- Email: [email protected]
Efficacy and tolerability of meloxicam in an observational, controlled cohort study in patients with rheumatic disease
CLINICAL THERAPELJTICYIVOL.
22, NO. 4,200O
Efficacy and Tolerability of Meloxicam in an Observational, Controlled Cohort Study in Patients with Rheumatic Disease Frank Degner, MD,’ RalfSigmund,2
and Henning Zeidler, MD3
‘Therupeutic Area CNSIGeneral Diseases, Boehringer Ingelheim GmbH, Ingelheitn, 2Depurtment of Medical Datu Services, Boehringer Ingelheirn Phurrna KG, Bibrrach, and ‘Division of Rhewnutology, Depurtment of Internul Medicine, kklicul School Hunnover, Hunnover, Germany
ABSTRACT Background: Meloxicam is a nonsteroidal anti-inflammatory drug (NSAID) that inhibits cyclooxygenase-2 without affecting platelet aggregation. Objective: The goal of this study was to assessthe efficacy and tolerability of meloxicam under natural prescribing conditions for up to 6 months. Methods: This was a multicenter, prospective, observational cohort study. Participating centers were randomized to I of 2 groups: the meloxicam-only group, and the group who received comparator NSAlDs (ie, diclofenac, ibuprofen, piroxicam, or indomethacin). Results: A total of 4526 patients received either meloxicam (n = 2530) or a comparator NSAID (n = 1996). Treatment groups were comparablein terms of observed efficacy measures.Meloxicam patients had significantly lower rates of reported gastrointestinal (GI) adverse drug reactions (I .80% vs 3.20%; P = 0.003), including dyspepsia(0.08% vs 0.35%; P = 0.049), abdominal pain (0.91% vs 1.90%; P = 0.006), gastritis (0.08% vs 0.60%; P = 0.002), and GI bleeding (0.08% vs 0.50%; P = 0.007), compared with those receiving comparator NSAIDs. Conclusion: There was evidence that meloxicam was prescribed preferentially to patients who had not responded to NSAIDs previously, who had previously experienced NSAID-induced side effects (and so were at high risk for developing NSAID-induced GI toxicity), or who were more seriously impaired. The nonrandomized and unblinded nature of this study limits the conclusionsthat can be drawn concerning efficacy or tolerability. Nevertheless,the study results are consistentwith the favorable GI tolerability seen with meloxicam in double-blind comparative trials. Key words: NSAID, COX-2 inhibition, cohort study, meloxicam, gastrointestinal adverse effects. (Clin Ther. 2000;22:400410) Accepfed for publication Februay Printed in the USA. Reproduction
7, 2000. in whole
or part
is not permitted.
F. DEGNER ET AL.
INTRODUCTION Although nonsteroidal anti-inflammatory drugs (NSAIDs) are integral to arthritis therapy, their use has been limited by gastrointestinal (GI) side effects.’ Singh et al2 found that 15% of rheumatoid arthritis patients treated with classic NSAIDs reported Cl side effects over an observation period of 2.5 years; other investigators have reported that NSAID users have an -3 times greater relative risk (RR) for serious GI events than those who do not use NSAIDs.” Thus, it is important to develop treatment strategies that minimize the GI side effects of these valuable medications. Meloxicam, an NSAID of the enolic acid class, inhibits the cyclooxygenase-2 (COX-2) isozyme more potently than the COX-I isozyme4x5 and lacks an inhibitory effect on platelet aggregation at therapeutic doses.6 Introduced in Europe in 1996, meloxicam is available in >90 countries worldwide. In the MELISSA (Meloxicam Large-Scale International Study Safety Assessment) triak7 which involved 9323 patients with osteoarthritis, meloxicam exhibited comparable efficacy and a significantly improved Gl tolerability profile compared with diclofenac, a drug with relatively low GI toxicity. Similar results were seen in the SELECT (Safety and Efficacy Large Scale Evaluation of COX Inhibition Therapies) ttiaLx in which meloxicam was associated with a significantly lower incidence of GI side effects compared with piroxicam. These results were also confirmed in a pooled analysis of all double-blind trials comparing meloxicam with diclofenac, piroxicam, and naproxen.’ However, comparisons with other NSAIDs have not been reported. The efficacy and tolerability of meloxicam have been demonstrated in doubleblind, randomized trials, but to date no in-
formation has been published regarding its use in actual clinical practice. The limitations of observational open-label studies are well known,” particularly with regard to the underreporting of side effects.” At the same time, such studies can complement the results of blinded trials by providing useful information about the efficacy and tolerability of medications under natural prescribing conditions. Also, underreporting can be controlled to some degree by reference to a control group. We therefore undertook this observational cohort study to assess the efficacy and tolerability of meloxicam compared with those of other commonly used NSAIDs. PATIENTS
AND METHODS
Study Design and Protocol This was a prospective, open-label, parallel-group, observational cohort study performed in Germany from August 1996 through July 1997. A total of lo,5 15 physicians specializing in family medicine or internal medicine who had not received sales presentations concerning meloxicam were recruited for the study; 683 agreed to participate and were randomized to 1 of 2 observation groups (group 1 = meloxicam only and group 2 = comparator NSAlDs) without their knowledge. Physicians were free to prescribe any NSAID they deemed appropriate, but the group I analysis included only efficacy and tolerability data from patients prescribed meloxicam, whereas the group 2 analysis included only data from patients prescribed the comparator NSAIDs. This approach was intended to control for a potential prescribing bias caused by the launch of meloxicam, which had been shown in double-blind clinical trials”%‘2 to have a favorable GI tolerability 401
CLINICAL
profile over classic NSAIDs. Physicians received a study protocol reflecting their randomization status. Patients were observed in the prescribing setting in a manner unaffected by the randomization procedure applied to the participating centers. To be eligible for the trial, patients had to be diagnosed with acute or chronic active rheumatic disease for which NSAID therapy was required. Patients were also eligible if they had been previously diagnosed with rheumatic disease and required a switch to another NSAID because of adverse reactions or lack of efficacy. They were excluded if they had any contraindication listed in the product information for the prescribed NSAID. The selection of an appropriate NSAID for each patient (either meloxicam or another NSAID) was left to the individual physicians. Concomitant therapy, including steroids, aspirin, anticoagulants, other NSAlDs, and acetaminophen, was also allowed at the physician’s discretion. Baseline comparability of treatment groups was considered important for assessment of efficacy and tolerability. Demographic and medical data were obtained at baseline regarding age; sex; indication for NSAID therapy; history of perforation, ulceration, or Cl bleeding (PUB); history of previous NSAID use with or without associated side effects; and lack of efficacy of previous NSAID therapy. According to guidelines issued by the European Agency for the Evaluation of Medicinal Products, this observational pharmacoepidemiologic study was not required to obtain patients’ informed consent. The study was performed in accordance with German drug law and European Union pharmacovigilance guidelines. Patients were observed for up to 6 months to assess the efficacy and tolera402
THERAPEIJTICS’
bility of the treatment regimens. After the initial visit at which patients received a prescription for 1 of the study drugs, 3 follow-up visits were scheduled at 1 month, between months I and 3, and between months 3 and 6. Additional visits could be scheduled if medically indicated. A physical examination was performed at each visit. Efficacy measures assessed by the patient at each visit were disease activity, pain at rest, pain on movement, and functional impairment, using a 4-point verbal rating scale (0 = none, I = mild, 2 = moderate, 3 = severe). Health status as evaluated by the patient and a global assessment of efficacy and tolerability by the physician were recorded using a 4-point verbal rating scale (1 = good, 2 = moderate, 3 = unsatisfactory, and 4 = bad). Adverse drug reactions occurring during the observation period were recorded for each patient. An adverse drug reaction was defined as an unwanted reaction or side effect that occurred in association with the use of a study drug. Reports of adverse reactions were made using the standard reporting form of the Bundesinstitut fur Arzneimittel und Medizinprodukte and the German Physicians’ Drug Commission for recording adverse drug reactions to marketed drugs. Clinical and laboratory tests were performed at the physician’s discretion. Statistical Analysis Baseline characteristics and demographic data were summarized using descriptive statistics and 2-sided t tests for continuous variables and the Fisher exact test for categoric variables. Efficacy end points were analyzed using two l-sided t tests. For patients who withdrew during
