ABSTRACT
Background
Psychotic disorders and use of antipsychotics prior to or after ischemic stroke (IS) may be associated with poor outcomes. However, data on antipsychotic use in young IS patients are limited. We aimed to characterize young patients purchasing antipsychotics prior to IS or de novo post‐stroke, and to examine their association with long‐term outcomes.
Methods
We analyzed data from the Helsinki Young Stroke Registry, including 1008 consecutive patients aged 15–49 with first‐ever IS 1994–2007. We considered patients without antipsychotic purchases as non‐users, those with at least one purchase any time before IS as prior users, and those who had purchases at any time after IS (but not before) as de novo users. Cox regression models assessed the association of antipsychotic purchases with any recurrent vascular event or all‐cause mortality.
Results
Of 966 included IS survivors (62.6% male, median age 44), 55 (5.7%) purchased antipsychotics before and 67 (6.9%) after index IS. Compared with de novo or non‐users, prior users more often had other/unknown socioeconomic status, a history of psychiatric hospitalization, drug abuse, smoking, heavy drinking, more severe stroke symptoms on admission, and limb paresis at discharge. Antipsychotics purchased before IS were associated with a heightened hazard of endpoint events when adjusted for sociodemographics and cardiovascular comorbidities. The association was not found for de novo users.
Conclusions
Around 6% of young IS patients had a history of antipsychotic use, while a similar proportion initiated antipsychotics post‐stroke. Pre‐stroke antipsychotic use was associated with recurrent vascular events and mortality.
Keywords: antipsychotics, long‐term, recurrence, stroke, young adult
Among 966 young ischemic stroke patients, approximately 6% had prior antipsychotic use and a similar proportion initiated treatment after stroke. Pre‐stroke antipsychotic users had more adverse socioeconomic and clinical risk profiles, including substance use and more severe stroke presentation. Pre‐stroke antipsychotic use was associated with higher risks of recurrent vascular events and mortality.
1. Introduction
Around 10%–15% of ischemic strokes (IS) affect individuals aged under 50 and their incidence is increasing [1]. In young patients, with families to care for and many years of active lives ahead, the long‐term social, psychological, and vocational consequences of IS can be devastating.
Recent data have reported psychotic disorders and antipsychotic treatment to be associated with increased risk of cerebrovascular and other cardiovascular events as well as all‐cause mortality [2, 3, 4]. Limited data suggest that this could apply also to young patients and that it could even be more significant in younger compared with older patients [4, 5]. Patients with psychosis prior to IS had a poorer short‐ and long‐term outcome as well as used less frequently antihypertensives after IS [6]. Furthermore, stroke recurrence is more common in these patients [6]. Additionally, use of antipsychotics before stroke may be associated with more severe strokes, longer hospitalization, and higher post‐stroke mortality [7]. In addition to psychiatric comorbidities known before stroke, post‐stroke psychiatric symptoms and complications, including psychosis, may have a negative effect on the recovery and quality of life [8, 9, 10]. Previous studies have reported that psychiatric disorders, for example depression, anxiety, insomnia, and other mental disturbances after stroke, affect up to nearly one‐third of all stroke patients and potentially worsen their outcome [10, 11, 12, 13, 14, 15, 16, 17, 18]. However, especially in young patients, data on antipsychotic use before and after IS and their effect on long‐term outcome are limited.
We aimed to characterize young IS patients who purchased antipsychotics prior to and de novo after IS, and to examine their association with long‐term recurrent vascular events and all‐cause mortality.
2. Patients and Methods
2.1. Study Population
The cohort originates from the Helsinki Young Stroke Registry (HYSR), including 1008 consecutive patients aged 15–49 with their first‐ever IS and treated at the Department of Neurology, Helsinki University Hospital between January 1994 and May 2007, as identified from a prospective computerized hospital discharge database. This study utilized the original World Health Organization (WHO) stroke definition, however, including patients with a short duration of symptoms when imaging‐positive findings of IS were present [19]. Transient ischemic attacks were excluded. The personal identification number assigned to every resident in Finland enabled further combining the data from HYSR with data from several national registries. The present analysis excluded patients with a false primary diagnosis, those who were lost to follow‐up, and those dying within 3 weeks from IS. After applying these exclusion criteria, the study population included 966 patients.
2.2. Baseline Data
All baseline laboratory and other diagnostic tests have been previously fully described [20]. A chest x‐ray, electrocardiogram, and brain imaging with computed tomography or magnetic resonance imaging were performed to all patients. Data collection included patients' sociodemographics: age, sex, and socioeconomic status; data on ischemic stroke risk factors, including the status of cigarette smoking at the time of index event, heavy alcohol use (defined as consumption over 200 g a week), obesity (body mass index ≥ 30 or patient clearly stated as heavily obese), cardiovascular disease, atrial fibrillation, hypertension, dyslipidemia, diabetes mellitus type 1 and 2; data on stroke‐related variables assessed at hospital admission, including NIH Stroke Scale (NIHSS), Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification, infarct size and laterality, silent infarcts, and leukoaraiosis; and data on limb paresis at hospital discharge. Patient's occupation categorized the socioeconomic status as upper white‐collar worker, lower white‐collar worker, blue‐collar worker, other (entrepreneur, student, pensioner, and unemployed) or unknown/missing (https://stat.fi/fi/luokitukset/sosioekon_asema/sosioekon_asema_1_19890101). NIHSS score at admission categorized the stroke severity as mild (NIHSS 0–6), moderate (NIHSS 7–14) and severe (NIHSS 15 or more). Categories for limb paresis at hospital discharge were no paresis, mild (1 point from NIHSS sections 5–6), and moderate to severe (2–4 points from NIHSS sections 5–6).
