Integrative Chinese and Western Medicine for patients after Percutaneous Coronary Intervention (ICWM-PCI): A Prospective Observational Real-World Cohort Study

Background Integrative Chinese and Western Medicine (ICWM) is widely used in coronary artery disease (CAD) patients after percutaneous coronary intervention (PCI) in China. However, the evidence-based on the long-term prognosis and large sample on this topic are weak. The purpose of this study is to evaluate the correlation between the therapeutic effect of ICWM and the prognosis of patients after PCI. Methods This study is a prospective observational real-world cohort study that was conducted from September 2016 to August 2019 in Fuwai Hospital. The study was reviewed and approved by the Ethics Review Committee of Fuwai Hospital, Chinese Academy of Medical Sciences. We consecutively screened 6000 patients after PCI and they were followed up for 2 years. ICWM were related to prognostic outcomes using unadjusted (Kaplan-Meier curves) and risk-adjusted (multivariable Cox regression) analyses. The primary endpoint was a composite of all-cause death, revascularization, and myocardial infarction. Results A total of 5942 patients after PCI were enrolled in this study, 5453 patients were included in the nal analysis (4189[76.8%] were male; mean [SD] age, 61.91[9.91] years). There were 2932 patients (53.8%) in western medicine group (WMG) and 2521 patients (46.2%) in integrated medicine group (IMG). Cox regression analysis showed that IMG had a 27% lower cumulative risk of the major adverse cardiovascular event (MACE) than WMG (hazard ratio [HR], 0.73; 95% CI, 0.63-0.85; P(cid:0)0.0001), especially in all-cause mortality and revascularization. Conclusions Among patients after PCI, ICWM compared with conventional western medicine was correlated with a lower risk of 2-year MACE. Further research is needed to provide higher levels of evidence. the function of rats after the Akt pathway 26 . Tongxinluo can alleviate myocardial ischemia/reperfusion injury by activating cardiac microvascular endothelial cells endothelial nitric oxide synthase 27 . Compound danshen dripping pills ameliorated myocardial ischemia, reversed the reprogramming of the metabolism induced by ISO and normalized the level of most myocardial substrates and the genes/enzymes associated with those metabolic changes 28


Background
Coronary artery disease (CAD) is a major public challenge and the leading cause of morbidity and mortality worldwide 1 . Percutaneous coronary intervention (PCI) is the most common strategy for treating CAD. In 2018, 915256 patients underwent PCI in the Chinese mainland 2 , which increased 21.5% year over year in 2017 3 . With the development of interventional techniques, medical apparatus and instruments, the survival risk of patients with CAD is greatly reduced, clinicians and researchers begin to pay more attention to the postoperative cardiac rehabilitation of patients. PCI can achieve patients' coronary artery reperfusion, but PCI can lead to vascular endothelial injury as a kind of invasive operation, it cannot change the etiology and pathogenesis of CAD. Even if the patients after PCI take conventional western drugs (such as aspirin, clopidogrel, statins, etc.), there are still about 5-15% risk of major adverse cardiovascular event (MACE) within one year after PCI [4][5][6] . In addition, conventional western drugs cannot be used to some patients due to allergies or resistance 7 . Therefore, the secondary prophylactic drug strategy after PCI needs to be improved.
Many CAD patients after PCI are treated with traditional Chinese medicine (TCM) based on conventional western medicine in China, and previous evidence-based studies on integrated Chinese and western medicine (ICWM) for CAD showed that it can reduce recurrence angina frequency 8 and the occurrence of adverse cardiovascular events [9][10][11] , alleviate the clinical symptoms 12 and the quality of life of patients 13 , ameliorating hemorheology and blood lipid parameters, in ammatory mediators 14 . But the sample size of these studies is small, and the follow-up time is short. On the other extreme, meta-analyses on the results have been conducted in [14][15][16] . Unfortunately, due to the inconsistency in the context and circumstances of these studies, further insight and consensus are to reach. Thus, the evidence for the role of TCM in long-term curative effects after PCI is not su cient. Additionally, the evidence of real-world study about integrated Chinese and western medicine treatment for CAD is insu cient. Therefore, we designed this large sample, long-term follow up, real-world study with the following objectives: (1) to observe the treatment effects of Chinese patients after PCI in the real-world, (2) to identify the incidence and trends in MACE after PCI, (3) to compare the risk of MACE within 2-year after PCI among western medicine group (WMG) and integrative medicine group (IMG).

