Association between Stroke Risk and Influenza Vaccination in Patients with Gout: A Nationwide Population-Based Study

The risk of stroke in patients with gout is high. The effect of vaccines in lowering the stroke risk in patients with gout remains unclear. We retrospectively analyzed 23,949 patients with gout (age ≥ 55 years) from the National Health Insurance Research Database over a 12-year period. The patients were divided into vaccinated (n = 11,649) and unvaccinated groups (n = 12,300). Overall, the vaccinated group had significantly lower risks of all stroke, hemorrhagic stroke, and ischemic stroke than the unvaccinated group (adjusted hazard ratio [aHR], 0.59 and 95% confidence interval [CI], 0.55–0.63; aHR, 0.60 and 95% CI, 0.49–0.73; and aHR, 0.60 and 95% CI, 0.55–0.65, respectively). The association appeared to be dose-dependent for both hemorrhagic and ischemic stroke (hemorrhagic stroke: aHR, 0.81 and 95% CI, 0.61–1.08; aHR, 0.80 and 95% CI, 0.62–1.02; and aHR, 0.37 and 95% CI, 0.28–0.48; ischemic stroke: aHR, 0.83 and 95% CI, 0.74–0.94; aHR, 0.73 and 95% CI, 0.65–0.81; and aHR, 0.42 and 95% CI, 0.38–0.47 for patients vaccinated 1, 2 or 3, and ≥4 times, respectively, during the follow-up period). Patients with a history of atrial fibrillation did not have a lower risk of hemorrhagic stroke even after receiving four vaccinations (aHR, 0.59; 95% CI, 0.25–1.38). Influenza vaccination was associated with a lower risk of all stroke in people with gout, and the association appeared to be dose-dependent.


Introduction
Gout is the most common type of inflammatory arthritis in adults [1]. The prevalence of gout varies from 0.4% to 6.1% in most developed countries, and that in Taiwan was estimated to be 4.92% in 2004 [2]. The development of gout is predominantly caused by chronic elevation of serum uric acid (SUA) levels above the saturation point for monosodium urate crystal formation [2]. Uric acid is an end product of purine metabolism in humans. The SUA concentration was strongly correlated with risk factors for cardiovascular disease, such as body weight, serum glucose, and lipid levels [3]. In addition, the association between SUA concentration and stroke risk has been well documented. A recent meta-analysis of 13 cohort studies indicated that an elevated SUA level was significantly associated with a modestly increased risk of stroke [4].
Stroke is a major cause of death and disability worldwide [5]. A temporal relationship between influenza infection and stroke has been observed [6], with acute infection being associated with a transient increase in the risk of vascular events, including stroke [7]. Influenza vaccination was discovered to not only reduce the risk of influenza infection, but also to lower the risk of stroke [8]. The possible mechanism could be explained by the protective effect of influenza vaccination from influenza infection, which can induce a systemic inflammatory response and endothelial injury, and enhance thrombotic tendencies [9]. In a previous study, the neuroprotective effect from influenza vaccination among patients with high risk factors, such as diabetes and older age, has been demonstrated [8]. However, the possible positive effect of influenza vaccination on the risk of stroke in patients with gout remains unclear. The aim of the present study was to investigate the association between the risk of stroke and influenza vaccination in patients with gout from a nationwide population.

