Dementia is characterized by substantial and progressive decline in cognitive and functional abilities. It is a major burden on patients and their families and constitutes a huge public health concern due to the aging world population (1). Its estimated global prevalence is 36 million, and this is expected to more than double in 2 decades to around 80 million (2,3). This will, in turn, translate into enormous healthcare costs; the global expenditure for the management of dementia was $818 billion in 2015, representing a 35% increase from 2010 (4). Therefore, there is great interest in identifying risk factors associated with the development of dementia. Medication use among elderly patients may constitute a modifiable risk factor for the development of dementia.
Proton pump inhibitors (PPIs) have been central to the management of acid-related upper gastrointestinal disorders for the past 3 decades (5). PPIs are among the most commonly prescribed drugs worldwide. Up to 1 in 5 older adults takes PPIs—and frequently on a long-term basis (6,7). However, PPIs may be used inappropriately in up to 40%–50% of patients (8). PPIs have been overprescribed in a variety of healthcare settings including outpatient clinics, nursing homes, and rehabilitation centers (9). Many patients are started on PPI therapy during hospitalization and may continue to be prescribed it—often unnecessarily—after discharge (10). Consequently, PPI use among the elderly has been rising steadily (9).
In recent years, PPIs have been associated with various adverse events, although often without compelling evidence of causation (11). A German study (12) first pointed out a possible association between PPI use and an increased risk of developing dementia among the elderly. Subsequent studies have evaluated this association but have reached conflicting conclusions. If there were such an association, it would have major clinical and socioeconomic consequences. We therefore conducted a systematic review and meta-analysis to evaluate any possible association between PPI therapy and development of dementia.
We conducted this systematic review and meta-analysis based on the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (13) and Meta-analysis of Observational Studies in Epidemiology (14).
Data sources and search strategy
An experienced medical librarian (W.M.L.) developed search strategies in consultation with 2 of the investigators (M.A.K. and Z.K.) and performed the literature search. The initial search strategy was devised in MEDLINE (through OvidSP) and translated to match the subject headings and keywords for EMBASE (through OvidSP), ISI Web of Science, and Cochrane databases from inception through February 2019. Search terms included proton pump inhibitors, PPI, gastric acid suppression, omeprazole, pantoprazole, rabeprazole, lansoprazole, esomeprazole, acid suppressive therapy, dementia, Alzheimer's dementia, vascular dementia, Parkinson's dementia, cognitive decline, cognition, senile dementia, and cognitive impairment. We searched MeSH terms and free-text words. The search accounted for plurals and variations in spelling with the use of appropriate wildcards. There was no restriction of language. Articles were selected for full-text review on the basis of their title and abstract. We conducted a manual search through the bibliographies of the retrieved publications to increase the yield of potentially relevant articles. All results were downloaded into EndNote 8.0 (Thompson ISI ResearchSoft, Philadelphia, PA), a bibliographic database manager; any duplicate citations were identified and removed.
Inclusion and exclusion criteria
Two investigators (M.A.K. and S.K.) screened original articles using predetermined inclusion criteria, independently collected information to assess the methodological validity of each included study, and extracted data using structured data extraction forms. Any disagreement between the reviewers was to be discussed with a third reviewer (C.W.H.), with an agreement to be reached by consensus. Studies had to be observational in nature (cohort, case-control, or cross-sectional) or randomized or nonrandomized controlled trials and had to have evaluated the association between the development of dementia and PPI use. Studies were included irrespective of the method of assessment of dementia development. We excluded studies when PPI therapy data could not be extracted or was not reported, outcomes not clearly evaluating dementia, animal studies, and studies that exclusively evaluated H2-receptor antagonists. We also excluded unpublished and abstract-based data, and only included data from published peer-reviewed articles because there have been discrepancies in the quality of data reported in abstracts (15,16).
Data extraction and quality assessment
Extracted data included study design, country and year of the study, patient demographics and clinical characteristics, dementia diagnostic criteria, adjusted confounding variables, and, wherever available, duration and dosage of PPI as well as the specific PPI(s) used.
