Adverse Drug Reactions Associated with the Use of Nsaids: a Case

doi: 10.1111/j.1472-8206.2011.00991.x

ORIGINAL Adverse drug reactions associated with the ARTICLE use of NSAIDs: a case/noncase analysis of

gy spontaneous reports from the French

lo pharmacovigilance database 2002–2006 o Maryse Lapeyre-Mestrea,b,c*, Sabrina Grolleaua,b, c Jean-Louis Montastruc a,b,c, Association Franc¸aise des Centres Re´gionaux a de Pharmacovigilance (CRPV)a a

m Service de Pharmacologie Clinique, Centre Midi-Pyre´ne´es de Pharmacovigilance, de Pharmacoe´pide´miologie et d’Information sur le Me´dicament, Hoˆpitaux de Toulouse, Toulouse, France r bFaculte´ de Me´decine, Universite´ Paul Sabatier, Universite´ de Toulouse, Toulouse, France a cUnite´ mixte INSERM 1027, Equipe de Pharmacoe´pide´miologie, Toulouse, Toulouse, France h P

Keywords ABSTRACT adverse drug reactions, gastrointestinal system, To evaluate the safety proﬁle of eight oral nonsteroidal anti-inﬂammatory drugs liver, (NSAIDs) available in France, using data reported through the French pharmaco- NSAID(s), vigilance system. Data (from 2002 to 2006) were analysed for aceclofenac, seriousness, diclofenac, ketoprofen, meloxicam, naproxen, nimesulide, piroxicam and tenoxicam, Clinical skin focusing on the reported rates of serious adverse drug reactions (ADRs) in the following system organ classes: gastrointestinal, hepatic, cutaneous, renal and

Received 8 October 2010; cardiovascular. A total of 42 389 serious ADR reports were identiﬁed, and 38 506 revised 27 June 2011; were included in a case/noncase analysis. Ketoprofen was associated with the highest ntal & accepted 11 August 2011 cumulative reported rate of serious ADRs (0.78 cases per million deﬁned daily doses), e followed by diclofenac (0.58), nimesulide (0.52), naproxen (0.50), piroxicam (0.47), tenoxicam (0.42), meloxicam (0.41) and aceclofenac (0.30). The most frequently *Correspondence and reprints: maryse.lapeyre-mestre@ reported serious ADRs were cutaneous, followed by gastrointestinal, hepatic, renal univ-tlse3.fr and rarely, cardiovascular events. In the case/noncase analysis, ketoprofen, piroxicam and naproxen were associated with the highest risk of serious gastrointestinal ADRs (odds ratios [ORs] of 6.87, 6.54 and 5.07, respectively). Nimesulide

Fundam and aceclofenac were associated with the highest risk of liver ADRs (adjusted ORs of 4.53 and 3.67, respectively), as was meloxicam for cutaneous ADRs (adjusted OR of 3.15) and tenoxicam for renal ADRs (adjusted OR of 3.17). The most frequent serious ADRs reported with the selected oral NSAIDs are cutaneous, followed by gastrointestinal, hepatic and renal events. The highest risks for serious gastrointestinal, hepatic, cutaneous and renal adverse events were linked, respectively, with ketoprofen, nimesulide, meloxicam and tenoxicam compared with the other NSAIDs.

possess different chemical and clinical profiles but INTRODUCTION essentially exert the same therapeutic properties and Nonsteroidal anti-inflammatory drugs (NSAIDs) are are associated with similar adverse effects. Gastrointes- widely used to treat a large number of common acute tinal injuries, which range from dyspepsia to fatal upper and chronic inflammatory conditions. These drugs gastrointestinal tract bleeding and perforation, are

