Abstract
Objective
To determine whether increased risk of acute ischemic stroke (AIS) following herpes zoster (HZ) might be modified by the status of zoster vaccine live (ZVL) vaccination and antiviral treatment following HZ.
Methods
We included 87,405 Medicare fee-for-service beneficiaries aged ≥66 years diagnosed with HZ and AIS from 2008 to 2017. We used a self-controlled case series design to examine the association between HZ and AIS, and estimated incidence rate ratios (IRRs) by comparing incidence of AIS in risk periods vs control periods. To examine effect modification by ZVL and antiviral treatment, beneficiaries were classified into 4 mutually exclusive groups: (1) no vaccination and no antiviral treatment; (2) vaccination only; (3) antiviral treatment only; and (4) both vaccination and antiviral treatment. We tested for interaction to examine changes in IRRs across 4 groups.
Results
Among 87,405 beneficiaries with HZ and AIS, 22.0%, 2.0%, 70.1%, and 5.8% were in groups 1 to 4, respectively. IRRs in 0–14, 15–30, 31–90, and 91–180 days following HZ were 1.89 (95% confidence interval [CI], 1.77–2.02), 1.58 (95% CI, 1.47–1.69), 1.36 (95% CI, 1.31–1.42), and 1.19 (95% CI, 1.15–1.23), respectively. There was no evidence of effect modification by ZVL and antiviral treatment on AIS (p = 0.067 for interaction). The pattern of association between HZ and risk for AIS was largely consistent across age group, sex, and race.
Conclusions
Risk of AIS increased significantly following HZ, and this increased risk was not modified by ZVL and antiviral treatment. Our findings suggest the importance of following recommended HZ vaccination in prevention of HZ and HZ-associated AIS.
Stroke is the 5th leading cause of death and causes serious long-term disability, with an estimated annual cost of $33.9 billion in 2015 in the United States.1 Accumulating evidence shows the importance of infectious causes of stroke.2–4 Herpes zoster (HZ), also known as shingles, is caused by reactivation of latent varicella-zoster virus infection generally acquired at young ages. Studies have suggested that HZ infection is associated with an increased risk of stroke, especially ischemic stroke.4–6 Almost 1 in 3 people in the United States will develop HZ during their lifetime, with an estimated 1 million HZ cases each year (cdc.gov/shingles/about/overview.html). More than half of individuals who developed HZ were aged 60 years or older.7,8 In 2006, the Advisory Committee on Immunization Practices (ACIP) recommended routine zoster vaccine live (ZVL) (Zostavax, a one-dose HZ live-attenuated vaccine manufactured by Merck & Co., Inc.) for preventing HZ for all persons aged >60 years,9 and in 2017, ACIP recommended a new HZ vaccine (Shingrix, manufactured by GlaxoSmithKline, a 2-dose, adjuvant, recombinant HZ vaccine [RZV]) for people aged 50 years and older.10 ZVL coverage increased continuously since 2006 and 33.0% of adults received ZVL in 2016.10 For patients with HZ, antiviral therapy is recommended.11
While many studies examined the association between HZ and risk for stroke, a few studies also examined if the risk of stroke after HZ might be modified by status of ZVL and antiviral treatment, with inconsistent findings. For example, some suggested that there was no difference in risk of stroke between patients with or without ZVL,12 others suggested that antiviral treatment was associated with reduced risk of stroke,13,14 and other studies suggested no difference in risk of stroke with antiviral treatment following HZ.15,16 To our knowledge, no study simultaneously examined the effect of ZVL and antiviral treatment on risk of acute ischemic stroke (AIS); the present study examined these effects among adults ≥66 years of age who enrolled in a Medicare fee-for-service (FFS) program from 2008 to 2017 in the United States.
Methods
Standard protocol approvals, registrations, and patient consents
The Centers for Disease Control and Prevention Human Subjects Coordinator determined that this study did not require review for human subjects protections because the data did not contain personal identifiers and were not originally collected specifically for this study. Therefore, the requirement of informed consent was waived.
Data source and study population
This population-based study utilized Medicare’s Enrollment Databases to generate study cohort and FFS Medicare Parts A (hospitalization), B (office-based care), and D (prescription drug coverage) claims data, and Medicare Provider Analysis and Review (MEDPAR) data to examine the association between HZ and risk of AIS from 2008 to 2017. We used the following procedures and definitions to select the final analytical cohort: (1) identified all Medicare FFS beneficiaries aged 65 years or older with at least 12 months continuous enrollment in Medicare Part A and B and eligible for Part D (≥1 month) during 2007–2017; (2) identified all diagnosed AIS among FFS beneficiaries during 2007–2017, including multiple admissions; (3) identified all diagnosed HZ among FFS beneficiaries during 2007–2017, including multiple admissions; (4) used ≥12 months lookback period of time to identify incident HZ (lookback period where no HZ codes were billed), and the length of lookback time varied by the years of Medicare enrollment, for example, 12 months for beneficiaries 66 years of age (Medicare eligible at age 65 years), 24 months for age 67 years, and so on. Among FFS beneficiaries with multiple admissions and/or >1 HZ occurrence as identified in step 3, we selected first HZ occurrence as the index case and the date of diagnosed HZ as index date; (5) used the same procedures to identify the incident AIS and its index date as we used for incident HZ and its index date; and (6) merged incident HZ and AIS data as identified in steps 4 and 5 to create the final analytical cohort. Because of ≥12 months lookback period, our final cohort included FFS beneficiaries aged ≥66 years or older (Medicare eligible at age 65 years) with diagnosed incident HZ and AIS from 2008 to 2017 (2007 served as look-back time). During 2007–2017, there were 210,197,235 FFS beneficiaries aged 65 years or older, among which there were 1,242,801 and 1,574,739 incident AIS and HZ among FFS beneficiaries aged 66 years or older from 2008 to 2017, respectively. Among FFS beneficiaries with incident AIS and HZ, we identified 87,405 beneficiaries who had both incident AIS and HZ during 2008–2017 (figure 1).
