OBJECTIVE To evaluate the anticoagulant efficacy and safety of nafamostat mesylate （NM） in the treatment of uremic patients at high risk of bleeding undergoing continuous veno-venous hemofiltration （CVVH） with different methods （pre- dilution and post-dilution）. METHODS A total of 130 uremic patients at high risk of bleeding who underwent CVVH treatment in the nephrology department of Chongqing University Three Gorges Hospital from July 2023 to September 2024 were selected. They were divided into pre-dilution group and post-dilution group according to the random number table method， with 65 cases in each group. Both groups of patients received CVVH treatment under NM anticoagulation. The pre-dilution group adopted the pre-dilution replacement method， while the post-dilution group adopted the post-dilution replacement method. The coagulation， pressure， and usage duration of the filter and dialysis circuit venous reservoirs were compared between the two groups. The changes in prothrombin time （PT）， prothrombin time-international normalized ratio （PT-INR）， activated partial thromboplastin time （APTT）， and fibrinogen （FIB） in the peripheral venous blood before the heparin pump and after the filter at 1， 4 and 7 h of CVVH treatment， as well as 20 min after the end of treatment， were compared between the two groups. The single-compartment urea clearance rate （spKt/V）， β2-microglobulin （β2-MG） clearance rate and the incidence of adverse reactions were duni2007@foxmail.com compared between the two groups. RESULTS Both the pre-dilution and post-dilution groups had 60 patients who completed the study. The incidence of grade Ⅱ-Ⅲ coagulation of the filter and venous reservoirs， as well as the number of patients with transmembrane and venous pressure alarm intervention in the post- dilution group were significantly higher or more than those in the pre-dilution group （P＜0.05）， while usage time of the filter and the pipeline in the post-dilution group was significantly shorter than that in the pre-dilution group （P＜0.05）. The APTT values before the heparin pump as well as PT and APTT values after the filter at 1 h， 4 h， and 7 h of CVVH treatment in the post-dilution group were significantly higher than those in the pre-dilution group （P＜0.001）. There were no significant differences in PT， PT- INR， APTT and FIB between the two groups of patients 20 min after the end of treatment （P＞0.05）. The spKt/v and β2-MG clearance rates in the post-dilution group were significantly higher than those in the pre-dilution group （P＜0.001）. There was no significant difference in the incidence of adverse reactions between the two groups （P＞0.05）. CONCLUSIONS When NM is used as an anticoagulant in the CVVH treatment of uremic patients at high risk of bleeding， compared with the pre-dilution treatment method， the post-dilution treatment method has a higher incidence of filter and dialysis tubing venous reservoir， a shorter usage time of the filter and pipeline， and a greater impact on extracorporeal coagulation， but has a higher solute clearance rate. Clinically， different dilution methods can be selected according to the different treatment needs of patients.

The Type III Secretion System (T3SS) serves as a pivotal virulence apparatus for numerous Gram-negative bacterial pathogens, enabling them to infect both animal and plant hosts. Functioning as a molecular syringe, the T3SS directly translocates bacterial effector proteins from the bacterial cytoplasm into the interior of eukaryotic host cells. These effectors are central weapons that precisely manipulate a wide spectrum of host cellular physiological processes, ranging from cytoskeletal dynamics to immune signaling, to establish a favorable niche for bacterial survival and proliferation. Among the diverse arsenal of T3SS effectors, the YopJ family constitutes a critical group of virulence factors. Members of this family are characterized by a conserved catalytic triad structure—a hallmark of the CE clan of cysteine proteases that has been evolutionarily repurposed to confer acetyltransferase activity. A defining and intriguing feature of these enzymes is their stringent dependence on a host-derived eukaryotic cofactor, inositol hexakisphosphate (IP₆), for allosteric activation. This requirement acts as a sophisticated molecular safeguard, ensuring enzymatic activity only within the appropriate host environment, thereby preventing detrimental effects on the bacterium itself. While seminal studies on individual members such as Yersinia’s YopJ and Salmonella’s AvrA have provided deep mechanistic insights, a systematic and integrative understanding of the structure-function relationships across the entire family remains fragmented. Key questions persist regarding how a conserved catalytic core has diverged to recognize distinct host substrates in different kingdoms of life. To address this gap, this article