Study Design and Patients
This study was a prospective observational real-world cohort study. During the period from enrollment to the end of follow-up, no intervention was carried out for patients. We observed patients' medication strategies and outcomes under real medical conditions in China. Follow-up was conducted on 3th, 6th, 12th, 18th, and 24th months after PCI (with a window of ±14 days). The study was reviewed and approved by the Ethics Review Committee of Fuwai Hospital, Chinese Academy of Medical Sciences.
Patients diagnosed with CAD after PCI within a week in Fuwai hospital Chinese Academy of Medical Sciences from September 2016 to August 2017 were included. Only stents were included, not balloon dilatation. Exclusion criteria include (1) age≤18 years old; (2) cognitive or communication impairment; (3) Participating in other clinical studies. All participants gave written informed consent.

Groups and Medicine Strategies
Patients were divided into WMG and IMG according to their medicine strategies during the 2-year followup. If the patients after PCI used western medicine in combination with TCM for more than half a year, they were divided into IMG; Otherwise, the patients were classi ed into WMG. Because the cohort distribution of patients was not determined before the end of the last follow-up, data on the lost patients were directly deleted.
Conventional western medicine refers to the routine medication for CAD patients after PCI, including aspirin, clopidogrel, statins, β-blocker, ACEI/ARB, etc. The speci c strategy was determined by the clinician. TCM refers to the drug treatment based on syndrome differentiation, including decoction, powder and Chinese patent medicine. Decoction and powder are composed by many kinds of herbal drugs, such as Ren Shen, San Qi, Angelica Sinensis, et al., the speci c drug types and doses were decided by clinicians according to the patients' TCM syndrome classi cation. Chinese patent medicines are certain forms according to the prescribed prescription and preparation technology, which were commercial TCM preparation approved by China Food and Drug Administration.

Outcome Measures
The primary endpoint was MACE, including all-cause mortality, myocardial infarction (MI), and revascularization, which was a composite of PCI, PTCA and CABG. The secondary endpoints were angina symptom score (ASS) and traditional Chinese medicine syndrome quanti cation score (TCMSS). The safety were evaluated by cardiovascular hospitalization and stroke. An independent endpoint committee adjudicated all potential events. All endpoint events of patients were evaluated by an independent endpoint committee.

Sample size calculation
In this study, it is expected that the data of 6,000 patients who were successively enrolled in Fuwai Hospital after PCI will be accumulated during the rst year. According to previous experience, the proportion of patients treated with ICWM can reach 1/3 level in Fuwai Hospital. Combined with the literature 17 and clinical data of our hospital, the incidence of primary endpoint events 2 years after PCI was about 15-20%. According to this incidence level, the number of primary endpoints that can be accumulated by the end of the study is expected to 1200. As an event-driven study, based on a bilateral signi cant level of 2.5% and an 80% degree of assurance, a statistically signi cant result can be obtained when a relative reduction of 20% (HR=0.8) or more in the risk of the primary endpoint in IMG compared to WMG with the above sample size.

Statistical Analysis
Continuous variables are expressed in the form of mean and standard deviation (SD), categorical variables are described in the form of frequency and percentage. According to the distribution characteristics of data, the corresponding parameter or non-parameter statistical method is selected.
Baseline characteristics were compared between two groups using T-tests or χ 2 tests. The primary endpoint and the safety endpoints were analyzed by the survival analysis. In addition to describing the occurrence of events in each group with the Kaplan Meier curve, Cox proportional hazard model was used to estimate the hazard ratio (HR) and its 95% con dence interval (CI). To further explore the prognostic differences between the two groups, we selected some baseline characteristics as covariables for multivariate Cox regression analysis based on previous studies and clinical signi cance. The secondary endpoints were analyzed by T-tests. The signi cance level of the statistical test was 5% on two sides, and the statistical analysis software was intended to use SAS 9.4.

Patient and Public Involvement statement
During the initial enrollment phase, the patients were involved in this study as the subjects and signed informed consent. They were followed-up by telephone 2 years after enrollment. The patients/the public were not involved in the design or plan of this study, and they were not asked to assess the burden of the intervention because this is an observational study.

