Reflux Esophagitis and the Risk of Stroke in Young Adults

Gastro-esophageal reflux disease (GERD) is a common gastrointestinal disorder in Western countries, with the prevalence ranging from 10% to 20%. A lower prevalence has been reported in Asia.¹ However, more recent studies suggest that the prevalence of GERD in Asia is increasing.^2–4 Reflux esophagitis (RE), the most common manifestation of GERD with esophageal injury, can be diagnosed endoscopically.⁵ Endoscopic studies shows that the prevalence of reflux esophagitis among the adult population in Taiwan is 14.5%,² which is similar to that reported in Western countries. There is a wide range of extra-esophageal complications or comorbid conditions associated with GERD, including pulmonary, laryngeal, gastrointestinal, and cardiovascular diseases, but other health problems or comorbidities potentially associated with GERD, such as cardiovascular disease, may yet be identified.⁶ Strokes of undetermined etiology account for one-third to one-fourth of ischemic strokes among young people⁷ and, to date, there has been no specific study to evaluate the risk of stroke after diagnosis of RE in young adults.

The aim of the present study is to evaluate the risk of stroke among RE patients aged 18 to 50 years during a 1-year period after the diagnosis of RE compared to a cohort of non-RE patients during the same period.

Materials and Methods

Database

This study used the Longitudinal Health Insurance Database 2005 derived from the Taiwan National Health Insurance program, which is made available to scientists in Taiwan for research purposes. National Health Insurance provides coverage for 22.6 million people, which account for 98.4% of the island’s total population of 22.96 million. The Taiwan National Health Research Institutes created the Longitudinal Health Insurance Database 2005, which consists of 1 000 000 subjects, by systematically selecting a representative database from all enrollees listed in the 2005 Registry of Beneficiaries. The Longitudinal Health Insurance Database 2005 contains all the original claims data from these 1 000 000 beneficiaries collected by the National Health Insurance program. There is no significant difference between the patients in the Longitudinal Health Insurance Database 2005, and all enrollees under the National Health Insurance program in terms of gender distribution, age distribution, or amount of average payroll-related insurance payments. The details of database generation are described on the website of the Taiwan National Health Research Institutes.⁸ The study was exempt from full review by the Institutional Review Board because the dataset used consists of deidentified secondary data released to the public for research purposes.

Study Sample

The study design was a retrospective cohort study. We selected patients between 18 and 50 years of age who visited ambulatory care centers for the treatment of RE between January 1, 2003 and December 31, 2005, as identified from the database by the diagnosis of RE (ICD-9-CM code 530.11 or 530.81). We excluded patients who had visited ambulatory care centers for the treatment of RE before the year 2003 (1996–2002) to increase the likelihood of identifying new cases (n=12). We also excluded patients who had stroke (ICD-9-CM codes 430–438) diagnosed before 2003 (n=53). However, because the National Health Insurance program in Taiwan was initiated in 1995, the dataset only allowed us to trace medical services utilization from 1996 to 2006. Therefore, we could not exclude patients who had stroke or RE before 1996. To increase the validity of the RE diagnosis, we selected only those patients with RE diagnosed by endoscopy or who received proton pump inhibitor prescriptions for >30 days. Finally, we excluded patients who used clopidogrel during the study period (n=8). In total, 2340 patients with RE were included in the study group. We assigned their first ambulatory care visits for the treatment of RE as their index ambulatory care visits.

Our comparison group was extracted from the remaining patients in the Registry of RE diagnosed between 1996 and 2006. We then randomly selected 11 700 subjects (5 for every RE patient) matched in terms of age, gender, and the year of index ambulatory care visits using the SAS program (SAS System for Windows, version 8.2; SAS). These selected patients did not have any type of stroke before their index ambulatory care visits. Each patient was individually tracked for 1 year from the index visit to identify those who had strokes.

