Upper GI bleeding: celecoxib led to fewer GI complications than ibuprofen or diclofenac

Clinical bottom line (level 1b)

  1. Patients with osteoarthritis or rheumatoid arthritis who took celecoxib compared with ibuprofen or diclofenac were less likely to stop medication due to adverse effects (NNT = 44 at 6 months) .
  2. Patients on celecoxib were less likely to have GI complications (NNT = 210 at 6 months) , bleeding-related complications (NNT = 35 at 6 months) , or renal complications (NNT = 64 at 6 months) , but more likely to develop cutaneous complications (NNH = 30 at 6 months) .
  3. There was no clear difference in cardiovascular events between the two groups.
Silverstein et al: JAMA 2000; 284 : 1247-1255
Expires January 2004

The study

Double-blinded concealed randomised trial with intention-to-treat
Setting: 386 clinics, USA and Canada

8059 patients (aged 18 to 90; mean 60, 69% female) with osteoarthritis or rheumatoid arthritis

Excluded if
  • known hypersensitivity to study drugs or sulphonamides
  • possibility of pregancy, or lactating
  • active GI, hepatic,renal or coagulation disorders
  • malignancy unless removed surgically with no recurrence within 5 years
  • oesophageal or gastroduodenal erosion within the previous 30 days
  • history of gastric or duodenal surgery other than an oversew
  • osteoarthritis or rheumatoid arthritis for < 3 months
  • aged < 18

Control Group: (n = 4028, 3981 analysed): diclofenac 75 mg twice daily or ibuprofen 800 mg three times a day
Experimental Group: (n = 4031, 3987 analysed): celecoxib 400 mg twice daily
Patients were not allowed to take other NSAIDs
99% followed for 6 months
Outcome notes:
  • GI complications : gastroduodenal ulcer, upper GI bleeding, perforation, gastric outflow obstruction
  • bleeding-related complications : anaemia, ecchymosis, haematochezia
  • renal complications : peripheral oedema, hypertension, increased creatinine level
  • cardiovascular complications : cerebrovascular accident, myocardial infarction, angina
  • cutaneous complications : rash, pruritis, urticaria

The evidence

Outcome Time to outcome CER EER RRR
(95% CI)		 ARR
(95% CI)		 NNT
(95% CI)
withdrawal from study 6 months 1784
(44.3%)		 1611
(40.0%)		 10%
(5% to 14%)		 4.32%
(2.17% to 6.48%)		 23
(15 to 46)
GI complications 6 months 51
(1.27%)		 32
(0.79%)		 37%
(3% to 60%)		 0.47%
(0.032% to 0.91%)		 210
(110 to 3200)
withdrawal due to adverse effects 6 months 822
(20.4%)		 732
(18.2%)		 11%
(3% to 19%)		 2.25%
(0.53% to 3.97%)		 44
(25 to 190)
bleeding-related complications 6 months 238
(5.91%)		 123
(3.05%)		 48%
(36% to 58%)		 2.86%
(1.96% to 3,76%)		 35
(27 to 51)
renal complications 6 months 263
(6.53%)		 200
(4.96%)		 24%
(9% to 36%)		 1.57%
(0.55% to 2.58%)		 64
(39 to 180)
cardiovascular complications 6 months 39
(0.97%)		 37
(0.92%)		 5%
(-48% to 39%)		 0.05%
(-0.38% to 0.47%)		 2000
(NNT = 270 to infinity;
NNH = 210 to infinity)
cutaneous complications 6 months 163
(4.05%)		 298
(7.39%)		 -83%
(-120% to -52%)		 -3.35%
(-4.36% to -2.33%)		 -30
(-43 to -23)

Citation

  1. Silverstein FE, Faich G, Goldstein JL, et al: gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study : a randomized controlled trial. JAMA 2000; 284 : 1247-1255
Search Terms:
Contributor: Chris Ball, January 2002
Reviewer:

Clinical Question.
Patient osteoarthritis, rheumatoid arthritis
Intervention or Exposure celecoxib, COX-2 inhibitor
Comparison ibuprofen, diclofenac
Outcome ulcer, GI bleeding, renal complications, rash