Eradication of Helicobacter Pylori May Be Associated With the Incidence of Alzheimer’s Disease in Diabetes Mellitus Patients

Introduction: Alzheimer’s disease (AD) is the most common form of dementia. Eradication of Helicobacter pylori (H. pylori) could affect the incidence and progression of many diseases; however, there are limited studies of the association between H. pylori eradication and AD outcome. We utilized the National Health Insurance Research Database (NHIRD) of Taiwan to determine the relationship between H. pylori eradication and AD in a diabetes mellitus (DM) population. Methods: We collected data from the NHIRD and the Diabetes Mellitus Health Database in Taiwan of patients without a prior diagnosis of AD. We specied three cohorts: patients with (1) peptic ulcer disease (PUD) but no H. pylori treatment, without DM (PUD-HPRx in GP); (2) PUD and DM, but no H. pylori eradication therapy (PUD-HPRx in DM); (3) PUD and DM, with H. pylori eradication therapy (PUD+HPRx in DM). All cohorts were matched according to age, sex, Charlson Comorbidity Index score, and comorbidities. Results: Data were collected from 2000 to 2010, and 157,231 patients were enrolled in total. We compared the effects of treatment for H. pylori infection on the incidence and mortality of AD. The patients with DM who received H. pylori eradication therapy had a higher incidence of AD than the general population (adjusted hazard ratio of incidence [aHR], 1.088). Subgroup analysis showed that the risk of AD was higher in the younger patients who received H. pylori eradication therapy as compared with those who did not (aHR for younger than 45 years, 1.071; aHR of age 45-54 years, 1.089; aHR of age 55-64 years, 1.079) However, a lower mortality rate was observed in the PUD+HPRx in DM group (aHR, 0.945, compared with PUD-HPRx in DM; P < 0.001). Conclusion: In this study, we demonstrated that DM patients who underwent treatment for eradication of H. pylori had a higher incidence of AD, especially younger patients. Nevertheless, there was a lower mortality rate in patients who received H. pylori treatment. Further study is needed to clarify the interrelated roles of AD and eradication therapy for H. pylori.


Introduction
Alzheimer's disease (AD), also known as "senile dementia" or "cognitive disorder", is a progressive, agein uenced neurodegenerative disease. It is the most common form of dementia, especially in industrialized countries, and affects approximately 30 million people around the world with a continuous rising trend. (1,2) AD is characterized by progressive impairments in diverse behavioral and cognitive functions that exert a prominent impact on AD patients, their families, and society. The pathophysiology of AD is complex, and has not yet been fully elucidated. AD can display an early or late onset depending upon the genome, diet and lifestyle of the patient.(3) Currently, identi ed risk factors of AD include age, sex, plasma homocysteine level, and genetic factors such as apolipoprotein E allele 4.(4) Several studies have shown an association between infection and AD, including HSV-1, chlamydia pneumonia, and spirochetes.(5, 6) Nonetheless, the signi cant heterogeneity of AD-related risk factors, etiologies, and neuropathologic processes makes the development of new therapeutic strategies to slow disease progression especially challenging. (7) Helicobacter pylori (H. pylori) is a gram-negative, spiral-shaped pathogenic bacterium that inhabits the gastrointestinal system in more than 50% of adults. It can cause chronic gastritis, peptic ulcer disease and gastric malignancies in humans.(8) In recent years, numerous studies have emphasized the relationships between H. pylori infection and the pathogeneses of extra-gastrointestinal disorders, such as coronary heart disease(9) and cerebrovascular disease.(10) Eradication of H. pylori using a proton pump inhibitor and antibiotics could lower the incidence and progression of many diseases, such as gastrointestinal tract malignancies (11) and even coronary heart disease. (12) Previously, one small-group case-control study identi ed an association between H. pylori and AD. (13) One hypothesis is that H. pylori might access the brain, mainly via the oral-nasal-olfactory pathway or the circulation, by circulating monocytes (infected with H. pylori due to defective autophagy) through a disrupted blood-brain barrier (BBB). (14) However, there are limited reports and evidence of associations between H. pylori eradication and the outcome of AD. Some of the greatest challenges include the relatively low prevalence of AD in regions with high rates of H. pylori infection and the requirement for a large sample size in order to draw meaningful conclusions. Therefore, in this study, we utilized the National Health Insurance Research Database (NHIRD) of Taiwan, a country in which the prevalence of H. pylori approaches 80%, (15) in order to determine the relationship between H. pylori eradication and Alzheimer's disease, and analyzed the outcome in a diabetes mellitus (DM) population.