Figure 1.
Flowchart of Medicare fee-for-service beneficiaries with herpes zoster and acute ischemic stroke 2008–2017
Exposure, vaccination and antiviral treatment, and outcome
The exposure variable was HZ, and the index date of HZ for each FFS beneficiary was identified in the inpatient and outpatient claims data from January 1, 2008, to December 31, 2017. National Drug Code data from the Medicare Part D events data and Current Procedural Terminology code from inpatient and outpatient claims data were used to identify the status of ZVL,12 and beneficiaries were classified as vaccinated if the date of ZVL was before the index date of HZ. The antiviral treatment status was determined by prescription of antiviral medications ±7 days from the index date of HZ based on Medicare Part D data, and the ±7 days criteria were used to accommodate the difference in reporting dates of HZ diagnosis and antiviral treatment in Medicare data.17 The antiviral medications included acyclovir, famciclovir, and valacyclovir (table 1).16 Beneficiaries who had ZVL before the index date of HZ were defined as having ZVL. Beneficiaries who received antiviral treatment ±7 days from index date of HZ were defined as having HZ antiviral treatment. To examine the possible effect modification by ZVL and antiviral treatment on risk for AIS, beneficiaries were classified into 4 mutually exclusive groups based on the status of ZVL and antiviral treatment: (1) no vaccination and no antiviral treatment (n = 19,250); (2) vaccination only (n = 1,769); (3) antiviral treatment only (n = 61,298); and (4) with both vaccination and antiviral treatment (n = 5,088).
Table1.
Zoster vaccine live codes and antiviral treatment medications, Medicare fee-for-service beneficiaries 2008–2017
| Code | Codes and website link | |
|---|---|---|
| Zoster vaccine live (Zostavax) | National Drug Code | 00006496300 (hipaaspace.com/medical_billing/coding/national.drug.codes/00064963-00) |
| 00006496301 (hipaaspace.com/medical_billing/coding/national.drug.codes/00064963-00) | ||
| 00006496341 (hipaaspace.com/medical_billing/coding/national.drug.codes/00064963-41) | ||
| 54868570300 (bluecrossnc.com/sites/default/files/document/attachment/common/pdfs/2009_MAPD_Vaccine_Claim_Form.pdf) | ||
| Current Procedural Terminology code | 90,736 | |
| Antiviral treatment medications for herpes zoster | Acyclovir Famciclovir Valacyclovir |
CMS Chronic Conditions Data Warehouse (CCW)–Codebook Medicare Part D Event (PDE)/Drug Characteristics File Version 1.1 (January 2019) (ccwdata.org/web/guest/data-dictionaries) |
We used MEDPAR files to identify AIS, the outcome of interest. The MEDPAR files contained inpatient hospital and skilled nursing facility stay records for all Medicare beneficiaries, and we used the primary diagnosis codes (ICD-9-CM for 2007–2015 and ICD-10-CM for last quarter of 2015 and 2016–2017) to identify beneficiaries with AIS. If the beneficiaries had a diagnosis of stroke (any type, including TIA) based on the Chronic Conditions Warehouse definition used by Centers for Medicare and Medicaid Services (CMS) (ccwdata.org/web/guest/condition-categories) that occurred >30 days before the index date of AIS, they were classified as having history of stroke. Table 2 provides the ICD-9-CM and ICD-10-CM codes for HZ and AIS, respectively.
Table 2.
ICD-9-CM and ICD-10-CM codes for herpes zoster and acute ischemic stroke, Medicare fee-for-service beneficiaries 2008–2017
| Disease | Inclusion criteria | CD-CM codes | Description |
|---|---|---|---|
| Herpes zoster | ICD-9-CM codes | ||
| Including | 053.00 | Herpes zoster with meningitis | |
| 053.10 | Herpes zoster with unspecified nervous system complication | ||
| 053.14 | Herpes zoster myelitis | ||
| 053.19 | Herpes zoster with other nervous system complications | ||
| 053.20 | Herpes zoster dermatitis of eyelid | ||
| 053.21 | Herpes zoster keratoconjunctivitis | ||
| 053.22 | Herpes zoster iridocyclitis | ||
| 053.29 | Herpes zoster with other ophthalmic complications | ||
| 053.71 | Otitis externa due to herpes zoster | ||
| 053.79 | Herpes zoster with other specified complications | ||
| 053.8 | Herpes zoster with unspecified complication | ||
| 053.9 | Herpes zoster without complication | ||
| Excluding | 053.12 | Postherpetic trigeminal neuralgia | |
| 053.13 | Postherpetic polyneuropathy | ||
| ICD-10-CM codes | |||
| Including | B02 | Zoster | |
| Excluding | B0222 | Zoster with other nervous system involvement | |
| B0223 | Zoster ocular disease | ||
| Acute ischemic stroke | ICD-9-CM codes | ||
| 433.01 | Occlusion and stenosis of basilar artery with cerebral infarction | ||
| 433.11 | Occlusion and stenosis of carotid artery with cerebral infarction | ||
| 433.21 | Occlusion and stenosis of vertebral artery with cerebral infarction | ||
| 433.31 | Occlusion and stenosis of multiple and bilateral precerebral arteries with cerebral infarction | ||
| 433.81 | Occlusion and stenosis of other specified precerebral artery with cerebral infarction | ||
| 433.91 | Occlusion and stenosis of unspecified precerebral artery with cerebral infarction | ||
| 434.01 | Cerebral thrombosis with cerebral infarction | ||
| 434.11 | Cerebral embolism with cerebral infarction | ||
| 434.91 | Cerebral artery occlusion unspecified with cerebral infarction | ||
| 436 | Acute but ill-defined cerebrovascular disease | ||
| ICD-10-CM codes | |||
| I63 | Cerebral infarction | ||
| I66 | Occlusion and stenosis of cerebral arteries | ||
| I6789 | Other cerebrovascular disease |
Abbreviations: ICD-9-CM = International Classification of Diseases, 9th revision; ICD-10-CM = International Classification of Diseases, 10th revision.
