A growing number of observational studies suggest that proton pump inhibitors (PPIs), one of the most commonly prescribed drug classes in the United States (1,2), may be associated with various adverse effects, including pneumonia (3), Clostridium difficile infection (4), bone fracture (5), and acute and chronic kidney diseases (6). While the clinical significance of these associations continues to be debated in the medical literature (7,8), lay media have prominently featured articles on possible adverse effects, often encouraging discontinuation or dose reduction (9–11). Moreover, the widely publicized Choosing Wisely campaign by the American Board of Internal Medicine also highlighted possible adverse effects of PPIs and suggested that most patients with gastroesophageal reflux disease (GERD) do not need them (12). In this survey of PPI users in the United States, we sought to evaluate the association between concern about PPI-related adverse effects and prior attempts to stop these medications. As secondary aims, we sought to evaluate patients' experiences discussing adverse effects with their physicians, knowledge of different PPI adverse effects, and willingness to try stopping their PPI in the future. These are consequential questions since they are likely to affect how patients use their PPI, whether or not they try to stop them, and whether or not it is appropriate to do so.
Study population and survey distribution
Our inclusion criteria were age ≥18 years, residence in the United States, use of a PPI for GERD, and having a primary care provider (PCP). Participants were identified using a commercial survey firm, Survey Sampling International (SSI, Shelton, CT). Survey Sampling International recruits members of its panel using multiple opt-in methods, including social media, e-mails, apps, and advertisements on television and websites in order to maximize panel diversity. It uses a probability-weighted random process, which accounts for known characteristics of panelists, to direct panel members to take specific surveys when they log on to the company's online portal. The survey remained active until at least 750 individuals had completed it. The survey became available online in July 2017 and closed within 1 month. Participants received a modest incentive through SSI. Full details of the sampling strategy can be found in Supplement 1 (Supplementary Digital Content, http://links.lww.com/AJG/A19).
Questionnaire development and content
The survey included 51 multiple-choice and free-response questions developed by the coauthors. It was pretested among four patients who used PPIs for GERD, identified through gastroenterology clinics at the University of Michigan. These patients were instructed to complete the survey while actively “thinking aloud” about how they understood the questions and any areas of ambiguity. In the survey, the term “long-term side effects” was used in place of “adverse effects” based on feedback. The survey was further refined based on feedback from members of the University of Michigan's Esophageal Disorders Program and the Center for Bioethics and Social Sciences in Medicine, which includes experts in survey design. Care was taken in ordering the survey questions to avoid introducing bias regarding awareness of adverse effects. The full survey can be found in Supplement 2 (Supplementary Digital Content, http://links.lww.com/AJG/A19).
For our primary outcome variable, patients were asked, “Have you ever tried stopping your [PPI name] because of concern about long term side effects?” with dichotomous (yes/no) response options. This was the second question in the survey following the screening questions. Respondents were also asked to rate their familiarity and concern about PPI adverse effects in general using a 4-point scale: “not at all,” “slightly,” “somewhat,” or “extremely.” The question regarding concern about adverse effects was the first question to follow the screening questions.
Patients were asked to write in any specific adverse effects that they believe had been associated with PPIs (“active recall”) and then later in the survey were asked to select from a list which specific adverse effects they were aware of having been associated with PPIs (“passive recall”). Finally, we asked respondents about prior discussion of PPI adverse effects with their providers and assessed patients' willingness to stop PPIs under varying conditions, including the type of provider recommending stopping (primary care provider vs gastroenterologist vs pharmacist) and different deprescribing strategies (tapering the PPI dose vs substituting an H2 blocker vs providing reassurance about ability to resume a PPI if needed). Response options were very unwilling, somewhat unwilling, somewhat willing, and very willing.
Additional data were collected regarding basic demographics, and frequency and duration of PPI use. Using a 5-point Likert-type scale, respondents were asked to rate their improvement in GERD symptoms after starting a PPI. We additionally queried non-GERD indications for PPI use, including history of peptic ulcer disease, esophagitis, Barrett's esophagus, gastrointestinal bleeding, and the daily use of antiplatelet drugs (including aspirin), anticoagulants, nonsteroidal anti-inflammatory drugs (NSAIDs), or corticosteroids. Not all results of the survey are included in this article.
