Response to Nonsteroidal Anti-inflammatory Drugs in African Americans with Osteoarthritis of the Knee a

Abstract

Objective:

This 6-week, randomized, double-blind, parallel-group study compared the analgesic efficacy, tolerability and safety of celecoxib, naproxen and placebo in African Americans with osteoarthritis (OA) of the knee.

Methods:

A total of 322 patients aged ≥ 45 years with OA of the knee in a flare state received 200 mg celecoxib orally once daily, 500 mg naproxen orally twice daily or placebo for 6 weeks. The primary endpoint was change from baseline in the Patient's Assessment of Arthritis Pain.

Results:

Celecoxib was as effective as naproxen in reducing OA pain. Similar efficacy was observed in many of the secondary outcome measures. Celecoxib was well tolerated and demonstrated favorable upper gastro-intestinal tolerability. Improvements in outcome measures were numerically greater in the active treatment groups compared with the placebo group, but did not reach statistical significance.

Conclusions:

Celecoxib was as effective as naproxen in relieving OA pain in African Americans and was well tolerated. Few significant differences were observed between active treatments and placebo, possibly because of a strong placebo effect.

Keywords

Celecoxib Cyclo-oxygenase-2 Ethnicity Knee Nonsteroidal anti-inflammatory drugs Osteoarthritis Pain Race

Introduction

Osteoarthritis (OA) is the most prevalent form of arthritis, and is a leading cause of pain and disability among adults and the elderly. This complex clinical condition affects more than 26 million people in the USA who are over the age of 25 years, and accounts for more than 25% of all arthritis-related healthcare visits.^1,2 It is estimated that 80% of people aged over 75 will have radiographic evidence of OA, although only 60% of them are likely to be symptomatic.³ As the US population ages the prevalence of OA is expected to increase, placing a considerable burden – already estimated to exceed $60 billion annually^1,4 – on healthcare resources.

Risk factors for the incidence and progression of OA vary considerably according to the type of joint affected; commonly recognized risk factors include age, sex, hormonal status, bone density, genetics, nutrition and obesity.^3,5 There is increasing evidence to suggest that ethnicity and race may also play important roles in the prevalence and variability of OA, as well as of other arthritic conditions.^3,5,6 Recent data suggest that African American people have a higher burden of multiple, large-joint OA and are more likely to have knee OA than white people.⁷ Data from the National Health Interview Survey indicate a higher prevalence of physician-diagnosed arthritis among non-Hispanic white than non-Hispanic black people; however, non-Hispanic black people have significantly higher age-adjusted activity limitation, work limitation and severe joint pain than non-Hispanic white people.⁸ Specific to OA, there is some evidence to suggest that the prevalence of OA in black people is lower than in white people.⁷

The management of OA is focused on the control of pain, joint stability and improvement in function and health-related quality of life. Modern approaches to treatment aim to slow disease progression, e.g. through the use of exercise, weight loss (if overweight/obese) and self-management programs.^5,9 However, pain is the most salient symptom for many patients with OA, and typical first-line treatment is the use of acetaminophen (paracetamol) and nonselective nonsteroidal anti-inflammatory drugs (nsNSAIDs) or cyclo-oxygenase-2 (COX-2)-selective NSAIDs. Celecoxib, a COX-2-selective NSAID, has established efficacy in patients with OA^{10 – 12} and has been shown to have a favorable gastrointestinal (GI) tolerability profile compared with nsNSAIDs, with fewer patients receiving celecoxib experiencing GI adverse events (AEs) such as dyspepsia.^{13 – 17} Despite being commonly prescribed for the relief of OA symptoms, all NSAIDs are associated with an increased risk of serious GI-related AEs such as gastric ulceration, perforation and bleeding, and all carry a warning in the USA prescribing information about this GI risk, as well as about cardiovascular risk factors.^18,19

There has been growing interest in the differing racial and ethnic experiences of clinical pain, and a number of studies have indicated that the experience varies considerably among different populations.^{20 – 23} For example, differences in pain intensity, higher rates of chronic pain and lower pain thresholds have been reported in African American people compared with non-Hispanic white people.²⁴ Disparities in the experience and tolerability of pain among ethnic and racial groups clearly have important implications for the management of pain control. Undertreatment of pain in minority populations has been reported,^{24 – 26} and OA treatments may need to be tailored to suit different racial and ethnic populations.

Despite the apparent variation in the prevalence and severity of OA among different racial and ethnic groups, and differences in their experiences of arthritis-related pain, nonwhite patient groups remain substantially under-represented in clinical trials.^{27 – 29} Studies suggest that people from racial/ethnic minorities, especially in the USA, are less willing to participate in clinical trials than other people, because of mistrust of the medical system and perceptions of racial discrimination.^28,29 Other factors include concerns regarding treatment risk, language barriers and costs associated with interpretation/translation, a shortage of investigators from different ethnic/racial groups, and sociocultural barriers.²⁷ In a review of health research studies conducted in the USA and Europe, Wendler et al.²⁹ identified significant differences in the number of individuals from minority populations being invited to participate in trials. Whatever the cause, under-representation of racial/ethnic groups within clinical trials has the potential to compromise the generalizability and external validity of trial findings, and may have important repercussions regarding the safety and efficacy of treatments in minority groups. Many payers and hospitals who serve substantial racial/ethnic minority populations often request data from clinical trials that mirror the racial/ethnic diversity of their beneficiaries, when making formulary decisions on new drugs.

This study evaluated the analgesic efficacy, safety and GI tolerability of celecoxib and the nsNSAID naproxen in patients with OA from an African American population.

