A Population-Based Cohort Study on Peripheral Arterial Disease in Patients with Schizophrenia

Purpose Peripheral arterial disease (PAD) is considered the leading cause of atherosclerotic cardiovascular morbidity. Several risk factors of PAD have been observed in patients with schizophrenia. Therefore, we hypothesize that the incidence of PAD is higher in the schizophrenia population than in the general population. Methods The patients in this population-based cohort study were selected from the Taiwanese National Health Insurance Research Database on the basis of the claims data from 2000 to 2011. We compared the incidence of PAD between schizophrenia and nonschizophrenia cohorts. Cox proportional hazard regression models were employed for analyzing the risk of PAD after adjustment for sex, age, and comorbidities. Results The adjusted hazard ratio (HR) for PAD in the schizophrenia cohort was 1.26-fold higher than that in the nonschizophrenia cohort. Furthermore, patients with schizophrenia using atypical antipsychotics exhibited a high adjusted HR for PAD. Conclusion Compared with the general population, the risk of PAD is higher among patients with schizophrenia. Early diagnosis and intervention can mitigate complications resulting from cardiovascular diseases and lower mortality.


Introduction
Peripheral arterial disease (PAD) is the narrowing of arteries, mostly in the lower limbs, leading to claudication. The prevalence of PAD in both men and women has increased from 3% among those aged 25-29 years to 24% among those 95-99 years in high-income countries [1]. PAD is considered the leading cause of atherosclerotic cardiovascular morbidity, followed by coronary artery disease and stroke [2,3]. The distal superficial femoral artery is the most common site of PAD occurrence, leading to claudication in the calf muscle region. However, patients do not always exhibit typical PAD symptoms. The incidence of asymptomatic PAD is higher than that of symptomatic PAD in the general population [4]. In a previous study, asymptomatic PAD was a crucial predictor of cardiovascular morbidity and mortality [5]. However, PAD is typically underrecognized and undertreated in clinical practice [6,7]. The ankle-brachial index (ABI) is a simple noninvasive test that can identify adults at a high risk of PAD. Prevention and early management of PAD may lower the risk of myocardial infarction, stroke, or death. Several risk factors, such as old age, diabetes, smoking, dyslipidemia, obesity, and hypertension, are associated with PAD [6,8]. Some modifiable risk factors may be helpful in the management of PAD [8], and some of these risk factors are easily identified in patients with schizophrenia.
Schizophrenia is a chronic, severe brain disorder causing disability. The global lifetime prevalence of schizophrenia is approximately 1%. Several studies have reported an increase in the risk of myocardial infarction, stroke, or death in the schizophrenia population [9][10][11]. Schizophrenia is associated with excess deaths from coronary heart disease and stroke in persons younger than 75 years. Cardiovascular disease occurs more frequently in people with schizophrenia and is the commonest cause of death [12]. Several PAD associated risk factors, such as smoking, low physical activity, diabetes, and hyperlipidemia, were higher prevalent in schizophrenia population. Several articles reported schizophrenia population has higher tobacco smoking rate than general population [13][14][15]. Smoking is the key risk factor related with PAD in our schizophrenia patients. Patients with schizophrenia have fewer physical activity [16]. Low physical activity might related with negative symptoms of schizophrenia. Patients with schizophrenia have higher prevalence of diabetes, and schizophrenia patients are at least double the risk of developing type 2 diabetes mellitus [15,17]. And, history of diabetes can deeply influence PAD evolution and cardiovascular diseases [18]. Patients with schizophrenia also had a much higher prevalence of hyperlipidemia in young adulthood than that in the general population [19], and increased risk of initiation of anti-hyperlipidemia medications was noted among patients with schizophrenia [20]. Owing to sharing these risk factors, higher PAD in schizophrenia population was supposed. However, noepidemiological study has studied the relationship between PAD and schizophrenia.
The cause for higher PAD incidence in the schizophrenia population than in the general population remains unknown. It is crucial to understand this cause for the prevention and treatment of related diseases in the schizophrenia population. Early awareness of cardiovascular risk in this population can facilitate early intervention.
We hypothesize that the incidence of PAD is higher in the schizophrenia population than in the general population. Accordingly, we conducted this population-based cohort study in Taiwan to investigate the prevalence of PAD are between the schizophrenia and general populations.