F. DEGNER ET AL.
the observation period, the last observation was carried forward. The incidence of adverse events in each treatment group was compared using the Fisher exact test. Additionally, RRs and the related 95% Cls were calculated, as well as a logistic regression model for risk-factor adjustment. Significance was set at 0.05 for all statistical tests. RESULTS After computer randomization of the centers to 1 of 2 treatment groups, 4526 patients were included in the final results: 2530 were treated with meloxicam and 1996 with the comparator NSAlDs. Treat-
ment groups were similar with regard to age, sex, and indication for NSAID therapy (rheumatoid arthritis, osteoarthritis, or other rheumatic conditions such as back pain and ankylosing spondylitis) (Table I). Diclofenac was the most commonly prescribed comparator NSAID (60% of cases), with 49% receiving doses 5100 mg and the rest receiving higher doses (Table II). The second most prescribed NSAID was ibuprofen, for which the majority of prescriptions (72%) were for a daily dose 5 1200 mg. The majority of meloxicam prescriptions (78%) were for a daily dose ~7.5 mg; no evidence of having to increase the initial
Table 1. Demographic and baseline characteristics.
Mean age, y (+ SD) Sex, no. (%) Female Male Diagnosis, no. (%) Rheumatoid arthritis Osteoarthritis Other Disease characteristics, no. (%) History of PUB Previous NSAID use Side effects Not effective Health status’ Pain on movement’ Pain at restl Disease activityi Functional impairmenti NSAIDs = nonsteroidal anti-inflammatory ‘P < 0.001 between groups. ‘Rated bad or unsatisfactory. *Rated moderate or severe. BP -C 0.01 between groups.
Meloxicam (n = 2530)
Comparator NSAIDs (n = 1996)
59* I5
58 + 16
1467 (58) 1063 (42)
10.58(53) 938 (47)
721 (29) 1544 (61) 265 (10)
681 (34) IO49 (53) 266 (13)
312 (12)* 1771 (70)* 474 (19)* 1200 (47)* 960 (38) 1584 (63)* 1000 (40)” 522 (21) 1218 (48;
I36 (7) 1053 (53) 114 (6) 783 (39) 749 (38) I I96 (60) 702 (35) 337 (17) 961 (48)
drugs; PUB = perforation,
ulceration,
or bleeding.
403
CLINICAL
THERAPEUTICS”
Table II. Number (%) of patients receiving low and high dosesof meloxicam and comparator NSAIDs. Treatment
Low Dose*
Meloxicam ComparatorNSAIDs.1 Diclofenac Ibuprofen Piroxicam lndomethacin
196X (78)
NSAlDs
= nonsteroidal
anti-inflammatory
‘Low doses were as follows: meloxicam ~~20 mg/d; indomethacin 5; IO0 mgid. ‘Other unspecified NSAlDs were received
583 216 109 66
562 (22)
(49) (72) (XI) (67)
610 84 26 32
(51) (28) (19) (33)
drugs. ~7.5
mgid;
by 270
diclofenac
~100
mgid;
ibuprofen
5~1200
mgid:
piroxicam
patients.
dose becauseof insufficient efficacy was observed. During the first month of treatment (ie, at visit 2), the prescribeddoseof meloxicam remainedunchangedin 92% of patients; it was increasedin 3% of patients and decreasedin 5%. Similar rates were observedafter the visits at 3 months(93%, 2%, and 5% for unchanged,increased,and decreased doses, respectively) and 6 months (93%, 2%, and 5%, respectively). Concomitant glucocorticosteroid therapy was reported in 9% of both the meloxicam and NSAID groups; smoking in 24% and 26%, respectively; and alcohol consumption in 12% and 15%, respectively. At baseline, treatment groups differed significantly (P < O.OOl),with moremeloxicam patientshaving a history of PUB ( 12% vs 7%); having a history of previousNSAID use(70% vs 53%); having experiencedside effects with previous NSAID use(19% vs 6%); andhaving beennonresponders to previous NSAID therapy (47% vs 39%) (Table I). Disease-relatedcharacteristicsat baseline also differed significantly between groups, with meloxicam patients having more pain on movement (P < 0.001), more 404
High Dose
pain at rest (P < O.Ol), and greater disease activity (P < 0.01) than the comparator NSAID group. Thus, it was concludedthat meloxicam was prescribedpreferentially to patientswho had not respondedto previous NSAID therapy, who had experienced GI sideeffectswith NSAIDs, or who weremore seriouslyimpaired. The proportions of patients in the meloxicam andcomparatorNSAID groups who completed the study-defined as either completing the planned study period of 6 months or being terminated from observation before this time due to successful completion of therapy (ie, no further NSAID treatment deemed necessary)were similar (88.2% and 90.2%, respectively) (Table 111).The between-groupdifference in the dropout rate due to lack of efficacy (P < 0.001) is consistent with baselinedifferences in diseaseactivity and previous responseto NSAIDs (see Table I). A subgroup analysis of nonresponders to previous NSAID therapy showed that 86 (3.4%) patients in the meloxicam group dropped out as early as visit 2 because of lack of therapeutic response,
F. DEGNER ET AL.
Table III.
Patient disposition. Meloxicam (n = 2530)
Completed, no. (%) Discontinued, no. (%) Adverse drug reaction Lack of efficacy Previous NSAID nonresponders/ early dropouts Others Other reasons Data not available, no. (%) NSAlDs = nonsteroidal anti-inflammatory “P < 0.001 versus meloxicam group.
Table IV. Change
187 (9.4) 38 (1.9)
1 I1 (X6)* 24(1.2)* 87 (4.4) 38(1.9) 8 (0.4)
86(3.4) 127 (5.0) 44(1.7) 3 (0.1) drugs.
in efficacy results (mean + SD) from baseline
Final global assessment* Health status* Pain on movement+ Pain at rest+ Disease activity? Functional impairment+ = nonsteroidal a 4-point verbal a 4-point verbal
1801 (90.2)
2231 (88.2) 296(11.7) 39(l.S) 213 (8.4)
Meloxicam
NSAlDs *Using ‘Using
Comparator NSAlDs (n = 1996)
(n = 2530)
Baseline
Last Visit
-
1.39kO.71 I.58 + 0.77 I .02 + 0.8 I 0.65 + 0.76
0.33 f 0.62
2.27 kO.84 2.73 + 0.78 2.26 kO.82
I .8 1 f 0.83 2.44 f 0.85
0.87 f 0.87
anti-inflammatory drugs. rating scale in which I = good, rating scale in which 0 = none,
to the last visit.
