Background : Lumiracoxib (Prexige®) is a cyclooxygenase-2 (COX-2) selective inhibitor.
Aim : To compare the gastroduodenal tolerability of lumiracoxib with placebo and naproxen in a randomized, parallel-group, double-blind study.
Methods : Sixty-five healthy male subjects were randomized to receive 8 days' dosing with lumiracoxib 200 mg twice daily (b.d.) (n = 21), placebo (n = 22) or naproxen 500 mg b.d. (n = 22). Endoscopic evaluations of gastric and duodenal mucosae were conducted at baseline and after 8 days' dosing. Serum was assayed for ex-vivo concentrations of thromboxane B2 (TxB2) to determine cyclooxygenase-1 (COX-1) inhibitory activity.
Results : Sixty subjects (20 per group) completed the study. No gastroduodenal erosions were observed in subjects receiving lumiracoxib. Thirteen subjects receiving naproxen developed duodenal erosions. At the gastric site, one subject in each of the naproxen and placebo groups had erosions; one subject receiving naproxen also developed a small asymptomatic gastric ulcer. Gastrointestinal adverse events accounted for 42.3% of all adverse events, occurring in 3/21, 4/22 and 6/22 of the lumiracoxib, placebo and naproxen groups, respectively. TxB2 levels were similar for patients receiving placebo or lumiracoxib, but were reduced by > 95% in patients receiving naproxen, compared with placebo.
Conclusions : Multiple doses of lumiracoxib resulted in gastroduodenal tolerability similar to placebo and superior to naproxen.
The management of pain associated with arthritis frequently involves the use of nonsteroidal anti-inflammatory drugs (NSAIDs).1,2 Traditional NSAIDs exert their analgesic and anti-inflammatory effects mainly through nonspecific inhibition of cyclooxygenase (COX)-mediated prostaglandin synthesis, by inhibiting both COX-1 and COX-2 isoenzymes.3–5
COX-1 is expressed throughout the body and plays an important role in the normal physiology of the gastrointestinal tract, kidneys and platelets.6–9 As a result, inhibition of COX-1 can result in adverse gastrointestinal outcomes, kidney dysfunction and inhibition of platelet aggregation.10–12 Gastropathy associated with traditional NSAIDs is well recognized and often serious.13,14 Endoscopy studies estimate that 15–30% of patients using traditional NSAIDs on a regular basis can be expected to develop a gastrointestinal ulcer.15 In severe cases, traditional NSAID-induced gastric and duodenal ulcers can perforate and bleed without warning, sometimes even leading to death.16
COX-2 is expressed in certain tissues; however, it is highly up-regulated in inflamed tissue and is thought to be the COX isoenzyme primarily responsible for prostaglandin-mediated pain and inflammation.7,17 One current hypothesis is that COX-1 inhibition underlies the gastropathy associated with traditional NSAIDs, and the development of COX-2 selective inhibitors may be expected to result in an improved gastrointestinal tolerability profile. Indeed it is now well recognized that COX-2 selective inhibitors, such as celecoxib and rofecoxib, are associated with a marked decrease in the incidence of gastrointestinal adverse effects and inhibition of platelet function, compared with traditional NSAIDs.12,18–24
Lumiracoxib is a novel COX-2 selective inhibitor in development for the management of osteoarthritis, rheumatoid arthritis and acute pain. Lumiracoxib is the only COX-2 selective inhibitor to have a carboxylic acid moiety, which may lead to its interesting pharmacokinetic/pharmacodynamic profile. Lumiracoxib is absorbed rapidly, and has a plasma elimination half-life of 3–6 h.25,26 Lumiracoxib ‘exposure’ increases with increasing dose in a dose-proportional manner up to 800 mg.26 In a multiple-dose study in patients with osteoarthritis, lumiracoxib demonstrated time-independent pharmacokinetics with no sign of accumulation upon 4-week multiple dosing.27 In this study, hysteresis observed in pain recorded on a visual analogue scale plotted against lumiracoxib plasma concentration suggests that the action of lumiracoxib is influenced by concentration in an effect compartment (e.g. an inflamed joint).27 The COX selectivity profile of lumiracoxib determined during preclinical studies and a single-dose study in humans suggests that its gastrointestinal toxicity may be reduced compared with traditional NSAIDs.26,28
