EQUINE VETERINARY JOURNALEquine vet. J. (2007)doi: 10.2746/042516407X159123

1

Summary

Reasons for performing study: The selective COX-2-inhibitornimesulide is used extra-label in equine veterinary practiceas an anti-inflammatory agent. However, there are no dataon which to base the rational use of the drug in this species.

Objectives: To determine the effective COX selectivity ofnimesulide in the horse, and suggest a suitable dosingschedule.

Methods: The pharmacokinetics of nimesulide in the horseafter oral administration (1 mg/kg bwt), and oral and i.v.administration (1.5 mg/kg bwt) were investigated, effects offeeding status on bioavailability determined, and plasmaprotein binding of the drug and its principal metabolitesmeasured. Compartmental and noncompartmentalpharmacokinetic analyses were performed. The plasmaconcentration-time profile was used, together with in vitroliterature data on nimesulide inhibition of COX isoforms, todetermine the effective COX selectivity of nimesulide in thehorse, and suggest a suitable dosing schedule.

Results and Conclusions: The findings suggest that 1.5 mg/kgbwt may produce adequate clinical effects and that thedosing interval should be 12–24 h depending on conditionseverity. However, at that dose, the concentration in theanimal exceeds the in vitro IC50 for both isoforms, so thatCOX-1/COX-2 selectivity is lost and side-effects due to COX-1 inhibition are a possibility. Nimesulide should therefore beused with caution in equine clinical practice.

Introduction

Subsequent to the discovery of the 2 main cyclooxygenase (COX)isoforms in mammals (Vane et al. 1998) the view developed thatnonsteroidal anti-inflammatory drugs (NSAIDs) that selectivelyinhibited the inducible COX-2 isoform, but had reduced inhibitoryeffects on the constitutive COX-1 isoform, would have fewergastrointestinal side-effects than substances that strongly inhibited

COX-1 (Vane and Botting 1998; Cannon 1999; Kaplan-Machlisand Klostermeyer 1999). New selective COX-2 inhibitors were,therefore, developed (Talley 1999) and available NSAIDsclassified according to their relative abilities to inhibit COX-1 andCOX-2 in man and animals of veterinary interest. It is now clear,however, that selective COX-2 inhibitors are not necessarily saferor more effective anti-inflammatory agents than conventionalNSAIDs (Feldman and McMahon 2000; Kam and See 2000;Wallace 2002). For example, highly selective COX-2 inhibitorshave reduced gastrointestinal side-effects in man, but are alsoassociated with increased frequency of severe cardiovascularadverse events (Psaty and Furberg 2005; Vonkeman et al. 2006).Adverse vascular effects have also been recently documented withhigh doses of the classic NSAIDs ibuprofen and diclofenac(Kearney et al. 2006).

NSAIDs are used extensively in equine veterinary practice. Ithas been suggested that selective COX-2 inhibitors may bepreferable to nonselective COX inhibitors in the management ofequine colic, as they seem to have fewer adverse effects on dorsaland ventral colon motility and may not hinder the repair ofischemically injured jejunal mucosa (Campbell and Blikslager2000; van Hoogmoed et al. 2002). However, selective COX-2inhibitors can slow the healing of existing ulcers (Moreau et al.2005) and horses, particularly racehorses, are prone toasymptomatic gastrointestinal ulcers (Dionne et al. 2003).

It is, therefore, important to determine the COX selectivity ofindividual NSAIDs (expressed as the ratio of COX-1 and COX-2IC50s in blood after treatment) used in equine clinical practice. In fact Lees et al. (2004a,b) advocated integration ofpharmacodynamic (PD) with pharmacokinetic (PK) data, in orderto obtain information for the rational use of NSAIDS in veterinarypractice. Such studies provide information on (a) species-specificactivity, (b) the dose necessary to produce an adequate anti-inflammatory effect and (c) indications of the risks of unwantedside effects.

Nimesulide is a sulphonanilide derivative shown to have anti-inflammatory, analgesic and antipyretic effects in animal

Oral and intravenous administration of nimesulide in thehorse: rational dosage regimen from pharmacokinetic andpharmacodynamic data

R. VILLA*, P. CAGNARDI, C. BELLOLI†, A. ZONCA, C. ZIZZADORO†, E. FERRO‡ and S. CARLI

Department of Veterinary Sciences and Technologies for Food Safety, University of Milan, Via Celoria, 10, 20133, Milan, Italy †Departmentof Animal Health and Welfare, University of Bari, Italy. ‡Department of Veterinary Clinical Sciences, University of Milan, Italy

Keywords: horse; nimesulide; NSAIDs; COX-2-inhibitors; pharmacokinetics; pharmacodynamics; oral administration; intravenousadministration; bioavailability; protein binding

*Author to whom correspondence should be addressed.[Paper received for publication 03.08.06; Accepted 15.09.06]