We obtained data on hospitalizations from psychiatric reasons prior to IS from the Care Register for Health Care, maintained by the National Institute for Health and Welfare, Finland, from 1969 to 2011. The diagnoses for psychiatric hospitalizations included International Classification of Diseases eighth revision (ICD‐8) categories 291–299, 300–301; ninth revision (ICD‐9) categories 295–298, 300–301, 309; and tenth revision (ICD‐10) categories F04–F09, F20–F48, F60–69 and excluded psychiatric hospitalizations due to substance abuse.
2.3. Antipsychotics
Data on filled prescriptions (i.e., purchases) between 1993 and 2011, including date of purchase and Anatomical Therapeutic Chemical (ATC) codes (WHO Collaborating Centre for Drug Statistics Methodology, 2018), originated from the Drug Prescription Register kept by the Social Insurance Institution of Finland. Information on all prescriptions assigned by any physician in Finland are included in this register. Prescribed medications in Finland are entitled to reimbursement of 40–100% of the medicine's price, and the Social Insurance Institution of Finland reimburses a 3 months' equivalent amount of medication at any one time. ATC codes N05A (antipsychotics), excluding lithium (N05AN01), identified those who purchased antipsychotics before and after the index IS. This study defined patients without any antipsychotic purchases prior to IS from 1993 onwards or after IS until the end of follow‐up as non‐users, those with at least one purchase of antipsychotics any time before index IS as prior users and de novo users as those who purchased antipsychotics any time after IS but with no former purchases at any point before IS.
2.4. Follow‐Up Data
Data on the outcomes occurring 1994–2011 came from the Care Register for Health Care. We also verified from patient records all possible cases of composite vascular events identified from this register (n = 245, 87.5%) [21]. Statistics Finland (https://stat.fi/) was the source of dates and causes of death. The follow‐up started at the index IS and ended at the occurrence of the event of interest (including death of any cause) or on 31st December 2011, whichever came first. The endpoint of interest was a composite of any nonfatal or fatal vascular event or all‐cause mortality, including recurrent (ischemic and hemorrhagic) strokes and other vascular events, which comprised cardiac events (acute coronary syndromes, cardiac death, and other cardiac events), peripheral arterial events, and venous events including pulmonary embolism and deep vein thrombosis. Table S1 presents the ICD codes and the respective clinical definitions of the events. We excluded from the relevant analysis those patients who initiated antipsychotics after the event of interest.
2.5. Statistical Analyses
To compare antipsychotic users at baseline, Pearson's chi‐square and Fisher's exact tests were used for categorical variables and Kruskal–Wallis test for non‐normally distributed continuous variable (age) across the groups with patients without antipsychotic purchases, with pre‐stroke purchases, and with de‐novo post‐stroke purchases.
We fitted univariable and multivariable Cox regression models to examine the risk of our endpoint event across the groups. We constructed multivariable Cox regression models by progressively adjusting for baseline variables with p‐values < 0.05 from univariable models and for potential confounders identified from previous literature [1, 20, 21, 22], adding one variable at a time and retaining any added variables that remained statistically significant. We presented three models: Model 1 adjusted for age, sex, and socioeconomic status, which we considered as essential to control for in any case; Model 2 further adjusted for cardiovascular risk factors (smoking, obesity, and cardiovascular comorbidity burden, including atrial fibrillation, cardiovascular disease, diabetes mellitus 1 or 2, dyslipidemia, and hypertension); and the final Model 3 further adjusted for stroke etiology (TOAST). Prior psychiatric hospitalization (for any reason) and heavy alcohol use were dropped during the model selection process. The proportional hazards assumption was violated in the models, and it was not possible to overcome this through allowing different time periods to have different hazard ratios (HRs), limiting follow‐up to 10 years, or further fitting the interaction between age or antipsychotic use and different follow‐up periods. Hence, to support clinically robust interpretation of the modeled HRs we present the survival curves based on the actual data alongside those predicted by the final multivariable model. No considerable multicollinearity (variance inflation factors > 5) existed between covariates assessed in the models. As a sensitivity analysis, the antipsychotic purchases were restricted to one‐year period before and after the index IS to ensure that all patients had an equal observation window for retrospective and follow‐up medication data, thereby reducing the potential for immortal time bias. In the sensitivity analysis we defined antipsychotic use as follows: Non‐users had no antipsychotic purchases within one year before or after the IS; prior users had at least one purchase within the year before IS; and the novo users had at least one purchase within the year after IS but none in the year preceding IS.
We performed statistical analyses with IBM SPSS Statistics for Macintosh, Version 28.0 (SPSS Inc., IBM, Armonk, NY: IBM) and with R: R Core Team (2020). R Foundation for Statistical Computing, Vienna, Austria.
2.6. Ethics Statement
Ethical approval for this study was obtained from The Ethics Committee of the Hospital District of Helsinki and Uusimaa (73/13/03/00/11). This present study is a continuation of the same research project including many prior studies performed on the Helsinki Young Stroke Registry and thus, the Ethics approval number remains the same as in the previous studies. Informed consent was not sought for the present study since the study is based on registry data without direct patient contact.
