Analysis of Pathogenic Bacteria and Associated Risk Factors of Bloodstream Infections in ICU Patients Undergoing ECMO Support Using mNGS Technology
Abstract
Objective: To analyze the characteristics of pathogens and related risk factors of bloodstream infections (BSI) in patients undergoing ECMO support using mNGS technology.
Methods: A retrospective analysis was conducted on 59 patients who received ECMO support in the Emergency Intensive Care Unit (EICU) of the Affiliated Hospital of Xuzhou Medical University from January 2021 to March 2024. The patients were divided into two groups based on the presence of BSI: the infection group (BSI group) and the noninfection group (N-BSI group). The clinical data of the two groups were compared, and the characteristics of pathogens were analyzed using mNGS technology. Logistic regression was used to analyze the related risk factors for BSI associated with ECMO.
Results: (1) The incidence of BSI in patients undergoing ECMO support was 20.34%. Compared to the N-BSI group, the BSI group had significantly higher levels of procalcitonin (14.9 ± 7.2 vs. 9.4 ± 4.7), C-reactive protein (140.58 ± 24.64 vs. 87.26 ± 11.06), blood lactate (8.55 ± 1.40 vs. 5.07 ± 0.55), and ECMO catheter indwelling time (12.00 ± 1.71 vs. 7.96 ± 0.76) (p < 0.05). (2) mNGS detection indicated that the BSI group mainly identified viruses and Gram-negative bacilli (G-), with Acinetobacter baumannii being the most prevalent pathogen. The resistance genes were predominantly blaTEM, which confers resistance to penicillins and cephalosporins. (3) Logistic regression analysis revealed that diabetes, ECMO indwelling time, blood lactate levels, and G-bacilli infection were risk factors for BSI during ECMO.
Conclusion: Gram-negative bacteria are the primary pathogens in BSI among patients undergoing ECMO support. Diabetes, ECMO catheter indwelling time, and elevated blood lactate are independent risk factors for these infections. Rational selection of antibiotics and strengthened management of related factors can effectively control the occurrence of BSI during ECMO support.
1. Introduction
Extracorporeal Membrane Oxygenation (ECMO) is an extracorporeal life support technology [1–3]. First applied clinically abroad in 1971, ECMO technology was introduced to China at the end of the 20th century and is now widely used in clinical critical care medicine. Depending on the cause of illness, it is mainly used for the treatment of respiratory failure or combined cardiac failure [4, 5]. As an invasive life support technology, ECMO is characterized by convenient operation and timely rescue. However, with the widespread adoption of ECMO support technology, many hospitals have observed an increase in infection rates during ECMO procedures [6–9].
Infectious diseases are the most common illnesses in intensive care units (ICUs) [10]. As the disease progresses, it can lead to septic shock, subsequently affecting organ function [11, 12]. This disease, which is of global concern, is characterized by rapid transmission and high mortality rates [13, 14]. Therefore, rapid and accurate detection of pathogens has become a focus of development in recent years. Metagenomic next-generation sequencing (mNGS) is a technique that conducts comprehensive nucleic acid detection through sample specimens [15]. The process involves extracting nucleic acids from the sample and sequencing them to obtain the sequences of pathogenic microorganisms, which are then identified using specific bioinformatics analysis software packages [16]. mNGS has advantages such as rapidity, unbiased analysis, comprehensiveness, and accuracy, and it is widely used in clinical and research settings [17].
A national key research project by Guizhou Medical University found that mNGS has high sensitivity in the etiological diagnosis of severe pneumonia [18], which can guide clinicians in adjusting antibiotic strategies. Multiple journal studies by the Chinese Medical Association have shown that mNGS has high sensitivity in pathogen detection, good consistency with microbial culture in detecting pathogenic bacteria, and can significantly shorten the detection time of pathogens, thus providing convenience for clinical diagnosis and treatment and reducing patient hospital stay [10, 19–21].
Given the increasing number of patients, effectively reducing infection rates in ECMO patients, rapidly diagnosing infections, and improving survival rates have become the current treatment priorities. Previous studies have reported bloodstream infection (BSI) incidence rates in ECMO patients ranging from 15% to 25%, with Gram-negative bacilli and Staphylococcus species being the predominant pathogens [22–24]. Additionally, recent literature has highlighted ECMO duration, central venous catheterization, and immune suppression as major risk factors for BSI [22, 25]. Our study builds on these findings by employing mNGS to provide a comprehensive pathogen profile and resistance gene analysis, offering insights into infection risk factors specific to our ICU population. This study is based on this premise and is conducted as follows.