2 Oral and i.v. administration of nimesulide in the horse

experimental models (Rabasseda 1996). The drug is marketed inseveral countries as an analgesic/anti-inflammatory for use in manand dogs, but has not been licensed for human use in the US,Canada, UK, Australia, New Zealand or Scandinavian countries.The European Union has issued precautionary advice on the marketing of this drug in man following serious complicationsafter use, and has withdrawn marketing authorisation for 200 mg tablets. (www.emea.eu.int/pdfs/human/referral/nimesulide/172404en.pdf). The drug exerts its effects through avariety of mechanisms (Bennett 1999), but the most importantseems to be selective inhibition of COX-2, as demonstrated in manand various animal species (Famaey 1997; Fabiola et al. 1998;Ricketts et al. 1998). In equine whole blood, nimesulidepreferentially inhibits COX-2 (Belloli et al. 2003) as also reportedin dogs (Toutain et al. 2001a) but is a substantially more potentinhibitor of both COX-1 and COX-2 in horses than dogs. A rationaldosage regimen for nimesulide in the dog (Toutain et al. 2001a,b)has been established by PK/PD studies, but the PK/PD profile ofnimesulide in the horse has not been determined, and there are nodata on which to base rational use. It is important to obtain such dataas nimesulide is used extra-label in equine clinical practice:European legislation (Commission Decision 93/623/CEE amendedby Commission Decision 2000/68/CE) permits its use in animalsnot destined for human consumption, provided such animals areidentified. Its use in racehorses is confirmed by the frequentdetection of nimesulide and its reduced nitro metabolite (4-amino-2-phenoxy-methanesulfonanilide, 4-AMPS) in blood and urinesamples taken at the race track (Sarkar et al. 1997). This reducednitro metabolite is weakly active, about 30-fold less so thannimesulide in the carrageenin oedema test in rats (Bernareggi 1998).

The aims of the present study were to determine thepharmacokinetics of nimesulide in horses and investigate theeffect of feeding status on drug bioavailability, since the drug isnormally administered orally. We also integrated previouslypublished in vitro data on nimesulide inhibition of COX-1 andCOX-2 isoforms with the plasma concentration-time profile, todetermine the effective COX selectivity of nimesulide in thehorse; and suggest an appropriate dosing schedule.

Materials and methods

The study was carried out in compliance with National Legislation(DL n. 116, 27th January 1992) deriving from DirectiveN.86/609/CEE 24th November 1986 on the protection of Animalsused for Experimental and other Scientific purposes, and all theanimal procedures were approved by the local Ethics Committee.

Animals

Fifteen healthy horses age 3–6 years were divided into 3 Groups.Group A, 4 Thoroughbreds, 2 males and 2 females, weight 385–466 kg; Group B, 4 Thoroughbreds, 2 males and 2 females,weight 410–550 kg, and Group C, 7 Thoroughbreds, 4 males and3 females, weight 390–520 kg. The animals were housed inindividual stalls throughout the study and received no medicationin the 30 days leading up to drug administration. Animals ofGroups A and B were fasted for 12 h before drug administration,while Group C had free access to food (pasture and hay). Fastedanimals were fed a small amount of hay about 4 h after drugadministration, as the only food, since the animals were not beingexercised. Water was supplied ad libitum.

Drugs

For i.v. administration, nimesulide1 was dissolved in aqueoussodium hydroxide (0.1 mol/l) and the pH adjusted to 9.5 (± 0.2)with hydrochloric acid (0.1 mol/l). The solution was prepared onthe day of administration, and was filtered through a disposable0.22 µm filter (Acrodisc PF)2 before injection. The finalnimesulide concentration, assayed by validated HPLC method,was 19.8 mg/ml. For oral administration a commercial granularpreparation (Aulin)3, was used.

Experimental design

Horses of Group A were administered nimesulide according to ani.v./oral cross-over design with a 2-week interval betweenadministrations. Horses of Groups B (fasted) and C (not fasted)were treated once orally.

Drug administration

Since nimesulide is not authorised for use in horses, doses werebased on those used in man. The typical oral dose in man is 1.4–2.8 mg/kg bwt; and the conservative oral dose of 1 mg/kg bwtwas used initially. Subsequently, oral and i.v. doses of 1.5 mg/kgbwt were used following lack of observed adverse effects at 1 mg/kg bwt. Group A received nimesulide orally and i.v. at 1.5 mg/kg bwt; Groups B and C received nimesulide orally, onceonly at 1 mg/kg bwt.

For oral administration the granular preparation was mixedwith a small quantity of oats and administered under veterinarysurveillance. The preparation was always totally consumed in afew minutes. For i.v. administration a venous catheter4 was usedand the solution injected slowly as single bolus.

Sampling

Blood samples were collected into nonheparinised tubes beforeadministration (time 0) and at various subsequent times between0.08 h and 24 h, depending on administration route. Samplingtimes (h) were:

• Group A i.v.: 0.08, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 5, 6, 8, 10, 12, 24

• Group A and B orally: 0.33, 0.67, 1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 24

• Group C orally: 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 24.

Serum samples were prepared by centrifugation (1500 g, 10 minat room temperature) and 1 ml aliquots were stored at –80°Cpending assay. Urine samples were collected manually fromGroup B horses up to 24 h after treatment when they urinatedspontaneously.

Assays

Nimesulide and its metabolites 4-hydroxynimesulide (4-HXN) and4-amino-2-phenoxymethanesulfonanilide (4-APMS) for analysiswere supplied by Helsinn Chemicals, Lugano, Switzerland. Serumand urine concentrations, 4-HXN and 4-APMS were determinedby HPLC using external standard calibration. Nimesulide andmetabolites were extracted into chloroform: 0.2 ml of serum or

R. Villa et al. 3

urine was added to 0.5 ml borate buffer (0.56 mol/l; pH 9) andshaken with 8 ml chloroform. The HPLC method used a reverse-phase column and isocratic conditions (Inertsil5 150-5 ODS-2, 5 µm, 250 x 4.6 mm). The mobile phase was: (A) 25 mmol/l ofpotassium dihydrogen phosphate and triethylamine 1 mmol/l,adjusted to pH 7.5 (± 0.1) with potassium hydroxide 10 mol/l; (B)acetonitrile plus buffer A, ratio 58:42 v:v. All runs were performedat a flow rate of 1 ml/min; run times were 11 min. A UV LC90detector6 was used, set to read at 230 nm.