Comparator NSAlDs (n = 1996) Mean Change
Baseline
Last Visit
-
-
0.69
1.38 + 0.63 1.66 * 0.75 I .03 + 0.80
1.62
2.25 f 0.81 2.64 + 0.83 2.17 kO.82
1.48 1.57
I .68 f 0.80 2.41 f 0.88
I.71
Mean Change 0.59
0.30 + 0.59
1.61 I .54 I.38
0.88 + 0.86
I.53
0.63 kO.72
2 = moderate, 3 = unsatisfactory, and 4 = bad. I = mild, 2 = moderate, and 3 = severe.
compared with only 24 (1.2%) in the comparator NSAID group (P < 0.001). The dropout rates for lack of efficacy in the subgroup of other patients were similar, with no statistically significant dif-
subgroups(5.0% and 4.8%, respectively). These findings suggestthat differences in the prescribeddoseshad only a minor effect on efficacy in the patient population studied. Overall, despite significant dif-
ference between groups (5.0% and 4.4%, respectively). Specifically, the dropout rates due to lack of efficacy in the meloxicam group were comparable with those in the diclofenac YZIOO-mg and >lOO-mg
ferences
in baseline
disease characteris-
tics and history of responseto previous NSAID therapy, efficacy was similar between groups, with no statistically
signif-
icant differences (Table IV). 405
CLINICAL
THERAPEUTICS”
Table V. Patients experiencing gastrointestinal (GI) adverse reactions. No. of Patients (%) Adverse Reaction Total’ GI system overall+ Abdominal pain Gastritis Nausea Dyspepsia Diarrhea Esophagitis Anorexia Flatulence Duodenitis Other Vomiting GI tract bleeding overall+ Any gastric or duodenal ulcer, hematemesis, or melena (PUB) Gastric ulcer Gastric ulcer, perforated Duodenal ulcer Melena Hemorrhagic gastritis GI hemorrhage
Meloxicam (n = 2530)
NSAIDs (n = 1996)
Relative Risk (9.5% Cl)
P’
63 (2.50) 45 (1.78) 23 (0.91) 2 (0.08) 10 (0.40) 2 (0.08) 8 (0.32) I (0.04) I (0.04) 0 (0.00) 0 (0.00) 3 (0.12) 5 (0.20) 2" (0.08)
71 (3.56) 63 (3.16) 38 (1.90) 12 (0.60) 9(0.45) 7 (0.35) 5 (0.25) 3 (0.15) I (0.05) I (0.05) 1 (0.05) I (0.05) 0 (0.00) 10%(0.50)
0.70 (0.50~.98) 0.56 (0.39-0.82) 0.48 (0.29-0.80) 0.13 (0.03-0.59) 0.88 (0.3f~2.15) 0.23 (0.05-l .08) 1.26(0.41-3.86) 0.26 (0.03-2.53) 0.79 (0.05-12.6) 2.37 (0.25-22.7) 0.16 (0.04-0.72)
0.04 0.003 0.006 0.002 NS 0.049 NS NS NS NS NS NS NS 0.007
0.32 (0.06-1.63)
0.007 NS NS NS NS NS NS
0 0 0 0 0 0 2
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.08)
6 5 I I I I 5
(0.30) (0.25) (0.05) (0.05) (0.05) (0.05) (0.25)
NSAIDs = nonsteroidal anti-inflammatory drugs; PUB = perforation, ulceration, or bleeding. *Represents difference between meloxicam and comparator NSAIDs (Fisher exact test). ‘Discrepancies between the total number of patients experiencing adverse events and the numbers experiencing individual adverse events are due to some individuals having experienced multiple adverse events. iIncluding I serious adverse reaction.
Global tolerancewasreportedasgood or satisfactory in meloxicam patients (2271 [89.8%]) more often than in comparator NSAID patients (1600 [80.2%]) (P < 0.001). The rate of adverse events was higher in the comparator NSAID group (3.56%) than in the meloxicam group (2.50%), with an RR of 0.70 (P = 0.04; 95% CI, 0.50-0.98). More patientshad Gl adverse events in the comparator NSAID 406
group (3.16%) than in the meloxicamgroup (1.80%), with an RR of 0.56 (P = 0.003; 95% CI, 0.39-0.82). Fewer meloxicam patients experienced dyspepsia (P = 0.049; RR, 0.23; 95% CI, 0.05-1.08) abdominal pain (P = 0.006; RR, 0.48; 95% CI, 0.29*.80), gastritis (P = 0.002; RR, 0.13; 95% CI, 0.03-0.59) or GI bleeding (P = 0.007; RR, 0.16; 95% Cl, 0.04-0.72) (Table V). In all, 6 casesof upper CI PUB were
F. DEGNER
ET AL.
Table VI. Effects of risk factors on gastrointestinal
(GI) toxicity. No. (%)
Meloxicam (n = 2530) Daily dose ~7.5 mg (n = 1968) Daily dose >7.5 mg (n = 562) History of PUB (n = 3 12) No history of PUB (n = 2218) Previous NSAID use (n = 1771) No previous NSAID use (n = 759) Side effects of previous NSAID therapy (n = 474) No side effects of previous NSAID therapy (n = 2056) Comparator NSAlDs (n = 1996) Low dose* (n = 974) High dose (n = 752) Not specified (n = 270) History of PUB (n = 136) No history of PUB (n = 1860) Previous NSAID use (n = 1053) No previous NSAID use (n = 943) Side effects of previous NSAID therapy (n = 114) No side effects of previous NSAID therapy (n = 1882) PUB = perforation, ulceration. *Diclofenac or indomethacin
or bleeding; ~~100 mg/d;
NSAID ibuprofen
observed: I with diclofenac (O.l%), I with ibuprofen (0.3%), 1 with indomethacin (1 .O%), and 3 with other NSAIDs (1. I %). We analyzed the effect of potential confounding factors on the observed GI toxicity. In contrast to findings in the comparator NSAID group, no dose effect on GI drug reactions or GI bleeding was observed in the meloxicam group (Table VI). History of PUB, previous NSAID use, and side effects of previous NSAID therapy were associated with higher rates of GI adverse reactions and GI bleeding in the comparator NSAID group compared with the meloxicam group. Using logistic regression to adjust the RR for a GI adverse drug reaction by the influence of side ef-
= nonsteroidal rl200 mgid;
Cl Adverse Reaction
GI Bleeding
45 (1.8) 35 (1.8) 10 (1.8) 4 (1.3) 41 (1.8) 41 (2.3) 4 (0.5) I2 (2.5) 33 (1.6) 63 (3.2) 25 (2.6) 22 (2.9) 16 (5.9) IO (7.4) 53 (2.8) 44 (4.2) 19 (2.0) 9 (7.9) 54 (2.9)
2 (0.08) 2 (0.10) 0 (0.00) 0 (0.00) 2 (0.09) 2 (0.11) 0 (0.00) 0 (0.00) 2 (0.10)
anti-inflammatory piroxicam a20
IO (0.50)
2 (0.21) 5 (0.66) 3 (1.11) 1 (0.74) 9 (0.48) 7 (0.66) 3 (0.32) 1 (0.88) 9 (0.48) drug.
mg.
