Herpes Zoster Infection Associated With Acute Coronary Syndrome: A Population-based Retrospective Cohort Study.

Herpes zoster infection associated with acute

coronary

syndrome: a population-based retrospective

cohort study

C.-C. Wang,

C.-L. Lin,

Y.-J. Chang,

G.-J. Wang,

F.-C. Sung

and C.-H. Kao

Varicella zoster virus (VZV) is a double-stranded DNA virus of the Herpesviridae family. VZV affects only humans, and the clinical manifestations of VZV infection are classified as varicella (chickenpox) or herpes zoster (shingles). The virus primarily infects children aged 5–9 years, manifesting as

chickenpox, which is characterized by the cutaneous distribution of diffuse and scattered maculopapular vesicles. Subsequently, VZV infections become latent in cranial nerve and

dorsal root ganglia and in autonomic ganglia. As cellular immunity to the VZV diminishes with age or in persons in an immunocompromised state, reactivation of the latent VZV may occur, resulting in shingles, which is characterized by unilateral vesicular eruptions within the affected dermatomes. Vasculopathy in VZV infection may result from the production of procoagulant antiphospholipid antibodies1 or immune

complex mediated vasculitis; an association between herpes virus infection and atherosclerosis has also been proposed.2

These relationships suggest a possible link between VZV infection and vascular thrombosis.

Previous studies of herpes zoster have focused primarily on its possible relationships with cerebrovascular stroke.3,4 Several

cases of herpes zoster-induced vasculopathy and stroke have been reported, and it has been proposed that the herpes virus spreads transaxonally to the nascent blood vessels, where it induces inflammation and thrombotic responses. Recently, a large population-based cohort study has reported that the risk of stroke and herpes zoster ophthalmicus increased following a herpes zoster attack during the 1-year follow-up period.5

Several case reports have described peripheral arterial thrombosis in patients with concurrent herpes zoster infection.6,7 To

our knowledge, no large population-based cohort study has investigated the association between herpes zoster infection and the development of acute coronary syndrome (ACS). We

hypothesized that the risk of ACS increased following the reactivation of VZV infection in an Asian population-based cohort.

Patients and methods

We analysed a subset of data from the reimbursement claims provided by the National Health Research Institute in Taiwan. The claims database contains information from 13 insurance programmes. Taiwan implemented a universal insurance system in March 1995. The National Health Insurance (NHI) system provides health care for 99% of the 23 million citizens of Taiwan through contracts with 97% of the health-care facilities.

8 We reviewed the reimbursement claims of the NHI that

had been previously extracted and compiled in

the NHI Research Database (NHIRD). The NHIRD claims

records contained information on both ambulatory and inpatient medical services from 1996 to 2010, and diagnoses were coded based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). The

patients’ personal information was encrypted to ensure privacy. To ensure the accuracy of the claims data, the Bureau of NHI (BNHI) invites experts to review claims by randomly sampling every 100 outpatient claims and every 15 inpatient claims. False claims and unnecessary treatment result in severe penalties from BNHI.9 Several studies have demonstrated the

high accuracy and validity of ICD-9 diagnosis in NHIRD.8,10

Our study was approved by the ethics review board of the China Medical University (CMU-REC-101-012).

Participants and outcome measures

Patients newly diagnosed with herpes zoster (ICD-9-CM code 053) between 1999 and 2010 were included in our study cohort. The index date was defined as the date of the first clinical visit for herpes zoster. We excluded patients with a history of ACS (ICD-9-CM codes 410, 411_1) or with missing age or sex information at the baseline. The herpes zoster

cohort to nonherpes zoster cohort was selected at a 1:4 ratio in order to enhance the power of statistical tests and to prevent an insufficient number of ACS cases in the stratified

analysis. The nonherpes zoster cohort was selected using a simple random sampling method to four randomly selected insured people from NHI beneficiaries without herpes zoster status for every person with herpes zoster in the same period, frequency matched by age, sex and index year of herpes zoster diagnosis. Subjects in both the herpes zoster and nonherpes zoster cohorts were followed up until they were diagnosed with ACS or censored because of loss of follow-up, withdrawal from insurance, death or until the end of 2010.