Data Acquisition
Since March 1995, the Taiwan National Health Insurance (NHI) has covered ambulatory care, outpatient care, dental services, hospital inpatient care, and drug prescriptions. It insured more than 96% of the whole population of 23 million in 2000, rising to 99% by the end of 2004. We also collected data from a subgroup database, the Longitudinal Health Insurance Database (LHID 2010), which details anonymized reimbursement claims data for medical research and includes 1-million randomly-selected individuals from the NHIRD from January 2000 to December 2010. We further collected data from the Diabetes Mellitus Health Database (to identify the DM cohort) from January 2000 to December 2015, DM being identi ed according to Ninth Revision of the International Classi cation of Diseases (ICD-9) code 250, and patients with ≥ 3 outpatient department follow-up visits or at least one hospitalization were included.
The rationale for creating this longitudinal database was to allow a 5-year window in both directions in order to analyze the long-term impact of H. pylori treatment on AD in the general population and DM patients.

Study Design
We selected peptic ulcer disease (PUD) patients over 18 years of age identi ed by ICD-9 codes 531 (gastric ulcer), 532 (duodenal ulcer), and 533 (nonspeci c peptic ulcer). Then, we speci ed three cohorts, as follows: (1) patients with PUD but no H. pylori treatment, without DM (PUD-HPRx in GP); (2) patients with PUD and DM, but no H. pylori eradication therapy (PUD-HPRx in DM); and (3) patients with PUD and DM who had received H. pylori eradication treatment (PUD + HPRx in DM). The de nition of patients who had received H. pylori treatment was that they were given a course of either triple or quadruple therapy (proton pump inhibitor or H2 receptor blocker, plus clarithromycin or metronidazole, plus amoxicillin or tetracycline, with or without bismuth)(16) for more than 7 days, rather than a negative result of an H. pylori test, due to the latter not being available in the NHIRD. The three groups were matched using the propensity-score method by age, gender, Charlson Comorbidity Index score (CCIS), comorbidities, and medications for underlying diseases. Subjects who were selected into one group were excluded from the database and were not selected in the other groups. After matching, there were apparent differences between the cohorts in terms of age (especially those under 35 years); comorbidities of hyperlipidemia, myocardial infarction, and chronic kidney disease (CKD); and the percentage of patients receiving antiplatelet drugs.

Study Sample
Diagnosis of AD was de ned by ICD-9 code 331.0, with ≥ 3 months of AD-related medical treatment including cholinesterase inhibitors such as Donepezil, Galantamine or Rivastigmine. The person-years of follow-up were estimated from the index date plus a 2-year lag time to the date of diagnosis of AD. We excluded any rst date of an AD-de ned event prior to the index date. Subjects with incomplete demographic data (e.g., lost to follow-up, withdrawal from the insurance system, or erroneous information) were also excluded from the study. The follow-up period began after H. pylori therapy (index date) plus a 2-year lag time, and ended on December 31, 2010.

Statistical Analysis
We analyzed the distribution of risk factors for the 3 study cohorts (PUD-HPRx in GP, PUD + HPRx in DM, PUD-HPRx in DM) by analysis of variance (ANOVA), the chi-square test, or Fisher's exact test. Cox proportional hazards regression analyses were performed to determine the crude and adjusted hazard ratios (aHRs) after adjustment for age, gender, comorbidities, CCIS, and nonsteroidal anti-in ammatory drugs or antiplatelet agents. A competing risk model was also performed to determine the risks of AD incidence and mortality. Kaplan-Meier curves were employed to estimate the probability of AD onset or mortality, and the log-rank test was used to analyze the differences between groups. Statistical analyses were performed using SAS 9.3 software (SAS Institute, Cary, NC, USA). Statistical signi cance was set at P < 0.05.

Results
Higher cumulative incidence of AD in PUD + HPRx in DM cohort as compared with PUD-HPRx in GP and younger DM patients To understand the impact on the development of AD of treatment for H. pylori and other underlying diseases (GP and DM), we rst compared the new incidence of AD patients with PUD diagnosed by esophagogastroduodenoscopy who received H. pylori therapy (PUD + HPRx) or never received H. pylori therapy (PUD-HPRx) in patients with DM. Then, we compared these cohorts with patients diagnosed with PUD but without DM in order to reduce confounding factors of different incidences. The patient characteristics are shown in Table 1. The average age and gender balance was 57.61 ± 14.37 years with 43.9% men in the PUD-HPRx in GP control group, 57.35 ± 13.73 years with 43.32% men in the PUD + HRPx in DM group, and 57.52 ± 14.08 years with 43.32% men in the PUD-HPRx in DM group.  Fig. 1). Besides, DM patients who received H. pylori eradication therapy had a lower mortality rate. We identi ed a signi cantly lower mortality rate in the DM cohorts (aHR of PUD + HRPx, 0.945; 95% CI, 0.918-0.972; P < 0.001) ( Table 2). *P-value < 0.05.