Statistical analysis
We calculated the mean age (±SD) and distribution of age group, sex, race, and AIS with and without history of stroke by 4 ZVL and antiviral treatment groups. We used self-controlled case series (SCCS) study design to estimate incidence rate ratio (IRR) and 95% confidence interval (CI) for risk of AIS following exposure to HZ. As shown in figure 2, the SCCS study design is based on within-person comparisons (self-matched) after exposure during an observation period subdivided into risk and control periods, and this method implicitly controls for all fixed confounders during the period of study.18 The null hypothesis, IRR = 1.0, implies that AIS event rates remained constant during the entire observation period of time and were not affected by HZ. An IRR >1.0 or <1.0 implies an increased or reduced risk of AIS following HZ. The start of the observation period was January 1, 2008, and for those beneficiaries who enrolled after January 1, 2008, it was the January 1 of the first enrollment year. The end of the observation period was December 31, 2017, or the date of death for those who died before December 31, 2017. We used the National Death Index linked to Medicare data to determine the date of death until December 31, 2016, and date of death in beneficiary file in 2017. We categorized the baseline risk after HZ into 4 periods: 0–14, 15–30, 31–90, and 91–180 days. The findings of other studies suggested that the elevated risk of stroke after HZ gradually returned to the baseline (IRR ≈ 1.0) within 1 year.12,14 One of 2 key assumptions of SCCS design is that the occurrence of events (AIS in our case) does not influence the length of observation period.19 Among beneficiaries with AIS, 39% of them (n = 34,119) died before the end of the observation period. Thus, the assumption of event being independent of observation period was violated (due to increased mortality after AIS), so we used the modified SCCS method that takes into account the event dependent observation period of time.19–21 We conducted a test for interaction in the modified SCCS model to examine whether the IRRs for risk of AIS changed significantly across 4 groups classified by status of ZVL and antiviral treatment based on the likelihood ratio tests.18,21 To correct for multiple hypothesis testing, we calculated the false discovery rate (FDR) and denoted significance by an FDR threshold of 5%, and reported the FDR-adjusted p value for interaction.22 Beneficiaries with and without history of stroke might have different risk profiles,23,24 so we also presented results by status of history of stroke (first vs recurrent stroke). The IRRs were adjusted for age as a categorical variable of 5-year age group from 66 to ≥95 years of age.
Figure 2.
Graphic representation of self-controlled case series study design
(A) A patient who had acute ischemic stroke in the risk period during exposure. (B) A patient who had acute ischemic stroke during the control period.
We performed analyses in subgroups defined according to age group (66–74 years, 75–84 years, and ≥85 years), sex, and race (non-Hispanic white, non-Hispanic black, Hispanics, and others). For analyses stratified by demographics, we combined the 4 vaccination and antiviral groups since there was no evidence of effect modification in risk for AIS and to increase the sample size for stable estimates. We evaluated the presence of interaction in these subgroups. SAS version 9.4 (SAS Institute, Cary, NC) was used for analysis, and R package SCCS was used for SCCS analyses.21
Data availability
Medicare data are available from CMS, Department of Health and Human Services, for any qualified investigator.
Results
Description of Medicare beneficiaries
Among 87,405 beneficiaries included in the study, the mean age ±SD was 80.0 ± 7.5 years, 19,250 (22.0%) did not receive ZVL and did not have antiviral treatment, 1,769 (2.0%) had ZVL only, 61,298 (70.1%) received antiviral treatment only, and 5,088 (5.8%) had both vaccine and antiviral treatment (table 3). ZVL coverage increased with age and the median time from ZVL to the index date of HZ was 33 months (interquartile range, 15–56). The antiviral treatment decreased with age, 57,806 (66.1%) of beneficiaries were women, and there were no difference in ZVL and receiving antiviral treatment between men and women. Non-Hispanic black and Hispanic beneficiaries were less likely to receive ZVL or to receive both ZVL and antiviral treatment as compared to non-Hispanic white or other participants, but more likely to receive antiviral treatment after HZ. Among all cases, 59,294 (68%) had no prior history of stroke (table 3).
Table 3.