We used the Newcastle-Ottawa Scale (NOS) to assess methodological quality of observational studies (17). This assessment scale has 2 different tools for evaluating case-control and cohort studies. Each tool measures quality in the 3 parameters of selection, comparability, and exposure/outcome and allocates a maximum of 4, 2, and 3 points, respectively. For comparability of cases and controls and of cohorts, a point was given if the study controlled for age and sex, and a second point was given if the study controlled for 3 or more common important factors for dementia (e.g., diabetes mellitus, midlife hypertension, obesity, smoking, depression, physical inactivity, and cognitive inactivity). We used the modified NOS cohort scale adapted from Herzog et al. (18) for a cross-sectional study. High-quality studies have scores of 7 or higher, moderate-quality studies between 4 and 6, and low-quality studies less than 4. Two investigators (Y.Y. and M.A.K.) assessed quality independently, with any disagreement to be resolved by consensus. No controlled clinical trial met the inclusion criteria.
We used the GRADE framework to interpret our findings (19). The Cochrane Collaboration has adopted the principles of the GRADE system for evaluating the quality of evidence for outcomes reported in systematic reviews. For systematic reviews, the GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the quantity of specific interest. Quality of a body of evidence involves consideration of study design, study limitations (risk of bias), indirectness of evidence, inconsistency of results, imprecision, and publication bias. Since this review eventually only included observational studies, certainty of evidence starts at low, with factors that can increase the quality of the evidence considered, including large magnitude of effect, dose-response gradient, and effect of plausible residual confounding. We used GRADEpro to generate overall quality of evidence (https://gradepro.org/).
Data synthesis and statistical analysis
Our primary outcome of interest was the association between PPI use and risk of development of dementia. This association could have been assessed as categorical variables such as adjusted hazard ratio (HR) and adjusted odds ratio (OR) or relative risk (RR). None of the included studies reported RR. HRs and ORs were pooled separately since HRs, ORs, and RRs can only be treated as equivalent measures of risk when the outcome of interest is rare; dementia is not considered a rare event since it had more than 20% prevalence in this review. Wherever possible, we used adjusted ratios (i.e. those that adjusted for most factors) unless only unadjusted data were provided in which case unadjusted ratios were used. One study provided both adjusted HR and adjusted OR (20). That study contributed data to both the pooled HR and pooled OR estimates. For studies that provided adjusted HR and raw data, we did not calculate unadjusted OR to pool with adjusted ORs provided by other studies. As a secondary analysis, we also evaluated the association between PPI use and the risk of development of Alzheimer dementia exclusively. For 1 study that included subjects with “normal cognition or mild cognitive impairment,” we used data reported for subjects with normal cognition at baseline (21). Sensitivity analysis was performed by excluding studies using only International Statistical Classification of Diseases and Related Health Problems (ICD) codes for identifying dementia. One study that defined outcome as “conversion to cognitive impairment (mild cognitive impairment or dementia)” was excluded in sensitivity analysis. We used the DerSimonian and Laird (22) random-effects model for meta-analyses. To estimate statistical heterogeneity, we used the Cochrane χ2 and I2 statistics. A P value of less than 0.1 for χ2 was defined as indicating the presence of heterogeneity. I2 values of greater than 50% indicated substantial heterogeneity (23). We did not perform an assessment of publication bias assessment because of the small number of studies (<10 for all effect estimates). All analyses were performed using Review Manager (RevMan, version 5.3 for Windows; The Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen, Denmark, 2014).
Search strategy yield, study characteristics, and quality assessment
Our search strategy identified 1,422 articles of which 212 were excluded as duplicate publications and a further 1,169 were excluded after title and abstract review. Manual search of the bibliographies of the remaining 41 articles identified no other relevant publications. The remaining 41 articles underwent full-text review, after which, 30 were excluded for the reasons stated in Figure 1. Two studies (24,25) did not present data as risk of dementia development, but rather reported cognitive impairment development using various dementia scales. These studies could not be incorporated in our analysis. Therefore, 11 studies (12,20,21,26–33) were included in this systematic review and meta-analysis; all were observational in design. There were 3 case-control studies (20,30,33), 1 cross-sectional study (31), and 7 population cohort studies (12,21,26–29,32).