fects of previous NSAID therapy resulted in a risk estimate for the meloxicam group compared with the comparator NSAID group of 0.49 (95% CI, 0.33-0.74). Incidences of adverse drug reactions in organ systems other than the GI system were similarly low in both groups, with event rates not exceeding 0.5%. Two serious adverse drug reactions were observed. A 75year-old man receiving diclofenac 100 mgld PO developed Gl bleeding from a perforated gastric ulcer, with subsequent hemorrhagic shock, from which he died. A 72-year-old man receiving meloxicam 7.5 mgid PO and concomitant low-dose aspirin ( 100 mgid) developed an upper GI hemorrhage. This 407
CLINICALTHERAPEUTICS'"
was controlled endoscopically with clips. Therapy to eradicate Helicobacter pylori was initiated, consisting of omeprazole, amoxicillin, and clarithromycin. DISCUSSION The results of the present observational study are consistentwith those of doubleblind comparative studies indicating that meloxicam is effective in treating the symptoms of arthritic diseaseand has a better GI tolerability profile than certain NSAIDs. Baseline analysis indicated that meloxicam was prescribed preferentially for patientswho had more seriousdisease or who hadfailed previous NSAID therapy due to GI side effects or lack of response. Theseresultssuggestthat physicians were aware of meloxicam’s improved GI tolerability profile and thus were more likely to prescribe meloxicam rather than classic NSAIDs to patients whose condition was refractory to treatment. Despite this selective prescription to NSAID nonresponders and high-risk patients, meloxicam’s efficacy was comparable to that of other NSAIDs. The higher baseline ratings for pain and diseaseactivity and the greater proportion of NSAID nonrespondersin the group receiving meloxicam are consistent with differencesin the rate of dropoutsdue to lack of efficacy. Overall efficacy at the final visit was similar, with no statistically significant differences between groups. Meloxicam had a more favorable GI tolerability profile than other NSAIDs, including a lower rate of ulceration and GI bleeding. Fewer patients receiving meloxicam reported dyspepsia,abdominal pain, or gastritis. The incidence of GI bleeding (ie, upper GI PUB, hemorrhagic gastritis, or GI hemorrhage)was 0.08% for meloxicam, which was comparable to the 0.1% 408
rate of upper GI PUB (including hematemesisand melena) reported in doubleblind studies.7-9Serious adverse drug reactions were rare, with I case in 1193 patientstreated with diclofenac and 1 case in 2530 patients treated with meloxicam. The former casehad age as a risk factor for serious NSAID-related GI toxicity,’ and the latter had this plus 2 additional risk factors: concomitant low-dose aspirin use and H pylori infection. The overall incidence of adverse drug reactions in the present study was lower than that reported in randomized doubleblind clinical trials.7-9This difference may be explained by differences in the reporting of adverse events and assessmentof the relation between adverse events and study drug in clinical trials comparedwith observational studies. Underreporting of adverse events by physicians may influence the findings of observational studies such as this”; also, the open-label design of the presentstudy does not rule out bias in physicians’ reporting of adverse drug reactions, despite use of a standard reporting form, sinceprescribing physicians may have been familiar with meloxicam’s favorable GI tolerability profile. However, in this regard, meloxicam’s improved GI tolerability also extended to GI bleeding, a serious medical condition that is presumably lesssusceptibleto reporting bias than lessseriousconditions. The dosesof the various NSAIDs differed, which might have influenced rates of adverse drug reactions. In epidemiologic studies,seriousGI toxicity hasbeen reported to vary by a factor of 3 to 10 over the range of recommendeddoses,depending on the NSAID studied.“,‘4 Nevertheless,the differences in GI adverseeffects between meloxicam and other NSAIDs observed in this study were com-
F. DEGNER ET AL.
parable to those reported in double-blind clinical trials.7-9 Postmarketing cohort studies have inherent limitations due to their lack of randomization and blinding.‘O They are considered an important adjunct to clinical trial results and to spontaneousreporting of adverse events outside of formal studies, however, because they capture the therapeutic experience under natural conditions. In this respect, the results of this comparative observational study seemto supportthe adverse-eventprofile of meloxicam in controlled clinical trials.
REFERENCES 1. Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs. N Engl J Med. 1999;340:1888-1899. 2.
nonsteroidal anti-inflammatorydrugtreatment in rheumatoid arthritis. Arch Med. 1996;156:153&1536. 3.
Address correspondence to: Frank Degner, MD, Therapeutic Area CNS/ General Diseases,Boehringer Ingelheim GmbH, Bingerstrasse173, D-55216 Ingelheim, Germany.
Gabriel SE, Jaakkimainen
intern
L, Bombardier
C. Risk for seriousgastrointestinal complicationsrelatedto useof nonsteroidal anti-inflammatorydrugs.Ann kern Med. 1991;115:787-796.
CONCLUSIONS The results of this comparative observational cohort study provide evidence for selective prescribing of meloxicam to patients at high risk for GI side effects of NSAID therapy. Meloxicam was preferentially prescribedto patientswith higher ratings of pain and diseaseactivity and those who had been unresponsive to previous NSAID treatment. Despite these differences in baseline characteristics, patients who received meloxicam experiencedsimilar efficacy with fewer GI drug reactions than did patientsprescribedother NSAIDs. However, the design of the study and the effects of preferential prescribingpreclude definitive conclusions about meloxicam’s efficacy or tolerability. Nevertheless, the resultsof this study are consistentwith the favorable GI tolerability profile of meloxicam seen in double-blind, randomized, comparative clinical trials.
Singh G, Ramey DR, Morfeld D, et al. Gastrointestinal tract complications of
4.
Pairet M, van Ryn J, Schierok H, et al. Differential inhibition of cyclooxygenases-I and -2 by meloxicam and its 4’isomer. lnflamm Res. 1998;47:270-276.
5.
PanaraMR, RendaG, Sciulli MG, et al. Dose-dependent inhibition of platelet cyclooxygenase-I and monocyte cyclooxygenase-2 by meloxicam in healthy subjects. J Phurmacol
6.
Exp
Ther.
1999;290:27iG280.
de Meijer A, Vollaard H, de Metz M, et al. Meloxicam I5 mg/day spares platelet
functionin healthyvolunteers.Clin Ther. 1999;66:425430.
Phar-
macol
Hawkey C, Kahan A, Steinbriick K, et al. Gastrointestinal tolerability of meloxicam compared to diclofenac in osteoarthritis patients. Br J Rheumatol. 1998;37:937-945. Dequeker J, Hawkey C, Kahan A, et al. Improvement in gastrointestinal tolerability of the selective cyclooxygenase COX-2 inhibitor, meloxicam, compared with piroxicam: Results of the Safety and Efficacy
LargeScaleEvaluationof COX Inhibition Therapies(SELECT)trial in osteoarthritis. Br J Rheumatol.
1998;37:94&95
I.
409
CLINICAL
9. Distel M, Mueller C, Bluhmki E, Fries J. Safety of meloxicam: A global analysis of clinical trials. Br J Rheumatol. 199635 (Suppl 1):68-77. 10. Brewer T, Colditz GA. Postmarketing sur veillance and adverse drug reactions. JAMA. 1999: 28 I :824-829. Il.
410
Ottervanger JP, Valkenburg HA, Grobbee DE, Stricker BH. Differences in perceived and presented adverse drug reactions in general practice. J Cfirr Epidenliol. 1998; 5 1:795-799.
THERAPEUTICS’
12. Barner A. Review of clinical trials and benefit/risk ratio of meloxicam. Stand J Rhrumarol. 1996;2S(Suppl 102):29-37. 13. Langman MJS, Weil J, Wainwright P, et al. Risks of bleeding peptic ulcer associated with individual non-steroidal anti-inflammatory drugs. Lmzcet.1994343: 1075-l 078. 14. Garcia Rodriguez LA, Jick H. Risk of upper gastrointestinal bleeding and perforation associated with individual non-steroidal anti-inflammatory drugs. Lancer.l994;343: 769-772.
