Strokes are the third leading cause of death and the most common cause of complex disability in Taiwan [1
]. Many medical complications of stroke are common and often lead to poor clinical outcomes, such as depression, known as post-stroke depression (PSD). PSD has a high prevalence and a negative impact on stroke patients’ long-term survival and well-being [2
]. Thus, it is important to reduce or prevent PSD in stroke survivors.
Many studies have evaluated various approaches to prevent PSD, including pharmacological therapy and psychotherapy. In some clinical trials, medicines such as Escitalopram and Duloxetine have shown some effectiveness in preventing PSD; however, most remedies had small treatment effect [3
]. Other studies have also revealed that rehabilitation after a stroke may have positive effects in preventing PSD [4
Traditional Chinese medicine (TCM) is classified as one form of complementary and alternative medicine (CAM) and is popular in Asian countries; for example, 52.7% of stroke patients receive both Chinese herbal remedies and acupuncture/traumatology treatment in Taiwan [5
]. Some studies have reported that acupuncture may be effective in preventing PSD; however, such studies have so far been either too small or lacking big data analysis. If a case is to be made for the usefulness of acupuncture in relation to PSD, it is imperative to provide empirical evidence showing its positive effects on stroke patients’ psychological well-being.
Potential confounders related to PSD have been explored, including gender (female) [6
], disabilities, comorbidities [7
], stroke severity [8
], pre-stroke depression [9
], cognitive impairment after stroke, dysphagia [10
], incontinence [10
], anxiety, and social isolation at follow-up [10
]. In order to evaluate the preventive effect of acupuncture after subjects have been diagnosed as having had a stroke, a more rigorous study must be performed. In response, the main purpose of this study was to investigate whether acupuncture treatment after stroke attack reduces the risk of PSD after empirically controlling for covariates during the observation period.
2. Materials and Method
2.1. Source of Data
Our study used reimbursement claims data obtained from the National Health Insurance Research Dataset (NHIRD) in Taiwan. The NHIRD covers more than 99% of the population and has contracts with 97% of the hospitals and clinics in Taiwan [11
]. The National Health Research Institute maintains and updates the NHIRD. The institute has publicly released a sub-dataset composed of claims data for 1,000,000 randomly selected insurance enrollees for research and administrative purposes. This random subgroup represents approximately 5% of the entire insured population in Taiwan. This sub-dataset, consisting of a longitudinal health insurance database for 2005 (LHID2005), was employed for this study after obtaining approval from the National Health Research Institute review committee. For data analysis, we retrieved information about patients’ characteristics and medical care records by linking ambulatory care visit claims, in-patient expenditures by admissions, and the registry for beneficiaries. Secondary data were collected and administered by the Taiwan National Institute of Family Planning (now the Bureau of Health Promotion, BHP), and approved by their IRB. All data analyzed in this study were anonymized.
2.2. Participants and End-Point
Patients who were newly diagnosed with a stroke (ICD-9-CM codes: 430–434, 436–437) between 2000 and 2005 were included in this study. We excluded patients who had had a head injury before 2000 (N
= 340); who had suffered from depression before their stroke (N
= 1381); who were not insured or had died within 3 months of suffering the stroke (N
= 256); had sought ambulatory care for depression within the first 3 months following their stroke (N
= 58); or had not had an acupuncture interval for more than six months (N
= 98). The study criteria, the exclusion criteria, and the follow-up procedure are presented in Figure 1
In Taiwan, TCM doctors must finish a training course in Chinese medicine and acupuncture and pass the national examination before they are certified to practice and, thereafter, qualified for filing NHI claims for acupuncture reimbursement. We classified the stroke patients into two groups based upon their acupuncture use or non-use after their stroke: (1) acupuncture users: those who received six or more or 1–5 acupuncture treatments after being diagnosed with a stroke from 2000 to 2007 (respectively N = 1036, 12.21%; and, N = 1053, 12.41%); these were respectively called frequent and infrequent users; and (2) acupuncture non-users: those who were defined as reporting no acupuncture received after their stroke from 2000 to 2007 (N = 6398, 75.39%).
