Use of gastric acid–suppressive agents increases the risk of dementia in patients with upper gastrointestinal disease: A population-based retrospective cohort study

Background Prescriptions for gastric acid–suppressive agents, including proton-pump inhibitors (PPIs) and histamine type-2 receptor antagonists (H2RAs), are rising. However, little data exist regarding their association with dementia in the Asian population. The objective of this study was thus to investigate the impact of the use of PPIs and H2RAs on the risk of dementia in an Asian population with upper gastrointestinal disease (UGID). Methods We conducted a population-based retrospective cohort study with a 10-year follow-up using data from 2000 to 2015 derived from Taiwan’s Longitudinal Health Insurance Database. We included 6711 patients with UGID receiving gastric acid–suppressive agents, 6711 patients with UGID not receiving agents, and 6711 patients without UGID or treatment thereof, all at least 20 years of age. Groups were matched for age, sex, and index date. The association between gastric acid–suppressive agent use and dementia was analyzed using a Cox proportional hazards regression model adjusted for potential confounders. Results The adjusted hazard ratio (aHR) of dementia for patients with UGID receiving gastric acid–suppressive agents compared with patients with UGID without gastric acid–suppressive agents was 1.470 (95% confidence interval [CI] 1.267–1.705, p < 0.001). Both PPIs and H2RAs increase the risk of dementia (PPIs: aHR 1.886 [95% CI 1.377–2.582], p < 0.001; H2RAs: aHR 1.357 [95% CI 1.098–1.678], p < 0.01), with PPIs exhibiting significantly greater risk (aHR 1.456 [95% CI 1.022–2.075], p < 0.05). Conclusions Our results demonstrate an increased risk of dementia in patients with UGID receiving gastric acid–suppressive agents, including PPIs and H2RAs, and the use of PPIs was associated with a significantly greater risk than H2RA use.


Introduction
According to the World Health Organization, dementia is a syndrome usually of a chronic or progressive nature caused by a variety of brain illnesses that affect memory, thinking, behavior, and the ability to perform everyday activities. In 2015, the number of people with dementia worldwide was estimated to be 47 million; this number is expected to increase to 75 million in 2030 and 132 million in 2050 [1]. The disease has an enormous socioeconomic cost. In 2010, the estimated worldwide cost of dementia was US$604 billion; it was US$818 billion in 2015 and is set to increase in coming years [1,2]. Hence, factors that potentially increase the risk of dementia must be identified.
The rising number of prescriptions for gastric acid-suppressive agents including protonpump inhibitors (PPIs) and histamine type-2 receptor antagonist (H2RAs), the standard treatment for upper gastrointestinal disease (UGID), has prompted research regarding whether either medication increases the risk of dementia. One study reported that PPIs are associated with greater risk of dementia in Germans aged 75 and older [3], and another reported an increased risk for dementia among Asian PPI users [4]. A third study determined that H2RA use more than doubles the risk of cognitive impairment in African Americans [5].
Because the previous study of an Asian population investigated only the effect of PPIs on the risk of dementia in the general population, we conducted a retrospective cohort study to investigate the effect of gastric acid-suppressive agents, including both PPIs and H2RAs, on the risk of dementia among an Asian population with UGID.

Data source
Analyses were conducted using a longitudinal sample of patients from Taiwan's Longitudinal Health Insurance Database (LHID), a national database containing 2 million enrollees randomly selected from the registry of National Health Insurance (NHI) beneficiaries in Taiwan

Study patients
This was a retrospective cohort study. We aggregated the data with a 5-year baseline from 2001 to 2005 and followed up by 10-year intervals for each individual. We identified patients newly diagnosed with UGID (ICD-9-CM codes 530-536) between January 1, 2001, and December 31, 2005. Two consecutive diagnoses were required to enhance diagnosis validity. For each patient, we assigned the first date of receiving gastric acid-suppressive agents for treatment of UGID as the index date. We also identified patients without UGID between January 1, 2000, and December 31, 2015, as a comparison group. The records of drug prescriptions were reviewed to determine which individuals had ever received prescriptions for gastric acidsuppressive agents during 2001-2015 after the respective index dates. We classified patients with UGID into two groups: those who received gastric acid-suppressive agents and those who did not. In the group of patients with UGID receiving gastric acid-suppressive agents, we excluded patients who had been diagnosed with dementia before the index date. We also excluded patients with UGID receiving gastric acid-suppressive agents after December 31, 2005, and those receiving gastric acid-suppressive agents for less than 90 days within 365 days after first being administered the agents. Patients with UGID receiving gastric acid-suppressive agents were selected as study group I, and patients with UGID but not receiving gastric acid-suppressive agents were study group II. Each patient in study group I was matched to one in study group II by age, sex, and index year. Patients without UGID were selected as a comparison group. Each patient in study group II was matched to one patient in the comparison group by age, sex, and index year. The outcome of interest was dementia defined as a diagnosis (ICD-9-CM codes 331.0, 331, 290.0, 290.20, 290.21, 290.3). We identified patients with at least two diagnoses in the follow-up period. All individuals were followed for 10 years or censored at the date of dementia diagnosis.

Medication usage and other relevant information
Study group I received gastric acid-suppressive agents before December 31, 2005. Use of gastric acid-suppressive agents was defined as a prescription for at least one type of PPI (omeprazole, esomeprazole, lansoprazole, dexlansoprazole, pantoprazole, or rabeprazole) or H2RA (cimetidine, ranitidine, famotidine, nizatidine, or roxatidine) for at least 90 days within 365 days after first being administered one of these agents.