3. Results
After applying exclusion criteria (Figure 1), a total of 966 patients (62.6% male; median age 44, interquartile range [IQR] 37–47) were followed up, of which 55 (5.7%) purchased antipsychotics any time prior to IS and 67 (6.9%) initiated antipsychotics de novo at some point after IS. Table 1 presents the baseline characteristics of each antipsychotic user group and the different types of antipsychotics purchased. The median time from the first pre‐stroke antipsychotic purchase to the index IS was 4.4 (IQR 1.6–6.9) years and from the index IS to the first de novo post‐stroke purchase 2.2 (IQR 0.66–6.9) years. A total of 25 (45.5%) prior antipsychotic users and 32 (47.8%) de novo users had their second purchase within 3 months and 35 (63.6%) and 37 (55.2%) within 6 months from the first filled prescription, correspondingly. Furthermore, 27 (49.1%) of the pre‐stroke users purchased antipsychotics within a year prior to the index IS and 14 (20.9%) of the post‐stroke de novo users purchased antipsychotics within a year prior to endpoint event.
FIGURE 1.
Flow chart of the study.
TABLE 1.
Baseline characteristics of those with and without purchase of antipsychotics.
| Characteristic | Antipsychotics (N = 966) | p | ||
|---|---|---|---|---|
| No purchases 844 (87.4) | Any time pre‐stroke 55 (5.7) | De novo post‐stroke 67 (6.9) | ||
| n (%) | n (%) | n (%) | ||
| Sociodemographic variables | ||||
| Age, median (IQR) | 44.0 (37.0–47.0) | 45.0 (41.0–47.0) | 43.0 (36.0–47.0) | 0.256 |
| Male sex | 525 (62.3) | 38 (69.1) | 42 (62.7) | 0.593 |
| Socioeconomic statusa | < 0.001*** | |||
| Blue‐collar worker | 350 (42.3) | 12 (22.6) | 27 (41.5) | |
| Lower white‐collar worker | 213 (25.8) | 10 (18.9) | 12 (18.5) | |
| Upper white‐collar worker | 103 (12.5) | 1 (1.9) | 3 (4.6) | |
| Other/unknown | 161 (19.5) | 30 (56.6) | 23 (35.4) | |
| Psychiatric hospitalization prior to IS (any) | 36 (4.3) | 29 (52.7) | 7 (10.4) | < 0.001*** |
| Psychotic disorder | 6 (0.7) | 14 (25.5) | 0 | |
| Mood disorder | 18 (2.1) | 19 (34.5) | 3 (4.5) | |
| Other psychiatric reason | 17 (2.0) | 16 (29.1) | 6 (9.0) | |
| Risk factors for IS | ||||
| Atrial fibrillation | 36 (4.3) | 1 (1.8) | 2 (3.0) | 0.864 |
| Cardiovascular disease | 83 (9.8) | 5 (9.1) | 7 (10.4) | 0.969 |
| Diabetes mellitus type 1 | 40 (4.7) | 3 (5.5) | 0 | 0.129 |
| Diabetes mellitus type 2 | 51 (6.0) | 2 (3.6) | 5 (7.5) | 0.704 |
| Dyslipidemia | 509 (60.3) | 35 (63.6) | 37 (55.2) | 0.617 |
| Hypertension | 350 (41.5) | 16 (29.1) | 18 (26.9) | 0.016* |
| Cardiovascular burdend | 0.472 | |||
| No comorbidities | 200 (23.7) | 15 (27.3) | 22 (32.8) | |
| ≤ 2 comorbidities | 558 (66.1) | 36 (65.5) | 40 (59.7) | |
| ≥ 3 comorbidities | 86 (10.2) | 4 (7.3) | 5 (7.5) | |
| Obesity | 99 (11.7) | 3 (5.5) | 5 (7.5) | 0.221 |
| Current cigarette smoking | 357 (42.3) | 36 (65.5) | 37 (55.2) | < 0.001*** |
| Heavy alcohol use | 93 (11.0) | 25 (45.5) | 18 (26.9) | < 0.001*** |
| History of drug abuse | 18 (2.1) | 5 (9.1) | 4 (6.0) | 0.005** |
| Stroke‐related variables measured at hospital admission | ||||
| Silent infarcts | 102 (12.1) | 11 (20.0) | 10 (14.9) | 0.200 |
| Leukoaraiosis | 47 (5.6) | 3 (5.5) | 4 (6.0) | 0.949 |
| Infarct size | 0.729 | |||
| Small | 381 (45.1) | 19 (34.5) | 30 (44.8) | |
| Medium | 237 (28.1) | 16 (29.1) | 19 (28.4) | |
| Large anterior | 125 (14.8) | 10 (18.2) | 11 (16.4) | |
| Large posterior | 101 (12.0) | 10 (18.2) | 7 (10.4) | |
| Lateralityb | 0.550 | |||
| Right | 364 (44.6) | 20 (38.5) | 26 (41.3) | |
| Left | 377 (46.2) | 24 (46.2) | 29 (46.0) | |
| Both | 75 (9.2) | 8 (15.4) | 8 (12.7) | |
| TOAST | 0.111 | |||
| Large‐artery atherosclerosis | 55 (6.5) | 7 (12.7) | 9 (13.4) | |
| Cardioembolism | 159 (18.8) | 11 (20.0) | 13 (19.4) | |
| Small‐vessel disease | 129 (15.3) | 4 (7.3) | 4 (6.0) | |
| Other | 221 (26.2) | 17 (30.9) | 17 (25.4) | |
| Undetermined causes | 280 (33.2) | 16 (29.1) | 24 (35.8) | |
| NIHSS at admission | 0.016* | |||
| 0–6, mild | 655 (77.6) | 34 (61.8) | 45 (67.2) | |
| 7–14, moderate | 123 (14.6) | 16 (29.1) | 16 (23.9) | |
| ≥ 15, severe | 66 (7.8) | 5 (9.1) | 6 (9.0) | |
| Disability at discharge | ||||
| Limb paresis at dischargec | 0.004** | |||
| No | 601 (71.9) | 29 (53.7) | 40 (59.7) | |
| Mild | 120 (14.4) | 10 (18.5) | 10 (14.9) | |
| Moderate–severe | 115 (13.8) | 15 (27.8) | 17 (25.4) | |
| Aphasia at dischargec | 186 (22.2) | 14 (25.9) | 12 (17.9) | 0.667 |
| Antipsychotic subgroups | ||||
| Typical antipsychotics | 53 (96.4) | 41 (61.2) | < 0.001*** | |
| Atypical antipsychotics | 20 (36.4) | 42 (62.7) | < 0.001*** | |
Note: Data missing in a21 (2.2%), b35 (3.6%), and c9 (0.9%) patients, dIncluding atrial fibrillation, cardiovascular disease, diabetes mellitus 1 or 2, dyslipidemia, and hypertension.
