By: Dr. Geoffrey Modest
A recent Spanish study looked at the risk of peptic ulcer bleeding in patients with H Pylori (HP) infection and in patients also using NSAIDs/low-dose aspirin (see Am J Gastroenterol 2015; 110:684–689). This case-control study looked at 666 patients with endoscopically-confirmed major peptic ulcer bleeding and 666 controls (matched by age, sex, month of admission), assessing medication use in the prior 7 days. HP was assessed by serology.
Results:
–mean age 60; 29% female; with cases having significantly more smokers, ulcer history, dyspepsia, use of aspirin or NSAIDs, being on anticoagulants, not being on PPIs, and having HP infections (the latter being in 74.3% of cases and 54.8% of controls, [RR: 2.6 (CI: 2.0-3.3)])
–aspirin use (<300 mg/d) was associated with 15.8% of cases vs 12% of controls [RR: 1.9 (CI: 1.3-2.7)].
–NSAID use was associated with 34.5% vs 13.4% of controls [RR: 4.0 (CI: 3.0-5.4)]
–aspirin use plus HP infection did not further increase the risk of bleeding [RR:3.5 (CI: 2.0-6.1)]
–NSAID use plus HP infection did increase the risk of bleeding in at least an additive manner [RR: 8.0 (CI: 5.0-12.8)]
–subgroup analysis of those on aspirin >500 mg/d found similar results to NSAIDs
–so, their conclusion: the risk of documented peptic ulcer bleeding was dramatically increased in those with H Pylori infection determined serologically who were on NSAIDs but not on low-dose aspirin
Here are a few older articles on this subject:
–the first I saw: a 1997 RCT in the Lancet looked at 202 patients with musculoskeletal pain who were going to be put on NSAIDs. These patients did not have prior NSAID exposure and did have endoscopically-documented but asymptomatic H Pylori infection. The researchers then assessed the incidence of GI bleeding in the group given H Pylori eradication meds vs those given placebo treatment, finding that the development of endoscopically-proven GI ulcers occurred in 26% of those with persistant HP and only 3% in those with HP eradicated (see Lancet 1997; 350: 975–79).
—The American College of Gastroenterology published practice guidelines in 2009 which supported the conclusion that there is an additive role of H Pylori infection in those on NSAIDs in the development of ulcers and that 2 systematic reviews have shown that HP eradication is superior to placebo in preventing peptic ulcers among NSAID users.They comment that there is a potential advantage to H pylori testing and that then using either a gastroprotective agent or eradicating H Pylori may be useful depending on the individual’s underlying GI risk (see Am J Gastroenterol 2009; 104:728 – 738)
–also there have been several articles in my BMJ blog (for articles on treatment issues, see here. For an article on potential benefits in reducing gastric cancer, see here.)
So, what can one conclude?
–One could adopt the posture, as I do, that we should be screening and treating patients with HP more rigorously, with part of the rationale being that many people (perhaps, way too many) take a lot of NSAIDs and the data are pretty impressive that treating the HP infection decreases the risk of bleeding (though another angle is to gastroprotect everyone with a PPI, or possibly high-dose histamine-blocker — though I personally don’t fancy the idea of treating a potential complication of a med with another med, which in itself may have some adverse effects, such as osteoporosis, increased pneumonia, etc. I do very strongly try to limit use of NSAIDs in my patients as much as I can persuade them, given not just the GI effects, but effects on hypertension, heart failure, kidneys… Also, as an aside and pretty anecdotal report: I saw a small clinical research study in the Lancet about 20-25 years ago showing dramatic resolution of ITP (idiopathic thrombocytopenic purpura) if one treats an underlying HP infection (in those infected). Then, about 2 weeks after seeing this, I had an Irish patient from Boston who had prednisone-resistant ITP, but (unexpectedly) had positive HP antibodies, got treatment for HP, and the ITP (with persistent platelet count in the 20,000 range and lots of ecchymoses) vanished and did not recur to date.
–The data on aspirin are reasonably consistent that there is not a clear additive effect with HP infection. One could still consider gastroprotection just because of the gastric effects of aspirin. I do reinforce with patients that low-dose aspirin is associated with GI bleeding, and that it is still prudent to avoid other GI irritants (smoking, alcohol, NSAIDs, etc). And, by the way, there are some concerns that enteric-coated aspirin may not be as effective for cardioprotection as regular 81mg aspirin and is no more gastroprotective. See here for details.