SEARCH

SEARCH BY CITATION

Keywords:

  • Celecoxib;
  • gastritis;
  • hypertension;
  • NSAIDs;
  • thromboembolism

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

Background

Cyclo-oxygenase (COX)-2 inhibitors have been the target of severe criticism, more so following the withdrawal of Rofecoxib. Post-marketing surveillance of Celecoxib in Asian Indians, who are predisposed to premature athero-thrombotic events, has not been studied.

Aims

To study the adverse effects of Celecoxib and compare them with those of other non-steroidal anti-inflammatory drugs (NSAIDs) in an Asian Indian cohort.

Materials and methods

This is a retrospective chart review with convenience sampling of patients on NSAIDs (at least five tablets a week, for at least 3 months prior to the study), attending the Rheumatology clinic of a tertiary care institution in south India between June 2004 and November 2004. Those with pre-existing heart disease, hypertension, thrombo-embolic disease, peptic ulcer and patients on corticosteroids were excluded. All the recorded adverse events were noted and compared between the Celecoxib and non-selective NSAID users. Univariate analysis using Chi-square test was performed.

Results

Of the 1387 patients included, 915 were on Celecoxib. In the NSAID group, 204 had used multiple NSAIDs in sequence. Of the Celecoxib users, 164 had switched over to an NSAID during the study period. New onset of hypertension was significantly higher in the Celecoxib users as compared to non-selective NSAID users (3.06% vs. 1.27%, = 0.04). However, those who had switched over to NSAIDs did not show this trend. NSAID users, on the other hand, had significant gastrointestinal (GI) toxicity (2.54% vs. 0.327%, = 0.001). A significant number of Celecoxib users who switched over to NSAIDs also developed GI toxicity (6.1% vs. 1.21%, = 0.018) over a shorter time span, as compared to the continuous NSAID users. Multiple NSAID users had higher adverse events (6.37% vs. 2.23%, = 0.023) as compared to single NSAID users.

Conclusion

Celecoxib significantly increased the incidence of new onset hypertension in this cohort of Indian patients with rheumatic diseases. No thromboembolic events were documented.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

Non-steroidal anti-inflammatory drugs (NSAIDs) are widely acclaimed for their anti-inflammatory, analgesic and antipyretic properties. The non-selective NSAIDs act by inhibiting both isoforms of the enzyme cyclo-oxygenase (COX-1 and COX-2). COX-2 inhibition is mainly responsible for anti-inflammatory actions and COX-1 inhibition leads to NSAID-induced gastrointestinal damage.[1] The hypothesis that selective inhibition of COX-2 isoform may help in reducing pain and inflammation without compromising the gastric mucosa led to discovery of the selective COX-2 inhibitors. Celecoxib was developed first in this group and was found to possess analgesic and anti-inflammatory efficacy comparable to the non-selective NSAIDs in treatment of inflammatory arthritic conditions.[2] In view of their gastrointestinal safety profile, within a short span of time COX-2 inhibitors gained popularity over non-selective NSAIDs.[3] However, COX-2 inhibition reduces vascular prostacyclin (PGI2) production, thus affecting the balance between prothrombotic and anti-thrombotic eicosanoids.[4] This property can tip the balance in favor of prothrombotic eicosanoids, which can lead to increased cardiovascular thrombotic events.[5]

Serious concerns regarding the cardiovascular safety of Rofecoxib were expressed following the Vioxx Gastrointestinal Outcomes Research (VIGOR) study.[6] Results of the Adenomatous Polyp Prevention on Vioxx (APPROVe) trial confirmed the danger,[7] and this led to withdrawal of Rofecoxib from the market in September 2004. The Asian Indian population, predisposed to premature coronary heart disease, with a high incidence of thrombogenic and atherogenic risk factors,[8] is likely to be vulnerable to the adverse effects of COX-2 inhibitors. The positive association of cardiovascular events and inflammatory rheumatic diseases has already been proven.[9-11] Thus, rheumatologists should be cautious in using COX-2 inhibitors in patients with inflammatory arthritis.

