Effect of Herbal Prescriptions in Accordance with Pattern Identification in Acute Cerebral Infarction Patients: Based on Fire-Heat Pattern

Objectives. This study was conducted to verify the necessity of corresponding prescription to the diagnosed pattern in acute cerebral infarction patients. Methods. We studied cerebral infarction patients hospitalized within 30 days after the ictus. Forty-four clinical indicators, Motricity Index (MI) score, Scandinavian Stroke Scale (SSS) score, and herbal prescriptions were checked twice, two weeks apart. The probability of each pattern was calculated based on the clinical indicators. Changes in MI score, SSS score, and the probability of fire-heat pattern were compared between the pattern-prescription correspondence group and the noncorrespondence group. Results. Increments of MI score and SSS score in the correspondence group were significantly greater than those of the noncorrespondence group (p = 0.003, p = 0.001) while the baseline score of the two groups showed no significant difference. Probability of fire-heat pattern decreased significantly in the correspondence group (p = 0.013) while the noncorrespondence group showed no significant difference after the treatment. Conclusion. Acute cerebral infarction patients who are diagnosed as fire-heat pattern showed better improvement in dysfunctions caused by the disease when they took the pattern corresponding prescriptions. This study provides evidence for the necessity and usefulness of pattern identification in Traditional Korean Medicine.


Introduction
In South Korea, the political and social status of Traditional Korean Medicine (TKM) almost equals that of Western medicine. This is due to a strong preference for TKM in Korean people. When 1,000 people living in Seoul were asked to choose between Western medicine doctors and TKM doctors if they had developed a stroke, 25% of the subjects chose TKM doctors, and 45% of the subjects responded that they were willing to receive TKM treatments [1]. Also, a survey demonstrated that 40% of stroke patients tried TKM treatments after they were discharged from a Western medicine hospital [2]. Among many diseases, TKM has been favored especially for the treatment of stroke.
Considering the large role in the treatment of stroke, significant amount of effort was put into accumulating scientific evidence of the efficacy and safety of TKM. As a result, questionnaire for the pattern identification and guidelines for assessing the clinical indicators were developed over the years [3,4]. These guidelines allowed us to accumulate coherent data, and by using these data, we were able to work out equations for standardization of pattern identification for stroke patients [5,6]. These works enhanced the objectivity and reproducibility of pattern identification and remedied its shortcomings. However, the necessity and usefulness of pattern identification have not been verified yet. Many herbal prescriptions have been proven to have a beneficial effect in acute stroke patients [7][8][9], but the relation between the efficacy and the pattern identification was not indicated.
In this study, to verify the necessity and usefulness of pattern identification, we compared the two groups of the patients diagnosed as fire-heat pattern. The group who took the herbal prescriptions in accordance with the pattern identification and the group of people who did not take the herbal prescriptions accordingly were compared to demonstrate if the corresponding prescription taking group shows better outcome. Also, we performed a correlation study between the changes in clinical indicators and the improvement in   [3,10] to reduce inconsistency in pattern identification carried out by different TKM doctors. Table 1 shows the forty-four clinical indicators contained in the CRF, and there were four possible patterns to choose from, which were fire-heat pattern, Yin Deficiency Pattern, Phlegm Dampness Pattern, and Qi Deficiency Pattern. The patients whose patterns were not decided because the opinions of the two TKM doctors differed were dropped out. After the pattern identification, the patients were allocated into each group according to their confirmed pattern.

Materials and Methods
All subjects were studied twice, 2 weeks apart. During the 2-week period, all participants received conventional Western medicine treatment such as antiplatelet agent, risk factor control (e.g., hypertension, diabetes mellitus, dyslipidemia, and cardiac disease), and rehabilitation exercise. TKM treatment was also administered to all of the patients, which includes herbal prescription, acupuncture, and electroacupuncture. The contents of acupuncture and electroacupuncture treatment are shown in Table 2. The herbal prescriptions applied to each patient were selected according to the patient's condition and associated symptoms by the TKM doctors who were irrelevant to the present study. The prescriptions used during the treatment period were checked, and we classified the prescriptions based on the guideline suggested in the SOP (Table 3).

