The Association of Selective Serotonin Reuptake Inhibitors Use and Stroke in Geriatric Population

BRIEF REPORT The Association of Selective Serotonin Reuptake Inhibitors Use and Stroke in Geriatric Population Chia-Chun Hung, M.D., Ching-Heng Lin, Ph.D., Tsuo-Hung Lan, Ph.D., Chin-Hong Chan, M.D.

Objectives: Selective serotonin reuptake inhibitor (SSRI) exposure has controversial results in increasing the stroke risk. With the risk of stroke increased with age, the safety of SSRI use among older adults attracts much concern. Methods: We analyzed 28,145 subjects older than 65 years from a subset of a 9-year cohort database from the National Health Insurance Research Database, Taiwan. Results: The survival analysis showed a greater probability of stroke in subjects with SSRI exposure after adjusting other covariates. Compared with other variables, SSRI exposure had the strongest effect (hazard ratio: 2.66, 95% confidence interval: 2.21e3.20). The risk was independent to depression-related stroke risk. Conclusions: The use of SSRIs independently increases the risk of stroke among older patients. SSRIs are still practically safe to most users, providing precautionary measures are taken. (Am J Geriatr Psychiatry 2013; 21:811e815) Key Words: Geriatric, hemorrhage, ischemia, selective serotonin reuptake inhibitors (SSRIs), stroke

S

troke is the third leading cause of death in the United States. Nearly three quarters of all strokes occur in people over the age of 65 years.1 The use

of second-generation antidepressants Monoamine oxidase inhibitors [MAOIs], Serotonin-norepinephrine reuptake inhibitors [SNRIs], selective serotonin reuptake inhibitor [SSRI] is increased in recent years, and SSRIs have been introduced to the world as antidepressants, anxiolytic, and prescriptions for broaden indications. For the last few years, the risk of stroke in patients using SSRIs has become an issue of attention. There are more than 100 cases of cerebrovascular adverse reactions to SSRIs in the World Health Organization database.2 Several studies exploring the relationship between SSRIs and stroke found no significant increased risk3 or very low risk.4 However, the possibility of an increased stroke risk with SSRI use cannot be totally ruled out in high-risk patients who are vulnerable for hemorrhagic stroke, bleeding diathesis, or vasospam.5 Because older adults have higher risk of cerebrovascular disorders, we should pay more attention to the SSRI use in this group. For exploring the stroke risk with SSRI use in the older adults group, we conducted a populationbased study by using the National Health Insurance data to analyze the relative risk.

METHODS Study Population and Design The National Health Research Institutes, Taiwan, manages the authorization of National Health Insurance Research Database (NHIRD) for research use. We obtained a subset of the NHIRD with 1,000,000 random subjects, accounted for about 5% of all enrollees in the National Health Insurance (NIH) program. There were no statistically significant differences in age, gender, or costs between the sample group and all enrollees (data not shown). The database contains information of medical claims for ambulatory care, inpatient care, dental services, and prescription drugs as well as registration files of the enrollees, insured from January 1996 to

Received November 3, 2011; revised February 17, 2012; accepted February 27, 2012. From the Department of Psychiatry (C-CH, T-HL, C-HC) and Department of Research (C-HL), Taichung Veterans General Hospital, Taichung, Taiwan, and the Department of Psychiatry, National Yang-Ming University (T-HL), Taipei, Taiwan. Send correspondence and reprint requests to Chin-Hong Chan, M.D., Department of Psychiatry, Taichung Veterans General Hospital, No. 160, Sec. 3, Chung-Kang Rd, Taichung, Taiwan 40705. e-mail: [email protected] Ó 2013 American Association for Geriatric Psychiatry http://dx.doi.org/10.1016/j.jagp.2013.01.018

