Statins can increase the risk of herpes zoster infection in Asia.

Statins can increase the risk of herpes zoster infection in Asia

H.-H. Chen

& C.-L. Lin

& C.-J. Yeh

& S.-Y. Yeh

& C.-H. Kao

Introduction

Several guidelines emphasize the importance of statins in treating dyslipidemia [1, 2]. The key pathway by which statins lower blood cholesterol levels is the inhibition of HMG-CoA

reductase. This direct effect of statins reduces the risk of cerebrovascular diseases. In addition to having lipid-lowering effects,

statins have been reported to reduce inflammation, stabilize plaques in the blood, and increase nitric oxide production

[3, 4]. In addition, recent data have shown that statins can increase new-onset diabetes mellitus, dementia, and myopathy.

Therefore, benefit–risk evaluation of statin use is recommended.

In the present study, we focused on the association between

statin use and herpes zoster (HZ) infection. Several studies conducted in Taiwan [5] and Israel [6] have shown that elderly

patients and patients with diabetes and immunocompromising conditions are at an increased risk of developing HZ infection.

Although previous studies conducted in Japan [7] and Canada [8] have discussed the relationship between statin and HZ infection, we considered additional factors, such as various comorbidities, the Charlson comorbidity index (CCI) score,

and different types of statins and their defined daily dose (DDD), to evaluate the association between statin use and HZ infection in Asia. Methods

Data source

Taiwan’s National Health Insurance (NHI) program is a mandatory universal health insurance program established in 1995 and covers approximately 99 % of Taiwanese citizens (http://w3.

nhri.org.tw/nhird//en/Background.html). The National Health Insurance Research Database (NHIRD), which contains the annual registration files and original claims data for

reimbursement, is managed by the National Health Research

Institutes (NHRI). To protect the privacy of patients, the NHRI encrypts patients’ personal information, such as patient identification numbers, sex, date of birth, registry of medical services, and prescription drug details, before releasing the data to researchers. This retrospective cohort study used the Longitudinal Health Insurance Database (LHID2000), which is a subset of the NHIRD. In brief, the LHID2000 comprises the original claims data of 1,000,000 randomly selected enrollees from the Registry for Beneficiaries of the NHI program during 1996 and 2000. According to an NHRI

report, there is no statistically significant difference between the LHID2000 and NHIRD data in terms of age, sex, annual birth rate, and health costs (http://w3.nhri.org.tw/nhird//en/

Data_Subsets.html#S3). The diagnoses of diseases were based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

Sampled patients

From the LHID2000, patients aged 20 years were divided into two cohorts according to their statin use: (1) a statin cohort comprising patients receiving statin therapy for at least

3 months and (2) a nonstatin cohort comprising patients receiving no statin therapy. We set the date on which statin

therapy commenced as the index date. Patients younger than 20 years of age, those having a history of HZ infection (ICD- 9-CM code 053), and those with incomplete information on age or sex were excluded from the study. The patients receiving statin therapy and those without statin therapy were

matched at a 1:1 ratio according to propensity scores, which were calculated by using a logistic regression to estimate the

probability of treatment assignment according to baseline variables, including the year inwhich statin therapy was received,

age, sex, CCI score, and comorbidities, namely hypertension

(ICD-9-CM codes 401–405), diabetes mellitus (ICD-9-CM

code 250), inflammatory bowel disease (ICD-9-CM codes

555 and 556), stroke (ICD-9-CM codes 430–438), cancer

(ICD-9-CM codes 140–208), systemic lupus erythematosus

(SLE) (ICD-9-CM code 710.0), rheumatoid arthritis (RA)

(ICD-9-CM code 714), and human immunodeficiency virus

(HIV) infection (ICD-9-CM codes 795.71, V08, 042, and 079.53) (Fig. 1).

Outcome measurement

The occurrence of HZ infection was set as the primary outcome.

All patients were followed from the index date to the

date of HZ infection, withdrawal from the insurance system, or the end of 2011.

Variables of interest

The cumulative defined daily dose (cDDD) was calculated using the total prescribed DDD of each type of statin used by the patients, including simvastatin (ATC C10AA01), lovastatin (ATC C10AA02), pravastatin (ATC C10AA03), fluvastatin (ATC C10AA04), atorvastatin (ATC C10AA05), and

rosuvastatin (ATC C10AA07). The annual mean DDD was calculated by determining the ratio of the cDDD of individual statin therapy to the follow-up period. For each type of statin, the annual DDD was separated into two levels by setting a cutoff value in the third quartile (highest 25 %).

Statistical analysis

The statin cohort and nonstatin cohort were matched on the basis of the propensity score. To estimate the propensity score, a logistic regression model was used, in

which statin treatment status was regressed on the baseline characteristics listed in Table 1. The standardized difference was used to quantify differences in means or prevalence between the two cohorts for continuous or categorical matching

variables, respectively.