Definition of exposure

At least three previous claims for ambulatory or inpatient care

for a single diagnosis before the index date represented a comorbidity at the baseline. The baseline comorbidities

considered were hypertension (ICD-9-CM codes 401–405), diabetes mellitus 9-CM code 250), hyperlipidaemia (ICD-9-CM code 272) and cerebral vascular disease (ICD-(ICD-9-CM codes 430–438), chronic obstructive pulmonary disease (COPD; ICD-9-CM codes 491–493, 494, 496), renal dysfunction (ICD-9-CM codes 580–589) and cancer (ICD-9-CM codes

140–208). Acyclovir and valacyclovir were the antiviral treatments for herpes zoster infection. Famciclovir was not considered

because it was not imported into Taiwan until 2011. Statistical analysis

The distributions of categorical demographics and clinical characteristics among the herpes zoster and the control groups were compared using the v2 test. Patients were classified based

on socioeconomic status; occupation [white-collar workers (professionals, managers and administrative workers), bluecollar workers (farmers, fishermen and industrial labourers) or

‘other’ (retired, unemployed and low-income patients)];

urbanization (urbanization level was categorized by the population density of the residential area into two groups, with

urban as the most urbanized and rural as the least urbanized) and monthly income (new Taiwan dollars 15 000, 15 000–

rates for the herpes zoster group

and the rates of the control group. Hazard ratios (HRs) and

95% confidence intervals (CIs) were calculated using a multivariable Cox proportional-hazards regression model to examine

associations between herpes zoster and the incidence of ACS. Variables considered in the multivariate analysis included sex, age, frequency of medical visits, urbanization, monthly income (in new Taiwan dollars), occupation, comorbidities and medication status. We have also addressed the attributable risk and 95% CI from herpes infection.11

All analyses were performed using SAS version 9.2 (SAS,

Cary, NC, U.S.A.), and two-tailed P-values < 0_05 were considered statistically significant.

Results

Baseline characteristics

A total of 57 958 herpes zoster patients and 231 832 control patients were included in our study (Table 1). Among all patients, 60_1% were < 55 years old and 52_6% were women. Over half were white-collar workers. The herpes

zoster group exhibited significantly higher rates of comorbidities at the baseline. Hypertension (28_5% vs. 25_1%), followed by hyperlipidaemia (18_6% vs. 15_5%), COPD (11_2% vs. 8_87%), renal dysfunction (7_05% vs. 5_23%), diabetes mellitus (10_9% vs. 9_3%), cancer (3_43% vs. 2_24%) and cerebral vascular disease (9_0% vs. 7_9%) showed discrepancies in prevalence.

Attributable risks and 95% CIs

Approximately 2_2% of the ACS incidence in the herpes cohort would be eliminated if the herpes exposure was eliminated (95% CI 1_1–3_2%). In other words, the excess ACS due to herpes zoster was 2_2%.

Incidence rate ratios and hazard ratios of acute coronary syndrome

Over the 12-year follow-up period, 1053 patients with herpes zoster and 3318 control patients were diagnosed with ACS (Table 2). The incidence rates for the herpes zoster and the control groups were 36_8 and 29_6 per 10 000 person-years, respectively. The overall adjusted HR (aHR) of ACS was 1_15

for patients with herpes zoster (95% CI 1_07–1_24). In both groups, ACS occurred more frequently in men than in

women. Herpes zoster patients over the age of 65 years had a higher incidence of ACS (25%) than the age-matched controls, and the risk of ACS progressively increased with increasing age. The risk of ACS was higher among people living in urban areas than in rural areas (aHR 1_08, 95% CI 1_02–1_15). The risk of ACS was highest among the lowest income group (aHR 1_31, 95% CI 1_19–1_45). The presence of hypertension, hyperlipidemia, diabetes mellitus, cerebral vascular disease, COPD and renal dysfunction was also associated with a higher incidence of ACS. We further divided the herpes zoster group into two subgroups according to treatment. Patients treated with antivirals exhibited a higher incidence rate of ACS than those who were not treated with antivirals, but this difference was not significant after adjusting for all potential variables (aHR 0_96, 95% CI 0_80–1_15) (Table 4).

