Thiamine and ascorbic acid have been proposed to mitigate the devastating consequences of sepsis and septic shock. To date, randomized controlled trials have failed to demonstrate a benefit of these therapies and heterogeneity of treatment effect is suspected. In this study, we aimed at assessing the heterogeneity of treatment effect of thiamine (B1) and the combination of B1 plus ascorbic acid (AA + B1) in critically ill patients with sepsis. We conducted a bi-centric retrospective cohort study. All adult patients admitted to the ICU with sepsis or septic shock between January 2012 and August 2022 were included. Patient clusters were identified using latent variable analysis based on demographics and physiological variables obtained within 24 h of admission. Within each cluster and using inverse probability weighted Cox models, we compared in-hospital mortality between patients who received standard treatment (control), standard treatment plus B1 (B1 group), and standard treatment plus a combination of thiamine and ascorbic acid (AA + B1 group). A total of 3465 septic patients were included, 2183, 1054 and 228 in the standard, B1 and AA + B1 groups respectively. Five clusters of patients were identified in an unsupervised manner. The “Cluster Severe” included the most severely ill patients, the “Cluster Resp” patients presented with predominantly respiratory failure, the “Cluster Old” included elderly patients with multiple comorbidities, the “Cluster Fit” patients were young, healthy with low severity indices and “Cluster Liver” included patients with predominant liver failure. B1 treatment was associated with different outcomes across the five clusters. It was associated with a lower in-hospital mortality in the “Cluster Severe” and “Cluster Resp”. On the other hand, the combination of thiamine and ascorbic acid was not associated with reduced mortality in any cluster but an increased mortality in”Cluster Old”. These results reinforce the lack of efficacy of the combination of AA + B1 reported in recent trials and even raise concerns about potential harm in older patients with comorbidities. On the contrary, we reported improved ICU survival associated with B1 supplementation in the most severe patients and those with predominant respiratory failure, supporting the need for further trials in this specific population.

中文翻译：

---

硫胺素和抗坏血酸对通过潜在变量分析确定的脓毒症患者集群的不同影响


硫胺素和抗坏血酸已被提议用于减轻脓毒症和感染性休克的破坏性后果。迄今为止，随机对照试验未能证明这些疗法的益处，并且怀疑治疗效果存在异质性。在这项研究中，我们旨在评估硫胺素 （B1） 和 B1 加抗坏血酸 （AA + B1） 联合治疗脓毒症患者治疗效果的异质性。我们进行了一项以双为中心的回顾性队列研究。包括 2012 年 1 月至 2022 年 8 月期间因脓毒症或感染性休克入住 ICU 的所有成年患者。根据入院后 24 小时内获得的人口统计学和生理变量，使用潜在变量分析确定患者聚集性病例。在每个聚类中，使用逆概率加权 Cox 模型，我们比较了接受标准治疗 （对照）、标准治疗加 B1 （B1 组） 和标准治疗加硫胺素和抗坏血酸联合治疗 （AA + B1 组） 的患者的院内死亡率。共纳入脓毒症患者 3465 例，标准组、B1 组和 AA + B1 组分别为 2183 例、1054 例和 228 例。以无监督的方式确定了 5 组患者。“集束重度”包括病情最严重的患者，“集束性反应”患者主要表现为呼吸衰竭，“集群老年”包括患有多种合并症的老年患者，“集束性适合”患者年轻、健康且严重程度指数较低，“集束性肝”包括以肝功能衰竭为主的患者。B1 治疗与 5 个集群的不同结局相关。它与 “Cluster Severe” 和 “Cluster Resp” 的院内死亡率较低相关。 另一方面，硫胺素和抗坏血酸的组合与任何集群的死亡率降低无关，但与“集群老”的死亡率增加有关。这些结果加强了最近试验中报道的 AA + B1 组合缺乏疗效，甚至引发了对患有合并症的老年患者潜在危害的担忧。相反，我们报告了最严重患者和主要呼吸衰竭患者与补充 B1 相关的 ICU 生存率提高，支持在该特定人群中进行进一步试验的必要性。