Study Patients and Baseline Characteristics
We screened 6,000 patients after PCI, 5942 of whom were included in this study. 5471 (Figure 2A). Based on previous studies and clinical signi cance, we selected some baseline characteristics as covariables for Cox regression analysis (age, gender, the course of CAD, history of MI, ACS, rst PCI, overweight or obesity, current smoker, hypertension, hyperlipemia, diabetes). After adjusting these baseline characteristics, multivariate Cox regression analysis showed a signi cant difference between WMG and IMG (HR, 0.72; 95% CI, 0.62-0.84; P 0.0001). Then, we excluded the patients with poor medication compliance during follow-up (12 patients in WMG and 33 patients in IMG), the results of multivariate Cox still showed that the risk of MACE was signi cantly lower in IMG than WMG (HR, 0.71; 95% CI, 0.61-0.82; P 0.0001).
The incidence of the components of MACE were further investigated. The differences of two groups in allcause mortality (HR, 0.11; 95%CI, 0.05-0.24; P<0.0001) and revascularization (HR, 0.82; 95%CI, 0.70-0.96; P=0.012) were statistically signi cant ( Figure 2B & C). The number of patients who experienced MI was so small that Cox regression analysis was not performed (6 patients in WMG vs. 3 patients in IMG).

Secondary Outcomes
After excluding 77 deaths during 2 years (69 patients in WMG and 8 patients in IMG), 5376 survival patients were followed up with ASS and TCMSS. We analyzed the difference value between the last follow-up and baseline. Table 2 showed that the variations of ASS (P<0.001) and TCMSS (P<0.001) were both signi cantly different between two groups. Figure 3 showed the cumulative risks of cardiovascular hospitalization and stroke between the two groups. 406 patients (13.85%) in WMG and 296 patients (11.74%) in IMG hospitalized for cardiovascular reasons during the 2-year follow up. Multivariate Cox regression analysis showed that the risk of cardiovascular hospitalization in IMG was lower 19.1% than WMG (HR, 0.81; 95%CI, 0.70-0.94; P=0.007).

Safety Outcomes
In terms of stroke, 24 patients (0.82%) in WMG and 23 patients (0.91%) in IMG experienced it. Multivariate Cox regression indicated that there was no signi cant difference in the incidence of stroke between the two groups (HR, 1.04; 95%CI, 0.58-1.86; P=0.896).