Characteristics of Medicare fee-for-service (FFS) beneficiaries aged ≥66 years with diagnosed herpes zoster (HZ) and acute ischemic stroke by status of zoster vaccine live (ZVL) and antiviral treatment, Medicare 2008–2017
| Status of ZVL and antiviral treatment | ||||||
|---|---|---|---|---|---|---|
|
|
||||||
| Characteristics | Medicare FFS beneficiaries, n | Total | No vaccination and no antiviral treatment | Vaccination onlya | Antiviral treatment only | Vaccination and antiviral treatmenta |
|
| ||||||
| All, n (%) | 87,405 | 100 | 19,250 (22.0) | 1,769 (2.0) | 61,298 (70.1) | 5,088 (5.8) |
|
| ||||||
| Age, y, mean ± SD | 80.0 ± 7.5 | 80.5 ± 7.7 | 79.9 ± 6.9 | 79.8 ± 7.5 | 80.4 ± 6.8 | |
|
| ||||||
| Age group, y, n (%) | ||||||
|
| ||||||
| 66–74 | 25,523 | 100 | 5,283 (20.7) | 485 (1.9) | 18,527 (72.6) | 1,228 (4.8) |
|
| ||||||
| 75–84 | 38,139 | 100 | 8,158 (21.4) | 826 (2.2) | 26,650 (69.9) | 2,505 (6.6) |
|
| ||||||
| ≥85 | 23,743 | 100 | 5,809 (24.5) | 458 (1.9) | 16,121 (67.9) | 1,355 (5.7) |
|
| ||||||
| Sex, n (%) | ||||||
|
| ||||||
| Men | 29,599 | 100 | 6,455 (21.8) | 626 (2.1) | 20,766 (70.2) | 1,752 (5.9) |
|
| ||||||
| Women | 57,806 | 100 | 12,795 (22.1) | 1,143 (2.0) | 40,532 (70.1) | 3,336 (5.8) |
|
| ||||||
| Race, n (%) | ||||||
|
| ||||||
| Non-Hispanic white | 75,364 | 100 | 16,252 (21.6) | 1,592 (2.1) | 52,839 (70.1) | 4,681 (6.2) |
|
| ||||||
| Non-Hispanic black | 4,933 | 100 | 1,315 (26.7) | 31 (0.6) | 3,508 (71.1) | 79 (1.6) |
|
| ||||||
| Hispanics | 4,107 | 100 | 1,010 (24.6) | 45 (1.1) | 2,957 (72.0) | 95 (2.3) |
|
| ||||||
| Other | 3,001 | 100 | 673 (22.4) | 101 (3.4) | 1,994 (66.4) | 233 (7.8) |
|
| ||||||
| Without history of stroke, n (%) | 59,294 | 100 | 12,639 (21.3) | 1,237 (2.1) | 41,768 (70.4) | 3,650 (6.2) |
|
| ||||||
| With history of stroke, n (%) | 28,111 | 100 | 6,611 (23.5) | 532 (1.9) | 19,530 (69.5) | 1,438 (5.1) |
The median month from ZVL to the index date of HZ was 33 (ranging from 1 to 130 months with an interquartile range 15–56).
IRR for AIS
The IRRs for AIS in 0–14, 15–30, 31–90, and 91–180 days following HZ were 1.89 (95% CI, 1.77–2.02), 1.58 (95% CI, 1.47–1.69), 1.36 (95% CI, 1.31–1.42), and 1.19 (95% CI, 1.15–1.23), respectively. There was no evidence of effect modification by the status of ZVL and antiviral treatment (FDR-adjusted p = 0.067 for interaction) (table 4). The pattern of null effect modification was similar among beneficiaries diagnosed with or without history of stroke (FDR-adjusted p = 0.171 and 0.097 for interaction, respectively [table 4]). The pattern of association between HZ and risk for AIS was largely consistent across demographic subgroups (table 5). However, the risk of stroke in 31–90 days following HZ were significantly higher among men compared to women (1.50 [95% CI, 1.40–1.60] vs 1.30 [95% CI, 1.23–1.36]) and was significantly higher among Hispanic patients in 90–180 days following HZ compared to other racial groups (table 5).
Table 4.
Incidence rate ratio (95% confidence interval) of acute ischemic stroke (AIS) among Medicare fee-for-service beneficiaries diagnosed with herpes zoster (HZ), Medicare 2008–2017
| Status of zoster vaccine live (ZVL) and antiviral treatment | |||||
|---|---|---|---|---|---|
|
|
|||||
| AIS and risk period | Total | No vaccination and no antiviral treatment | Vaccination only | Antiviral treatment only | Vaccination and antiviral treatment |
|
| |||||
| Total AIS a | |||||
|
| |||||
| AIS, n (%) | 87,405 (100) | 19,250 (22.0) | 1,769 (2.0) | 61,298 (70.1) | 5,088 (5.8) |
|
| |||||
| 1–14 days | 1.89 (1.77–2.02) | 2.08 (1.83–2.36) | 1.66 (1.01–2.74) | 1.91 (1.76–2.07) | 1.39 (0.99–1.95) |
|
| |||||
| 15–30 days | 1.58 (1.47–1.69) | 1.51 (1.31–1.74) | 1.47 (0.86–2.49) | 1.68 (1.55–1.83) | 1.10 (0.77–1.59) |
|
| |||||