Survey Sampling International panelists are subject to systematic quality control measures before inclusion in a survey sample. Such measures include digital fingerprinting to prevent duplicate responses and ongoing monitoring to identify inappropriately rapid question responses or inattention (13).
To investigate our primary question, we generated a multivariable logistic regression model to identify independent associations between having attempted to stop PPI because of concern about adverse effects (dependent variable) and the following prespecified variables: level of concern about adverse effects, a provider's recommendation to stop, and risk of upper gastrointestinal bleeding (UGIB), as well as age and gender. Patients were considered at high risk of UGIB if they reported using ≥weekly NSAIDs plus antiplatelet therapy (APT), dual-APT, or APT/NSAIDs plus either anticoagulation, prior peptic ulcer disease, or daily oral corticosteroid (14). Based on the regression results, we calculated predicted probabilities of stopping PPIs at different levels of concern about adverse effects and having received a recommendation to stop or not.
Patients' written responses about specific adverse effects associated with PPIs were reviewed manually and coded into the same 12 adverse effect categories queried in the question on passive recall. Responses that were overly general (e.g., “cancer” or “infection”), acute symptomatic reactions (e.g., “rash”), or not known to be associated with PPIs (e.g., “stomach ulcers”) were not further categorized (see coding schema in Supplement 3 [Supplementary Digital Content, http://links.lww.com/AJG/A19]). Responses to questions about willingness to stop the PPI under different clinical conditions were dichotomized as unwilling (somewhat/very) vs willing (somewhat/very) for the purposes of analysis, and differences in rates of willingness to stop under varying conditions were evaluated using chi-square tests.
The survey was administered using Qualtrics software (Provo, UT). Statistical analyses were carried out using Stata 15 (College Station, TX). This study was deemed exempt from Institutional Review Board review at the University of Michigan, which funded the study. Survey Sampling International had no input on survey design, content, or analysis.
Among the 1,431 individuals who answered the screening questions, 755 (53%) met all inclusion criteria and completed the survey. Ninety-three additional individuals opened the survey but did not complete any questions. Among excluded individuals, 11 were <18 years old, 604 reported not using a PPI for reflux, and 121 did not have a PCP; these categories were not mutually exclusive. A formal response rate could not be calculated because of the probability-weighted random process SSI uses for sampling, which does not allow for calculation of the number of SSI panel members eligible for the survey, and because of a cap on the number of participants.
Among patient participants, mean age was 49 years (s.d. 16), 71% were women, and 91% were white. Patient characteristics are summarized in Table 1. A majority (62%) used PPIs at least daily, and 73% reported taking PPIs for more than 2 years. The most commonly used PPIs were omeprazole (57%) and esomeprazole (16%). Subjects endorsed substantial symptom improvement on PPIs, with almost 90% reporting moderate to complete resolution of GERD symptoms. Eighty-nine percent of patients started a PPI based on a provider's recommendation, most commonly their PCP (77%). Nearly half of the patients (46%) reported seeing a gastroenterologist for GERD, but only 14% had been started on a PPI by a gastroenterologist. Secondary indications for using a PPI included peptic ulcer disease (15%), esophagitis (15%), and Barrett's esophagus (7%). Based on reported risk factors, 24% of patients were categorized as at high risk of UGIB.
Awareness and perception of PPI adverse effects
Twenty-nine percent of patients were slightly familiar, 20% somewhat familiar, and 5% extremely familiar with reports of PPI adverse effects, while 46% of respondents endorsed being “not all familiar.” Twenty percent of patients wrote in an adverse effect that has been associated with PPIs, most often kidney disease (6%), osteoporosis/osteopenia (5%), and dementia (5%; Figure 1). When asked to select all known adverse effects from a list (passive recall), patients most often selected chronic kidney disease (17%), vitamin D deficiency (15%), and bone fracture (12%; Figure 1). Fifty-four percent were not aware of any of the listed adverse effects being associated with PPIs. Nonetheless, most patients reported some degree of concern about PPI adverse effects, with 33% slightly, 32% somewhat, and 14% extremely concerned.
A provider had discussed the risks and benefits of PPI use with 24% of patients. Nine percent reported a provider had recommended that they stop their PPI. Most patients felt comfortable discussing whether to stop PPIs with their provider (83% somewhat or very comfortable).