ª 2011 The Authors Fundamental and Clinical Pharmacology ª 2011 Socie´ te´ Française de Pharmacologie et de The´ rapeutique Fundamental & Clinical Pharmacology 27 (2013) 223–230 223 224 M. Lapeyre-Mestre et al. among the most common adverse drug reactions (ADRs) corticosteroids, aspirin or anticoagulants, serious com- associated with the use of NSAID [1–3]. Other important orbidity or a history of upper gastrointestinal events ADRs include skin reactions, renal complications, aller- [10]. gic reactions, alteration in hepatic enzyme levels and, The aim of this pharmacoepidemiological study was to rarely, hepatopathies [4–7]. Hepatic ADRs associated compare the safety profile of eight-first-generation NSA- with NSAIDs are quite uncommon when compared with IDs available in France, by assessing the reporting of other pharmacological classes such as antibiotics [6,8]. serious ADRs in several system organ classes [SOCs] from Epidemiological studies have reported the incidence of data in the French pharmacovigilance system from 2002 acute liver injury to be 1–9 cases per 100 000 NSAID to 2006. users [6,7]. As recommended by the European Medicine Agency METHODS (EMA) [9], NSAIDs should be prescribed at the lowest effective dose and for the shortest time necessary to This study analysed serious ADRs spontaneously re- control symptoms. Moreover, other factors such as the ported to the French pharmacovigilance system between overall safety profile of each compound (e.g. potential 1st January 2002 and 31st December 2006 for NSAIDs gastrointestinal and other concerns) and individual risk available in France. It has been compulsory since 1995 factors must be considered for improved patient man- for all prescribers in France to report any ADR defined as agement [10]. ‘serious’ and/or ‘unexpected’ [16]. ‘Serious’ ADRs are Differences in safety profiles between NSAIDs are a key defined as those being fatal, life-threatening, causing discriminator for choosing between one NSAID and hospitalization, resulting in persistent or significant another, and not all NSAIDs have the same level of risk disability or incapacity, requiring intervention to prevent for inducing gastrointestinal ADRs. First-generation permanent damage or causing a congenital anomaly NSAIDs are nonselective and block the activity of both [16]. This study focused on serious ADRs associated with COX-1 and COX-2. Second-generation agents, the COX-2 gastrointestinal, hepatic, cutaneous, renal and cardio- inhibitors, were developed with the aim of reducing the vascular systems. incidence of gastrointestinal adverse effects by sparing Serious ADRs of interest were identified from specific the gastrointestinal protective functions of COX-1 while codes of the World Health Organisation Adverse Reac- still inhibiting inflammation [11,12]. However, it was tions Terminology (WHO-ART) classification of selected clear reasonably soon after their commercialization that organ class disorders (Table I). the use of these COX-2 inhibitors (i.e. rofecoxib and valdecoxib) was associated with an increase in cardio- Study design vascular risk, leading to their withdrawal from the The data collected were analysed for the reporting of market [13,14]. These events led to a revision of the serious ADRs in the population treated with the selected overall safety profile of all NSAIDs, particularly in regard NSAIDs and to calculate the odd ratios (ORs) of the to their renal and cardiovascular risk profile. Moreover, association of these ADRs with the NSAIDs in this a number of studies evaluating the gastrointestinal population, using a case/noncase analysis. tolerability of COX-2 inhibitors did not demonstrate a For the estimation of ADR reporting, data were lower risk of upper or lower gastrointestinal events analysed and compared for eight oral formulations of [15,16]. NSAIDs: aceclofenac, diclofenac, ketoprofen, meloxicam, Several factors influence a patient’s risk of developing naproxen, nimesulide, piroxicam and tenoxicam. different ADRs while receiving NSAIDs; for example, Case/noncase analysis has been widely used in the elderly patients or those with cardiovascular disorders, last decade in several spontaneous reporting data- chronic renal disease, rheumatoid arthritis or chronic bases, including the French pharmacovigilance database obstructive pulmonary disease are at greater risk of [17,18]. For example, this analytical method was applied cardiovascular ADRs [15]. Gastrointestinal ADRs are to data from the French and Spanish pharmacovigilance important indicators when evaluating the overall risk national databases to compare the risk of hepatic profile of NSAIDs because of the risk of gastrointestinal damage with the administration of different NSAIDs bleeding and perforation. Risk factors for gastrointestinal [19]. In this method, ‘cases’ are the ADRs of interest and complications, according to the National Institute for the ‘noncase’ control group corresponds to all other Clinical Excellence, include age over 65 years, the use of ADRs recorded in the database. The principle is to

Table I World health organisation adverse reactions terminology (WHO-ART) codes for selected system organ class (SOC) disorders.