22, NO. 4,200O
Efficacy and Tolerability of Meloxicam in an Observational, Controlled Cohort Study in Patients with Rheumatic Disease Frank Degner, MD,’ RalfSigmund,2
and Henning Zeidler, MD3
‘Therupeutic Area CNSIGeneral Diseases, Boehringer Ingelheim GmbH, Ingelheitn, 2Depurtment of Medical Datu Services, Boehringer Ingelheirn Phurrna KG, Bibrrach, and ‘Division of Rhewnutology, Depurtment of Internul Medicine, kklicul School Hunnover, Hunnover, Germany
ABSTRACT Background: Meloxicam is a nonsteroidal anti-inflammatory drug (NSAID) that inhibits cyclooxygenase-2 without affecting platelet aggregation. Objective: The goal of this study was to assessthe efficacy and tolerability of meloxicam under natural prescribing conditions for up to 6 months. Methods: This was a multicenter, prospective, observational cohort study. Participating centers were randomized to I of 2 groups: the meloxicam-only group, and the group who received comparator NSAlDs (ie, diclofenac, ibuprofen, piroxicam, or indomethacin). Results: A total of 4526 patients received either meloxicam (n = 2530) or a comparator NSAID (n = 1996). Treatment groups were comparablein terms of observed efficacy measures.Meloxicam patients had significantly lower rates of reported gastrointestinal (GI) adverse drug reactions (I .80% vs 3.20%; P = 0.003), including dyspepsia(0.08% vs 0.35%; P = 0.049), abdominal pain (0.91% vs 1.90%; P = 0.006), gastritis (0.08% vs 0.60%; P = 0.002), and GI bleeding (0.08% vs 0.50%; P = 0.007), compared with those receiving comparator NSAIDs. Conclusion: There was evidence that meloxicam was prescribed preferentially to patients who had not responded to NSAIDs previously, who had previously experienced NSAID-induced side effects (and so were at high risk for developing NSAID-induced GI toxicity), or who were more seriously impaired. The nonrandomized and unblinded nature of this study limits the conclusionsthat can be drawn concerning efficacy or tolerability. Nevertheless,the study results are consistentwith the favorable GI tolerability seen with meloxicam in double-blind comparative trials. Key words: NSAID, COX-2 inhibition, cohort study, meloxicam, gastrointestinal adverse effects. (Clin Ther. 2000;22:400410) Accepfed for publication Februay Printed in the USA. Reproduction
7, 2000. in whole
or part
is not permitted.
F. DEGNER ET AL.
INTRODUCTION Although nonsteroidal anti-inflammatory drugs (NSAIDs) are integral to arthritis therapy, their use has been limited by gastrointestinal (GI) side effects.’ Singh et al2 found that 15% of rheumatoid arthritis patients treated with classic NSAIDs reported Cl side effects over an observation period of 2.5 years; other investigators have reported that NSAID users have an -3 times greater relative risk (RR) for serious GI events than those who do not use NSAIDs.” Thus, it is important to develop treatment strategies that minimize the GI side effects of these valuable medications. Meloxicam, an NSAID of the enolic acid class, inhibits the cyclooxygenase-2 (COX-2) isozyme more potently than the COX-I isozyme4x5 and lacks an inhibitory effect on platelet aggregation at therapeutic doses.6 Introduced in Europe in 1996, meloxicam is available in >90 countries worldwide. In the MELISSA (Meloxicam Large-Scale International Study Safety Assessment) triak7 which involved 9323 patients with osteoarthritis, meloxicam exhibited comparable efficacy and a significantly improved Gl tolerability profile compared with diclofenac, a drug with relatively low GI toxicity. Similar results were seen in the SELECT (Safety and Efficacy Large Scale Evaluation of COX Inhibition Therapies) ttiaLx in which meloxicam was associated with a significantly lower incidence of GI side effects compared with piroxicam. These results were also confirmed in a pooled analysis of all double-blind trials comparing meloxicam with diclofenac, piroxicam, and naproxen.’ However, comparisons with other NSAIDs have not been reported. The efficacy and tolerability of meloxicam have been demonstrated in doubleblind, randomized trials, but to date no in-
formation has been published regarding its use in actual clinical practice. The limitations of observational open-label studies are well known,” particularly with regard to the underreporting of side effects.” At the same time, such studies can complement the results of blinded trials by providing useful information about the efficacy and tolerability of medications under natural prescribing conditions. Also, underreporting can be controlled to some degree by reference to a control group. We therefore undertook this observational cohort study to assess the efficacy and tolerability of meloxicam compared with those of other commonly used NSAIDs. PATIENTS
AND METHODS
Study Design and Protocol This was a prospective, open-label, parallel-group, observational cohort study performed in Germany from August 1996 through July 1997. A total of lo,5 15 physicians specializing in family medicine or internal medicine who had not received sales presentations concerning meloxicam were recruited for the study; 683 agreed to participate and were randomized to 1 of 2 observation groups (group 1 = meloxicam only and group 2 = comparator NSAlDs) without their knowledge. Physicians were free to prescribe any NSAID they deemed appropriate, but the group I analysis included only efficacy and tolerability data from patients prescribed meloxicam, whereas the group 2 analysis included only data from patients prescribed the comparator NSAIDs. This approach was intended to control for a potential prescribing bias caused by the launch of meloxicam, which had been shown in double-blind clinical trials”%‘2 to have a favorable GI tolerability 401
CLINICAL
profile over classic NSAIDs. Physicians received a study protocol reflecting their randomization status. Patients were observed in the prescribing setting in a manner unaffected by the randomization procedure applied to the participating centers. To be eligible for the trial, patients had to be diagnosed with acute or chronic active rheumatic disease for which NSAID therapy was required. Patients were also eligible if they had been previously diagnosed with rheumatic disease and required a switch to another NSAID because of adverse reactions or lack of efficacy. They were excluded if they had any contraindication listed in the product information for the prescribed NSAID. The selection of an appropriate NSAID for each patient (either meloxicam or another NSAID) was left to the individual physicians. Concomitant therapy, including steroids, aspirin, anticoagulants, other NSAlDs, and acetaminophen, was also allowed at the physician’s discretion. Baseline comparability of treatment groups was considered important for assessment of efficacy and tolerability. Demographic and medical data were obtained at baseline regarding age; sex; indication for NSAID therapy; history of perforation, ulceration, or Cl bleeding (PUB); history of previous NSAID use with or without associated side effects; and lack of efficacy of previous NSAID therapy. According to guidelines issued by the European Agency for the Evaluation of Medicinal Products, this observational pharmacoepidemiologic study was not required to obtain patients’ informed consent. The study was performed in accordance with German drug law and European Union pharmacovigilance guidelines. Patients were observed for up to 6 months to assess the efficacy and tolera402
THERAPEIJTICS’
bility of the treatment regimens. After the initial visit at which patients received a prescription for 1 of the study drugs, 3 follow-up visits were scheduled at 1 month, between months I and 3, and between months 3 and 6. Additional visits could be scheduled if medically indicated. A physical examination was performed at each visit. Efficacy measures assessed by the patient at each visit were disease activity, pain at rest, pain on movement, and functional impairment, using a 4-point verbal rating scale (0 = none, I = mild, 2 = moderate, 3 = severe). Health status as evaluated by the patient and a global assessment of efficacy and tolerability by the physician were recorded using a 4-point verbal rating scale (1 = good, 2 = moderate, 3 = unsatisfactory, and 4 = bad). Adverse drug reactions occurring during the observation period were recorded for each patient. An adverse drug reaction was defined as an unwanted reaction or side effect that occurred in association with the use of a study drug. Reports of adverse reactions were made using the standard reporting form of the Bundesinstitut fur Arzneimittel und Medizinprodukte and the German Physicians’ Drug Commission for recording adverse drug reactions to marketed drugs. Clinical and laboratory tests were performed at the physician’s discretion. Statistical Analysis Baseline characteristics and demographic data were summarized using descriptive statistics and 2-sided t tests for continuous variables and the Fisher exact test for categoric variables. Efficacy end points were analyzed using two l-sided t tests. For patients who withdrew during
