A population-based cohort study on deep vein thrombosis and pulmonary embolism among schizophrenia patients.

A population-based cohort study on deep vein thrombosis

and

pulmonary embolism among schizophrenia patients

Wen-Yu Hsu

, Hsien-Yuan Lane

, Cheng-Li Lin

, Chia-Hung Kao

⁎

1. Introduction

In the United States, the annual incidence of venous thromboembolisms (VTEs), including deep vein thrombosis (DVT) and pulmonary

embolism (PE), is approximately 1 per 1000 adults, causing discomfort,

suffering, and occasionally death (White, 2003). Death occurs within

one month of incidence in approximately 6% of patients with DVT

and in 10% of those with PE (Cushman et al., 2004). Several studies

have identifed age, immobilization, obesity, smoking status, allergy, autoimmune disease, heart failure, lower leg fracture, surgery, diabetes, pregnancy, antipsychotics, physical restraint and cancer as acquired risk factors for VTE.

Few studies have investigated DVT and PE in patients with schizophrenia. Schizophrenia is a chronic, severe, and disabling brain disorder. Worldwide, the lifetime prevalence of schizophrenia is approximately 1%. Schizophrenia patients also have more physical problems than general population. In the recent Canada study, they found that schizophrenia patients are at increased risk of death compared to the general population, and the majority of these deaths are

occurring in older age from physical disease processes (Kredentser

et al., 2014). As we have known, there are several risk factors of DVT and PE in schizophrenia patients. These factors might increase risk of DVT and PE in this population. Patients with schizophrenia may exhibit long-term positive, negative, cognitive, or affective symptoms. Negative symptoms include blunted or fat affect, alogia,

avolition, apathy, anticipatory anhedonia, and asociality (Choi and

Medalia, 2010; Foussias and Remington, 2010; Blanchard et al.,

2011). Avolition was found to be a signifcant predictor for motor

activity levels in patients with schizophrenia (Docx et al., 2013).

Immobilization was associated with risk of DVT and PE in previous

study (Beam et al., 2009). And, increased risk of VTE among subjects

with current antipsychotic use was found (Wu et al., 2013). Besides,

Cooper et al., 2012; Blondon

et al., 2013; Minichino et al., 2013). High metabolic syndrome risk in this population also increases the risk of DVT and PE (Kawasaki et al., 1997; Vazzana et al., 2012; Chang et al., 2013; Klovaite et al., 2014).

We hypothesize that the incidence of VTE is higher among schizophrenia patients in comparison with the general population. Accordingly, this population-based cohort study was conducted to investigate whether DVT and PE are more prevalent among schizophrenia patients in Taiwan.

2. Methods 2.1. Data source

This study used data that were obtained from the National Health Insurance Research Database (NHIRD), which contains claims data from 1996 to 2011 related to inpatient care, ambulatory care, dental care, and prescription drugs. In Taiwan, national health care is obligatory, and few people are excluded from the system. The National Health Insurance (NHI) program, which was implemented on March 1, 1995, provides insurance coverage for approximately 99% of Taiwan's

population of approximately 23.74 million people (http://www.nhi.

gov.tw/english/index.aspx).

The NHI catastrophic illness fles—the Registry of Catastrophic Illnesses Patient Database (RCIPD)—were established to track patients with major or catastrophic illnesses, including cancer, end-stage renal disease, autoimmune diseases, congenital illness, and several mental illness, such as schizophrenia. To validate the diagnosis of patients, the Bureau of NHI routinely reviews the original medical charts of all patients who apply for catastrophic illness registration. This study analyzed depersonalized secondary data; thus, no informed consent was required. This study was approved by the Ethics Review Board of China Medical University (CMU-REC-101-012). Diagnostic codes are reported based on the International Classifcation of Disease, 9th Revision, Clinical Modifcation (ICD-9-CM).