In the present placebo-controlled pilot study, the safety and tolerability of multiple oral doses of lumiracoxib 200 mg b.d. was evaluated in healthy subjects, using endoscopic evaluation to focus particularly on gastroduodenal tolerability compared with both placebo and the traditional NSAID, naproxen. A dose of lumiracoxib 200 mg b.d. was chosen based on the pharmacokinetic and pharmacodynamic results of the first human study with lumiracoxib, in which lumiracoxib demonstrated rapid absorption, achieving maximum plasma concentrations within 2–3 h of dosing.26 Based on this, assessment of the possible effects of lumiracoxib on the COX-1 enzyme (measured via thromboxane B2[TxB2] levels) was performed at 2 and 4 h post-dose in this study, in order to evaluate the impact at expected peak plasma concentrations. Naproxen was chosen as the active control on the basis that in previous studies it produced gastroduodenal erosions in 72% of healthy volunteers within 7 days of multiple dosing.29
Healthy male and female subjects aged 18–45 years were recruited to take part in this single-centre, 8-day, double-blind, double-dummy, multiple-dose, placebo-controlled study. The study was conducted in accordance with Good Clinical Practice and the declaration of Helsinki (and subsequent revisions) and was approved by the local ethics review board. All subjects provided written informed consent.
Subjects were screened for eligibility during a 21-day period before the start of the study and were required to be in good health as determined by past medical history, physical examination, electrocardiogram (ECG), laboratory tests and urinalysis. Females must have undergone documented surgical sterilization at least 6 months before the study and were required to provide a negative pregnancy test. Subjects were nonsmokers or light smokers (< 10 cigarettes per day).
Subjects were required to have normal gastric mucosa at prebaseline endoscopy and a negative test result for Helicobacter pylori infection. Individuals had to be a minimum weight of 45 kg. Body weight, body frame and elbow breadth were used to determine whether subjects were within 20% of their ideal weight;30 subjects outside this range were excluded from the study. Individuals were excluded if they had used an NSAID within 2 weeks of study start, any other prescription medication within 1 month of study start, or any over-the-counter medication or vitamins within 7 days of study start. Furthermore, individuals were not eligible for inclusion in the study if they had any surgical or medical condition likely to affect drug disposition or safety, such as a history of inflammatory bowel syndrome, gastritis, ulcers and gastrointestinal or rectal bleeding, major gastrointestinal surgery, pancreatic injury or pancreatitis, urinary obstruction, or NSAID- or aspirin-induced gastritis.
Eligible subjects were randomized in the ratio 1 : 1 : 1 to receive 14 doses of lumiracoxib 200 mg b.d., naproxen 500 mg b.d. or placebo during an 8-day dosing phase in a double-blind, double-dummy fashion. Subjects were resident in the test centre for the entire course of the study, from at least 10 h before dosing on day 1 until day 9. All study medication was given with 200 mL of water shortly before or after food (with the exception of day 6, when fasting for 10 h was required before blood sampling for the pharmacokinetic and pharmacodynamic analyses). The first dose of medication was given on the evening of day 1. Thereafter, subjects were dosed at 12-hourly intervals, until the last dose on the morning of day 8. Individuals were discharged from the test centre on day 9. Caffeine-containing food and drinks, spicy foods, alcohol, other gastric irritants and strenuous physical exercise were not permitted for the duration of the study.
