Four Subgroups of Blood Stasis Syndrome Are Identified by Manifestation Cluster Analysis in Males

Blood stasis syndrome (BSS) is an important pathological condition in traditional East Asian medicine and is associated with ischemic heart disease, cerebral vascular accident, diabetes mellitus, chronic renal failure, severe traumatic injury, and dysmenorrhea. However, previous studies have been unable to reveal the clinical and biological characteristics or biological markers of BSS. We hypothesized that the heterogeneity among the manifestations of BSS or non-BSS could interfere with an analysis to describe the characteristics of BSS. In this study, male participants based on the severity of BSS-associated symptoms and signs were clustered and classified into four subgroups: BSS subgroups (1), (2), (3), and (4). Non-BSS core subgroup was redefined using manifestation cluster analysis. Biological characteristics of subgroups BSS(1) and BSS(2) belong to the range of the non-BSS core subgroup (1), whereas that of subgroups BSS(3) and BSS(4) are characterized by different biological parameters such as systemic inflammatory conditions and elevated D-dimer level. Our results suggested that patients in subgroups of BSS(3) and BSS(4) are more likely to be exposed in an inflammatory state than other BSS subgroups. We found the heterogeneity among the manifestations which could mask the characteristics of BSS and identified the clinical and biological profiles of the four BSS subgroups through comparisons of the redefined non-BSS and BSS subgroups. This finding could provide accurate diagnostic criteria and new approaches for BSS treatments in different subgroups.


Introduction
Blood stasis is a pathological concept in traditional East Asian medicine that refers to stagnant blood that has lost its physiological function within the body [1][2][3] and leads to blood stasis syndrome (BSS), which is characterized by multiple signs and symptoms, such as sublingual varicosis, angiotelectasis, a slow and choppy pulse, local fixed pain, nyctalgia, menstrual cramps, a dark-purple tongue, or infraorbital darkness [2,4]. Clinical studies have reported that these manifestations are observed in patients with ischemic heart disease, cerebral vascular accident, diabetes mellitus, chronic renal failure, severe traumatic injury, and dysmenorrhea [3,5]. In addition, traditional herbal formulas for BSS were effective in relieving the severity of these diseases [6][7][8][9][10]. However, previous studies could not determine the clinical and biological characteristics or biological markers of BSS [5,6,[10][11][12].
In psychoneurological filed, many studies have clustered pattern of symptoms with a psychoneurological symptom cluster intensity score because they have showed high heterogeneity which lead diagnosis and therapeutic failures [13,14]. Clustering analysis is a method to define subgroups of individuals with high heterogeneity to explore clinical phenotypes in patients with various diseases [15]. Classification of disease into subtypes which have different clinical signs in terms of prognosis and individual differences might be needed to explain between clinical phenotype and biological mechanisms [16].
To overcome the limitations of previous studies, we diagnosed and classified into subgroups of non-BSS and BSS participants based on BSS-associated manifestation cluster Each bar represents the mean ± standard deviation of the average total score of each BSS manifestation. The total score is calculated by subtracting 31 from the sum of each original variable score. The P-value was calculated using an independent t-test.
analysis. We found that the heterogeneity among the manifestations of BSS within individuals could be considered to be an independent parameter for determining the characteristics of BSS. This approach identified subgroup-specific clinical characteristics and could lead to other studies on the biological markers of BSS. We utilized the data of 476 participants with the same diagnosis from two KMDs to guarantee the reliability of the diagnoses. However, we could not enroll enough female participants to perform manifestation cluster analysis with meaningful results. Female data could not be combined with male data because of the differences in clinically important symptoms, such as menstrual cramps and lumps in menstrual blood. Therefore, we excluded female participants from this analysis. The data from 219 of male participants were enrolled for a cluster analysis (supplementary Figure 1). The eligibility criteria were as follows: males aged between 20 and 70 years who gave their written informed consent to participate and agreed to comply with the study regulations. The exclusion criteria were patients with any psychiatric condition that rendered them unable to communicate, critically ill patients, pregnant women, or patients with any conditions that could influence the study assessment.

