Depression events associated with proton-pump inhibitors in postmarketing drug surveillance data

Proton-pump inhibitors, PPIs, are widely prescribed and are available over the counter for prolonged reduction of stomach acid production and related disorders. PPIs irreversibly inhibit the hydrogen/potassium ATPase in gastric parietal cells. Recent retrospective studies have described an association between PPI use and depression. However, there is conflicting evidence that PPI therapy improves depressive symptoms. Considering the widespread use and over the counter availability of these drugs, further investigation into depression adverse event was warranted with a larger scale postmarketing set of reports. Here we analyzed over 68,178 reports from the FDA Adverse Event Reporting System consisting of PPI and histamine-2 receptor antagonist monotherapy records and found a statistically significant association between use of PPIs and depression. Additionally, we analyzed each of the six currently marketed PPIs individually and observed the association with the depression adverse reaction for all of them.

Common adverse drug reactions (ADRs) associated with PPIs observed in controlled trials include nausea, diarrhea, flatulence, vomiting, and headache.Serious ADRs include throat tightness, rash, and difficulty breathing [13][14][15][16][17][18].Post-approval studies have found additional association of PPI use with the following ADRs: Clostridium difficile-associated diarrhea, calcium deficiency, bone fractures, and hypomagnesemia, which are now listed as precautions in the FDA package inserts.
PPI drug use has been associated with neurological disorders such as Alzheimer's and non-Alzheimer's dementia [19][20][21].Additionally, recent studies analyzing the FDA Adverse Event Reporting System (FAERS) data have observed multiple electrolyte abnormalities and a broad spectrum of neurological disorders disproportionally reported after PPI use when compared to H2RAs [22,23].
The physiology behind neurological and psychiatric conditions often shares common molecular mechanisms [24,25], resulting in depression's comorbidity with neurological disorders and vice versa [26].There are multiple hypothesized mechanisms behind the psychiatric/neurologic adverse reactions of PPIs [27][28][29].In an epidemiological study by Laudisio et al., PPI use was associated with depression in a geriatric population, while the association was not observed in the H2RA cohort [30].In another study conducted in Sweden by Wang et al. [31], PPI use was associated with increased risk of anxiety and depression in children.Furthermore, a Taiwanese nationwide population-based study by Huang et al. [32] confirmed this association.There is conflicting evidence suggesting improvement of depressive symptoms following PPI therapy [33].
Depression was declared by the World Health Organization (WHO) to be the third greatest cause of burden for disease worldwide [34], and considering the common use of PPIs, this association needed to be further investigated.In this study, we performed an analysis of millions of FAERS reports and identified a rare but statistically significant increase in depression reports in patients taking PPIs as monotherapy when compared to H2RA-related reports.

FDA Adverse Event Reporting System
The FDA Adverse Event Reporting System (FAERS) and its older version AERS contain the FDA's postmarketing drug and biologic product safety data.Reports are submitted by product vendors, and, on a voluntary basis, by legal representatives, healthcare professionals, and consumers to the FDA through MedWatch [35,36], the FDA Safety Information and Adverse Event Reporting Program.Pharmaceutical industry/manufacturers are legally required to forward the information to the FDA.

Data Preparation
Quarterly FAERS/AERS ASCII reports were downloaded in their original format.The data from different years were unified into a consistent format.Additionally, since the ADR reports were collected from around the world, it was necessary to translate drug brand names into generic ones using online drug databases.A total of 19,190,582 combined FAERS/AERS reports were used for the analysis [37,38].The reports were further narrowed down to exclude reports submitted by consumers and legal representatives to avoid potential bias.All the reports submitted by physicians, pharmacists, nurses, and other healthcare professionals were included in the analysis datasets.