WHO-ART codes SOC disorder

0700 Liver and biliary system disorders (hepatic failure or necrosis, abnormal hepatic functions, hepatitis, cholestasis, increase in liver enzymes) 0600 Gastrointestinal system disorders (ulcer, gastritis, gastrointestinal haemorrhage, intestinal perforation…) 0100 Skin and appendages disorders (dermatitis, eruption, toxic epidermic necrosis, drug-related eruption, Lyell and Stevens–Johnson syndrome…) 1300 Urinary system disorders (nephritis, renal failure, decrease in renal clearance, tubular nephropathy…) 1020 Myo-, endo-, pericardial and valve disorders (angor, myocardial infarction, coronary thrombosis, coronary arteriosclerosis) 1040 Vascular (extracardiac) disorders (cerebral infarction, embolism; arterial, peripheral embolism/ischaemia; arterial, carotid, cerebral, upper limb thrombosis…) 1810 Body as a whole – general disorders (ischaemic necrosis)

compare drug exposure (i.e. all drugs taken before the risk of confounding. All analyses and calculations were drug-related event) among cases and noncase controls. executed using SAS Version 9.2 of the SAS System for Windows (SAS Institute Inc., Cary, NC, USA). Statistical analysis Reporting rate of ADRs RESULTS The reporting rate of the total number of serious ADRs and the reporting of serious ADRs related to the selected Reporting rate of ADRs system organ class (SOC) was computed for each NSAID Of the 42 389 ‘serious’ ADRs that were recorded in the by dividing the number of ADRs reported by the total French pharmacovigilance database between 1st Janu- number of defined daily doses (TDDDs) consumed in the ary 2002 and 31st December 2006, 34 748 (81.9% of same period and multiplying by 106 (reporting rates are the total number) involved hospitalization; 1067 (2.5%) expressed as number of adverse reactions per million led to disability or sequelae, 5152 (12.2%) were life- DDD). For each reporting rate, a 95% confidence interval threatening, and 1422 (3.4%) resulted in death. (CI) was also computed using the exact method based on In terms of serious ADRs overall reporting rate (number the Poisson distribution. of adverse reactions per million DDDs), ketoprofen had the The TDDDs of each NSAID considered in the study were highest reporting rate of NSAIDs (0.78), followed by estimated from the sales figures in standard units obtained diclofenac (0.58), nimesulide (0.52), naproxen (0.50), from the IMS-Health, while the DDD for each compound piroxicam (0.47), tenoxicam (0.42), meloxicam (0.41) was obtained from the WHO Collaborating Centre for Drug and aceclofenac (0.30) (Table II). Statistics Methodology. The DDDs for oral formulations of Of the eight NSAIDs examined, nimesulide, tenoxicam the drugs of interest, checked on 29th June 2007 (http:// and piroxicam had the lowest proportions of reported www.whocc.no/atcddd/), were aceclofenac, 200 mg; diclo- ADRs that were life-threatening or resulted in death fenac, 100 mg; ketoprofen, 150 mg; naproxen, 500 mg; (9.6%, 8.0% and 6.6%, respectively) compared with nimesulide, 200 mg; piroxicam, 20 mg; tenoxicam, 20 mg; aceclofenac (22.7%), naproxen (16.1%), diclofenac and meloxicam, 15 mg. (14.5%), meloxicam (13.0%) and ketoprofen (11.9%).

Case/noncase analysis Reporting of ADRs by SOCs The case/noncase method was used to determine whether The cumulative reporting rates of ADRs according to the there is a statistically significant association between the SOCs for the eight NSAIDs evaluated are summarized in event and the NSAIDs’ exposure. In this study, the case/ Table III. noncase method was applied to compare the reports corresponding to specific ADRs in different SOCs, defined Gastrointestinal ADRs as ‘cases’, with the other reports in the database, called Ketoprofen had the highest cumulative reporting rate for ‘noncases’, for each NSAID examined. The association gastrointestinal ADRs (0.17 cases per million DDDs), between each ADR and the selected NSAIDs was quan- followed by piroxicam (0.10), diclofenac (0.10), nap- tified by calculating an OR of exposure to each drug with roxen (0.09), nimesulide and tenoxicam (both 0.05), its 95% CI, adjusted on age and gender, to minimize the aceclofenac (0.04) and meloxicam (0.04).

Table II Annual reporting rates [with the 95% CI] of serious adverse reactions (ADRs); the number of ADRs per million deﬁned daily doses (DDDs) for each of the eight nonsteroidal anti-inﬂammatory drugs (NSAIDs), as reported in the French pharmacovigilance database from 2002 to 2006.