provides a systematic review of the YopJ family, focusing on three interconnected aspects: their structural features, their catalytic mechanism, and their divergent immunosuppressive strategies in animal versus plant hosts. By conducting a comparative analysis of the sequences and resolved three-dimensional structures of three representative members (e.g., HopZ1a, PopP2, AvrA), we elucidate regions of significant variation embedded within the conserved core catalytic architecture. These variable regions, often involving surface loops and substrate-binding interfaces, are crucial determinants of target specificity and functional specialization. The functional divergence of this effector family is most apparent when comparing their modes of action in different hosts. In animal hosts, YopJ-family effectors primarily sabotage innate immune signaling pathways. They achieve this by acetylating key serine and threonine residues within the activation loops of critical kinases in the MAPK and NF‑κB pathways. This post-translational modification blocks the phosphorylation and subsequent activation of these kinases, leading to potent suppression of inflammatory cytokine production. Conversely, in plant hosts, the strategy broadens to dismantle the two-tiered plant immune system. YopJ homologs target a more diverse set of substrates, including immune-associated receptor-like cytoplasmic kinases (RLCKs), microtubule networks via tubulin acetylation (which disrupts cellular trafficking and signaling), and transcription factors central to defense gene regulation. This multi-target approach effectively suppresses both Pattern-Triggered Immunity (PTI) and Effector-Triggered Immunity (ETI). In conclusion, this synthesis aims to deepen the mechanistic understanding of YopJ family-mediated pathogenesis by integrating structural biology with cellular function across host kingdoms. Elucidating the precise molecular basis for substrate selection—how conserved platforms achieve target diversity—is a major frontier. Furthermore, this knowledge provides a vital theoretical foundation for developing novel anti-virulence strategies. Targeting the conserved IP₆-binding pocket or the catalytic acetyltransferase activity itself represents a promising avenue for designing broad-spectrum inhibitors that could disarm this critical family of bacterial effectors, potentially offering new therapeutic approaches against a range of pathogenic bacteria.

OBJECTIVE To assess the bleeding risk signals associated with the concomitant use of direct oral anticoagulants （DOACs） and triazole antifungals， and to provide pharmacovigilance evidence for the safety evaluation and monitoring of combined clinical use. METHODS Adverse event reports involving the concomitant use of DOACs and triazole antifungals were extracted from the US FDA Adverse Event Reporting System （FAERS） from the first quarter of 2004 to the third quarter of 2025. Nine bleeding-related preferred terms （PTs） were selected. The Ω shrinkage measure， additive model， multiplicative model， and combined risk ratio method were employed to detect drug-drug interaction signals. The strength of positive signals was further analyzed based on the Ω shrinkage measure. RESULTS A total of 790 adverse event reports involving the concomitant use of DOACs and triazole antifungals were included， among which 229 reports involved nine bleeding-related PTs. A total of 13 signals were consistently identified as posit ive by all four methods. These signals involved six drug combinations： apixaban-fluconazole， apixaban-posaconazole， rivaroxaban-itraconazole， dabigatran etexilate-fluconazole， apixaban-voriconazole， and dabigatran etexilate-itraconazole. The Ω shrinkage measure showed that the apixaban-posaconazole combination exhibited stronger signals for bleeding （ Ω ＝2.73， Ω 025 ＝2.05） and hemoptysis （ Ω ＝2.17， Ω 025 ＝0.83）； the apixaban-fluconazole combination exhibited stronger signals for hematoma （ Ω ＝2.30， Ω 025 ＝1.47） and hematuria （ Ω ＝1.71， Ω 025 ＝0.74）； the rivaroxaban-itraconazole combination exhibited stronger signals for epistaxis （ Ω ＝2.01， Ω 025 ＝0.90） and hematoma （ Ω ＝1.93， Ω 025 ＝0.42）； no positive Ω signals were observed for intracranial hemorrhage or upper gastrointestinal hemorrhage. CONCLUSION S This study suggests that the concomitant use of DOACs and triazole antifungals may increase the risk of bleeding-related events， with differences in signal strength and signal distribution across various drug combinations. In clinical practice， particular attention should be paid to the concomitant use of apixaban or rivaroxaban with strong cytochrome P450 3A4 or P-glycoprotein inhibitors such as posaconazole and itraconazole. For other DOAC-triazole antifungal combinations， close monitoring for bleeding-related manifestations and timely adjustment of anticoagulation or antifungal regimens are also warranted.