Six studies (20,26,27,29,31,32) did not differentiate types of dementia, while the remaining 5 evaluated Alzheimer dementia separately from all causes of dementia (12,21,28,30,33). One study also evaluated the risk of vascular dementia (33). Three studies (26,29,32) derived data from insurance claims databases, 5 (12,21,27,28,31) derived data from prospective population-based cohorts, and the remaining 3 case-control studies (20,30,33) used national databases. Four studies (20,26,29,32) relied on ICD coding exclusively for identifying dementia cases; in the remaining 7 (12,21,27,28,30,31,33), dementia was diagnosed after various cognitive screening tests and neuropsychological testing.
A total of 642,949 subjects were included in our analysis, among whom 64% were women. There were 158,954 PPI users and 483,995 nonusers. Twenty four percent (159,557) of patients were diagnosed with dementia. Detailed study characteristics are outlined in Table 1. As per NOS assessment, most studies showed no evidence of selection bias, with good comparability of exposed and unexposed groups of each cohort and outcome assessment. All studies adjusted for common confounding factors. All studies were assessed as high quality except the cross-sectional study, which was moderate quality (27,31).
Seven studies (12,20,21,26,28,29,32) with 200,892 subjects presented data as adjusted HR. One study had presented PPI usage data for 1, 3, and 5 years. We used data for 5 years counting it as regular long-term PPI usage as dementia has a long latent period. Booker et al. (20) had presented results as both adjusted HR and OR; data from that study were included in our calculations of pooled adjusted HR and pooled adjusted OR. Pooled HR with 95% confidence interval (CI) for association of PPI use with dementia was 1.10 (0.88–1.37), Cochran Q test P < 0.0001, I2 = 96% (Figure 2). Sensitivity analysis after excluding studies using only ICD codes for identifying dementia revealed a pooled adjusted HR of 1.04 (95% CI 0.71–1.53), Cochran Q test P = 0.02, I2 = 76%. Sensitivity analysis excluding study (21) that defined outcome as “conversion to mild cognitive impairment or dementia” showed adjusted HR 1.17 (0.93–1.48). We also evaluated any association with Alzheimer dementia exclusively (12,21,28); pooled adjusted HR was 1.06 (0.72–1.55), Cochran Q test P = 0.03, I2 = 72% (Figure S1, Supplementary Digital Content 1, http://links.lww.com/AJG/B352).
Four studies (20,27,30,31,33) with 465,895 subjects presented data as adjusted OR for dementia. One study (33) had presented separate data for Alzheimer disease and vascular dementia. These were pooled as 2 separate entities in the same analysis of all causes of dementia. Pooled adjusted OR = 1.03 (95% 0.84–1.25), Cochran Q test P < 0.0001, I2 = 94% (Figure 3). Although exclusively evaluating the association with Alzheimer dementia reported in 2 studies (30,33), pooled OR (95% CI) was 0.96 (0.82–1.11), Cochran Q test P = 0.02, I2 = 89%.
Per the GRADE framework, the level of quality of evidence was very low for both outcomes because of study design (observational studies only) and significant heterogeneity in all effect estimates.
In this systematic review and meta-analysis, we found no significant or clinically meaningful association between PPI use and the development of dementia. Although summary outcomes had variously been expressed as OR or HR, we found no statistically significant association using either.
PPIs are among the most commonly used medicines worldwide and are frequently prescribed to elderly patients (9). Therefore, any strong, consistent, and clinically meaningful association between PPI use and an increased risk of the development of dementia would have serious implications for health care. Since different observational studies reported conflicting results for this association, there was a need for this systematic review and meta-analysis.
The first study (12) evaluating this association was a population-based cohort study of 3,327 patients from primary care clinics in Germany; it reported that PPI users had a 44% increased risk of the development of dementia (i.e. a reported HR of 1.44). This report attracted substantial media attention, presumably prompting several other groups to study the proposed association. Subsequently, the same group conducted a large insurance data claims-based study (26) to evaluate this association and found similar results. The biologic plausibility presented to explain this association was that PPIs may augment amyloid beta (Aβ) peptide levels in an amyloid cell model and in the mouse brain (34). PPIs might act as inhibitory modulators for [GREEK UPSILON WITH ACUTE AND HOOK SYMBOL]-secretase and activators for β-secretase which would in turn lead to accumulation of Aβ levels (34). These Aβ accumulations are a major pathological marker for dementia in Alzheimer disease and are cytotoxic to endothelial cells (35). Furthermore, it has been postulated that PPIs might be involved in decreased degradation of Aβ peptides by lysosomes in microglia through inhibition of the proton pumps required for acidification of lysosomes (36–39). In addition, Lam et al. had linked PPI use to the development of vitamin B12 deficiency. That in turn may cause cognitive decline and neurological damage (40,41).