Acute coronary syndrome in relation to time since herpes zoster diagnosis

Analysis divided by the length of follow-up periods showed

that the risk of ACS among the two cohorts did not differ significantly within 1 year after herpes zoster was diagnosed.

After the 1-year follow-up period, the herpes zoster group had a significantly higher risk of ACS than the nonherpes zoster group (aHR 1_10, 95% CI 1_02–1_19) (Table 5). Kaplan–

Meier survival analysis showed that patients with herpes zoster infections had significantly higher ACS rates than the control patients (log-rank test, P < 0_001; Fig. 1).

Discussion

Our study demonstrated that herpes zoster infection is associated with an increased risk of ACS. After adjusting for all

potential covariates, including baseline comorbidities, socioeconomic status and medication, the association between the

infection and the development of ACS remained significant. Association between varicella zoster virus infection and ischaemic heart disease

In the literature, the increased risk of cerebrovascular events and peripheral vascular thromboembolism following a herpes

zoster attack has been established.12–15 Massano et al.14

described the development of acute occlusions in the left common iliac and right femoral arteries in a 39-year-old man

1 week after VZV infection. In addition, Peyton et al.15

reported two cases with varicella pneumonia and coexisting

severe lower limb ischaemia caused by thrombosis of the profounda femoris and tibial arteries. However, the relationship

between herpes zoster infection and ACS has been rarely discussed. A study of the association between herpes virus infection

and the incidence of coronary heart disease (CHD) found no significant difference in the risk of CHD in patients with a baseline history of shingles compared with control patients.16

However, this lack of an association may have resulted from a smaller sample size and a shorter follow-up duration than in our study. A case–control study showed that a higher proportion of patients was seropositive to four or more herpes virus (high herpetic burden) in the ischaemic heart disease group than in the control group.17 Therefore, the relationship

between VZV infection and CHD remains controversial. Compared with previous studies, our study considered a larger

population and longer duration of follow-up. We proposed that following a herpes zoster attack the risk of ACS increased

in the long-term follow-up period. The mechanism of varicella zoster virus-induced vasculopathy

Herpes zoster infection may induce vasculopathy in several

ways. Firstly, it may induce the production of several prothrombotic autoimmune antibodies. IgM and IgG anticardiolipin

antibodies have been reported in a patient with concurrent

deep vein thrombosis and VZV infection,18 and lupus anticoagulant,

IgM and IgG antiphospholipid antibodies were reported in two patients with acute thrombosis of the profunda femoris and tibial arteries following varicella pneumonia.15 Secondly,

following VZV infection, delayed vasculopathy may develop because of an autoimmune phenomenon induced by circulating immune complexes.19 Finally, the direct spread of VZV transaxonally

from the dorsal root ganglia may cause vasculopathy

characterized by the disruption of the internal elastic lamina, intimal hyperplasia and decreased smooth muscle cells in the

medial layer.20 We demonstrated that the cumulative risk of

ACS in the herpes zoster group increased significantly compared with the nonherpes zoster group after the first year follow-up period. Our data were not consistent with those of a previous study,5 in which the risk of stroke in the herpes zoster group

increased significantly within the 1-year follow-up period. This difference may be due to differences in underlying pathogenesis. Cerebral vascular and peripheral vascular thromboembolism may be caused directly by the spread of VZV transaxonally from

the nerve ganglia. However, it is unlikely that coronary thromboembolism is caused by a similar mechanism. We hypothesized

that the persistent low-grade productive infection12 after

the acute stage may result in the chronic, increased inflammation and production of prothrombotic antibodies, which may

ultimately lead to the delayed presentation of ACS events. The incidence of acute coronary syndrome in relation to age and socioeconomic status

The incidence rate of ACS was higher among older rather than

younger patients. This implied that the reactivation of VZV infection among older people may be associated with more

serious complications. In a large-scale epidemiological study conducted in the U.K. and Canada, herpes zoster infection in patients over 65 years of age was associated with a higher rate of hospitalization per case and a longer average length of stay.21 Therefore, the reactivation of VZV infection among

older people may contribute to more complications and poorer prognosis.

