Effects of (Combined) Chinese Medicine on the Risk of Death From COVID-19: A Retrospective Data Analysis Based on 4567 Patients

The coronavirus disease 2019 (COVID-19) epidemic is still ongoing, but the optimal treatment remains unclear. China adopted a series of measures, including widespread screening, strict quarantine and early treatment, combining western medicine with Chinese medicine, leading to rapid control of its spread. Nevertheless, the effects of ( combined ) Chinese medicine in reducing the toll of COVID-19 lack proof from statistics. We conducted a retrospective data analysis to determine whether ( combined ) Chinese medicine is able to affect patient outcomes and to decrease the risk of death in COVID-19 patients. following The diagnostic criteria refer to the published A positive laboratory test was de�ned as a con�rmed case.


Introduction
Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, has led to a pandemic of more than 40 million infected individuals and 1.11 million deaths [1]. The COVID-19 pandemic not only poses a massive challenge to the global health system but also has handicapped economic development and social stability worldwide. Thus far, the e cacy and the safety of drugs and vaccines needs to be proven, and the best methods to control the pandemic are still early detection and quarantine, with the proper treatment approach being unclear.
China insisted upon early diagnosis, quarantine and treatment using western medicine along with Chinese medicine, and it quickly controlled the disease. In the ght against COVID-19, China also went from panic and disorder to calm and order. Treatment protocols for COVID-19 matured in clinical practice as well. Since the fth version of the protocol published on February 4, 2020, the Chinese government has alleviated Wuhan City's pandemic pressure from the aspects of necessary items, equipment, personnel and a shortage of sickbeds. Based on the experience of using Chinese medicine in the early stages, the National Health Commission and the State Administration of Traditional Chinese Medicine promoted and comprehensively strengthened the use of Chinese medicine in the COVID-19 epidemic. According to the incomplete statistics, the utilization rate of Chinese medicine in the treatment of COVID-19 in all provinces in China is approximately 90%.
Albeit, currently, clinical trials of Chinese medicine are mostly observational, there is no denying that Chinese medicine plays an important role in relieving symptoms and preventing the progression of the disease. A newly published systematic review and meta-analysis of 1474 patients with COVID-19 treated with a combination of western and Chinese medicine showed that, in the combined Chinese medicine group, the overall clinical response rate (OR=2.67, 95% CI 1.83-3.89, I 2 =0%), CT scan improvement (OR=2.43, 95% CI 1.80-3.29, I 2 =0%), percentage of severe/critical cases (OR=0.40, 95% CI 24-0.67, I 2 =17.1%) , negative rate of reverse transcription polymerase chain reaction (RT-PCR) (OR=2.55, 95% CI 1.06-6.17, I 2 =56.4%), and symptom disappearance rate (fever, cough and fatigue) were better than in the control group [4]. However, there is still a lack of research evidence supporting the treatment effect of combined Chinese medicine in reducing death and critical illness.
Therefore, based on Wuhan's large-scale introduction of Chinese medicine treatment for COVID-19, the related data were retrospectively analyzed, and further study of traditional Chinese medicine combined with western medicine for COVID-19 is very important. Reducing mortality is the key to clinical treatment, and whether (combined) Chinese medicine can affect COVID-19 clinical outcomes needs further study.
Our study is a retrospective analysis, and the application for an exemption from informed consent was approved by the Ethics Committee of the Guangdong Hospital of Traditional Chinese Medicine (batch number: ZE2020-049-01).

Research objective
Data source: COVID-19 inpatients in four designated hospitals (the following are the four centers) in Wuhan City from February 4, 2020, to April 30, 2020. The diagnostic criteria refer to the protocol published by the National Health Committee. A positive laboratory test was de ned as a con rmed case.
Inclusion criteria: Laboratory-con rmed cases Exclusion criteria: Time of admission was missing or was prior to February 4, 2020; the discharge status was missing; and a lack of information about medication.

Supervision:
This study was funded by the National Administration of Traditional Chinese Medicine and was designed by the researchers. The data were analyzed and interpreted by the authors. All of the authors reviewed the manuscript and ensured the accuracy and completeness of the data.
Information extraction: In this study, COVID-19 medical records from Centers 1 and 4 were directly exported as formatted information by the HIS system and then checked and classi ed by professionals engaged in data management. COVID-19 medical records from Centers 2 and 3 were extracted and entered into the database from the original medical records after authorization, and then were exported as a formatted database. During the data extraction, the information generated or required to be clinically judged by professionals was calculated and discussed by clinical experts, clinical research methodology experts, statistical experts and other experts, and then clearly de ned. The extracted information included 1) demographic information, such as sex, age and basic diseases; 2) case information, such as the time of onset, time of admission, and clinical classi cation of vital signs; 3) treatment information, including whether antiviral drugs, antibiotics, hormone immunotherapy, Traditional Chinese medicine decoction granules, etc., were used; and 4) prognostic information, including survival/death.