Am J Geriatr Psychiatry 21:8, August 2013

811

The Association of SSRIs Use and Stroke in Geriatric Population December 2008. The International Classification of Diseases, Ninth Revision (ICD-9) diagnosis system was further incorporated into the data since 2000. This is a 9-year retrospective observational study. Subjects were included if they were age 65 years or older on January 1, 2001. Subjects were defined to have SSRI exposed when received SSRI prescription (paroxetine, fluoxetine, sertraline, citalopram, escitalopram, or fluvoxamine) for at least two consecutive months during January 1, 2001, to December 31, 2008. Those who used only 1-month SSRI have been removed from the analysis because of instability of SSRI use; otherwise, they were defined to have no SSRI exposure. For subjects with SSRI exposure, the date of first SSRI prescription was defined as entered date for this study. The primary outcome was the first stroke event within 1 year after SSRI exposure. Those who had started on SSRIs after suffering a stroke were excluded from the beginning of the study. The stroke diagnosis was identified by meeting the occurrence of ICD-9, codes from 430 to 438 in any inpatient treatment or outpatient treatments during the 9-year period. Patients who had accompanied diagnosis of head injury (ICD-9 code: 800-804, 850854) and stroke within 2 weeks of head injury events were excluded. The endpoint of follow-up was occurrence of first stroke event, death, or end of study on December 31, 2009. The date of death was defined as the ending date of coverage. As shown by Lien et al.,5 probably for a practical reason, most NHI coverage periods of stroke patients ended within a month on life termination. This is evident from that comparison of the date of death and the date end of NHI coverage in stroke patients who died within 1 year after discharge. Ninety-seven percent of the records demonstrated the same date of death and date of end of NHI coverage. Thus, the end of NHI coverage is a good proxy for a patient’s survival. Comorbidities for stroke, including myocardial infarction, angina pectoris, atrial fibrillation, peripheral arterial occlusive disease, congestive heart failure, diabetes mellitus, hypertension, hyperlipidemia, and renal diseases in the last 12 months before first SSRI prescription, were included in analysis. Myocardial infarction (ICD-9 code: 410e412), angina pectoris (ICD-9 code: 413), atrial fibrillation (ICD-9 code: 427), peripheral arterial occlusive disease (ICD-9 code: 444), congestive heart failure (ICD-9 code: 428), diabetes mellitus (ICD-9 code: 250),

812

hypertension (ICD-9 code: 401e405), and hyperlipidemia (ICD-9 code: 272) were confirmed by the records of 1 or more prescriptions for the diagnoses, while renal diseases were confirmed when ICD-9 code 390-438 or 585 appeared in the NHI records. For subjects who did not use SSRIs, comorbidities in the last 12 months from January 1, 2001, were retrieved from the database. The subjects were excluded if they had any inpatient diagnosis of stroke or 3 outpatient records of stroke before January 1, 2001. The subjects with any diagnosis of mental disorder (ICD-9 code: 290e319) before January 1, 2001, were also excluded. Finally, a total of 28,145 subjects were used for statistical analysis.

Statistic Analysis Distributions of SSRI-exposed and non-exposed subjects in age, gender, and clinical comorbidities were examined using c2 tests for categorical variables and Student’s t-tests for continuous variables. Multivariable Cox proportional-hazards models were used to explore the relation between SSRI exposure and occurrence of stroke, adjusted for age, gender, and comorbidities. The proportional hazard assumption was tested graphically and by including the interaction of time with each covariate. All statistical tests were two-sided, conducted at significance level of 0.05, and reported using p value and/or 95% confidence intervals (CIs). All analyses were performed using the SAS software, version 9.1 (SAS Institute, Cary, NC).

RESULTS Because the frequencies of age, sex, and clinical variables were significantly different between subjects exposed and not exposed to SSRI, we built up a regression model, adjusting for all variables mentioned to avoid the possible confounding effects. Independent hazard ratios for SSRI exposure, age, sex, and clinical comorbidities of stroke in multivariate Cox proportional-hazards regression model are shown in Table 1. The independent hazard ratio of stroke among older adults after adjusted for contribution of age, sex, and clinical comorbidities reached 2.66 (95% CI: 2.21e3.20) for subjects with SSRI exposure. In all

Am J Geriatr Psychiatry 21:8, August 2013

Hung et al.

TABLE 1. The Cox Proportion Hazard Regression Model of Stroke (Wald c2 Test, N [ 26,549, df [ 1)

Variable SSRI (use/nonuse) Age (1 year) Sex (male/female) Physical condition AF (yes/no) MI (yes/no) CHF (yes/no) AP (yes/no) PAOD (yes/no) Diabetes (yes/no) Hypertension (yes/no) Hyperlipidemia (yes/no) Renal disease CKD (yes/no) ESRD (yes/no) Medication Aspirin (yes/no) Heparin (yes/no) Warfarin (yes/no) NSAID (yes/no) Antipsychotics (yes/no)a

95% CI

Hazard Ratio

Lower

Upper

p

2.66 1.04 1.20

2.21 1.03 1.03

3.20 1.05 1.39

<0.001 <0.001 0.0174

1.05 1.45 1.01 1.13 3.77 1.81 1.57 0.93

0.62 0.98 0.72 0.93 1.41 1.51 1.34 0.75

1.81 2.15 1.41 1.39 10.09 2.18 1.84 1.15

0.8496 0.0633 0.9529 0.2248 0.0082 <0.001 <0.001 0.5029

1.00 0.67

0.51 0.21

1.96 2.10

0.9962 0.4901

1.36 1.29 1.74 0.93 1.06

1.11 0.73 0.91 0.79 0.84

1.66 2.27 3.32 1.08 1.33

0.0034 0.3833 0.0925 0.3259 0.6181

Notes: AF: atrial fibrillation; AP: angina pectoris; CHF: congestive heart failure; CI: confidence interval; CKD: chronic kidney disease; ESRD: end-stage renal disease; MI: myocardial infarction; NSAID: nonsteroidal anti-inflammatory drugs; PAOD: peripheral arterial occlusive disease; SSRI: selective serotonin reuptake inhibitor. a Antipsychotics included typical (haloperidol, trifluoperazine) and atypical agents (amisulpride, aripiprazole, clozapine, Sulpride, olanzapine, quetiapine, risperidone, ziprasidone, zotepine).