Data Source
The Taiwan National Health Insurance (NHI) program was implemented in 1995 and offers comprehensive medical coverage to all residents of Taiwan (http://www.nhi.gov.tw/english/index. ) before the index date, and those with incomplete age or sex information. Patients in the nonschizophrenia cohort were randomly selected from the insured patients without any history of schizophrenia and PAD at baseline. The exclusion criteria for the patients in the nonschizophrenia cohort were identical to those for patients in the schizophrenia cohort. The nonschizophrenia cohort was frequency matched with the schizophrenia cohort on the basis of age (in 5-y bands), sex, the year of index date and comorbidities of diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma, and medication of statin and aspirin.

Ethics Statement
The NHIRD encrypts patient personal information to protect privacy and provides researchers with anonymous identification numbers associated with relevant claims information, including sex, date of birth, medical services received, and prescriptions. We excluded all individually identifying or patient demographic information. Therefore, the patient consent is not required to access the NHIRD. This study was approved by the Institutional Review Board (IRB) of China Medical University (CMUH104-REC2-115). The IRB specifically waived the consent requirement.

Statistical Analysis
We compared the distributions of age, sex, comorbidities and medications between the schizophrenia and nonschizophrenia cohorts using the chi-square test. Student t test was used to compare the differences in the mean age and mean follow-up years between the 2 cohorts. The incidence density rate of PAD (per 10 000 person-y) was calculated on the basis of sex, age and, comorbidities for each cohort. Univariate and multivariate Cox proportional hazard regression models were employed to examine and compare the risk of PAD associated with schizophrenia between the 2 cohorts, and the risk was evaluated as a hazard ratio (HR) with a 95% confidence interval (CI). Multivariate models were employed after adjustment for age, sex, and comorbidities (diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma) and medication of statin and aspirin. Further analysis was performed to assess the role of antipsychotics in influencing the PAD outcomes. The Kaplan-Meier method was employed for estimating the cumulative incidence of PAD between the schizophrenia and nonschizophrenia cohorts, and the differences were assessed using a log-rank test. Statistical analysis was performed using the SAS 9.3 statistical package (SAS Institute Inc., NC, USA), and P<0.05 in 2-tailed tests was considered significant.

Results
We identified 59 234 patients with newly diagnosed schizophrenia from 2000 to 2011 as the schizophrenia cohort and 59 234patients without any schizophrenia diagnosis as the nonschizophrenia cohort ( Table 1). The mean (± standard deviation [SD]) age of patients in the schizophrenia and nonschizophrenia cohorts was 38.6(±13.2) and 38.5(±13.6) years, respectively, and approximately 45% of these patients were in the age group of 20-34 years. The number of men (52.3%) was higher than that of women (47.3%). The prevalence of preexisting comorbidities (diabetes, COPD, heart failure, stroke, obesity and asthma)and statin use and aspirin use were similar in the schizophrenia cohort than in the nonschizophrenia cohort. The mean follow-up duration was 6.40 (SD = 3.37) and 6.46 (SD = 3.32) years for the schizophrenia and nonschizophrenia cohorts, respectively. Fig 1 shows that the cumulative incidence of PAD was higher in the schizophrenia cohort than in the nonschizophrenia cohort (log-rank test, P = 0.001). The overall incidence of PAD was 1.26-fold higher in the schizophrenia cohort than that in the nonschizophrenia cohort (16.8 and 13.9 per 10 000 person-years, respectively), with a crude HR of 1.21 (95% CI = 1.08-1.36, Table 2). After adjustment for age, sex, comorbidities (diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma) and medication of statin and aspirin, patients in the schizophrenia cohort had a 1.26-fold higher risk of PAD than those in the nonschizophrenia cohort (HR = 1.26, 95% CI = 1.13-1.42). Sex-specific analysis showed that the risks of PAD were higher in women than in men, and these risks were higher in the schizophrenia cohort than in the nonschizophrenia cohort (HR = 1.08, 95% CI = 1.03-1.14 for men; HR = 1.50, 95% CI = 1.26-1.77, P for interaction = 0.01). The incidence densities of PAD increased with age in both cohorts. The age-specific schizophrenia-to-nonschizophrenia relative risk of PAD was higher for younger patients (age = 20-34 y, HR = 1.72, 95% CI = 1.25-2.37) than for older patients (age >50 y, HR = 1.06, 95% CI = 0.89-1.26, P for interaction = 0.02). The comorbidity-specific schizophrenia-to-nonschizophrenia risk of PAD was significantly higher for patients without comorbidities   Table 3 shows the association of the risk of PAD with treatments between the schizophrenia and nonschizophrenia cohorts. The patients in the schizophrenia cohort were at a significantly higher risk of PAD than those in the nonschizophrenia cohort. Among the patients in the schizophrenia cohort, those who received combined typical and atypical antipsychotic treatment exhibited a significantly higher risk of PAD (HR = 1.86, 95% CI = 1.34-2.59) than those in the comparison cohort, followed by those who received only atypical antipsychotic treatment (HR = 1.28, 95% CI = 1.14-1.44) than those in the nonschizophrenia cohort. Compared with the nonschizophrenia cohort, schizophrenia who received atypical antipsychotics treatment for less than 1 year were associated with higher risk of PAD (HR = 1.55, 95% CI = 1.30-1.86), followed by those who received atypical antipsychotic treatment more than 1 year (HR = 1.19, 95% CI = 1.05-1.36).