Protein binding

Serum protein binding was determined in vitro for nimesulide, 4-HXN and 4-APMS in the concentration range 0.5–10 µg/ml.The free compounds were separated by ultrafiltration (Villa et al.1994, 1997) using a disposable device (Amicon)7 and analysed byHPLC as described above.

Pharmacokinetics

Pharmacokinetic parameters were deduced from serumconcentration-time data using the WinNonLin 4.0.1 software8

which allows compartmental and noncompartmental analyses ofthe experimental data. Minimum information criterion estimates(MAICE; Yamaoka et al. 1978) were used to choose the modelbest fitting the data (Groups A and B). All data points wereweighted by the inverse square of the fitted value.

Serum concentrations of nimesulide after i.v. administrationwere fitted to the following bi-exponential equation:

C(t)= Y1 exp(-λ1t) + Y2 exp (-λ2t)

where Ct (µg/ml) is serum drug concentration at time t; Y1 and Y2are serum concentrations extrapolated to time zero of thenimesulide distribution and elimination phases, respectively; λ1and λ2 are the slopes of the distribution and elimination phases ofthe drug, respectively. The distribution half-time (t1/2λ1) andterminal half-time (t1/2λ2) were calculated as loge2/λn; serum concentration at time 0 (C0) was calculated as the sum ofthe intercepts.

The volume of distribution Vc in the central compartment wascalculated as:

Vc = Dose/C0

The area under the serum concentration-time curve (AUC

(0-∞)) and area under the first moment curve (AUMC) werecalculated by the method arithmetic trapezes with extrapolation toinfinity as follows:

AUC(tlast-∞) = Clast/λ2AUMC (tlast-∞) = tClast/λ2 + Clast/λ2

2

where tlast is the last time with measurable concentration (Clast)and λ2 is the rate constant for the elimination phase. Meanresidence time (MRT), mean absorption time (MAT), bodyclearance (ClB) and volume of distribution at steady state (Vdss)were determined from the following equations (Gibaldi andPerrier 1982):

MRT = AUMC/AUC

MAT = MRToral-MRTi.v.

ClB = Dose/AUCVdss = Cl x MRT

After oral administration data were fitted to the equation:

C (t) = FD x λ3 x (e-λ4t - eλ3t)V(λ3 − λ4)

where D is the dose, F the bioavailability, V the volume ofdistribution, and λ3 and λ4 the first order rate constants for theabsorption and elimination phases, respectively. Half-timeswere calculated as loge2/λn.

The bioavailability (F%) of nimesulide after oraladministration for both doses was calculated as the ratio of thearea under serum concentration-time curves (AUC(0-∞)) afteroral and i.v. administration:

F% = (AUCoral/AUCi.v.) x 100

Statistical analysis

Pharmacokinetic parameters are reported as means with s.d.,harmonic means were calculated for terminal half-lives, andpseudostandard deviations (SE) were calculated using a jack-knifetechnique (Lam et al. 1985). InStat 3.09 was used to perform theanalyses. Cmax, Tmax, AUC and MRT values for nimesulide afteroral administration of 1.5 mg/kg bwt in fasted and 1 mg/kg bwt infasted and fed conditions were compared by one-way ANOVAwith Bonferroni test for multiple comparisons. Differences withP<0.05 were considered significant. Terminal half-lives after oraladministration of

100.0

10.0

1.0

0.1

Con

c. (

ug/m

l)

0 5 10 15 20 25

Time (h)

10.0

1.0

0.1

Con

c. (

ug/m

l)

0 5 10 15 20 25

Time (h)

Fig 1: Semi-logarithmic plot of individual nimesulide concentrations inhorses of Group A after i.v. (a) and oral administration (b) at 1.5 mg/kg bwt.

a)

b)

4 Oral and i.v. administration of nimesulide in the horse

1 and 1.5 mg/kg bwt, and terminal half-lives, AUC and MRT afteroral and i.v. administration of 1.5 mg/kg bwt were compared byunpaired t test with Welch correction (variances unequal).Differences with P<0.05 were considered significant.

Results

Serum and urine concentrations

HPLC retention times for 4-HXN, nimesulide and 4-APMS wereabout 4.3, 7.0 and 8.1 min, respectively. The analytical method wassubjected to intralaboratory validation and found to be specific,linear (in the range 0.05–15 µg/ml), precise and accurate, with alimit of quantification of 0.05 µg/ml for all the compoundsinvestigated in serum and urines. Nimesulide was detected inserum at all sampling times, the 4-HXN metabolite was detectedonly as non-quantifiable traces and 4-APMS was never detected.Pharmacokinetic analysis was performed for nimesulide only.

Serum concentrations of nimesulide after i.v. and oraladministration (1.5 mg/kg bwt) to individual animals in Group A areshown in Figure 1. Serum concentrations after oral administration (1mg/kg bwt) to individual animals of Group B are shown in Figure 2.

After i.v. administration (1.5 mg/kg bwt, Group A) meannimesulide concentration (± s.d.) in serum was 13.35 ± 1.05 µg/mlat first sampling (0.08 h), decreased to 9.93 ± 0.68 µg/ml at 2 h post treatment, and subsequently declined more slowly to reach0.83 ± 0.21 µg/ml at last sampling (24 h).

Following oral administration at 1.5 mg/kg bwt to Group A(fasted), peak concentrations (5.34 ± 1.11 µg/ml) were obtainedbetween 3 and 4 h, and declined slowly to 0.81 ± 0.21 µg/ml at 24 h.