variants we analyzed, SSRI exposure contributed the highest risk of stroke by 2.66-fold, more than hypertension, diabetes, and end stage of renal disease contributed. Furthermore, KaplaneMeier survival analysis demonstrated a greater probability of stroke in subjects with SSRI exposed than those without SSRI exposed. The survival rates of the two groups were significantly different, with 97.86% of non-SSRI exposure and 92.83% of SSRI users at time point of 1 year (p value of log-rank test <0.001, df ¼ 1). After a total period of 9 years, the survival rates of the two groups were 80.7% of non-SSRI exposure and 56.6% of SSRI users (p value of log-rank test <0.001, df ¼ 1; data not shown). Further analysis of stroke subtype was done. Among the 182 patients with stroke events, there were 119 patients diagnosed of ischemic stroke and 63 patients diagnosed of hemorrhagic stroke. The risk of ischemic stroke was 2.54-fold (95% CI: 2.03e3.19, Wald test, c2 ¼ 44.81, df ¼ 1, p <0.001) and

Am J Geriatr Psychiatry 21:8, August 2013

hemorrhagic stroke was 3.03-fold (95% CI: 2.19e4.19, Wald test, c2 ¼ 166.33, df ¼ 1, p <0.001). The risk of both ischemic and hemorrhagic strokes decreased gradually with time.

DISCUSSION The safety of antidepressant treatment in older adults is still questioned now. The chief finding of present study is that patients older than 65 years receiving SSRIs treatment have increased risk of both ischemic and hemorrhagic strokes. SSRIs are one of the most popular prescribed medications among drugs-involved serotonin. The major molecular function of SSRIs is to bind on serotonin transporter (SERT), inhibit recycling of serotonin to presynaptic cell membrane, and increase the concentration of synaptic 5-HT. Serotonin may promote vasoconstriction and platelet activation, with following thromboembolic formation. Serotonin might promote arterial contractility and be involved in the pathogenesis of migraine and hypertension6 and possibly induce transient ischemic attack.7 In the older adults group, the aging vessels may get more vulnerable to injury or thromboembolism. It supposed that these patients had higher risk of vasoconstrictive or ischemic stroke under serotonin effects. On the contrary, there was evidence of SSRI use with increased bleeding tendency. SSRIs inhibit the function of SERT. Beside its therapeutic effect within the central nervous system, SSRIs inhibit the platelets to uptake serotonin from plasma. Because platelets do not synthesize serotonin, SSRIs attenuate platelets activation by depleting serotonin storage and may increase the risk of coagulopathy.8 A recent detailed review explored the relationship of SSRI antidepressants and abnormal bleeding postulated that SSRIs associated with an increased risk of bleeding, especially from the upper gastrointestinal tract.4 Considering the mechanism of SSRIs, it is reasonable to pay attention to the risk of hemorrhagic stroke and other bleeding tendency among older adult patients with SSRI treatment. In our analysis, people older than 65 years taking SSRI show elevated risk of both ischemic and hemorrhagic strokes. Serotonin has complicated effects on the vascular system. In line with our observation,

813

The Association of SSRIs Use and Stroke in Geriatric Population a previous brief report postulated that the temporal relation of the use of SSRI on platelets dysfunction might be bidirectional.9 The quick and early effect of SSRI use on platelets might be an increase in tendency for thrombosis, whereas the late effect after repeated dosing might be an increase in tendency to bleed.9 Ni et al.6 suggested that serotonin reuptake transporter inhibition may cause dual effects in thrombosis and bleeding. It may reflect on the clinical finding in this study that SSRI may increase risks of both hemorrhagic and ischemic strokes. On the contrary, serotonin may have different effects on different vascular conditions. Golino et al.10 observed that serotonin has a vasodilating effect on normal human coronary arteries; when the endothelium is damaged, as in coronary artery disease, serotonin has a direct, unopposed vasoconstricting effect. In atherosclerosis, the damaged endothelium is unable to produce sufficient amounts of endothelium-derived relaxing factor to counteract the vasoconstrictive effects of serotonin.10 It suggests that physicians may consider the vascular condition before prescribing SSRIs among older adult patients.

major depressive disorder after adjusted for other possible confounders. Because there are more clinical comorbidities in the older adult group, there is a higher risk of stroke. But SSRI exposure still contributes additional risk to stroke among older people.