Discussion
Our results show that the patients in the schizophrenia cohort have a 1.26-fold higher risk of PAD than those in the nonschizophrenia cohort after controlling for age, gender, and comorbidities of diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma, and medication of statin and aspirin. Peripheral vasculature has not been explored exclusively in patients with schizophrenia. This is the first study to report that patients with schizophrenia have a higher risk of PAD than their comparable controls. Patients with PAD are at a high risk of adverse cardiovascular events [21]. PAD, a manifestation of systemic atherosclerosis, leads to segmental narrowing and occlusion of arteries. Our results are in agreement with those of Israel et al, in which peripheral endothelial dysfunction was reported in unmedicated patients with acute schizophrenia [22]. Besides, schizophrenia patients have higher prevalence of tobacco smoking, low physical activity, diabetes, and hyperlipidemia. These would worst their vascular health. Our results showed that the incidence rate of PAD was higher in women than in men in the nonschizophrenia cohort, which is similar to those of previous reports [23,24]. In addition, we found that the risk of PAD increased in both men and women in the schizophrenia cohort. Men in the schizophrenia cohort had a 1.50-fold higher adjusted HR for PAD than those in the nonschizophrenia cohort. Women in the schizophrenia cohort had a 1.08-fold higher adjusted HR for PAD than those in the nonschizophrenia cohort. In a previous study, the rate of smoking was higher in men than in women in patients with schizophrenia [25]. Early onset of : adjusted hazard ratio controlling for age, sex, and comorbidities of diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma, and medication of statin and aspirin; *p<0.05, schizophrenia in males might result in early process of peripheral endothelial dysfunction. A high rate of smoking and early onset age of schizophrenia in men in the schizophrenia cohort may explain our results and the rate of smoking was accounted for by adjusting for smokingassociated diseases (such as asthma).
Age plays a crucial role in PAD. The prevalence of PAD increased in older patients in a previous study [1]. In our study, we observed that the incidence of PAD was higher in older patients in both the cohorts. However, patients aged between 20 and 34 years in the schizophrenia cohort have a 1.72-fold higher adjusted HR for PAD than those in the nonschizophrenia cohort. Age between 35 and 49 schizophrenia cohort have a 1.31-fold higher adjusted HR for PAD development than the age between 35 and 49 years in the nonschizophrenia cohort. No significant difference in both the age more than 50 year old cohorts. These results indicate that younger patients with schizophrenia have a high PAD risk. The results of our study showed that schizophrenia may facilitate the progress of atherosclerosis and is consistent with the report of Israel et al [22], in which peripheral endothelial dysfunction was reported in unmedicated patients with schizophrenia. The anti-inflammatory mediator nitric oxide exerts a protective effect on endothelial functioning. Burghardt et al suggested that patients with schizophrenia may lose the genetic protection provided by nitric oxide once their condition progresses to the pro-inflammatory state of metabolic syndrome [26].
In our study, we investigated the effect of different antipsychotics. We found that patients in the schizophrenia cohort who received atypical antipsychotic treatment had a higher HR for PAD (1.28) than those in the nonschizophrenia cohort after adjustment for age, sex, and comorbidities of diabetes, hypertension, hyperlipidemia, COPD, heart failure, CAD, stroke, obesity and asthma, and medication of statin and aspirin. No significant difference was observed for typical antipsychotic treatment after adjustment for confounding factors between the schizophrenia and nonschizophrenia cohorts for PAD. Compared with patients in the nonschizophrenia