At the beginning of this millennium when Celecoxib was introduced in the Indian market we had switched our inflammatory arthritis patients to the COX-2 inhibitor. Safety concerns regarding Rofecoxib prompted us to look into the cardiovascular, renal and gastrointestinal (GI) safety profile of Celecoxib in comparison with non-selective NSAIDs.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

This was a retrospective, case-sheet-based study using convenience sampling. Patients attending the outpatient and inpatient services of the department of Clinical Immunology and Rheumatology of our large tertiary care teaching hospital, who were prescribed either Celecoxib or non-selective NSAIDs (naproxen, indomethacin or diclofenac sodium) for at least 3 months between June 2004 and November 2004, were included. Patients below the age of 12 years and those with pre-existing cardiovascular disease, hypertension, diabetes, renal failure, acid peptic disease, esophageo-gastro-duodenitis, thrombo-embolic events or in a prothrombotic state, were excluded. All the selected patients were broadly divided into the Celecoxib group (Group I) and the NSAID group (Group II). Group I patients were further divided into those who had used Celecoxib throughout the period of study (Group Ia) and those who had switched to non-selective NSAIDs after taking Celecoxib for at least 3 months and had continued the non-selective NSAIDs for another 3 months (Group Ib). Similarly, patients in Group II were divided into subgroups of those who had taken a single NSAID throughout (Group IIa) and those who had taken multiple NSAIDs sequentially (Group IIb). Demographic data and all the documented cardiovascular, renal and GI side effects of these selected patients were extracted from the case sheets.

A thrombo-embolic event was defined as cardiac arrest due to coronary artery disease, myocardial infarction, angina pectoris, valvular heart disease with in situ thrombus, cerebro-vascular accident in the form of thrombotic or embolic stroke or transient ischemic attack, retinal artery thrombosis, deep vein thrombosis, pulmonary embolism, pulmonary infarction and hepatic vein thrombosis.[12] GI side effects defined in this study included non-specific dyspepsia, ulceration, upper GI bleed or death related to any of these events.[13] New onset hypertension was defined as increase in systolic blood pressure (BP) by more than 20 mmHg, increase in baseline or absolute level of systolic BP above 140 mmHg, and rise in diastolic blood pressure by more than 15 mmHg, increase in baseline or absolute level of diastolic BP above 90 mmHg. Renal toxicity was defined as toxicity directly associated with the intake of Celecoxib or non-selective NSAIDs, including acute tubular necrosis, acute tubulointerstitial nephritis, glomerulonephritis, renal papillary necrosis, chronic renal failure or salt and water retention.

Comparisons were done between the Celecoxib users and non-selective NSAID users within the main groups, as well as within the sub-groups as mentioned above, in relation to the demographic parameters and toxicities. Chi-square test was used to locate any significant differences between the groups.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

A total of 5850 patients’ charts were reviewed, of which 3121 patients had taken non-selective NSAIDs or Celecoxib continuously, at least for 3 months. From this group, 1881 patients were being followed up in the Department of Clinical Immunology and Rheumatology. Based on the exclusion criteria, 494 patients were excluded and finally 1387 patients' charts were included in the study.

The number of patients within each sub-group, with their demographic data and diagnostic categories, are given in Table 1. There was a female preponderance in all the groups, as expected in systemic autoimmune connective tissue diseases. Age group and duration of disease were comparable in all the groups. Rheumatoid arthritis (RA) patients constituted more than half the number in all groups. This was followed by spondyloarthritis, psoriatic arthritis, other connective tissue disorders, osteoarthritis and crystal arthritis.

Table 1. Demography and diagnosis of the study patients
ParametersCelecoxib group (I)NSAID group (II)
 Group IaGroup IbTotalGroup IIaGroup IIbTotal
 n = 751n = 164n = 915n = 268n = 204n = 472
  1. a

    Osteoporosis, chronic pain syndrome, metabolic arthropathies.

Mean age ± SD (years)40.54 ± 16.9338.5 ± 13.3738.91 ± 14.0535.62 ± 14.0336.82 ± 13.4237.34 ± 13.68
Male: female282 : 46953 : 111335 : 580114 : 15486 : 118200 : 272
Mean duration ± SD (months)7.54 ± 6.535.95 ± 5.217.54 ± 6.537.21 ± 5.856.82 ± 5.627.21 ± 5.85
Diagnosis
Rheumatoid arthritis455 (60.58%)86 (52.43%)541 (59.12%)142 (52.98%)115 (56.37%)257 (54.44%)
Spondyloarthritis161 (21.43%)38 (23.17%)199 (21.74%)59 (22.01%)44 (21.56%)103 (21.82%)
Psoriatic arthritis43 (5.72%)11 (6.7%)54 (5.9%)13 (4.85%)11 (5.39%)24 (5.08%)
Crystal arthritis15 (1.99%)5 (3.04%)20 (2.18%)9 (3.35%)3 (1.47%)12 (2.54%)
Osteoarthritis13 (1.73%)7 (4.26%)20 (2.18%)9 (3.35%)10 (4.9%)19 (4.02%)
Connective tissue disease35 (4.66%)8 (4.87%)43 (4.69%)19 (7.08%)8 (3.92%)27 (5.72%)
Others†a29 (3.86%)9 (5.48%)38 (4.15%)17 (6.34%)13 (6.37%)30 (6.35%)

No thrombo-embolic event was recorded in any of the included patients in the adverse effect profile (Table 2). Major side effects documented were new onset hypertension, GI toxicities leading to discontinuation of the medication and renal failure. Minor side effects included edema and headache.