2. Materials and Methods
2.1. Patients Samples
Clinical data were collected from 59 ECMO-supported patients admitted to the Emergency ICU of Xuzhou Medical University Affiliated Hospital from January 2021 to March 2024. Patients with a body temperature above 38.5°C underwent blood mNGS and were divided into two groups based on the presence of BSI: the BSI group and the non-BSI (N-BSI) group.
-
Inclusion criteria: (1) ECMO support received. (2) Complete clinical data available. (3) No evidence of BSI prior to ECMO treatment. (4) ECMO treatment duration of more than 48 h.
-
Exclusion criteria: (1) Patients undergoing ECMO support for the first time. (2) Hospital stay of less than 24 h. (3) Patients and families who refused to participate in the study.
This study was approved by the Ethics Committee of Xuzhou Medical University Affiliated Hospital (Approval Notice: XYFY2019-JS008-01). Formal written informed consent to participate in the study has been obtained from participants.
2.2. ECMO Cannulation
ECMO was performed using MAQUET centrifugal pumps and oxygenators. Heparin-coated catheters were inserted into the central vein and femoral artery. All patients in the V-A ECMO group had a side branch circulation sheath placed in the distal femoral artery to prevent distal limb ischemia and necrosis.
2.3. mNGS
mNGS sequencing was performed using the Illumina high-throughput sequencing platform. Nucleic acid sequences of all microorganisms (including bacteria, viruses, fungi, and parasites) in the sample were extracted and sequenced. Specific bioinformatics algorithms and pathogen databases were then used to identify the species of pathogenic microorganisms.
2.4. Other Treatments for Patients
Rational use of crystalloids, colloids, and vasopressors, as well as appropriate use of analgesics and sedatives, was employed to maintain patient stability, including endotracheal intubation or bronchoalveolar lavage.
2.5. Statistics
All data were analyzed using SPSS 26.0 statistical software. Normally distributed variable were expressed as mean ± standard deviation ( ± σ) and compared using t-tests. Non-normally distributed variables were expressed as median values and analyzed using the Mann–Whitney U test. Categorical data were presented as frequencies (n, %) and compared using chi-square (χ2) test.
To identify independent risk factors for BSI in ECMO-supported patients, we performed a multivariate logistic regression analysis, adjusting for potential confounders, including PCT, CRP, blood lactate (Lac), diabetes status, ECMO catheterization duration, and the presence of Gram-negative bacteria. The odds ratios (OR) and 95% confidence intervals (CI) were calculated, and the significance level was set at α = 0.05, with p < 0.05 indicating statistical significance.
3. Results
3.1. Comparison of Baseline Information Between Two Groups
A total of 59 patients who received ECMO support were included in the study, of which 23 were female and 36 were male, with an average age of 53.06 ± 17.55 years. There were 12 patients with BSI and 47 patients with N-BSI, resulting in an incidence rate of BSI in ECMO patients of 20.34%. Univariate analysis showed no statistically significant differences between the two groups in gender, age, chronic underlying diseases, APECHE II scores, ECMO cannulation methods, white blood cell (WBC) counts, or length of hospital stay (p > 0.05). However, significant differences were observed in PCT levels, CRP levels, blood Lac values, ECMO catheter indwelling time, and prognosis (Table 1, p < 0.05).
| Item | BSI (n = 12) | N-BSI (n = 47) | t/χ2 value | p value |
|---|---|---|---|---|
| Sex (number of cases, male/female) | 8/4 | 28/19 | 0.749 | |
| Age (years) | 53.25 ± 18.30 | 54.18 ± 17.17 | −0.164 | 0.807 |
| Underlying diseases (number of cases) | 8 | 23 | p > 0.05 | |
| APECHE II | 21.08 ± 5.43 | 19.13 ± 9.40 | 0.939 | 0.355 |
| WBC count | 17.98 ± 2.29 | 17.34 ± 1.58 | 1.189 | 0.851 |
| C-RP | 140.58 ± 24.64 | 87.26 ± 11.06 | 2.120 | 0.038∗ |
| PCT | 14.9 ± 7.2 | 9.4 ± 4.7 | 1.44 | 0.023∗ |
| Blood lac value | 8.55 ± 1.40 | 5.07 ± 0.55 | 2.311 | 0.009∗ |
| ECMO duration (days) | 12.00 ± 1.71 | 7.96 ± 0.76 | 2.322 | 0.024∗ |
| ECMO catheterization method (V-A mode vs. V-V mode) | 7 versus 5 | 28 versus 19 | p > 0.05 | |
| Hospitalization duration (days) | 22.08 ± 3.35 | 16.09 ± 1.45 | 1.804 | 0.076 |
| Survival rate (%) | 41.67% (5/12) | 48.94% (23/47) | 0.041∗ |
- Note: The bold values indicate ∗p < 0.05; ∗∗p < 0.01, statistically significant.
