Advertisement

Leaving No Stone Unturned in the Search for Adverse Events Associated With Use of Proton Pump Inhibitors

Published:April 29, 2020DOI:https://doi.org/10.1016/j.cgh.2020.04.053
      Proton pump inhibitor (PPI) therapy has improved the quality of life of numerous patients with acid-related disorders.
      • Moayyedi P.
      • Armstrong D.
      • Hunt R.H.
      • et al.
      The gain in quality-adjusted life months by switching to esomeprazole in those with continued reflux symptoms in primary care: EncomPASS--a cluster-randomized trial.
      There is no such thing as a free lunch, however, and this is exemplified by PPIs. When PPI therapy was introduced the focus was on monetary cost because these drugs were relatively expensive. PPI therapy now is mostly generic, with monetary cost no longer an issue, but another opportunity cost has emerged to take its place. The current concern with PPI therapy is that a plethora of studies have suggested these drugs are associated with long-term harm.
      • Vaezi M.
      • Yang Y.-X.
      • Howden C.W.
      Complications of proton pump inhibitor therapy.
      It seems that every few months a new issue arises, with the list of problems that PPIs might cause becoming ever longer, including pneumonia, fracture, heart disease, Clostridium difficile–associated diarrhea, dementia, chronic kidney disease, low B12 levels, gastric cancer, and even all-cause mortality.
      • Elias E.
      • Targownik L.E.
      The clinicians guide to proton pump inhibitor related adverse events.
      In this issue of Clinical Gastroenterology and Hepatology, Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      raise a new concern that PPI therapy, and acid suppression in general, may be associated with kidney stones. This was a well-conducted study evaluating more than 465,000 participants from the Women’s Veteran’s cohort study. Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      found that the PPI users had an increased risk of developing kidney stones with a mean follow-up period of more than 5 years, and more than 2.5 million patient-years of follow-up evaluation in this retrospective cohort study. Overall, 2.4% of the cohort developed kidney stones, with PPI users having an increased risk resulting in a hazard ratio (HR) of 1.74 (95% CI, 1.67–1.82). This translates to a number needed to harm (NNH) of 365 patients who need to take PPIs for 1 year to observe 1 extra episode of kidney stones. The study noted that the risk of kidney stones increased with each 30-day increase in defined daily doses of a PPI in the 3 months before the diagnosis of nephrolithiasis. The study found an almost identical effect with H2-receptor antagonist use, suggesting this may relate to acid suppression rather than a specific effect of PPI therapy.
      These data are derived from a well-conducted cohort study and it appears that this is another concern that clinicians should have before prescribing PPI therapy. Indeed, Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      should be congratulated on a thorough analysis of the data and a thoughtful discussion of what their results may mean. There is an increasing trend for investigators to use the discussion section of PPI harms reports to briefly mention any weaknesses in the study design, but then pivot quickly to why these are unlikely to apply in their case and congratulate themselves on the conduct of their research and caution on the use of PPI therapy. This is not the case with this article, which appropriately discusses the issues and concludes that “the effects are small and should not change (PPI) prescribing in most patients.”
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      What are the concerns with these data? Should this change our prescribing patterns in any patient and is the association between PPI therapy and kidney stones likely to be causal?
      Observational studies can be powerful tools to evaluate factors that may protect or cause a disease, but they have limitations. In particular, any result found in observational studies could be the result of bias and/or confounding and not owing to the factor causing or protecting against a disease. There are approaches that can mitigate against confounding or bias, but they can never eliminate these issues, so researchers almost never can claim cause and effect on the basis of 1 observational study. PPI therapy provides an excellent example of the problems encountered when attempting to make a causal inference from observational data. Studies usually show that patients who are prescribed PPIs are, on average, sicker than those who are not taking these drugs.
      • Moayyedi P.
      • Leontiadis G.I.
      The risks of PPI therapy.
      This is illustrated by Simonov et al,
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      in which patients on PPI therapy tended to be older and have more comorbidities including diabetes mellitus, which may predispose them to develop kidney stones. Of course, Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      adjusted for these confounding factors, but this highlights the limitation of this approach. The unadjusted HR for PPI and kidney stones was 1.74 (95% CI, 1.67–1.82), the adjusted HR was 1.46 (95% CI, 1.28–1.56), and the propensity matched analysis HR was 1.25 (95% CI, 1.19–1.33): with each degree of adjustment for confounding factors the HR becomes closer to unity. Although the HR always remains above 1, if is likely that if all confounders could be identified and adjusted the HR may not be statistically significant. This phenomenon is seen for many of the other harms associated with PPI therapy,
