Stroke in Traditional Korean Medicine: A Nine-Year Multicentre Community-Based Study in South Korea

In Korea, patients with stroke are commonly treated using traditional Korean medicine (TKM). The aim of this study was to provide information on the clinical characteristics of the pattern identification (PI) of stroke used in TKM. Stroke patients admitted to 15 TKM university hospitals from April 2005 through December 2013 were evaluated. The measured variables included the following factors as they related to the PI: (a) stroke etiology; (b) distribution of symptoms/signs; (c) physical characteristics and lifestyle parameters; (d) medical history; and (e) stroke-related laboratory results. Among 4912 stroke patients, 3466 patients received the same PI by two experts with the following distribution: Qi-Deficiency pattern (n = 810), Fire-Heat (FH) pattern (n = 1031), Dampness-Phlegm (DP) pattern (n = 1127), and Yin-Deficiency pattern (n = 498). Approximately 89.9% of subjects enrolled in this study had cerebral infarction. Some of specific symptoms were related to each type of PI, and obese phenotypes and blood lipids were significantly related to DP and FH. These results showed the characteristics of each type of PI and should lead to the standardization of diagnosis for stroke in TKM.

(n = 810, 23.37%), Fire-Heat (FH) pattern (n = 1031, 29.75%), Dampness-Phlegm (DP) pattern (n = 1127, 32.52%), and Yin-Deficiency (YD) pattern (n = 498, 14.37%). The different stroke types and NIHSS scores in relation to PI are presented in Table 1. Approximately 89.9% of subjects enrolled in this study had cerebral infarction (CI). In the TOAST classification of CI type, many subjects exhibited small vessel occlusion (SVO) rather than large artery atherosclerosis (LAA). The NIHSS score of the majority of patients was below 15. Among PI, the frequency of CI and SVO in the DP pattern was slightly higher than the YD pattern, but the NIHSS score among PI was not different.
Distribution of symptoms/signs in relation to PI. The Korean standard PI (K-SPI) of stroke consists of four PIs (QD, FH, DP, and YP pattern), and a total of 44 symptoms/signs for determining PI were reported in a previous study 6 . Variables are classified into four types: 19 FH variables, 7 DP variables, 11 QD variables, and 11 YD variables. Table 2 shows the distribution of the variables according to PI. In the FH pattern, headache-like flush, red tongue and strong pulse were the major symptoms. In the DP pattern, obesity (bi-sup) and a white "fur" on the tongue were common. Appearing powerless and lethargic and weak pulse symptoms were commonly observed in the QD pattern. Tidal fever, dry mouth and gauntness were more common in the YD pattern than in other PIs. Table 3 shows the physical characteristics and lifestyle patterns in relation to PIs. The BMI and WHR, which are related with obesity, were higher in the DP pattern than in other PIs. For lifestyle patterns, subjects that were currently smokers and drinkers were more likely to exhibit the FH pattern than the QD or YD pattern. Food preferences were also different according to PI. Medical history in relation to PI. The subjects' medical histories are presented in Table 4. With regard to past medical history, more than 58% of the patterns showed hypertension, which was the most common and triglycerides, were higher in the DP pattern compared to the other groups. Homocysteine and vitamin B12, which are indirect indicators of stroke, showed a different tendency among the pattern groups. The levels of homocysteine were higher in the FH group compared to the other groups, but the level of vitamin B12 was higher in the YD group.

