NSAIDs: topical administration did not clearly increase the risk of GI bleed or perforation, but oral did.

Clinical bottom line (level 3b)

  1. Patients who took oral NSAIDs are at increased risk of having GI bleeding or perforation (NNH = 100 at 45 days) .
  2. Patients who take anti-ulcer medication are at increased risk of developing GI bleeding or perforation (NNH = 130 at 2 years) .
  3. There is no clear increase in risk of GI bleeding or perforation with use of topical NSAIDs.
Evans et al: British Medical Journal 1995; 311: 22-26
Expires January 2003

The study

Case-control study with objective outcomes, adjusted for confounding factors, not validated in an independent set of patients.

Setting: university hospital, UK

3280 patients (aged ?, ?% male) discharged from hospital with GI disorders

Excluded if
  • recurrent GI bleed or perforation


Cases: 1096 patients (52% male, mean age ): discharged from hospital following a first episode of upper GI bleeding or perforation; aged < 50, who had a prescription for NSAIDs in the previous 2 years.
Controls: 2184 patients (% male, mean age ): matched for age and sex; discharged from hospital with any diagnosis other than GI bleeding or perforation within 90 days of matched case

Factors studied:

exposure to NSAIDs excluding aspirin



Factors summarised:

oral NSAIDs within last 45 days

topical NSAIDs within 45 days

ulcer healing drugs any exposure

oral NSAIDs within last 45 days

topical NSAIDs within last 45 days

ulcer healing drugs any exposure



Logistic regression analysis performed on risk factors.

Outcomes studied:

GI bleeding or perforation (community controls)

GI bleeding or perforation (hospital controls)


The evidence

Patient expected event rate for GI bleeding or perforation (community controls): 1%
risk factor for
GI bleeding or perforation (community controls) 		adjusted OR
(95% CI)		 NNH
(95% CI)
oral NSAIDs 2.59
(2.12 to 3.16) 		65
(48 to 91)
topical NSAIDs 1.45
(0.84 to 2.50) 		230
(-630 to 68)
ulcer healing drugs 4.21
(3.63 to 4.88) 		32
(27 to 39)

Patient expected event rate for GI bleeding or perforation (hospital controls): 1%
risk factor for
GI bleeding or perforation (hospital controls) 		adjusted OR
(95% CI)		 NNH
(95% CI)
oral NSAIDs 2.00
(1.60 to 2.50) 		100
(68 to 170)
topical NSAIDs 1.06
(0.60 to 1.88) 		1700
(-250 to 120)
ulcer healing drugs 1.76
(1.51 to 2.07) 		130
(95 to 200)

Comments

  1. See Moore et al 1998 for a review of efficacy of topical NSAIDs.

Citation

  1. Evans JM, McMahon AD, McGilchrist MM, et al: topical non-steroidal anti-inflammatory drugs and admission to hospital for upper gastrointestinal bleeding and perforation: a record linkage case-control study. British Medical Journal 1995; 311: 22-26
Contributor: Chris Ball and Clare Wotton, October 1999
Reviewer: Daniel Sontheimer

Clinical Question.
    Patient patients
    Intervention or Exposure topical or oral NSAID
    Outcome GI bleeding or perforation