP, KAPOOR 1999; 96: 7563-8. tology practice. J Hum Hypertens 2005; 19: SC et al.: Cyclooxygenase inhibitors and the 45. TACCONELLI S, CAPONE ML, SCIULLI MG, 1-5. antiplatelet effects of aspirin. N Engl J Med RICCIOTTI E, PATRIGNANI P: The biochemi- 58. WONG D, WANG M, CHENG Y, FITZGER- 2001; 345: 1809-17.

818