Abbreviations: IQR, interquartile range; IS, ischemic stroke; NIHSS, NIH Stroke Scale; TOAST, Trial of Org 10172 in Acute Stroke Treatment.
* p < 0.05, ** p < 0.01, *** p < 0.001. Bolded values indicate statistical significance at the p < 0.05 level.
Patients who purchased antipsychotics prior to IS had more often other/unknown socioeconomic status, history of psychiatric hospitalization, current smoking at IS occurrence, heavy alcohol use, history of drug abuse, more severe stroke symptoms on admission, and more severe limb paresis at discharge compared with others. In addition, patients without antipsychotic purchases had more often hypertension than those purchasing antipsychotics prior to IS or de novo post‐stroke (Table 1).
During a median follow‐up of 8.2 (IQR 5.1–12.7) years, follow‐up time resulted in a total of 8325 person‐years and a total of 328 (34.0%) patients had any recurrent vascular event or died; 141 (15.6%) had recurrent stroke, 192 (19.9%) had other vascular events, and 151 (15.6%) patients died from any cause. Among all deaths, 10 (6.6%) were accidental (four due to accidental poisoning, two to food inhalation, two to falls, one to drowning, and one to exposure to excessive heat of man‐made origin) and 4 (2.5%) were due to suicide. Figure 2 presents the cumulative risks and numbers of patients at risk for primary and secondary endpoint events by each antipsychotic user group.
FIGURE 2.
Cumulative risks and numbers of patients at risk for primary (A) and secondary endpoint events (B–D).
The univariable and multivariable models 2 and 3 indicated that purchase of antipsychotics any time prior IS was associated with a heightened hazard of recurrent vascular events and all‐cause mortality compared with patients with no purchases (Table 2). The association was not found between de‐novo‐initiated antipsychotics and the outcome event. Figure 3 shows that due to the violation of the proportional hazards assumption, compared to the actual data, the final multivariable Cox regression model underestimates the difference between the “no purchase” reference group and the “de novo” group in the earlier part of follow‐up and overestimates it in the later part. Moreover, the model fitted has a higher cumulative survival overall and thus seems to overestimate the survival of all subgroups, which may be explained by the predicted values being based only on mean values of the adjustment predictors and therefore do not show effects such as high age that could decrease survival.
TABLE 2.
Association between purchase of antipsychotics and recurrent vascular events or all‐cause mortality in Cox regression analyses.
| N = 966 | Univariable | Model 1 | Model 2 | Model 3 |
|---|---|---|---|---|
| Adjusted for age, sex, and socioeconomic status a | Adjusted for Model 1 + current smoking, obesity and cardiovascular burden b | Adjusted for Model 2 + TOAST | ||
| HR (95% CI) | aHR (95% CI) | aHR (95% CI) | aHR (95% CI) | |
| Purchase of antipsychotics | ||||
| No | Ref | Ref | Ref | Ref |
| Any time prior IS | 2.32 (1.61–3.35) | 1.45 (0.98–2.15) | 1.54 (1.03–2.30) | 1.54 (1.03–2.30) |
| De novo any time post IS only | 0.88 (0.57–1.35) | 0.76 (0.49–1.18) | 0.74 (0.48–1.15) | 0.73 (0.47–1.17) |
Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; HR, hazard ratio; IS, ischemic stroke; TOAST, Trial of Org 10172 in Acute Stroke Treatment.
Data missing in 21 (2.2%) patients.
Including atrial fibrillation, cardiovascular disease, diabetes mellitus 1 or 2, dyslipidemia, and hypertension.
FIGURE 3.
Survival probability and patients at risk based on (A) the actual data and (B) final cox regression model 3. While plotting the survival probabilities from the final cox model, except for the antipsychotic use, all the other variables were set to fixed or mean values. In (B), n = 945 because of missing socioeconomic status data in 21 (2.2%) patients.
The sensitivity analysis yielded results consistent with the main analysis. When antipsychotic purchases were restricted to the one‐year period before and after the index IS, prior antipsychotic use remained associated with a heightened hazard of recurrent vascular events and all‐cause mortality in Models 2 and 3, and this association was also observed in Model 1 (Table S2).
4. Discussion
In this registry‐based follow‐up study on 966 young adults with first‐ever IS, 6% had purchased antipsychotics at some point prior to IS and 7% initiated antipsychotics de novo after IS during the follow‐up. Compared with patients who purchased antipsychotics de novo after IS and with those without antipsychotic purchases, prior to IS antipsychotic users had more often other/unknown socioeconomic status, history of psychiatric hospitalization, and poor lifestyle habits, as well as more severe index events. Furthermore, as many as 58% of prior users and 33% of de novo users had any recurrent vascular event or died from any cause during the follow‐up. We also found an association between antipsychotic use prior to IS and these endpoint events in the long term.