2.2. Participant selection

The study subjects were identifed from the following two subsets of the NHIRD: First, cases were confrmed from the RCIPD. The schizophrenia cohort comprised patients aged 20 years who were newly diagnosed with schizophrenia (ICD-9-CM Code 295 except

295.8) between 2000 and 2011. The index date was set as the date at which the schizophrenia patients were registered for catastrophic illness; second, the control patient data were obtained from the LHID2000, which is a database containing longitudinally linked claims data (from 1996 to 2011) of one million people randomly

sampled fromthe NHIRD. The National Health Research Institutes reported that the difference in age, sex, and health care costs between

the LHID2000 and all enrollees is nonsignifcant. A nonschizophrenia comparison cohort was randomly selected from the LHID2000, comprising patients without schizophrenia and frequency matched with

the schizophrenia cohort at a 4:1 ratio based on age (in 5-year increments), sex, and year of diagnosis. For both cohorts, patients with a history of DVT (ICD-9-CM 453.8) or PE (ICD-9-CM 415.1, excluding ICD-9-CM 415.11) before the index date, as well as those with incomplete age or sex information were excluded.

2.3. Variables

We considered atrial fbrillation (ICD-9 Code 427.31), hypertension (ICD-9 Codes 401–405), diabetes (ICD-9 Code 250), cerebral vascular disease (CVA; ICD-9 Codes 430–438), heart failure (ICD-9 Code 428), all types of cancer (ICD-9 Codes 140–208), pregnancy (ICD-9 Codes 640.×1–676.×1, 640.×2–676.×2, and 650–659, as well as Procedure Codes 72–74), and lower leg fracture or surgery (ICD-9 Codes 820–823, as well as Procedure Codes 81.51, 81.52, 81.53, and 81.54) as preexisting comorbidities that were potential confounders in the association with DVT and PE. Information on benzodiazepine (BZD) treatment at the baseline was obtained. We hypothesized that frst-generation antipsychotics, and second-generation antipsychotics have different effects on DVT and PE in schizophrenia patients. 2.4. Main outcome

The study participants were followed from the index date

until DVT or PE diagnosis, withdrawal from the insurance program, censoring because of death, or the end date of the database (December 31, 2011).

2.5. Statistical analysis

All analyses were performed using SAS for Windows (version 9.3; SAS Institute, Inc., Cary, NC, USA). A two-tailed P value of 0.05 was considered statistically signifcant. Pearson's chi-square test was used for categorical variables, such as age group (20–34, 35–49, and

≥50 years), sex, comorbidities, and BZD treatment between the schizophrenia and nonschizophrenia cohorts. Continuous variables were analyzed using a two-sample t test. The cumulative incidence of DVT or PE among the schizophrenia and nonschizophrenia cohorts was assessed using the Kaplan–Meier method, and the differences were assessed with a log-rank test. The overall, gender-, age-, and comorbidity-specifc incidence rates (per 10 000 person-years) were calculated for each cohort. Univariable and multivariable Cox proportional hazard regression analyses were used to assess the risk of DVT

and PE development associated with schizophrenia, compared with that of the nonschizophrenia cohort. Hazard ratios (HRs) and 95% confdence intervals (CIs) were estimated in the Cox models. For the multivariable model, the effects of sex, age, and comorbidities were controlled to determine whether these factors cause any signifcant

difference between the two cohorts (Table 1). 3. Results

Our study included 60,264 patients with schizophrenia and 60,264 control patients without schizophrenia. Among the schizophrenia patients, 52.5% of them were men and 44.4% of them were younger

than 34 years of age (Table 1). The mean ages of the schizophrenia

and nonschizophrenia cohorts are 38.8 ± 13.3 years and 38.7 ± 13.8 years, respectively. Compared with the nonschizophrenia cohort, the schizophrenia patients exhibited a higher prevalence of diabetes, CVA, heart failure, lower leg fracture or surgery, and BZD treatment (P b 0.001). Compared with schizophrenia patients, those without schizophrenia typically had a higher prevalence of hyperlipidemia, cancer, and pregnancy (P b 0.001).