Table 2. Side effect profile in each group
ParametersCelecoxib group (I)NSAID group (II)
 Group IaGroup IbTotalGroup IIaGroup IIbTotal
 = 751= 164= 915= 268= 204= 472
  1. Group I, Celecoxib group; Group II, non-steroidal anti-inflammatory drug (NSAID) group; Group Ia, patients who used Celecoxib throughout the study period; Group Ib, those who switched to non-selective NSAIDs after taking Celecoxib for at least 3 months and had continued the non-selective NSAIDs for another 3 months; Group IIa, those who had taken a single NSAID throughout; Group IIb, those who had taken multiple NSAIDs sequentially.

New onset hypertension (%)23 (3.06%)5 (3.05%)28 (3.06%)4 (1.49%)2 (0.98%)6 (1.27%)
Gastrointestinal side effects (%)1 (0.13%)2 (1.21%)3 (0.33%)2 (0.74%)10 (4.9%)12 (2.54%)
Renal failure (%)1 (0.13%)01 (0.1%)01 (0.49%)1 (0.21%)
Thromboembolic events (%)000000
Total (%)25 (3.32%)7 (4.26%)32 (3.49%)6 (2.23%)13 (6.37%)19 (4.02%)

The Celecoxib group (Group I) had significantly higher incidence of new onset hypertension (Table 3) as compared to the non-selective NSAID group (Group II) (= 0.04). This difference was not seen when continuous Celecoxib users were compared with those Celecoxib users who switched over to non-selective NSAIDs (Groups Ia and Ib) (= 0.993). There was no difference between those who used Celecoxib continuously (Group Ia) and those patients who switched over to non-selective NSAIDs after a minimum of 3 months use of Celecoxib (Group Ib) in terms of any side effects (= 0.553). Non-selective NSAID users (Group II), on the other hand, had significantly higher GI toxicity when compared to all Celecoxib users (Group I) (= 0.001) and those who continued only on Celecoxib throughout the study period (Group Ia) (< 0.001).

Table 3. Difference in adverse effects in both study groups
GroupsAll side effectsNew onset hypertensionGI toxicityRenal failure
  1. Group I, Celecoxib group; Group II, non-steroidal anti-inflammatory drug (NSAID) group; Group Ia, patients who used Celecoxib throughout the study period; Group Ib, those who switched to non-selective NSAIDs after taking Celecoxib for at least 3 months and had continued the non-selective NSAIDs for another 3 months; Group IIa, those who had taken a single NSAID throughout; Group IIb, those who had taken multiple NSAIDs sequentially.

Group I vs. Group II

32/915 (3.49%) vs. 19/472 (4.02%)

= 0.6

28/915 (3.06%) vs. 6/472 (1.27%)

= 0.04

3/915 (0.327%) vs. 12/472 (2.54%)

= 0.001

1/915 (0.109%) vs. 1/472 (0.21%)

= 1.00

Group Ia vs. Group II

25/751 (3.32%) vs. 19/472 (4/02%)

= 0.03

23/751 (3.06%) vs. 6/472 (1.27%)

= 0.045

1/751 (0.133%) vs. 12/472 (2.54%)

≤ 0.001

1/751 (0.133%) vs. 1/472 (0.21%)

= 1.0

Group Ib vs. Group II

7/164 (4.26%) vs. 19/472 (4.02%)

= 0.89

5/164 (3.05%) vs. 6/472 (1.27%)

= 0.13

2/164 (1.21%) vs. 12/472 (2/54%)

= 0.49

0/164 (0) vs. 1/472 (0.21%)

= 1.0

Significant number of Celecoxib users who had switched over to non-selective NSAIDs developed gastritis after the change-over (6.1% vs 1.21%; p = 0.018). (6.1% vs 1.21%; p = 0.018). Adverse effects during the non-selective NSAID period appeared much earlier (6.08 ± 5.3 months) as compared to 15.75 ± 9.82 months during the Celecoxib period (p = 0.001) (Table 4). On the other hand, patients who were on multiple non-selective NSAIDs (Group IIb) showed significantly higher overall side effects (13/204, 6.37% vs. 6/268, 2.23%; = 0.023) and GI side effects (10/204, 4.9% vs. 2/268, 0.74%; = 0.04), as compared to patients who were only on a single NSAID (Group IIa).