The cumulative incidences of HZ infection in the statin

cohort and nonstatin cohort were assessed using the Kaplan–

Meier method, and the differences between the cohorts were estimated using the log-rank test. The overall incidence and incidence in both cohorts were measured according to gender, age group, and CCI score. Cox proportional hazards models,

in which the matched pairs were stratified, were used for estimating the hazard ratios (HRs) and 95 % confidence intervals

(CIs) of HZ infection associated with statin therapy compared

with the nonstatin cohort. Further analysis was conducted to

assess the association between the annual mean DDD of each

type of statin and HZ infection. All analyses were performed using SAS statistical software (version 9.3 forWindows; SAS Institute, Inc., Cary, NC, USA). All statistical test results were considered to be statistically significant when two-tailed pvalues were <0.05.

Results

The present study included 53,069 patients on statin therapy as the statin cohort and 53,069 patients without statin therapy

as the nonstatin cohort. The mean ages of the patients in the statin cohort and those in the nonstatin cohort were 59.6 [standard

deviation (SD)=12.5] years and 60.6 (SD=13.5) years, respectively. Approximately 22.4 % of the patients were

<49 years old. The number of women was higher than the number of men in both cohorts (50.5 % vs. 49.5 %), and 5.92 % of the patients received statin therapy and had a CCI score ≥3.Among the 53,069 patients in the statin cohort, some patients had a history of comorbidities, such as hypertension (70.00 %), diabetes mellitus (28.50 %), inflammatory bowel disease (1.49 %), stroke (9.93 %), cancer (3.14 %), SLE (0.14%), RA (0.23%), and HIVinfection (0.08 %). The mean follow-up periods for the patients in the statin cohort and the nonstatin cohort were 4.89 (SD=2.86) years and 4.75 (SD=

2.90) years, respectively (data not shown). Figure 2 shows that the patients in the statin cohort had a higher incidence of HZ infection than the patients in the nonstatin cohort, and the occurrence of HZ infection was significantly different between the two cohorts (log-rank test, p<0.001). Overall, the

patients in the statin cohort (8.80 per 1,000 person-years) had

a higher incidence of HZ infection than the patients in the nonstatin cohort (7.48 per 1,000 person-years) (Table 2). The

patients in the statin cohort had a 21 % higher risk of

contracting HZ infection than the patients in the nonstatin

cohort (95 % CI=1.13–1.29). The incidence of HZ infection

was higher in women than in men in both cohorts. The HRs

for HZ infection of the statin cohort relative to the nonstatin

cohort were significantly higher for women than for men

(HR=1.31, 95 % CI=1.20–1.43). The incidence of HZ infection

increased with the age of patients in both cohorts. The

HRs for HZ infection of the statin cohort were higher than those of the nonstatin cohort in all age groups (HR=1.35, 95 % CI=1.12–1.62 for 20–49 years; HR=1.33, 95 % CI=

1.19–1.49 for 50–64 years; HR=1.24, 95 % CI=1.11–1.38 for>65 years). The incidence of HZ infection increased with the CCI score in both cohorts. The risk of HZ infection was higher in patients in the statin cohort than in patients in the nonstatin cohort when both cohorts had CCI score=0 (HR=

1.21, 95 % CI=1.12–1.31) and CCI score=1 (adjusted HR=

1.21, 95 % CI=1.02–1.44). The associations between the annual mean DDD of individual statins and the risk of HZ infection

are shown in Table 3. For all six types of statins, the risk of HZ infection was significantly higher in patients

with a high annual mean DDD (cutoff value in the third quartile) than in nonusers of statins. For simvastatin,

fluvastatin, and lovastatin, patients with a low annual mean DDD exhibited a lower risk of HZ infection than nonusers of statins.

Discussion

The strengths of our study are its population-based design, generalizability of findings, and use of population-based data and NHIRD records using a large sample size and having low loss to follow-up in the longitudinal design, including study and control cohorts. In addition, the NHIRD covers a highly representative sample of Taiwan’s general population because the reimbursement policy is universal and operated by a single buyer, the government in Taiwan. All insurance claims should be scrutinized by medical reimbursement specialists and peer review.