Statistical analysis
Statistical analyses were performed using SAS version 9.4. Student's t test and Pearson's chisquared test were applied to evaluate differences in baseline characteristics such as age, sex, comorbidities, and drugs among different groups. Cox proportional hazard ratios (HRs) were used to estimate HRs and 95% confidence intervals (CIs). The Kaplan-Meier method and logrank test were used to examine the differences in 10-year dementia occurrence rates between the study groups and comparison group. The p values were two-sided, and p < 0.05 was considered statistically significant.

Results
We identified 506 668 patients newly diagnosed with UGID and 823 903 patients without UGID between January 1, 2001, and December 31, 2005. Fig 1 summarizes the enrollment criteria for study group I, study group II, and the comparison group. The latter two groups were 1:1 matched with group I by age, sex, and index year. We identified 6711 patients with UGID receiving gastric acid-suppressive agents as study group I, 6711 patients with UGID but not receiving gastric acid-suppressive agents as study group II, and 6711 patients without UGID as a comparison group. Table 1 lists the demographic characteristics of patients with and without UGID. After matching, the distributions of age and sex were similar among the three groups. The mean age was 55.02 ± 15.67 years, and 57.06% of the individuals were male. We estimated the crude HRs and adjusted HRs (aHRs) for the risk of dementia between study group I and study II and between study group II and the comparison group. During the 10-year follow-up, the aHR of dementia for study group I compared with study group II was 1.  Table 3).
The Kaplan-Meier curve of the cumulative occurrence of dementia among the three groups revealed that patients with UGID receiving gastric acid-suppressive agents had the highest cumulative incidence of dementia (Fig 2).

Discussion
This is the first retrospective cohort study to evaluate the risk of both PPIs and H2RAs on dementia in Asian patients with UGID. Our results demonstrate that the use of gastric acidsuppressive agents, both PPIs and H2RAs, increases the risk of dementia in patients with UGID and the use of PPIs involves significantly greater risk than the use of H2RAs.
Several mechanisms are related to our finding. For PPIs, first, the use of PPIs is associated with a higher β-amyloid (Aβ) level in the brain by affecting the enzymes β-and γ-secretases [6,7], which are involved in the pathogenesis of Alzheimer disease, the most common type of dementia. Second, PPIs may reduce the degradation of Aβ by lysosomes in microglia by inhibiting V-ATPases, thus increasing Aβ levels [8,9]. For H2RA, the anticholinergic effect caused by certain H2RAs, such as ranitidine and cimetidine, may lead to delirium and other cognitive deficits [10][11][12]. For both PPIs and H2RA, two kinds of medications could result in vitamin B12 deficiency, which is associated with cognitive decline [13]. In addition, gut microbiota play a major role in bidirectional communication between the gastrointestinal tract and central nervous system-called the microbiota-gut-brain axis. Thus, alteration of the composition of gut microbiota caused by the disruption of the low-pH environment by gastric acid-suppressive agents may negatively affect central nervous system function [14,15].
Our result can be supported by Makunts T, et al. study [16], which use data base from FDA adverse event reporting system, showed that PPI use significantly increase the memory impairment risk compared to H2RA use. However, in Makunts's study due to the limitations stem from the occasionally missing demographic variables, treatment duration and lack of medical record data, not only the baseline characteristics including age and sex were significantly different between two groups, but also the confounding factors such as concurrent medications and comorbidities were not adjusted, which may cause bias in statistical analysis and results. In addition, the indication bias caused by PPI and H2RA were appeared, thus the memory impairment caused by initial disease state cannot be ruled out.

PLOS ONE
Gastric acid-suppressive agents increases the risk of dementia, a retrospective cohort study Our study has several noteworthy strengths. First, it is the first cohort study to evaluate the impact of gastric acid-suppressive agents, both PPIs and H2RAs, on dementia risk in Asian patients based on a large sample. Individuals over the age of 20 who met the selection criteria were included, and thus, the results are not limited to the elderly population. Second, excluding those receiving gastric acid-suppressive agents fewer than 90 days within 365 days after being first administered an agent may have ensured that all drug users had used a PPI or H2RA drug for a sufficient period of time (rather than accidentally), making the relationship between gastric acid-suppressive agents and dementia clearer. Third, a 10-year follow-up is in line with the pathogenesis of dementia and may have reduced the bias caused by censoring. Fourth, by using Taiwan's LHID, we could both perform the analysis under real-life circumstances and minimize selection and recall bias. Fifth, we have included UGID in the analysis to prevent indication bias from affecting the results. Finally, we adjusted for numerous potential

PLOS ONE
factors that influence the occurrence of dementia, including age, sex, alcohol abuse, tobacco dependence, comorbidities, and drug use.
Nevertheless, limitations remain. One is that factors such as educational attainment and the ApoE4 allele could not be estimated because the data source lacks sociodemographic and genetic records. Another is that information regarding patients' adherence to gastric acid-suppressive agents could not be evaluated; nevertheless, the definition of PPI and/or H2RA use was the use of PPIs and/or H2RAs for at least 90 days within 365 days after the first administration, and thus, these patients all exhibited long-term use of PPIs and/or H2RAs to treat UGID. The cause-consequence relationship between gastric acid-suppressive agents and the risk of dementia was greatly increased. In conclusion, we conducted a retrospective cohort study with 10-year follow-up using data derived from the LHID and confirmed that using gastric acidsuppressive agents, including PPIs and H2RAs, is associated with an increased risk of dementia. Thus, restrictive and appropriate use of PPIs and H2RAs may help prevent the occurrence of dementia, and patients should be educated on medication adherence to shorten the treatment period during which they take gastric acid secretion inhibitors. The possible underlying