Patients and methods

Study Population

African American patients aged ≥ 45 years, with OA of the knee (diagnosed according to American College of Rheumatology guidelines³⁰) in a flare state, and with a functional capacity classification of I – III were eligible for study participation. For patients receiving NSAID or analgesic therapy, a flare was demonstrated if the Physician's Global Assessment of Arthritis and the Patient's Global Assessment of Arthritis were both ‘fair', ‘poor' or ‘very poor' at the baseline visit, and if the baseline Patient's Assessment of Arthritis Pain visual analog scale (VAS) measurement was between 40 and 90 mm (out of 100 mm; 0 representing no pain and 100 representing very severe pain), the Patient's Global Assessment of Arthritis showed an increase of one or more grades and the Physician's Global Assessment of Arthritis showed an increase of one or more grades. For patients who were not receiving treatment, a flare was defined if the Patient's Assessment of Arthritis Pain VAS was between 40 and 90 mm, the Patient's and Physician's Global Assessment of Arthritis was ‘poor' or ‘very poor', and the Global Assessment of Arthritis was ‘poor' or ‘very poor'. Physicians classified patients with a functional capacity classification ranging from I (complete functional capacity with ability to carry on all usual duties without handicaps) to IV (largely or wholly incapacitated with the subject bedridden or confined to wheelchair, permitting little or no self-care). Subjects with a functional capacity classification of IV were not eligible for participation. Patients receiving NSAID or analgesic therapy were required to discontinue medication ≥ 48 h before the baseline assessments.

Patients were excluded from participation if they had any of the following: inflammatory arthritis (except those with gout or pseudo-gout who had not experienced an acute flare in the past 2 years, and those with regional pain syndrome or fibromyalgia); acute joint trauma with active symptoms within 3 months prior to the start of the study; oral corticosteroid therapy within 4 weeks, intramuscular or soft-tissue injections of corticosteroids within 2 months, intra-articular injections of corticosteroids within 3 months, intra-articular injections of hyaluronic acid in the index joint within 9 months before the first dose of study medication or acetaminophen within 24 h of the baseline visit; a diagnosis of esophageal, gastric, pyloric channel or duodenal ulceration within 60 days before the first dose of study medication; unstable cardiovascular disease within 6 months before screening; history of GI perforations, obstructions or bleeding; active GI, renal, hepatic or coagulation disorders; malignancy (unless surgically removed and no recurrence within 5 years); known sensitivity to NSAIDs, sulfonamides, aspirin or related compounds.

Study Design

This 6-week randomized, double-blind, placebo-controlled, active-comparator, parallel-group study was conducted at 28 centers in the USA between 31 October 2001 and 20 November 2002, in accordance with the principles of Good Clinical Practice³¹ and the Declaration of Helsinki.³² The protocol was approved by the institutional review board at each study site, and all patients provided written informed consent prior to participation in the study.

There were four clinic visits: screening; baseline; week 2; week 6. Before enrollment, each patient provided a complete medical and medication history. At the screening visit (which occurred 1 – 14 days before the first dose of study medication) patients had a physical examination, underwent clinical laboratory testing (hematology, bio -chemistry and urinalysis) and, if female and of childbearing potential, received a urine pregnancy test. Evaluation of arthritis, consisting of both patient and physician assessments, was performed, and patients completed a Patient Health Questionnaire (PHQ-9; used to evaluate syndrome/suicidal ideation [available online at http://www.depression-primarycare.org/clinicians/toolkits/materials/forms/phq9/]), a Pain Satisfaction Scale and a Complementary and Alternative Medicines Questionnaire (to capture use of alternatives to conventional medication). Patients were instructed to discontinue the use of any NSAID and/or analgesic therapy. Acetaminophen (up to 2 g/day) was permitted for rescue analgesia during the screening period but must have been discontinued ≥ 24 h before baseline assessments.

Eligible patients completed baseline arthritis assessments, consisting of Patient's and Physician's Global Assessment of Arthritis, Patient's Assessment of Arthritis Pain (0 – 100 mm VAS), Functional Capacity Classification and Western Ontario McMasters (WOMAC) OA index, and underwent verification of flare criteria. Flare criteria were verified if the patient met the prespecified flare criteria described above. For the Patient's Global Assessment of Arthritis, patients rated their condition based on a scale from 1 to 5, score 1 representing ‘very good' (asymptomatic and no limitation of normal activities) and score 5 ‘very poor' (very severe symptoms that are intolerable and inability to carry out normal activities) in answer to the question: ‘Considering all the ways the OA in your knee affects you, how are you doing today?'. The Physician's Global Assessment of Arthritis was the physician's assessment of how the OA of the knee appeared at the time of the patient's visit and was based on the patient's disease signs. It was recorded on a five-point scale, 1 representing ‘very good' (asymptomatic and no limitation of normal activities) and 5 representing ‘very poor' (very severe symptoms that are intolerable and inability to carry out all normal activities). The WOMAC scale includes three subscales that are combined to generate a total score: pain (five questions, each response ranging from 0 to 4); stiffness (two questions, each response ranging from 0 to 4); physical function (17 questions, each response ranging from 0 to 4). A reduction in score indicates an improvement.

Patients received an American Pain Society (APS) pain-measure diary. The APS pain measure consisted of five questions; the first question required a yes/no response and for the remaining questions patients were required to rate their pain and its interference with daily activities on a scale of 0 (no pain) to 10 (worst pain possible). Patients completed diary entries in the evening for 7 days during week 1.

Patients were randomized in a 2 : 2 : 1 ratio to one of three treatments: 200 mg celecoxib orally once daily, 500 mg naproxen orally twice daily or placebo. The treatment regimen was assigned using a predetermined computer-generated randomization schedule. Patients were instructed to take two capsules daily with the morning meal (celecoxib group, 200 mg celecoxib plus placebo; naproxen group, 500 mg naproxen plus placebo) and one capsule daily with the evening meal (celecoxib group, placebo; naproxen group, 500 mg naproxen). Use of any medication other than the study drugs was considered concomitant therapy. NSAIDs and other analgesics for relief of arthritis symptoms (except for acetaminophen ≤ 2 g/day for ≤ 3 consecutive days, and not within 24 h before any arthritis assessment) were prohibited for the duration of the study. Patients were allowed to continue taking ≤ 325 mg/day aspirin for the duration of the study if they had been taking a stable dose for ≥ 30 days before the first dose of study medication.