De nition of different clinical types:
According to the diagnosis and treatment protocol of COVID-19, patients were clinically classi ed as: 1) mild: clinical symptoms were mild, and no manifestations of pneumonia were observed on imaging; 2) common: with symptoms such as fever and respiratory tract symptoms, with imaging ndings of pneumonia; 3) severe: met any of the following criteria: RR 30 times/min with shortness of breath; at rest, oxygen saturation was 93%; arterial partial pressure of oxygen (Pa0 2 )/oxygen concentration (Fi0 2 ) 300 mg, 1 mmHg= 0.1 Pa; and 4) critical: respiratory failure and needing mechanical ventilation or going into shock, combined with other organ failure, requires care in the ICU.
De nition of different groups: 1. Intervention measures included basic treatment (oxygen therapy, liquid therapy, etc.), antivirals, antibiotics, hormones or immune support, identi ed as the western medicine group.
2. On the basis of the abovementioned western medicine group, those who used Chinese patent medicine or Chinese medicine decoctions or Chinese medicine granules for COVID-19 treatment at the same time were identi ed as the Chinese Medicine/Chinese Medicine combined group.
Statistical analysis methods: The clinical outcomes, the disease population characteristics, the clinical symptoms and signs, etc., were calculated based on the medication group. A descriptive analysis of the mean ± SD median (interquartile interval), composition ratio and rate was performed for the comparison between the data groups. The counting data were compared between groups using the chi-square test, Fisher's exact probability method, etc., and propensity matching was adopted so that the general data between the groups was balanced. A Cox proportional hazard model was used to correct for factors such as demographic admission classi cation, and the death risk of the different groups was investigated. The Schoenfeld residual test was used to check the proportionality assumption of the Cox regression.

Results
Generation conditions: In this study, the clinical medical records of 2554 patients from center 1 (Wuhan Huoshenshan Hospital), center 2 (Hubei Province Hospital of Integrated Traditional Chinese and Western Medicine), center 3 (Wuhan Raytheon Hospital), and center 4 (Wuhan Hankou Hospital) were obtained. The clinical medical records of 6076 patients were obtained by combining the data from the 4 hospitals. A total of 13 patients were excluded due to not having COVID-19, 180 patients were excluded due to a lack of admission time data, 803 patients were excluded because their hospital admission was before February 4, 2020, 449 patients were excluded due to missing treatment data, and 645 patients were excluded due to missing hospital discharge status data. Finally, 4,567 patients were included in the descriptive analysis ( Figure 1, Comparison of general conditions between the different intervention groups: A total of 4,567 COVID-19 patients were divided into the western medicine group and the traditional Chinese medicine/combined traditional Chinese medicine and western medicine group for general situational comparisons according to whether traditional Chinese medicine, decoctions or granules for COVID-19 treatment were used together in the intervention measures (Table 2). In addition to sex and combining with tumor-based diseases, there were statistically signi cant differences between the two groups in age, clinical classi cation, onset to admission time, chronic liver disease, chronic respiratory system disease, and chronic kidney disease (P<0.05), with imbalances between the groups. Therefore, only the prognosis (crude case fatality rate) of different clinical types between the two groups was described (Table 3). Among the 4,567 patients in this group, the crude case fatality rate was 18.5% in the western medicine group and 2.2% in the Chinese medicine/Combined Chinese medicine group. Comparison of general conditions and prognosis between groups after propensity matching: To ensure balance for comparability between groups, propensity matching was adopted. With age, sex, the onset of admission type and the number of days after admission combined with basic diseases (cardiovascular and cerebrovascular disease, diabetes, liver disease, chronic respiratory disease, renal disease and tumor) as matching variables, the western medicine group and combined Chinese medicine group were matched at a 1:2 ratio, the tolerance of matching was 0.01, and data for 243 patients in the western medicine group and 486 patients in the (combined) Chinese medicine group were obtained. The results showed that after matching, there were no statistically signi cant differences between the two groups regarding sex, age, clinical classi cation of admission, diabetes, liver disease, chronic lung disease, kidney disease, tumors and cardiovascular and cerebrovascular diseases (P≤0.05), except for the time from onset to admission, which was comparable between the groups (Table 4). Case fatality rate comparison between groups after propensity matching: The fatality rates of the different intervention groups were compared among the data sets after bias matching (Table 5). Among the 243 patients in the western medicine group, 45 died, with a fatality rate of 18.5%. Among the 486 patients in the (combined) Chinese medicine group, 39 patients died, with a fatality rate of 8.0%. The difference in fatality rates between the two groups was statistically signi cant (P≤0.05). Furthermore, the case fatality rate of the common type of COVID-19 in the western medicine group was higher than that in the (combined) Chinese medicine group (8/144, 5.6% vs 2/281, 0.7%), and the difference was statistically signi cant (P<0.05). The case fatality rate of severe COVID-19 in the western medicine group was also higher than that in the (combined) Chinese medicine group (25/83, 30.1% vs 25/179, 14.0%), and the difference between the two groups was statistically signi cant (P<0.05), suggesting that (combined) Chinese medicine treatment may affect the prognosis of common and severe COVID-19. The chi-square test, CMH test and Cox regression analysis showed that there was no center effect; that is, the data from the four centers could be combined for analysis. To further investigate the prognostic effect of the (combined) Chinese medicine treatment on COVID-19 and to correct for the in uence of factors such as general admission classi cation combined with underlying diseases, a Cox proportional hazard regression model was used for the analysis, with outcome variable Y=1 indicating death and Y=0 indicating survival. Sex, age strati cation (<65 years old, ≥65 years old), type of admission (mild + normal, severe and critical), onset to days of admission (<14 days, 14-28 days, ≥28 days), groups (western medicine and (combined) Chinese medicine) and basic diseases (cardiovascular and cerebrovascular diseases, diabetes, liver diseases, chronic lung diseases, kidney diseases, tumors) were screened by the Forward Wald method. The results showed that after adjusting for other factors, combined Chinese medicine still reduced the risk of death from COVID-19 compared with western medicine alone (HR=0.135, 95% CI 0.088, 0.208). The risk of death was 6.991 times higher for the severe type and 67.338 times higher for the critical type than for the mild/normal type. Compared with onset to days of admission>28 days, the risk death of <14 days was 4.688 times higher, and the risk of death at 14-28 days was 1.965 times higher. The risk of death in patients >65 years old was 2.557 times that in ≤65-year-old patients; the risk of death in those with CKD was 1.705 times higher than that in those without CKD (Table 6, Figure 2).