NSAID 2002 2003 2004 2005 2006 Overall (2002–2006)

Aceclofenac aa0.16 [0.02–0.64] 0.52 [0.28–0.88] 0.18 [0.07–0.42] 0.30 [0.19–0.46] Diclofenac 0.59 [0.46–0.75] 0.63 [0.49–0.80] 0.65 [0.51–0.82] 0.56 (0.28–0.88) 0.47 [0.36–0.61] 0.58 [0.52–0.64] Ketoprofen 0.96 [0.78–1.17] 0.70 [0.55–0.87] 0.95 [0.79–1.11] 0.71 [0.58–0.86] 0.65 [0.53–0.79] 0.78 [0.72–0.85] Naproxen 0.49 [0.34–0.69] 0.50 [0.34–0.70] 0.61 [0.44–0.83] 0.58 [0.40–0.80] 0.29 [0.17–0.47] 0.50 [0.42–0.58] Nimesulide 0.42 [0.21–0.76] 0.48 [0.26–0.80] 0.71 [0.45–1.07] 0.56 [0.34–0.87] 0.40 [0.22–0.66] 0.52 [0.41–0.64] Meloxicam 0.66 [0.26–1.37] 0.49 [0.16–1.15] 0.31 [0.08–1.01] 0.39 [0.12–0.91] 0.24 [0.06–0.78] 0.41 [0.26–0.62] Piroxicam 0.52 [0.39–0.67] 0.42 [0.31–0.56] 0.47 [0.35–0.62] 0.55 [0.42–0.71] 0.39 [0.28–0.53] 0.47 [0.42–0.53] Tenoxicam 0.34 [0.11–0.80] 0.62 [0.26–1.22] 0.33 [0.10–0.93] 0.26 [0.06–0.85] 0.53 [0.17–1.25] 0.42 [0.27–0.61] aDrug became available in 2004.

Table III Cumulative reporting rate [95% CI] of serious adverse drug reactions (ADRs) per million deﬁned daily doses (DDDs) according to selected system organ class (SOCs) for each of eight nonsteroidal anti-inﬂammatory drugs (NSAIDs) as reported in the French pharmacovigilance database from 2002 to 2006.

Gastrointestinal Liver Skin Renal Cardiovascular

Aceclofenac 0.04 [0.009–0.12] 0.08 [0.03–0.18] 0.05 [0.01–0.14] 0.01 [0.000–0.14] 0.000 [0.000–0.05] Diclofenac 0.10 [0.07–0.13] 0.09 [0.06–0.11] 0.10 [0.07–0.13] 0.08 [0.05–0.10] 0.003 [0.000–0.012] Ketoprofen 0.17 [0.14–0.20] 0.09 [0.06–0.11] 0.13 [0.10–0.16] 0.11 [0.07–0.08] 0.002 [0.000–0.009] Naproxen 0.08 [0.05–0.12] 0.04 [0.02–0.07] 0.08 [0.05–0.11] 0.03 [0.01–0.06] 0.003 [0.000–0.017] Nimesulide 0.05 [0.02–0.09] 0.15 [0.10–0.23] 0.12 [0.07–0.19] 0.05 [0.02–0.09] 0.000 [0.000–0.023] Meloxicam 0.03 [0.004–0.13] 0.03 [0.004–0.13] 0.16 [0.07–0.31] 0.03 [0.004–0.13] 0.000 [0.000–0.067] Piroxicam 0.10 [0.07–0.13] 0.06 [0.04–0.08] 0.14 [0.11–0.17] 0.03 [0.02–0.05] 0.000 [0.000–0.007] Tenoxicam 0.05 [0.01–0.14] 0.03 [0.004–0.12] 0.13 [0.05–0.26] 0.06 [0.01–0.17] 0.000 [0.000–0.062]

Liver ADRs ketoprofen and naproxen over the entire period covered For liver ADRs, nimesulide had the highest cumulative by the study. No case associated with aceclofenac, reporting rate (0.16), followed by diclofenac (0.09) and nimesulide, meloxicam, piroxicam and tenoxicam was ketoprofen (0.09); tenoxicam exhibited the lowest rate reported during the study period. (0.03). Case/noncase analysis Skin ADRs Of the original 42 389 serious ADR reports examined, Meloxicam had the highest reporting of skin ADRs over 476 cases (1.12%) were excluded because of missing the entire period (0.16), followed by piroxicam (0.14) data concerning age and/or gender (required to calculate and ketoprofen (0.14); aceclofenac had the lowest adjusted odds ratios); 3407 reports (8.04%) were reporting rate (0.06). discarded because patients were under 15 years of age. Reports included in the case/noncase analysis totalled Renal ADRs 38 506 (90.84%) (Figure 1). Ketoprofen was associated with the highest reporting Cases were then stratiﬁed by each single SOC accord- rate of renal ADRs (0.12), followed by diclofenac (0.08) ing to gender and age (15–45, 46–65, 66–80 and and tenoxicam (0.07); aceclofenac had the lowest rate >80 years). Results of the analysis for each NSAID (0.01). according to the different SOC are presented in Table IV.