Gastroduodenal endoscopy was carried out at screening, before dosing on day 1 and after the final dose in the evening of day 8 by a single endoscopist. Subjects were administered a local anaesthetic [lignocaine (lidocaine) hydrochloride], a short-acting sedative (midazolam) and a drug to reverse sedation (flumazenil), on an as-required basis at the discretion of the endoscopist before, during and after endoscopy, respectively, following consultation with each subject in advance of the procedure. The endoscopist provided a written assessment of the number of lesions or erosions in the gastric and duodenal mucosa. Endoscopy results were graded according to an existing eight-point scale, modified from a scoring system developed by Lanza and colleagues (Table 1).29,31
Table 1. Endoscopic grading of gastric and duodenal lesions,29,31
> 10 petechiae
1–5 erosions (defined as a lesion producing a definite discontinuance in the mucosa, but without depth)
> 25 erosions
Ulcer (defined as any lesion of any size with unequivocal depth)
All subjects who received at least one dose of study medication were included in the safety and tolerability analyses. Adverse events, serious adverse events and vital signs were monitored throughout the study in addition to routine biochemistry, haematology, urinalysis and physical examinations. Faecal occult blood tests were also conducted at screening and on each study day using the Fecatwin SST/Fecatwin EIA assays according to the manufacturer's instructions (Labsystems). ECG recordings were performed at baseline (day 1) and day 8, before the endoscopy procedures, and after the study was completed.
On day 6, blood samples (2 × 3 mL) were collected into heparinized tubes by either direct venepuncture or an indwelling cannula inserted in a forearm vein. Samples taken pre-dose and at 1, 2, 4, 8 and 12 h after the morning dose on day 6 were centrifuged between 3 and 5 °C for 15 min at approximately 800 × g within 30 min of sampling. Approximately 1.5 mL of plasma were then stored at −20 °C or below within 60 min of sampling, pending analysis.
A high-performance liquid chromatographic mass spectrometry (HPLC/MS) method was used to determine lumiracoxib plasma concentrations with a lower limit of quantification of 10 ng/mL.
Naproxen plasma concentrations were determined using HPLC tandem mass spectrometry (HPLC/MS/MS) in selected reaction monitoring mode using electrospray ionization as the interface.
Thromboxane enzyme immunoassay
Serum TxB2 concentrations were determined ex vivo in order to assess inhibition of COX-1 in vivo, using an established assay.32 Separate blood samples (2 mL) were collected by venepuncture into regular glass vacutainers pre dose, 2 and 4 h after dosing on day 6. Samples were placed at 37 °C within 5 min of sampling and left for 1 h to allow the blood to clot. Supernatant serum was then collected following centrifugation at 12 000 × g for 5 min, and two 0.1 mL aliquots were dosed with 0.4 mL methanol to enable extraction of prostaglandins and precipitation of proteins. Samples were centrifuged again at 12 000 × g for 5 min and the supernatants were stored at −70 °C until the TxB2 enzyme immunoassay (Cayman Chemicals) was carried out.
Sample size was determined based on the expected percentage occurrence of gastroduodenal erosions in each group. The χ2-test was used to compare lumiracoxib and naproxen with respect to these percentages. An erosion rate of 72% has been reported for naproxen in a previous similar study in 32 healthy subjects within 7 days of multiple dosing and a rate of 9–13% has been reported for individuals taking placebo.29 Lumiracoxib was not expected to induce endoscopically detectable lesions in more than 25% of subjects based on documented evidence of rates with other COX-2 selective inhibitors. Therefore, this value was used as a maximum for the purposes of sample size determination. Based on these expected frequencies, it was calculated that with a minimum of 19 subjects per group there would be a probable incidence of 14 subjects with erosions in the naproxen group. The study would then provide 80% power to detect differences between the groups at a 5% level of significance.
Descriptive statistics are presented for demographic variables, vital signs, ECG evaluation and laboratory values. The following pharmacokinetic values were determined using noncompartmental methods for each drug: maximum plasma concentration at steady state (Cmaxss); time to maximum plasma concentration (Tmax); and the area under the concentration–time curve over the 12-h dosing period (AUCτ). Plasma concentrations below the lower limit of quantitation were treated as zero.
An analysis of variance was used to compare TxB2 levels (log transformed) between dosing groups. The relationship between plasma drug concentration and level of TxB2 inhibition was also assessed.