Our results are consistent with a large case-control study (30) and certain population cohort studies (21,28) that used robust methods for detection of dementia in the form of regular cognitive screening, standardized dementia evaluations, neuropsychological evaluations, and/or computed tomography/MRI scans and are based on Diagnostic and Statistical Manual of Mental Disorders, fourth edition. Such methods are superior to ICD-based detection of incident dementia, which is prone to under-recognition and undercoding of dementia inherent to insurance claims-based studies. Therefore, we conducted sensitivity analysis based on studies using the aforementioned more robust methods for detecting dementia and excluding studies based only on ICD coding. The results remained unchanged.
The study by Taipale et al. (30) was the only one to use the lag-window approach to account for the long latency of dementia. With a lag window of 5 years, and more than 1.5 defined daily doses of PPI use, the adjusted OR for risk of dementia was 1.03 (0.92–1.14). Likewise, Gray et al. (28) evaluated the risk of dementia development after calculating total standardized daily doses dispensed (rather than prescribed) to patients in the preceding 10 years. Even with the highest total standardized daily dose studied, there was no significant risk of development of dementia. Claims-based studies are somewhat limited with regard to accurate identification of drug use because there are no structured interviews about medication use and duration of treatment. By contrast, the purchased-based model used by the large case-control study is considered more valid than a prescription-based model because it avoids primary nonadherence problems.
There are several limitations to our findings. First, all the included studies were observational in nature and, therefore, have intrinsic shortcomings including differences in populations and possible unidentified confounders. Therefore, we have rated the pooled OR- and HR-based results as very low quality as per the GRADE framework. Some have suggested (20) that randomized controlled trials should be conducted to evaluate this proposed association further. However, the logistics of, and ethical justification for, such trials would be considerable. They would require the randomization of elderly patients to PPI use or nonuse and subsequent long-term follow-up. Since such an intervention would be hard to justify on clinical, ethical, or economic grounds, no such randomized study has been conducted.
Second, our analyses based on HR or OR had considerable heterogeneity. Although studies adjusted for certain common confounding factors, mainly claims-based studies were limited in adjusting for use of anticholinergics (29). This was also compounded by the various methods for detection of dementia. Therefore, to have confidence in our estimate, we conducted a sensitivity analysis limited to studies reporting defined methods for detection of dementia rather than relying on ICD coding alone. Third, the dose and duration of PPI use was not uniformly reported in all studies so could be a source of heterogeneity. Furthermore, we were unable to assess any biological gradient for the risk of dementia based on dose of PPI used or duration of exposure. Finally, all studies were conducted in developed countries, which may limit the generalizability of our findings to developing nations.
In summary, we found no statistically significant or clinically meaningful association between PPI use and the development of dementia from all known causes or from Alzheimer disease specifically. Overall, the evidence was rated as of very low quality. As for all drugs, PPIs should be given in the lowest effective dose, and their use should be restricted to patients with valid indications. If PPI use in an elderly patient is for an appropriate indication, there is no justification to stop it because of concerns about dementia.
A recent randomized controlled trial compared up to 3 years of treatment with pantoprazole or placebo among a large group of patients receiving rivaroxaban, aspirin, or both (42) and found no association between PPI use and a number of possible adverse outcomes, including a diagnosis of dementia. A web-based survey of US patients on PPI therapy has revealed substantial levels of concern related to PPI safety issues with many patients opting to discontinue treatment without medical advice (43). Our important negative finding—of no proven association between relatively short-term PPI use and the development of dementia—should be relevant for clinicians caring for elderly patients and can hopefully reassure some patients and their family members who may have concerns about this proposed association.
CONFLICTS OF INTEREST
Guarantor of the article: Colin W. Howden, MD, FACG.
Specific author contributions: M.A.K. and Y.Y.: concept and design, data collection, statistical analysis, manuscript drafting, and final approval of manuscript. U.I.: data collection, data interpretation, manuscript drafting, and final approval. S.K., Z.K., and M.K.: data collection, manuscript drafting, and final approval. W.L.: search strategy development and final approval of manuscript. C.W.H.: design and concept, critical revision of manuscript, and final approval of manuscript.