      • Moayyedi P.
      • Eikelboom J.W.
      • Bosch J.
      • et al.
      Safety of proton pump inhibitors based on a large, multi-year, randomized trial of patients receiving rivaroxaban or aspirin.
      and some, if not all, of the associations these studies report may relate to residual confounding.
      Biases are another issue with observational studies that can make the results difficult to interpret. An example of this is the association between PPI and pneumonia. The effect was strongest within the first week of prescription when the odds ratio was approximately 4, although this was reduced to approximately 1.5 after 1 month.
      • Moayyedi P.
      • Leontiadis G.I.
      The risks of PPI therapy.
      This marked reduction in risk over a relatively short period of time is not biologically plausible and a more likely explanation is that the association is the result of protopathic bias.
      • Faillie J.-L.
      Indication bias or protopathic bias?.
      Patients presenting to the clinician with a cough may be diagnosed with silent reflux and given a PPI. A few days later other symptoms develop, and pneumonia is made as the final diagnosis. This will not be apparent when simply interrogating a database where the researcher will observe that a PPI was prescribed before the onset of pneumonia and will imply the association is causal when this is not the case. The same may apply to the association between PPI therapy and kidney stones. Patients may present with abdominal pain and be given a PPI as a therapeutic trial, assuming the pain may be acid-related when subsequently it is found that the pain relates to kidney stones. Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      reported that 3% were prescribed PPIs within 1 month of the diagnosis of kidney stones, but did not provide the analysis that would allow us to interpret whether protopathic bias may play a role in the associations observed. It is interesting that the dose response is seen if PPI therapy is evaluated over 3 months before the diagnosis, but this is almost lost if the evaluation period is over 12 months.
      Randomized placebo-controlled trials (RCTs), when rigorously conducted, are free from confounding and bias and are the most powerful design to determine whether PPIs may cause long-term harm. An RCT
      • Moayyedi P.
      • Eikelboom J.W.
      • Bosch J.
      • et al.
      Safety of proton pump inhibitors based on a large, multi-year, randomized trial of patients receiving rivaroxaban or aspirin.
      enrolled 17,598 participants, randomized to pantoprazole 40 mg once daily or placebo, and followed them up for a mean of 3 years. The primary aim of the trial was to evaluate whether PPI therapy protected against serious upper gastrointestinal events in patients taking aspirin and/or apixaban,
      • Moayyedi P.
      • Eikelboom J.W.
      • Bosch J.
      • et al.
      Pantoprazole to prevent gastroduodenal events in patients receiving rivaroxaban and/or aspirin in a randomized, double-blind, placebo-controlled trial.
      but a defined secondary aim was to evaluate whether PPIs caused any harm. Patients were interviewed every 6 months and there was no difference in pneumonia, Clostridium difficile infection, other enteric infections, fracture, gastric atrophy, chronic kidney disease, dementia, cardiovascular disease, cancer, hospitalizations, and all-cause mortality in the PPI compared with the placebo arms. For some outcomes, such as C difficile infection, there were insufficient numbers of events to be certain PPIs caused no harm, but for many events any serious harm was unlikely. The only exception was enteric infections, which were slightly more common in patients randomized to PPIs, but even there the NNH was more than 900 per year. This RCT highlights that we should treat observational data cautiously. This more rigorous design does not support the observational data and suggests that in most cases small increases in risk seen in observational studies relate to residual confounding, bias, or both. This is particularly true when there are imbalances between patients taking PPI therapy and nonusers, which cannot fully be adjusted for because there is imperfect information of cases and controls. This type of data reported by Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      should not change clinical practice, as the authors acknowledge. Although it remains possible that PPIs cause kidney stones, it is unlikely, and if true any effect is likely to be extremely small. The NNH given in the article was 365, but this used an unadjusted HR, if the adjusted HR is used with US data for the incidence of kidney stones
      • Kittanamongkolchai W.
      • Vaughan L.E.
      • Enders F.T.
      • et al.
      The changing incidence and presentation of urinary stones over 3 decades.
      the NNH was 1550, a very small effect. The article by Simonov et al
      • Simonov M.
      • Abel E.A.
      • Skanderson M.
      • et al.
      Use of proton pump inhibitors increases risk of incident kidney stones.
      is interesting and well done, but should not change our approach to PPI prescribing. As with all medications, PPI therapy should be prescribed only for evidence-based indications,
      • Moayyedi P.
      • Lacy B.E.
      • Andrews C.N.
      • et al.
      ACG and CAG clinical guideline: management of dyspepsia.
      at the lowest dose and shortest duration possible.
      • Boghossian T.A.
      • Rashid F.J.
      • Thompson W.
      • et al.
      Deprescribing versus continuation of chronic proton pump inhibitor use in adults.

