2012-03-12 18:04:02 UTC

Is It Time to Dismiss the PPI Safety Concerns?

Aug. 2, 2011

Yu-Xiao Yang, Md, MSCe, FACP

Yu-Xiao Yang, MD, MSCE, FACP

Assistant Professor of Medicine and Epidemiology, Division of Gastroenterology, Center for Clinical Epidemiology and Biostatistics, and Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia


PPIs have been widely used since their introduction in the late 1980s because they are highly effective for acid-related conditions. However, potential safety concerns related to these drugs have started to emerge in recent years. These concerns have drawn much attention from the medical community and the general public. There is no doubt that the catchy headlines put out by the lay media have amplified the issue, but the primary reasons for the general interest in this issue are:

  • 1) The enormous popularity of PPIs and their widely recognized overuse in practice.
  • 2) The significant physiologic changes known to be induced by PPI therapy.
  • 3) The considerable morbidity and mortality associated with the potential adverse effects.

PPIs are among the most commonly prescribed classes of medications, with more than a hundred million prescriptions issued each year in the U.S. alone.1 Furthermore, recent data indicated that a large proportion of the PPI prescriptions in practice were written for either inappropriate or unclear indications.2 An all-too-familiar scenario is when a course of PPI therapy is started empirically for non-specific GI symptoms, but the patient is left on it indefinitely. An additional complicating factor that may make withdrawal of empirical PPI therapy difficult is the hypothesized rebound hypersecretion phenomenon, which may lead to de novo dyspeptic symptoms after as little as four weeks of standard dose PPI therapy.3 Furthermore, among patients with a proper indication for PPI therapy, too often patients are left on high-dose or continuous therapy unnecessarily despite data suggesting that step-down or on-demand therapy may be effective in appropriate patients.4, 5

I often hear from some of my esteemed colleagues or read in prominent medical journals comments proclaiming that the safety concerns associated with PPI therapy, whether real or not, are clinically irrelevant simply because the reported overall risk estimates for some of the adverse effects were modest. The associated “large” number needed to harm (NNH) based on average population baseline risk is sometimes presented to illustrate this point. However, this approach is clearly inadequate for a proper assessment of the clinical and public health relevance of any potential safety-related associations. First, the public health impact of a potential drug safety concern must be considered in the context of the prevalence of use. For example, a hypothetical medication that could kill one of every two people taking it once, i.e., a lethal NNH of two per dose, would have no real public health impact if nobody actually uses it in practice. On the other hand, a medication that is associated with a “large” NNH of 2,000 per year of use for a serious adverse event (e.g., pneumonia, hip fracture) may be an important public health concern if the medication is used by millions of people on a long-term or indefinite basis. Furthermore, the relative risk may be even higher (and the NNH might thus be much smaller) with high-dose and/or longterm therapy. In addition, depending on the specific baseline risk of an individual patient, the NNH for him/her may be higher or lower than that derived from an average population background risk.

While it is obvious that the reported potential adverse effects associated with PPI therapy — even at modest to moderate magnitudes — are important public health concerns, what is unclear is whether these potential associations are causal. A number of physiologic changes are known to be induced by PPI therapy. PPIs are highly potent acid suppressants, and the strong acid-suppressive effect is maintained as long as the PPI therapy is continued. Furthermore, the potency of acid suppression induced by PPI therapy may be accentuated among the elderly due to potentially decreased PPI clearance and Helicobacter pylori infection. Because the PPIs are such potent inhibitors of acid secretion, they cause a significant increase in serum gastrin levels in the vast majority of patients receiving long-term therapy. In addition, the interaction of the PPIs with CYP2C19 metabolism also appears to be a real pharmacodynamic phenomenon. Except for the few idiosyncratic reactions (e.g., acute interstitial nephritis), nearly all of the potential PPI-related safety concerns arise from these well-established physiologic changes. They generally fall into three major categories:

  • 1) The direct effect of gastric acid suppression itself (e.g., vitamin B12 deficiency, pneumonia, enteric infections &mdash including Clostridium difficile-associated disease [CDAD] &mdash mineral malabsorption, etc.).
  • 2) The physiologic response to the acid suppression (e.g., hypergastrinemia leading to increased cancer risk or hyperparathyroidism, rebound hypersecretion, etc.).
  • 3) The pharmacodynamic interaction with the metabolism of other medications (i.e., PPI and clopidogrel interaction).

Overall, there is sufficient biologic plausibility to justify a careful investigation to determine whether these theoretical mechanistic connections translate into clinically important adverse effects. Almost the entire epidemiologic evidence base regarding the safety of PPI therapy to date is composed of retrospective non-randomized studies. Post-marketing surveillance is a critical part of drug safety evaluation. The post-marketing surveillance mechanism generally consists of spontaneous adverse event reporting systems and formal Phase IV studies. Spontaneous reporting systems are inexpensive, vital for hypothesis generation and rarely may be sufficient for some regulatory actions (e.g., PPI-associated hypomagnesemia). Although post-marketing randomized controlled trial (RCT) data to inform safety evaluation are occasionally available (e.g., the Clopidogrel and the Optimization of Gastrointestinal Events trial for the PPI and clopidogrel interaction, de novo dyspepsia after PPI withdrawal), observational studies involving large populations with extensive person years of drug use are the most common type of Phase IV studies conducted. For some of the safety outcomes, observational studies showed either consistently no effect (e.g., colon cancer) or highly conflicting results (e.g., community-acquired pneumonia). For some of the other PPI safety outcomes (e.g., hip fractures, CDAD), the majority of observational studies showed a consistently positive association. Due to potential unmeasured and residual confounding, a positive association observed from such studies &emdash particularly when the effect size is relatively modest &emdash cannot be taken as definitive evidence supporting a causal relationship. The evidence supporting other safety outcomes is also less than definitive.