Financial support: None to report.
Potential competing interests: C.W.H.: consultant for Phathom, Ironwood, RedHill, Biopharma, ISOThrive, and Otsuka. The remaining authors have no conflicts of interests.
WHAT IS KNOWN
- ✓ PPIs have been associated with a wide variety of potentially serious long-term adverse effects including a possible increased risk of dementia.
- ✓ However, studies evaluating the association with dementia have reached different conclusions.
WHAT IS NEW HERE
- ✓ Systematic review and meta-analysis of all published observational studies examining this proposed association finds no evidence for it.
- ✓ Concerns about development of dementia do not justify withholding PPI therapy from patients with a genuine need for it.
1. Fratiglioni L, Launer LJ, Andersen K, et al. Incidence of dementia and major subtypes in Europe: A collaborative study of population-based cohorts. Neurologic Diseases in the Elderly Research Group. Neurology 2000;54(11 Suppl 5):S10–5.
2. Prince M, Bryce R, Albanese E, et al. The global prevalence of dementia: A systematic review and metaanalysis. Alzheimers Dement 2013;9(1):63–75.e2.
3. Ferri CP, Prince M, Brayne C, et al. Global prevalence of dementia: A Delphi consensus study. Lancet 2005;366(9503):2112–7.
4. Wimo A, Guerchet M, Ali GC, et al. The worldwide costs of dementia 2015 and comparisons with 2010. Alzheimers Dement 2017;13(1):1–7.
5. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol 2013;108(3):308–28; quiz 29.
6. Qato DM, Wilder J, Schumm LP, et al. Changes in prescription and over-the-counter medication and dietary supplement use among older adults in the United States, 2005 vs 2011. JAMA Intern Med 2016;176(4):473–82.
7. Gawron AJ, Pandolfino JE, Miskevics S, et al. Proton pump inhibitor prescriptions and subsequent use in US veterans diagnosed with gastroesophageal reflux disease. J Gen Intern Med 2013;28(7):930–7.
8. Heidelbaugh JJ, Goldberg KL, Inadomi JM. Magnitude and economic effect of overuse of antisecretory therapy in the ambulatory care setting. Am J Manag Care 2010;16(9):e228–34.
9. Forgacs I, Loganayagam A. Overprescribing proton pump inhibitors. BMJ 2008;336(7634):2–3.
10. 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(4):270–3.
11. Vaezi MF, Yang YX, Howden CW. Complications of proton pump inhibitor therapy. Gastroenterology 2017;153(1):35–48.
12. Haenisch B, von Holt K, Wiese B, et al. Risk of dementia in elderly patients with the use of proton pump inhibitors. Eur Arch Psychiatry Clin Neurosci 2015;265(5):419–28.
13. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: Explanation and elaboration. BMJ 2009;339:b2700.
14. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis of observational studies in Epidemiology (MOOSE) group. JAMA 2000;283(15):2008–12.
15. Scherer RW, Langenberg P, von Elm E. Full publication of results initially presented in abstracts. Cochrane Database Syst Rev 2007(2):MR000005.
16. Taddio A, Pain T, Fassos FF, et al. Quality of nonstructured and structured abstracts of original research articles in the British Medical Journal, the Canadian Medical Association Journal and the Journal of the American Medical Association. CMAJ 1994;150(10):1611–5.
17. Wells G, Shea B, O' Connell D, et al. The Newcastle-Ottawa scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses (http://wwwohrica/programs/clinical_epidemiology/oxfordasp)
. Accessed May 29, 2019.
18. Herzog R, Alvarez-Pasquin MJ, Diaz C, et al. Are healthcare workers' intentions to vaccinate related to their knowledge, beliefs and attitudes? A systematic review. BMC Public Health 2013;13:154.
19. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336(7650):924–6.
20. Booker A, Jacob LE, Rapp M, et al. Risk factors for dementia diagnosis in German primary care practices. Int Psychogeriatr 2016;28(7):1059–65.
21. Goldstein FC, Steenland K, Zhao L, et al. Proton pump inhibitors and risk of mild cognitive impairment and dementia. J Am Geriatr Soc 2017;65(9):1969–74.
