静脉注射维生素C作为脓毒症或脓毒性休克患者的微量营养素治疗：一项随机对照试验的系统评价与Meta分析的更新。
全文（开放获取） · 中文译文
摘要 目的：维生素C在脓毒症或脓毒性休克患者中的辅助治疗价值仍存在争议。本研究旨在通过荟萃分析评估静脉注射维生素C对脓毒症或脓毒性休克患者的治疗效果和安全性。 方法：在PubMed、Embase和Scopus数据库及其他相关数据库中检索相关文献。主要结局为28天死亡率。次要结局包括血管加压药的使用时间和总剂量、序贯器官衰竭评估（SOFA）评分从基线至72至96小时之间最早报告时间点的变化（ΔSOFA）、住院时间、重症监护病房住院时间、24小时尿量及不良事件发生率。 结果：14项随机对照试验共纳入1,958例患者符合评估条件。荟萃分析显示，与对照组相比，静脉注射维生素C降低了脓毒症或脓毒性休克患者的28天死亡率（RR 0.70；95% CI 0.53至0.93；P = 0.02；I 2 = 52%），基于低质量证据，亚组分析显示该获益由单中心研究驱动，在多中心试验中未观察到。维生素C缩短了血管加压药的使用时间（SMD = -0.42，95% CI -0.77至-0.08，P = 0.02），并与SOFA评分的更大降幅相关（MD = -1.04，95% CI -1.83至-0.25，P = 0.01），但未减少血管加压药剂量、住院时间或24小时尿量。未报告显著不良事件。 结论：这项荟萃分析提示，静脉注射维生素C可能与脓毒症患者死亡率降低相关，但这一获益未在低偏倚风险的多中心试验中得到证实，且主要受单中心研究驱动。静脉注射维生素C似乎是安全的。鉴于这些亚组分析的探索性性质以及所使用的数据驱动阈值，有必要开展高质量的多中心随机对照试验。 系统综述注册：http://www.crd.york.ac.uk/PROSPERO/view/CRD4204261367808，标识符CRD4204261367808。 引言 脓毒症是一种由感染诱发的系统性炎症反应综合征，是重症监护病房（ICU）中发病率和死亡率极高的危重病症。全球每年约有3200万脓毒症患者，死亡人数达530万（）。脓毒症的根本特征在于宿主对感染的反应失调导致的危及生命的器官功能障碍（）。在此过程中，过度的氧化应激在损伤后内皮修复过程中发挥作用，促进对缺氧的适应性反应和细菌清除（）。然而，感染诱导免疫系统激活后大量释放的炎症介质和活性氧（ROS）超出抗氧化防御能力，导致细胞损伤、微循环功能障碍和多器官衰竭（）。脓毒症或脓毒性休克患者的管理通常包括联合治疗方案，包括抗菌药物、液体复苏、血管升压药以及适当时机的机械通气。然而，脓毒性休克的死亡率高达近40%（）。 作为一种强效抗氧化剂，维生素C具有独特的氧化还原特性，可直接中和自由基并保护细胞膜完整性。基于既往研究的证据，维生素C通过恢复血管对血管活性物质的反应性、改善微循环血流和保持内皮功能，发挥潜在的保护作用（–）。此外，维生素C是多种酶促反应的重要辅因子，参与包括胶原合成、儿茶酚胺生成和肉碱生物合成在内的关键生理过程（）。然而，维生素C作为一种人体不能合成或储存的必需水溶性营养素，完全依赖外源性补充以维持维生素C水平。脓毒症患者由于代谢需求增加、摄入减少和丢失过多，常常出现严重的维生素C缺乏。近90%的脓毒症患者存在维生素C不足（定义为<23 μmol/L），约40%符合维生素C缺乏标准（定义为<11 μmol/L）（）。较低的维生素C水平与脓毒症患者器官功能障碍的严重程度密切相关（）。Marik等人的研究（）提出了包括大剂量维生素C联合氢化可的松和硫胺素的“脓毒症鸡尾酒疗法”，并证明该联合方案显著降低了脓毒症患者的死亡率。这项研究引发了全球范围内对维生素C在脓毒症或脓毒性休克中潜在治疗价值的广泛探索。 然而，目前关于脓毒症患者应用维生素C的临床研究结果复杂且相互矛盾。虽然一些研究显示了阳性结果，但一项纳入5,957例患者的系统评价表明，补充维生素C和镁可降低脓毒症患者的短期死亡率（）。但也有研究报道了阴性结果。LOVIT试验显示，与安慰剂相比，大剂量维生素C给药与持续性器官功能障碍甚至死亡的发生率增加相关（）。近期，C-EASIE试验将292例急诊科脓毒症患者随机分配至静脉注射维生素C组或安慰剂组，未发现序贯器官衰竭评估（SOFA）评分有显著改善（）。这种异质性导致维生素C在脓毒症管理中的作用引起了相当大的争议。拯救脓毒症运动发布了一项弱推荐，反对在成人脓毒症或脓毒性休克患者中常规使用静脉注射维生素C（）。与此同时，维生素C的剂量和给药方案存在显著差异，这可能是导致结果不一致的重要原因之一。结合最新临床研究，研究者进行了一项探索性荟萃分析，以重新评估维生素C在脓毒症或脓毒性休克患者中的应用。以28天死亡率、从基线至72 h或96 h的SOFA评分（ΔSOFA）、24-h尿量、住院时间、ICU住院时间、血管加压药的剂量和持续时间，以及 oc 以不良事件发生作为结局指标，本研究比较了不同剂量静脉注射维生素C作为脓毒症或脓毒性休克患者辅助治疗的安全性和疗效。 材料与方法 研究方案与注册 本荟萃分析遵循系统综述和荟萃分析优先报告条目（PRISMA）指南（）进行。研究方案已在PROSPERO平台前瞻性注册（CRD420261367808）。 数据来源与检索策略 系统检索了PubMed、Embase、Scopus和Cochrane图书馆中自数据库建库至2025年6月1日发表文献。为确保文献检索全面，手工筛查了原始研究引用的参考文献。同时通过ClinicalTrials.gov识别未发表的临床试验。检索策略如下：[（脓毒症 OR “脓毒性休克” OR “全身炎症反应综合征” OR SIRS OR “脓毒症综合征”）AND（“维生素C” OR “抗坏血酸” OR 抗坏血酸盐）AND（“治疗” OR “疗法” OR “治疗性” OR “管理”）] AND（“随机对照试验” OR “对照临床试验” OR “随机” OR “安慰剂” OR “临床试验作为主题” OR “随机地” OR “试验”）。 纳入标准 纳入标准如下：（1）研究类型：随机对照试验（RCT）；（2）研究对象：年龄≥18岁、伴有或不伴有休克的脓毒症患者；（3）干预措施：比较任何剂量静脉使用维生素C与标准脓毒症治疗（联合或不联合安慰剂）。对于包含维生素C单药治疗组、联合治疗组和安慰剂组的三臂试验，仅提取维生素C单药治疗组和安慰剂组进行分析；（4）结局指标：28天病死率、SOFA评分从基线至72至96小时之间最早报告时间点的变化（ΔSOFA）、24小时尿量、住院时间、ICU住院时间、血管升压药的剂量和持续时间、维生素C水平和不良事件。排除标准如下：（1）研究对象为18岁以下的儿科脓毒症患者或孕妇；（2）综述类文章；（3）病例报告；（4）动物研究；（5）会议摘要；（6）无法获取全文的研究；（7）使用相同研究数据的重复发表；（8）经口服或鼻胃管途径给予维生素；（9）维生素C与其他疗法（如硫胺素、糖皮质激素或其他抗氧化剂）联合给药的研究，但纳入标准中规定的具有独立维生素C单药治疗组的三臂试验除外，此类试验仅提取单药治疗组和安慰剂组。 研究选择 两位研究者独立根据预设的纳入和排除标准识别并筛选相关文献。筛选过程中出现的任何分歧均通过与第三位研究者仔细商议解决。 数据提取 两位研究者独立提取：第一作者、发表年份、研究设计、样本量、基线特征、维生素C剂量、28天死亡率（或作为替代指标的院内死亡率）、血管升压药使用持续时间及剂量（去甲肾上腺素当量）、ICU及住院时间（天）、SOFA评分（基线至72–96小时）、24小时尿量、血浆维生素C浓度及不良事件。中位数（IQR）数据使用验证公式转换为均数±标准差（）。对于三臂试验，仅提取单药治疗组和安慰剂组。