Table 4. Follow-up of adverse effect profile in Group Ib patients
ParametersCelecoxib periodNSAID periodP-value
 = 164= 164 
  1. Group 1b, patients who had switched to non-selective NSAIDs after taking Celecoxib for at least 3 months and had continued the non-selective NSAIDs for another 3 months; NSAIDs, non-steroidal anti-inflammatory drugs.

New onset hypertension5 (3.05%)2 (1.21%)0.44
GI toxicity2 (1.21%)10 (6.1%)0.018
Renal failure01 (0.61%)1.00
Total side effects7 (4.26%)13 (7.92%)0.166
Mean duration (months) of drug usage in patients with side effects (±SD)15.75 ± 9.826.08 ± 5.300.001*

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

NSAIDs are widely prescribed for pain relief in all rheumatological conditions because of their ability to curb inflammation and optimize function. They have been proven to be more efficacious than paracetamol for management of pain and improvement of quality of life.[14] This study was undertaken in the wake of the Rofecoxib controversy, to study the toxicity profile of Celecoxib in an Asian Indian population. Globally there was a steep decline in the use of COX-2 inhibitors following withdrawal of Rofecoxib.[15] As compared to Rofecoxib, COX-2 inhibition is less with Celecoxib.[16] Thus, thrombogenic effects of Celecoxib are expected to be less than Rofecoxib.

No thrombo-embolic events were reported with the use of Celecoxib for more than 3 months in our patients with rheumatic diseases. The most significant observation in this cohort was the development of new onset hypertension in young patients using Celecoxib, as compared to those who had used non-selective NSAIDs. This finding is in stark contrast to two other studies which have shown Celecoxib to have a significantly lower incidence of hypertension when compared to ibuprofen,[17] and an equal risk of developing new onset hypertension as compared to diclofenac.[18] No significant hypertension was observed in those Celecoxib users who had switched over to other non-selective NSAIDs. This may suggest a cause–effect relationship between the two in this population. Muscara et al. have described elevation of blood pressure and leukocyte adherence in rats on suppression of COX-2. They have proposed that the hypertensive effects of Celecoxib may be due to its effects on renal function and on postacyclin synthesis.[19] However, this needs to be tested prospectively. Ambulatory blood pressure data has suggested a 2–4 mmHg increase in systolic blood pressure over 4 h after dosing with Celecoxib.[20] Due to the relatively short half-life of Celecoxib, Solomon et al. have hypothesized that the timing of blood pressure measurements in relation to the dosing of Celecoxib may be the reason for the varied results in different studies.[21]

The incidence of GI toxicity among the non-selective NSAID users was, as expected, higher when compared to the Celecoxib users. Relatively low occurrence of gastric side effects of Celecoxib is related to its low propensity to inhibit the COX-1-mediated production of prostaglandins involved in the maintenance of GI mucosal integrity.[22] Renal failure was rare and similar in both the groups, as reported in the literature.[23] The finding of lack of thrombo-embolic events in our patients on Celecoxib cannot be generalized at the moment without a prospective cohort study. Celecoxib is known to reduce endothelial tissue factor expression, a key initiator of the coagulation cascade.[24]

Methodological limitations of this study included its retrospective, case-sheet-based, convenience sampling, which relied a lot on the accuracy of written records and was therefore, prone to selection bias. For Asian Indians, who are already prone to premature atherosclerosis and cardiovascular mortality, a systemic autoimmune inflammatory condition by itself acts as a second hit. Use of Celecoxib in this subset could act as a third hit as per the biological basis mentioned earlier, which is further confirmed by the results of this study.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References

Asian Indian patients with inflammatory rheumatic diseases on Celecoxib alone at dosages of 400 mg/day, for 3 months or longer, have significantly high risks of developing new onset hypertension. In comparison, patients on non-selective NSAIDs for similar duration develop more GI toxicity, more so when they use multiple conventional NSAIDs. Those patients on Celecoxib who switched over to conventional NSAIDs also had significantly higher GI toxicity. There was no thromboembolic event recorded in this study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. References