The Diagnostic and Statistical Manual (DSM) IV categorizes PSD as a “mood disorder due to a general medical condition (i.e., stroke)” with certain depressive features serving as specifiers, for example, major depressive-like episodes, manic features, or mixed features [12
]. The patients were linked to the ambulatory care visit claims and inpatient expenditures by admissions claims during the years 2000–2007 to identify possible treatment for depression. The follow-up ended on the date of depression diagnosis (diagnosed according to ICD-9-CM code 296, 309, or 311, or A-code A212 or A219) in outpatient care or on the date of censoring, which was either the date of withdrawal (including death) from the NHI program or the date of the follow-up end (i.e., 31 December 2007).
The covariates considered in our analysis include socio-economic factors and covariates related to stroke severity or progress. The socio-economic factors include gender, age, living area (categorized as “urban area”, “satellite city”, and “rural area”) [13
], and insurance premium (categorized as “<15,000 New Taiwan Dollars (NTD)”, and “≥15,000 NTD”).
The covariates related to stroke severity or progress include type of stroke (categorized as hemorrhagic, occlusion or others (e.g., transient ischemic attack)), length of hospital stay, rehabilitation after stroke in 3 months, comorbidities that are correlated to depression or disability, such as cancer (ICD-9-CM code: 140–208), arthritis or rheumatism (ICD-9-CM code: 714.0, 729.0), chronic obstructive pulmonary disease (ICD-9-CM code: 490–496), peripheral arterial disease (ICD-9-CM code: 440–449), diabetes (ICD-9-CM code: 250) chronic kidney disease (ICD-9-CM: 585) and ranking according to the stroke severity index (SSI). The SSI is an index that estimates a stroke’s severity by using six items listed in the hospitalization data in the NHI database [14
], namely airway suction, bacterial sensitivity test, general ward stay, intensive care unit stay, nasogastric intubation, osmotherapy, and urinary catheterization.
2.4. Statistical Analysis
A chi-square test was used to compare differences in age, sex, residential area, insurance premium, type of stroke, length of hospital stay, and the aforementioned comorbidities between groups. A Kaplan–Meier analysis was performed for censored graft survival. To assess the independent effects of acupuncture on the risk of depression, we conducted a Cox proportional hazard regression analysis with age, sex, insurance premium, urbanization level, SSI, rehabilitation after stroke in three months, and selected comorbidities adjusted simultaneously in the model. We also adjusted the urbanization level to account for the urban rural difference in accessibility to medical care in Taiwan. All statistical analyses were performed using SAS (version 9.4, SAS Institute, Inc., Cary, NC, USA), in which a p-value < 0.05 was considered statistically significant.
Our results indicate that using acupuncture after stroke diagnosis lowers the risk of PSD by an estimated 52.5% (1–0.475) in frequent acupuncture users and 28.2% (1–0.718) in infrequent acupuncture users. Applying acupuncture to treat stroke complications is common in Taiwan, and to our knowledge, our study is the first to investigate the effects of acupuncture in preventing PSD in a large population-based cohort. Within a 7-year observation period, the results indicate that recently-diagnosed stroke patients using acupuncture after being diagnosed were likely to have a lower risk of PSD. The overall risk reduction is estimated to be about 30.6%.
Strokes may increase vulnerability to the development of depression through a variety of neurobiological mechanisms. Three possible explanations for the association between physical illness and depression include a coincidental relationship; a negative mood reaction to the physical consequences of the stroke; and a neurotransmitter imbalance as a result of cerebral damage caused by the stroke [15
]. Based on these PSD mechanisms, our hypothesis is that using acupuncture after a stroke might improve stroke complications and result in the prevention of PSD. Studies suggest that acupuncture after a stroke may improve problems with pain [16
], spasticity [16
], physical functions [17
], quality of life [18
] and cognitive functions [19
]. Our findings seem to support the hypothesis that acupuncture reduces complications caused by a stroke and thus lowers the risk of PSD.
However, the short-term or long-term effects of using acupuncture to prevent PSD are unknown. To address this issue, a sensitivity analysis controlling for acupuncture over a three-month or six-month period after a stroke diagnosis was performed (table not shown). The results did not show that receiving acupuncture offered benefits in terms of depression symptoms. Nevertheless, the study did reveal that TCM-use benefited patients with higher depressive symptoms by attenuating their worsening [20
]. These results suggest that the preventive effect from a continuous acupuncture treatment period may vary widely after a stroke. Further randomized control trials between short-term or long-term follow-up times are needed for verification.