Measurements.
Baseline characteristics such as age, sex, Body Mass Index (BMI), period from onset to admission, medical history, alcohol and smoking habits, and Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification [11,12] of the stroke types were checked. To estimate the motor function, we used Motricity Index (MI) score [13], which is a reliable scale in assessing motor impairment after stroke. Scandinavian Stroke Scale (SSS) score [14] was used to evaluate the degree of dysfunctions in the subjects. The assessors for the MI and SSS scores did not have the information about the herbal prescriptions the patients are taking.
To assess the changes in the clinical indicators, we used the logistic equations for calculating the probability of each pattern suggested by Kim et al. [5]. The same CRF and SOP used in the present study were used in their research, and the logistic equations were derived based on the clinical data of 480 stroke patients as a result of regression analysis. The equations for the probability of four patterns are as follows: Put in "1" for the existing clinical indicators and "0" for the nonexisting clinical indicators.

Probability of Fire-Heat Pattern. Consider
Probability of Yin Deficiency Pattern. Consider

Probability of Phlegm Dampness Pattern. Consider
Probability of Qi Deficiency Pattern. Consider Patients that display more fire-heat pattern related symptoms show higher probability of fire-heat pattern. The discriminant validity of the equations for the probability of the four patterns is shown in Table 4.

Statistical
Analysis. Statistical analysis was performed by using the Statistical Package for the Social Sciences version 12.0 for Windows (SPSS, Chicago, IL). Chi-square test was used for the categorical variables, and Mann-Whitney test was used for the continuous variables when comparing the two groups. Wilcoxon signed rank test was used for statistical comparisons between the values before and after the treatment. We correlated the changes in the probability of fire-heat pattern with the changes of SSS score and MI score, respectively, using Spearman's rank correlation. A < 0.05 was considered significant.

Results
Of the 300 patients enrolled in the study, 68 patients were discharged before the second checkup, and 44 patients with perfect MI and SSS score were excluded as they could not expect further improvement. 40 patients were unable to determine the pattern because the diagnosis of the two TKM   doctors differed. 11 patients were dropped out due to missing data. Also, many patients diagnosed as Yin Deficiency Pattern, Phlegm Dampness Pattern, and Qi Deficiency Pattern received different types of herbal prescriptions during the treatment period. Only two patients in the Yin Deficiency Pattern, three patients in the Phlegm Dampness Pattern, and two patients in the Qi Deficiency Pattern received pattern corresponding herbal prescription for the whole 2-week period, so we were unable to secure a sufficient sample size for statistical analysis for those three patterns. Among the remaining 57 patients who were diagnosed as fire-heat pattern, we considered 40 patients who received herbal prescriptions targeting fire-heat pattern related symptoms into correspondence group and the other 17 patients who received herbal prescriptions focusing on clinical indicators of other patterns into noncorrespondence group (Figure 1). After 2-week period of treatment, no aggravation of the neurologic deficit was observed in the patients.

Baseline Assessment.
General characteristics, period from onset to admission, medical history, alcohol and smoking experience, and proportion of ischemic stroke type according to TOAST classification showed no significant difference between the two groups ( Table 5).

MI Score and the SSS Score before and after the Treatment.
Both groups showed increase in the MI score and the SSS score, but the increments of the MI score and the SSS score in the correspondence group were significantly greater than those of the noncorrespondence group ( = 0.003, = 0.001) while the baseline scores of the two groups showed no significant difference (Table 6).

Changes in the Probability of Fire-Heat Pattern.
The probability of fire-heat pattern was significantly higher than the probability of other patterns in both groups ( < 0.0001). The baseline probability of fire-heat pattern between the two groups showed no significant difference. The probability of fire-heat pattern decreased significantly after the treatment in the correspondence group ( = 0.013) while the probability of fire-heat pattern in noncorrespondence group showed no significant change. Probability of other patterns showed no significant change after the treatment in both groups (Table 7).