Analysis and Control Cohort Selection
FAERS/AERS reports where PPIs and H2RAs were used as monotherapy were selected into the respective cohorts.The term "monotherapy" pertains to records where only a single treatment is listed in the report.The query included US approved brand and generic names of PPIs and H2RAs.The dictionary for brand vs. generic names was compiled from resources such as USAN and WHO Drug Global, which were employed for a non-US brand name report recognition.Additionally, frequent misspellings of both generic and brand names were manually added to the translation dictionary.PPI monotherapy cohorts collectively included 8488 reports, and H2RA monotherapy cohorts included 117,435 reports.
Reporting odds ratio (ROR) analysis was performed by comparing the reported PPI ADRs in relation to H2RA report numbers with and without an ADR of interest.The PPI monotherapy cohort was further split into individual PPI cohorts, which included omeprazole (n = 2651), esomeprazole (n = 1993), pantoprazole (n = 1846), lansoprazole (n = 1170), dexlansoprazole (n = 405), and rabeprazole (n = 423).Individual PPI-reported depression ADRs were calculated and compared to the H2RA cohort to screen for potential ADR variability within individual PPIs in the cohort.Demographic analysis was performed (Tables 1-3).Additionally, co-occurring adverse events were analyzed for presence of any additional events that may be related psychiatric disorders (Tables 4 and 5) Standard Error of Log Reporting Odds Ratio: 95% Confidence Interval: Haldane-Anscombe correction was not applied to cohorts with zero depression reports [39].

Results
Combined PPI monotherapy reports had a significantly higher number of depression ADRs in comparison with H2RA monotherapy reports.The reported frequency of depression AE reports in the PPI cohorts ranged from 0.25% for dexlansoprazole to 1.42% for rabeprazole, and 0.73% for the combined cohort (Figure 1), with ROR being 50

Depression Reports in Males
There was a significant increase in the number of reported depression ADRs in the combined PPI monotherapy cohort when compared to the H2RA monotherapy cohort (ROR39.40

Depression Reports in Males
There was a significant increase in the number of reported depression ADRs in the combined PPI monotherapy cohort when compared to the H2RA monotherapy cohort (ROR39.

Depression Reports in Males
There was a significant increase in the number of reported depression ADRs in the combined PPI monotherapy cohort when compared to the H2RA monotherapy cohort (ROR39.40,95% CI [20.47, 119.21]) when only male patients taking PPIs were analyzed.Each of the individual PPI treatments analyzed had a significant increase in the number of depression ADRs in comparison with the H2RA control, with the exception of dexlan-  3 and 4).

Depression Reports in Females
There was a significantly higher number of depression ADR reports in the combined PPI monotherapy female-only cohort when compared to the H2RA monotherapy cohort (OR 27.54, 95% CI [13.35, 56.

Depression Reports in Females
There was a significantly higher number of depression ADR reports in the combined PPI monotherapy female-only cohort when compared to the H2RA monotherapy cohort (OR 27.54, 95% CI [13.35, 56.80]).All individual PPI cohorts had a significant increase in the number of depression ADR reports when compared to the H2RA control, as follows: omeprazole (ROR 51.24, 95% CI [23.54, 111.51]), pantoprazole (ROR 12.30, 95% CI [3.78,

Depression Reports in Females
There was a significantly higher number of depression ADR reports in the combined PPI monotherapy female-only cohort when compared to the H2RA monotherapy cohort (OR 27.54, 95% CI [13.35, 56.80]).All individual PPI cohorts had a significant increase in the number of depression ADR reports when compared to the H2RA control, as follows: omeprazole (ROR 51.24, 95% CI [23.54, 111 5 and 6).