After oral administration at 1 mg/kg bwt to Group B (fasted),peak serum concentration (2.86 ± 0.94 µg/ml) was reached between4 and 5 h and then declined slowly to 0.38 ± 0.05 µg/ml at 24 h.

Oral administration of 1 mg/kg bwt nimesulide to Group C (fedhorses) resulted in marked inter- and intraindividual variability (Fig 3). Peak serum concentration (1.19 ± 0.20 µg/ml) was reachedbetween 7 and 12 h later. At 24 h post dosing only 4/7 horses hadserum concentrations above the limit of quantification (mean 0.24 ±

TABLE 1: Total amounts (mg) of nimesulide (NIM) and 4-amino-2-phenoxymethanesulfonanilide (4-APMS) recorded in urine collected afterspontaneous urination from horses administered 1 mg/kg bwt of nimesulide orally (Group B)

Horse 1 Horse 2 Horse 3 Horse 4 Mean ± s.d.Time NIM 4-APMS NIM 4-APMS NIM 4-APMS NIM 4-APMS NIM 4-APMS

0–2h 0.23 0.22 0.29 0.26 0.26 ± 0.04 0.24 ± 0.032–4h 0.23 0.65 0.77 1.63 0.87 1.78 0.62 ± 0.34 1.35 ± 0.614–6h 0.45 1.10 0.65 2.48 0.68 1.35 0.59 ± 0.13 1.64 ± 0.746–8h 0.30 0.97 0.30 0.978–12h 0.50 1.43 0.59 5.67 1.11 4.38 0.73 ± 0.33 3.83 ± 2.1712–24h 3.04 12.44 0.20 0.28 0.23 0.99 0.40 0.33 0.97 ± 1.38 3.51 ± 5.96

Totals 3.04 12.44 1.91 4.65 2.53 11.03 3.06 7.84 2.50 ± 0.58 7.84 ± 3.19

TABLE 2: Pharmacokinetic parameters (mean ± s.d.) after i.v. and oraladministration in the horse (compartmental analysis) Groups A and B

Group A Group BParameter (n = 4, fasted) (n = 4, fasted)

1.5 mg/kg bwt i.v. 1.5 mg/kg bwt oral 1 mg/kg bwt oral

AUC(0-∞) (h.µg/ml) 110.09 ± 10.16a 59.76 ± 11.29a,b 31.75 ± 7.21b

t1/2λ1 (h) 0.73 ± 1.07§

t1/2λ2 (h) 6.32 ± 0.91§

t1/2λ3 (h) 0.68 ± 0.22c§ 1.20 ± 0.28c§

t1/2λ4 (h) 7.63 ± 1.94§ 6.80 ± 1.46§

C0 (µg/ml) 13.42 ± 0.82Vc (ml/kg bwt) 112.11 ± 6.69ClB (ml/h/kg bwt) 13.71 ± 1.20AUC after tlast (%) 6.54 ± 1.95d 18.90 ± 5.88d 16.73 ± 7.46MRT(0-∞) (h) 8.96 ± 0.77e 15.19 ± 2.75e 13.61 ± 3.18Vdss (ml/kg bwt) 122.12 ± 1.30MAT (h) 6.57 ± 2.88F % 54.25 ± 8.82Cmax (µg/ml) 5.34 ± 1.11 2.86 ± 0.94Tmax (h) 3.50 ± 0.58f 4.25 ± 0.50f

a, b, c, d, e, f P<0.05; § harmonic mean ± pseudo SE; AUC(0-∞) = area underserum concentration-time curve; t1/2λ1 = distribution half-time; t1/2λ2 = terminalhalf-time; t1/2λ3 = absorption half-time; t1/2λ4 = elimination half-time; C0 = serum concentration at time 0;Vc = volume of distribution in centralcompartment; ClB = serum clearance; AUC after tlast = percentage of AUCextrapolated to infinity after last sampling time; MRT(0-∞) = mean residencetime; Vdss = volume of distribution at steady state; MAT = Mean absorptiontime; F = bioavailability; Tmax = observed time for Cmax.

10.0

1.0

0.1

Con

c. (

ug/m

l)

0 5 10 15 20 25

Time (h)

10.0

1.0

0.1

Con

c (u

g/m

l)

0 5 10 15 20 25

Time (h)

Fig 2: Semi-logarithmic plot of individual nimesulide concentrations inhorses of Group B after oral administration at 1 mg/kg bwt.

Fig 3 : Semi-logarithmic plot of individual nimesulide concentrations inhorses of Group C after oral administration at 1 mg/kg bwt in fed conditions.

R. Villa et al. 5

0.14 µg/ml).Nimesulide and the 4-APMS metabolite were always detected in

urine samples from Group B animals, given 1 mg/kg bwt orally,while the 4-HXN metabolite was detected only in trace amounts.From 2 h post administration, 4-APMS was recovered at higherconcentrations than nimesulide in all samples (Table 1). Urine wascollected after spontaneous urination only; samplings were notuniform across individuals. It was not possible to collect the totalurine output.

Protein binding

Nimesulide (94.54 ± 1.25%) and 4-HXN (90.40 ± 2.26%) werebound strongly to serum proteins, while 4-APMS binding wasconsiderably lower (45.00 ± 4.42%).

Pharmacokinetic analyses

The time course of nimesulide serum concentrations after i.v.administration was best described by a 2-compartment open modelin all animals. A one-compartment open model best fitted the dataafter oral administration in Groups A and B. Data from Group Cwere best described by noncompartmental analysis. To compareresults in the fasting and nonfasting states, serum concentrationdata in Group B and orally treated Group A horses were alsoanalysed by the noncompartmental model used in Group C.