Arthritis assessments (Patient's Assessment of Arthritis Pain [VAS], Patient's and Physician's Global Assessment of Arthritis) were carried out at week 2 and week 6. At week 6 or early termination, patients also completed the WOMAC OA index, Pain Satisfaction Scale and PHQ-9, and received a physical examination and laboratory testing.

Efficacy Evaluation

The primary efficacy outcome was the change from baseline to week 6 in the Patient's Assessment of Arthritis Pain measured on a VAS from 0 mm (no pain) to 100 mm (worst pain). Patients selected one knee to be the ‘index joint' for the pain assessments, which were performed at baseline, week 2 and week 6. Secondary efficacy outcomes included change in Patient's and Physician's Global Assessments of Arthritis and WOMAC OA index from baseline to week 6, change in APS pain scores from baseline to day 7 (week 1), change in Pain Satisfaction Scale and PHQ-9 scores from screening to week 6, and measurement of upper gastrointestinal (UGI) tolerability. The tertiary efficacy variable was the Complementary and Alternative Questionnaire worksheet, which was completed at the screening visit.

Safety Evaluation

Safety was assessed by monitoring treatment-emergent AEs, serious AEs and physical examinations. All observed or volunteered AEs and serious AEs were recorded at weeks 2 and 6, regardless of treatment group or suspected causality. In addition to individually recorded AEs, a composite measure of UGI tolerability was calculated, defined as patients having at least one of the following events as assessed by the investigator: moderate/severe nausea; moderate/severe abdominal pain; moderate/severe dyspepsia.

Statistical Analyses

All statistical analyses were performed using the SAS^® statistical package, version 8 (SAS Institute, Cary, NC, USA) for Windows^®. Sample size calculation was based on the maximum clinically acceptable difference for declaring noninferiority, which was compared with the lower boundary of the two-sided 95% confidence interval (CI) for the difference between the two treatment groups. The maximum clinically acceptable difference was 10 mm on the basis of the VAS scale, which ranged from 0 to 100 mm. Assuming a standard deviation of 25 mm, a sample size of 100 patients per active treatment group provided 80% power to find a 95% CI within the acceptable margins if the true treatment difference was zero. A total of 120 patients per active treatment group was required in order to adjust for the differences between the intention-to-treat and efficacy-evaluable populations due to nonevaluable patients. Sixty patients were randomized to placebo in order to have 80% power to detect a difference of 15 mm between the active treatment group and placebo on the VAS scale. A common standard deviation of 25 mm was assumed and hypothesis testing was carried out at the 5% significance level.

The primary efficacy analysis was carried out on the evaluable population, defined as treated patients with > 70% treatment compliance, no major protocol violations and having both baseline and week 6 VAS assessments. Change in VAS score from baseline to week 6 was analyzed using a generalized linear model with treatment and center effects in the model and baseline score as a covariate. Pairwise comparisons were conducted. Celecoxib was declared as effective as naproxen if the lower boundary of the two-sided 95% CI of the treatment difference (naproxen - celecoxib) lay above -10 mm. As a test of internal control, differences in the mean change in VAS score were also analyzed between celecoxib and placebo, and between naproxen and placebo.

All secondary analyses were performed using the modified intention-to-treat populations, comprising randomized patients receiving at least one dose of study medication and having at least one post baseline follow-up efficacy measure. The 24-item WOMAC scale and subscales were analyzed using a generalized linear model with treatment and center effects in the model and baseline WOMAC score as a covariate. The Patient's and Physician's Global Assessments were recorded as ‘improved', ‘no change' or ‘worsened', using the Cochran–Mantel–Haenszel (CMH) test, stratified by center. APS questions were analyzed using the CMH test stratified by center (question 1: Have you experienced any pain in the past 24 h?; yes or no). Change from baseline for the remaining questions was analyzed using a general linear model with treatment, center and baseline APS value (questions 2 – 5) as a covariate. For the PHQ-9, a variable was calculated for each patient to represent the categorical change in depression status (‘improved', ‘no change' or ‘worsened'), and analyzed using the CMH test, stratified by center. Change in PHQ-9 score was analyzed using a generalized linear model with treatment, center and screening PHQ-9 score as a covariate. Patient pain satisfaction was analyzed at screening and week 6 using the CMH test, stratified by center.

Safety and UGI events were analyzed in the safety population, defined as randomized patients receiving at least one dose of study medication. The incidence of UGI events was analyzed using the two-tailed Fisher's exact test. Hypothesis testing was carried out using a type-I error rate of 0.05. All statistical tests were performed as two-sided tests.

Results

Patient Disposition and Baseline Demographics

A total of 322 patients were randomized to treatment (127, 128 and 67 patients in the celecoxib, naproxen and placebo treatment groups, respectively); 316 patients received treatment, of whom 69 discontinued prematurely. Reasons for discontinuation included AEs, lack of efficacy, abnormal laboratory test results, protocol violation and defaulting (e.g. the patient was lost to follow-up or was no longer willing to participate) (Fig. 1).

Figure 1:

Patient screening, enrollment, randomization and withdrawal in a 6-week, randomized, double-blind, parallel-group study of African American patients with osteoarthritis of the knee who received celecoxib, naproxen or placebo. ^aIncludes ‘lost to follow-up' and ‘patient no longer willing to participate'. ^bIncludes ‘protocol violation'. AE, adverse event; bid, twice daily; lab, laboratory; mITT, modified intention-to-treat; qd, once daily

Baseline characteristics of the patients were similar among the three treatment groups (Table 1). Patients were predominantly female (80%) and had a mean age of 58 years (range 45 – 83 years). All patients reported that they were of African American descent. Most subjects were compliant in taking the study medication as instructed (mean compliancy 87.8% for celecoxib, 90.9% for naproxen and 76.7% for placebo).