3.2. Pathogen Distribution in Two Groups
Blood samples from patients in the BSI group were tested using mNGS, which identified 26 strains of pathogens. The predominant pathogens included viruses and Gram-negative bacilli (G-), with human herpesvirus being the most frequently detected (19.23%, 5/26), followed by Acinetobacter baumannii (15.38%, 4/26), Staphylococcus epidermidis (7.69%, 2/26), Klebsiella pneumoniae (3.85%, 1/26), Haemophilus influenzae (3.85%, 1/26), Aspergillus (3.85%, 1/26), and Tropheryma whipplei (3.85%, 1/26). Additional detected pathogens included Staphylococcus aureus (SA), Circovirus, and Rhinovirus.
Based on clinical manifestations, sequencing read counts, and imaging findings, viruses were considered non-pathogenic, whereas Gram-negative bacteria were identified as pathogenic. Notably, Acinetobacter baumannii exhibited resistance genes predominantly of blaTEM type, indicating resistance to penicillins and cephalosporins, such as ceftazidime and cefotaxime, with a resistance rate exceeding95%.
In the BSI group, bronchial lavage fluid was tested using mNGS, which identified 43 strains of pathogens. Bacteria accounted for 74.42% (32/43), viruses for 16.28% (7/43), and fungi for 9.30% (4/43). The most common bacterial pathogen was Acinetobacter baumannii (40.63%, 13/32), which was found in respiratory infections. Drug resistance profiling revealed that Acinetobacter baumannii harbored prodominantly blaOXA-23 genes, indicating resistance to most antibiotics, including carbapenems (imipenem and meropenem), with sensitivity only to polymyxins. The burden of drug resistance gene varied between the respiratory and BSI. Viruses were was 16.28% (7/43) of cases, with human herpesvirus and cytomegalovirus being the most common.
For the N-BSI group, bronchial alveolar lavage fluid tested using mNGS identified 80 strains of pathogens, including 61 strains of bacteria (76.25%), 13 strains of viruses (16.25%), and 6 strains of fungi (7.50%). The predominant bacteria were Gram-negative, including Acinetobacter baumannii (32.79%, 20/61), Pseudomonas aeruginosa (9.83%, 6/61), Klebsiella pneumoniae (13.11%, 8/61), and Stenotrophomonas maltophilia (8.20%, 5/61). Common Gram-positive bacteria included SA (16.40%, 10/61) and Streptococcus pneumoniae (6.56%, 4/61). Other pathogens included Escherichia coli, Burkholderia cepacia, Serratia marcescens, and atypical pathogens such as Mycobacterium tuberculosis and Tropheryma whipplei. mNGS testing identified mecA and mecC as the primary drug resistance genes in Acinetobacter baumannii, indicating intrinsic resistance. Virus strains included Epstein–Barr virus (30.77%, 4/13), human herpesvirus (38.46%, 5/13), novel coronavirus (23.08%, 3/13), and cytomegalovirus (7.69%, 1/13). Fungal strains included Candida albicans (33.33%, 2/6), Candida glabrata (16.67%, 1/6), and Aspergillus (33.33%, 2/6) (Table 2).