F. DEGNER ET AL.
the observation period, the last observation was carried forward. The incidence of adverse events in each treatment group was compared using the Fisher exact test. Additionally, RRs and the related 95% Cls were calculated, as well as a logistic regression model for risk-factor adjustment. Significance was set at 0.05 for all statistical tests. RESULTS After computer randomization of the centers to 1 of 2 treatment groups, 4526 patients were included in the final results: 2530 were treated with meloxicam and 1996 with the comparator NSAlDs. Treat-
ment groups were similar with regard to age, sex, and indication for NSAID therapy (rheumatoid arthritis, osteoarthritis, or other rheumatic conditions such as back pain and ankylosing spondylitis) (Table I). Diclofenac was the most commonly prescribed comparator NSAID (60% of cases), with 49% receiving doses 5100 mg and the rest receiving higher doses (Table II). The second most prescribed NSAID was ibuprofen, for which the majority of prescriptions (72%) were for a daily dose 5 1200 mg. The majority of meloxicam prescriptions (78%) were for a daily dose ~7.5 mg; no evidence of having to increase the initial
Table 1. Demographic and baseline characteristics.
Mean age, y (+ SD) Sex, no. (%) Female Male Diagnosis, no. (%) Rheumatoid arthritis Osteoarthritis Other Disease characteristics, no. (%) History of PUB Previous NSAID use Side effects Not effective Health status’ Pain on movement’ Pain at restl Disease activityi Functional impairmenti NSAIDs = nonsteroidal anti-inflammatory ‘P < 0.001 between groups. ‘Rated bad or unsatisfactory. *Rated moderate or severe. BP -C 0.01 between groups.
Meloxicam (n = 2530)
Comparator NSAIDs (n = 1996)
59* I5
58 + 16
1467 (58) 1063 (42)
10.58(53) 938 (47)
721 (29) 1544 (61) 265 (10)
681 (34) IO49 (53) 266 (13)
312 (12)* 1771 (70)* 474 (19)* 1200 (47)* 960 (38) 1584 (63)* 1000 (40)” 522 (21) 1218 (48;
I36 (7) 1053 (53) 114 (6) 783 (39) 749 (38) I I96 (60) 702 (35) 337 (17) 961 (48)
drugs; PUB = perforation,
ulceration,
or bleeding.
403
CLINICAL
THERAPEUTICS”
Table II. Number (%) of patients receiving low and high dosesof meloxicam and comparator NSAIDs. Treatment
Low Dose*
Meloxicam ComparatorNSAIDs.1 Diclofenac Ibuprofen Piroxicam lndomethacin
196X (78)
NSAlDs
= nonsteroidal
anti-inflammatory
‘Low doses were as follows: meloxicam ~~20 mg/d; indomethacin 5; IO0 mgid. ‘Other unspecified NSAlDs were received
583 216 109 66
562 (22)
(49) (72) (XI) (67)
610 84 26 32
(51) (28) (19) (33)
drugs. ~7.5
mgid;
by 270
diclofenac
~100
mgid;
ibuprofen
5~1200
mgid:
piroxicam
patients.
dose becauseof insufficient efficacy was observed. During the first month of treatment (ie, at visit 2), the prescribeddoseof meloxicam remainedunchangedin 92% of patients; it was increasedin 3% of patients and decreasedin 5%. Similar rates were observedafter the visits at 3 months(93%, 2%, and 5% for unchanged,increased,and decreased doses, respectively) and 6 months (93%, 2%, and 5%, respectively). Concomitant glucocorticosteroid therapy was reported in 9% of both the meloxicam and NSAID groups; smoking in 24% and 26%, respectively; and alcohol consumption in 12% and 15%, respectively. At baseline, treatment groups differed significantly (P < O.OOl),with moremeloxicam patientshaving a history of PUB ( 12% vs 7%); having a history of previousNSAID use(70% vs 53%); having experiencedside effects with previous NSAID use(19% vs 6%); andhaving beennonresponders to previous NSAID therapy (47% vs 39%) (Table I). Disease-relatedcharacteristicsat baseline also differed significantly between groups, with meloxicam patients having more pain on movement (P < 0.001), more 404
High Dose
pain at rest (P < O.Ol), and greater disease activity (P < 0.01) than the comparator NSAID group. Thus, it was concludedthat meloxicam was prescribedpreferentially to patientswho had not respondedto previous NSAID therapy, who had experienced GI sideeffectswith NSAIDs, or who weremore seriouslyimpaired. The proportions of patients in the meloxicam andcomparatorNSAID groups who completed the study-defined as either completing the planned study period of 6 months or being terminated from observation before this time due to successful completion of therapy (ie, no further NSAID treatment deemed necessary)were similar (88.2% and 90.2%, respectively) (Table 111).The between-groupdifference in the dropout rate due to lack of efficacy (P < 0.001) is consistent with baselinedifferences in diseaseactivity and previous responseto NSAIDs (see Table I). A subgroup analysis of nonresponders to previous NSAID therapy showed that 86 (3.4%) patients in the meloxicam group dropped out as early as visit 2 because of lack of therapeutic response,
F. DEGNER ET AL.
Table III.
Patient disposition. Meloxicam (n = 2530)
Completed, no. (%) Discontinued, no. (%) Adverse drug reaction Lack of efficacy Previous NSAID nonresponders/ early dropouts Others Other reasons Data not available, no. (%) NSAlDs = nonsteroidal anti-inflammatory “P < 0.001 versus meloxicam group.
Table IV. Change
187 (9.4) 38 (1.9)
1 I1 (X6)* 24(1.2)* 87 (4.4) 38(1.9) 8 (0.4)
86(3.4) 127 (5.0) 44(1.7) 3 (0.1) drugs.
in efficacy results (mean + SD) from baseline
Final global assessment* Health status* Pain on movement+ Pain at rest+ Disease activity? Functional impairment+ = nonsteroidal a 4-point verbal a 4-point verbal
1801 (90.2)
2231 (88.2) 296(11.7) 39(l.S) 213 (8.4)
Meloxicam
NSAlDs *Using ‘Using
Comparator NSAlDs (n = 1996)
(n = 2530)
Baseline
Last Visit
-
1.39kO.71 I.58 + 0.77 I .02 + 0.8 I 0.65 + 0.76
0.33 f 0.62
2.27 kO.84 2.73 + 0.78 2.26 kO.82
I .8 1 f 0.83 2.44 f 0.85
0.87 f 0.87
anti-inflammatory drugs. rating scale in which I = good, rating scale in which 0 = none,
to the last visit.