分歧由第三位研究者解决；无法获得的数据记为NA。 偏倚风险 使用Cochrane偏倚风险2.0（RoB 2）工具，由两位评价者独立评估纳入RCT的方法学质量，该工具评估五个领域的偏倚：随机化过程、对预期干预的偏离、结局数据缺失、结局测量以及报告结果的选择。 临床结局 主要结局是28天死亡率。收集文章中报告的28天死亡率；若无法获得，则报告院内死亡率。当死亡率时间范围未明确时，假定为院内死亡率。次要结局包括维生素C的血浆浓度、24小时尿量、住院时间、ICU住院时间、血管升压药的剂量和持续时间，以及SOFA评分的变化。同时分析了与静脉注射维生素C相关的严重不良事件。 亚组分析 维生素C的剂量和患者病情是结果不一致性的主要原因。进行了探索性亚组分析，以识别异质性的潜在来源并为未来研究提出假设。基于关键临床试验结果（–），我们比较了以下干预措施：极高剂量维生素C（≥ 12 g/day，或 ≥ 150 mg/kg/day）；高剂量维生素C（< 12 g，≥ 6 g/day，或 < 150 mg/kg/day，≥ 75 mg/kg/day）；低剂量维生素C（< 6 g/day，或 < 75 mg/kg/day）。患者病情严重程度是治疗反应的主要决定因素，也是异质性的潜在来源。由于在治疗效应修饰的背景下，尚无普遍接受的序贯器官衰竭评估（SOFA）阈值来区分“中度”与“重度”脓毒症，我们采用了数据驱动的方法。所有纳入研究的基线SOFA评分中位数为9.75。平均基线SOFA评分≥10的纳入研究被归入高严重程度亚组（n=6项研究，相当于“严重脓毒症”），而平均基线SOFA评分<10的研究被归入低严重程度亚组（n=7项研究，相当于“普通脓毒症”）。维生素C给药持续时间在各研究间也存在差异。基于纳入研究报告的给药方案，我们将患者分为≤3天组和>3天组进行探索性亚组分析。根据参与机构数量，我们比较了单中心组与多中心组（参与医院≥2家）。证据质量 采用推荐意见分级的评估、制定及评价（GRADE）方法（）评估各结局的证据质量。为了透明和清晰，GRADE系统将证据确定性分为四个等级：高：进一步研究极不可能改变我们对效应估计的置信度。中：进一步研究很可能对效应估计的置信度产生重要影响，并可能改变该估计。低：进一步研究极有可能对效应估计的置信度产生重要影响，并很可能改变该估计。极低：任何效应估计都非常不确定。最后，由GRADEpro GDT生成的GRADE结果汇总表（）提供研究结果摘要及相应的GRADE评级。 统计分析 采用Review Manager（版本5.4）进行数据分析。二分类数据使用相对危险度（RR）及95%置信区间（CI）进行分析，连续数据使用均数差（MD）或标准化均数差（SMD）进行分析。使用I2评估研究结果间的异质性。当p ≥ 0.1且I2 < 50%时，采用固定效应模型；当P < 0.1且I2 ≥ 50%时，采用随机效应模型，并进行敏感性分析以确定异质性来源。组间差异以双侧P < 0.05为差异有统计学意义。 结果 文献检索结果与特征 如 所示，最初确定了约 3,372 篇相关文章。在去除重复和非临床试验文章后，剩余 285 篇文章。经过标题和摘要筛选，保留 41 篇文章。经过全文评估，应用了以下排除标准：(1) 无脓毒症患者的研究（N = 7）；(2) 已撤稿文章（N = 1）；(3) 无法获得关键终点数据的研究（N = 1）；(4) 涉及联合治疗或口服给药途径的研究（N = 18）。最终，共纳入 14 项随机对照试验进行荟萃分析（, , , , –）。 选择流程图。 纳入研究的特点见 所示。所有 14 项纳入研究均为随机对照试验，共涵盖 1,958 例患者。其中，991 例患者被分配到维生素 C 组，968 例分配到安慰剂组 [另有 1 例患者在延迟知情同意下接受随机化，并在获得同意前死亡。研究伦理委员会允许该患者仅纳入主要分析（) ]。纳入研究的样本量范围为 24 至 862 例患者。总体而言，平均年龄范围为 52 至 71 岁，其中 764 名（39.02%）参与者为女性。患者基线平均 SOFA 评分为 8.76。各纳入研究中维生素 C 的给药剂量和持续时间不同，详见 和 。脓毒症、严重脓毒症和脓毒性休克的定义由各纳入研究独立确定。 纳入研究的特点。 &范围, *中位数 (IQR)；#MD (SE)。SOFA，序贯器官衰竭评估。 纳入研究的偏倚风险评估 14项纳入的RCT的方法学质量由两名评审员使用Cochrane偏倚风险2.0（RoB 2）工具独立评估，该工具评估五个领域的偏倚：随机化过程、偏离预期干预、缺失结局数据、结局测量和报告结果的选择。大多数领域的总体偏倚风险较低()。 主要结局的Meta分析 所有14项纳入的试验均报告了28天死亡率或死亡人数(, , , , –)。一项研究()报告了估算的28天死亡率，并从中推断出死亡人数。两项研究(, )包含单独的维生素C高剂量组和低剂量组，且与共同的安慰剂组比较；对于每项研究，我们在纳入主要Meta分析前，将两个维生素C组的死亡率数据合并为一个估计值。 合并数据分析显示，对照组的死亡率为29.44% (285/968)，维生素C组为25.53% (253/991)。与对照组相比，静脉注射维生素C与总体死亡率降低相关（RR 0.70; 95% CI: 0.53至0.93; P = 0.02; I2 = 52%）()。漏斗图显示轻微不对称，左侧研究较多，尽管Egger检验不显著（P = 0.07, ）。 森林图显示维生素C治疗组对死亡率的影响。 亚组分析 维生素C的剂量和持续时间 七项研究(, , , , , , )采用了极高剂量静脉注射维生素C，四项研究(, , , , )采用了高剂量静脉注射维生素C，另有五项研究(, , , , )采用了低剂量静脉注射维生素C。两项研究中极高剂量和低剂量静脉注射维生素C组的数据在其各自亚组内分别进行了分析(, )。 低剂量静脉注射维生素C与总死亡率降低显著相关（RR 0.57；95% CI 0.39至0.83；P = 0.003；I2 = 0%）。相比之下，极高剂量组（RR 0.83；95% CI 0.59至1.18；P = 0.31；I2 = 42%）和高剂量组（RR 0.63；95% CI 0.27至1.48；P = 0.29；I2 = 66%）未显著降低死亡率。三组间的亚组差异检验显示出一致趋势（P = 0.34，I2 = 7.9%）()。维生素C给药的持续时间在各研究间也有所不同。根据纳入研究报告的给药方案，患者被分为≤3天组和>3天组。亚组分析显示，静脉注射维生素C≤3天可降低脓毒症患者的死亡率（RR 0.67；95% CI 0.48至0.95；P = 0.03；I2 = 63%）。然而，亚组内的高异质性提示应谨慎解读这些结果（）。 在纳入的14项研究中，13项报告了总体基线SOFA评分，并被纳入该亚组分析。7项基线SOFA评分<10的研究（, , –, , ）纳入普通脓毒症组，6项基线SOFA评分≥10的研究（, , , , , ）被分入重症脓毒症亚组。一项研究（）未报告总体SOFA评分，被排除。在重症脓毒症亚组中，静脉注射维生素C显示出接近统计学显著性的潜在死亡率获益（RR 0.64; 95% CI 0.41 to 1.01; P = 0.05; I2 = 69%），而普通脓毒症亚组未观察到显著获益（RR 0.82; 95% CI 0.57 to 1.17; P = 0.34; I2 = 11%）。亚组差异无显著性（P = 0.42; I2 = 0%）（）。 参与机构数量 为进一步分析参与医院数量对接受维生素C辅助治疗的脓毒症患者死亡率的影响，进行了一项比较单中心与多中心研究的亚组分析。在纳入的研究中，9项为单中心研究，5项为多中心研究。在单中心亚组中，观察到静脉注射维生素C可显著降低脓毒症患者的28天死亡率（RR 0.54; 95% CI 0.38 to 0.78; P = 0.0009; I2 = 0%）。相比之下，未观察到类似效应。