Table 1:

Baseline demographic and clinical characteristics of African American patients (n = 322) with osteoarthritis (OA) of the knee who received celecoxib, naproxen or placebo for 6 weeks

Variable	Celecoxib group n = 127	Naproxen group n = 128	Placebo group n = 67
Age, years	58.0 ± 8.8	58.0 ± 8.1	58.0 ± 8.8
	(45 – 83)	(45 – 79)	(45 – 82)
Female	102 (80)	105 (82)	51 (76)
Duration of OA, years	5.4 ± 5.0	5.1 ± 5.6	6.2 ± 7.3
Patient's Global Assessment
Very good	0	0	0
Good	0	1 (1)	0
Fair	34 (27)	28 (22)	17 (25)
Poor	75 (59)	83 (65)	41 (61)
Very poor	18 (14)	16 (13)	9 (13)
Physician's Global Assessment
Very good	0	0	0
Good	0	0	0
Fair	33 (26)	33 (26)	18 (27)
Poor	84 (66)	87 (69)	45 (67)
Very poor	10 (8)	7 (6)	4 (6)
Functional capacity classification
I	5 (4)	4 (3)	2 (3)
II	63 (50)	75 (59)	39 (58)
III	58 (46)	49 (38)	26 (39)
IV	0	0	0
VAS score, mm	67.4 ± 12.7	68.4 ± 13.2	69.6 ± 12.6
WOMAC total domain score^a	55.6 ± 16.2	57.7 ± 17.0	60.2 ± 14.8

Data presented as mean ± SD (range) or n (%) of patients.

No statistically significant between-group differences (P ≥ 0.05); Cochran–Mantel–Haenszel (row mean scores differ) test stratified by center.

Baseline data were missing for one subject in the Naproxen group for Physician's Global Assessment and for one subject in the Celecoxib group for functional capacity classification.

WOMAC, Western Ontario McMasters osteoarthritis index; VAS, visual analog scale.

WOMAC total domain score is the sum of pain, stiffness and physical function domain scores.

No significant differences in baseline arthritis assessments were observed between treatment groups (Table 1). Most patients had mild-to-moderate OA, being assessed as ‘fair' or ‘poor' in the baseline Patient's and Physician's Global Assessments of Arthritis, and most had a functional capacity classification of II or III (Table 1). However, a greater proportion of patients in each of the treatment groups provided baseline global assessment ratings of ‘very poor' compared with the physician assessments.

Efficacy Outcomes

The primary efficacy outcome, Patient's Assessment of Arthritis Pain (VAS), improved in all three treatment groups (Table 2). Celecoxib treatment was observed to be as effective as naproxen in reducing OA pain, since the lower boundary of the two-sided 95% CI of the treatment difference (naproxen - celecoxib) was above -10 mm (lower boundary of the two-sided 95% CI, -8.9 mm). No statistically significant difference was observed between the active treatment groups and placebo (mean treatment difference naproxen - placebo -5.5 mm, 95% CI -14.0, 2.9; mean treatment difference celecoxib - placebo -3.3 mm, 95% CI -11.7, 5.2).

Table 2:

Patient's Assessment of Arthritis Pain at week 6 in the efficacy-evaluable population of African American patients with osteoarthritis of the knee who received celecoxib, naproxen or placebo for 6 weeks

VAS score	Celecoxib group n = 100	Naproxen group n = 106	Placebo group n = 46
Baseline, mean ± SE	67.7 ± 1.2	68.0 ± 1.3	70.0 ± 1.8
Week 6, mean ± SE	31.5 ± 2.4	29.9 ± 2.4	36.5 ± 4.2
Change from baseline
Least-squares mean ± SE	-36.2 ± 2.7	-38.0 ± 2.5	-33.5 ± 4.2
95% confidence interval	-8.9, 4.4	-14.0, 2.9	-11.7, 5.2

No statistically significant between-group differences (P ≥ 0.05); change in visual analog scale (VAS) score from baseline to week 6 was analyzed using a generalized linear model with treatment and center effects in the model and baseline score as a covariate. Pairwise comparisons were conducted.

Results for the secondary efficacy outcomes were similar between the celecoxib and naproxen groups. Mean change from baseline did not differ significantly between the celecoxib and naproxen treatment groups in the Patient's or Physician's Global Assessment of Arthritis scores. By week 6/early termination, most patients in each of the treatment groups described their arthritis condition as ‘good' or ‘very good' (celecoxib 51%, naproxen 57%, placebo 50%).

By the week 6/early termination visit, physicians described the arthritis condition of most patients in the celecoxib and naproxen groups as ‘good' or ‘very good' (celecoxib 58%, naproxen 70%) compared with 49% in the placebo group (Fig. 2). In addition, greater proportions of patients on celecoxib or naproxen were described by their physicians as being ‘improved' (a reduction of at least two grades or a change to ‘very good') by the week 6/early termination visit compared with the placebo group (celecoxib 52%, naproxen 56%, placebo 39%). Between-treatment differences were statistically significant at week 6/early termination in favor of celecoxib and naproxen over placebo (P = 0.047 and P = 0.039, respectively).

Figure 2:

Physician's Global Assessment of Arthritis: overall ratings in African American patients with osteoarthritis of the knee who received celecoxib, naproxen or placebo for 6 weeks: modified intention-to-treat population. NS, not statistically significant (P ≥ 0.05); Cochran–Mantel–Haenszel test

Changes in WOMAC OA index scores from baseline to week 6/early termination did not differ significantly between the celecoxib and naproxen groups. No between-treatment differences were observed (Table 3).

Table 3:

Change from baseline to week 6/early termination of the Western Ontario McMasters osteoarthritis index in African American patients with osteoarthritis of the knee who received celecoxib, naproxen or placebo for 6 weeks: modified intention-to-treat population

Domain	Celecoxib group n = 124	Naproxen group n = 125	Placebo group n = 65
Total	-22.6(1.8)	-26.0(1.8)	-20.8 (2.4)
Pain	-4.9 (0.4)	-5.7 (0.4)	-4.7 (0.6)
Stiffness	-1.8(0.2)	-2.0 (0.2)	-1.5 (0.2)
Physical function	-16.0(1.3)	-18.3 (1.3)	-14.4 (1.7)

Data presented as least-squares mean (SE).