Comparator NSAlDs (n = 1996) Mean Change
Baseline
Last Visit
-
-
0.69
1.38 + 0.63 1.66 * 0.75 I .03 + 0.80
1.62
2.25 f 0.81 2.64 + 0.83 2.17 kO.82
1.48 1.57
I .68 f 0.80 2.41 f 0.88
I.71
Mean Change 0.59
0.30 + 0.59
1.61 I .54 I.38
0.88 + 0.86
I.53
0.63 kO.72
2 = moderate, 3 = unsatisfactory, and 4 = bad. I = mild, 2 = moderate, and 3 = severe.
compared with only 24 (1.2%) in the comparator NSAID group (P < 0.001). The dropout rates for lack of efficacy in the subgroup of other patients were similar, with no statistically significant dif-
subgroups(5.0% and 4.8%, respectively). These findings suggestthat differences in the prescribeddoseshad only a minor effect on efficacy in the patient population studied. Overall, despite significant dif-
ference between groups (5.0% and 4.4%, respectively). Specifically, the dropout rates due to lack of efficacy in the meloxicam group were comparable with those in the diclofenac YZIOO-mg and >lOO-mg
ferences
in baseline
disease characteris-
tics and history of responseto previous NSAID therapy, efficacy was similar between groups, with no statistically
signif-
icant differences (Table IV). 405
CLINICAL
THERAPEUTICS”
Table V. Patients experiencing gastrointestinal (GI) adverse reactions. No. of Patients (%) Adverse Reaction Total’ GI system overall+ Abdominal pain Gastritis Nausea Dyspepsia Diarrhea Esophagitis Anorexia Flatulence Duodenitis Other Vomiting GI tract bleeding overall+ Any gastric or duodenal ulcer, hematemesis, or melena (PUB) Gastric ulcer Gastric ulcer, perforated Duodenal ulcer Melena Hemorrhagic gastritis GI hemorrhage
Meloxicam (n = 2530)
NSAIDs (n = 1996)
Relative Risk (9.5% Cl)
P’
63 (2.50) 45 (1.78) 23 (0.91) 2 (0.08) 10 (0.40) 2 (0.08) 8 (0.32) I (0.04) I (0.04) 0 (0.00) 0 (0.00) 3 (0.12) 5 (0.20) 2" (0.08)
71 (3.56) 63 (3.16) 38 (1.90) 12 (0.60) 9(0.45) 7 (0.35) 5 (0.25) 3 (0.15) I (0.05) I (0.05) 1 (0.05) I (0.05) 0 (0.00) 10%(0.50)
0.70 (0.50~.98) 0.56 (0.39-0.82) 0.48 (0.29-0.80) 0.13 (0.03-0.59) 0.88 (0.3f~2.15) 0.23 (0.05-l .08) 1.26(0.41-3.86) 0.26 (0.03-2.53) 0.79 (0.05-12.6) 2.37 (0.25-22.7) 0.16 (0.04-0.72)
0.04 0.003 0.006 0.002 NS 0.049 NS NS NS NS NS NS NS 0.007
0.32 (0.06-1.63)
0.007 NS NS NS NS NS NS
0 0 0 0 0 0 2
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.08)
6 5 I I I I 5
(0.30) (0.25) (0.05) (0.05) (0.05) (0.05) (0.25)
NSAIDs = nonsteroidal anti-inflammatory drugs; PUB = perforation, ulceration, or bleeding. *Represents difference between meloxicam and comparator NSAIDs (Fisher exact test). ‘Discrepancies between the total number of patients experiencing adverse events and the numbers experiencing individual adverse events are due to some individuals having experienced multiple adverse events. iIncluding I serious adverse reaction.
Global tolerancewasreportedasgood or satisfactory in meloxicam patients (2271 [89.8%]) more often than in comparator NSAID patients (1600 [80.2%]) (P < 0.001). The rate of adverse events was higher in the comparator NSAID group (3.56%) than in the meloxicam group (2.50%), with an RR of 0.70 (P = 0.04; 95% CI, 0.50-0.98). More patientshad Gl adverse events in the comparator NSAID 406
group (3.16%) than in the meloxicamgroup (1.80%), with an RR of 0.56 (P = 0.003; 95% CI, 0.39-0.82). Fewer meloxicam patients experienced dyspepsia (P = 0.049; RR, 0.23; 95% CI, 0.05-1.08) abdominal pain (P = 0.006; RR, 0.48; 95% CI, 0.29*.80), gastritis (P = 0.002; RR, 0.13; 95% CI, 0.03-0.59) or GI bleeding (P = 0.007; RR, 0.16; 95% Cl, 0.04-0.72) (Table V). In all, 6 casesof upper CI PUB were
F. DEGNER
ET AL.
Table VI. Effects of risk factors on gastrointestinal
(GI) toxicity. No. (%)
Meloxicam (n = 2530) Daily dose ~7.5 mg (n = 1968) Daily dose >7.5 mg (n = 562) History of PUB (n = 3 12) No history of PUB (n = 2218) Previous NSAID use (n = 1771) No previous NSAID use (n = 759) Side effects of previous NSAID therapy (n = 474) No side effects of previous NSAID therapy (n = 2056) Comparator NSAlDs (n = 1996) Low dose* (n = 974) High dose (n = 752) Not specified (n = 270) History of PUB (n = 136) No history of PUB (n = 1860) Previous NSAID use (n = 1053) No previous NSAID use (n = 943) Side effects of previous NSAID therapy (n = 114) No side effects of previous NSAID therapy (n = 1882) PUB = perforation, ulceration. *Diclofenac or indomethacin
or bleeding; ~~100 mg/d;
NSAID ibuprofen
observed: I with diclofenac (O.l%), I with ibuprofen (0.3%), 1 with indomethacin (1 .O%), and 3 with other NSAIDs (1. I %). We analyzed the effect of potential confounding factors on the observed GI toxicity. In contrast to findings in the comparator NSAID group, no dose effect on GI drug reactions or GI bleeding was observed in the meloxicam group (Table VI). History of PUB, previous NSAID use, and side effects of previous NSAID therapy were associated with higher rates of GI adverse reactions and GI bleeding in the comparator NSAID group compared with the meloxicam group. Using logistic regression to adjust the RR for a GI adverse drug reaction by the influence of side ef-
= nonsteroidal rl200 mgid;
Cl Adverse Reaction
GI Bleeding
45 (1.8) 35 (1.8) 10 (1.8) 4 (1.3) 41 (1.8) 41 (2.3) 4 (0.5) I2 (2.5) 33 (1.6) 63 (3.2) 25 (2.6) 22 (2.9) 16 (5.9) IO (7.4) 53 (2.8) 44 (4.2) 19 (2.0) 9 (7.9) 54 (2.9)
2 (0.08) 2 (0.10) 0 (0.00) 0 (0.00) 2 (0.09) 2 (0.11) 0 (0.00) 0 (0.00) 2 (0.10)
anti-inflammatory piroxicam a20
IO (0.50)
2 (0.21) 5 (0.66) 3 (1.11) 1 (0.74) 9 (0.48) 7 (0.66) 3 (0.32) 1 (0.88) 9 (0.48) drug.
mg.