全文（开放获取） · 英文原文
【Abstract】 OBJECTIVE: The adjunctive therapeutic value of vitamin C in sepsis or septic shock patients remains controversial. This study aimed to evaluate the therapeutic efficacy and safety of intravenous administration of vitamin C in sepsis or septic shock patients through meta-analysis. METHODS: Relevant literature was searched in PubMed, Embase, and Scopus databases and other related databases. The primary outcome was 28-day mortality. Secondary outcomes included duration and total dose of vasopressors, change in sequential organ failure assessment (SOFA) scores from baseline to the earliest reported time point between 72 and 96 h (ΔSOFA), hospital length of stay, intensive care unit length of stay, urine output in the 24-h, and incidence of adverse events. RESULTS: Fourteen randomized controlled trials enrolling 1,958 patients were eligible for assessment. Meta-analysis demonstrated that intravenous vitamin C reduced 28-day mortality in sepsis or septic shock patients compared to the control group (RR 0.70; 95% CI 0.53 to 0.93; P = 0.02; I 2 = 52%), based on low quality of evidence, and subgroup analysis showed the benefit was driven by single-center studies, not observed in multicenter trials. Vitamin C shortened the duration of vasopressor use (SMD = -0.42, 95% CI -0.77 to -0.08, P = 0.02) and was associated with a greater reduction in SOFA score (MD = -1.04, 95% CI -1.83 to -0.25, P = 0.01), but did not reduce vasopressor dose, length of stay, or 24-h urine output. No significant adverse events were reported. CONCLUSION: This meta-analysis suggests that intravenous vitamin C might be associated with reduced mortality in sepsis, but the benefit was not confirmed in low-risk-of-bias multicenter trials and was primarily driven by single-center studies. Intravenous vitamin C appeared safe. Given the exploratory nature of these subgroup analyses and the data-driven thresholds used, high-quality, multicenter RCTs are warranted. SYSTEMATIC REVIEW REGISTRATION: http://www.crd.york.ac.uk/PROSPERO/view/CRD4204261367808, identifier CRD4204261367808. 【Introduction】 Sepsis, a systemic inflammatory response syndrome triggered by infection, represents a critical condition with exceedingly high incidence and mortality rates in intensive care units (ICUs). Sepsis affects approximately 32 million patients globally each year, with fatalities reaching 5.3 million (). Sepsis is fundamentally characterized by life-threatening organ dysfunction resulting from the host dysregulated response to infection (). During this process, excessive oxidative stress plays a role in postinjury endothelial repair processes, promoting adaptive responses to hypoxia and bacterial clearance (). However, massive release of inflammatory mediators and reactive oxygen species (ROS) after infection-induced immune system activation overwhelms antioxidant defenses, leading to cellular damage, microcirculatory dysfunction, and multiple organ failure (). The management of patients with sepsis or septic shock typically includes a combination treatment regimen comprising of antimicrobials, fluid resuscitation, vasopressors, and mechanical ventilation