In general, APS pain scores improved from baseline to day 7 (data not shown). There were no statistically significant differences between naproxen and placebo, and between celecoxib and naproxen, in mean APS pain scores. Celecoxib mean APS pain scores were only significantly different from placebo mean APS scores in the following instances: ‘any pain in the past 24 h' on day 6 (P = 0.017), ‘pain interference in normal work' on day 3 (P = 0.008) and ‘worst pain in 24 h' on day 3 (P = 0.007).

Overall, greater percentages of patients in the celecoxib and naproxen groups responded positively to the questions on the Pain Satisfaction Scale at the week 6/early termination visit, compared with placebo (data not shown). Mean scores for the sum of question 1 in the PHQ-9 (‘Over the past 2 weeks how often have you been bothered by loss of pleasure in activity, depression, problems with sleep, lack of energy, changes in appetite, feeling like a failure, trouble concentrating, moving slowly or becoming restless, or thoughts of being better off dead or hurting yourself?') improved in both the celecoxib and the naproxen group from screening to week 6/early termination, and remained the same in the placebo group (data not shown). No significant differences were observed between the active treatments.

As reported in the Complementary and Alternative Medicines Questionnaire, prescription medicines, over-the-counter (OTC) medicines and OTC medicines prescribed by a physician were used by 79, 57 and 53% of the screened population, respectively. The most commonly used alternatives to medical OA treatment were store-bought lotions, oils and creams (69%), vitamins or minerals (25%) and glucosamine and/or chondroitin sulfate (23%). Prayer (60%), massage (29%), nutritional therapy (25%) and spiritual healing (22%) were common alternative treatments. Herbal medicine, reflexology, acupuncture and homeopathy were used by < 17% of the population.

Safety Outcomes

There was a low incidence of the composite of UGI tolerability AEs (data not shown). There were significant differences in UGI tolerability in favor of celecoxib (P = 0.029, celecoxib versus naproxen; P = 0.013, celecoxib versus placebo). UGI events, specifically moderate/severe nausea, moderate/severe abdominal pain and/or moderate/severe dyspepsia, were experienced by 10 patients (six in the naproxen group; four in the placebo group). No UGI events occurred in the celecoxib group.

The incidence of AEs and treatment-emergent AEs was similar among the study groups (n = 76, 104 and 55 for celecoxib, naproxen and placebo treatment-emergent AEs, respectively). AEs considered to be treatment related occurred in 26 patients in the celecoxib group, 21 patients in the naproxen group and 15 patients in the placebo group. The most commonly reported treatment-related AEs were depression, headache, nausea, diarrhea, abdominal pain and dyspepsia (Table 4). A total of 16 patients discontinued the study because of AEs; 12 were withdrawn because of AEs that were considered treatment-related (six, four and two patients in the celecoxib, naproxen and placebo groups, respectively). There were only two serious AEs: one in the naproxen group (accidental nasal fracture due to a gunshot wound) and one in the celecoxib group (prostate biopsy, which indicated a malignancy, with subsequent prostatectomy) that occurred after the patient had completed the study treatment. Both were deemed unrelated to the study treatment. There were no deaths.

Table 4:

Treatment-related adverse events occurring in ≥ 2% of African American patients with osteoarthritis of the knee who received celecoxib, naproxen or placebo for 6 weeks (in decreasing order of occurrence in the celecoxib group)

Adverse events by preferred term	Celecoxib group n = 125	Naproxen group n = 125	Placebo group n = 66
Depression	5 (4)	4 (3)	3 (5)
Dizziness	3 (2)	0 (0)	0 (0)
Headache	3 (2)	0 (0)	2 (3)
Diarrhea	3 (2)	2 (2)	1 (2)
Gastroesophageal reflux	3 (2)	0 (0)	1 (2)
Nausea	3 (2)	3 (2)	2 (3)
Abdominal pain	1 (1)	3 (2)	3 (5)
Dyspepsia	1 (1)	4 (3)	1 (2)

Data presented as n (%) of patients.

Discussion

Previous studies have shown celecoxib to be effective and well tolerated in the treatment of OA pain;^{10 – 12} however, the efficacy and safety of celecoxib have not been specifically assessed in nonwhite populations because of the under-representation of such people in clinical trials.^{27 – 29} The present study was undertaken to extend the understanding of the efficacy, safety and GI tolerability of celecoxib in African American patients with knee OA.

In this population, mean changes in the Patient's Assessment of Arthritis Pain (VAS) did not differ significantly between the active treatments. The lower confidence limit of the treatment difference (naproxen – celecoxib) was above -10 mm; therefore, according to the study protocol, 200 mg celecoxib orally once daily was declared to be as effective as 500 mg naproxen orally twice daily in the reduction of OA pain. The results of the secondary efficacy analyses, notably the Patient's and Physician's Global Assessment of Arthritis, WOMAC scores, Pain Satisfaction Scale and Patient Health Questionnaire, were also similar between the celecoxib and naproxen treatment groups.

These results were consistent with previous OA studies demonstrating comparable efficacy of celecoxib with naproxen and other nsNSAIDs.^{10 – 12} In knee OA, the analgesic efficacy of celecoxib (100 mg and 200 mg twice daily) was shown to be comparable with that of naproxen (500 mg twice daily) in a 12-week study.¹⁰ In hip OA, Kivitz et al.¹² demonstrated that celecoxib at a daily dose of 200 mg and 400 mg was as effective as a standard therapeutic dose of naproxen (1000 mg daily) in reducing pain and improving functional capacity. Furthermore, the efficacy of celecoxib (100 mg twice daily and 200 mg twice daily) compared with that of diclofenac (100 mg once daily) or naproxen (500 mg twice daily) was reported in a mixed population of patients with OA (of the knee, hip or hand) treated for 12 weeks.¹¹ Although these studies demonstrated the comparability of celecoxib and nsNSAIDs in the treatment of the symptoms associated with OA, most patients were either white or race/ethnicity was undefined, and no subanalyses were carried out according to racial/ethnic origin.