fects of previous NSAID therapy resulted in a risk estimate for the meloxicam group compared with the comparator NSAID group of 0.49 (95% CI, 0.33-0.74). Incidences of adverse drug reactions in organ systems other than the GI system were similarly low in both groups, with event rates not exceeding 0.5%. Two serious adverse drug reactions were observed. A 75year-old man receiving diclofenac 100 mgld PO developed Gl bleeding from a perforated gastric ulcer, with subsequent hemorrhagic shock, from which he died. A 72-year-old man receiving meloxicam 7.5 mgid PO and concomitant low-dose aspirin ( 100 mgid) developed an upper GI hemorrhage. This 407
CLINICALTHERAPEUTICS'"
was controlled endoscopically with clips. Therapy to eradicate Helicobacter pylori was initiated, consisting of omeprazole, amoxicillin, and clarithromycin. DISCUSSION The results of the present observational study are consistentwith those of doubleblind comparative studies indicating that meloxicam is effective in treating the symptoms of arthritic diseaseand has a better GI tolerability profile than certain NSAIDs. Baseline analysis indicated that meloxicam was prescribed preferentially for patientswho had more seriousdisease or who hadfailed previous NSAID therapy due to GI side effects or lack of response. Theseresultssuggestthat physicians were aware of meloxicam’s improved GI tolerability profile and thus were more likely to prescribe meloxicam rather than classic NSAIDs to patients whose condition was refractory to treatment. Despite this selective prescription to NSAID nonresponders and high-risk patients, meloxicam’s efficacy was comparable to that of other NSAIDs. The higher baseline ratings for pain and diseaseactivity and the greater proportion of NSAID nonrespondersin the group receiving meloxicam are consistent with differencesin the rate of dropoutsdue to lack of efficacy. Overall efficacy at the final visit was similar, with no statistically significant differences between groups. Meloxicam had a more favorable GI tolerability profile than other NSAIDs, including a lower rate of ulceration and GI bleeding. Fewer patients receiving meloxicam reported dyspepsia,abdominal pain, or gastritis. The incidence of GI bleeding (ie, upper GI PUB, hemorrhagic gastritis, or GI hemorrhage)was 0.08% for meloxicam, which was comparable to the 0.1% 408
rate of upper GI PUB (including hematemesisand melena) reported in doubleblind studies.7-9Serious adverse drug reactions were rare, with I case in 1193 patientstreated with diclofenac and 1 case in 2530 patients treated with meloxicam. The former casehad age as a risk factor for serious NSAID-related GI toxicity,’ and the latter had this plus 2 additional risk factors: concomitant low-dose aspirin use and H pylori infection. The overall incidence of adverse drug reactions in the present study was lower than that reported in randomized doubleblind clinical trials.7-9This difference may be explained by differences in the reporting of adverse events and assessmentof the relation between adverse events and study drug in clinical trials comparedwith observational studies. Underreporting of adverse events by physicians may influence the findings of observational studies such as this”; also, the open-label design of the presentstudy does not rule out bias in physicians’ reporting of adverse drug reactions, despite use of a standard reporting form, sinceprescribing physicians may have been familiar with meloxicam’s favorable GI tolerability profile. However, in this regard, meloxicam’s improved GI tolerability also extended to GI bleeding, a serious medical condition that is presumably lesssusceptibleto reporting bias than lessseriousconditions. The dosesof the various NSAIDs differed, which might have influenced rates of adverse drug reactions. In epidemiologic studies,seriousGI toxicity hasbeen reported to vary by a factor of 3 to 10 over the range of recommendeddoses,depending on the NSAID studied.“,‘4 Nevertheless,the differences in GI adverseeffects between meloxicam and other NSAIDs observed in this study were com-
F. DEGNER ET AL.
parable to those reported in double-blind clinical trials.7-9 Postmarketing cohort studies have inherent limitations due to their lack of randomization and blinding.‘O They are considered an important adjunct to clinical trial results and to spontaneousreporting of adverse events outside of formal studies, however, because they capture the therapeutic experience under natural conditions. In this respect, the results of this comparative observational study seemto supportthe adverse-eventprofile of meloxicam in controlled clinical trials.
REFERENCES 1. Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs. N Engl J Med. 1999;340:1888-1899. 2.
nonsteroidal anti-inflammatorydrugtreatment in rheumatoid arthritis. Arch Med. 1996;156:153&1536. 3.
Address correspondence to: Frank Degner, MD, Therapeutic Area CNS/ General Diseases,Boehringer Ingelheim GmbH, Bingerstrasse173, D-55216 Ingelheim, Germany.
Gabriel SE, Jaakkimainen
intern
L, Bombardier
C. Risk for seriousgastrointestinal complicationsrelatedto useof nonsteroidal anti-inflammatorydrugs.Ann kern Med. 1991;115:787-796.
CONCLUSIONS The results of this comparative observational cohort study provide evidence for selective prescribing of meloxicam to patients at high risk for GI side effects of NSAID therapy. Meloxicam was preferentially prescribedto patientswith higher ratings of pain and diseaseactivity and those who had been unresponsive to previous NSAID treatment. Despite these differences in baseline characteristics, patients who received meloxicam experiencedsimilar efficacy with fewer GI drug reactions than did patientsprescribedother NSAIDs. However, the design of the study and the effects of preferential prescribingpreclude definitive conclusions about meloxicam’s efficacy or tolerability. Nevertheless, the resultsof this study are consistentwith the favorable GI tolerability profile of meloxicam seen in double-blind, randomized, comparative clinical trials.
Singh G, Ramey DR, Morfeld D, et al. Gastrointestinal tract complications of
4.
Pairet M, van Ryn J, Schierok H, et al. Differential inhibition of cyclooxygenases-I and -2 by meloxicam and its 4’isomer. lnflamm Res. 1998;47:270-276.
5.
PanaraMR, RendaG, Sciulli MG, et al. Dose-dependent inhibition of platelet cyclooxygenase-I and monocyte cyclooxygenase-2 by meloxicam in healthy subjects. J Phurmacol
6.
Exp
Ther.
1999;290:27iG280.
de Meijer A, Vollaard H, de Metz M, et al. Meloxicam I5 mg/day spares platelet
functionin healthyvolunteers.Clin Ther. 1999;66:425430.
Phar-
macol
Hawkey C, Kahan A, Steinbriick K, et al. Gastrointestinal tolerability of meloxicam compared to diclofenac in osteoarthritis patients. Br J Rheumatol. 1998;37:937-945. Dequeker J, Hawkey C, Kahan A, et al. Improvement in gastrointestinal tolerability of the selective cyclooxygenase COX-2 inhibitor, meloxicam, compared with piroxicam: Results of the Safety and Efficacy
LargeScaleEvaluationof COX Inhibition Therapies(SELECT)trial in osteoarthritis. Br J Rheumatol.
1998;37:94&95
I.
409
CLINICAL
9. Distel M, Mueller C, Bluhmki E, Fries J. Safety of meloxicam: A global analysis of clinical trials. Br J Rheumatol. 199635 (Suppl 1):68-77. 10. Brewer T, Colditz GA. Postmarketing sur veillance and adverse drug reactions. JAMA. 1999: 28 I :824-829. Il.
410
Ottervanger JP, Valkenburg HA, Grobbee DE, Stricker BH. Differences in perceived and presented adverse drug reactions in general practice. J Cfirr Epidenliol. 1998; 5 1:795-799.
THERAPEUTICS’
12. Barner A. Review of clinical trials and benefit/risk ratio of meloxicam. Stand J Rhrumarol. 1996;2S(Suppl 102):29-37. 13. Langman MJS, Weil J, Wainwright P, et al. Risks of bleeding peptic ulcer associated with individual non-steroidal anti-inflammatory drugs. Lmzcet.1994343: 1075-l 078. 14. Garcia Rodriguez LA, Jick H. Risk of upper gastrointestinal bleeding and perforation associated with individual non-steroidal anti-inflammatory drugs. Lancer.l994;343: 769-772.