      References

        • Moayyedi P.
        • Armstrong D.
        • Hunt R.H.
        • et al.
        The gain in quality-adjusted life months by switching to esomeprazole in those with continued reflux symptoms in primary care: EncomPASS--a cluster-randomized trial.
        Am J Gastroenterol. 2010; 105: 2341-2346
        • Vaezi M.
        • Yang Y.-X.
        • Howden C.W.
        Complications of proton pump inhibitor therapy.
        Gastroenterology. 2017; 153: 35-48
        • Elias E.
        • Targownik L.E.
        The clinicians guide to proton pump inhibitor related adverse events.
        Drugs. 2019; 79: 715-731
        • Simonov M.
        • Abel E.A.
        • Skanderson M.
        • et al.
        Use of proton pump inhibitors increases risk of incident kidney stones.
        Clin Gastroenterol Hepatol. 2021; 19: 72-79
        • Moayyedi P.
        • Leontiadis G.I.
        The risks of PPI therapy.
        Nat Rev Gastroenterol Hepatol. 2012; 9: 132-139
        • Faillie J.-L.
        Indication bias or protopathic bias?.
        Br J Clin Pharmacol. 2015; 80: 779-780
        • Moayyedi P.
        • Eikelboom J.W.
        • Bosch J.
        • et al.
        Safety of proton pump inhibitors based on a large, multi-year, randomized trial of patients receiving rivaroxaban or aspirin.
        Gastroenterology. 2019; 157: 682-691
        • Moayyedi P.
        • Eikelboom J.W.
        • Bosch J.
        • et al.
        Pantoprazole to prevent gastroduodenal events in patients receiving rivaroxaban and/or aspirin in a randomized, double-blind, placebo-controlled trial.
        Gastroenterology. 2019; 157: 403-412
        • Kittanamongkolchai W.
        • Vaughan L.E.
        • Enders F.T.
        • et al.
        The changing incidence and presentation of urinary stones over 3 decades.
        Mayo Clin Proc. 2018; 93: 291-299
        • Moayyedi P.
        • Lacy B.E.
        • Andrews C.N.
        • et al.
        ACG and CAG clinical guideline: management of dyspepsia.
        Am J Gastroenterol. 2017; 112: 988-1013
        • Boghossian T.A.
        • Rashid F.J.
        • Thompson W.
        • et al.
        Deprescribing versus continuation of chronic proton pump inhibitor use in adults.
        Cochrane Database Syst Rev. 2017; 3: CD011969

      Linked Article