LABORATORIO DE URGENCIAS

Septic shock: Past, Present, and Perspectives

Erika P. Plata-Menchaca;  Ricard Ferrer; Orlando Ruben Perez-Nieto;  Eduardo Kattan; Toshiaki Iba et al,
* Shock, Organ Dysfunction and Resuscitation Group, Intensive Care Department,Vall’d’Hebron  University Hospital, Vall d’Hebron Hospital Campus, Spain  
* Department of Medicine, Universitat Autonoma de Barcelona (Cerdanyola del Valles),Bellaterra, Barcelona, Spain
* The Latin American Intensive Care Network (LIVEN), www.livenresearch.com, Chile
* Intensive Care Unit, Hospital General San Juan del Río, Queretaro, Mexico Universidad Autonoma de Queretaro, Mexico
* Departamento de Medicina Intensiva, Facultad de Medicina, Pontificia Universidad Catolica de Chile, Chile
* Faculty of Medical Science, Juntendo University, Urayasu, Chiba, Japan Intensive Care

                                                                                                                     J Critical Care 91 (2026) 155269

1. Past: Historical evolution and milestones in septic shock Septic shock has long stood as the ultimate complication of infection, recognized by profound circulatory, cellular, and metabolic disturbance with significantly increased mortality [1]. Its roots trace back to early observations of “blood poisoning,” but systematic understanding was limited until the modern era. In the late 20th century, sepsis was rede fined as a syndrome of dysregulated host response, not merely direct microbial invasion, recognized in the first consensus in 1991 [2]. Sub sequent decades delivered pivotal milestones: refinement of definitions (Sepsis-2 in 2001, and Sepsis-3 in 2016 [1]), and the adoption of in ternational guidelines and sepsis bundles to standardize care [3]. The Rivers trial in 2001 introduced early goal-directed therapy (EGDT), spurring widespread attention to rapid resuscitation, antibi otics, and invasive monitoring [4]. The Surviving Sepsis Campaign (SSC), launched in 2002, facilitated global guidelines and large-scale education, resulting in sepsis recognition as a time-critical medical emergency [3]. Through the 2010s, randomized trials such as ProCESS, ARISE,and ProMISe, challenged EGDT’s superiority over usual care, shifting focus toward prompt recognition, bundled care, and individualization [5]. Meanwhile, advances in supportive management, including rational fluid therapy, vasopressor selection (notably norepi nephrine), and modulation of immunologic dysregulation, incremen tally improved outcomes [6]. Despite these gains, septic shock is associated with persistently high mortality, historically ranging between 40 and 50 % in severe cases [7]. Key research milestones furthered understanding of pathophysiology: microcirculatory dysfunction, mitochondrial failure, immune de rangements, and the role of the endothelial glycocalyx [8]. Yet, the heterogeneity of patient populations and underlying infections contribute to treatment limitations.
One therapeutic target in patients with septic shock has been lactate clearance, as elevated lactate levels are commonly observed in the context of hypoperfusion. However, targeting lactate reduction has not consistently translated into improved outcomes, in part because lactate may also be elevated due to non-hypoperfusion mechanisms, such as the adrenergic stress response frequently associated with sepsis [9]. Several therapeutic strategies have also failed to improve outcomes in septic shock. A prominent example is recombinant activated protein C (drotrecogin alfa), which was approved in 2001 following the PROWESS trial that suggested a reduction in mortality. Subsequent systematic re views, however, failed to confirm this benefit and highlighted an increased risk of serious bleeding, ultimately leading to its market withdrawal ten years later [10].
2. Present: Guidelines, current management, and uncertaintie
2.1. Fluid resuscitation The systematic use of fluid responsiveness assessment using dynamic measures to guide fluid administration has been increasingly adopted over static measures. Fluid overload is now recognized to worsen out comes by disrupting the vascular barrier and precipitating organ edema [11]. Individualized targets, with fluid responsiveness testing, are emphasized [12].
2.2. Vasopressor therapy Norepinephrine is first-line, initiated early if mean arterial pressure (MAP) remains
The 2021 SSC guidelines underscore the urgency of septic shock care: early broad-spectrum antibiotics, intravenous crystalloids within three hours, frequent reassessment of hemodynamics, and norepinephrine as the initial vasopressor when shock persists [3]. Serum lactate is measured to gauge severity and guide resuscitation. Key current approaches include: adjunctive agent selection (e.g., vasopressin, angiotensin II) remain uncertain [13,14].
2.3. Infection source control Timely intervention remains critical, with delays directly correlating with mortality [3]. However, timely diagnostics remain major chal lenges, especially in settings with limited resources.