The results of the present study are complicated by the fact that while improvements in pain and functionality were numerically greater in the active treatment groups than in the placebo group, few of these improvements reached statistical significance. This was particularly evident in that no significant differences were observed between the active treatments and the placebo group in the primary endpoint of Patient's Assessment of Arthritis (VAS), and in many of the secondary outcome measures.

Several reasons may account for this apparent lack of effect in the active treatments compared with placebo. Populations of patients with OA are highly heterogeneous, and differences between the patients recruited into the present study may have affected the study outcomes. Additionally, differences in study medication compliance may have contributed to the results observed. For example, patients took a mean of 77 – 91% of the protocol-specified dosage throughout the study. The doses used in this study were based on the daily recommended doses for treating the signs and symptoms of OA; noncompliance with the recommended dose may reduce the effectiveness of the study medication and thus affect the clinical outcome. It is also possible that the relatively short duration of this study limited its ability to detect latent effects of study treatment. Mean changes from baseline in the primary and secondary outcome measures continued to improve at interim assessments, and it is possible that the optimum response to treatment may have been met after 6 weeks. Few studies have assessed the onset of clinical efficacy of celecoxib in patients with OA,^{33 – 35} and only one study used naproxen as a comparator.³⁵ That study was pooled with data from two other randomized, placebo-controlled, 12-week trials of patients with OA with flare activity, and demonstrated the onset of analgesia (measured on the APS pain scale) with 200 mg and 400 mg celecoxib per day to be within 24 h; efficacy was maintained for ≥ 12 weeks and remained equivalent to that of naproxen 1000 mg per day.³⁵ Again, no subanalyses were carried out by race/ethnicity.

The improvements observed within the placebo treatment group cannot be disregarded, and the results of this study may be indicative of a high placebo effect. A meta-analysis demonstrated that the placebo effect in arthritis is robust and approaches the treatment effect, especially for pain, stiffness and self-reported function.³⁶ The significance of placebo effects within different racial/ethnic groups is becoming increasingly recognized; however, most evidence appears to be anecdotal and arises from discussions between clinicians. There has been some evidence of an increased placebo response among African American patients in clinical trials for depression and schizophrenia, compared with white patients.^37,38

Elevated rates of placebo response have been reported in other OA trials,^39,40 and a study to evaluate baseline factors as a prediction of placebo response in a large OA study concluded that placebo responders were more likely to report moderate levels of pain at baseline.⁴¹ In the latter study, patients from the placebo arm of a trial evaluating the effects of risedronate on pain and joint space narrowing in knee OA were stratified according to 20% and 50% declines in WOMAC pain scores between baseline and 6 months. Almost half of the patients receiving placebo reported clinically detectable improvements in pain (20% improvement in WOMAC score), and more than one-third of the patients had clinically meaningful improvements in pain (50% improvement in WOMAC score). Sociodemographic factors, including race and ethnicity, were not found to be reliable predictors of the placebo response.

A pooled analysis of placebo-controlled, randomized trials identified from Pfizer's Clinical Trials Registry (Pfizer Inc., New York, NY, USA) evaluated the effects of patient characteristics, such as age, sex, race and body mass index (BMI), on the efficacy of NSAIDs for the treatment of OA.⁴² A total of 15 studies were identified, comprising more than 21000 patients with OA. The NSAID response (nsNSAIDs and COX-2-selective NSAIDs) was shown to be influenced by age, BMI, sex and race, and the placebo response was influenced by age and race. Predictors of the highest benefits of pain reduction with NSAIDs, compared with placebo, were identified as older age (≥ 61 years), female sex, higher BMI (> 25 kg/m²) and white race. However, the analysis was of white versus nonwhite patients, nonwhite being a composite of black, Asian, Hispanic, Native American and other people; attribution of the difference to one race/ethnic group was therefore not possible.

If the results observed in the present study can be attributed to a placebo response, this may relate to the increased utilization of complementary and alternative medicines (e.g. homeopathy, meditation, prayer, massage, dietary supplements) reported among African American people.⁴³ Approximately 70% of the patients in this study reported the use of store-bought lotions, oils and creams. Prayer (60%), massage (29%), nutritional therapy (25%) and spiritual healing (22%) were also acknowledged as being part of the patients' healthcare utilization. This is consistent with what has been reported in the literature; data from the 2002 National Health Interview Survey indicated that 67.6% of African Americans reported using complementary and alternative medication.⁴⁴ However, the therapeutic effects of these medications remain controversial; while there is increasing evidence that alternative medicines and complementary therapies are more effective than placebo, the strength of this evidence is low, and benefits are typically attributed to subject expectation or placebo effect.^45,46

In common with many other OA studies, eligibility for this study included patients with OA in a flare state. It has been suggested that differences in flare state prior to screening may influence conclusions on treatment efficacy.⁴⁷ Patients with OA are known to fluctuate in their levels of pain and functioning;⁴⁸ thus, patients with high-intensity flares before treatment are likely to show improvement of symptoms irrespective of the treatment group to which they are assigned. Data from a placebo-controlled, randomized clinical trial of darbufelone and naproxen in patients with OA were reanalyzed and stratified by flare intensity at screening (low, moderate and high).⁴⁷ In patients with low-to-moderate flare intensity, significant improvements in mean WOMAC scores were observed in both active treatments; in patients with high-intensity flares at screening, a strong placebo effect was observed, with no significant differences between the active treatments and placebo.⁴⁷ In the present study, the baseline patient and physician global assessments (and mean VAS and WOMAC scores) would indicate that most patients' flares were of moderate-to-high intensity. If a large percentage of patients randomized to placebo experienced high-intensity flares at screening, this may provide an explanation for the improvements observed within this treatment group. Additionally, if more patients with high-intensity flares were randomized to the placebo group, this would increase the likelihood of a type II error, in which the results were falsely attributed to a placebo response. This being the case, the effect of the active treatments may have been underestimated.