Sixty-five subjects were recruited into the study and were randomized to receive lumiracoxib (200 mg b.d., n = 21), naproxen (500 mg b.d., n = 22) or placebo (n = 22) for 8 days. Five subjects were withdrawn from the study after receiving a single dose of study medication because of abnormal pre-dose gastric mucosa revealed by endoscopy. The remaining 60 subjects (20 per group) all completed the study.
The study population consisted of males (although the protocol was open to the enrolment of women) with a mean age of 27.2 years. With one exception, all subjects were Caucasian. The three groups were well matched with respect to baseline demographic characteristics, with no apparent differences between groups (Table 2).
Table 2. Study population characteristics
Lumiracoxib 200 mg b.d. (n = 21)
Naproxen 500 mg b.d . (n = 22)
Placebo (n = 22)
Total (n = 65)
Values are mean ± s.d., unless otherwise stated.
25.7 ± 6.2
28.4 ± 6.7
27.5 ± 5.9
27.2 ± 6.3
79.0 ± 10.4
73.6 ± 11.0
76.2 ± 10.5
76.2 ± 10.7
181.5 ± 9.6
177.1 ± 5.5
178.7 ± 7.4
179.1 ± 7.7
Elbow breadth (cm)
7.5 ± 0.4
7.4 ± 0.3
7.4 ± 0.3
7.4 ± 0.3
Body frame, n (%)
Fifty-two adverse events were reported during the study in 25 of the 65 subjects (38%) who received at least one dose of study drug (Table 3). Of these, eight (15%) were classified as moderate in severity; the remainder were classified as mild and all were transient. Nineteen adverse events (37%) were suspected to be related to study drug by the clinical investigator. The gastrointestinal tract was associated with the most adverse events (13 subjects with a total of 22 adverse events). More subjects in the naproxen group (n = 6) reported gastrointestinal adverse events than in the placebo (n = 4) or lumiracoxib (n = 3) groups. No serious adverse events were reported during the study.
Table 3. Summary of adverse events (AEs)
Type/location of adverse event
Number of events (number of individuals)
Lumiracoxib 200 mg b.d. (n = 21)
Naproxen 500 mg b.d. (n = 22)
Placebo (n = 22)
Total (n = 65)
Nervous system disorders
Renal and urinary disorders
No clinically significant changes were observed in vital signs, physical examinations or ECGs during the study.
Nine subjects experienced transient elevations in liver enzymes [aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT)] during the study: four subjects receiving lumiracoxib, four subjects receiving placebo and one subject receiving naproxen. The n-fold increase over the upper limit of normal (ULN) ranged from 1.1 to 4.2 for AST and from 1.1 to 2.2 for ALT. An increase in AST > 4-fold over the ULN was observed in one subject only at study end. This was considered by the investigator to be most likely a result of intense physical exercise, since this subject also had increased creatinine phosphokinase and lactate dehydrogenase concentrations. There were no clinical events, discontinuations or interruptions of therapy associated with these increases, and all were followed to resolution or to the satisfaction of the investigator. In addition, the liver enzyme increases were not associated with increased bilirubin levels.
One subject receiving naproxen experienced a decrease in haemoglobin of 2 g/dL during the study (12.3 g/dL on day 8 from 14.3 g/dL at baseline). Values returned to normal by the final study evaluation. A positive faecal occult blood result was reported for one subject in the lumiracoxib group on day 8. However, there were no clinical signs or symptoms, or endoscopic findings relating to this event and no adverse events were reported.
All other deviations in biochemistry and urinalysis were minor and not considered clinically significant.
Incidence of gastric erosions and ulcers
After the final dose of the dosing phase on day 8, no duodenal erosions (endoscopy Grade 3 or above) were observed in subjects in the lumiracoxib or placebo groups (Figure 1). This was a significant reduction in frequency compared with the 13/20 (65%) subjects having duodenal erosions in the naproxen group.
With respect to the gastric sites, one subject in the naproxen group developed an asymptomatic gastric ulcer and one subject in the placebo group developed a Grade 4 gastric erosion. No gastric erosions were reported in the lumiracoxib group (Figure 1).