Materials and Methods
We used Taiwan's National Health Insurance Research Database (NHIRD) to analyze the effect of influenza vaccination on stroke risk in patients with gout. The National Health Insurance (NHI) has been established since 1995, and the NHIRD has been described in detail in previous studies [10,11]. The NHIRD comprises the detailed health care data of >23 million enrollees, representing >99% of Taiwan's entire population. The National Health Research Institutes further encrypt these data before researchers can access the database, although theoretically, NHIRD data alone are insufficient to identify any individuals. All researchers using the NHIRD and its data subsets must sign a written agreement declaring that they have no intention of obtaining information that could potentially violate the privacy of patients or care providers. Consequently, our data were delinked and deidentified. The present study was approved by the Joint Institutional Review Board of Taipei Medical University (approval no. N201804043).
We screened all patients who visited health care facilities in Taiwan with a diagnosis of gout (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code V77.5) between 1 January 2001 and 31 December 2012 (n = 111,066). Patients with no subsequent outpatient visit, emergency department visit, or inpatient hospitalization for gout within 1 year of the first presentation were excluded (n = 38,530) because they were considered to not have gout. An additional 48,587 patients were excluded because of the following: patients aged <55 years (n = 43,227), patients with any inpatient or outpatient diagnosis related to stroke before the cohort entry date (n = 3569), and patients who had been vaccinated within the 6 months before the cohort entry date (n = 1791). Finally, 23,949 patients were included in the study (Figure 1).
In Taiwan, influenza vaccination has been free for high-risk individuals (patients aged ≥50 years with type 2 diabetes, chronic liver infection, liver cirrhosis, cardiovascular disease, or chronic pulmonary disease) since 1998, and for all adults aged >65 years old since 2001. Vaccination status was identified by ICD-9-CM code V048 and/or the use of vaccine (confirmed by drug codes). Our primary endpoint was incidence of stroke (ICD-9-CM code V17.1) in patients with gout in all seasons during the follow-up years. The patients were followed until stroke diagnosis, withdrawal from the NHI, loss to follow-up, death, or as of 31 December 2012.  In Taiwan, influenza vaccination has been free for high-risk individuals (patients aged ≥50 years with type 2 diabetes, chronic liver infection, liver cirrhosis, cardiovascular disease, or chronic pulmonary disease) since 1998, and for all adults aged >65 years old since 2001. Vaccination status was identified by ICD-9-CM code V048 and/or the use of vaccine (confirmed by drug codes). Our primary endpoint was incidence of stroke (ICD-9-CM code V17.1) in patients with gout in all seasons during the follow-up years. The The potential confounders of this cohort study included sociodemographic characteristics (age, sex, urbanization level, and monthly income), comorbidities (Charlson comorbidity index, DM, hypertension, dyslipidemia, and atrial fibrillation (AF)), and medication use (allopurinol, benzbromarone, colchicine, aspirin, statin, renin-angiotensin antagonist (RAA), and metformin).
We used the propensity score method to reduce selection bias in the comparison between vaccinated and unvaccinated patients by accounting for covariates using a logistic regression model. The chi-square test was used for categorical variables, and the t-test was used for continuous variables. We also analyzed the association between influenza vaccination and stroke in patients with gout by using Cox proportional hazards regression. The influence of the dose effect of influenza vaccination on the incidence of stroke was also determined. Patients were categorized into four groups in accordance with their vaccination status (0, 1, 2, 3, or ≥4 vaccinations). These data were stratified in accordance with patients' age, sex, comorbidity, and chronic medication use. A sensitivity analysis was performed to evaluate the differences and similarities between groups regarding influenza vaccination and risk of stroke. All statistical analyses were performed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA) and SAS 9.4 software (SAS Institute Inc., Cary, NC, USA). Statistical significance was set at p < 0.05.

Baseline Characteristics
Of the 23,949 eligible individuals, 11,649 (48.64%) received influenza vaccination, and the remaining 12,300 (51.35%) were unvaccinated. A significant difference was discovered between the two groups regarding their age distribution, level of urbanization, and monthly income. The prevalence of pre-existing medical comorbidities, including hypertension and AF, was higher in the vaccinated group; in contrast, the prevalence of dyslipidemia was higher in the unvaccinated group. Significant differences were also discovered with regard to comorbidity-associated medication use. Higher proportions of vaccinated patients than unvaccinated patients used allopurinol, benzbromarone, colchicine, aspirin, statin, RAA, and metformin for >28 days (Table 1).

Risk of Stroke in Vaccinated and Unvaccinated Patients in Accordance with Age and Sex
After adjustment for potential confounders, the incidence rate of all stroke was significantly lower in the vaccinated group than in the unvaccinated group (adjusted hazard ratio [aHR], 0.59; 95% CI, 0.55-0.63; p < 0.001; Table 2). Lower risks of hemorrhagic stroke (aHR, 0.60; 95% CI, 0.49-0.73; p < 0.001) and ischemic stroke (aHR, 0.60; 95% CI, 0.55-0.65; p < 0.001) were also discovered in the vaccinated group. Among the patients aged 55-64 years and 65-74 years, the risks of all stroke, hemorrhagic stroke, and ischemic stroke were significantly lower in the vaccinated group. The risk of hemorrhagic stroke was not significantly different between vaccinated and unvaccinated patients aged ≥75 years (aHR, 0.67; 95% CI, 0.44-1.02). In male patients, the risks of all stroke, hemorrhagic stroke, and ischemic stroke were significantly lower if they had been vaccinated. In female patients, the risks of all stroke and ischemic stroke were lower if they had been vaccinated, but the risk of hemorrhagic stroke was not significantly different between the vaccinated and unvaccinated patients (aHR, 0.76; 95% CI, 0.54-1.08).