Objective To construct machine learning models based on preoperative inflammatory and radiological features for the prediction of glioma grading and isocitrate dehydrogenase (IDH) mutation status, and to analyze application values of these models and identify the optimal predictive models. Methods A retrospective analysis was conducted on the data of pathologically confirmed glioma patients admitted to Tongji Hospital Affiliated to Tongji University from March 2019 to March 2023. LASSO regression was used to screen feature variables, and predictive models were constructed based on logistic regression (LR), random forest (RF), support vector machine (SVM), gradient boosting decision tree (XGBoost) and K-nearest neighbor (KNN) algorithms. The model performance was comprehensively evaluated using metrics including discrimination ability, area under the precision-recall curve (AUC), accuracy, F1 score and Brier score. The DeLong test was adopted to compare the AUC values among different models; Friedman rank-sum test was used to determine the overall performance differences of the models, with the Nemenyi test applied for multiple comparison correction. Results In the task of glioma grading prediction, the LR model achieved the highest comprehensive score (0.726), and no significant difference was observed between the LR model and the other four models; age was positively correlated with glioma grading (P=0.003). In the task of IDH mutation status prediction, the XGBoost model obtained the highest comprehensive score (0.832), which was superior to the LR (0.762, P=0.035) and KNN models (0.754, P=0.025), while no statistical differences were found between the XGBoost model and the RF or SVM models. Conclusions The LR model for glioma grading prediction and XGBoost model for IDH mutation prediction constructed based on a task-oriented strategy achieve a favorable interpretability while ensuring optimized performance, thereby providing reliable decision support for the individualized diagnosis and treatment of glioma.

Objective To investigate the temporal changes in pathological damage and macrophage polarization in liver and spleen tissues of mice infected with Angiostrongylus cantonensis, and to preliminarily unravel the peripheral immune responses during the early stage of A. cantonensis infection. Methods Forty female BALB/c mice at ages of 6 to 8 weeks were randomly divided into four groups, including the control group and 7-, 14-, and 21-day infection groups, with 10 mice in each group. Each mouse in the infection groups was inoculated with 30 third-stage (L3) larvae of A. cantonensis by oral gavage, and five mice were randomly selected from each infection group on days 7, 14, and 21 post-infection, while mice in the control group were given the same volume of physiological saline and five mice were randomly selected from the control group on the day of oral gavage. Mouse liver and spleen tissues were sampled. The histopathological changes of mouse liver and spleen tissues were observed using hematoxylin and eosin (HE) staining, and the percentage of positive staining area and the co-localization positive rates of the macrophage surface antigens F4/80, CD86, and CD206 were quantified in mouse liver and spleen tissues using immunohistochemical and immunofluorescence staining. In addition, five mice were collected from each infection group on days 7, 14, and 21 post-infection, and five mice were collected from the control group on the day of oral gavage. Mouse liver and spleen tissues were sampled for detection of macrophage markers CD86 and CD206 and macrophage phenotyping using flow cytometry, and the expression of M1 macrophage markers, including inducible nitric oxide synthase (Nos2), tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and M2 markers, including arginase 1 (Arg1), mannose receptor C-type 1 (Mrc1) and chitinase-like protein 3 (Chil3) was quantified in mouse liver and spleen tissues using real-time quantitative PCR (RT-qPCR) assay. Results Proliferative lesions of the hepatocyte were observed in mouse liver tissues and the follicular structures of the mouse spleen white pulp were disrupted 21 days post-infection with A. cantonensis. Immunohistochemical staining showed that