| Strains | BSI (strains) | N-BSI (strains) | |
|---|---|---|---|
| Respiratory system | Blood system | Respiratory system | |
| Bacteria | 32 | 9 | 61 |
| Acinetobacter baumannii | 9 | 4 | 20 |
| Klebsiella pneumoniae | 3 | 1 | 8 |
| Pseudomonas aeruginosa | 2 | 0 | 6 |
| Stenotrophomonas maltophilia | 4 | 0 | 5 |
| Staphylococcus | 3 | 1 | 10 |
| Streptococcus pneumoniae | 2 | 1 | 4 |
| Others | 9 | 2 | 8 |
| Viruses | 7 | 16 | 13 |
| Human herpesvirus | 2 | 5 | 5 |
| COVID-19 | 2 | 2 | 3 |
| Epstein–Barr virus | 2 | 4 | 4 |
| Circovirus | 0 | 2 | 0 |
| Others | 1 | 3 | 1 |
| Fungi | 4 | 1 | 6 |
| Candida | 2 | 0 | 3 |
| Aspergillus | 2 | 1 | 2 |
| Mucor | 0 | 0 | 1 |
3.3. Infection Risk Factors and Logistic Analysis in Two Groups
To assess the independent risk factors for BSI in ECMO-supported patients, we conducted a multivariate logistic regression analysis, adjusting for relevant clinical variables. The analysis identified elevated blood Lac levels (OR = 1.33, 95% CI: 1.10–1.75, p = 0.003), diabetes (OR = 3.41, 95% CI: 1.14–10.24, p = 0.03), prolonged ECMO catheterization duration (OR = 1.22, 95% CI: 1.03–1.61, p = 0.025), and the presence of Gram-negative bacteria (OR = 4.26, 95% CI: 1.56–11.64, p = 0.004) as significant independent risk factors for BSI in ECMO patients.
Conversely, PCT (p = 0.091), CRP (p = 0.825), and viral infections (p = 0.598) were not significantly associated with an increased risk of BSI. These findings highlight the critical role of metabolic imbalances, prolonged catheterization, and bacterial infections in ECMO-associated BSI.
The detailed results of the logistic regression analysis, including OR, 95% CI, and p values, are summarized in Table 3.
| Risk factors | PCT | CRP | Lac | Diabetes | ECMO catheterization duration | Gram-negative bacteria | Viruses |
|---|---|---|---|---|---|---|---|
| Coefficient β | −0.098 | 0.003 | −0.284 | −0.011 | −0.195 | 0.35 | −0.235 |
| Wald (χ2) | 2.85 | 0.051 | 8.931 | 4.732 | 5.041 | 13.65 | 0.27 |
| p value | 0.091 | 0.825 | 0.003∗ | 0.03∗ | 0.025∗ | 0.004∗ | 0.598 |
| OR (95% CI) | 0.91 (0.78–1.11) | 1.00 (0.98–1.02) | 1.33 (1.10–1.75) | 3.41 (1.14–10.24) | 1.22 (1.03–1.61) | 4.26 (1.56–11.64) | 0.79 (0.36–1.68) |
- Note: The bold values indicate ∗p < 0.05, ∗∗p < 0.01, statistically significant.
- Abbreviations: CRP, C-reactive protein; Lac, lactate; OR, odds ratios; PCT, procalcitonin.
4. Discussion
ECMO has rapidly developed both domestically and internationally as the last line of defense in the treatment of heart and lung failure [28, 29]. In recent years, with the widespread adoption and continuous improvement of ECMO support technology in China, the number of cases treated and clinical complications have also increased annually, with infections ranking as the top complication, followed by thrombosis, hepatic and renal dysfunction, hemolysis, and other complications [30–33]. This study divided ECMO patients in our department into two groups based on the presence of BSI, with an incidence rate of 20.34% for BSI, consistent with previous studies that estimate rates between 15% and 25% [22]. The incidence in the V-A ECMO group was slightly higher than in the V-V ECMO group (7 vs. 5).
High-throughput gene sequencing detected that viruses were the most common pathogens in BSI, followed by Gram-negative bacilli (G-), with Acinetobacter baumannii being the most prevalent. The resistance genes identified in Acinetobacter baumannii, such as blaTEM, indicated resistance to penicillins and cephalosporins, including ceftazidime and cefotaxime, with a resistance rate exceeding 95% [34]. The identification of resistance genes has significant clinical implications for the management of BSI in ECMO patients. The presence of blaTEM and blaOXA-23 genes in Acinetobacter baumannii highlights the challenge of treating infections caused by carbapenem-resistant strains [34]. These genes encode β-lactamase enzymes, which hydrolyze β-lactam antibiotics, rendering them ineffective. Clinically, this necessitates the use of alternative antibiotics, such as polymyxins, which may be associated with nephrotoxicity and other adverse effects [35].