Association between Number of Influenza Vaccinations and Reduction of Ischemic Stroke Risk
In the main model, the reduction of ischemic stroke risk was significant after the first vaccination, and a trend of risk reduction was observed as the number of vaccinations increased (aHR, 0.83 and 95% CI, 0.74-0.94; aHR, 0.73 and 95% CI, 0.65-0.81; and aHR, 0.42 and 95% CI, 0.38-0.47, all p < 0.001, for patients vaccinated 1, 2, 3, and ≥4 times, respectively). Similar findings were obtained for the main model with additional covariates and in most of the subgroup analyses. Patients older than 65 years and female patients benefited from vaccinations when two or three doses had been administered (aHR, 0.76 and 95% CI, 0.67-0.87; and aHR, 0.67 and 95% CI, 0.56-0.79 for patients aged >65 years and female patients, respectively). Patients with a history of DM had lower risk of developing ischemic stroke after two or three vaccination doses had been administered (aHR, 0.80; 95% CI, 0.66-0.96). Patients on medication for <28 days had a lower risk of developing ischemic stroke after two or three vaccination doses had been administered (aHR, 0.75 and 95% CI, 0.65-0.85; aHR, 0.78 and 95% CI, 0.67-0.90; aHR, 0.76 and 95% CI, 0.67-0.88; aHR, 0.68 and 95% CI, 0.54-0.86; aHR, 0.77 and 95% CI, 0.67-0.90; and aHR, 0.79 and 95% CI, 0.64-0.99 for allopurinol, benzbromarone, colchicine, aspirin, statin, and RAA, respectively). In contrast, patients on metformin for >28 days had a lower risk of developing ischemic stroke after two or three vaccination doses had been administered (aHR, 0.78; 95% CI, 0.65-0.95; Table 4).

Main Findings
This nationwide population-based cohort study obtained several major findings. First, overall, people with gout had a lower risk of all stroke after receiving an influenza vaccination, but the reduction of hemorrhagic stroke risk was not highly significant in people with gout who were either female or aged >75 years. Second, the association between influenza vaccination and lower risk of stroke in patients with gout tended to be dose-dependent. The effect of reducing hemorrhagic stroke became apparent after four doses of the vaccine had been administered. In contrast, the effect of reducing ischemic stroke became apparent after one dose of the vaccine had been administered, and the trend in risk reduction became significant as the number of vaccinations increased. Third, regardless of comorbidities, patients with gout had a lower risk of hemorrhagic stroke after receiving more than four influenza vaccinations, except for those with gout and AF, in whom this effect was not discovered even after the patient had received more than four influenza vaccinations. In addition, regardless of comorbidities, patients with gout had lower risk of ischemic stroke after receiving one influenza vaccination, except if they had diabetes; these patients had a lower risk of ischemic stroke only after receiving two or three influenza vaccinations. Finally, among the patients with gout who had used metformin for <28 days, having received ≥1 vaccination was associated with a lower risk of hemorrhagic stroke. Those who had used metformin for >28 days required more than two or three vaccinations to obtain a protective effect against ischemic stroke. To the best of our knowledge, this is the first study to investigate the association between influenza vaccination and stroke risk in patients with gout.

Association between Stroke Development and Gout
Studies have reported an association between gout and stroke. In 2013, a large population-based cohort study conducted in England concluded that gout was associated with increased risk of all stroke [12]. Reaching a similar conclusion, a retrospective cohort study conducted in the United States in 2017 demonstrated that gout was an incident stroke risk factor equivalent to DM [13]. Compared with patients with DM only, patients with both gout and DM had higher hazard ratios for incident stroke [13]. A populationbased study conducted in the United Kingdom also indicated that gout was independently associated with higher risk of AF-a risk factor for stroke-at diagnosis and that the risk of AF was higher after the diagnosis [14]. Another reason gout is associated with increased stroke risk is the pathophysiological effects of a high level of SUA, which may induce an increase in blood pressure [15,16], endothelial inflammation and dysfunction [17,18], impaired nitric oxide production [19,20], and atherosclerosis [21][22][23].