In addition, our study also explored potential confounders related to PSD. Females have a lower risk of depression than males, and the overall risk reduction is estimated to be about 22.1% in our study. Although these findings are similar to those found in [4
], other research has reported that female stroke patients have a higher risk of PSD. However, our findings that stroke patients with advanced age [21
], greater stroke severity [8
] and compounded comorbidities [7
] have a higher risk of PSD are consistent with much of the related prior literature. Be that as it may, a summed measure of comorbidities may ignore potentially important relationships between diseases and PSD. For example, our study revealed that stroke patients comorbid with myocardial infarction (HR = 1.301), cancer (HR = 1.414), diabetes (HR = 1.139), chronic kidney diseases (HR = 1.518), chronic obstructive pulmonary disease (HR = 1.218) or head traumatic injury (HR = 1.658) are at higher risk than their counterparts without these comorbidities. Furthermore, stroke patients with peripheral arterial disease or hypertension have a lower risk of depression after stroke diagnosis, while comorbid rheumatoid arthritis is not significantly correlated with PSD. Meanwhile, the benefit of 3-months rehabilitation was evaluated after stroke diagnosis in a previous study [4
], but those results do not correspond with ours. This might be due to the short-term and long-term ability of acupuncture or rehabilitation to prevent PSD being unknown. Accordingly, it is suggested that future research include these comorbid factors or the long-term effects of rehabilitation to further our understanding in this line of research. In addition, there was little evidence that compared the risk of PSD between occlusion-stroke and hemorrhagic stroke. Our study provided some information to fill this research gap. The results also indicated that patients, with occlusion-stroke or hemorrhagic stroke, who received frequent acupuncture treatments, may minimize the risk of PSD (HR = 0.446–0.499).
Our study has some strong points. One is that it is the first population-based study to investigate acupuncture’s potential in preventing PSD based upon the real clinical conditions of stroke patients. A second is that acupuncture is a relatively low-cost treatment and, when performed by a TCM doctor, has none of the potential side effects associated with taking medicines. A third is that our study is an observational study based on reimbursement claims data obtained from the National Health Insurance Research Dataset (NHIRD). Compared to studies using questionnaires to detect depressive symptoms, our study reduced the potential of being confounded by the Hawthorne Effect. A fourth strong point is that stroke severity was evaluated in our study [14
] to a greater degree than in other studies [4
Nevertheless, this study still has some limitations. First, acupoints are hard to define using the NHI database; moreover, different acupuncturists use different acupoints in different patients, depending on the symptoms and severity of stroke complications. However, some clinical studies have shown that even sham-acupuncture may have treatment effects in some diseases. Thus, there may be a positive effect even without considering the location of acupoints in different individuals in our study. Second, disability is an important predictor of PSD, but the severity of disabilities is hard to define using the NHI database. However, we controlled for stroke severity and comorbidities to minimize the effects of the severity of disability. Third, potential confounders, such as brain lesion of stroke [22
], cognitive impairment after stroke, dysphagia [10
], incontinence [10
], anxiety, and social isolation [10
], are not included in the NHI database, and so may confound our research results. Accordingly, future work is needed to further this understanding. Fourth, patients in the acupuncture group may have also received other complementary treatments such as massage, herbs, and aromatherapy, among others. Thus, we may overestimate the treatment effect of acupuncture. As such, the effects of the more specific acupuncture mechanisms or interactions deserves future investigation when data is available. Moreover, we conducted the exclusion criteria for depression within the first 3 months following their stroke for the purpose of minimizing survivorship bias. Further sensitivity analysis (table not shown) was conducted and was found to be similar to the present results, suggesting the possible role of acupuncture in preventing PSD development. Fifth, because the benefits of medicines and psychotherapy or pre-acupuncture depression screening in preventing PSD were not available in our dataset or relevant to acupuncture treatment, we did not control for these variables in our study. Sixth, we cannot know the total number of dropouts and the main reasons for dropping out because our study used NHIRD, and so the reasons that patients choose not to continue acupuncture were based on individual will. Accordingly, further research examining the choice to dropout or remain is therefore warranted.