So what does this all mean for researchers and practitioners moving forward? In my opinion, unless there are consistently negative observational data (e.g., colorectal cancer), simply dismissing the potential risk in the face of clearly suggestive and plausible, but methodologically limited, epidemiologic evidence cannot possibly be a sensible approach. I believe the rational course of action should be to pursue more definitive evidence through carefully designed clinical research. Future studies should move beyond simply looking at the gross epidemiologic associations or generating uninterpretable summary estimates from the almost invariably heterogenous pool of observational studies. They should focus on the effect of PPI therapy on biomarkers relevant to the underlying mechanisms. The preferred study design would be either a prospective cohort design to allow better control of potential confounders or, if possible, RCTs.

Regarding the issue of fracture risk, the recent study from Canada using bone mineral density (BMD) data is a good example of such efforts.6 Although it was limited by the use of a convenience sample and a crude BMD assessment approach, it represents an important step in the right direction. We are currently conducting an NIH-funded prospective cohort study comparing the volumetric BMD measured by peripheral quantitative CT between long-term PPI users and nonusers with tight control of calcium and vitamin D intake. I believe this type of study can address the methodological limitations inherent to the existing observational studies and ultimately help unravel the nature of the observed epidemiologic link between PPI therapy and the adverse effects. If a true causal association is confirmed, such studies will also provide useful information regarding potential preventive measures.

While awaiting more definitive evidence, prescription of PPI therapy can be guided by several general principles to minimize potential risks. First, we should only use PPI therapy in patients who will benefit from it. On the other hand, no patients with proper indications for PPI therapy should be deprived of these highly effective medications. It is obvious from the existing evidence that the most consistent and largest increase in the risk of adverse effects is generally associated with long-term and/or high-dose PPI therapy. Therefore, in patients with appropriate indications for PPI therapy, one should always use the lowest effective dose. Unnecessary long-term and/or continuous therapy can be avoided by considering on-demand therapy in suitable patients and conducting periodic review of treatment indications.

I also provide all PPI users in my practice several recommendations regarding calcium intake that although unproven, are based on theoretical considerations of the relevant mechanisms and involve no added risk or cost. Specifically, since the actual daily calcium intake among the U.S. population across all ages is significantly lower than the recommended daily allowance, I instruct all my patients requiring PPI therapy to ensure that they take the recommended daily allowance of calcium and vitamin D appropriate for their age. Regarding the source of calcium intake, I recommend dairy products as the preferred source for most efficient calcium absorption, if possible. Waterinsoluble calcium carbonate supplement should be taken with a meal, which has been shown to reverse the calcium malabsorption in achlorhydric patients. Calcium citrate is also a good option, if financially feasible, because its absorption should not be affected by acid suppression whether taken with a meal or fasting.

1. Katz MH. Failing the acid test: benefits of proton pump inhibitors may not justify the risks for many users. Arch Intern Med 2010;170:747-8.

2. Forgacs I, Loganayagam A. Overprescribing proton pump inhibitors. BMJ 2008;336:2-3.

3. Reimer C, Sondergaard B, Hilsted L, Bytzer P. Proton-pump inhibitor therapy induces acidrelated symptoms in healthy volunteers after withdrawal of therapy. Gastroenterology 2009;137:80-7.

4. Inadomi JM, McIntyre L, Bernard L, Fendrick AM. Step-down from multiple- to single-dose proton pump inhibitors (PPIs): a prospective study of patients with heartburn or acid regurgitation completely relieved with PPIs. Am J Gastroenterol 2003;98:1940-4.

5. Metz DC, Inadomi JM, Howden CW, van Zanten SJV, Bytzer P. On-demand therapy for gastroesophageal reflux disease. American Journal of Gastroenterology 2007;102:642-53.

6. Targownik LE, Lix LM, Leung S, Leslie WD. Proton-pump inhibitor use is not associated with osteoporosis or accelerated bone mineral density loss. Gastroenterology 2010;138:896-904.

More on GERD

Psychotherapy helps patients with chronic digestive disorders

April 17, 2018

New AGA Clinical Practice Update explains benefits of psycho- gastroenterology for patients.

Course for APPs benefits all experience levels

April 16, 2018

AGA's popular course for advanced practice providers will take place Aug. 10 through 12 in Chicago, Illinois.

APPs: Enhance your delivery of patient care

March 28, 2018

AGA's popular course for advanced practice providers will be held this summer in Chicago.