LIMITATIONS Although the study uses a representative and large sample from the National Health Insurance Database, there are still some limitations left. First, since our information comes from the database, we could not know exactly about drug adherence of the patients. Second, although we adjusted several factors in the odds ratio analysis, we could not know the body weight, body mass index, cigarette use, or the lifestyle of individuals. For the limitation of observational studies, we could only suggest a possible causality between the SSRIs use and stroke among older adults. Further prospective, welldesigned studies should be conducted to replicate our findings.

MAJOR DEPRESSIVE DISORDERSeRELATED FACTORS Depression itself was considered potential risk factors of cerebrovascular accidents. The Framingham Heart study published in 2007 revealed that the risk of developing stroke/transient ischemic attack was 4.21 times greater in participants younger than 65 years with symptoms of depression. However, the relationship was not significant among participants older than 65 years.11 Because a large proportion of patients with major depressive disorder receiving SSRIs treatment, one may consider the contribution of depressions to the elevated risk of stroke. For clarifying the independent role of SSRI exposure, we conducted further analysis from the same database. People were included with age greater than 65 years but had no diagnosis of major depressive disorder during 2000e2009. Among a total of 26,549 people older than 65 years without major depressive disorder, SSRI exposure explains a 2.57-fold risk of stroke (95% CI: 1.98e3.35). It again confirms that SSRI alone would increase risk of stroke among older adult patients without

814

CONCLUSION The use of SSRIs increases the risk of stroke among patients older than 65 years. We call for an alert to emphasize on the potential risk of the use of SSRIs prescribed from medical professionals, especially psychiatrists and whoever is prescribing SSRIs in practice. We also remind that there is a possible difference across ethnic populations on SSRI to invoke stroke. While prescribing SSRI among elder patients, a balance between benefits and withdrawals should be sincerely considered. To the result of this study, it is worthy of paying attention to the increased risk of stroke with SSRI treatment among older adult patients. The statistical analysis is supported by the Biostatistics Task Force of Taichung Veterans General Hospital. This study is based on the data obtained from NHIRD that were provided by the Bureau of National Health Insurance of the Department of Health and managed by National Health Research Institutes (registration #99278). The

Am J Geriatr Psychiatry 21:8, August 2013

Hung et al. interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health, or National Health Research Institutes.

This study was presented at 15th International Congress of International Psychogeriatric Association, The Hague, the Netherlands, September 6e9, 2011. Disclosures of interest: No disclosures to report.

References 1. American Heart Association: Heart Disease and Stroke Statistics. Dallas, TX, American Heart Association, 2010-update 2. Edwards IR, Olsson S: WHO programme global monitoring, in Pharmacovigilance. Edited by Mann S, Andrews A. Chichester, England, Wiley & Sons, 2002, pp 169e182 3. Bak S, Tsiropoulos I, Kjaersgaard JO, et al: Selective serotonin reuptake inhibitors and the risk of stroke: a population-based caseecontrol study. Stroke 2002; 33(6):1465e1473 4. Ramasubbu R: Cerebrovascular effects of selective serotonin reuptake inhibitors: a systematic review. J Clin Psychiatry 2004; 65(12):1642e1653 5. Lien HM, Chou SY, Liu JT: Hospital ownership and performance: evidence from stroke and cardiac treatment in Taiwan. J Health Econ 2008; 27(5):1208e1223 6. Ni W, Watts SW: 5-Hydroxytryptamine in the cardiovascular system: focus on the serotonin transporter (SERT). Clin Exp Pharmacol Physiol 2006; 33(7):575e583

Am J Geriatr Psychiatry 21:8, August 2013

7. Manos GH, Wechsler SM: Transient ischemic attack reported with paroxetine use. Ann Pharmacother 2004; 38(4): 617e620 8. Huang ZS, Chiang TL, Lee TK: Stroke prevalence in Taiwan. Findings from the 1994 National Health Interview Survey. Stroke 1997; 28(8):1579e1584 9. Kurne A, Ertugrul A, Anil Yagcioglu AE, et al: Venous thromboembolism and escitalopram. Gen Hosp Psychiatry 2004; 26(6): 481e483 10. Golino P, Piscione F, Willerson JT, et al: Divergent effects of serotonin on coronary-artery dimensions and blood flow in patients with coronary atherosclerosis and control patients. N Engl J Med 1991; 324(10):641e 648 11. Salaycik KJ, Kelly-Hayes M, Beiser A, et al: Depressive symptoms and risk of stroke: the Framingham Study. Stroke 2007; 38(1): 16e21

815