Moreover, we demonstrated that the lower-income patients had a higher risk of ACS than higher-income patients. This result may be explained by the policies regarding antiviral therapy for herpes zoster infection that have been implemented by the NHI in Taiwan. The NHI reimbursement for

the cost of antiviral therapy for herpes zoster is limited to immunocompromised patients and patients for whom lack of treatment may contribute to retinal necrosis, or faecal or urinary incontinence. Patients not in these categories must pay

for antiviral therapy outside of insurance. Therefore, those with lower incomes may be reluctant to seek treatment for

herpes zoster infection; the lack of treatment possibly contributes to persistent viral infection and, subsequently, more complications. This result may imply that early aggressive therapy may reduce the future complications of vasculopathy, although further studies are required to confirm this

hypothesis.

The effect of antiviral treatment on the incidence of acute coronary syndrome

This study failed to demonstrate that antiviral treatment of herpes zoster infection successfully reduced the cumulative risk of ACS. This result may have been confounded by several factors. Firstly, the NHI reimburses only the cost of antiviral treatment for immunocompromised patients or those with complications. Patients who received antiviral treatments may have had more baseline comorbidities and a higher inherent risk of ACS than patients who did not receive antiviral treatment. Secondly, the NHI reimburses the cost of antiviral treatment only for 5–10 days. Certain patients may experience

persistent or low-grade productive viral infection in ganglia. Whether prolonged treatment could reduce the incidence of ACS requires further investigation.

The association between varicella zoster virus infection and conventional cardiovascular risk factors

In our study, the herpes zoster group had significantly higher proportions of hypertension, diabetes mellitus, hyperlipidemia and cerebrovascular events than the control group. Diminished cellular immunity in older patients with diabetes mellitus or

hyperlipidemia may have increased their susceptibility to reac- tivation of VZV infection.22,23 However, after adjusting for

these multiple potential confounding factors, herpes zoster

infection remained independently associated with the development of ACS in our study cohort.

Study limitations

The strength of our study lies in the fact that we analysed a population-based dataset highly representative of a general population, which allowed for 12 years of follow-up observations. However, the study has certain limitations. Firstly, the

habits, alcohol consumption, body mass index or family history of systemic diseases, all of which may be risk factors for ACS. Secondly, the evidence derived from a retrospective cohort study is generally lower in statistical quality than that from randomized trials because of potential biases related to

adjusting for confounding variables. Despite the meticulous study design and control measures for confounding factors,

bias resulting from unknown confounding factors may have affected our results. Thirdly, all data in the NHIRD are anonymous. Thus, relevant clinical variables, such as blood pressure,

imaging results, pathology findings and serum laboratory data were unavailable for our study patient cases. However, the

data regarding herpes zoster and ACS diagnoses were nonetheless reliable. Fourthly, our study focused on the association

between reactivation of VZV infection and the risk of ACS because we hypothesized that reactivation of VZV infection increased inflammation, the production of procoagulants and the induction of vasculitis. Because of the limited follow-up period, our results cannot be extrapolated to determine further whether patients with active shingles or during the incubation period have an increased risk of ACS. However, based on our literature review, we believe that reactivation of VZV infection could result in vasculopathy and is associated with an

increased risk of ACS. Finally, we could only identify patients with VZV infection based on the ICD-9 codes and had no access to serology data for clinical validation. Therefore, this study was unable to clearly differentiate the suspected cases from the verified cases.

In conclusion, patients with a history of herpes zoster infection are more likely to develop ACS than patients without such a history. Whether early aggressive antiviral treatment may prevent subsequent VZV-associated vasculopathy requires further investigation.