Cardiovascular ADRs Gastrointestinal ADRs For arterial thrombosis (cardiac and peripheral), a very The case/noncase comparison showed that ketoprofen small number of cases were reported with diclofenac, and piroxicam were associated with the highest risk of

Total number of ADRs 72 834 (with at least 1 AR from the French DB)

Total number of ADRs between 2002–2006 42 389 (with at least 1 AR from the French DB)

Age < 15 years 3407

Missing data 476

Total number of eligible ADRs 38 506 (100%)

LIVER ADRs GI ADRs SKIN ADRs RENAL ADRs CV ADRs 3542 1805 6052 2231 337 (9.89%) (5.04%) (16.90%) (6.20%) (0.94%)

Figure 1 Population of the case/noncase analysis. ADRs, adverse drug reactions; CV, cardiovascular; DB, database; GI, gastrointestinal.

Table IV A case/noncase analysis showing age- and gender-adjusted odds ratios (OR) for association of nonsteroidal anti-inﬂammatory drugs (NSAIDs) with serious gastrointestinal, liver, skin, renal and cardiovascular adverse reactions (ADRs) as reported in the French pharmacovigilance database from 2002 to 2006. For each ADR, the comparator for the exposure to each NSAID is the whole database (among 38 506 cases with age and gender). Values in boldface indicate a level of signiﬁcance <0.05.

Gastrointestinal Liver Skin Renal Cardiovascular

NSAID Total n OR [95% CI] n OR [95% CI] n OR [95% CI] n OR [95% CI] n OR [95% CI]

Aceclofenac 22 3 3.74 [1.10–12.7]63.67 [1.43–941] 4 1.09 [0.37–3.24] 1 0.84 [0.11–6.23] 0 – Diclofenac 338 59 4.55 [3.40–6.08]531.93 [1.44–2.60] 60 1.17 [0.88–1.55] 48 2.86 [2.10–3.90] 2 0.69 [0.17–2.80] Ketoprofen 506 111 6.87 [5.51–8.57] 58 1.27 [0.96–1.67] 87 1.03 [0.81–1.30] 74 2.85 [2.20–3.70] 1 0.21 [0.03–1.50] Meloxicam 23 2 2.26 [0.53–9.66] 2 0.89 [0.21–3.80] 9 3.15 [1.36–7.31] 2 1.65 [0.39–7.09] 0 – Naproxen 162 28 5.07 [3.33–7.73] 14 0.82 [0.46–1.49] 26 1.01 [0.66–1.55] 11 1.34 [0.72–2.48] 1 0.70 [0.10–5.05] Nimesulide 83 8 2.33 [1.12–4.86]254.53 [2.83–7.27]201.69 [1.02–2.81] 8 1.81 [0.87–3.77] 0 – Piroxicam 257 57 6.54 [4.84–8.84]351.54 [1.07–2.20]772.22 [1.69–2.91] 21 1.56 [0.99–2.44] 0 – Tenoxicam 25 3 2.94 [0.87–9.92] 2 0.86 [0.20–3.68] 8 2.22 [0.92–5.38] 4 3.17 [1.08–9.33]0 – gastrointestinal ADRs, followed by naproxen, diclofe- Skin ADRs nac, aceclofenac and nimesulide. Meloxicam and ten- The level of risk for skin ADRs was lower than those oxicam were not signiﬁcantly associated with observed for liver or gastrointestinal disorders. Meloxi- gastrointestinal ADRs, and few cases were reported cam, piroxicam and nimesulide were signiﬁcantly asso- with these drugs. ciated with skin ADRs.

Liver ADRs Renal ADRs Nimesulide, aceclofenac, diclofenac and piroxicam were Tenoxicam, diclofenac and ketoprofen were signiﬁcantly signiﬁcantly associated with liver ADRs. associated with renal ADRs.