Discussion
In this study, we quantified and confirmed a statistically significant association between PPI exposure and depression ADRs using 125,923 PPI and H2RA ADR reports from the FAERS/AERS database.To our knowledge, it is the first study investigating the relationship of PPI use with depression utilizing the most recent population-scale postmarketing surveillance data.
Although the fraction of depression reports was relatively low, ranging from 0.25% to 1.42% of PPI monotherapy ADR reports, the depression ADR ROR numbers were statistically significant when compared to the H2RA control cohort.This study expanded on the previous epidemiological evidence provided in the geriatric study by Laudisio et al. [30] and the pediatric study by Wang et al. [31].Additionally, in our study, depression ADRs were analyzed across age groups.The depression ADR signal prevailed in all groups above 20-29 to ≥70-year-old cohorts (Table 2).
The molecular mechanisms of the observed association may originate from several pathways and are beyond the scope of this study.Due to complex pharmacology of PPIs, it is difficult to pinpoint a single physiological mechanism involved in the etiology of depression ADRs.In a previously published mini-review, we discussed some of the mechanisms by which PPIs might be causing neurological adverse events [27], including hypomagnesemia and impaired vitamin B absorption.
A systematic review published by Derom et al. [40] found an association between higher dietary magnesium intake and lower depression symptoms.In another publication by Cheungpasitporn et al. [41], the investigators evaluated the effect of high magnesium and found a similar association.Magnesium is known to play multiple important roles in the central nervous system (CNS) [42], and inadequate levels affecting psychiatric symptoms seem intuitive.
In a four-year longitudinal study by Laird et al. [43] in adults 50 years and older, vitamin B12 deficiency was found to be associated with increased depressive symptoms.In another study by Syed et al. [44], when supplemented along with antidepressants, vitamin B12 significantly decreased depressive symptoms.A systematic review by Almeida et al. [45] has found that the severity of depression symptoms in not decreased over a short period.However, there are long-term benefits in vitamin B12 supplementation for depression management.
Coincidentally, the PPI users reported both hypomagnesemia and vitamin B12 deficiency.The report counts were as follows: 35 reports of decreased vs. 1 report of increased magnesium levels, and 43 reports of decreased vs. 1 report of increased vitamin B12 levels (Tables 4 and 5).This associations of PPIs with vitamin B12 and magnesium absorption abnormalities [46,47] are plausible mechanisms that could explain the elevated reporting numbers of depression ADRs.Moreover, regardless of specific molecular mechanisms related to depression ADRs, it is not surprising to see psychiatric ADRs, given the previous evidence of a wide range of neurological damage associated with PPI use [21,23,[48][49][50].
The observed increased risk of depression with use of PPIs calls for more careful consideration in using PPIs for people at high risk of depression.PPIs should be used for the shortest time necessary as recommended by the FDA and other regulatory authorities.Prescribers should be aware of this potential risk and educate the patients to avoid deviations on recommended dose, frequency, and use duration.Further studies are needed to evaluate the neurologic and psychiatric adverse events of PPIs in a controlled setting, identify at-risk populations, and identify the molecular mechanisms associated with the PPIs' neuropsychiatric adverse events.

Study Limitations
Only a fraction of population incidences is represented in FAERS/AERS due to the mostly voluntary nature of the reporting.FAERS/AERS reporting can be biased based on multiple factors such as legal challenges or news events influencing submissions [51].Additionally, significant underreporting of AEs has been observed depending of drug, type of ADR, and seriousness [52].Causality of depression cannot be inferred from the association found in this study since the cases were not individually adjudicated by a clinician.However, the study provides a statistically significant signal that needs to be further investigated in a controlled setting.Drug brand names from some countries may not have been translated if they did not pass the FDA AE submission application.Only reports where the brand name has been registered with WHO Drug were included in the dictionary.
, 95% CI [20.47, 119.21]) when only male patients taking PPIs were analyzed.Each of the individual PPI treatments analyzed had a significant increase in the number of depression ADRs in comparison with the H2RA control, with the exception of dexlansoprazole, which had zero depression ADR reports.The ROR values were as follows: omeprazole (ROR 47.26, 95% CI [15.86, 140.84]), pantoprazole (ROR 82.59, 95% CI [27.68,
40, 95% CI [20.47, 119.21]) when only male patients taking PPIs were analyzed.Each of the individual PPI treatments analyzed had a significant increase in the number of depression ADRs in comparison with the H2RA control, with the exception of dexlansoprazole, which had zero depression ADR reports.The ROR values were as follows: omeprazole (ROR 47.26, 95% CI [15.86, 140.84]), pantoprazole (ROR 82.59, 95% CI [27.68,

Table 1 .
Report numbers by sex.

Table 2 .
Report numbers by age.

Table 3 .
Report numbers by country.