Osteoarthritis flare intensity may vary between ethnic/racial populations. Studies in systemic lupus erythematosus indicate that the incidence of disease flares differs across ethnic groups: some reports suggest that patients of Southeast Asian ethnicity are more likely to experience disease flares than those of white ethnicity.⁴⁹ Although there are currently no known studies in arthritic patients, differences in disease flares between ethnic/racial groups may provide a plausible explanation for the response to treatment seen in the African American patients enrolled in the present study. Flare intensity may be an important determinant of treatment outcome, and should be taken into consideration when evaluating trial findings.

Safety and tolerability are key considerations of analgesic therapies, and the composite measure of UGI tolerability presented in this paper, in addition to individually recorded AEs, provides a clinically relevant assessment of treatment. No UGI tolerability events (defined as moderate or severe nausea, abdominal pain and/or dyspepsia) were reported in the celecoxib treatment group. This finding, coupled with the low incidence of patient discontinuation in the celecoxib group due to treatment-related AEs, provides further demonstration of the favorable GI tolerability profile of celecoxib.

The present study, which evaluated the efficacy and GI tolerability of celecoxib in African American patients with OA, provides insight into the management of OA symptoms within this under-represented population, and highlights the need for further studies looking at the role of race or ethnicity in the treatment response.

In conclusion, 200 mg celecoxib orally once daily was found to be as effective as 500 mg naproxen orally twice daily in relieving the symptoms associated with OA of the knee, and had better UGI tolerability than the naproxen regimen, in African American patients.

Footnotes

Acknowledgments

This study was sponsored by Pfizer Inc. Editorial support was provided by Dr C Campbell of PAREXEL, Uxbridge, UK, and was funded by Pfizer Inc.

Conflicts of interest: Margaret Noyes Essex, Michael O'Connell and Pritha Bhadra Brown are full-time employees of Pfizer Inc.

References

Lawrence

, Felson

, Helmick

: Estimates of the prevalence of arthritis and other rheumatic conditions in the United States, part II. Arthritis Rheum 2008; 58: 26–35.

American Academy of Orthopaedic Surgeons: Arthritis and related conditions. In: The Burden of Musculoskeletal Disease in the United States: Prevalence, Societal and Economic Cost. Rosemont, IL: American Academy of Orthopaedic Surgeons, 2008; pp 71–96.

Arden

, Nevitt

: Osteoarthritis: epidemiology. Best Pract Res Clin Rheumatol 2006; 20: 3–25.

Yelin

, Murphy

, Cisternas

: Medical care expenditures and earnings losses among persons with arthritis and other rheumatic conditions in 2003, and comparisons with 1997. Arthritis Rheum 2007; 56: 1397–1407.

National Institute for Health and Clinical Excellence: NICE Clinical Guideline 59. Osteoarthritis: the Care and Management of Osteoarthritis in Adults, 2008 (available at: http://www.nice.org.uk/nicemedia/live/11926/39557/39557.pdf).

Centres for Disease Control and Prevention (CDC):

Prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation - United States, 2003-2005.

MMWR Morb Mortal Wkly Rep 2006; 55: 1089–1092.

Nelson

, Renner

, Schwartz

: Differences in multijoint radiographic osteoarthritis phenotypes among African Americans and Caucasians: the Johnston County Osteoarthritis project. Arthritis Rheum 2011; 63: 3843–3852.

Bolen

, Schieb

, Hootman

: Differences in the prevalence and severity of arthritis among racial/ethnic groups in the United States, National Health Interview Survey, 2002, 2003, and 2006. Prev Chronic Dis 2010; 7: A64.

Hochberg

, Altman

, April

: American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee. Arthritis Care Res 2012; 64: 465–474.

10.

Bensen

, Fiechtner

, McMillen

: Treatment of osteoarthritis with celecoxib, a cyclooxygenase-2 inhibitor: a randomized controlled trial. Mayo Clin Proc 1999; 74: 1095–1105.

11.

Singh

, Fort

, Goldstein

: Celecoxib versus naproxen and diclofenac in osteoarthritis patients: SUCCESS-I Study. Am J Med 2006; 119: 255–266.

12.

Kivitz

, Moskowitz

, Woods

: Comparative efficacy and safety of celecoxib and naproxen in the treatment of osteoarthritis of the hip. J Int Med Res 2001; 29: 467–479.

13.

Bensen

, Zhao

, Burke

: Upper gastrointestinal tolerability of celecoxib, a COX-2 specific inhibitor, compared to naproxen and placebo. J Rheumatol 2000; 27: 1876–1883.

14.

Deeks

, Smith

, Bradley

: Efficacy, tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials. BMJ 2002; 325: 619.

15.

Mallen

, Essex

, Zhang

: Gastrointestinal tolerability of NSAIDs in elderly patients: a pooled analysis of 21 randomized clinical trials with celecoxib and nonselective NSAIDs. Curr Med Res Opin 2011; 27: 1359–1366.

16.

Niculescu

, Li

, Huang

: Pooled analysis of GI tolerability of 21 randomized controlled trials of celecoxib and nonselective NSAIDs. Curr Med Res Opin 2009; 25: 729–740.

17.

Silverstein

, Faich

, Goldstein

: Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA 2000; 284: 1247–1255.

18.

US Food and Drug Administration: FDA Memorandum: Analysis and Recommendations for Agency Action Regarding Non-steroidal Anti-inflammatory Drugs and Cardiovascular Risk, 2005 (available at: http://www.fda.gov/downloads/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm106201.pdf).

19.

US Food and Drug Administration: CELEBREX. Highlights of Prescribing Information, 2012 (available at: http://pfizer.com/files/products/uspi_celebrex.pdf).