Statistical Analysis

The SAS statistical package was used for all statistical analyses. Pearson χ² tests were performed to explore differences between the 2 groups in terms of sociodemographic characteristics and select comorbid medical disorders (hypertension, diabetes, coronary heart disease, renal disease, heart failure, and hyperlipidemia) and the incidence of stroke. We then used the Kaplan-Meier method and log-rank test to estimate the 1-year stroke-free survival rate and to examine differences in the risk of stroke between these 2 groups. Stratified Cox proportional hazard regressions (stratified by age and gender) were also performed to calculate the 1-year stroke-free survival rate after adjusting for patient’s monthly income (New Taiwan [NT] $0, NT $1–$15 840, NT $15 841–$25 000, and >NT $25 000), level of urbanization (ranging from most urbanized [level 1] to least urbanized [level 5]), the geographical location of the patient’s residence (Northern, Central, Eastern, and Southern Taiwan), and whether a patient had hypertension, diabetes, coronary heart disease, renal disease, heart failure, and hyperlipidemia. NT $15 840 was used as the first income level cut-off point because that is the government’s definition of minimum wage for full-time employees in Taiwan. We present hazard ratios along with 95% confidence intervals using a significance level of 0.05.

Results

Table 1 presents the distribution of demographic characteristics and select comorbid medical disorders for patients with RE and patients in the comparison group. Of the total sample of 14 040 patients, the mean age was 35.7 (standard deviation, 9.7 years), and 57.1% were male. After matching for gender and age, patients with RE had a greater tendency to have comorbidities of renal disease (P<0.001), heart failure (P<0.001), hypertension (P<0.001), diabetes (P<0.001), coronary heart disease (P<0.001), and hyperlipidemia (P<0.001) at the time of their index ambulatory care visits compared to the patients in the comparison group. Table 1 also shows that patients with RE were more likely to have higher monthly incomes (P<0.001) and to reside in the central part of Taiwan (P<0.001) than the comparison group.

The distribution of stroke between patients with RE and patients in the comparison group is shown in Table 2. Of the 14 040 sampled patients, 78 patients (0.56%) had strokes during the 1-year follow-up period: 22 (0.94% of the RE patients) from the study group and 56 (0.48% of patients without RE) from the comparison group. The log-rank test reveals that patients with RE had significantly lower 1-year stroke-free survival rates than patients in the comparison group (P<0.001). The results of Kaplan-Meier survival analysis are presented in the Figure.

• Download figure
• Download PowerPoint

The crude and adjusted hazard ratios for stroke by group are also shown in Table 2. It suggests that the crude hazard of stroke during the 1-year follow-up period was 1.97-times greater (95% confidence interval, 1.20–3.24; P=0.007) for patients with RE than for those without. The stratified Cox proportional hazard regressions (stratified by age and gender) shows that patients with RE were more likely to have stroke during the 1-year follow-up period (hazard ratio, 1.68; 95% confidence interval, 1.03–2.76; P=0.041) for patients with RE compared with the comparison group after adjusting for patients’ monthly income, hypertension, diabetes, coronary heart disease, renal disease, heart failure, hyperlipidemia, level of urbanization, and the geographical location the patient’s residence.

Discussion

Some evidence suggests that GERD plays a role in stimulating anginal attacks in patients with impaired coronary circulation through the vagal visceral reflex.^9,10 There also have been a few epidemiological studies exploring the link between GERD and cardiovascular diseases.^11,12 Using the UK General Practice Research Database, Ruigómez et al¹¹ demonstrated that patients with GERD had an increased risk of subsequent diagnosis of angina. However, there has been only 1 epidemiological study exploring the link between GERD and stroke.⁶ Jansson et al⁶ demonstrated a possible link between myocardial infarction, angina pectoris, stroke, and GERD in a population-based, cross-sectional study in Norway. A causal relationship between GERD and stroke cannot be established because of the cross-sectional study design.