Cardiovascular ADRs measure the contribution of these types of factors, it The majority of NSAIDs (aceclofenac, meloxicam, nime- can be surmised that reporting rates of some ADRs may sulide, piroxicam and tenoxicam) did not have reports of be overestimated or erroneously attributed to NSAIDs. cardiovascular ADRs during the period of the study. Two Nevertheless, despite the well-known limits of spontane- cases of thrombosis were reported for diclofenac, and one ous reporting systems, the results of this analysis are in case each for ketoprofen and naproxen. agreement in many ways with the scientific literature. Of the serious ADRs attributed to the oral NSAIDs examined in this study, the most frequently reported DISCUSSION involved the skin, followed by the gastrointestinal This analysis of the French Pharmacovigilance database system, liver and the renal system, with far fewer reports evaluated and compared the overall safety profile of eight involving the cardiovascular system. It is well known widely used NSAIDs, in terms of the overall reporting that cutaneous adverse reactions are frequently observed rate of serious ADRs, as well as the reporting rate of with NSAIDs, with cross-reactive reactions between co- specific classes of serious ADRs. As far as we know, this is administered drugs. Oxicams and ketoprofen seem to the first study providing data on the safety profile of eight present the highest cutaneous risk, as observed in studies different NSAIDs under real conditions of use in France. investigating rare but serious ADRs such as Stevens– The main strength of the present study was the use of Johnson syndrome or toxic epidermolysis [20]; nimesu- a national database, which collected standardized infor- lide and diclofenac present an intermediate risk, and mation on adverse reactions related to drug use in a aceclofenac and naproxen the lowest risk. ‘real-world’ population. However, the study presents The second highest reporting of ADRs concerns several limitations resulting from the spontaneous gastrointestinal reactions. The rank of the studied reporting of ADRs, which include the inability to NSAIDs investigated, in term of serious gastrointestinal accurately identify drug-related events (particularly ADRs, reveals that ketoprofen and piroxicam exhibited cardiovascular ones), bias owing to media coverage, the highest levels of risk, whereas nimesulide exhibited the time elapsed since the marketing of a drug and the the lowest [2,18]. In a case–control study, Laporte et al. fact that some ADRs are better known than others compared the gastrointestinal bleeding risk associated (resulting in under-reporting). with first- and second-generation NSAIDs, and the rank The overall reporting rates of serious ADRs were of risk was very similar to that obtained in our study similar for the selected NSAIDs (0.31–0.58 cases per [21]. million DDDs) in this French context, with the exception The reporting of hepatic ADRs for a number of NSAIDs of ketoprofen (0.78), which was associated with a higher assessed in this study confirms that the liver is one of the reporting rate of serious ADRs. Annual reporting rates in potential organ targets of ADRs when NSAIDs are used France were similar for the years 2002–2006, although [5–7,22], even though the reporting of these ADRs is the rates of ADRs were high in 2004, which may be lower when compared with the reporting of cutaneous or explained by increasing reporting because of the gastrointestinal ADRs. Nimesulide presented the highest increased media coverage of the COX-2 inhibitors and reporting of hepatic ADRs in terms of cases per million conventional NSAID safety following the withdrawal of DDDs (0.157). A comparative analysis of the French and rofecoxib [13–15]. These reporting rates must be inter- Spanish pharmacovigilance databases conducted in the preted with caution, because not all reactions have a period 1982–2001 had shown a nonsignificant OR for constant risk over time, and if the reaction occurs early, liver injuries and nimesulide in France, whereas it was there may be large difference between long-term and statistically significant in Spain [19]. Actually, nimesu- short-term users. A recent study in France using data lide was withdrawn from the Spanish and the Finnish from the Health Insurance system indicated that the markets in 2002 because of an unfavourable benefit–risk median duration of NSAIDs’ prescription should be balance related to liver injuries [23–25]. In the present estimated 30 days (Fournier JP, Lapeyre-Mestre M, study, it was found that nimesulide and aceclofenac were Sommet A, Pathak A, Oustric S, Montastruc JL unpub- associated with the highest risk of liver ADRs, followed lished data). The number of adverse event reports for a by diclofenac and piroxicam, which is in agreement with drug may be influenced by advertising and may depend the literature [6,7,26,27]. on expectations and the degree of awareness of health Study results revealed that serious renal ADRs are professionals. Although there is no reliable way to significantly associated with the use of tenoxicam,