Table 2 shows pharmacokinetic parameters obtained bycompartmental analyses after i.v. and oral administration at 1.5 mg/kg bwt and after oral administration at 1.0 mg/kg bwt.Table 3 compares pharmacokinetic parameters obtained bynoncompartmental analyses in fasting and nonfasting states,following administration of nimesulide at 1.0 mg/kg bwt.

Discussion

The pharmacokinetics of nimesulide in the horse is similar to thatin the dog (Toutain et al. 2001a). Specifically, after i.v.administration nimesulide distributed slowly and slightly (t1/2λ1 0.73 ± 1.07 h, Vdss 122.12 ± 6.69 ml/kg bwt and Vc 112.11± 1.30 ml/kg bwt) indicating it was probably mainly confined toextracellular fluid. This finding is probably attributable to itsextensive protein binding (94.54 ± 1.25%) – a characteristic ofNSAIDs in general – and partly to nimesulide’s weak acidity (pKa = 6.56) (Singh et al. 1999), which limits its passage into cells.

Nimesulide was eliminated (t1/2λ2 = 6.32 ± 0.91 h; ClB = 13.71

± 1.20 ml/h/kg bwt) at similar or slower rates than some otherNSAIDs in the horse (elimination half lives: phenylbutazone 3–8 h; naproxen 4–5 h; flunixin 1.6 h; meclofenamic acid 1 h)(Lees and Higgins 1985). According to Lees et al. (2004a),between species differences in NSAIDs clearance and terminalhalf-life are due mainly to differences in hepatic clearance, sinceextensive plasma protein binding limits glomerular ultrafiltrationand hence renal clearance. In man, nimesulide is almost entirelyeliminated by hepatic biotransformation, with negligible amountsof the parent drug eliminated in urine. The elimination half-lifeand clearance after oral administration in man (t1/2β‚ 1.96–4.75 h;Cl/f 35.2–90.9 ml/h/kg bwt, Bernareggi 1993) appear faster thanin the horse, suggesting that metabolic transformation is slower asan elimination route in the latter species. Furthermore, hepaticclearance in the horse was close to body clearance (0.23 ml/min/kg bwt) while liver blood flow is around 24 ml/min/kg bwt (Dyke et al. 1998), so hepatic extraction is verylow (E about 0.01), and lower than in man (E = 0.1) (Bernareggi1993), again indicating that hepatic transformation is slower in thehorse than man.

The 4-APMS metabolite was the major nimesulide-relatedcomponent of horse urine, with levels over 3 times higher thannimesulide itself, as expected from the greater polarity and thelower plasma protein binding (45 ± 4.42%) of this metabolitecompared to both nimesulide (94.54 ± 1.25%) and 4-HXN (90.40 ±2.26%). In man, 5 active (anti-inflammatory and analgesic)metabolites with lower potency than the parent drug, have beenidentified. The main ones are 4-HXN, present in plasma and urine,and 2-(4’-hydroxyphenoxy)-4-N-acetylamino-methanesulfonanilide,found in urine and faeces; the others, including 4-APMS are presentmainly in faeces, but also urine (Bernareggi 1998). Therefore,although the investigation of nimesulide biotransformation wasincomplete, reduction (to 4-APMS) appears to be the mainmetabolic conversion in the horse, in contrast to man wherehydroxylation to 4-HXN is the main transformation (Bernareggi1998). Our findings on 4-APMS in horse urine support the use ofthis compound as marker for detecting illicit nimesulide use inracehorses (Sarkar et al. 1997).

The bioavailability of nimesulide observed after the oraladministration of 1.5 mg/kg bwt in fasted horses (54 ± 8 %) wasfairly similar to that in fasted dogs administered 5 mg/kg bwt (47 ± 12%) (Toutain et al. 2001a), while absorption was muchslower in horse than in dog (MAT horse 6.57 ± 2.88 vs. MAT dogclose to 0) (Toutain and Bousquet-Melou 2004). After the 2 different oral doses, the significant (P<0.05) reductions in Cmaxand AUC after 1 mg/kg bwt vs. 1.5 mg/kg bwt were dose-relatedsuggesting bioavailability is unlikely to be affected.

Although data on absolute bioavailability in man have notbeen published, from Bernareggi (1993) one may infer thatnimesulide absorption is almost complete after oral administrationand is not influenced by formulation (tablet, suspension,granules); furthermore absorption is little influenced by foodintake – as is also the case in the dog (Toutain et al. 2001a). Bycontrast, in the horse we found that the nonfasted state resulted invariably delayed and incomplete absorption and consequentlyreduced bioavailability. Comparison of AUCs(0-∞) (fasted: 34.06± 7.02 h.µg/ml; fed: 13.98 ± 4.56 h.µg/ml) and MRTs (fasted:13.61 ± 3.18 h; fed: 12.22 ± 3.28 h) indicated that nimesulidebioavailability in fed horses was about half that of fasted horses.Previous in vitro and in vivo investigations (Gerring et al. 1981;Maitho et al. 1986; Landoni and Lees 1995; Lees et al. 2004a)

TABLE 3: Pharmacokinetic parameters (mean ± s.d.) after oraladministration in fasted and not fasted horses (noncompartmentalanalysis) Groups B and C

Group B Group CParameter (n = 4, fasted) (n = 7, fed)

1 mg/kg bwt oral 1 mg/kg bwt oral

AUC(0-∞) (h.µg/ml) 34.06 ± 7.02a 13.98 ± 4.56a

AUC after tlast (%) 16.73 ± 7.46 18.97 ± 11.14MRT(0-∞) (h) 13.61 ± 3.18 12.22 ± 3.28Cmax (µg/ml) 2.86 ± 0.94b 1.19 ± 0.20b

Tmax (h) 4.25 ± 0.50c 8.86 ± 1.68c

a, b, c P<0.05; AUC(0-∞) = area under serum concentration-time curve; AUCafter tlast = percentage of AUC extrapolated to infinity after last samplingtime; MRT(0-∞) = mean residence time; Tmax = observed time for Cmax.