cohort, patients in the schizophrenia cohort who received combined typical and atypical antipsychotic treatment had a higher HR (1.86) for PAD after adjustment. Antipsychotics, particularly atypical antipsychotics, may play a role in the development of PAD. The Clinical Antipsychotic Trials of Intervention Effectiveness study revealed an association between the use of atypical antipsychotics in schizophrenia and metabolic syndrome [27]. The American Diabetes Association, the American Psychiatric Association, the American Association of Clinical Endocrinologists, and the North American Association for the Study of Obesity suggest that baseline screening and follow-up monitoring is essential for mitigating the likelihood of developing CVD, diabetes, or other diabetes-related complications while prescribing atypical antipsychotics [28]. Atypical antipsychotics may counteract some vascular health benefits of a diet high in omega 3 fatty acids, as reported in the study by Ellingrod [29]. Our study results were in agreement with those of the aforementioned reports.
However, some limitations were encountered while conducting this study. First, the accuracy of the incidence of PAD was not validated by reviewing the medical charts, and PAD cases were identified only on the basis of the ICD-9-CM codes(440.0, 440.2, 440.3, 440.8, 440.9, 443, 444.0, 444.22, 444.8, 447.8, and 447.9). Second, the use of ICD system should be considered as a limitation of this study. The use of ICD could be considered as a further limitation of this study. 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. If these doctors or hospitals make wrong coding for the diagnoses, they will be punished with a lot of penalties. Therefore, the diagnoses based on ICD-9 codes in this study should be highly reliable. Third, the incidence of asymptomatic PAD is higher than that of symptomatic PAD in the general population. In our study, asymptomatic PAD was not considered as part of the study design. Fourth, lack of data (e.g, smoking status, obesity, and family history) may have influenced the results of our study. Therefore, we applied proxy measures, such as hyperlipidemia and diabetes as indicators of obesity and COPD as an indicator of smoking, for controlling the potential confounding effects. However, the effects of some unmeasured confounders could not be addressed in our study. Fifth, the adherence of treatment (statin and aspirin) in schizophrenia cohort might be poorer than in nonschizophrenia cohort. Sixth, the experimental design is another limitation. Our study is a retrospective cohort study. The evidence derived from a retrospective cohort study is generally of lower methodological quality than that from randomized trials by the experimental design because a retrospective cohort study is subject to many biases related to the necessary adjustments for confounding factors. Despite the meticulous design of this study and its adequate control of confounding factors, biases could remain because of possibly unmeasured or unknown confounding factors.
In conclusion, this nationwide population-based cohort study shows that patients in the schizophrenia cohort have a 1.26-fold higher adjusted HR for PAD than that in the nonschizophrenia cohort after adjustment. This is the first study to report the association between schizophrenia and the risk of PAD. Patients with schizophrenia who were treated using atypical antipsychotics have a high risk of PAD. However, most patients with the new onset schizophrenia patients received atypical antipsychotics. It is crucial to adequately assess the risk of PAD among patients with schizophrenia, particularly those who are unable to exercise adequate personal care or gain access to health care by themselves because of their illness. In-depth ABI analyses may facilitate PAD diagnosis. Early diagnosis and management of PAD may further facilitate the prevention of myocardial infarction, stroke, and other cardiovascular diseases. Additional studies are warranted to examine the role of PAD in schizophrenia.