ª 2011 The Authors Fundamental and Clinical Pharmacology ª 2011 Socie´ te´ Française de Pharmacologie et de The´ rapeutique Fundamental & Clinical Pharmacology 27 (2013) 223–230 Adverse drug reactions related to oral NSAIDs’ use 229 ketoprofen and diclofenac. Despite these results needing 2 Hernandez-Diaz S., Garcia Rodriguez A. Association between to be interpreted with caution because of the low number nonsteroidal anti-inflammatory drugs and upper gastrointesti- of reported cases, they may suggest that the expected nal tract bleeding/perforation. Arch. Intern. Med. (2000) 160 2093–2099. pharmacological impact of NSAID on renal function is 3 Helin-Salmivaara A., Saarelainen S., Gronroos J.M., Vesalainen not always taken into account when treating in partic- R., Klaukka T., Huupponen R. Risk of upper gastrointestinal ular old patients. Indeed, NSAIDs may affect arterial events with the use of various NSAIDs: a case-control study in blood pressure via the renin–angiotensin system and a general population. Scand. J. Gastroenterol. (2007) 42 alterations in sodium levels and water retention in the 923–932. kidneys [9]. 4 Mandell B.F. General tolerability and use of nonsteroidal The last type of serious ADR investigated, thrombotic anti-inflammatory drugs. Am. J. Med. (1999) 107 72S–76S; cardiovascular events, was very rarely reported to the discussion 76S–77S. 5 Garcia Rodriguez L.A., Williams R., Derby L.E., Dean A.D., French pharmacovigilance system, as in other systems Jick H. Acute liver injury associated with nonsteroidal anti- [26]. inflammatory drugs and the role of risk factors. Arch. Intern. Med. (1994) 154 311–316. 6 Garcia Rodriguez L.A., Ruigomez A., Jick H. A review of CONCLUSION epidemiologic research on drug-induced acute liver injury using This study shows that the most frequent serious ADRs the general practice research data base in the United Kingdom. reported for oral NSAIDs were cutaneous, followed by Pharmacotherapy (1997) 17 721–728. 7 Garcia Rodriguez L.A., Perez Gutthann S., Walker A.M., Lueck gastrointestinal, hepatic and renal events and rarely, L. The role of non-steroidal anti-inflammatory drugs in acute cardiovascular ones. Overall, the highest risk for serious liver injury. [erratum appears in BMJ 1992 Oct ADRs was linked to ketoprofen, the lowest to aceclo- 17;305(6859):920]. BMJ (1992) 305 865–868. fenac, with the other drugs presenting intermediate risks. 8 Hussaini S.H., Farrington E.A. Idiosyncratic drug-induced The risks differed for the specific drug and system liver injury: an overview. Expert Opin. Drug Saf. (2007) 6 involved; the greatest gastrointestinal, hepatic, cutane- 673–684. ous and renal risks were observed, respectively, with 9 European Medicine Agency EMA Press Release EMEA/247323/ 2005 ‘‘General reccomendation about NSAIDs.’’ http:// ketoprofen, nimesulide, meloxicam and tenoxicam. This www.ema.europa.eu/docs/en_GB/document_library/Press_ emphasizes the importance of evaluating the risks of release/2009/11/WC500014477.pdf (accessed 16 February specific NSAIDs in relation to the patient’s history, in 2011). order to accurately determine their benefit–risk profiles 10 National Institute for Clinical Excellence. Guidance on the use of and improve patient safety. cyclo-oxygenase (Cox) II selective inhibitors, celecoxib, rofecoxib, meloxicam and etodolac for osteoarthritis and rheumatoid arthritis. National Institute of Clinical Excellence, London, ACKNOWLEDGEMENTS 2001; Technology Appraisal Guidance No. 27. 11 Deviere J. Do selective cyclo-oxygenase inhibitors eliminate the Editorial assistance for the preparation of this manuscript adverse events associated with nonsteroidal anti-inflammatory was provided by inScience Communications, a Wolters drug therapy? Eur. J. Gastroenterol. Hepatol. (2002) 14(Suppl. 1) Kluwer business. Funding for this assistance was pro- S29–S33. vided by Helsinn Healthcare SA. 12 Bennett A., Tavares I.A. COX-2 inhibitors compared and contrasted. Expert Opin. Pharmacother. (2001) 2 1859–1876. 13 Clark D.W.J., Layton D., Shakir S.A.W. Do some inhibitors of FUNDING DISCLOSURE COX-2 increase the risk of thromboembolic events? Linking pharmacology with pharmacoepidemiology. Drug Saf. (2004) This study was supported by Helsinn Healthcare SA and 27 427–456. by Therabel Lucien Pharma. 14 Sommet A., Grolleau S., Bagheri H., Lapeyre-Mestre M., Montastruc J.L., French Network of Regional Pharmacovigi- lance Centres. Was the thrombotic risk of rofecoxib predictable REFERENCES from the French Pharmacovigilance Database before 30 1 Garcia Rodriguez L.A., Cattaruzzi C., Troncon M.G., Agostinis L. September 2004? Eur. J. Clin. Pharmacol. (2008) 64 829–834. Risk of hospitalization for upper gastrointestinal tract bleeding 15 Solomon D.H., Glynn R.J., Rothman K.J. et al. Subgroup associated with ketorolac, other nonsteroidal anti-inflammatory analyses to determine cardiovascular risk associated with drugs, calcium antagonists, and other antihypertensive drugs. nonsteroidal antiinflammatory drugs and coxibs in specific Arch. Intern. Med. (1998) 158 33–39. patient groups. Arthritis Rheum. (2008) 15 59.