| 31–90 days | 1.36 (1.31–1.42) | 1.49 (1.37–1.61) | 1.30 (0.97–1.74) | 1.33 (1.26–1.39) | 1.50 (1.28–1.77) |
|
| |||||
| 91–180 days | 1.19 (1.15–1.23) | 1.22 (1.13–1.31) | 1.33 (1.04–1.70) | 1.19 (1.14–1.25) | 1.09 (0.93–1.28) |
|
| |||||
| Without history of stroke only b | |||||
|
| |||||
| AIS, n (%) | 59,294 (100) | 12,639 (21.3) | 1,237 (2.1) | 41,768 (70.4) | 3,650 (6.2) |
|
| |||||
| 1–14 days | 1.87 (1.72–2.03) | 1.99 (1.70–2.34) | 1.78 (1.00–3.16) | 1.90 (1.72–2.10) | 1.10 (0.71–1.72) |
|
| |||||
| 15–30 days | 1.62 (1.49–1.77) | 1.53 (1.28–1.83) | 1.64 (0.91–2.99) | 1.71 (1.55–1.90) | 0.98 (0.62–1.54) |
|
| |||||
| 31–90 days | 1.38 (1.32–1.45) | 1.47 (1.34–1.63) | 1.15 (0.79–1.67) | 1.36 (1.28–1.44) | 1.42 (1.17–1.72) |
|
| |||||
| 91–180 days | 1.20 (1.15–1.25) | 1.24 (1.13–1.36) | 1.09 (0.79–1.50) | 1.20 (1.14–1.27) | 1.03 (0.85–1.24) |
|
| |||||
| With history of stroke only c | |||||
|
| |||||
| AIS, n (%) | 28,111 (100) | 6,611 (23.5) | 532 (1.9) | 19,530 (69.5) | 1,438 (5.1) |
|
| |||||
| 1–14 days | 1.97 (1.76–2.20) | 2.21 (1.81–2.71) | 1.42 (0.53–3.85) | 1.90 (1.65–2.18) | 1.87 (1.10–3.17) |
|
| |||||
| 15–30 days | 1.52 (1.34–1.71) | 1.47 (1.16–1.87) | 1.03 (0.33–3.22) | 1.57 (1.36–1.82) | 1.22 (0.65–2.28) |
|
| |||||
| 31–90 days | 1.32 (1.23–1.42) | 1.50 (1.32–1.71) | 1.62 (1.00–2.60) | 1.24 (1.14–1.35) | 1.45 (1.07–1.96) |
|
| |||||
| 91–180 days | 1.18 (1.11–1.26) | 1.18 (1.04–1.34) | 1.95 (1.34–2.83) | 1.17 (1.08–1.26) | 1.07 (0.80–1.43) |
False discovery rate–adjusted p value = 0.067 for interaction between HZ and the status of ZVL and antiviral treatment after HZ based on likelihood ratio test.
False discovery rate–adjusted p value = 0.097 for interaction between HZ and status of ZVL and antiviral treatment after HZ on risk of AIS based on likelihood ratio test.
False discovery rate–adjusted p value = 0.171 for interaction between HZ and status of ZVL and antiviral treatment after HZ on risk of AIS among beneficiaries with history of stroke based on likelihood ratio test.
Table5.
Incidence rate ratio (95% confidence interval [CI]) for acute ischemic stroke (AIS) among Medicare fee-for-service beneficiaries diagnosed with herpes zoster (HZ) by selected characteristics, Medicare 2008–2017
| Incidence rate ratio (95% CI) after HZ | |||||
|---|---|---|---|---|---|
|
|
|||||
| Characteristics | AIS, n | 0–14 days | 15–30 days | 31–90 days | 91–180 days |
|
| |||||
| Age group, y a | |||||
|
| |||||
| 66–74 | 25,523 | 2.01 (1.77–2.28) | 1.64 (1.43–1.87) | 1.43 (1.32–1.54) | 1.28 (1.20–1.37) |
|
| |||||
| 75–84 | 38,139 | 1.81 (1.63–2.01) | 1.42 (1.27–1.60) | 1.34 (1.26–1.42) | 1.11 (1.05–1.18) |
|
| |||||
| ≥85 | 23,743 | 1.75 (1.56–1.97) | 1.57 (1.39–1.76) | 1.21 (1.13–1.31) | 1.12 (1.05–1.20) |
|
| |||||
| Sex b | |||||
|
| |||||
| Men | 29,599 | 1.75 (1.55–1.97) | 1.64 (1.45–1.85) | 1.50 (1.40–1.60) | 1.23 (1.16–1.31) |
|
| |||||
| Women | 57,806 | 1.99 (1.84–2.16) | 1.57 (1.44–1.71) | 1.30 (1.23–1.36) | 1.18 (1.13–1.23) |
|
| |||||
| Race c | |||||
|
| |||||
| Non-Hispanic white | 75,364 | 1.84 (1.71–1.97) | 1.60 (1.48–1.72) | 1.40 (1.34–1.46) | 1.20 (1.15–1.25) |
|
| |||||
| Non-Hispanic black | 4,923 | 2.31 (1.82–2.94) | 1.70 (1.30–2.24) | 1.18 (1.00–1.41) | 1.05 (0.90–1.23) |
|
| |||||
| Hispanic | 4,107 | 2.43 (1.86–3.17) | 1.40 (1.00–1.96) | 1.22 (1.00–1.48) | 1.45 (1.25–1.69) |
|
| |||||
| Other | 3,001 | 2.08 (1.47–2.94) | 1.36 (0.91–2.05) | 1.08 (0.85–1.37) | 0.95 (0.77–1.17) |
False discovery rate–adjusted p value = 0.067 for interaction between HZ and age group on risk for AIS based on likelihood ratio test.
False discovery rate–adjusted p value = 0.012 for interaction between HZ and sex on risk for AIS based on likelihood ratio test.
False discovery rate–adjusted p value = 0.008 for interaction between HZ and race on risk for AIS based on likelihood ratio test.