The Kaplan–Meier analysis results show that the cumulative

incidence curves of DVT were signifcantly higher in the schizophrenia cohort than in the nonschizophrenia cohort (log-rank test P b 0.001;

Fig. 1A). During the mean follow-up period of 6.40 years for the schizophrenia cohort and 6.48 years for the non-schizophrenia cohort, the overall incidence density of DVT was signifcantly higher (2.06-fold) in schizophrenia cohort than in the nonschizophrenia cohort (4.17 vs 2.02 per 10 000 person-years) with an adjusted HR of 2.02 (95% CI = 1.52–2.70) after adjusting the model to control

the effects of age, sex, comorbidities, and BZD treatment (Table 2).

In both cohorts, the DVT incidence was higher in women than in men. The sex-specifc schizophrenia cohort to nonschizophrenia

cohort relative risk of DVT development is signifcantly higher for both women (adjusted HR = 1.65, 95% CI = 1.15–2.36) and men (adjusted HR = 2.94, 95% CI = 1.79–4.83). The age-specifc schizophrenia cohort to nonschizophrenia cohort relative risk of DVT development is signifcantly higher for patients younger 54 years (adjusted HR = 2.64, 95% CI = 1.07–6.53) and those aged 55–69 years (adjusted HR = 2.63, 95% CI = 1.61–4.30). Compared with the nonschizophrenia cohort, the schizophrenia patients exhibited a higher risk of developing DVT for those patients without

comorbidities (adjusted HR = 2.10, 95% CI = 1.41–3.15). Fig. 1B

shows the cumulative PE incidence curve for the two cohorts, as

well as cases where the schizophrenia cohort incidence curve is signifcantly higher than that of the nonschizophrenia cohort (log-rank

test P = 0.002). Overall, the risk of developing PE was 1.99-fold higher among the schizophrenia cohort in comparison with the nonschizophrenia cohort (95% CI = 1.21–3.27) after adjusting the model to control the effects of age, sex, comorbidities, and BZD treatment. The sex-specifc analysis results indicate that the risk of PE development is signifcantly higher among women in the schizophrenia cohort compared with those in the nonschizophrenia cohort (adjusted HR = 1.97; 95% CI = 1.03–3.78).

Further analysis was performed to assess whether the antipsychotic medications infuenced the DVT and PE outcomes. (Supplementary

table). We observed that the schizophrenia patients receiving frstgeneration antipsychotics or receiving second-generation antipsychotic

treatment exhibited a higher risk of DVT and PE development than those without schizophrenia.

4. Discussion

Our results show that the schizophrenia patients have a 2.02-fold (1.99-fold) higher adjusted HR for DVT (PE) development than the nonschizophrenia cohort. In the United States from 2002 to 2007, Khaykin et al. reported that hospitalizations of patients with secondary diagnosis of schizophrenia were associated with a higher probability of

developing PE or DVT compared to the general population (Khaykin

et al., 2010). Daumit et al. reported that hospitalizations of schizophrenia patients had a relatively higher adjusted odd for postoperative DVT

development in comparison with thosewithout schizophrenia (Daumit

development among non-psychiatric hospitalizations of schizophrenia patients compared with non-schizophrenia patients. Few studies have investigated the association between schizophrenia and DVT or PE development. This study is the frst showing that the risk of DVT and PE development is higher among schizophrenia patients compared with non-schizophrenia patients.

Furthermore, the schizophrenia cohort receiving frst-generation or second-generation antipsychotics exhibited a higher adjusted HR for DVT and PE development in comparison with the non-schizophrenia cohort. This result is similar to fndings of several previous studies (Liperoti et al., 2005; Masopust et al., 2007; Wu et al., 2013). Masopust et al. indicated a potentially increased risk of VTE development in

patients using antipsychotic drugs (Masopust et al., 2007). Wu et al.

reported that current antipsychotic drug use was associated with a

52% increased risk of developing VTE (Wu et al., 2013). Furthermore,

they showed that all classes of antipsychotics were associated with increased risk of VTE.

