Patterns of High-Dose and Long-Term Proton Pump Inhibitor Use: A Cross-Sectional Study in Six South Australian Residential Aged Care Services

Proton pump inhibitors (PPIs) are indicated for gastroesophageal reflux disease (GERD) and dyspepsia, Barrett’s esophagus, and to prevent adverse gastrointestinal effects of non-steroidal anti-inflammatory drugs (NSAIDs) [1‐4]. An estimated 5.5% of all Australians use PPIs, with the highest prevalence of use among those aged over 80 years [5]. Internationally, there is considerable variation in the prevalence of PPI use in residential aged care services (RACSs), ranging from 14 to 93% of residents [6‐8].

The Australian Therapeutic Guidelines and American Gastroenterological Association recommend PPI use be limited to 4–8 weeks’ duration for the symptomatic relief of uncomplicated GERD [9, 10]. If symptoms are not alleviated, the PPI dose may be increased or administered twice daily until symptoms are adequately controlled [10, 11]. However, periodic reassessment of the possibility to reduce the dose to the lowest effective dose once symptom relief is achieved is recommended. Long-term PPI use is not recommended for symptomatic relief of uncomplicated symptoms without further investigation, particularly if dose reduction or cessation is unsuccessful [9‐11]. However, prophylactic PPI use may be suitable for older people at high risk of an ulcer who take concomitant gastric irritant medications such NSAIDs, including aspirin [10]. Medications for symptom relief, such as PPIs, should also be routinely reviewed in people living with dementia because people living with dementia may under-report both disease-related symptoms and medication side effects [12].

Proton pump inhibitors are considered generally safe and well tolerated medications with infrequent adverse events. However, vitamin B12 deficiencies, hypomagnesaemia, Clostridium difficile infection, community-acquired pneumonia, fractures, and renal injury have been associated with PPI use [13]. Much of the evidence for these and other adverse events is limited or inconclusive. Nevertheless, frail older residents may be at increased risk of experiencing adverse events, particularly if high-dose PPIs are prescribed. Given the high prevalence of PPI use in RACSs, even rare adverse events may affect large numbers of residents [14].

Existing literature suggest older people may be prescribed PPIs for long periods and without clear indications for use [6]. Few studies have assessed duration of PPI use among residents of RACSs and whether residents have documented clinical indications for PPI use. A previous study utilizing data collected from 22 skilled nursing facilities in the USA found that one-quarter of residents prescribed a PPI did not have a valid indication for use and were not taking a concomitant NSAID or anticoagulant [7].

There is increasing focus on regularly reviewing the need for PPI treatment in older people. Over the last decade, multiple national quality improvement initiatives directed at healthcare professionals and consumers have been implemented to improve PPI use in Australia [14]. As part of the Choosing Wisely Australia initiative, the Gastroenterological Society of Australia and the Royal Australian College of General Practitioners both recommend that PPIs are not prescribed long term without attempting to reduce to the lowest possible dose or ceasing the PPI [15]. The objective of this study was to characterize PPI use. We also determined factors associated with high-dose PPI use among older adults in RACSs.

The 383 residents who participated in this study had an average age of 88 years, 76% were female, and 44% had a documented diagnosis of dementia (Table 2). The characteristics of residents who did and did not receive a PPI are compared in Table 2.

There were 196 (51%) residents who were prescribed a PPI. Of these, seven residents (3.6% of residents receiving a PPI) were prescribed a low-dose PPI, 143 (73.0%) received a standard dose and 45 (23%) residents received a high-dose PPI (Table 3). Pantoprazole and esomeprazole together represented approximately 80% of PPI use. Twice-daily administration was identified for 21 residents (10% of PPI users). The PPI was prescribed for regular use for all 196 residents, with no residents prescribed PRN PPIs.

Overall, 145 (74%) residents receiving a PPI had at least one documented clinical indication for PPI use, of which the most frequent was GERD (Table 3). Approximately half of all residents (55%) prescribed a PPI were also prescribed one or more medications associated with gastric irritation or increased risk of gastrointestinal bleeding. Neither an indication nor a medication associated with increased bleeding risk or gastric irritation was identified for 23 (12%) residents prescribed a PPI.

Among the 196 residents who received a PPI, the duration of PPI prescribing could be established for 147 (75%) residents based on the current medication chart. A PPI was taken for longer than 8 weeks in almost all residents (n = 141, 96%) where the current medication chart included at least 8 weeks of PPI prescribing data. Of the 124 residents with documented GERD or dyspepsia who were prescribed a PPI, 123 residents had sufficient information recorded on the medication chart to assess whether the PPI had been prescribed for more than 8 weeks, and three-quarters (n = 93, 76%) had received a PPI for more than 8 weeks (Table 3).

Overall, 142 (72%) residents receiving a PPI were able to self-report their willingness for medications to be deprescribed. Of these, few residents (8%) were comfortable with the number of medications taken and most residents (83%) indicated a willingness to stop one or more regular medications if their doctor said it was possible (Table 4).

Characteristics associated with high-dose PPI use compared to low or standard doses are shown in Table 5. A greater number of medications prescribed for regular administration and increasingly complex medication regimens were significantly associated with the use of a high-dose PPI. High-dose PPI use was not found to be associated with dementia, frailty, independence in ADLs, or increased NPI-NH scores in the age- and sex-adjusted analyses.