Discussion
Our findings from this large Medicare cohort study using SCCS design suggested that the risk of AIS increased significantly in the first 6 months following HZ. The risk was highest within the first 2 weeks following HZ and reduced gradually over time, consistent with the findings of previous studies.4–6 The pattern of association was consistent among beneficiaries with and without history of stroke. To our knowledge, this is the first study to examine simultaneously the effect modification by the status of ZVL and antiviral treatment following HZ on risk of AIS. Our results provided no evidence of such modification on risk of AIS after HZ. It appears that having HZ vaccine before symptoms of HZ appear might be the most effective way to prevent HZ-associated stroke risk. It is worth noting that only a small fraction of beneficiaries (5.8%) among this Medicare cohort received recommended ZVL and were treated with antiviral medications after HZ.9
Previous findings of effect modification by ZVL or antiviral treatment were inconsistent. Minassian et al.12 studied Medicare beneficiaries ≥65 years of age with HZ diagnosis and either an ischemic stroke or myocardial infarction between 2006 and 2011, and found no evidence of difference in IRRs for ischemic stroke between vaccinated and unvaccinated beneficiaries. Two studies suggested that antiviral treatment following HZ was associated with reduced risk of stroke.13,14 One of these studies included patients with autoimmune disease.13 Two matched cohort studies from Taiwan and Denmark suggested that antiviral treatment of HZ had no effect on the incidence of stroke.15,16 The Denmark study used antiviral treatment following HZ to identify the patients with HZ and compared patients with HZ with those without HZ.16 Both studies were subject to the effects of uncontrolled confounders. Differences in study designs, sample size, patients studied (e.g., patients with autoimmune disease), and definition of stroke included in the study (ischemic or hemorrhagic stroke or TIA) might contribute to the inconsistent findings. Our study included a large number of Medicare beneficiaries diagnosed with AIS, and with an adequate number to classify beneficiaries into 4 mutually exclusive groups based on status of ZVL and antiviral treatment. Both ZVL and antiviral treatment are proven to reduce the severity of HZ.9,11 Our findings indicated that reducing the HZ severity by these means has no impact on AIS risk and the elevated risk of AIS after HZ might be independent of ZVL history and antiviral treatments after HZ.
Our results suggested that the risk of AIS in 31–90 days following HZ was higher among men compared to women (IRR 1.50 [95% CI, 1.40–1.60] vs 1.30 [95% CI, 1.23–1.36]). HZ incidence was generally higher in women than in men, as observed in previous studies,6,25,26 but the reasons for the observed higher risk for stroke in 31–90 days among men compared to women were unclear; however, this difference in risk of AIS has limited clinical significance in prevention of AIS following HZ. In addition, the pattern of association between HZ and risk of AIS following HZ among Hispanic patients appeared to be different compared to that of other racial groups. The risk for AIS among Hispanic patients was consistently higher from within 2 weeks to 91–180 days following HZ, especially in 91–180 days (IRR 1.45 [95% CI, 1.25–1.67]); the significant likelihood ratio test for interaction by status of ZVL and antiviral treatment was mainly driven by this difference in IRR in 91–180 days following HZ. We are not aware of any publication on racial differences in effects of HZ on risk of AIS. Other studies suggested that Hispanic patients were significantly less likely to seek health care after HZ,27 and there were racial disparities in general health care utilizations among older US adults that might contribute to the persistent risk of AIS following HZ among Hispanic patients.28
HZ infection involves viral invasion of the arterial walls, and can lead to granulomatous and necrotizing vasculitis.29 Possible biological explanation for the association between HZ and stroke is the HZ viral ability to replicate in cerebral arteries where the infection is spread along the nerve fibers to the blood vessels with subsequent inflammation leading to pathologic vascular remodeling and increased risk of stroke.30,31 This mechanism might provide partial explanation for the null effect modification by ZVL, for example, after the reactivation of HZ, its viral ability to replicate in cerebral arteries, and the associated inflammation responses might not be altered by ZVL status. In addition, the recipients of ZVL had a 51% overall reduction in HZ and the vaccine efficacy reduced significantly by age from 64% (95% CI, 56% to 71%) for 60–69 years, 41% (95% CI, 28% to 52%) for 70–79 years, to 18% (95% CI, −29 to 48%) among persons ≥80 years of age (merck.com/product/usa/pi_circulars/z/zostavax/zostavax_pi2.pdf). The mean age of the beneficiaries in our study was 80.0 years, which might also contribute to the null effect of ZVL on risk for AIS. However, this might not explain the null effect of antiviral treatment following HZ, and reasons for the null effect of antiviral treatment on risk for stroke remain unclear. The main goal of treatment after HZ among elderly patients is the reduction or elimination of pain. Three antiviral drugs (acyclovir, famciclovir, and valacyclovir) are guanosine analogs that are phosphorylated by viral thymidine kinase and cellular kinases to a triphosphate form that inhibits varicella-zoster virus DNA polymerase.32 However, we cannot rule out the findings by chance or effect of other factors such as the differences in seeking health care for HZ by age, sex, race, ethnicity, and HZ severity, and further studies are needed to elucidate the possible mechanisms underlying null effect of ZVL and antiviral treatment, and racial and sex differences in HZ-related risk for AIS. Our study of a large national cohort of Medicare beneficiaries had adequate numbers to classify all beneficiaries into mutually exclusive groups to examine the effect modification of ZVL and antiviral treatment following HZ on risk of AIS. As each individual treats himself or herself as the control (within-individual comparison) under the SCCS study design, only individuals who had HZ and AIS were included, therefore all fixed confounding effects were eliminated.