Our findings align with global trends, indicating that Gram-negative bacilli remain the predominant causative agents of BSI in ECMO patients. The high detection rate of Acinetobacter baumannii in our study is consistent with findings from large ICU-based studies, which have shown its significant role in ECMO-related infections [36, 37]. Additionally, our data provide further insight into the presence of resistant strains, such as carbapenem-resistant Acinetobacter baumannii, which has been increasingly reported as a major contributor to ECMO-associated infections in recent literature [38, 39]. These results highlight the urgent need for inproved infection control strategies and antimicrobial stewardship in ECMO-supported patients.
In the BSI group, bronchial lavage fluid was tested using mNGS, which identified carbapenem-resistant Acinetobacter baumannii as the most common pathogen in respiratory infections. The presence of blaOXA-23 genes, which confer resistance to most carbapenems and other antibiotics, poses a significant therapeutic challenge [40]. The resistance conferred by these genes highlights the need for early detection of multidrug-resistant organisms. For example, polymyxin was the only antibiotic to which Acinetobacter baumannii showed sensitivity, underscoring the clinical importance of resistance gene analysis for guiding empirical therapy [41]. Resistance genes like blaOXA-23 are often located on mobile genetic elements, such as plasmids, which facilitate horizontal gene transfer among pathogens, potentially escalating resistance within ICU settings [42, 43]. This reinforces the need for rigorous infection control measures to prevent the spread of resistant pathogens.
In the N-BSI group, bronchial lavage fluid test uring mNGS identified Klebsiella pneumoniae and SA as common pathogens, with resistance gene such as mecA and mecC detected in SA [44–46]. These genes confer methicillin resistance by encoding an altered penicillin-binding protein (PBP2) that reduces the efficacy of β-lactam antibiotics [47]. The clinical implication of these findings is the need to use alternative agents, such as vancomycin or linezolid, for treating methicillin-resistant SA (MRSA) [48]. The widespread distribution of resistance genes in critical care settings further emphasizes the importance of using mNGS to detect resistance mechanisms promptly, enabling the selection of targeted therapies while minimizing the development of further resistance.
However, the detection rate of bacteria using mNGS maybe influenced by prior antibiotic exposure, which can alter microbial load and affect sequencing results. To assess this potential impact, we reviewed patients’ antibiotic history before sample collection. In our cohort, 85% of patients had received broad-spectrum antibiotics, primarily carbapenems or glycopeptides, as part of standard empirical therapy for critically ill ECMO patients. Whenever feasible, a 24 h washout period was implemented before sample collection to allow partial clearance of exogenous microbial DNA and minimize potential bias. However, in cases of severe illness, continuous antibiotic administration was unavoidable. Despite this, mNGS successfully identified a high proportion of bacterial and viral pathogens, suggesting that while antibiotic exposure may reduce bacterial DNA load, it does not entirely suppress microbial detection. These findings align with previous studies indicating that mNGS remains a valuable diagnostic tool even in patients undergoing antibiotic therapy, providing crucial information for timely and targeted treatment adjustments.
Acinetobacter baumannii, an opportunistic pathogen, caused infection in these patients due to varying degrees of immunosuppression, where viral infections directly damaged cells, followed by immune-mediated cellular damage, which further induced the invasion of opportunistic pathogens [49]. This led to inherent properties of bacterial biofilms and bacterial translocation [50], resulting in pneumonia combined with BSI [51, 52]. Therefore, enhancing immunity, reducing exposure, and wearing masks are of great significance in preventing viral invasion and reducing infection rates [53, 54].
Mechanical ventilation time is an important risk factor for infections with carbapenem-resistant Acinetobacter baumannii, which has been extensively documented as a major cause of nosocomial infections in ECMO patients [55]. Studies have suggested that patient prognosis is correlated with the duration of mechanical ventilation [56]. Unfortunately, in this study, patients in the V-A ECMO group did not have respiratory symptoms. With the prolonged duration of ECMO support, the mechanical ventilation time also extended, leading to ventilator-associated pneumonia (VAP) in 9 patients, with an infection rate of 25.71%. The main pathogen detected was carbapenem-resistant Acinetobacter baumannii, with blaOXA-23 being the main resistance gene, indicating pan-drug resistance to most antibiotics, except sensitivity to polymyxin. It has been suggested that the OXA gene, located on the plasmid or chromosome of the pathogen, facilitates easier transfer and drug escape, leading to higher resistance. The blaOXA-23 gene is located in carbapenem-resistant Acinetobacter baumannii and is a key resistance gene detected in these pathogens [57–59]. This highlights the need for clinical healthcare providers to reduce mechanical ventilation time in ECMO patients to lower the incidence of ventilator-associated pneumonia, which is crucial for controlling VAP [60].