Association between Stroke Risk and Influenza Vaccination
Early research revealed that influenza vaccination may be associated with reduced stroke risk [24]. In 2017, a meta-analysis also indicated that vaccination against influenza was associated with a lower risk of stroke [8]. The present study focused on the effect of influenza vaccination in reducing the stroke risk in people with gout. Our findings revealed that people with gout who were vaccinated had a lower risk of all stroke than those who were unvaccinated. Recent population-based cohort studies using the NHIRD have reached similar conclusions; they have reported that influenza vaccination may reduce the risks of hemorrhagic stroke and ischemic stroke in people with AF [10,11], although the reasons for this finding remain unclear. The possible pathophysiological mechanisms may be related to protection from influenza infection. Influenza infection may cause inflammation, smooth muscle cell proliferation, fibrin deposition in atherosclerotic plaque, and loss of antiinflammatory properties in high-density lipoproteins [9,25]. These changes may contribute considerably to the development of stroke. A nationwide case-control study conducted in Taiwan also demonstrated that influenza infection was significantly associated with the development of AF, where the AF risk was found to be increased by 18% [26]. Increased risk of stroke for the aforementioned reasons may be reduced by influenza vaccination. Future research should focus on the mechanisms through which influenza vaccination protects against stroke development in people with gout.

Different Postvaccination Stroke Risk Depending on Age and Sex
The most prevalent subtype of hemorrhagic stroke is intracranial hemorrhage (ICH). A systemic review concluded that age was a risk factor for ICH; elderly people had a higher risk of ICH, and this risk approximately doubled with each decade of life [27]. However, the influence of sex on hemorrhagic stroke remains controversial. A study suggested that the risk of ICH before 65 years was generally lower in women than in men, but that this disparity may disappear after the age of 64 years [28]. Conversely, another study reported no significant difference in incident ICH events between the sexes [29]. In our study, people with gout aged >75 years who had been vaccinated did not exhibit a reduced risk of hemorrhagic stroke. Women were not observed to have a lower risk of hemorrhagic stroke after receiving vaccination (Table 3). Through sensitivity analysis, we discovered a lower risk of hemorrhagic stroke in women with gout after they had received ≥4 influenza vaccinations. Regarding ischemic stroke, older age and male sex are risk factors [28]. Both male patients and patients aged >75 years had a lower risk of ischemic stroke after receiving influenza vaccination. We therefore recommend that people with gout who are male or older than 75 years receive influenza vaccination to help prevent ischemic stroke.

Effect of Medications on Stroke Risk
A retrospective cohort study conducted using the NHIRD reported lower risks of hospitalized stroke and all-cause mortality in urate-lowering medication users than in nonusers among patients with gout [30]. In our study, people with gout who used a uratelowering agent (allopurinol or benzbromarone) for >28 days had significantly lower risks of hemorrhagic and ischemic strokes, indicating that urate-lowering agents and influenza vaccination have an addictive effect on protection from stroke.
Another important medication for stroke risk is aspirin, which has been well reported for stroke primary prevention before [31]. To clarify the possible effect of aspirin on ischemic stroke prevention in the present study, we conduct a subgroup analysis for the sensitivity analysis. The results revealed a potential neuroprotective effect among patients with aspirin usage more or less than 28 days after receiving more than two vaccinations. (Table 4) Therefore, although aspirin may be important for ischemic stroke prevention in the present selected cohort, the potential effect of stoke prevention from influenza vaccination was still observed.

Limitations
The present study had several limitations. First, the NHIRD does not contain data regarding the SUA level, blood sugar, lipid profile, body mass index, or smoking status. Second, we enrolled patients aged >55 years because of the free influenza vaccination policy. Future study is required to investigate younger patients with gout. Third, healthy user bias may have been present [32], although we did adjust for sociodemographic characteristics such as urbanization and monthly income as potential confounders. Fourth, the data were collected until the end of 2012; future studies enrolling more recent data are required to validate our results. Finally, although we controlled for and minimized bias by using propensity score analysis and sensitivity analysis, bias from the uneven prevalence of comorbidities between vaccinated and unvaccinated groups, residual unmeasured confounders, and healthy user bias may still have been present.

Conclusions
Influenza vaccination was associated with a lower risk of all stroke in patients with gout, and this association appeared to be dose-dependent. Further research is required to determine the potential mechanisms through which influenza vaccination protects against stroke in this patient population.