Helicobacter pylori infection and the risk of
acute
coronary syndrome: a nationwide retrospective
cohort study
C.-Y. Lai & T.-Y. Yang & C.-L. Lin & C.-H. Kao
Introduction
Cardiovascular disease (CVD) is the second leading cause of death in Taiwan [1]. Acute coronary syndrome (ACS) is a critical stage of the clinical manifestation of coronary artery disease (CAD) inCVD[2]. ACS results in substantialmorbidity and mortality, accounting for half of all deaths from CVD and contributing to the heavy economic burden of the disease [3]. The effective treatment of ACS is guided by early diagnosis and risk stratification to predict those who are at high risk of short- and long-term adverse outcomes. A family history of premature coronary heart disease (CHD), modifiable risk factors such as hyperlipidemia [4], hypertension, diabetes, and
metabolic syndrome [5], and non modifiable risk factors such as sex and age, can be used to predict atherosclerosis development and the risk of presenting with ACS [6].
Atherosclerosis is by far the most frequent cause of CAD. Life-threatening manifestations are typically precipitated by
acute thrombosis superimposed on ruptured or eroded atherosclerotic plaque [7]. Multiple factors contribute to the
pathogenesis of atherosclerosis, including endothelial dysfunction, dyslipidemia, inflammatory and immunological factors,
plaque rupture, and smoking [8].
The importance of inflammation in the pathogenesis of
atherosclerosis derives from the fact that markers of increased or decreased systemic inflammation are associated with the
risk of atherosclerosis [9]. Evidence for inflammation in atherosclerotic lesions was obtained from early histological observations,
and inflammation is central to understanding the pathogenesis of atherosclerosis [10].
Microbe infection could act according to a number of mechanisms, including direct vascular injury and induction of a systemic inflammatory state, such as that induced by Helicobacter pylori (HP) infection (HPI) [11].
HPI is the most common chronic bacterial infection of the
human upper gastrointestinal tract [12]. Conservative estimates suggest that half of the world’s population is infected
with HP [13]. HP is a microaerophilic spiral-shaped Gramnegative bacteriumthat colonizes the gastric lumen of humans
and other primates and is of major etiological importance in peptic ulcer disease and gastric cancer [14]. Increasing evidence from both clinical and experimental observations suggests that inflammation plays a crucial role in CAD Fibre15].
However, subsequent studies have produced conflicting findings. Whether this relationship with inflammation arises from
the bacterium itself or its association with other confounding factors related to atherosclerosis, such as a low socioeconomic class, old age, and smoking, primarily modifiable risk factors, such as hyperlipidemia, hypertension, diabetes, and metabolic syndrome, or non-modifiable risk factors, such as sex and age, remains uncertain [17–19]. Confounding by the strong relationship of HPI to other coronary heart disease risk factors, such as
age and socioeconomic class may, at least partially, explain the conflicting results that have been obtained [20].
A large population-based study may help clarify the effect of such confounding factors. Therefore, we conducted an investigation by using records from the Taiwan National Health Insurance Research Database (NHIRD) to evaluate whether HPI patients are at risk of developing ACS after
accounting for traditional risk factors of ischemic heart disease. The results of this study are provided as a reference to the public and medical professionals.
Materials and methods
Study design and data source
We used the Taiwan NHIRD to conduct a population-based retrospective cohort study. This database was established in
1995 when the Taiwan National Health Insurance (NHI), a comprehensive insurance program, was launched by the Taiwan Department of Health. The NHI covers 99.9 % of
the population of Taiwan. The Taiwan National Health Research Institute is responsible for managing various databases,
including the registration files of the insured and claims data for reimbursement. Before releasing the electronic files for study, the personal identification numbers were encrypted to protect patient privacy. We used the inpatient claims data as the datasets, and the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) was used to define diseases. The Institutional Review Board of China Medical University (CMU-REC-101-012) approved this study from full ethical review.