when appropriate. However, the mortality of septic shock reaches up to nearly 40% (). As a potent antioxidant, vitamin C possesses unique redox properties that directly neutralize free radicals and protect cellular membrane integrity. Given the evidence from previous studies, vitamin C exerts potential protective effects by restoring vascular responsiveness to vasoactive substances, improving microcirculatory blood flow, and preserving endothelial function (–). Furthermore, vitamin C serves as an essential cofactor for multiple enzymatic reactions, participating in critical physiological processes including collagen synthesis, catecholamine production, and carnitine biosynthesis (). However, vitamin C as an essential water-soluble nutrient which cannot be synthesized or stored by humans, entirely relies on exogenous supplementation to maintain vitamin C levels. Sepsis patients frequently develop severe vitamin C deficiency due to increased metabolic demands, reduced intake, and excessive losses. Nearly 90% of sepsis patients with hypovitaminosis C (defined as < 23 μmol/L), and approximately 40% meeting criteria for vitamin C deficiency (defined as < 11 μmol/L) (). Lower vitamin C levels are closely correlated with the severity of organ dysfunction in sepsis patients (). The study by Marik et al. () proposed the “sepsis cocktail therapy” including high-dose vitamin C combined with hydrocortisone and thiamine demonstrated that this combination regimen significantly reduced mortality in sepsis patients. This research sparked extensive worldwide exploration into the potential therapeutic value of vitamin C in sepsis or septic shock. However, current clinical studies on vitamin C application in sepsis patients present complex and contradictory outcomes. While some studies demonstrated positive outcomes, a systematic review encompassing 5,957 patients indicated that supplementation with vitamin C and magnesium could reduce short-term mortality in sepsis patients (). However, negative findings have also been reported. The LOVIT trial revealed that high-dose vitamin C administration was associated with increased incidence of persistent organ dysfunction or even mortality compared with placebo (). More recently, the C-EASIE trial randomized 292 emergency department patients with sepsis to intravenous vitamin C or placebo, finding no significant improvement in sequential organ failure assessment (SOFA) scores (). Such heterogeneity has resulted in considerable controversy regarding the role of vitamin C in sepsis management. The Surviving Sepsis Campaign issued a weak recommendation against the routine use of intravenous vitamin C in adults with sepsis or septic shock (). Meanwhile, significant variations exist in the dosage and administration regimens of vitamin C, which may be one of the important reasons for the inconsistent results. Incorporating the latest clinical research, the investigators conducted an exploratory meta-analysis to re-evaluate the application of vitamin C in sepsis or septic shock patients. With 28-day mortality rate, SOFA scores from baseline to 72 h or 96 h (ΔSOFA), urine output in the 24-h, hospital length of stay, ICU length of stay, the dosage and duration of vasopressors, and occurrence of adverse events as outcomes, this study compared the safety and therapeutic efficacy of different dosages of intravenous vitamin C as adjunctive therapy for patients with sepsis or septic shock. 