Discussion
Around the world, epidemiologists are setting short and long-term projections of COVID-19 as a way to mitigate the spread and impact of SARS-CoV-2. Although their forecasts and timetables vary, they all agree that COVID-19 is still present and the future depends on many unknown factors [5]. Therefore, the prevention and treatment of COVID-19, the expansion of intensive care capacity, the detection of effective treatments, and the reduction of mortality from COVID-19 remain long-term clinical and public health issues.
According to the current global COVID-19 data report, the crude case fatality rate is approximately 2.9% (1,060786/36 4440433) [1]. In this study, in order to avoid any in uence of the early confusion, shortages, and the COVID-19 case fatality rate, we eliminated all patients admitted before February 4, 2020. At the same time, to eliminate other important causes of missing data, we included a descriptive analysis of 4567 cases, and the crude mortality was 3.0%, in line with the global outbreak data.
Importantly, in this study, we found that, after adjusting for other prognostic risk factors, the risk of death from COVID-19 was reduced by (combined) Chinese medicine compared with western medicine alone, while any bias caused by a shortage of medical supplies was avoided as much as possible. However, in this set of data, there may be some defects in our de nition of a (combined) Chinese medicine intervention. First, due to the incomplete information about the original medical advice, it was not possible to de ne the timing of the Chinese medicine intervention and to discuss the dose-time effect.
Interestingly, we analyzed the data with detailed medication information from Center 1 and found that there was a certain correlation between the time of using Chinese medicine and the prognosis. Second, there may be differences in the clinical effects of Chinese medicine decoctions, granules and proprietary Chinese medicine. Due to the limited number of cases and the incomplete original medical advice information, this study failed to further stratify the (combined) Chinese medicine treatment scheme and evaluated only the impact of (combined) Chinese medicine treatment on the prognosis at the overall level. An in-depth analysis and discussion is needed to further clarify which programs, drugs or prescriptions are effective. Third, in this study, the Chinese medicine diagnosis and treatment schemes, drugs and prescriptions adopted by the different centers may be different, and the diagnosis and treatment levels in different centers may also be different. However, in the Cox risk regression analysis, we did not nd any in uence of a center effect, which may be related to the early formation of a uni ed treatment scheme in China.
Multivariate Cox regression also showed that being clinically categorized as severe/critical, ≥65 years of age, and complicated with chronic kidney disease all increased the risk of death in COVID-19 patients, which is basically consistent with the current reports on the prognostic factors of COVID-19 at home and abroad [6-7]. However, due to the lack of laboratory examination results and dynamic monitoring data, indicators such as the D2-cluster level, IL-6, SOFA score, total bilirubin, myocardial markers, platelets and so on were not included in the analysis of prognostic factors of COVID-19 in this study. In addition, this study found that a time from onset to admission to the hospital <14 days or 14-28 days increased the risk of death from COVID-19. When analyzing the reasons for this nding, even if we excluded patients admitted before February 4 th , during the period of resource shortages, we note that resources after that time were still limited, and the Chinese government, based on humanitarian needs, used a priority strategy for severe patients. As a result, patients admitted in the early stage might have a higher risk of death.
Therefore, we further analyzed and compared the time of onset to admission of patients among the different clinical types (enclosure Table 3), and we found that the clinical type of admission in the severe group and the proportion of patients with onset to admission days <14 days had the highest rate of death. The admission classi cation of the critical type group and the proportion of days from onset to admission was higher, indicating that the number of days from onset to admission was indeed correlated with the clinical classi cation of disease admission, which also reasonably explained the results of the multivariate Cox regression.
In addition, the cases in this study were all from Wuhan. Due to the in uence of hospitalization, diagnosis and treatment strategy and other factors, the number of cases treated by pure western medicine was relatively small, so the number of cases that could be included in the propensity matching was limited. Enclosure.docx