Discussion
In the practice of TEAM, including TCM and TKM, a unique decision-making process called Bian Zheng Lun Zhi (PI or syndrome differentiation followed by treatment) is widely used 11 . This method, also called traditional Chinese medical diagnostics, is the procedure and practice of examining patients, determining diseases and differentiating syndromes/identifying patterns of signs and symptoms of diseases 4 . Through the comprehensive analysis of symptoms and signs, which has implications in determining the cause, nature and location of the illness and the patient's physical condition, their treatment is determined and confirmed 4,11 . According to TEAM theory, even patients with the same disease receive different treatments based on PI results. Although this  diagnostic system has many advantages in that it uses a comprehensive analysis of symptoms and signs to assess the cause of the diseases, there are many variations in the diagnostic process 7,12 . KIOM focuses on the importance of standardization in PI and has been engaged in relevant research 7 . The SOPI-Stroke project was conducted by KIOM from April 2005 to December 2013 to standardize and objectify PI for stroke through a scientific process 5,6,10 . In a previous study, our team described the four standard patterns, FH pattern, DP pattern, YD pattern, and QD pattern, in stroke and the forty-four indices used to determine the pattern 6 . This paper provides the characteristics, laboratory results and symptoms and signs of stroke patients who were treated according to TKM in the SOPI-Stroke project.
Among the symptoms/signs for PI diagnosis, some were common in each of the types of PI (Table 2). Over 80% of subjects with the FH pattern exhibited a reddened complexion, red tongue and a strong pulse. A white "fur" on the tongue and obesity (bi-sup) were major symptoms in the DP pattern, and many patients with the QD pattern appeared powerless and lethargic and had a weak pulse pattern. The YD pattern was associated with paleness, a red zygomatic site and dry mouth. These results were similar to other studies presented by our team and other TCM researchers 13,14 .
Many previous studies were performed to evaluate the relationship between Western indicators, such as body composition or blood parameters, and PI patterns [15][16][17][18][19] . Tables 3 and 5, respectively, show the distribution of body composition and blood parameters among PI. Among these, the level of obesity indices, including BMI and waist circumference, were significantly higher in the DP and FH patterns than in the QD and YD patterns (Table 3). Total cholesterol, triglycerides and total lipids, which are blood parameters positively related to obesity, were also higher in the DP and FH (Table 5). These results are similar to those previously reported by Min et al. 15 , and other studies showed that DP was significantly related to the obese phenotype [16][17][18] .
Genetic studies showed that many of the gene polymorphisms were related to obesity and obese phenotypes [20][21][22][23] . Similar studies that show the association between genetic polymorphisms and PI were performed in TKM and TCM 17,[24][25][26] . Specifically, the polymorphisms in UCP2 and UCP3, which were significantly associated with BMI and serum cholesterol in Korean female subjects, were also associated with DP pattern 24,27,28 . Some studies showed controversial results. For example, Kim et al. reported that subjects with the C allele of − 607 G > C in wnt10b had lower BMI levels than subject with the GG type 29 , but Ko et al. showed that the − 607 G > C polymorphism was related to the YD pattern, not DP pattern 25 . Those results suggested that the PI, a diagnostic system used in TEAM, exhibited similar or different phenotypes compared to the phenotypes discussed in Western medicine.
There are some limitations to our study. To obtain the characteristics of stroke patients treated with TKM, observational studies are very important. This is a potentially interesting observational study because it provided evidence for clinical practice. This paper describes the baseline characteristics of stroke according to PI but there is no follow-up data describing long-term outcomes. Therefore, it is necessary to conduct improved randomized controlled trial studies or cohort studies to obtain reliable results on PI phenotypes and the effect of treatment based on PI compared with Western medicine. Second, the population enrolled in this study focused on the CI type and SVO type according to the TOAST classification. Additionally, the severity of stroke in the enrolled patients was lower than that of stroke patients visiting Western medical hospitals. For this reason, it is difficult to generalize the characteristics of PI. TKM uses four methods of diagnosis, which include diagnosis by observation,  hearing and smelling, interrogation, and palpation. Diagnosis depends on the clinician's experience and knowledge, along with a variety of environmental factors. It is essential to establish an objective diagnostic standard for tongue and pulse measurements, along with other reliable diagnostic tools. Although this study had some limitations, the significance of our study was to show the characteristics of the patients receiving TKM treatment. In TKM treatment, PI comprises a series of processes that involve not only identifying specific neurological symptoms but also unspecific symptoms and indicators with four examinations, as well as determining treatment goals after integrating all the data for stroke diagnosis. These results should to lead to the standardization of PI for stroke in TKM.

Subjects.
This study was a community-based multicentre trial that was part of the SOPI-Stroke project 5,6 .
Stroke patients who were admitted to 15 TKM university hospitals participated in this study from 2005 through 2013 (The entire list of hospitals can be found as Supplementary Table S1). Each patient provided written informed consent to undergo procedures that were approved by the respective institutions' Institutional Review Boards (IRB). This study was conducted in accordance with approved guidelines by the IRB of the KIOM and by each TKM university hospital's IRB. All patients provided informed consent after a thorough explanation of the details. Figure 1 shows the PI distribution among stroke patients by gender according to region.
Inclusion/exclusion criteria. We enrolled stroke patients within 30 days of the onset of their symptoms if their diagnosis was confirmed by an imaging diagnosis such as computerized tomography (CT) or magnetic resonance imaging (MRI) 5,6 . Patients with traumatic stroke such as subarachnoid, subdural, and epidural haemorrhage were excluded from the study. This study was approved by the IRB of the KIOM and by each TKM university hospital's IRB.
Measured variables. Each patient was seen by two experts in the same department within each site. All experts were well trained in standard operation procedures (SOPs). The experts had at least three years of clinical experience with stroke after finishing a regular college education of TKM for six years. The examination parameters were extracted from parts of a case report form (CRF) for the standardization of stroke diagnosis developed by an expert committee organized by the KIOM. The measured variables used included the following key subjects in relation to PI: (a) stroke etiology; (b) distribution of signs/symptoms; (c) physical characteristics and lifestyle parameters; (d) medical history; and (e) laboratory results. Specifically, as suggested by the KIOM, the clinicians were required to measure the stroke PI of each patient according to the FH pattern, DP pattern, QD pattern, or YD pattern.
Statistics. Data were statistically analysed with SAS software, version 9.1.3 (SAS Institute Inc., Cary, NC, USA).
Categorical variables were compared with the chi-square test or Fisher's exact test, and differences in continuous variables were determined by one-way analysis of variance (ANOVA). Statistical significance was set at P < 0.05.