Statins and infection

An 11-year mortality follow-up of the lipid-lowering arm in the UK fromthe Anglo-Scandinavian Cardiac Outcomes Trial showed that there was a significant reduction in the relative risk of deaths from infections or respiratory illnesses (36 %, p=0.04), but for the infection-related deaths, the reduction in the risk of death was of borderline significance (40 %, p=

0.06) [9]. In addition, the trial results suggested that atorvastatin

can prevent only certain respiratory infections. Previous

metaanalysis data have shown that statin therapy can prevent or treat many common bacterial pneumonia [10, 11]. Several

possible mechanisms explain these results. First, the immunomodulatory effects of statins have been observed in other clinical

studies, such as those on atherosclerosis, organ transplantation, dementia, and RA [12–14]. Statins can control immune

responses such as T-cell activation by reducing interferon-γ- induced major histocompatibility complex class II expression [12]. In addition, statins can block the effect of leukocyte

function-associated antigen 1 and simultaneously affect leukocyte endothelial interaction [13]. These effects of statins

cause a maladaptive dysregulated immune response and, thus, prevent sepsis. Second, the antiinflammatory response of statins [14] has been observed in several

studies; for example, simvastatin can inhibit the inflammatory response to Staphylococcus aureus [15]. Third, previous studies have shown that statins have direct antimicrobial effects; for example, lovastatin exerts antimicrobial effects on Salmonella typhimurium in cultured macrophages [16] and fluvastatin exerts antimicrobial effects on cytomegalovirus [17]. Statins and HZ infection

Previous studies that have discussed the relationship between herpes and statins have mainly focused on herpes simplex

virus type 1 (HSV-1) and Alzheimer’s disease. One study observed that statins can reduce the neuronal spread of HSV-1

[18]. Another study suggested that herpes virus infection can cause viral protein kinase-mediated cholesterol accumulation [19]. In addition, a previous study suggested that high cholesterol levels in blood were associated with HZ infection after

heart transplantation. The study observed that cholesterol can cause the reactivation and spread of the varicella zoster virus in vivo [20]. The present study revealed that statin users are at a higher risk of HZ infection than nonstatin users (HR=1.21,

95 % CI=1.13–1.29, p<0.001) (Table 2). In addition, the results

showed that statin users in Asia are at a higher risk of HZ

infection than those in Canada (HR=1.13) [8]. In the present

study, female statin users were at a higher risk of HZ infection

(HR=1.31, 95 % CI=1.20–1.43, p<0.001) than male statin

users (HR=1.10, 95 % CI=1.00–1.22, p>0.05). ATaiwanese metaanalysis showed no significant gender difference in the incidence of HZ infection [5]. The mechanism is unclear, but for public issue, female statin users are at a higher risk than male statin users in Asia. Another distinct result of the present study is that younger statin users are at a higher risk of HZ infection (<49 years, HR=1.35, p<0.01 vs.<64 years, HR=

1.33, p<0.001 vs.>65 years, HR=1.24, p<0.001). However, in the general population, the incidence of HZ infection increased with age [5]. We hypothesize that younger patients

consume a lipid-rich diet and, thus, have elevated blood cholesterol levels, leading to an increased possibility of the statins

being prescribed to treat dyslipidemia.

Different statins and herpes

A previous study suggested that different statins have different effects on HZ infection [11]. However, the present study

found that the dose was the only factor influencing HZ infection among various statins (Table 3). A higher DDD of statins

increased the risk of HZ infection. A high dose of statins is prescribed to patients with high cholesterol levels, and high cholesterol levels are associated with HZ infection [20]. In addition, we observed that the intensity or potency [2] of a

statin is not a crucial factor for HZ infection. Another factor for HZ infection is the statin exposure time. Longer exposure

to statins increases the risk of HZ infection (Fig. 2).

Conclusions

In Asia, statin users who are female, younger, consume a higher DDD of statins, and have a longer exposure time to statins are at a higher risk of HZ infection. We agree that statins should be prescribed to prevent and

reduce the incidence of cardiovascular as well as cerebrovascular diseases worldwide. However, we also suggest

that clinicians in Asia pay more attention to statin users with a high risk of HZ infection. For public health concerns in Asia, HZ vaccine may be a good option available to patients carrying a high risk of HZ infection before or after statin use.

Limitations

First, our data set did not contain information on some potential confounding factors, including body weight, nutritional

conditions, and smoking and alcohol consumption. Second,

although we examined the DDD of statins, we could not determine a detailed lipid profile before or after statin therapy.

This may be another factor confounding the results regarding HZ infection. Third, some medications, such as corticosteroids for rheumatic disease, may play an essential role in HZ

infection. Forth, we have to admit this is a possible bias that underdiagnosed HZ infection. However, the NHIRD covers a highly representative sample of Taiwan’s general population because the reimbursement policy is universal and

operated by a single buyer, the government in Taiwan.

All insurance claims should be scrutinized by medical reimbursement specialists and peer review according to the standard diagnosed criteria in the study. Therefore, the diagnoses of HZ infection in this study were highly reliable. Fifth, he NHIRD cannot provide the virus titers for individual study subjects. Therefore, the severity of HZ infection was not available in the study. In addition, asymptomatic HZ infection could be missed and

underdiagnosed in the analyses. However, in the present study,

we adjusted only the comorbidities and basic demographic

data in our analysis.