Liperoti et al. reported that second-generation antipsychotics,

including risperidone, olanzapine, clozapine, and quetiapine, appear to

increase the risk of developing VTE (Liperoti et al., 2005). However,

this study observed no increased risk associated with using phenothiazine or other frst-generation antipsychotic agents. Many patients

with schizophrenia would have had different antipsychotic during their life-time, and there is no correction for duration of potential

exposure in this study. But, antipsychotics play a critical role in DVT and PE development in schizophrenia patients. Based on the fndings of this study, antipsychotics should be considered a contributing factor of DVT and PE development.

In this study, the schizophrenia cohort exhibited a higher incidence of DVT and PE compared with the non-schizophrenia cohort, which could be related to several factors. First, the negative symptoms of schizophrenia could lead to a decrease in activities of daily living among schizophrenia; moreover, antipsychotics could also lead to a reduction in daily activity. Hypoactivity and immobility increase the

risk of developing DVT and PE (Beam et al., 2009). Second, antipsychotic

exposure could increase the risk of DVT and PE (Wu et al., 2013; Barbui

et al., 2014). Third, schizophrenia patientswho are physically restrained

Cecchi et al., 2012). Fourth, higher rates of smoking and metabolic syndrome among schizophrenia patients could also increase the risk (Kawasaki et al., 1997; de Leon and Diaz, 2005; Severinsen et al., 2009; Vazzana et al., 2012; Blondon et al., 2013; Chang et al., 2013; Klovaite et al., 2014).

The strength of our fndings is the result of the nationwide

population-based design of this longitudinal study on the risks of DVT and PE development among in schizophrenia patients in Taiwan. In addition, the diagnosis of all of schizophrenia patients identifed in the NHIRD claims data was confrmed through their inclusion in the RCIPD, indicating that the data are highly reliable. Because all residents in Taiwan are assigned a unique personal identifcation number, the medical history of all patients could be traced throughout the entire follow-up period. Thus, our fndings are generalizable to the entire population of Taiwan.

However, some limitations were encountered while conducting this study. First, the accuracy of DVT and PE incidence was not validated by reviewingmedical charts, and VTE cases were identifed based only on ICD-9-CM Codes 415.1 (PE) and 453.8 (DVT). Second, in our analysis, a lack of data (e.g., smoking status, obesity, physical restraint, hormone replacement therapy, the use of contraceptive drugs, and glucocorticosteroid treatment) may have infuenced the results of this study. Therefore, we used proxy measures (e.g., such as hyperlipidemia and diabetes as indicators of obesity) to control the potential effects of unmeasurable confounders. However, some unmeasured confounders could not be addressed by this study. Third, our comparison cohort included inpatients, which may have reduced the potential for overestimating the risk, specifcally because VTE is a typical complication during and after hospitalization

for acute medical illness or surgery.

In conclusion, this nationwide population-based cohort study shows that schizophrenia patients have a 2.02-fold (1.99-fold) higher adjusted HR for DVT (PE) development in comparisonwith the non-schizophrenia cohort. These fndings indicate the importance of using a multidisciplinary approach in managing the potential risk factors for DVT and PE

development among schizophrenia patients, particularly those who, because of their illness, are unable to maintain adequate personal care or gain access to health care by themselves. Occasionally, DVT is a “silent”

disease; for patientswith schizophrenia, identifying the early signs of DVT may be diffcult. In this study, higher incidences of DVT and PE were observed in the schizophrenia cohort than in non-schizophrenia cohort. Regular physical examinations and history taking are critical to detect DVT and PE. Early diagnosis and intervention could mitigate complications or and reduce mortality.

Supplementary data to this article can be found online at http://dx.