Assessment of Clinical Symptoms and Signs.
The KMD was trained twice with the standard operating procedures to estimate the severity of 31 BSS manifestations including discoloration within the body, local pain and tenderness, and disorder of blood circulation for accurately diagnosis. Thirtyone indicators for BSS were derived through 3 meetings by Evidence-Based Complementary and Alternative Medicine 3 The data are expressed as the mean ± standard deviation or a number with a frequency. BMI: body mass index; SBP: systolic blood pressure; DBP: diastolic blood pressure. The P-value was calculated using an ANOVA or the Chi-square test.
an expert committee in three countries: Korea, China, and Japan [3,18,19]. And two KMDs independently scored the participants using the "case report form for the diagnostic technology of blood stasis questionnaire-1". Scores were given according to the following scale: 1 = none, 2 = slight, 3 = moderate, 4 = severe, and 5 = very severe. We used the average scores of individual variables for the analysis.

Analysis of the Biological Parameters.
Blood was collected from each participant at the hospitals and transported to the Samkwang Medical Laboratory (Seoul, Korea) for analysis.

Statistical Analysis.
All statistical analyses were performed using the Statistical Analysis System (SAS version 9.1.3, SAS Institute Inc., Cary, NC, USA). All P-values were two-sided, and P < 0.01 was considered statistically significant. Continuous variables were expressed using the mean ± standard deviation (SD), and categorical variables, such as smoking, drinking, and medication usage, were described by a number (percentage; Table 1). The significance of the differences in the general characteristics between groups was calculated using one-way analysis of variance (ANOVA) or the Chi-square test. To identify subgroups of participants based on their BSS manifestations, we carried out Ward's Minimum Variance Cluster analysis with Eigenvalue and pseudo T-squared statistics and determined the number of subgroups as well as used hierarchical clustering with squared Euclidean distances [20,21]. Differences in the manifestations and 34 biological parameters between the redefined non-BSS and BSS subgroups were examined by the independent t-test or Wilcoxon Rank Sum test (Table 2).

Manifestation Heterogeneity of the BSS and Non-BSS
Participants. We clustered male participants based on the severity of their BSS-associated symptoms and signs to reveal the manifestation heterogeneity among the BSS and non-BSS groups. A diagnosis with non-BSS included three manifestation clusters, and participants with BSS were divided into four subgroups (Figure 1(a)), showing that the manifestation heterogeneity of diagnosis could mask the clinical and biological profile of BSS. BSS and non-BSS participants classified in the same cluster could have substantial similarities in their biological characteristics.

The Redefined Non-BSS Core and Four Subgroups of BSS.
To overcome the limitations of previous studies that did not consider the masking effect of manifestation heterogeneity, we redefined the cluster 1 group with non-BSS as a core of non-BSS, which had 73.1% (90/123) of participants with non-BSS (Figure 1(a)) and the lowest scores for BSS manifestations (Figure 1(b), p < 0.001). Participants with BSS were divided into four subgroups. The non-BSS core and subgroups of BSS are represented by diagnosis with clustering numbers as BSS (1) to BSS (4), and the general characteristics of the groups are presented in Table 1. The non-BSS(1) group had a lower proportion of participants who smoke, but age, BMI, SBP, DBP, drinking status, and medication were similar among the five groups (Table 1).