The effects of antipsychotics or psychotic disorders, especially in young IS patients, have been poorly characterized. Studies have shown that patients with psychotic disorders have an increased risk for stroke, cardiometabolic outcomes and all‐cause mortality [2, 3, 4, 23]. Additionally, these patients tend to be younger [3, 4, 24] and it has been hypothesized that traditional stroke‐related risk factors, which have also been associated with psychiatric disorders, might be one explanation of this connection [5]. Also, individuals with a history of diagnosed psychosis or schizophrenia often have generally poorer lifestyle habits, including high rates of smoking, drug and alcohol abuse [25]. They also tend to have poorer somatic health, with more cardiovascular risk factors, a higher likelihood of developing comorbidities, and excess cardiovascular mortality, than those without these disorders [24, 26, 27]. Furthermore, these patients often have sub‐optimal treatment for hypertension and dyslipidemia [6, 24]. Also, second‐generation antipsychotics are reported to be associated with metabolic side effects such as weight gain and negative effects on glucose and lipid metabolism [28, 29]. Our results align with these previous studies, although we did not observe a difference in the prevalence of cardiovascular comorbidities between patients using antipsychotics prior IS or de novo post‐stroke and non‐users. This might be due to some methodological differences, for example, that the actual psychiatric diagnoses were not available in our study. The association between prior antipsychotic use and the composite primary endpoint in this study appears to be primarily driven by vascular events other than stroke and by all‐cause mortality. This may partly reflect our definitions of vascular endpoint events but likely also reflects the substantial cardiovascular risks associated with underlying psychiatric comorbidities and antipsychotic medications themselves, mediated through both direct and indirect mechanisms. Both typical and atypical antipsychotics are known to prolong the QT interval—particularly in patients with pre‐existing cardiac disease—thereby increasing the risk of ventricular arrhythmias and sudden cardiac death [30, 31]. Moreover, antipsychotics with high or intermediate metabolic risk, particularly in older patients, have been associated with an increased long‐term risk of major cardiovascular events, including thromboembolic events and myocardial infarction, compared with low‐risk agents [32]. The combined direct and indirect effects of antipsychotics likely contribute to increased cardiovascular mortality in patients with severe mental disorders, with the risk rising in a dose‐dependent manner [33, 34, 35].
Our study found that young patients who purchased antipsychotics prior to IS had more often more severe index events than other IS patients, consistent with findings from a large Danish study [7] and a retrospective German study [36]. Our study also revealed that antipsychotic use prior to IS was associated with recurrent vascular events, including strokes and all‐cause mortality. In line with this, a Swedish registry‐based study showed that, in addition to poorer short‐ and long‐term functional outcomes, patients with psychosis prior to IS had a higher stroke recurrence compared with patients without [6]. The proportion of antipsychotic users prior to stroke was somewhat higher in the Danish study; 7.5% of the patients had filled prescriptions of antipsychotics within 3–12 months prior to admission [7], whereas in our cohort 5.7% had purchased antipsychotics at any point and 2.8% within one year prior to IS. However, the Danish study also included hemorrhagic strokes and older patients, which might increase the number of users.
In our study, 6.9% of young IS survivors purchased antipsychotics de novo post‐stroke. However, we found no association between these patients newly starting antipsychotics after IS and recurrent vascular events or all‐cause mortality in the long term. Psychiatric disorders, including depression, anxiety, insomnia, and other mental disturbances after stroke affect nearly one‐third of all stroke patients and can potentially worsen recovery, outcome, and quality of life [8, 9, 10, 11, 12, 13, 14, 15, 16]. Furthermore, previous studies report post‐stroke psychotic symptoms to be rare but having an association with poor functional outcomes and increased long‐term mortality [10, 15]. In a follow‐up study of all first‐ever stroke patients in Western Australia, the cumulative incidence of psychosis in patients without prior psychiatric disorder was 6.7% [15], corresponding to our results. In addition, compared with patients with other post‐stroke psychiatric disorders, those with psychosis had the lowest survival rates, with cardiovascular disease being the most frequent cause of death [15].
Our study has several strengths, including a large, consecutive cohort of young IS patients and detailed baseline data on stroke characteristics. As an observational study, it benefits from reflecting a real‐life setting and is free from consent bias. We collected data from Finnish registers, which are obligatory and known to be of good to high quality [37, 38], reducing information bias and allowing the evaluation of outcomes of interest. Furthermore, the outcome data were mostly verified from patient records [21]. While we could not confirm whether the purchase of antipsychotics led to the actual use of the medication, we believe our study is likely to provide a reliable reflection of actual drug use.