6 Oral and i.v. administration of nimesulide in the horse

have shown clearly that food influences the absorption of variousNSAIDs (phenylbutazone, ketoprofen, meclofenamic acid, andflunixin); in particular bioavailability, Cmax and Tmax in the horsemay all be influenced by drug binding to hay and digesta.

The activity of nimesulide against horse COX-1 and COX-2has been investigated previously by Belloli et al. (2003) using awhole blood assay that takes account of drug binding to plasmaproteins, and supplies IC50 values more suitable for comparisonwith plasma concentrations derived from in vivo studies (Pairetand van Ryn 1998; Blain et al. 2002). Nimesulide was found to bepreferential for horse COX-2 (IC50 COX-1/IC50 COX-2 = 12.17),in agreement with findings in the dog (IC50 COX-1/IC50 COX-2 =13) (Toutain et al. 2001a). However, concentrations inhibitingboth COX-1 (IC50 6.3 µg/ml) and COX-2 (IC50 0.49 µg/ml) wereabout 7 times higher in the dog than the horse (COX-1 0.86 µg/ml,COX-2 0.07 µg/ml).

High potency against equine cyclooxygenases, compared tothe human and canine enzymes (Cryer and Feldman 1998; Pairetand van Ryn 1998; Brideau et al. 2001; Toutain et al. 2001a), hasbeen described for several other NSAIDs (Belloli et al. 2003) andwould explain the lower doses required to obtain therapeuticeffects (Lees et al. 1991).

Assuming low binding of nimesulide to erythrocytes (Bree et al. 1993), blood IC50s can be converted into plasma IC50s whichare more representative of the in vivo situation. The PCV in the horseis about 0.38, and plasma IC50s for nimesulide would therefore beabout 1.4 µg/ml for COX-1 and 0.1 µg/ml for COX-2.

Using PK/PD approach Toutain et al. (2001b) showed that in vivo the plasma EC50 for nimesulide for antipyretic effect wassignificantly higher than the IC50 calculated in vitro for COX-2inhibition, while the plasma EC50 for anti-inflammatory effectwas close to the IC50 for COX-1 inhibition; the plasma IC50calculated for COX-1 is therefore probably a good estimate of thein vivo anti-COX-1 activity of nimesulide. Taking this estimate,and considering the daily plasma clearance of 336 ml/kg bwt/day,estimated in the present study, the daily doses of nimesuliderequired to maintain steady state concentrations (product of dailyclearance and IC50) at IC50 plasma levels for COX-1 and COX-2in the horse can be estimated at about 0.47 mg/kg bwt and 0.04 mg/kg bwt, respectively. As the oral bioavailability ofnimesulide in the horse is about 50% in fasted animals and about25% in fed horses, these doses have to be increased to about 1.0and 0.1 mg/kg bwt in fasted and 2.0 and 0.2 mg/kg bwt in fedhorses, respectively.

Toutain et al. (2001b) suggested previously that a plasmaconcentration of 1 µg/ml, near to the plasma IC50 for equine COX-1, would produce adequate anti-inflammatory effects, butwould be low enough to minimise side-effects. After oraladministration of 1 mg/kg bwt to fasted horses in the present study,the plasma concentration of nimesulide exceeded 1 µg/ml for about10–12 h (Cmax = 2.86 ± 0.94 µg/ml), while in fed horses the drugwas over this level in only 4 of the 7 animals and between 6 and 10 h after treatment. After giving 1.5 mg/kg bwt orally, meanplasma concentration was the range 2.7–0.8 µg/ml from 40 min–24h after dosing in fasting animals. Based on these findings andconsidering that several mechanisms in addition to COX inhibitionmay contribute to the anti-inflammatory action of nimesulide, wesuggest that a dose lower than 1.5 mg/kg bwt would be sufficient forantipyretic activity and mild inflammatory conditions; and that 1.5 mg/kg bwt would produce satisfactory anti-inflammatoryeffects in both fasted and fed animals.

The elimination half-life of nimesulide in the horse (about 7 h)indicates that steady-state concentration is reached in about 2 days(5 times elimination half-life). When administered at a dose in the range 1–1.5 mg/kg bwt there is no substantial accumulationwhen the drug is given at 12–24 h intervals. The predicted accumulation factor is 1.2 and 1.1 for dosing intervals of 12 and24 h respectively, favourable values for NSAIDs, which arecharacterised by limited tolerance, and indicating that a 12–24 hdosing interval is appropriate, depending on the severity of theinflammatory condition.

To conclude, the data obtained for nimesulide provide usefulinformation for the rational and safe use of the drug in horses, inserious and mild inflammatory conditions. In particular, theysuggest that the dose required to obtain adequate clinical effects isof the order of 1.5 mg/kg bwt and the dosing interval should be12–24 h depending on condition severity. However, at this dose,the concentration in the animal exceeds the in vitro IC50 for bothisoforms, COX-1/COX-2 selectivity is therefore lost and side-effects due to COX-1 inhibition are a possibility. Nimesulideshould, therefore, be used with caution in equine clinical practice.