Willingness to stop PPIs
In the base case, 71% of patients were willing to stop PPIs if recommended by a PCP, vs 77% if recommended by a gastroenterologist or 59% if recommended by a pharmacist (P < 0.001). Compared to the base case, willingness to stop increased to 79% if patients were allowed to stop PPI by switching to a less potent acid-blocking medication, 82% if they could resume the PPI if needed, and 83% if they could stop by tapering the PPI dose (P < 0.001).
Prior attempts to stop PPIs and association with concern about adverse effects
Thirty-nine percent of all patients reported a prior attempt to stop their PPI because of concern about adverse effects, and most of these (83%) did so without a provider's recommendation to stop. In multivariable regression, factors independently associated with an attempt to stop PPIs included: (i) provider recommendation to stop (odds ratio [OR] 3.26 [1.82–5.83]); (ii) concern about adverse effects (OR 5.13 [2.77–9.51] for slightly, 12.0 [6.51–22.2] for somewhat, and 19.4 [9.75–38.70] for extremely concerned); and (iii) female gender (OR 1.64 [1.12–2.39]; Table 2). Notably, UGIB risk was not inversely associated with attempts to stop (OR 0.98 [0.66–1.44]), such that individuals at high risk were no less likely to stop than others. Figure 2 displays predicted probabilities of having attempted to stop using a PPI at varying levels of concern about adverse effects and having received a provider's recommendation to stop.
In this national survey, we found that most patients using PPIs for GERD endorse some concern about adverse effects and that these concerns are strongly associated with prior attempts to stop these drugs. We also found that most patients who attempted to stop PPIs did so without the recommendation of their physician and that patients at high risk of UGIB because of pharmacologic and clinical risk factors, who benefit from ongoing PPI use (8,14), were as likely to have tried stopping as others. Taken together, these results suggest that patients are taking matters into their own hands to reduce what they perceive to be the risks of PPIs but without an appreciation of the full clinical picture. Importantly, if the wrong patients stop their PPIs, such as those at high risk of UGIB, they may suffer net harm from doing so.
The reason why such a large proportion of our sample had tried to discontinue their PPI without the recommendation of their providers requires further exploration. Patients appear to regard PPIs as a class of medications that may be self-managed, a perception that may be abetted by availability over the counter. The American Gastroenterological Association's contribution to the American Board of Internal Medicine Choosing Wisely campaign, which urges patients to reconsider the necessity of PPIs, may validate this tendency to self-titrate these drugs. Notably, few other recommendations in the Choosing Wisely campaign, which largely focused on reduced use of imaging and certain prescription medications, urged steps that patients can take on their own without the involvement of a physician. Patients should always be cautioned to talk with their providers before making any medication changes, even those involving over-the-counter drugs.
Only 24% of the patients in our sample reported that a provider had discussed the risks and benefits of PPIs with them. However, our findings provide reassuring evidence about the role providers can play to help appropriately selected patients stop PPIs. Patients whose provider had recommended that they stop were 3 times as likely to have attempted a trial off PPI. In addition, most patients endorsed a willingness to stop PPIs in the future if recommended by a PCP (71%) or gastroenterologist (79%). It is noteworthy that while 54% of patients endorsed some familiarity with reported PPI adverse effects, only 20% could actively recall specific adverse effects, suggesting that patients may benefit from education about the true nature of possible risks.
The main limitation of this study is that our sample of Internet users, while geographically diverse, may not be representative of the broader population of PPI users since they may have greater exposure to news media and specifically reports about PPI adverse effects. Internet-based surveys are also subject to self-selection bias based on topics of interest; however, relatively few individuals dropped out of the survey upon learning of the topic. Women were disproportionately represented in our survey, which has been found in the past with health-related Internet surveys and online health information seeking (15,16). These types of biases would likely have the greatest effect on prevalence estimates reported herein, but less effect on the outcomes of our primary analysis, which report the associations between variables (e.g., stopping a PPI and concern about adverse effects), since the mechanism of association is not hypothesized to differ between survey participants and nonparticipants. To our knowledge, this is the first study to evaluate the impact of concerns about PPI adverse effects on PPI use, including appropriateness of stopping.