Following multiple daily dosing with lumiracoxib 200 mg b.d. the mean (± s.d.) pre-dose plasma concentration on day 6 was 311.7 ± 129.8 ng/mL and the mean (± s.d.) Cmaxss was 2550.4 ± 1267.7 ng/mL. At 12 h after dosing, the mean (± s.d.) plasma concentration was 189.0 ± 59.9 ng/mL and the mean (± s.d.) AUCτ was 12 502 ± 2427.3 ng·h/mL. The median Tmax occurred 2.0 h after dosing. Figure 2 illustrates the mean plasma concentration over time on day 6.
In the naproxen group the mean (± s.d.) pre-dose plasma concentration on day 6 was 39 036 ± 8491 ng/mL and the mean (± s.d.) Cmaxss was 93 712 ± 24 223 ng/mL. At 12 h after dosing, the mean (± s.d.) plasma concentration was 34 370 ± 8484 ng/mL and the mean (± s.d.) AUCτ was 661 711 ± 140 929 ng·h/mL. The median Tmax occurred 1.5 h after dosing.
Mean serum TxB2 concentrations were similar for the placebo and lumiracoxib groups before dosing as well as 2 and 4 h after dosing on day 6 (Figure 3; Table 4). However, TxB2 concentrations in the naproxen group were significantly decreased compared with those observed in either the placebo or lumiracoxib groups at each timepoint (Figure 3; Table 4). Compared with placebo, mean TxB2 levels were reduced > 95% in the naproxen group at 2 and 4 h after dosing on day 6 of the study. Pre-dose TxB2 concentrations were also low in the naproxen group, suggesting that inhibition lasted over the dosing interval. There was no apparent relationship between lumiracoxib and TxB2 serum concentrations, even at the highest observed lumiracoxib concentration (∼5000 ng/mL or 17 µm) (Figure 4a). In contrast, TxB2 serum concentrations were reduced at all pre- and post-dose naproxen plasma concentrations (Figure 4b).
Table 4. Analysis of the differences in plasma thromboxane inhibition between groups (lumiracoxib, placebo and naproxen), before and after dosing
Day 6 (pre-dose)
Lumiracoxib vs. naproxen
Placebo vs. naproxen
Lumiracoxib vs. placebo
Day 6 (2 h post-dose)
Lumiracoxib vs. naproxen
Placebo vs. naproxen
Lumiracoxib vs. placebo
Day 6 (4 h post-dose)
Lumiracoxib vs. naproxen
Placebo vs. naproxen
Lumiracoxib vs. placebo
In this study, lumiracoxib, a novel COX-2 selective inhibitor, was well tolerated in healthy male subjects at a dose of 200 mg b.d. Multiple doses of lumiracoxib 200 mg b.d. over an 8-day period were associated with an incidence of gastrointestinal adverse events and gastroduodenal erosions/ulcers which was similar to placebo and superior to a standard dose (500 mg b.d.) of a traditional NSAID, naproxen. Indeed, no erosions were detected in any subjects who received lumiracoxib, compared with 13 subjects who developed duodenal erosions (Grade 3 and above) and one subject who developed a gastric ulcer in the naproxen group. These data are corroborated by a second lumiracoxib study in which it was demonstrated that no gastric or duodenal erosions were observed in healthy volunteers after dosing with lumiracoxib 800 mg once daily (o.d.) for 8 days.33 In addition, a longer-term 13-week, endoscopic study in more than 1000 patients with osteoarthritis demonstrated that the cumulative incidence of gastroduodenal ulcers and number of patients with > 10 erosions was also significantly reduced with lumiracoxib (200 or 400 mg o.d.) than was observed for the traditional NSAID ibuprofen.34 These results are also consistent with the findings of a previous 1-week endoscopy study comparing the effects of celecoxib, naproxen and placebo on gastroduodenal mucosa, where the incidence of gastric erosions and/or ulcers ranged from 9 to 13% in patients receiving celecoxib or placebo, compared with 72% in patients who received naproxen.29 Previous studies have shown that naproxen is associated with mucosal injury/lesions in both duodenal and gastric regions of the gastrointestinal tract.35,36 Although the gastric site was affected, the predominance of erosions in the duodenum described here may possibly relate to the conditions of the study mandating domiciliation, resting and food intake with study drugs. These conditions may have favoured delivery of high concentrations of the study drugs to the duodenal site, leading for naproxen to impairment of prostaglandin (COX-1)-dependent mucosal protective mechanisms.