Our study had several limitations. First, HZ diagnosis was based on administrative data and was not able to be verified. Other studies have combined ICD-9 codes and antiviral treatment in Medicare data to identify HZ cases,12,13 and had a positive predicted value >85% for HZ.13 In order to examine effect modification by antiviral treatment, we used ICD-9-CM and ICD-10-CM codes to identify HZ and misclassification is possible. However, when we restricted our analysis to the beneficiaries identified by combined ICD-9-CM and ICD-10CM codes and antiviral treatment, the pattern of association remained unchanged (IRRs in 0–14, 15–30, 31–90, and 91–180 days following HZ were 1.87 [95% CI, 1.73–2.02], 1.63 [95% CI, 1.51–1.77], 1.33 [95% CI, 1.27–1.40], and 1.18 [95% CI, 1.14–1.23], respectively). Second, Zostavax, a live and attenuated vaccine, was the only vaccine that was approved by the Food and Drug Administration for preventing HZ for all persons aged >60 years during the study period.9 The usual age of eligibility for Medicare is 65 years, and if beneficiaries had ZVL before age 65 years, this information was not captured in Medicare data. The efficacy of ZVL is higher among adults aged 60–65 compared to older age groups, and misclassification resulting from missing information for adults aged 60–65 years might affect our results. This is an important limitation, and further study is needed to clarify its effect on risk of AIS. In 2017, ACIP recommended RZV (Shingrix) for people ≥50 years of age with higher efficacy for prevention of HZ, and its possible effect on AIS following HZ remains to be determined.10 Third, similar to ZVL vaccination, if beneficiaries had any stroke before age 65 years, this information was not available in Medicare data and some recurrent stroke could be classified as first stroke. Fourth, dose and duration of antiviral therapy for HZ varied between patients, and this information was not available in Medicare; further study is needed to examine their effect on risk of stroke following HZ. Fifth, another study suggested that there were moderate differences in seeking care for HZ by age, sex, race/ethnicity, and HZ severity.27 Medicare data might miss those who did not seek for care after HZ. Sixth, the efficacy of ZVL declines over time, and the median time of ZVL to HZ was 33 months (ranging from 1 to 130 months with an interquartile range 15–56). In our study, we have limited sample size for a detailed analysis of timing of ZVL on risk of AIS. Seventh, the AIS diagnosis was based on ICD-9-CM and ICD-10-CM codes from the Medicare claims data and was not verified in this study, and this could also lead to misclassification. However, other studies validated the selected diagnostic codes for AIS with relatively high positive predictive values.33,34 Finally, other studies suggested stronger association between HZ ophthalmicus and risk of stroke.4,15 The sample size for HZ ophthalmicus from the Medicare data is limited to examine the effect modification by ZVL and antiviral treatment.
Risk of AIS increased significantly following HZ, and this increased risk was not modified by ZVL and antiviral treatment. Our findings underscore the importance of following HZ vaccination recommendations in order to prevent HZ and HZ-associated AIS.
Acknowledgments
Study funding
No targeted funding reported.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Glossary
- ACIP
Advisory Committee on Immunization Practices
- AIS
acute ischemic stroke
- CI
confidence interval
- CMS
Centers for Medicare and Medicaid Services
- FFS
fee-for-service
- HZ
herpes zoster
- ICD-9-CM
International Classification of Diseases, 9th revision, Clinical Modification
- ICD-10-CM
International Classification of Diseases, 10th revision, Clinical Modification
- IRR
incidence rate ratio
- MEDPAR
Medicare Provider Analysis and Review
- RZV
recombinant herpes zoster vaccine
- SCCS
self-controlled case series
- ZVL
zoster vaccine live
APPENDIX
Appendix
Appendix Authors
| Name | Location | Contribution |
| Quanhe Yang, PhD | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Designed and conceptualized study, conducted statistical analysis, drafted the manuscript for intellectual content |
| Mary G. George, MD, MSPH | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Interpreted the data, revised the manuscript for intellectual content |
| Anping Chang, MPH | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Generated the analytical datasets |
| Xin Tong, MPH | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Validated and generated analytical datasets |
| Robert Merritt, MS | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Interpreted the data, revised the manuscript for intellectual content |
| Yuling Hong, MD, PhD | Division for Heart Disease and Stroke Prevention, Centers for Disease Control and Prevention, Atlanta, GA | Interpreted the data, revised the manuscript for intellectual content |
Footnotes
Disclosure
The authors report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.