Our study also has some unavoidable limitations due to its retrospective and observational design. Since antipsychotic use was defined based on filled prescriptions, we were unable to assess the exact dosage, indication, or underlying diagnosis; thus, antipsychotic use served as a proxy measure. As these medications are prescribed not only for schizophrenia and other psychotic or affective disorders but also for conditions such as personality and anxiety disorders, insomnia, minor mood symptoms, and agitation in dementia among elderly patients [39, 40], their use does not necessarily indicate the presence of a psychiatric disorder. Consequently, the study population may include a broad spectrum of psychiatric disease severity. Furthermore, data on concomitant use and adherence to secondary stroke prevention medications—and thus the ability to assess their effect—were not available in this study. Previous Finnish studies utilizing HYSR data have reported suboptimal antihypertensive use in approximately one‐third of young IS patients, and statin use in fewer than half by the end of follow‐up [41, 42]. Assessment of adherence to antithrombotic therapy is also challenging, as aspirin, for example, is available over the counter in Finland. Nonetheless, evaluation of secondary preventive medication use was beyond the scope of the present study. As the follow‐up time started at the index event, immortal time bias might be present. A total of 109 (11.3%) of those not purchasing any antipsychotics died during the follow‐up, with a median time of 6.0 (IQR 2.0–10.0) years. In contrast, the median time to the first purchase of antipsychotics de novo post‐stroke was 2.2 (IQR 0.66–6.9) years, meaning that those dying before starting antipsychotics were less likely to have been potential new users. Additionally, some patients included earlier in the study, with a shorter retrospective look‐back period, may have been misclassified as non‐prior users. However, in the sensitivity analysis, the results were consistent with and supported the main analysis. Regarding statistical methods, we thoroughly assessed the proportional hazards assumption and found that it was violated in the Cox regression model. Despite exploring several alternative statistical approaches to correct for this, no improvement was made. Therefore, while the Cox proportional hazards may not be the most ideal model for this data, it remains the best option for assessing the effect of antipsychotic use over the follow‐up period while controlling for known confounders. Furthermore, comparing survival probability curves from the actual data and the final multivariable Cox regression model suggested that, if anything, the model rather seemed to overestimate the survival of all groups. This indicates that patients using antipsychotics prior IS might experience the endpoint event at a higher rate than the model suggests.
5. Conclusion
In conclusion, our study suggests that young IS patients with prior antipsychotic use form a high‐risk group for recurrent vascular events and death. A closer post‐stroke follow‐up should be considered to improve long‐term outcomes for these patients.
Author Contributions
Jenna Broman: conceptualization, writing – original draft, formal analysis, methodology, data curation. Hanna Granroth‐Wilding: formal analysis, writing – review and editing, methodology. Ivan Marinkovic: conceptualization, methodology, writing – review and editing. Markku Kaste: conceptualization, methodology, data curation, writing – review and editing. Jukka Putaala: conceptualization, methodology, data curation, supervision, formal analysis, writing – review and editing. Karoliina Aarnio: conceptualization, supervision, data curation, methodology, writing – review and editing, formal analysis. Turgut Tatlisumak: conceptualization, methodology, data curation, writing – review and editing.
Funding
Jenna Broman received support from Helsinki and Uusimaa Hospital District (Y223524003) and from Maire Taponen Foundation and Karoliina Aarnio from Maire Taponen Foundation.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: ICD‐9 and ICD‐10 codes used to identify outcome events.
Table S2: Association between antipsychotics purchased within one year prior to or after the index IS and recurrent vascular events or all‐cause mortality in Cox regression sensitivity analysis.
Acknowledgements
Anu Eräkanto, Department of Neurology, Helsinki University Hospital, Helsinki, Finland. Open access publishing facilitated by Helsingin yliopisto, as part of the Wiley ‐ FinELib agreement.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Feigin V. L., Stark B. A., Johnson C. O., et al., “Global, Regional, and National Burden of Stroke and Its Risk Factors, 1990–2019: A Systematic Analysis for the Global Burden of Disease Study 2019,” Lancet Neurology 20, no. 10 (2021): 1–26, 10.1016/S1474-4422(21)00252-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Li M., Fan Y. L., Tang Z. Y., and Cheng X. S., “Schizophrenia and Risk of Stroke: A Meta‐Analysis of Cohort Studies,” International Journal of Cardiology 173, no. 3 (2014): 588–590, 10.1016/j.ijcard.2014.03.101. [DOI] [PubMed] [Google Scholar]
- 3. Tsai K. Y., Lee C. C., Chou Y. M., Su C. Y., and Chou F. H. C., “The Incidence and Relative Risk of Stroke in Patients With Schizophrenia: A Five‐Year Follow‐Up Study,” Schizophrenia Research 138, no. 1 (2012): 41–47, 10.1016/J.SCHRES.2012.02.013. [DOI] [PubMed] [Google Scholar]