Discussion
In this study, we found that the patients with DM who received H. pylori eradication therapy had a higher incidence of AD than the general population after a 2-year follow-up period. In addition, the risk of AD was signi cantly higher in the younger patients who received H. pylori eradication therapy as compared with those who did not in the DM group. Despite the higher risk of developing AD, there was a lower mortality rate in the DM patients who received H. pylori eradication therapy.
The rst important nding of this study is that after 2 years of follow-up, the incidence of AD in patients with DM who received Helicobacter pylori eradication therapy was higher than that of the general population. In addition, compared with the untreated DM group, young patients who received H. pylori eradication therapy had a signi cantly higher risk of AD. Previous reports showed that H. pylori can induce systemic in ammation and increase homocysteine levels, contributing to worsening of AD. (17)(18)(19) However, the contribution of H. pylori to the neuro-in ammatory process in AD has not yet been con rmed or denied.(20) Seiler et al. (21) revealed that elevated levels of ammonia could be taken into account as a factor contributing to manifestations and the progression of AD. Also, existing evidence suggests that the accumulation of ammonia in the brain affects nerve function and may cause some neurological abnormalities. (22,23) H. pylori obtains nutrition from food residues after travel of the meal from the stomach and drop of the acid to a minimum in a blink-like momentum protected by a shield of ammonia and leaving some scattered ammonia after it in the gastric lumen. (24) This scattered ammonia will stimulate the stomach wall to secrete acid to buffer the ammonia. (24,25) This biological balance between the ammonia of H. pylori and gastric acid ensures a residual level of ammonia that could further account for a constant systemic serum ammonia level. (26,27) However, under the in uence of antibiotic treatment, H.pylori may be forced to migrate to the colon, (25,28) where it will continue to produce ammonia, causing a large accumulation that is toxic to the brain.(25, 28) One population-based study also found that eradication of H. pylori was associated with progression of dementia. (29) Another important nding of this study was that the DM patients who received H. pylori eradication therapy had a lower mortality rate. One possible reason is that eradication of H. pylori could lower the incidence of related gastrointestinal tract malignancies. (11,30) In addition, H. pylori may also confer a mortality risk to AD patients through its involvement in a variety of extradigestive vascular conditions, including ischemic heart disease(31, 32) and cerebrovascular disease. reported that an H. pylori eradication regimen may improve the long-term survival rate of AD patients.

Limitation
There were several limitations of this study. First, the results of tests showing eradication of H. pylori were not available. We instead obtained data from patient medical records, and patients documented as undergoing an H. pylori eradication regimen for more than 7 days were deemed to have received H. pylori treatment, but hard data showing successful H. pylori eradication did not exist. Second, although we matched the study cohorts by age, sex, comorbidities, CCIS, and medications, a generally acceptable method, it was not possible to fully account for and correct other potential confounders (e.g., diet, alcohol consumption), thus preventing causal inferences from being drawn from the observed associations. Third, we excluded a general population with PUD that received H. pylori eradication treatment, a cohort that consisted of fewer cases and was di cult to match with the other groups, which may have led to potential confounding bias. Finally, as per the general limitations that apply to retrospective data analyses, we were unable to correct data errors such as incorrect diagnoses or errors in data entry, and we were also unable to review patient medical records to double-check the accuracy of the data. Therefore, there may have been some unadjusted data in this cohort study.
In conclusion, the results of this study demonstrated that eradication of H. pylori in DM patients resulted in a higher incidence of AD, especially in the younger patients. Nevertheless, there was a lower mortality rate in the patients who received H. pylori treatment, even in light of the higher AD prevalence, in the DM group. Further longitudinal study is needed to clarify the interrelated roles of cognition, eradication therapy for H. pylori infection, and AD. Figure 1 AD incidence Forest plot for PUD+HPRx and PUD-HPRx in DM groups. Subgroup analysis showed that the incidence of AD was signi cantly higher in the patients aged under 65 years who received H. pylori eradication therapy as compared with those who did not in the DM group (aHR for younger than 45 years,