The duration of ECMO support was identified as a critical factor influencing the risk of infection. These findings corroborate prior reports that ECMO duration and metabolic abnormalities, such as hyperglycemia, significantly contribute to infection susceptibility and poor outcomes [61, 62]. Long-term high blood glucose levels disrupt the body’s homeostasis and increase the risk of pulmonary infections, which are common complications in type 2 diabetes, primarily caused by bacterial infections [63]. This retrospective study found that diabetes is a significant risk factor for high mortality rates in BSI. Beijing Anzhen Hospital’s research suggested that ECMO infections are related to immune metabolic abnormalities [64]. Additionally, prolonged treatment time and central catheter placement are also significantly associated with increased hospital-acquired BSI rates in ICU patients [65]. Our study showed that BSI in critically ill patients during ECMO support were highly related to the duration of ECMO operation, which could be associated with the destruction of the patient’s immune barrier and the environmental disinfection of the ward.
BSI patients often required large doses of vasopressors, and our study indicated significantly elevated blood Lac levels in these patients. Lac, a metabolic product of exacerbated blood acidification, increases under conditions of hypovolemia and microcirculatory disturbances, triggering elevated Lac levels and leading to an acidic environment that disrupts homeostasis and immune function [66, 67], facilitating the proliferation of opportunistic pathogens. Numerous studies have shown the prognostic value of Lac in patients with septic shock, with critical care journals highlighting the prognostic evaluation value of blood Lac in septic shock [68–71]. Our analysis of patient data showed that the destruction of the immune barrier and the negative pressure oxygenation of ECMO further impair immune function, aligning with the conditions for BSI in ECMO patients. Therefore, clinical healthcare providers should be vigilant for potential BSI in patients with elevated blood Lac levels, promptly identifying infection sources to prevent further exacerbation.
This study is limited by its single-center design, requiring multicenter studies to further verify the results. Additionally, there is a lack of long-term follow-up data for some patients’ prognosis. Finally, further rigorous prospective controlled studies are needed to confirm the infection risk factors in ECMO patients.
In conclusion, ECMO support therapy is associated with a high incidence of infections, including BSI, catheter-associated BSI, and ventilator-associated pneumonia, with Gram-negative bacilli being the main pathogens. Reducing catheter-associated infections requires attention to clinical blood Lac levels and ECMO catheter indwelling time [72]. Additionally, attention to patients’ underlying conditions, especially long-term poor blood glucose control, is crucial for preventing infections during ECMO therapy.
Ethics Statement
All procedures performed in studies involving human participants were following the ethical standards of Xuzhou Medical University Affiliated Hospital (ethical approval notice: XYFY2019-JS008-01), and with the 1964 Helsinki Declaration. Informed consent to participate in the study has been obtained from participants.
Consent
Please see the Ethics Statement.
Conflicts of Interest
The authors declare no conflicts of interest.
Author Contributions
Y.W., S.S. and X.H. conceived and designed the research. J.K., Y.H., Z.L. and M.Y. acquired the data. X.Q., Y.Y. and Y.Z. performed the analysis. J.K. and S.S. prepared the draft. J.K. prepared the tables. Y.W., S.S. and X.H. reviewed and edited the manuscript. All authors read and approved the final manuscript. All authors agree on the publication of this research article.
Funding
This study was supported by the National Natural Science Foundation of China (81802504), the Sichuan Science and Technology Program (2025YFHZ0123), Chengdu Science and Technology Program (2024-YF05-01315-SN), and a grant from Shenzhen Weixin (2024HX008) for Dr. Yi Wang. It is also supported by an open fund of Key Laboratory for Hubei Provincial (2023KFZZ026) for Dr. Man Yuan.
Acknowledgments
The authors have nothing to report.
Open Research
Data Availability Statement
Please contact the authors for data requests.