Discussion
To our knowledge, this study is the largest prospective clinical study on the ICWM for CAD patients after PCI with MACE as the primary outcome index, and it is also the rst study focusing on the real-world e cacy on this topic. The results demonstrated that compared to conventional western medicine alone, ICWM showed a signi cantly lower risk of MACE (especially in all-cause mortality and revascularization) and milder clinical symptoms.
Our study found that ICWM had a signi cant advantage in the prognosis of patients after PCI, which was consistent with the results of previous RCT studies. 5C trial 9 of 808 patients with acute coronary syndrome after PCI reported that ICWM can bring a signi cant bene t to MACE. Similarly, another study 10 of 1054 patients with stable CAD showed that ICWM reduced the incidence of MACE. In terms of observational studies, a prospective cohort study of 334 AMI patients showed that ICWM is one of the protective factors. 18 Another prognostic analysis of 5284 patients with CAD got a similar result, which showed that ICWM has a potential role in reducing MACE, but there was a great difference in sample size between the two groups, which may reduce the statistical power. 19 However, some of studies showed that ICWM can only bene t clinical symptom scores, but cannot reduce the rate of cardiovascular events 12,13 . The different results may be related to the sample size, follow-up time, drugs and de nition of MACE.
Overall, the 2-year incidence of MACE in this study is 13.3% (15.0% in WMG vs. 11.2% in IMG), which is consistent with the incidence in recent relevant studies [20][21][22][23] . However, this study showed a signi cant difference in mortality between two groups, which has an obvious discrepancy with previous studies about ICWM for patients after PCI 8-10, 13 . The results differed from the previous studies may be caused by the following reasons: rstly, the sample size of those studies was small and the follow-up time was short. Secondly, most of the previous studies focused on cardiovascular mortality, rather than not allcause mortality. TCM is not only aimed at the heart or coronary artery but regards the patient as a whole, many patients in this study have other life-threatening diseases such as cancers, heart failure, abnormal liver or kidney function, or other complications. The holistic view of Chinese medicine may be bene cial to the prognosis of patients. Thirdly, the medication and endpoint events information for the post-mortem follow-up of the deceased patient was obtained from the family, rather than from themselves. This problem may lead to deviation in the classi cation of the deceased, but this is inevitable. In addition, we asked multiple family members of each patient to reduce their memory bias.
In this study, the actual clinical medication of patients within two years after PCI was recorded. The TCM taken by the patients in this study has been extensively utilized in clinical practice in China, some of which have been proved to have a signi cant effect on CAD in previous clinical research 8, 14,24 . In terms of the mechanism of drug action, the previous studies showed that Shexiang Baoxin Pill can reduce myocardial infarction area and protect cardiac function through promoting angiogenesis 25 . Qishen capsules can increase angiogenesis to improve the cardiac function of rats after MI by the Akt signaling pathway 26 . Tongxinluo can alleviate myocardial ischemia/reperfusion injury by activating cardiac microvascular endothelial cells endothelial nitric oxide synthase 27 . Compound danshen dripping pills ameliorated myocardial ischemia, reversed the reprogramming of the metabolism induced by ISO and normalized the level of most myocardial substrates and the genes/enzymes associated with those metabolic changes 28 .
On the one hand, this approach re ects the effect of ICWM in clinical practice, where the TCM may not be continually taken, thus evaluating the effect of ICWM in a more practical and comprehensive manner. On the other hand, the di culties of standardizing the intervention factors in TCM treatment and drugs, which may vary according to syndromes in response to time, place, climate etc. beyond the disease in TCMs are avoided.
Notably, this study, based on real-world data, can help to augment and extrapolate data obtained in randomized controlled trials and provide information about the safety and effectiveness of the medication in heterogeneous and large populations. This is an attempt to explore and evaluate the e cacy of ICWM based on the changeable clinical characteristics of TCM application. The patients usually take TCM only when they feel uncomfortable in China, few patients take TCM continually as they did in RCTs. Therefore, compared to the previous RCTs, this study can better re ect the clinical practice of ICWM in the treatment of CAD in the real world. In addition, the disease types of patients with CAD are the same, but the syndromes of them are different, and dialectical treatment is the fundamental principle of TCM, so the drugs in the sight of TCM are also different for them. Furthermore, the syndromes of patients will change with time, place, climate and other factors, TCM also needs to be adjusted accordingly. As a result, on the one hand, all patients with CAD using the same TCM are not in line with the theoretical basis of TCM and clinical usage of the actual situation. On the other hand, RCTs need to accurately de ne the intervention, to determine the impact of a speci c intervention factor on the study endpoints. Two of them are contradictory and it is one of the di culties in the development of evidence-based studies for TCM. So, we try to use this type of study to discuss this issue.
This study has the following limitations: rstly, as a prospective, observational, real-world study, inclusion criteria were wider than the RCT study, and patient heterogeneity was greater. Although multivariate adjustment was performed in Cox analysis of endpoint events, other unknown factors that may have in uenced the results may still exist; secondly, this study is single-center clinical research, which may limit the generalizability of the results, although it may promote consistency of treatment and evaluation. We are looking forward to relevant RCTs with a large sample size and long follow-up time to verify our study.
As for the cohort study, the main shortage was bias and confounding factors, which in uences the reliability of the study. In the endpoint events analysis present study, we adjusted for baseline characteristics that might affect prognosis to make a good match between two groups, such as demographic characteristics, disease history, and comorbidities. Moreover, the losing follow-up rate is only 4.0% in the present study. Therefore, it is reasonable to believe that the present clinical trial had a well-controlled confounding bias, which indicated the results had good reliability and credibility.

Conclusions
Compared to western medicine alone, ICWM had a signi cant lower risk of 2-year MACE among the patients after PCI, especially in terms of all-cause mortality and revascularization. The additional bene ts of TCM based on conventional western medicines can be anticipated to further improve the second prevention of CAD in the future. Median (Q1, Q3) 0.00 0.00 0.00 -1.00 -1.00 0.00

Abbreviations
Minimum, Maximum -4.00 8.00 -6.00 7.00 a ASS, Angina Symptom Score, consists of 4 items with a total score ranging from 0 to 24, with higher