16 Code de la Sante´ Publique, Article R5121-153, version en nimesulide and other non-steroidal anti-inflammatory drugs. vigueur au 8 aout 2004. http://www.legifrance.gouv.fr/ [see comment] BMJ (2003) 327 18–22. (accessed 16 February 2011). 23 Macia M.A., Carvajal A., del Pozo J.G., Vera E., del Pino A. 17 van Puijenbroek E.P., Bate A., Leufkens H.G., Lindquist M., Hepatotoxicity associated with nimesulide: data from the Orre R., Egberts A.C. A comparison of measures of dispropor- Spanish Pharmacovigilance System. Clin. Pharmacol. Ther. tionality for signal detection in spontaneous reporting systems (2002) 72 596–597. for adverse drug reactions. Pharmacoepidemiol. Drug Saf. 24 European Medicine Agency. 2004. EMEA statement following (2002) 11 3–10. withdrawal of nimesulide from the Spanish market. EMEA/ 18 Lugardon S., Lapeyre-Mestre M., Montastruc J.L. Upper CPMP/1724/04/2004. Available from: http://www.ema. gastrointestinal adverse drug reactions and cyclo-oxygenase-2 europa.eu/docs/en_GB/document_library/Referrals_document/ inhibitors (celecoxib and rofecoxib): a case/non-case study from Nimesulide_31/WC500013818.pdf (accessed 16 February the French Pharmacovigilance Database. Eur. J. Clin. Pharma- 2011). col. (2004) 60 673–677. 25 National Agency for Medicine (Finland). NAM statement 19 Lapeyre-Mestre M., de Castro A.M.R., Bareille M.-P. et al. following withdrawal of Nimed (nimesulide) from the Finnish Non-steroidal anti-inflammatory drug-related hepatic damage in market. Available from: http://www.fimea.fi/whats_new/news/ France and Spain: analysis from national spontaneous reporting 1/0/the_sale_of_nimed_an_anti-inflammatory_analgesic_ systems. Fundam. Clin. Pharmacol. (2006) 20 391–395. is_temporarily_suspended_due_to_its_adverse_liver_effects 20 Mockenhaupt M., Kelly J.P., Kaufman D., Stern R.S., SCAR (accessed 16 February 2011). Study Group. The risk of Stevens-Johnson syndrome and toxic 26 Raber A., Heras J., Costa J., Fortea J., Cobos A. Incidence of epidermal necrolysis associated with nonsteroidal antiinflam- spontaneous notifications of adverse reactions with aceclofenac, matory drugs: a multinational perspective. J. Rheumatol. meloxicam, and rofecoxib during the first year after marketing (2003) 30 2234–2240. in the United Kingdom. Ther. Clin. Risk Manag. (2007) 3 225– 21 Laporte J.R., Ibań˜ez L., Vidal X., Vendrell L., Leone R. Upper 230. gastrointestinal bleeding associated with the use of NSAIDs: 27 Sanchez-Matienzo D., Arana A., Castellsague J., Perez-Gutthann newer versus older agents. Drug Saf. (2004) 27 411–420. S. Hepatic disorders in patients treated with COX-2 selective 22 Traversa G., Bianchi C., Da Cas R., Abraha I., Menniti-Ippolito inhibitors or nonselective NSAIDs: a case/noncase analysis of F., Venegoni M. Cohort study of hepatotoxicity associated with spontaneous reports. Clin. Ther. (2006) 28 1123–1132.

ª 2011 The Authors Fundamental and Clinical Pharmacology ª 2011 Socie´ te´ Franc¸aise de Pharmacologie et de The´ rapeutique Fundamental & Clinical Pharmacology 27 (2013) 223–230 Copyright of Fundamental & Clinical Pharmacology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.