Should PPIs be routinely co-prescribed with long-term NSAIDs?

This article was published in the July 2017 edition of Medical Observer, under the title, “Protecting against ulceration” (pp. 66-67). (PDF)

Clinical scenario

Josef, a 68-year-old retiree saw me recently with knee osteoarthritis, and he commenced a therapeutic trial of naproxen.  Afterwards, a discussion on the GPs Down Under online group made me wonder whether I should have co-prescribed a PPI (proton pump inhibitor) for gastroprotection prophylactically.  This hasn’t been my usual practice for individuals with no specific risks or history of peptic ulcers.  What is the evidence?


Clinical question

When commencing NSAIDs, what is the effect of using PPIs as prophylaxis on the risk of gastrointestinal toxicity?


What does the research evidence say?

Step 1: The Cochrane Library

The Cochrane Library has a systematic review published in 2002 on the prevention of NSAID-induced gastroduodenal ulcers [1].  It was edited in 2011, and content was assessed as up-to-date to May 2009.  I tried to see whether there was a more recent review.

Step 2: TripDatabase

I conducted a search using the TripDatabase PICO search tool (Participant: “NSAID”, Intervention: “PPI”, Comparator: “placebo”, Outcomes: blank).  No newer systematic review was identified.  A useful clinical review article on the topic from 2013 still cited the Cochrane systematic review as the primary evidence [2].

Let’s look at the Cochrane systematic review by Rostom et al. (2002) in detail [1].


Critical appraisal

I will use the systematic reviews critical appraisal sheet from the Centre for Evidence Based Medicine [3].

What PICO question does the systematic review ask?

In people who had taken NSAIDs for greater than 3 weeks (Participants); what is the effect of PPIs (and also H2-antagonists or misoprostol), used as prophylaxis (Intervention); compared to placebo (Comparator); on the primary outcome of endoscopic ulcers, and clinical ulcer complications (Outcome).

Is it clearly stated?


Is it unlikely that important studies were missed?

Probably.  Like most Cochrane systematic reviews, the search strategy was exhaustive, involving multiple electronic databases, and is well documented.  The systematic review was assessed as up-to-date to 2009 – it is possible that there may be newer primary research not included.  It should be noted that a funnel plot comparing PPI vs placebo (figure 2, p. 8) [1] possibly indicates some publication bias favouring PPIs.

Were the criteria used to select articles for inclusion appropriate?

Yes.  The authors only included randomised trials.

Were the included studies sufficiently valid for the question asked?

Probably/possibly.  The authors formally assessed the risk of bias of the included studies using Jadad’s scale (see Stat Facts) [4].  Of the 6 studies in the systematic review that looked at PPIs, all but one had a score less than three, but, only one had a score greater than 3.

Were the results similar between studies?

Yes.  There was minimal heterogeneity measured (I2 = 0% for multiple outcome comparisons) between the studies for PPI vs placebo on total endoscopic ulcers.  All studies demonstrated uniform benefit from the PPIs.


What were the results?

Comparing PPI vs placebo on total endoscopic ulcers for patients taking NSAIDs:

  • 8 weeks or longer: risk ratio 0.34 (95% CI 0.28 to 0.42), P < 0.00001

Other results:

  • PPIs might cause more diarrhoea: risk ratio 1.66 (95% CI 0.85 to 3.22), P = 0.13
  • PPIs probably reduce dyspepsia: risk ratio 0.50 (95% CI 0.30 to 0.82), P = 0.0059

One small study looked at “clinical ulcers”.  Although a statistically significant result wasn’t found, clinical ulcers were observed only in the placebo group, with none in the PPI group (4 vs 0) [5].


Discussion and conclusion

The clinical evidence points clearly towards PPIs having a gastro-protective effect in individuals taking long-term NSAIDs, insofar as endoscopic ulcers.  My rough calculation based on the data in this paper is that there is an NNT (number-needed-to-treat) value of 5, to prevent one endoscopic ulcer in patients treated with NSAIDs for 8 weeks or longer.

However, it is crucial to recognise that endoscopic ulcers are very common in long-term NSAID use (up to 40%) [6] and the majority of these will never present clinically [1].

Guidelines around the topic are a little vague.  NICE guidance documents are relatively assertive in recommending co-prescribing of PPIs with NSAIDs, especially for people with arthritis [7].  eTG Complete recommends that prophylaxis be considered for patients with “risk factors for increased gastrointestinal toxicity” [8].  In the absence of large randomised trials using clinical outcomes [2], this may be a pragmatic recommendation.

Josef had low gastrointestinal risk so I opted not to start a PPI.


Stat Facts

Jadad’s scale

This instrument is commonly used to assess the methodological quality of randomised trials [4]. It rates three domains, randomisation, blinding, and patient withdrawals/dropouts, producing a score from 0 (worst) to 5 (best).  A Jadad score of 3 is often used as a threshold of “reasonable” quality, though whether 3 is included or excluded varies between publications.




  1. Rostom A, Dube C, Wells G, et al. Prevention of NSAID-induced gastroduodenal ulcers. Cochrane database of systematic reviews 2002(4):CD002296.
  2. Scheiman JM. The use of proton pump inhibitors in treating and preventing NSAID-induced mucosal damage. Arthritis Res Ther 2013;15 Suppl 3:S5.
  3. Centre for Evidence Based Medicine. Systematic Review: Are the results of the review valid? Oxford: University of Oxford, 2005.
  4. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996 Feb;17(1):1-12.
  5. Lai KC, Lam SK, Chu KM, et al. Lansoprazole reduces ulcer relapse after eradication of Helicobacter pylori in nonsteroidal anti-inflammatory drug users–a randomized trial. Aliment Pharmacol Ther 2003 Oct 15;18(8):829-36.
  6. Stalnikowicz R, Rachmilewitz D. NSAID-induced gastroduodenal damage: is prevention needed? A review and metaanalysis. J Clin Gastroenterol 1993 Oct;17(3):238-43.
  7. Non-steroidal anti-inflammatory drugs [key therapeutic topic]. National Institute for Health and Care Excellence. Published: 2015/1. Updated: 2017/1. Accessed: 2017/6/15. Available:
  8. Principles of analgesic and anti-inflammatory drug use for musculoskeletal conditions in adults. eTG complete (Rheumatology), eTG March 2017 edition. Therapeutic Guidelines Ltd.  Accessed: 2017/6/15.

Leave a Reply