This study characterizing PPI use in 6 Australian RACSs showed that half of all residents receive a PPI on a regular basis. Our study findings are consistent with other research in RACSs suggesting that PPI use is highly prevalent [6‐8]. Furthermore, our finding that residents taking high-dose PPIs had high levels of polypharmacy and medication regimen complexity compared to low/standard-dose users is consistent with previous RACS research that found PPIs were taken by 72% of residents receiving nine or more regular medications compared to 36% receiving fewer than nine medications [26]. Previous research has predominantly investigated prevalence of PPI use in RACSs using defined daily doses or without assessing dose. We found that three in four residents who took a PPI received a standard dose. Less than 4% of residents who took a PPI received a low-dose PPI. It is also important for prescribers to document the indication and intended duration when prescribing a PPI for a resident of an aged care service. Nearly nine in ten residents taking a PPI had a documented indication consistent with PPI use. However, three-quarters of residents with GERD or dyspepsia charted a PPI had taken the PPI for more than 8 weeks. These findings are similar to a recent Canadian study involving 147 residents who were prescribed a PPI, of whom 62% of residents with GERD had received the PPI for more than 8 weeks, and one in five residents taking a PPI had no documented indication for PPI use [27].

Our findings suggest that stepping-down therapy to lower doses may be appropriate for up to one-quarter of residents who received a high-dose PPI. The finding that many residents with GERD or dyspepsia had received a PPI for more than 8 weeks also suggests there are opportunities for stepping-down therapy. Further, residents who took PPIs long term without an apparent indication and residents with documented dyspepsia or GERD who had received a PPI for more than 8 weeks may be suitable candidates for deprescribing [6, 28]. Deprescribing refers to “reducing medications after consideration of therapeutic goals, benefits and risks, and medical ethics” [29]. Deprescribing PPIs can be challenging, as abrupt cessation may result in rebound symptoms, possibly due to acid hypersecretion [30, 31]. Deprescribing for residents living with dementia can also be complicated by medical practitioner and systems-related barriers such as lack of time or resources, or fear of negative consequences [32]. Difficulties with decision-making, comprehension, and communication among residents with dementia may also limit deprescribing [32]. However, in our study, the residents’ willingness to consider deprescribing indicates that it may be feasible in this population to attempt PPI dose de-escalation or deprescribing. Residents taking PPIs in the present study also had high levels of polypharmacy and medication regimen complexity, and therefore could potentially benefit from deprescribing. Translational research could test and implement targeted strategies for stepping-down therapy among residents receiving high-dose PPIs and for deprescribing PPIs prescribed for treating GERD or dyspepsia for more than 8 weeks. This process could be informed by recommendations from both the Canadian guidelines for deprescribing PPIs and a recent systematic review [28, 33], and known resident, caregiver, and prescriber facilitators and barriers to deprescribing, particularly in people with dementia [34, 35], as well as previous deprescribing controlled trials in RACSs. Interdisciplinary approaches may be more likely to facilitate the optimization of PPI use among older people. For example, previous Australian research has indicated that repeated multifaceted quality improvement interventions targeting clinicians and older veterans taking a PPI to encourage PPI dose reductions have achieved sustained improvements in PPI use [14].

Study strengths include investigation of PPI use in RACSs, including medication selection and dose, within a modest sample size of residents. Our sample included residents from metropolitan and rural areas and was representative of all residents in the six RACSs from which the participating residents were recruited. The six RACSs were maintained by the same aged care provider organization, and therefore findings may not be generalizable to all Australians receiving residential aged care services. In Australia, 59% of residents of aged care services are aged 85 years and over, 68% of residents are female, and 52% of residents are living with dementia [36, 37].

To our knowledge, this is one of the first studies to investigate guideline-appropriate PPI use in RACSs that has considered PPI dose and duration. Because only three-quarters of residents receiving a PPI had medication charts with more than 8 weeks of prescribing data available, our study may have underestimated the proportion of residents using PPIs for longer than 8 weeks. The cross-sectional nature of this study also meant we could not assess causation or changes in PPI use over an extended period. We were unable to reliably assess previous attempts at dose changes or withdrawals as while we could extract the date the PPI was prescribed from some medication charts, we lacked continuous information on PPI dose and administration for all residents prior to study entry. We were also unable to determine the circumstances under which the PPI was originally prescribed and any adverse events experienced by study participants. Although older age is a risk factor for gastrointestinal bleeding associated with NSAIDs or anticoagulants [10], we were unable to determine whether PPI prophylaxis was warranted for all residents who received a gastric irritant medication or a medication associated with increased risk of gastrointestinal bleeding. There was also a small number of residents taking a PPI with a diagnosis of gastritis recorded in the medical history, and this was defined as a potential indication for PPI use in this study. We recognize that gastritis may not be an approved indication for PPI use in all settings, and therefore the proportion of residents who had at least one documented clinical indication for PPI treatment may be slightly overestimated. Finally, we explored associations using age- and sex-adjusted models, but our modest sample size prevented exploration using multivariable models, and therefore residual confounding may still exist.

Despite PPI use being highly prevalent, many residents taking a PPI had documented clinical indications for treatment. This is contrary to the popular belief that PPIs should be targeted for deprescribing due to lack of clear clinical indications in RACSs and could be further explored in future studies. However, there remains an opportunity to review the continuing need for PPI use among residents with GERD who are treated for more than 8 weeks, and identify if residents using higher PPI doses are suitable for “step-down” approaches to treatment. There is also a need for clinicians to specify when PPIs are prescribed for gastroprotection and clearly document whether indications for PPI use are still current.