【Materials and methods】 【Study protocol and registration】 This meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (). The protocol was prospectively registered on PROSPERO (CRD420261367808). 【Data sources and search strategy】 Literature published from database inception to June 1, 2025, was systematically searched across PubMed, Embase, Scopus, and the Cochrane Library. To ensure comprehensive literature retrieval, references cited in the original studies were manually screened. Unpublished clinical trials were concurrently identified through ClinicalTrials.gov. The search strategy was as follows: [(sepsis OR “septic shock” OR “systemic inflammatory response syndrome” OR SIRS OR “sepsis syndrome”) AND (“vitamin C” OR “ascorbic acid” OR ascorbate) AND (“treatment” OR “therapy” OR “therapeutic” OR “management”)] AND (“randomized controlled trial” OR “controlled clinical trial” OR “randomized” OR “placebo” OR “clinical trials as topic” OR “randomly” OR “trial”). 【Inclusion criteria】 The inclusion criteria were as follows: (1) Study type: Randomized controlled trials (RCTs); (2) Participants: Sepsis patients aged ≥ 18 years old with or without shock; (3) Interventions: compared the intravenous use of vitamin C at any dosage against standard sepsis care with or without placebo. For three-arm trials that included a vitamin C monotherapy arm, a combination therapy arm, and a placebo arm, only the vitamin C monotherapy and placebo arms were extracted for analysis; (4) Outcome measures: 28-day mortality rate, change in SOFA score from baseline to the earliest reported time point between 72 and 96 h (ΔSOFA), urine output in the 24-h, hospital length of stay, ICU length of stay, the dosage and duration of vasopressors, the level of vitamin C and adverse events. Exclusion criteria were as follows: (1) Study subjects were pediatric sepsis patients under 18 years old or pregnant women; (2) Review articles; (3) Case reports; (4) Animal studies; (5) Conference abstracts; (6) Studies with inaccessible full texts; (7) Duplicate publications using identical research data; (8) Administration of vitamins via oral route or nasogastric tube; (9) Studies in which vitamin C was administered in combination with other therapies (e.g., thiamine, corticosteroids, or other antioxidants), except for three-arm trials with a separate vitamin C monotherapy arm as specified in the Inclusion criteria, from which only the monotherapy and placebo arms were extracted. 【Study selection】 Two investigators independently identified and screened relevant literature according to the predefined inclusion and exclusion criteria. Any disagreements arising during the screening process were resolved through careful deliberation with a third investigator. 