there were significant differences in the percentages of F4/80, CD86 and CD206 positive staining areas in the liver and spleen tissues among the four groups of mice (F = 242.40, 197.14, 183.19, 157.65, 242.35 and 146.24; all P values < 0.001), and the percentages of positive staining in the liver and spleen tissues of mice in the 14-day infection group [(4.45 ± 0.51)%, (3.74 ± 0.67)%, (8.32 ± 0.72)%, (16.56 ± 1.14)%, (11.62 ± 0.52)%, and (8.29 ± 0.72)%, respectively] and the 21-day infection group [(3.70 ± 0.11)%, (3.22 ± 0.43)%, (11.53 ± 1.03)%, (12.59 ± 1.05)%, (9.02 ± 0.83)%, and (11.67 ± 1.10)%, respectively] were higher than in the control group [(0.35 ± 0.16)%, (0.40 ± 0.02)%, (0.93 ± 0.05)%, (2.78 ± 0.26)%, (2.33 ± 0.20)%, and (1.85 ± 0.20)%, respectively] (all P values < 0.05). Immunofluorescence staining showed significant differences in the positive rates of F4/80 co-localization with CD86 and CD206 in mouse liver and spleen tissues among the four groups (F = 24.42, 25.28, 54.51 and 130.55; all P values < 0.001). Flow cytometry detected significant differences in the proportions of CD86⁺ and CD206⁺ macrophages in mouse liver and spleen tissues among the four groups (F = 67.98, 18.41, 29.77, 172.80; all P values < 0.001), and the proportions of CD206⁺ macrophages in the liver and spleen of the 21-day infection group were significantly higher than those in the control group [(9.25 ± 2.55)% vs (3.83 ± 0.72)%, and (4.22 ± 0.56)% vs (0.47 ± 0.18)%, respectively] (both P values < 0.05). In addition, RT-qPCR assay quantified significant differences in the relative mRNA expression of M1 macrophage markers (IL-1β, TNF-α and Nos2) and M2 macrophage markers (Arg1, Chil3 and Mrc1) in mouse liver and spleen tissues among the four groups (F = 41.30, 31.82, 199.33, 19.96, 62.01, 119.76, 23.67, 95.90, 72.27, 82.59, 123.41 and 29.75; all P values < 0.05). Conclusions A. cantonensis infection may cause progressive pathological damage in mouse liver and spleen tissues, accompanied by dynamic temporal changes in macrophage polarization. M1 macrophage polarization predominates at the early stage of A. cantonensis infection and shifts towards M2 polarization at the later stages, suggesting that M2 polarization may participate in immune regulation at late stages of A. cantonensis infection by suppressing excessive inflammatory responses and promoting tissue repair.

[摘 要] 目的：探究半乳糖凝集素-9（Gal-9）对胰腺癌细胞的侵袭与迁移功能的促进作用及其机制。方法：基于公开单细胞转录组数据分析胰腺癌中免疫检查点分子的表达谱，结合Western blotting和ELISA验证Gal-9在胰腺癌细胞及临床相关模型中的表达与分泌特征。建立C57BL/6小鼠胰腺原位移植瘤模型，通过免疫组化、免疫荧光染色和流式细胞术分析Gal-9对肿瘤细胞增殖及其对瘤内CD8+ T细胞浸润与功能状态的影响。Gal-9重组蛋白及肿瘤细胞条件培养基（CM）处理肿瘤相关成纤维细胞（CAF），Western blotting、RT-qPCR和免疫荧光检测CAF中Ⅰ型胶原表达变化。免疫共沉淀、免疫荧光染色和Western blotting分析Gal-9与CD44的相互作用及JAK2/STAT3通路相关蛋白的表达。构建过表达Gal-9胰腺癌细胞与CAF共移植的裸鼠皮下荷瘤模型，结合天狼星红和马松染色评估小鼠肿瘤组织中胶原沉积情况。脱细胞外基质法提取CAF中沉积的细胞外基质（ECM）蛋白；通过细胞划痕、Transwell迁移及侵袭实验，检测CAF来源Ⅰ型胶原对胰腺癌细胞侵袭、迁移功能的影响，并通过siRNA干扰Gal-9表达进行功能反向验证。结果：与经典免疫检查点分子PD-L1和CTLA-4相比，Gal-9在胰腺癌中显著高表达（P < 0.000 1）。体内实验表明，Gal-9促进肿瘤生长（P < 0.001），上调Ki-67表达，同时减少肿瘤内CD8⁺ T细胞浸润并诱导肿瘤组织中TIM-3⁺细胞比例升高，Granzyme B、IFN-γ及TNF-α表达下降（P < 0.05）。体外实验显示，外源性Gal-9及肿瘤细胞来源的CM均可上调CAF中Ⅰ型胶原表达（P < 0.001）。机制研究表明，在CAF中抗Gal-9抗体免疫沉淀复合物中检测到CD44，且JAK2/STAT3蛋白表达水平增加。CAF来源的Ⅰ型胶原处理胰腺癌细胞后，肿瘤细胞的迁移和侵袭数量增加。相反，肿瘤细胞中敲低Gal-9表达后，CAF来源Ⅰ型胶原水平下降，同时肿瘤细胞侵袭和迁移数量下降。结论：胰腺癌来源的Gal-9具有双重作用：一方面，通过促进肿瘤细胞增殖和诱导CD8+ T细胞耗竭促进肿瘤进展；另一方面，通过结合CAF上的CD44受体，激活JAK2/STAT3信号通路，刺激CAF合成并分泌Ⅰ型胶原，进而增强胰腺癌细胞的侵袭迁移能力。

ObjectiveTo investigate the association of the serum levels of Golgi protein 73 （GP73）， hypoxia-inducible factor-1α （HIF-1α）， and matrix metalloproteinase-9 （MMP-9） before and after transcatheter arterial chemoembolization （TACE） with the treatment outcome of patients with hepatitis B virus （HBV）-related hepatocellular carcinoma （HCC）. MethodsA total of 135 patients with HBV-related HCC who received TACE for the first time in The First Affiliated Hospital of Xinjiang Medical University from January 2023 to October 2025 were enrolled as subjects， and the serum levels of GP73， HIF-1α， and MMP-9 were measured before TACE and at 1 month after TACE. According to the treatment outcome after TACE， the patients were divided into remission group and non-remission group. The two groups were compared in terms of general information and the serum levels