22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clin Trials 1986;7(3):177–88.
23. Higgins JPT, Thomas J, Chandler J, et al. (editors). Assessing risk of bias in included studies. Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019). Cochrane, 2019. Available from http://www.training.cochrane.org/handbook
24. Wod M, Hallas J, Andersen K, et al. Lack of association between proton pump inhibitor use and cognitive decline. Clin Gastroenterol Hepatol 2018;16(5):681–9.
25. Lochhead P, Hagan K, Joshi AD, et al. Association between proton pump inhibitor use and cognitive function in women. Gastroenterology 2017;153(4):971–9 e4.
26. Gomm W, von Holt K, Thome F, et al. Association of proton pump inhibitors with risk of dementia: A pharmacoepidemiological claims data analysis. JAMA Neurol 2016;73(4):410–6.
27. Herghelegiu AM, Prada GI, Nacu R. Prolonged use OF proton pomp inhibitors and cognitive function IN older adults. Farmacia 2016;64(2):262–7.
28. Gray SL, Walker RL, Dublin S, et al. Proton pump inhibitor use and dementia risk: Prospective population-based study. J Am Geriatr Soc 2018;66(2):247–53.
29. Tai SY, Chien CY, Wu DC, et al. Risk of dementia from proton pump inhibitor use in Asian population: A nationwide cohort study in Taiwan. PLoS One 2017;12(2):e0171006.
30. Taipale H, Tolppanen AM, Tiihonen M, et al. No association between proton pump inhibitor use and risk of Alzheimer's disease. Am J Gastroenterol 2017;112(12):1802–8.
31. de Souto Barreto P, Lapeyre-Mestre M, Mathieu C, et al. Prevalence and associations of the use of proton-pump inhibitors in nursing homes: A cross-sectional study. J Am Med Dir Assoc 2013;14(4):265–9.
32. Hwang IC, Chang J, Park SM. A nationwide population-based cohort study of dementia risk among acid suppressant users. Am J Geriatr Psychiatry 2018;26(11):1175–83.
33. Imfeld P, Bodmer M, Jick SS, et al. Proton pump inhibitor use and risk of developing alzheimer's disease or vascular dementia: A case-control analysis. Drug Saf 2018;41(12):1387–96.
34. Badiola N, Alcalde V, Pujol A, et al. The proton-pump inhibitor lansoprazole enhances amyloid beta production. PLoS One 2013;8(3):e58837.
35. Querfurth HW, LaFerla FM. Alzheimer's disease. N Engl J Med 2010;362(4):329–44.
36. Majumdar A, Cruz D, Asamoah N, et al. Activation of microglia acidifies lysosomes and leads to degradation of Alzheimer amyloid fibrils. Mol Biol Cell 2007;18(4):1490–6.
37. Fallahzadeh MK, Borhani Haghighi A, Namazi MR. Proton pump inhibitors: Predisposers to Alzheimer disease? J Clin Pharm Ther 2010;35(2):125–6.
38. Mattsson JP, Vaananen K, Wallmark B, et al. Omeprazole and bafilomycin, two proton pump inhibitors: Differentiation of their effects on gastric, kidney and bone H(+)-translocating ATPases. Biochim Biophys Acta 1991;1065(2):261–8.
39. Pillay CS, Elliott E, Dennison C. Endolysosomal proteolysis and its regulation. Biochem J 2002;363(Pt 3):417–29.
40. Reynolds E. Vitamin B12, folic acid, and the nervous system. Lancet Neurol 2006;5(11):949–60.
41. O'Leary F, Allman-Farinelli M, Samman S. Vitamin B(1)(2) status, cognitive decline and dementia: A systematic review of prospective cohort studies. Br J Nutr 2012;108(11):1948–61.
42. Moayyedi P, Eikelboom JW, Bosch J, et al. Safety of proton pump inhibitors on a large, multi-year, randomized trial of patients receiving rivaroxaban or aspirin. Gastroenterology 2019;157:682–91.
43. Kurlander JE, Kennedy JK, Rubenstein JH, et al. Patients' perceptions of proton pump inhibitor risks and attempts at discontinuation: A national survey. Am J Gastroenterol 2019;114(2):244–9.