Manufacturers’ addresses

1Sigma-Aldrich, Milan, Italy.2Gelman Sciences, Ann Arbor, Michigan, USA.3Roche, Milan, Italy. 4Terumo Europe, Leuven, Belgium. 5Macherey-Nagel, Düren, Germany.6Perkin Elmer, Monza, Italy. 7Millipore, Milan, Italy. 8Pharsight Corporation, Mountain View, California, USA.9GraphPad Software, San Diego, California, USA.

Acknowledgements

We thank Helsinn, Lugano, Switzerland, for the support provided,and Don Ward for help with the English. This work was partiallysupported by funds from the Italian Ministry for University andScientific and Technological Research (FIRB – 2000).

References

Belloli, C., Zizzadoro, C., Badino, P., Odore, R., Ormas, P. and Re, G. (2003) Potencyand selectivity of non steroidal anti-inflammatory drugs for equine COX-1 andCOX-2. J. vet. Pharmacol. Therap. 26 Suppl. 1, 175-176.

Bennett, A. (1999) Overview of nimesulide. Rheumatology 38 Suppl. 1, 1-3.

Bernareggi, A. (1993) The pharmacokinetic profile of nimesulide in healthyvolunteers. Drugs 46 Suppl. 1, 64-73.

Bernareggi, A. (1998) Clinical pharmacokinetic of nimesulide. Clin. Pharmacokinet.35, 247-274.

Blain, H., Boileau, C., Lapicque, F., Nedelec, E., Loeuille, D., Guillaume, C.,Gaucher, A., Jeandel, C., Netter, P. and Jouzeau, J.Y. (2002) Limitation of the invitro whole blood assay for predicting the COX selectivity of NSAIDs in clinicaluse. Br. J. Clin. Pharmacol. 53, 255-265.

Bree, F., Nguyen, P., Urien, S., Albengres, E., Macciochi, A. and Tillement, J.P. (1993)Nimesulide binding to components within blood. Drugs 46 Suppl. 1, 83-90.

Brideau, C., van Staden, C. and Chung Chan, C. (2001) In vitro effects ofcyclooxygenase inhibitors in whole blood of horses, dogs, and cats. Am. J. vet.Res. 62, 1755-1760.

Campbell, N.B. and Blikslager, A.T. (2000) The role of cyclooxygenase inhibitors inrepair of ischaemic-injured jejunal mucosa in the horse. Equine vet. J. 32, 59-64.

Cannon, G.W. (1999) Cyclooxygenase-2 selective inhibitors. Drugs Today 35, 487-496.

Cryer, B. and Feldman, M. (1998) Cyclooxygenase-1 and cyclooxygenase-2selectivity of widely used nonsteroidal anti-inflammatory drugs. Am. J. Med.104, 413-421.

Dionne R.M., Vrinis A., Doucet M.Y. and Pare J. (2003) Gastric ulcers instandardbred racehorses : prevalence, lesion description and risk factors. J. vet.

R. Villa et al. 7

intern. Med. 17, 218-222.

Dyke, T.M, Hubbell, J.A., Sams, R.A. and Hinchcliff, K.W. (1998) Hepatic bloodflow in horses during the recuperative period from maximal exercise. Am. J. vet. Res. 59, 1476-1480.

Fabiola, G.F., Pattabhi, V. and Nagarajan, K. (1998) Structural basis for selectiveinhibition of COX-2 by nimesulide. Bioorganic Medicin. Chem. 6, 2337-2344.

Famaey, J.P. (1997) In vitro and in vivo pharmacological evidence of selectivecyclooxygenase-2 inhibition by nimesulide: an overview. Inflamm. Res.46, 437-446.

Feldman, M. and McMahon, A.T. (2000) Do cyclooxygenase-2 inhibitors providebenefits similar to those of traditional nonsteroidal anti-inflammatory drugs, withless gastrointestinal toxicity? Ann. Int. Med. 132, 134-143.

Gerring E.L., Lees P. and Taylor J.B. (1981) Pharmacokinetics of phenylbutazoneand its metabolites in the horse. Equine vet. J. 13, 152-157.

Gibaldi, M. and Perrier, D. (1982) Pharmacokinetics. Marcel Dekker Inc., New York.

Kam, P.C. and See, A.U. (2000) Cyclo-oxygenase isoenzymes: physiological andpharmacological role, Anaesthesia 55, 442-449.

Kaplan-Machlis, B. and Klostermeyer, B.S. (1999) The cyclooxygenase-2 inhibitors:safety and effectiveness. Ann. Pharmacother. 33, 979-988.

Kearney, P.M., Baigent, C., Godwin, J., Halls, H., Emberson, J.R. and Patrono, C.(2006) Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidalanti-inflammatory increase the risk of atherothrombosis? Meta-analysis ofrandomised trials drugs traditional non-steroidal anti-inflammatory, Br. med. J.332, 1302-1308.

Lam, F.C., Hung, C.T. and Perrier, D.G. (1985) Estimation of variance for harmonicmean half-lives. J. Pharm. Sci. 74, 229-231.

Landoni, M.F. and Lees, P. (1995) Influence of formulation on the pharmacokineticsand bioavailability of racemic ketoprofen in horses. J. vet. Pharmacol. Therap.18, 446-450.

Lees, P. and Higgins, A.J. (1985) Clinical pharmacology and therapeutic use of nonsteroidal anti-inflammatory drugs in horse. Equine vet. J. 17, 83-96.