Study population
From the inpatient claims data, we selected patients who were newly diagnosed with HPI (ICD-9-CM 041.86) from 1998 to 2010 as the HPI group. For the comparison group, we randomly selected people from the general population who were
4-fold frequency-matched by age, sex, and diagnosis year.We excluded patients with a history of HPI and those with incomplete age or sex information at the baseline. Patients who had a
history of comorbidities, such as hypertension (ICD 401– 405), diabetes 9-CM codes 250), hyperlipidemia (ICD-9-CM codes 272), stroke (ICD-(ICD-9-CMcodes 430–438), chronic obstructive pulmonary disease (COPD) (ICD-9-CM codes 490–492,494, and 496), and heart failure (ICD-9-CM 428) were included in this study. Both groups were followed up until the observation of an ACS (ICD-9-CM 410, 411.1, and 411.8) event or the end of the study period (December 31, 2010).
Statistical analysis
To compare the distributions of demographic variables and comorbidities between the HPI group and non-HPI group, we analyzed the categorical variables by using a chi-squared test. Two sample t tests were used to determine differences in the continuous variables, such as mean age, between the two groups. The sex-, age- and comorbidity-specific incidence
densities of ACS per 1,000 person-years of follow-up for each cohort were estimated. The crude hazard ratio (HR) with a 95 % confidence interval (CI) of the HPI group to the comparison group was calculated using univariable Cox proportional
hazards regression. Multivariable Cox proportional hazards regression was used tomeasure the adjusted HR (aHR) of ACS for the HPI group, compared with non-HPI group, after controlling for sex, age, and comorbidities. The interactional effects between the comorbidities and HPI were also assessed using Cox models. All of the statistical analyses in this study were performed using the SAS 9.1 statistical package (SAS Institute Inc., NC, USA). P<0.05 in 2-tailed tests was considered significant. Results
We identified 17 075 patients for the HPI group and selected 68 300 participants for the comparison group (Table 1). After frequency matching, both groups had similar sex and age distributions and more men (62.9 %) and younger participants (34.8 %, ≤49 y). Comorbidities were more prevalent in the HPI group than in the non-HPI group (P<.0001).
The incidence of ACS was 1.93-fold higher in the HPI group than in the non-HPI group (6.41 vs 3.33 per 1,000 person-years; aHR=1.48, 95 % CI=1.30–1.69) (Table 2). In both groups, the risks of developing HPI were higher for men than for women. Analysis by sex yielded an aHR of 1.38 (95 % CI=1.09–1.76) for the women and 1.53 (95 % CI= 1.31–1.79) for the men, compared with the non-HPI group. The risks of developing ACS increased with increasing age in both groups, with a much greater gradient in the HPI group. The participants with comorbidities exhibited higher incidences of ACS in both groups (12.2 vs 9.35 per 1,000
person-years), and the HPI group exhibited a higher aHR of 1.53 (95%CI=1.29–1.80) comparedwith the non-HPI group. Table 3 shows that the HPI group was significantly associated with the development of ACS. The risks of developing
ACS increased as age increased. The aHR of ACS was 1.63 times higher among men than among women (95%CI=1.44– 1.85). Compared with the participants without comorbidities, patients with comorbidities such as hypertension (aHR=1.59;
95 % CI=1.37–1.84), diabetes (aHR=2.03; 95 % CI=1.75– 2.36), hyperlipidemia (aHR=1.59; 95 % CI=1.28–1.97), COPD (aHR=1.41; 95 % CI=1.16–1.71), heart failure (aHR=1.50; 95 % CI=1.12–2.00), and stroke (aHR=1.14; 95 % CI=0.95–1.36) exhibited a greater risk for developing ACS (all P<0.05 except for stroke).
We also estimated risk of ACS in relation to HPI, hypertension, diabetes, hyperlipidemia, COPD, and heart failure,
and the interaction among these factors (Table 4). The risk of developing ACS in patients with HPI increased with the existence of any comorbidity. HPI was observed to have an interaction with hyperlipidemia (P=0.009) and COPD (P= 0.01). When HPI and hyperlipidemia were present, the aHR was 4.20 (95 % CI=3.19–5.54), compared with the participants without HPI and hyperlipidemia. The aHR was
2.66-fold higher in the HPI patients with COPD than in the non-HPI/non-COPD participants (95 % CI=2.01–3.52).