Our findings should motivate providers to take several steps to ensure appropriate PPI use: (i) make explicit recommendations about the intended duration of use when a PPI is first started, a practice often overlooked (17); (ii) discuss the possible risks of PPI cessation for patients with definite indications, for example UGIB in patients on multiple antithrombotic drugs or stricture in patients with severe esophagitis; and (iii) engage in education and shared decision making with the nearly one half of patients on long-term PPI therapy without an indication to ensure that they truly choose wisely when they decide whether to continue PPIs (18,19).
CONFLICTS OF INTEREST
Guarantor of the article: Jacob E. Kurlander, MD, MS.
Specific author contributions: J.E.K. and J.K.K.: co-primary authors, study concept and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content. J.H.R., C.R.R., S.L.K., and R.D.V.: analysis and interpretation of data, critical revision of the manuscript for important intellectual content. S.D.S.: study concept and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content. Critical revision of the manuscript for important intellectual content and approval of final version: all.
Financial support: This project was supported by the University of Michigan. The contents do not represent the views of the U.S. Department of Veterans Affairs or the U.S. Government.
Potential competing interests: J.E.K. has received research funding from Ironwood Pharmaceuticals. S.D.S. has served as consultant for FMS.
WHAT IS KNOWN
- ✓ Proton pump inhibitors are one of the most commonly used medications in the United States.
- ✓ They have been associated with multiple possible adverse effects.
WHAT IS NEW HERE
- ✓ Most patients endorse some degree of concern about PPI-related adverse effects.
- ✓ Patient attempts to stop PPIs are strongly associated with concern about adverse effects, as well as a physician's recommendation to stop.
- ✓ Most patients who have attempted to stop PPIs did so without a provider's recommendation.
- ✓ Patients at high risk of UGIB were no less likely to try stopping PPIs than others.
1. Kantor ED, Rehm CD, Haas JS, et al. Trends in prescription drug use among adults in the United States from 1999–2012. JAMA 2015;314:1818–30.
3. Lambert AA, Lam JO, Paik JJ, et al. Risk of community-acquired pneumonia with outpatient proton-pump inhibitor therapy: A systematic review and meta-analysis. PLoS One 2015;10:e0128004.
4. McDonald EG, Milligan J, Frenette C, et al. Continuous proton pump inhibitor therapy and the associated risk of recurrent Clostridium difficile
infection. JAMA Intern Med 2015;175:784–91.
5. Zhou B, Huang Y, Li H, et al. Proton-pump inhibitors and risk of fractures: An update meta-analysis. Osteoporos Int 2016;27:339–47.
6. Lazarus B, Chen Y, Wilson FP, et al. Proton pump inhibitor use and the risk of chronic kidney disease. JAMA Intern Med 2016;176;238–46.
7. Vaezi MF, Yang YX, Howden CW. Complications of proton pump inhibitor therapy. Gastroenterology 2017;153:35–48.
8. Freedberg DE, Kim LS, Yang YX. The risks and benefits of long-term use of proton pump inhibitors: Expert review and best practice advice from the American Gastroenterological Association. Gastroenterology 2017;152:706–15.
14. Bhatt DL, Scheiman J, Abraham NS, et al. ACCF/ACG/AHA 2008 expert consensus document on reducing the gastrointestinal risks of antiplatelet therapy and NSAID use a report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents. Circulation 2008;118:1894–909.
16. Eysenbach G, Wyatt J. Using the internet for surveys and health research. J Med Internet Res 2002;4:e13.
17. Pasina L, Urru SaM, Mandelli S, et al. Evidence-based and unlicensed indications for proton pump inhibitors and patients' preferences for discontinuation: A pilot study in a sample of Italian community pharmacies. J Clin Pharm Ther 2016;41:220–3.
18. Batuwitage BT, Kingham JGC, Morgan NE, et al. Inappropriate prescribing of proton pump inhibitors in primary care. Postgrad Med J 2007;83:66–8.
19. Leri F, Ayzenberg M, Voyce SJ, et al. Four-year trends of inappropriate proton pump inhibitor use after hospital discharge. South Med J 2013;106:270–3.
Supplemental Digital Content
© The American College of Gastroenterology 2019. All Rights Reserved.