- 4. Lee Y. B., Kim H., Lee J., et al., “Psychotic Disorders and the Risk of Type 2 Diabetes Mellitus, Atherosclerotic Cardiovascular Diseases, and All‐Cause Mortality: A Population‐Based Matched Cohort Study,” Diabetes & Metabolism Journal 48, no. 1 (2024): 122–133, 10.4093/DMJ.2022.0431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Chiu Y. C., Bai Y. M., Su T. P., Chen T. J., and Chen M. H., “Ischemic Stroke in Young Adults and Preexisting Psychiatric Disorders: A Nationwide Case‐Control Study,” Medicine 94, no. 38 (2015): e1520, 10.1097/MD.0000000000001520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Willers C., Sunnerhagen K. S., Lekander I., and von Euler M., “The Association of Pre‐Stroke Psychosis and Post‐Stroke Levels of Health, Resource Utilization, and Care Process: A Register‐Based Study,” Frontiers in Neurology 9 (2018): 1042, 10.3389/fneur.2018.01042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Prior A., Laursen T. M., Larsen K. K., et al., “Post‐Stroke Mortality, Stroke Severity, and Preadmission Antipsychotic Medicine Use ‐ A Population‐Based Cohort Study,” PLoS One 9, no. 1 (2014): e84103, 10.1371/journal.pone.0084103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Chemerinski E. and Robinson R. G., “The Neuropsychiatry of Stroke,” Psychosomatics 41, no. 1 (2000): 5–14, 10.1016/S0033-3182(00)71168-6. [DOI] [PubMed] [Google Scholar]
- 9. Angelelli P., Paolucci S., Bivona U., et al., “Development of Neuropsychiatric Symptoms in Poststroke Patients: A Cross‐Sectional Study,” Acta Psychiatrica Scandinavica 110, no. 1 (2004): 55–63, 10.1111/j.1600-0447.2004.00297.x. [DOI] [PubMed] [Google Scholar]
- 10. Stangeland H., Orgeta V., and Bell V., “Review: Poststroke Psychosis: A Systematic Review,” Journal of Neurology, Neurosurgery, and Psychiatry 89, no. 8 (2018): 879, 10.1136/JNNP-2017-317327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hackett M. L., Yapa C., Parag V., and Anderson C. S., “Frequency of Depression After Stroke,” Stroke 36, no. 6 (2005): 1330–1340, 10.1161/01.STR.0000165928.19135.35. [DOI] [PubMed] [Google Scholar]
- 12. Ayerbe L., Ayis S., Wolfe C. D. A., and Rudd A. G., “Natural History, Predictors and Outcomes of Depression After Stroke: Systematic Review and Meta‐Analysis,” British Journal of Psychiatry 202, no. 1 (2013): 14–21, 10.1192/bjp.bp.111.107664. [DOI] [PubMed] [Google Scholar]
- 13. Towfighi A., Ovbiagele B., El Husseini N., et al., “Poststroke Depression: A Scientific Statement for Healthcare Professionals From the American Heart Association/American Stroke Association,” Stroke 48, no. 2 (2017): e30, 10.1161/STR.0000000000000113. [DOI] [PubMed] [Google Scholar]
- 14. Ayerbe L., Ayis S., and Crichton S., “The Long‐Term Outcomes of Depression Up to 10 Years After Stroke; the South London Stroke Register,” Journal of Neurology, Neurosurgery, and Psychiatry 85, no. 5 (2014): 514–521, 10.1136/jnnp-2013-306448. [DOI] [PubMed] [Google Scholar]
- 15. Almeida O. P. and Xiao J., “Mortality Associated With Incident Mental Health Disorders After Stroke,” Australian and New Zealand Journal of Psychiatry 41, no. 3 (2007): 274–281, 10.1080/00048670601172772. [DOI] [PubMed] [Google Scholar]
- 16. Leppävuori A., Pohjasvaara T., Vataja R., Kaste M., and Erkinjuntti T., “Insomnia in Ischemic Stroke Patients,” Cerebrovascular Diseases 14, no. 2 (2002): 90–97, 10.1159/000064737. [DOI] [PubMed] [Google Scholar]
- 17. Ferro J. M., Caeiro L., and Figueira M. L., “Neuropsychiatric Sequelae of Stroke,” Nature Reviews Neurology 12, no. 5 (2016): 269–280, 10.1038/NRNEUROL.2016.46. [DOI] [PubMed] [Google Scholar]
- 18. Hackett M. L., Köhler S., O'Brien J. T., and Mead G. E., “Neuropsychiatric Outcomes of Stroke,” Lancet Neurology 13, no. 5 (2014): 525–534, 10.1016/S1474-4422(14)70016-X. [DOI] [PubMed] [Google Scholar]
- 19. Aho K., Harmsen P., Hatano S., Marquardsen J., Smirnov V. E., and Strasser T., “Cerebrovascular Disease in the Community: Results of a WHO Collaborative Study,” Bulletin of the World Health Organization 58, no. 1 (1980): 113–130. [PMC free article] [PubMed] [Google Scholar]
- 20. Putaala J., Metso A. J., Metso T. M., et al., “Analysis of 1008 Consecutive Patients Aged 15 to 49 With First‐Ever Ischemic Stroke: The Helsinki Young Stroke Registry,” Stroke 40, no. 4 (2009): 1195–1203, 10.1161/STROKEAHA.108.529883. [DOI] [PubMed] [Google Scholar]
- 21. Aarnio K., Siegerink B., Pirinen J., et al., “Cardiovascular Events After Ischemic Stroke in Young Adults: A Prospective Follow‐Up Study,” Neurology 86, no. 20 (2016): 1872–1879, 10.1212/WNL.0000000000002689. [DOI] [PubMed] [Google Scholar]
- 22. Putaala J., Haapaniemi E., Metso A. J., et al., “Recurrent Ischemic Events in Young Adults After First‐Ever Ischemic Stroke,” Annals of Neurology 68, no. 5 (2010): 661–671, 10.1002/ANA.22091. [DOI] [PubMed] [Google Scholar]
- 23. Shen Q., Mikkelsen D. H., Luitva L. B., et al., “Psychiatric Disorders and Subsequent Risk of Cardiovascular Disease: A Longitudinal Matched Cohort Study Across Three Countries,” EClinicalMedicine 61 (2023): 102063, 10.1016/j.eclinm.2023.102063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Lahti M., Tiihonen J., Wildgust H., et al., “Cardiovascular Morbidity, Mortality and Pharmacotherapy in Patients With Schizophrenia,” Psychological Medicine 42, no. 11 (2012): 2275–2285, 10.1017/S0033291712000396. [DOI] [PubMed] [Google Scholar]