To our knowledge, this study is the first of its kind to investigate the risk of stroke among young adults with RE during the year after a diagnosis of RE and adjusting for patient demographic characteristics and comorbid medical disorders. Our study shows that the likelihood of stroke was 1.68-times greater among young adults with RE during the 1-year follow-up period compared to age- and gender-matched subjects after adjusting for other risk factors for stroke. Our findings support the link between GERD and stroke found in the previous study.⁶

The mechanisms contributing to the link between RE and stroke are unclear. An irritant esophago-gastric stimulus could induce impairment in the cardiac conduction signaling or autonomic modulation of the heart rate, resulting in cardiac dysrhythmia.¹³ In addition, reflex coronary vasoconstriction and reduction of coronary blood flow may be initiated by acid stimulation of the esophagus through the vagal esophago-cardiac reflex.¹⁴ Cardioembolic stroke may occur in these 2 situations. In GERD, there is a high prevalence of vagus nerve dysfunction, which relates to delayed esophageal transit, abnormal peristalsis, and increased frequency of transient lower esophageal sphincter relaxations.^15,16 Because the immune system is under the direct control of the vagus nerve via the cholinergic anti-inflammatory pathways, vagus nerve dysfunction may set off excessive inflammatory responses and the spread of inflammatory mediators into the bloodstream, which then contribute to the triggering of the common atherosclerotic process and could lead to cardiovascular events.¹⁷ In addition, cardiovascular autonomic dysfunction mediated by the vagus nerve in GERD patients may lead to a change in the sympathetic and parasympathetic balance in the cerebral vasculature and defective autoregulation of the cerebral blood flow in diabetes patients, increasing the risk of stroke.^18,19 Risk factors regarding the link is risk factors shared by both GERD and stroke patients, such as diet, smoking, obesity, and metabolic syndrome, are a possible explanation for this.^20–22

One of this study’s strengths is the use of a population-based dataset, which enables us to trace all cases of RE and stroke during the study period. Moreover, the large sample size affords considerable statistical advantage for detecting real differences between the 2 cohorts. Nevertheless, this study suffers from a few limitations. First, RE and stroke diagnoses that rely on administrative claims data reported by physicians or hospitals may be less accurate than diagnoses made according to standardized criteria, and there is a lack of data regarding the severity of the endoscopic findings. However, to avoid misdiagnoses, we selected only patients with RE diagnosed by endoscopy and who have received proton pump inhibitor prescriptions for >30 days. The use of endoscopy rather than self-reported symptoms as diagnostic tools may provide RE diagnosis in a standardized fashion and lessen the possibility of misclassification.^23,24 In addition, virtually all hospitals in Taiwan capable of admitting stroke patients are equipped with CT or MRI scanners, which increases the validity of stroke diagnoses considerably. Second, patient information, such as smoking, alcohol consumption, dietary factor, body mass index, all of which may contribute to stroke, was not available through the administrative dataset. Thus, the association between RE and stroke may be partially explained by residual confounding by these factors. Increased obesity and westernization of the diet (high fat intake) among young Asians may increase the risk of RE and stroke.^2,4 However, to decrease confounding in the estimation of hazard rations for stroke, we did take into consideration conventional risk factors for stroke, such as hypertension, diabetes, coronary heart disease, and hyperlipidemia. Third, there may be a surveillance bias in that those patients with RE are more likely to have frequent check-ups, which may lead to early detection of stroke. However, because the clinical evidence linking RE to stroke in clinical practice has been lacking, patients with RE are usually followed-up by their gastroenterologists. Patients visit a neurologist or emergency specialist only when they have neurological symptoms, and then they undergo brain CT/MRI studies to diagnose stroke. Fourth, the possible increase in cardiovascular events in patients using proton pump inhibitors with clopidogrel may suggest a possible role of proton pump inhibitors in the occurrence of strokes in RE patients. However, to avoid the possible interaction of proton pump inhibitors and clopidogrel, we have excluded patients who used clopidogrel during the study period. Fifth, there may be the possibility of underestimating RE in the comparison group. However, this would lead to an underestimation of the strength of association between RE and stroke. Finally, as a further potential limitation, the study population mainly comprised Taiwanese of Chinese descent who had RE diagnosed based on endoscopy, and the results may not be capable of generalization to Western populations or patients with GERD identified by reflux symptoms.

Conclusions

Our study shows that there may be an association between RE and the risk of subsequent stroke in young adults. This study used a 1-year follow-up period to explore the relationship between RE and subsequent stroke attributable to limitations of the database. It is suggested that further studies using a longer follow-up period should be conducted to confirm the relationship found in the present study.

Footnotes