Discussion
We compared patients with HPI matched by age, sex, and history of comorbidities at the baseline, such as hypertension, hyperlipidemia, stroke, COPD, and heart failure. The risk of developing ACS in patients with HPI increased with the
presence of any comorbidity. HPI had a synergistic interaction with hyperlipidemia (P= 0.009) and COPD (P=0.01). The presence of both HPI and hyperlipidemia yielded an aHR of 4.20 (95 % CI=3.19–5.54) compared with participants without HPI and hyperlipidemia. Patients with COPD yielded an
aHR 2.66-fold higher than among non-HPI/non-COPD participants (95 % CI=2.01–3.52). After age and other comorbidities
were adjusted for, the results of this study suggested
an inverse significant biological gradient for the development of the risk of ACS between the HPI and non-HPI groups. Previous studies [21] on South Korean adults have shown
that HPI is associated with cardiovascular risk factors, particularly levels of triglycerides, HDL-cholesterol, and apolipoproteins, independent of the presence of peptic ulcers. A study from India [22] reported that HPI and diabetes exacerbated glucose tolerance and CHD development. The colonization of
CagA-positive HP does not seem to be an independent risk factor for severe CHD [27]. Previous studies provided evidence that HPI stimulates a cascade of inflammation, and
subsequent immune responses include cellular and humoral immunity between T and B lymphocytes [24, 25], even bacterial colonization, persistence, virulence, and resulting innate
and adaptive immunity play a crucial role in HP-related disease [26, 27]. However, HPI may influence atherogenesis
through low-grade, persistent inflammatory stimulation [28] and is significantly associated with risk of short-term adverse outcomes [29]. The reduction in restenosis of coronary vessels
after HP eradication could be interpreted as evidence of the involvement of an HP inflammatory process. Kowalski et al.
stated that HP-eradication therapy may prolong the
hospitalization-free period for patients with recurrent chest pain [30, 31]. Taiwan and other developing and
underdeveloped countries [32, 33] have a high prevalence of HPI; thus, the eradication of HPI may reduce the emerging burden of CVD. Investigating more aggressive HPI and treatment for the disease may be necessary. Randomized controlled
trials are required to evaluate the role of HP eradication in these patients. PPIs (proton pump inhibitors) have been shown to decrease the antiplatelet effects of clopidogrel ex vivo [34], raising concerns about the cardiovascular safety of this drug combination. Therefore, more caution when using PPIs for HP eradication is necessary.
This study used NHIRD claims data and as such has
several limitations. The NHI database does not disclose patients’ personal histories, socioeconomic status, serum laboratory data, smoking status, or inflammatorymarkers. And the diagnoses of HPI were involved in several different methods, including blood antibody test, stool antigen test, or carbon urea breath test [35]. However, the most reliable protocol for HPI detection is a biopsy check during endoscopy with a rapid urease test, histological examination, and microbial culture which is only one way for requested insurance benefit in the NHIRD of Taiwan. Therefore the investigation of HPI activity or severity might not be accurately determined in claim data.
We could not adjust for environmental factors, such as
smoking habits, etc. The evidence suggests that smokers have a risk of ACS, which is compatible with the results of other studies [36]. This may account for much of the risk reduction for smoking-related ACS. However, HPI is not directly related to smoking. Data unadjusted for smoking status may reflect
this possible relationship more adequately. In several case–control studies for proved association of HPI and ACS, there
were less causal relationship evidence for clarification of HPI and ACS. We used the NHIRD to improve the evidence for clinical significance between HPI and ACS [36]. However, we need further approaches to clarify the relationship between different environmental factors, such as drugs, smoking habits and other individual exposed information, to account for the excessive risk for HPI-related ACS after adjustment for common risk factors for ACS. In the future it would be possible to
clarify the detailed exposed information and intervening genetic factors of the host by the National Taiwan Biobank [37].
Conclusion
In this study, we determined a significant association between ACS and comorbidities and provide evidence to encourage clinicians to observe ACS-related comorbidities which might improve preventive ACS complications in the HPI population.
