Skip navigation
Acute Pain | Chronic Pain | General

Safety of topical NSAIDs


Clinical bottom line

When other drug use is taken into account there was no hint of association between topical NSAIDs and upper gastrointestinal bleeding Despite limitations of this sort of study, the evidence is that topical NSAIDs do not cause gastrointestinal problems.


JM Evans et al. Topical non-steroidal anti-inflammatory drugs and admission to hospital for upper gastrointestinal bleeding and perforation: a record linkage case-control study. BMJ 1995 311: 22-26.


This study covered 320,000 people living in Tayside, Scotland between 1989 and 1994. A case was any patient with a diagnostic code for upper gastrointestinal bleeding and perforation. This included acute, chronic or unspecified gastric ulcer, duodenal ulcer, with haemorrhage or perforation, or haematemesis or melena. Accuracy was determined in a validation study.

Controls were from the community or hospital. Community controls were matched for age and sex and were generated randomly for each case (up to six controls for each case). Up to two hospital controls were also selected, matched for age and sex, and generated randomly for each case. They could have been admitted to hospital for any reason apart from gastrointestinal bleeding and perforation.

Previous exposure to any or oral and topical NSAIDs, and gastrointestinal healing drugs, was examined for association with admission with GI bleeding. Topical NSAIDs was a wide definition that also included rubefacients (salicylates).


Over four years, 23,100 patients had been prescribed topical NSAIDs (7% of the population).

There were 1101 cases and 6593 controls. Association with exposure was assessed for 45-day and ever exposure, with community and hospital controls, and for all cases, bleeding and perforation separately, and with crude odds ratios and odds ratios calculated using conditional logistic regression.

Oral NSAIDs, topical NSAIDs and ulcer healing drugs were always significantly associated with higher bleeding rates with community controls, and without any adjustment for use of other drugs (Table 1). Oral NSAIDs and ulcer healing drugs, but not topical NSAIDs were associated with higher bleeding rates with hospital controls without any adjustment.

Table 1: Association between drug use and GI bleeding, without adjustment for use of other drugs (cyan shows not significant)

Crude odds ratios
Oral NSAID 45-day
Topical NSAID 45-day
Ulcer healing drugs 45-day

Using adjustment for the use of other drugs with conditional logistic regression, the following emerged:


The problem with examining safety of topical NSAIDs is that patients may well have used oral NSAIDs, or be also using oral NSAIDs, and may be using topical NSAIDs because they are high risk of bleeding, and may therefore be taking ulcer healing drugs. This complicates matters when looking for simple association between drug use and bleeding.

In this study, for instance, use of ulcer healing drugs was associated with higher rates of gastrointestinal bleeding. Yet we would not take this as evidence that ulcer healing drugs caused gastrointestinal bleeding, rather that people take these drugs because they are at higher risk of bleeding, and the use of ulcer healing drugs do not completely remove this risk.

So we have to take other criteria into account. When this is done, the association of oral NSAIDs with bleeding remains. But for topical NSAIDs, there was no hint of association. Despite limitations of this sort of study, the evidence is that oral NSAIDs do not cause gastrointestinal problems.