of GP73， HIF-1α， and MMP-9 before and after treatment. The Logistic regression analysis was used to investigate the influencing factors for the efficacy of TACE， and the receiver operating characteristic （ROC） curve was used to assess the value of the serum levels of GP73， HIF-1α， and MMP-9 before treatment used alone or in combination in predicting the efficacy of TACE. The independent-samples t test was used for comparison of continuous data between two groups， and the chi-square test was used for comparison of categorical data between two groups. ResultsAmong the 135 patients after treatment， 78 （57.78%） were enrolled in the remission group and 57 （42.22%） were enrolled in the non-remission group. Before treatment， the remission group had significantly lower serum levels of GP73， HIF-1α， and MMP-9 than the non-remission group （all P<0.05）； after treatment， both groups had reductions in the serum levels of GP73， HIF-1α， and MMP-9， and the remission group had significantly lower levels than the non-remission group （all P<0.05）. The binary Logistic regression analysis showed that increases in the serum levels of GP73， HIF-1α， and MMP-9 before treatment were independent influencing factors for the efficacy of TACE （all P<0.05）. The ROC curve analysis showed that the serum levels of GP73， HIF-1α， and MMP-9 before treatment used alone had an area under the ROC curve （AUC） of 0.799， 0.835， and 0.777， respectively， in predicting the efficacy of TACE， among which HIF-1α showed the highest predictive value， and the combination of these three indicators had an AUC of 0.950， a Youden index of 0.787， an optimal cut-off value of 0.39， a sensitivity of 87.72%， and a specificity of 91.03%. ConclusionTACE can reduce the serum levels of GP73， HIF-1α， and MMP-9 in patients with HBV-related HCC， and patients with a good outcome tend to have greater reductions. The high serum levels of GP73， HIF-1α， and MMP-9 before treatment are independent risk factors for poor short-term efficacy of TACE， and the combination of these three indicators has a relatively high value in predicting the efficacy of TACE.

Objective:To explore the effects of Yishen Tonglong Granules (YSTLG) on cell proliferation and apoptosis in a nude mouse model of androgen independent prostate cancer subcutaneous transplantation based on non classical Wnt signaling pathway.Methods:The tumor-bearing nude mouse model was established using the human prostate cancer bone metastatic cell line PC-3. After successful modeling, the mice were equally divided into six groups using the random number table method: model group, Western medicine group, Chinese materia medica low-, medium-, and high-dosage groups, and Chinese materia medica-Western medicine combination group. Corresponding drug interventions were administered to each group. Following 28 consecutive days of drug administration, the nude mice were euthanized. Tumor tissues were harvested for pathological observation via HE staining. Cell proliferation was assessed by immunohistochemistry; apoptosis was detected using TUNEL assay; the expressions of non-canonical Wnt signaling pathway-related proteins (Wnt5a, Rac1, RhoA, NFATc1) were analyzed through Western blot and immunohistochemical methods.Results:THE staining results demonstrated that YSTLG could effectively ameliorate pathological alterations in tumor tissues. Compared with the model group, the Chinese materia medica low-, medium-, and high-dosage groups, as well as the Chinese materia medica-Western medicine combination group, exhibited reduced proliferation indices ( P<0.01), elevated apoptosis indices ( P<0.01), down-regulated protein expressions of Wnt5a, Rac1, RhoA, and NFATc1 ( P<0.01), and decreased optical density values of Wnt5a, Rac1, RhoA, and NFATc1 ( P<0.01). These effects displayed a dosage-dependent trend. The Chinese materia medica-Western medicine combination group achieved the most pronounced therapeutic outcomes. Conclusions:YSTLG may exert inhibitory effects on the proliferation of androgen-independent prostate cancer cells and promote apoptosis, possibly through suppression of the non-canonical Wnt signaling pathway. Furthermore, its combination with Wnt signaling inhibitors may exhibit synergistic therapeutic effects.