References
- 1.Benjamin EJ, Blaha MJ, Chiuve SE, et al. Heart Disease and Stroke Statistics: 2017 update: a report from the American Heart Association. Circulation 2017;135:e146–e603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fugate JE, Lyons JL, Thakur KT, Smith BR, Hedley-Whyte ET, Mateen FJ. Infectious causes of stroke. Lancet Infect Dis 2014;14:869–880. [DOI] [PubMed] [Google Scholar]
- 3.Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial infarction and stroke after acute infection or vaccination. N Engl J Med 2004;351:2611–2618. [DOI] [PubMed] [Google Scholar]
- 4.Zhang Y, Luo G, Huang Y, Yu Q, Wang L, Li K. Risk of stroke/transient ischemic attack or myocardial infarction with herpes zoster: a systematic review and metaanalysis. J Stroke Cerebrovasc Dis 2017;26:1807–1816. [DOI] [PubMed] [Google Scholar]
- 5.Schink T, Behr S, Thone K, Bricout H, Garbe E. Risk of stroke after herpes zoster: evidence from a German self-controlled case-series study. PLoS One 2016;11: e0166554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kim MC, Yun SC, Lee HB, et al. Herpes zoster increases the risk of stroke and myocardial infarction. J Am Coll Cardiol 2017;70:295–296. [DOI] [PubMed] [Google Scholar]
- 7.Hales CM, Harpaz R, Joesoef MR, Bialek SR. Examination of links between herpes zoster incidence and childhood varicella vaccination. Ann Intern Med 2013;159:739–745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sampathkumar P, Drage LA, Martin DP. Herpes zoster (shingles) and postherpetic neuralgia. Mayo Clin Proc 2009;84:274–280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Harpaz R, Ortega-Sanchez IR, Seward JF; Advisory Committee on Immunization Practices, Centers for Disease Control and Prevention. Prevention of herpes zoster: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2008;57:1–30; quiz CE32–34. [PubMed] [Google Scholar]
- 10.Dooling KL, Guo A, Patel M, et al. Recommendations of the Advisory Committee on Immunization Practices for use of herpes zoster vaccines. MMWR Morb Mortal Wkly Rep 2018;67:103–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Schmader K. Herpes zoster. Ann Intern Med 2018;169:ITC19–ITC31. [DOI] [PubMed] [Google Scholar]
- 12.Minassian C, Thomas SL, Smeeth L, Douglas I, Brauer R, Langan SM. Acute cardiovascular events after herpes zoster: a self-controlled case series analysis in vaccinated and unvaccinated older residents of the United States. PLoS Med 2015;12: e1001919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Calabrese LH, Xie F, Yun H, et al. Herpes zoster and the risk of stroke in patients with autoimmune diseases. Arthritis Rheumatol 2017;69:439–446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Langan SM, Minassian C, Smeeth L, Thomas SL. Risk of stroke following herpes zoster: a self-controlled case-series study. Clin Infect Dis 2014;58:1497–1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lin HC, Chien CW, Ho JD. Herpes zoster ophthalmicus and the risk of stroke: a population-based follow-up study. Neurology 2010;74:792–797. [DOI] [PubMed] [Google Scholar]
- 16.Sreenivasan N, Basit S, Wohlfahrt J, et al. The short- and long-term risk of stroke after herpes zoster: a nationwide population-based cohort study. PLoS One 2013;8:e69156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhang J, Xie F, Delzell E, et al. Association between vaccination for herpes zoster and risk of herpes zoster infection among older patients with selected immune-mediated diseases. JAMA 2012;308:43–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Whitaker HJ, Farrington CP, Spiessens B, Musonda P. Tutorial in biostatistics: the self-controlled case series method. Stat Med 2006;25:1768–1797. [DOI] [PubMed] [Google Scholar]
- 19.Whitaker HJ, Ghebremichael-Weldeselassie Y, Douglas IJ, Smeeth L, Farrington CP. Investigating the assumptions of the self-controlled case series method. Stat Med 2018;37:643–658. [DOI] [PubMed] [Google Scholar]
- 20.Farrington CP, Anaya-Izquierdo K, Whitaker HJ, Hocine MN, Douglas I, Smeeth L. Self-controlled case series analysis with event-dependent observation periods. J Am Stat Assoc 2011;106:417–426. [Google Scholar]
- 21.Farrington P, Whitaker H, Ghebremichael Weldeselassie Y. Self-Controlled Case Series Studies: A Modelling Guide with R. Boca Raton: CRC Press, Taylor & Francis Group; 2018. [Google Scholar]
- 22.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Stat Methodol 1995;57:289–300. [Google Scholar]
- 23.Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2014;45: 2160–2236. [DOI] [PubMed] [Google Scholar]
- 24.Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol 2019;74:1376–1414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Esteban-Vasallo MD, Dominguez-Berjon MF, Gil-Prieto R, Astray-Mochales J, Gil de Miguel A. Temporal trends in herpes zoster-related hospitalizations in Madrid (Spain), 2003–2013. J Infect 2015;71:85–92. [DOI] [PubMed] [Google Scholar]
- 26.Yawn BP, Saddier P, Wollan PC, St Sauver JL, Kurland MJ, Sy LS. A population-based study of the incidence and complication rates of herpes zoster before zoster vaccine introduction. Mayo Clin Proc 2007;82:1341–1349. [DOI] [PubMed] [Google Scholar]
- 27.Hales CM, Harpaz R, Bialek SR. Self-reported herpes zoster, pain, and health care seeking in the Health and Retirement Study: implications for interpretation of health care-based studies. Ann Epidemiol 2016;26:441–446 e443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dunlop DD,Manheim LM,Song J,Chang RW.Gender andethnic/racial disparitiesin health care utilization among older adults. J Gerontol B Psychol Sci Soc Sci 2002;57:S221–S233. [DOI] [PubMed] [Google Scholar]
- 29.Grau AJ, Urbanek C, Palm F. Common infections and the risk of stroke. Nat Rev Neurol 2010;6:681–694. [DOI] [PubMed] [Google Scholar]
- 30.Gilden D, Nagel MA, Mahalingam R, et al. Clinical and molecular aspects of varicella zoster virus infection. Future Neurol 2009;4:103–117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nagel MA, Gilden D. The relationship between herpes zoster and stroke. Curr Neurol Neurosci Rep 2015;15:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Yoshikawa TT, Schmader K. Herpes zoster in older adults. Clin Infect Dis 2001;32:1481–1486. [DOI] [PubMed] [Google Scholar]
- 33.Leibson CL, Naessens JM, Brown RD, Whisnant JP. Accuracy of hospital discharge abstracts for identifying stroke. Stroke 1994;25:2348–2355. [DOI] [PubMed] [Google Scholar]
- 34.Tirschwell DL, Longstreth WT Jr. Validating administrative data in stroke research. Stroke 2002;33:2465–2470. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Medicare data are available from CMS, Department of Health and Human Services, for any qualified investigator.