Differences in the Manifestations between the Redefined
Non-BSS and BSS Subgroups. We identified BSS subgroupspecific clinical symptoms and signs through a comparison of BSS manifestation severities between non-BSS(1) and BSS subgroups. Subgroup BSS(1) had a tendency to bruise easily, and subgroup BSS (2)

The Differences in the Biological Parameters between the Redefined Non-BSS and BSS Subgroups.
Biological characterization of the BSS subgroups was performed through a comparison of the biological parameters between the non-BSS(1) and BSS subgroups. Subgroups BSS(1) and BSS (2) were not different from non-BSS(1) relative to our biological

Discussion
To identify the significant clinical and biological characteristics, data of enrolled participants were analyzed using Ward's Minimum Variance Cluster analysis with Eigenvalue and  Local sharp pain   31  Nocturnal pain   2  5  4  1  3  2  5  4  1  3  2  5  4  1  3  2  5  4  1  3  2  5  4  1 3 Figure 2: Differences in manifestations between the redefined non-BSS and BSS subgroups. Each bar represents the mean ± standard deviation of the average BSS manifestation score. A filled box represents a mean score ≥ 3 and a P-value < 0.01 compared to the non-BSS(1) group.
CCC. Firstly, male BSS participants were classified into 2, 4, or 6 clusters, and clinically relevant phenotypes have been identified in 4 clusters with Eigenvalue ≥ 1, which indicate positive definite to further analysis. As shown in Figure 1, we found four subgroups of BSS and non-BSS core group based on the severity of their BSS-associated manifestations using cluster analysis. The almost BSS manifestation severity of subgroup BSS(1) belongs to the range of the non-BSS(1) group, and subgroup BSS(2) exhibits abdominal tenderness and chronic pain ( Figure 2). However, these subgroups are not biologically different from the non-BSS(1) group (Table 2). Therefore, subgroups BSS(1) and BSS(2) should be excluded from analysis to identify biological characteristics and markers of BSS. Subgroups BSS(3) and BSS(4) are commonly characterized by a rough pulse, discoloration of the skin, chronic pain, and increased D-dimer level. The BSS(3) subgroup is distinguished from the other subgroups by traumatic pain, decreased RBCs and albumin level, and increased CRP level. Further clinical studies on subgroups BSS(3) and BSS(4) are necessary to evaluate the effectiveness of BSS-specific treatments based on these biological parameters.
The symptoms and signs of BSS is correlated with pathologic properties of senescent RBCs [1]. Accelerated RBC senescence could induce low levels of hemoglobin and hematocrit in subgroups of BSS (3). Also, the accumulation of senescent RBCs causes thrombosis and blood clotting. The D-dimer level is elevated in patients with blood clotting disorders, and it increased the risk for thrombosis [22,23]. Neutrophils and monocytes play a role in thrombus formation, and it contributes thrombo-inflammatory conditions [24]. Systematic inflammatory condition may elevate the level of CRP which is shown in subgroups of BSS(3) [25]. Therefore, according to our data, patients in subgroups of BSS(3) and BSS(4) might be more susceptible to develop severe diseases than those of BSS(1) and BSS (2). However, the prospective cohort study needs to prove prognosis of diseases depending on BSS subgroups.
When comparing two noncore subgroups of non-BSS(non-BSS(2), non-BSS(4)) with two subgroups of BSS (2) and BSS(4), biological characteristics were not different, except for WBC count in the comparison between non-BSS(4) and BSS(4) subgroups (Supplementary Table 1). It may mask the biological profiles of patients with BSS. Identification of clinical and biological profiles in a BSS needs to be considered the heterogeneity of manifestations in further studies.
Participants in four BSS subgroups smoked more cigarettes than non-BSS core group as shown in Table 1 (p < 0.05). People who smoke have showed increased erythrocyte sedimentation rate (ESR) which is a certain biomarker of inflammation [26]. Elevated ESR level has correlation with raised D-dimer level in patients with thrombosis and hemostasis [27]. Smoking habit may be a potential risk marker for blood stasis.
Although intense clinical studies are still necessary to identify the biological characteristics and markers of BSS, this new approach was able to verify the heterogeneity among the manifestations of a BSS diagnosis that could mask the clinical and biological profiles of BSS and identify four subgroups of BSS in males.

Conclusions
We verified the heterogeneity among the manifestations of a BSS diagnosis that could mask the clinical and biological profiles of BSS and identify four subgroups of BSS in males. This new approach could lead to other studies on the pathologic mechanism and biomarker of BSS.