【Data extraction】 Two investigators independently extracted: first author, publication year, study design, sample size, baseline characteristics, vitamin C dosage, 28-day mortality (or in-hospital mortality as surrogate), vasopressor duration and dosage (norepinephrine equivalents), ICU and hospital length of stay (days), SOFA score (baseline to 72–96 h), 24-h urine output, plasma vitamin C concentration, and adverse events. Median (IQR) data were converted to mean ± SD using validated formulas (). For three-arm trials, only monotherapy and placebo arms were extracted. Disagreements were resolved by a third investigator; unavailable data were recorded as NA. 【Risk of bias】 The methodological quality of included RCTs was assessed independently by two reviewers using the Cochrane Risk of Bias 2.0 (RoB 2) tool, which evaluates bias across five domains: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. 【Clinical outcomes】 The primary outcome was 28-day mortality. Collected 28-day mortality reported in the articles; if unavailable, report in-hospital mortality. When the mortality timeframe was unspecified, in-hospital mortality was presumed. Secondary outcomes included plasma concentration of vitamin C, urine output in the 24-h, hospital length of stay, ICU length of stay, the dose and duration of vasopressor, and changes in SOFA score. Serious adverse events related to intravenous vitamin C administration were simultaneously analyzed. 【Subgroup analysis】 The dosage of vitamin C and the conditions of patients are major causes of outcome inconsistency. Exploratory subgroup analysis was conducted to identify potential sources of heterogeneity and generate hypotheses for future research. Based on key clinical trial findings (–), we compared the following interventions: very high dose vitamin C (≥ 12 g/day, or ≥ 150 mg/kg/day); high dose vitamin C (< 12 g, ≥ 6 g/day, or < 150 mg/kg/day, ≥ 75 mg/kg/day); low dose vitamin C (< 6 g/day, or < 75 mg/kg/day). The severity of patient illness is a major determinant of treatment response and a potential source of heterogeneity. Since no universally accepted SOFA threshold exists for distinguishing “moderate” from “severe” sepsis in the context of treatment effect modification, we adopted a data-driven approach. The median baseline SOFA score across all included studies was 9.75. Included studies with a mean baseline SOFA score ≥ 10 were classified into the higher-severity subgroup (n = 6 studies, analogous to “severe sepsis”), while those with mean baseline SOFA < 10 were classified into the lower-severity subgroup (n = 7 studies, analogous to “general sepsis”). The duration of vitamin C administration also varied across studies. Based on the administration regimens reported in the included studies, we categorized patients into a ≤ 3-day group and a > 3-day group for exploratory subgroup analysis. According to the number of participating institutions, we compared the single-center group with multi-center group (participating hospitals ≥ 2). 【Quality of evidence】 The quality of evidence for the outcomes was determined using the grading of recommendations assessment, development and evaluation (GRADE) approach (). To achieve transparency and clarity, the GRADE system classifies the certainty of evidence in one of four grades: High: Further research is very unlikely to change our confidence in the estimate of effect. Moderate: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low: Any estimate of effect is very uncertain. Finally, the GRADE results summary table generated by GRADEpro GDT presents a summary of the study findings along with the corresponding GRADE ratings (). 