Lees, P., May, S.A. and McKellar, Q.A. (1991) Pharmacology and therapeutics of nonsteroidal anti-inflammatory drugs in the dog and cat: 1. General Pharmacology.J. small anim. Pract. 32, 183-193.

Lees, P., Landoni, M.F., Giraudel, J. and Toutain, P.L. (2004a) Pharmacodynamicsand pharmacokinetics of nonsteroidal anti-inflammatory drugs in species ofveterinary interest. J. vet. Pharmacol. Therap. 27, 479-490.

Lees, P., Giraudel, J., Landoni, M.F. and Toutain, P.L. (2004b) PK-PD integration andPK-PD modelling of nonsteroidal anti-inflammatory drugs: principles andapplications in veterinary pharmacology. J. vet. Pharmacol. Therap. 27, 491-502.

Maitho, T.E., Lees, P. and Taylor, J.B. (1986) Absorption and pharmacokinetics ofphenylbutazone in Welsh Mountain ponies. J. vet. Pharmacol. Therap. 9, 26-39.

Moreau, M., Daminet, S., Martel-Pelletier, J., Fernandes, J. and Pelletier, J.P. (2005)Superiority of the gastroduodenal safety profile of licofelone over rofecoxib, aCOX-2 selective inhibitor, in dogs. J. vet. Pharmacol. Therap. 28, 81-86.

Pairet, M. and van Ryn, J. (1998) Experimental models used to investigate thedifferential inhibition of cyclooxygenase-1 and cyclooxygenase-2 by non-steroidal anti-inflammatory drugs. Inflamm Res. 47 Suppl. 2, S93-101.

Psaty, B.M. and Furberg, C.D. (2005) COX-2 inhibitors - lessons in drug safety. N. Engl. J. Med. 352, 1133-1135.

Rabasseda, X. (1996) Nimesulide: a selective cyclooxygenase 2 inhibitorantiinflammatory drug. Drugs Today, 32 Suppl. D, 1-23.

Ricketts, A.P., Lundly, K.M. and Seibel, S.B. (1998) Evaluation of selectiveinhibition of canine cyclooxygenase 1 and 2 by carprofen and other nonsteroidalanti-inflammatory drugs, Am. J. vet. Res. 59, 1441-1446.

Sarkar, P., McIntosh, J.M., Leavitt, R. and Gouthro, H. (1997) An unique metaboliteof Nimesulide. J. Anal. Toxicol. 21, 197-202.

Singh, S., Sharda, N. and Mahajan, L. (1999) Spectophotometric determination ofpKa of nimesulide. Int. J. Pharmaceut. 176, 261-264.

Talley, J.J. (1999) Selective inhibitors of cyclooxygenase-2 (COX-2). Progr. medicin.Chem. 36, 201-234.

Toutain, P.L., Cester, C.C., Haak, T. and Metge, S. (2001a) Pharmacokinetic profileand in vitro selective cyclooxigenase-2 inhibition by nimesulide in the dog. J.vet. Pharmacol. Therap. 24, 35-42.

Toutain, P.L., Cester, C.C., Haak, T. and Laroute, V. (2001b) Apharmackinetic/pharmacodynamic approach vs. a dose titration for thedetermination of a dose regimen: the case of nimesulide, a COX-2 selective non-steroidal anti-inflammatory drug in the dog. J. vet. Pharmacol. Therap. 24, 43-55.

Toutain, P.L. and Bousquet-Melou, A. (2004) Bioavailability and its assessment. J.vet. Pharmacol. Therap. 27, 455-466.

van Hoogmoed, L.M., Snyder, J.R. and Harmon, F.A. (2002) In vitro investigation ofthe effects of cyclooxygenase-2 inhibitors on contractile activity of the equinedorsal and ventral colon. Am. J. vet. Res. 65, 1496-1500.

Vane, J.R., Bakhle, Y.S. and Botting, RM. (1998a) Cyclooxygenase 1 and 2. Ann.Rev. Pharmacol. Toxicol. 38, 97-120.

Vane, J.R. and Botting, R.M. (1998b) Mechanism of action of anti-inflammatorydrugs. Int. J. Tissue React. 20, 3-15.

Villa, R., Prandini, E., Montesissa, C. and Carli, S. (1994) Serum protein binding ofβ-lactamine derivatives in farm and domestic animals. Proceedings of the 6thInternational Congress of EAVPT, 7-11 August 1994, Edinburgh UK., J. vet. Pharmacol. Therap. 216-217.

Villa, R., Prandini, E., Caloni, F. and Carli, S. (1997) Serum protein binding of somesulphonamides, quinolones and floroquinolones in farm and domestic animals. J.vet. Pharmacol. Therap. 20 Suppl. 1, 34-35.

Vonkeman, H.E., Brouwers, J.R.B.J. and van de Laar, M.A.F.J. (2006) Understandingthe NSAID related risk of vascular events. BMJ 332, 895-898.

Wallace, J.L. (2002) Selective cyclooxygenase-2 inhibitors: after the smoke hascleared. Digest. Liver Dis. 34, 89-94.

Yamaoka, K., Nakagawa, T. and Uno, T. (1978) Application of Akaike’s informationcriterion (AIC) in the evaluation of linear pharmacokinetic equations, J.Pharmacokin. Biopharmac. 6, 165-175.

Author contributions The initiation, conception and planning ofthis study were by R.V., C.B., E.F. and S.C. Its execution was byR.V., P.C., C.B., A.Z., C.Z. and E.F., statistics by R.V., P.C., C.B. andA.Z., and writing by R.V., P.C., C.B. and S.C.