Correlation Analysis in the Fire-Heat Pattern Corresponding Prescription Group.
In the correlation study, the decrease in the probability of fire-heat pattern showed significant correlation with the increase in the SSS score ( = 0.027) and missed statistical significance with the increase in the MI score ( = 0.058) (Figures 2 and 3).

Discussion
The aim of this study was to verify the usefulness of the pattern identification. To achieve this goal, we compared the outcome of the treatments in pattern-prescription correspondence group and the noncorrespondence group. While the baseline scores did not differ significantly between the two groups, increments of MI score and SSS score after the treatment were significantly higher in the correspondence group than the noncorrespondence group ( = 0.003, = 0.001). This suggests that taking herbal prescriptions in accordance with the diagnosed pattern is more effective in improving functional impairments of acute ischemic stroke patients diagnosed as fire-heat pattern.
A type of herbal prescription is selected based on the clinical indicators a patient is showing, and when used, the herbal prescriptions are expected to alleviate the clinical symptoms. We used the probability of fire-heat pattern as a scale to evaluate the changes in the clinical symptoms of patients diagnosed as fire-heat pattern. As expected, the probability of fire-heat pattern decreased significantly in the correspondence group ( = 0.013) while there were no significant changes in the noncorrespondence group.
Also, correlation study indicates that the patients with larger increment in the SSS score showed larger decrement in the probability of fire-heat pattern ( = 0.027). In our previous study concerning the treatment of acute ischemic stroke patients, motor function recovery in the patients correlated significantly with the improvement in the symptoms related to fire-heat pattern [15], which is consistent with the results of the present study. These results suggest that patients with improved functional impairments tend to show alleviation of clinical symptoms related to fire-heat pattern.
Usage of herbal prescriptions on acute ischemic stroke patients has been studied over the years [7][8][9], but no research was carried out to verify the necessity and usefulness of the pattern identification. Pattern identification is a meaningful diagnostic tool of TKM as it allows individualized treatment, maximizing its effectiveness and minimizing its adverse effects. We tried comparing the pattern-prescription corresponding group and noncorresponding group in our study in 2011, but the sample size was too small and the treatment period was too short, and the results showed no statistical significance [16]. This is the first study to evaluate the effectiveness of pattern identification in acute ischemic stroke patients diagnosed as fire-heat pattern. We could not verify the effectiveness of pattern identification in patients diagnosed as Yin Deficiency Pattern, Phlegm Dampness Pattern, and Qi Deficiency Pattern, because they were not consistent in consuming the herbal prescriptions corresponding with their pattern and therefore were not suitable for the subjects of this study. This was not expected when we designed the study, but due to this outcome, the application of the results in the present study should be limited only in the patients diagnosed as fire-heat pattern. Another limitation is that the probability of pattern does not properly evaluate the severity of the clinical symptoms related to each pattern since the scale was made to determine the pattern not to assess the clinical indicators. We used the probability of patterns in this study because this scale was created based on the data collected using the same CRF and SOP used in the present study. There is no widely accepted scale assessing the severity of the symptoms related to pattern identification, and it should be developed in the future for more researches concerning pattern identification.
In the present study, correspondence group displayed better outcome than the noncorrespondence group, and in the correspondence group, patients with lesser clinical indicators related to fire-heat pattern after the treatment showed better improvement in the recovery of functional impairment. These results imply that herbal prescriptions in accordance with the diagnosed pattern alleviate the clinical symptoms in relation with the diagnosed pattern and are more effective in restoring the dysfunctions caused by the disease than the noncorresponding prescriptions. The results provide evidence for the necessity and usefulness of pattern identification in TKM, but further research is needed to confirm the effectiveness of pattern identification for the other pattern groups.