【Statistical analysis】 Data analysis was performed using Review Manager (Version 5.4). Dichotomous data were analyzed using relative risk (RR) with 95% confidence interval (CI), while continuous data were analyzed using mean difference (MD) or standardized mean difference (SMD). Heterogeneity among study results was assessed using the I2. When p ≥ 0.1 and I2 < 50%, a fixed-effects model was employed; When P < 0.1 and I2 ≥ 50%, a random-effects model was applied, and sensitivity analysis was conducted to identify sources of heterogeneity. A two-sided P < 0.05 was considered statistically significant between groups. 【Results】 【Literature search results and characteristics】 As shown in , about 3,372 relevant articles were initially identified. After removing duplicates and non-clinical trial articles, 285 articles remained. Following title and abstract screening, 41 articles were retained. After full-text assessment, the following exclusions were applied: (1) studies without sepsis patients (N = 7), (2) retracted articles (N = 1), (3) studies with unavailable critical endpoint data (N = 1), (4) studies involving combination therapies or oral administration routes (N = 18). Ultimately, 14 randomized controlled trials were included for meta-analysis (, , , , –). The selection flow diagram. The characteristics of included studies are presented in . All 14 included studies were randomized controlled trials, encompassing a total of 1,958 patients. Among them, 991 patients were allocated to the vitamin C group and 968 to the Placebo group [One additional patient underwent randomization under deferred consent and died before consent was obtained. The research ethics board allowed the inclusion of this patient only in the primary analysis ()]. The sample sizes of the included studies ranged from 24 to 862 patients. Overall, the mean age ranged between 52 and 71 years, and 764 (39.02%) participants were female. The mean SOFA score at baseline among patients was 8.76. The dose and duration of vitamin C administration differed across the included studies, as detailed in and . The definitions of sepsis, severe sepsis, and septic shock were independently established in each included study. Characteristics of included studies. &Range, *Median (IQR); #MD (SE). SOFA, sequential organ failure assessment. 【Risk of bias assessment for included studies】 The methodological quality of the 14 included RCTs was assessed independently by two reviewers using the Cochrane Risk of Bias 2.0 (RoB 2) tool, which evaluates bias across five domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. The overall risk of bias was low across most domains (). 【Meta-analysis of primary outcomes】 All 14 included trials reported either 28-day mortality rates or death counts (, , , , –). One study () reported the estimated 28-day mortality rate, from which the number of deaths was inferred. Two studies (, ) included separate high-dose and low-dose vitamin C arms versus a common placebo group; for each study, we pooled the mortality data from both vitamin C arms into a single estimate before inclusion in the primary meta-analysis. Pooled data analysis revealed a mortality rate of 29.44% (285/968) in the control group versus 25.53% (253/991) in the vitamin C group. Compared with the control group, intravenous vitamin C administration was associated with reduced overall mortality (RR 0.70; 95% CI: 0.53 to 0.93; P = 0.02; I2 = 52%) (). The funnel plot showed slight asymmetry, with more studies on the left side, although Egger’s test was not significant (P = 0.07, ). Forest plot displaying the effect of vitamin C treatment group on the mortality. 