- 25. Kelly D. L., McMahon R. P., Wehring H. J., et al., “Cigarette Smoking and Mortality Risk in People With Schizophrenia,” Schizophrenia Bulletin 37, no. 4 (2011): 832–838, 10.1093/SCHBUL/SBP152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Bresee L. C., Majumdar S. R., Patten S. B., and Johnson J. A., “Prevalence of Cardiovascular Risk Factors and Disease in People With Schizophrenia: A Population‐Based Study,” Schizophrenia Research 117, no. 1 (2010): 75–82, 10.1016/j.schres.2009.12.016. [DOI] [PubMed] [Google Scholar]
- 27. Jin H., Folsom D., Sasaki A., et al., “Increased Framingham 10‐Year Risk of Coronary Heart Disease in Middle‐Aged and Older Patients With Psychotic Symptoms,” Schizophrenia Research 125, no. 2–3 (2011): 295–299, 10.1016/j.schres.2010.10.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Howes O. D., Bhatnagar A., Gaughran F. P., Amiel S. A., Murray R. M., and Pilowsky L. S., “A Prospective Study of Impairment in Glucose Control Caused by Clozapine Without Changes in Insulin Resistance,” American Journal of Psychiatry 161, no. 2 (2004): 361–363, 10.1176/appi.ajp.161.2.361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Pillinger T., McCutcheon R. A., Vano L., et al., “Comparative Effects of 18 Antipsychotics on Metabolic Function in Patients With Schizophrenia, Predictors of Metabolic Dysregulation, and Association With Psychopathology: A Systematic Review and Network Meta‐Analysis,” Lancet Psychiatry 7, no. 1 (2020): 64–77, 10.1016/S2215-0366(19)30416-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Brown C. S., Farmer R. G., Soberman J. E., and Eichner S. F., “Pharmacokinetic Factors in the Adverse Cardiovascular Effects of Antipsychotic Drugs,” Clinical Pharmacokinetics 43, no. 1 (2004): 33–56, 10.2165/00003088-200443010-00003. [DOI] [PubMed] [Google Scholar]
- 31. Edinoff A. N., Ellis E. D., Nussdorf L. M., et al., “Antipsychotic Polypharmacy‐Related Cardiovascular Morbidity and Mortality: A Comprehensive Review,” Neurology International 14, no. 1 (2022): 294–309, 10.3390/NEUROLINT14010024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Szmulewicz A. G., Angriman F., Pedroso F. E., Vazquez C., and Martino D. J., “Long‐Term Antipsychotic Use and Major Cardiovascular Events: A Retrospective Cohort Study,” Journal of Clinical Psychiatry 78, no. 8 (2017): e905‐e912, 10.4088/JCP.16m10976. [DOI] [PubMed] [Google Scholar]
- 33. Osborn D. P. J., Levy G., Nazareth I., Petersen I., Islam A., and King M. B., “Relative Risk of Cardiovascular and Cancer Mortality in People With Severe Mental Illness From the United Kingdom's General Practice Research Database,” Archives of General Psychiatry 64, no. 2 (2007): 242–249, 10.1001/ARCHPSYC.64.2.242. [DOI] [PubMed] [Google Scholar]
- 34. Leung J. Y. T., Barr A. M., Procyshyn R. M., Honer W. G., and Pang C. C. Y., “Cardiovascular Side‐Effects of Antipsychotic Drugs: The Role of the Autonomic Nervous System,” Pharmacology & Therapeutics 135, no. 2 (2012): 113–122, 10.1016/j.pharmthera.2012.04.003. [DOI] [PubMed] [Google Scholar]
- 35. Howell S., Yarovova E., Khwanda A., and Rosen S. D., “Cardiovascular Effects of Psychotic Illnesses and Antipsychotic Therapy,” Heart 105, no. 24 (2019): 1852–1859, 10.1136/heartjnl-2017-312107. [DOI] [PubMed] [Google Scholar]
- 36. Hoyer C., Schmidt H. L., Kranaster L., and Alonso A., “Impact of Psychiatric Comorbidity on the Severity, Short‐Term Functional Outcome, and Psychiatric Complications After Acute Stroke,” Neuropsychiatric Disease and Treatment 15 (2019): 1823–1831, 10.2147/NDT.S206771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Tolonen H., Salomaa V., Torppa J., et al., “The Validation of the Finnish Hospital Discharge Register and Causes of Death Register Data on Stroke Diagnoses,” European Journal of Cardiovascular Prevention and Rehabilitation 14, no. 3 (2007): 380–385, 10.1097/01.HJR.0000239466.26132.F2. [DOI] [PubMed] [Google Scholar]
- 38. Sund R., “Quality of the Finnish Hospital Discharge Register: A Systematic Review,” Scandinavian Journal of Public Health 40, no. 6 (2012): 505–515, 10.1177/1403494812456637. [DOI] [PubMed] [Google Scholar]
- 39. Maher A. R., Maglione M., Bagley S., et al., “Efficacy and Comparative Effectiveness of Atypical Antipsychotic Medications for Off‐Label Uses in Adults: A Systematic Review and Meta‐Analysis,” Journal of the American Medical Association 306, no. 12 (2011): 1359–1369, 10.1001/JAMA.2011.1360. [DOI] [PubMed] [Google Scholar]
- 40. Maglione M., Maher A. R., Hu J., et al., “Off‐Label Use of Atypical Antipsychotics: An Update,” Comparative Effectiveness Review Summary Guides for Clinicians 43 (2012): 1–437, http://www.effectivehealthcare.ahrq.gov/reports/final.cfm. [Google Scholar]
- 41. Van Dongen M. M. E., Aarnio K., Martinez‐Majander N., et al., “Use of Antihypertensive Medication After Ischemic Stroke in Young Adults and Its Association With Long‐Term Outcome,” Annals of Medicine 51, no. 1 (2019): 68–77, 10.1080/07853890.2018.1564358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Van Dongen M. M. E., Aarnio K., Martinez‐Majander N., et al., “Use of Statins After Ischemic Stroke in Young Adults and Its Association With Long‐Term Outcome,” Stroke 50, no. 12 (2019): 3385–3392, 10.1161/STROKEAHA.119.026992. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: ICD‐9 and ICD‐10 codes used to identify outcome events.
Table S2: Association between antipsychotics purchased within one year prior to or after the index IS and recurrent vascular events or all‐cause mortality in Cox regression sensitivity analysis.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.