20.

Creamer

, Lethbridge-Cejku

, Hochberg

: Determinants of pain severity in knee osteoarthritis: effect of demographic and psychosocial variables using 3 pain measures. J Rheumatol 1999; 26: 1785–1792.

21.

Edwards

, Fillingim

, Keefe

: Race, ethnicity and pain. Pain 2001; 94: 133–137.

22.

Golightly

, Dominick

: Racial variations in self-reported osteoarthritis symptom severity among veterans. Aging Clin Exp Res 2005; 17: 264–269.

23.

Jordan

: Effect of race and ethnicity on outcomes in arthritis and rheumatic conditions. Curr Opin Rheumatol 1999; 11: 98–103.

24.

Dominick

, Bosworth

, Jeffreys

: Racial/ethnic variations in non-steroidal anti-inflammatory drug (NSAID) use among patients with osteoarthritis. Pharmacoepidemiol Drug Saf 2004; 13: 683–694.

25.

Todd

, Samaroo

, Hoffman

: Ethnicity as a risk factor for inadequate emergency department analgesia. JAMA 1993; 269: 1537–1539.

26.

Todd

, Deaton

, D'Adamo

: Ethnicity and analgesic practice. Ann Emerg Med 2000; 35: 11–16.

27.

Hussain-Gambles

, Atkin

, Leese

: Why ethnic minority groups are under-represented in clinical trials: a review of the literature. Health Soc Care Community 2004; 12: 382–388.

28.

Shavers

, Lynch

, Burmeister

: Factors that influence African-Americans' willingness to participate in medical research studies. Cancer 2001; 91 (suppl): 233–236.

29.

Wendler

, Kington

, Madans

: Are racial and ethnic minorities less willing to participate in health research? PLoS Med 2006; 3: e19.

30.

Altman

, Asch

, Bloch

: Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986; 29: 1039–1049.

31.

International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH): Guideline for Good Clinical Practice. ICH Harmonised Tripartite Guideline, 1996 (available at: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E6_R1/Step4/E6_R1__Guideline.pdf).

32.

World Medical Association (WMA): World Medical Association Declaration of Helsinki - Ethical Principles for Medical Research Involving Human Subjects, 2008 (available at: http://www.wma.net/en/30publications/10policies/b3/17c.pdf).

33.

Battisti

, Katz

, Weaver

: Pain management in osteoarthritis: a focus on onset of efficacy - a comparison of rofecoxib, celecoxib, acetaminophen, and nabumetone across four clinical trials. J Pain 2004; 5: 511–520.

34.

Bianchi

, Broggini

, Balzarini

: Effects of nimesulide on pain and on synovial fluid concentrations of substance P, interleukin-6 and interleukin-8 in patients with knee osteoarthritis: comparison with celecoxib. Int J Clin Pract 2007; 61: 1270–1277.

35.

Moskowitz

, Sunshine

, Brugger

: American Pain Society pain questionnaire and other pain measures in the assessment of osteoarthritis pain: a pooled analysis of three celecoxib pivotal studies. Am J Ther 2003; 10: 12–20.

36.

Zhang

, Robertson

, Jones

: The placebo effect and its determinants in osteoarthritis: meta-analysis of randomised controlled trials. Ann Rheum Dis 2008; 67: 1716–1723.

37.

Escobar

, Tuason

: Antidepressant agents - a cross-cultural study. Psychopharmacol Bull 1980; 16: 49–52.

38.

Goldberg

, Schooler

, Davidson

: Sex and race differences in response to drug treatment among schizophrenics. Psychopharmacologia 1966; 9: 31–47.

39.

Altman

, Moskowitz

: Intraarticular sodium hyaluronate (Hyalgan) in the treatment of patients with osteoarthritis of the knee: a randomized clinical trial. Hyalgan Study Group. J Rheumatol 1998; 25: 2203–2212.

40.

Lehmann

, Brzosko

, Kopsa

: Efficacy and tolerability of lumiracoxib 100 mg once daily in knee osteoarthritis: a 13-week, randomized, double-blind study vs. placebo and celecoxib. Curr Med Res Opin 2005; 21: 517–526.

41.

Bartlett

, Solinger

, Bingham

III : Identifying factors to predict placebo responders in knee osteoarthritis clinical trials. Osteoarthritis Cartilage 2007; 15 (suppl C): C141–C142.

42.

Lakshminarayanan

, Niculescu

, Andrade-Ortega

: Influence of age, gender, race and body mass index (BMI) on the efficacy of nonsteroidal anti-inflammatory drugs (NSAIDs) in osteoarthritis (OA) trials: 81. J Clin Rheumatol 2006; 12: S27.

43.

Katz

, Lee

: Racial/ethnic differences in the use of complementary and alternative medicine in patients with arthritis. J Clin Rheumatol 2007; 13: 3–11.

44.

Brown

, Barner

, Richards

: Patterns of complementary and alternative medicine use in African Americans. J Altern Complement Med 2007; 13: 751–758.

45.

Cucherat

, Haugh

, Gooch

: Evidence of clinical efficacy of homeopathy. A meta-analysis of clinical trials. HMRAG. Homeopathic Medicines Research Advisory Group. Eur J Clin Pharmacol 2000; 56: 27–33.

46.

Linde

, Clausius

, Ramirez

: Are the clinical effects of homeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834–843.

47.

Scott-Lennox

, McLaughlin-Miley

, Lennox

: Stratification of flare intensity identifies placebo responders in a treatment efficacy trial of patients with osteoarthritis. Arthritis Rheum 2001; 44: 1599–1607.

48.

Bellamy

, Sothern

, Campbell

: Rhythmic variations in pain perception in osteoarthritis of the knee. J Rheumatol 1990; 17: 364–372.

49.

Boers

, Li

, Wong

: Differences in SLE disease activity between patients of Caucasian and South-East Asian/Chinese background in an Australian hospital. APLAR J Rheumatol 2006; 9: 43–48.