【Subgroup analysis】 【Dose and duration of vitamin C】 Seven studies (, , , , , , ) employed very high-dose intravenous vitamin C administration, four studies (, , , ) utilized high-dose intravenous vitamin C administration, and another five studies (, , , , ) adopted low-dose intravenous vitamin C administration. Data from the very high-dose and low-dose intravenous vitamin C groups in two studies were analyzed separately within their respective subgroups (, ). Low-dose intravenous vitamin C administration was significantly associated with reduced overall mortality (RR 0.57; 95% CI 0.39 to 0.83; P = 0.003; I2 = 0%). In contrast, the very high dose group (RR 0.83; 95% CI 0.59 to 1.18; P = 0.31; I2 = 42%) and the high-dose group (RR 0.63; 95% CI 0.27 to 1.48; P = 0.29; I2 = 66%) did not significantly reduce mortality. Subgroup difference testing among the three groups revealed a consistent trend (P = 0.34, I2 = 7.9%) (). The duration of vitamin C administration also varied across studies. Based on the administration regimens reported in the included studies, patients were categorized into a ≤ 3-day group and a > 3-day group. Subgroup analysis indicated that intravenous vitamin C administration for ≤ 3 days reduced mortality in patients with sepsis (RR 0.67; 95% CI 0.48 to 0.95; P = 0.03; I2 = 63%). However, the high heterogeneity within the subgroup warrants cautious interpretation of these results (). 【SOFA score】 Of the 14 included studies, 13 reported overall baseline SOFA scores and were included in this subgroup analysis. Seven studies (, , –, , ) with baseline SOFA scores < 10 were included in the general sepsis group, six studies (, , , , , ) with baseline SOFA scores ≥ 10 were assigned to the severe sepsis subgroup. One study () did not report an overall SOFA score and was excluded. In the severe sepsis subgroup, intravenous vitamin C showed a potential mortality benefit that approached statistical significance (RR 0.64; 95% CI 0.41 to 1.01; P = 0.05; I2 = 69%), whereas no significant benefit was observed in the general sepsis subgroup (RR 0.82; 95% CI 0.57 to 1.17; P = 0.34; I2 = 11%). The subgroup difference was not significant (P = 0.42; I2 = 0%) (). 【The number of participating institutions】 To further analyze the impact of the number of participating hospitals on the mortality of sepsis patients receiving adjunctive vitamin C therapy, a subgroup analysis was conducted comparing single-center and multi-center studies. Among the studies included, nine studies were single-center, and five studies involved multi-center. In the single-center subgroup, intravenous vitamin C administration was observed to significantly reduce 28-day mortality in sepsis patients (RR 0.54; 95% CI 0.38 to 0.78; P = 0.0009; I2 = 0%). In contrast, no similar effect was obse