Clinically,the incidence of portal vein thrombosis(PVT)in patients with cirrhosis can be up to 10％-23％.When PVT is not treated promptly,it may develop to cavernous transformation of the portal vein(CTPV).CTPV can aggravate portal hypertension,accelerate the progression of esophagogastric varices bleeding,refractory ascites,refractory peritonitis,biliary tract diseases,and hepatic insufficiency.At present,noninvasive imaging techniques such as portal vein reconstruction,enhanced CT and ultrasound are mostly used to make the diagnosis and evaluation of CTPV.It is rather difficult to perform portosystemic shunt surgery in patients with CTPV complicated by portal hypertension,which was once regarded as a contraindication for interventional portosystemic shunt procedures.With the improvement of related technologies and surgical instruments,the transjugular intrahepatic portosystemic shunt(TIPS)has become an important treatment for CTPV.This paper aims to make a comprehensive review about the relevant researches concerning the portosystemic shunt surgery in patients with CTPV so as to clarify the importance of TIPS in the treatment of CTPV.

Objective:To investigate the relationship between oxygen consumption (VO 2), carbon dioxide production (VCO 2), and Oxygen Consumption/lactate (VO 2/Lac) with risk of death in patients with severe ARDS. Methods:A retrospective cohort study method was used, and the study subjects were hospitalized for >5 days adult patients with severe ARDS in the central intensive care unit of Henan Provincial People's Hospital from 1 March 2020 to 30 June 2023. The following patients were excluded: IC test was not completed on the 4th day of ICU admission, IC test results were unreliable, mechanical ventilation duration had exceeded 48 h at the time of ICU transfer or admission, palliative care patients and pregnant and parturient women. Using indirect calorimetry to determine VO 2 and VCO 2 values on the 4th day of admission, reviewing medical records to obtain general condition, disease information, blood gas analysis (including lactate value), diagnostic and therapeutic measures, and following up deaths by telephone and time of death. The primary outcome measure was death at 90 days, and the secondary outcome measure was death at 28 days, length of stay in ICU, total length of stay, and total hospitalization cost. Cox regression analysis and linear regression analysis were used to investigate the relationship between VO 2, VCO 2, VO 2/Lac and primary and secondary outcome indexes. Results:A total of 216 patients were enrolled, 78 patients (36.1%) died and 138 patients (63.9%) survived at 90 days. After correction for confounders, the results of multifactorial Cox regression analysis suggested that compared with the Q4 group, HR (95% CI) for 90-day risk of death in the VO 2 Q1 and Q2 groups was 3.21 (1.38, 7.49) and 3.24 (1.42, 7.38), and HR (95% CI) for 90-day risk of death in the VCO 2 Q1, Q2 and Q3 groups was 5.88 (2.33, 14.84), 4.26 (1. 60, 11.34) and 3.54 (1.34, 9.35), respectively, and the HR (95% CI) for 90-day risk of death in the VO 2/Lac Q1, Q2 and Q3 groups were 8.72 (3.01, 25.25), 8.43 (2.91, 24.47) and 4.04 (1.34, 12.17) respectively. P-trends were all <0.05, indicating that VO 2, VCO 2 and VO 2/Lac were linearly and negatively associated with the risk of 90-day mortality. In addition, VO 2, VCO 2, and VO 2/Lac were negatively associated with 28-day risk of death and higher VO 2/Lac was negatively associated with length of ICU stay. Conclusions:VO 2, VCO 2 and VO 2/Lac were negatively associated with 90-day mortality risk and 28-day mortality risk in patients with severe ARDS and may be independent risk factors predicting mortality risk of such patients.