The results of endoscopy safety studies provide a useful insight into short-term gastrointestinal safety; however, these findings need to be confirmed in larger studies evaluating long-term gastrointestinal outcomes in order to determine fully the clinical impact of lumiracoxib on gastrointestinal safety and tolerability. Notably, several studies with celecoxib and rofecoxib have confirmed that COX-2 selective inhibition is associated with a decreased incidence of gastric and duodenal ulcers and a markedly lower rate of gastrointestinal adverse events compared with traditional NSAIDs, including naproxen.19, 20, 23, 24, 37, 38. Indeed, a systematic review of nine randomized controlled studies involving more than 15 000 patients with rheumatological disease found that the incidence of ulcers detectable by endoscopy was 71% lower with celecoxib compared with traditional NSAIDs.39 Debate still exists as to the relevance of ulcers compared with complicated ulcers, obstructions and bleeding in the clinical setting, with complicated ulcers and their associated effects considered to provide a more realistic assessment of clinical impact, particularly for NSAID users.40 Results from the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET), which is assessing the incidence of perforations, obstructions and bleeds in > 18 000 patients with osteoarthritis receiving lumiracoxib 400 mg o.d., naproxen 500 mg b.d. or ibuprofen 800 mg three times daily over 52 weeks, is designed to provide important, robust conclusions about the relative safety of COX-2 selective inhibitors and traditional NSAIDs in the clinical setting.
In this study, pharmacokinetic analysis revealed that maximal plasma drug concentrations of lumiracoxib 200 mg were achieved 2 h after dosing. This result is in agreement with pharmacokinetic analyses performed in both healthy volunteers and patients with osteoarthritis, in which a median time to maximal drug concentration of 2–3 h was observed.26,27
Lumiracoxib and placebo were associated with similar serum TxB2 concentrations at all timepoints assessed. These were significantly different from those observed for naproxen, which demonstrated profound and sustained inhibition of TxB2 production, reaching a mean of > 95% inhibition compared with placebo at 2 and 4 h after dosing. COX-1-derived prostaglandins mediate the synthesis of TxB2 in platelets, and because COX-1 is produced at a fairly constant level in cells throughout the body, measurement of TxB2 inhibition can be used as a marker for inhibition of platelet activity and, by extrapolation, COX-1 activity.17 TxB2 inhibition was therefore selected as a convenient, well established and minimally intrusive assay for the activity of COX-1. TxB2 assays were carried after 5 days' multiple dosing both before, and 2 and 4 h after dosing in order to correspond with maximal plasma drug concentrations. These data clearly suggest that lumiracoxib has no effect on COX-1 activity. In addition, naproxen was shown to inhibit TxB2 over the dosing interval, suggesting that a sustained effect may be associated with development of gastroduodenal erosions. Similar results have been observed in a study of rofecoxib and naproxen, where naproxen was shown to inhibit plasma TxB2 synthesis by > 95%, yet no such inhibition was observed for rofecoxib.41
In conclusion, this preliminary study in healthy male subjects has shown that multiple daily dosing with lumiracoxib did not inhibit TxB2 production, indicating a lack of effect on the COX-1 isoenzyme. Consistent with COX-2 selectivity, lumiracoxib demonstrated a superior gastrointestinal safety profile over the course of the study as assessed by endoscopy when compared with the nonselective NSAID naproxen. Furthermore, lumiracoxib was well tolerated and was associated with similar rates of gastroduodenal erosions (no erosions in the lumiracoxib group compared with one in the placebo group) and gastrointestinal AEs to placebo.