2.4. Adjunctive therapies The use of corticosteroids remains controversial, though current guidelines support hydrocortisone for shock refractory to fluids and vasopressors [15]. Stress ulcer prophylaxis and glycemic control are applied selectively to avoid iatrogenic harm. The role of immunomod ulators and blood purification remains unproven despite encouraging preclinical and early clinical data. Importantly, several large trials of blood purification and other adjunctive interventions have yielded negative or inconclusive results, underscoring the limitations of a “one- size-fits-all” approach. These findings highlight the urgent need for better patient phenotyping and stratification, so that therapies can be targeted to subgroups most likely to benefit. Recent perspectives, further emphasize the heterogeneity of septic shock, the promise of precision medicine, and the need to prioritize research that aligns specific thera pies with defined patient subgroups [16].
2.5. Recognition of heterogeneity Sepsis is increasingly seen as a spectrum syndrome with variable host-pathogen interactions and immune responses [17]. Risk stratifica tion tools guide diagnosis but lack precision for individualized man agement [1].
Areas of uncertainty include the optimal resuscitation endpoint, optimal antibiotic administration and duration, heterogeneity of he modynamic profiles, and the best markers for guiding therapy. Multi plicity in care settings, resource and personnel limitations, and patient factors further compound variability in outcomes.
3. Perspectives: Future directions,
AI and machine learning in septic shock care The next decade is poised for transformative changes in septic shock management, largely driven by machine learning (ML), including the integration of earlier biomarkers, multi-omics, and precision medicine.
3.1. From macrohemodynamics to microcirculatory-guided resuscitation Hemodynamic resuscitation is expected to evolve beyond its tradi tional focus on macrocirculatory targets, such as blood pressure and cardiac output, toward the integration of microcirculatory assessment into routine practice. Techniques such as capillary refill time and sub lingual visualization may enable real-time detection of persistent perfusion deficits that predict organ dysfunction and mortality, even after macrohemodynamic optimization. Furthermore, future research addressing the role of venous congestion as a driver of microcirculatory dysfunction could open further therapeutic alternatives [18]. Incorpo rating these tools into resuscitation algorithms could allow for more precise, individualized interventions, help prevent overtreatment, and ultimately improve outcomes in patients with septic shock [19,20].
3.2. AI and machine learning ML models, leveraging vast electronic health record (EHR) data, already outperform clinical scoring systems in early sepsis and septic shock prediction, with some reporting AUROCs >0.9 [21,22]. These systems are expanding to automate risk stratification, predict hemodynamic deterioration, simulate fluid/vasopressor responsiveness or optimize second-line vasopressors [23]. The integration of structured and unstructured EHR data allows continuous real-time risk assessment, potentially triggering earlier and more personalized interventions. The real-world barriers such as high costs, external validation of machine learning tools, and the training needed for bedside integration are challenging to fully implement these strategies.
3.3. Precision phenotyping and dynamic decision-support ML-aided clustering of patient data is uncovering subphenotypes (“hyperinflammatory,” “immunoparalytic”) that may guide tailored therapy beyond one-size-fits-all approaches [17,24]. Multi-omic ana lyses, such as transcriptomics, metabolomics, proteomics, promise to identify new therapeutic targets for the development of new potential drugs [25]. Bedside ML tools may soon help triage patients, recommend in terventions, and predict complications (e.g., acute kidney injury, acute respiratory distress syndrome - ARDS), continuously updating as new data emerge [22,24].
3.4. Novel therapeutics and biomarkers With the aid of computational modelling, the rapid validation of candidate pathways, biomarkers, and therapeutic targets is becoming increasingly feasible. Systems biology and multi-omics integration (ge nomics, transcriptomics, proteomics, and metabolomics) provide un precedented insights into the host response to infection, enabling the identification of endotypes that may respond differently to specific in terventions. In parallel, advances in machine learning are accelerating the discovery and validation of predictive signatures, supporting the shift toward precision medicine in sepsis. At the clinical level, efforts are underway to develop rapid bedside assays for immune profiling, endothelial function, and organ dysfunc tion, which could enable near–real-time assessment of disease trajec tory. Point-of-care biomarkers are beginning to revolutionize the early diagnosis of sepsis, as well as bedside risk stratification and prognosti cation. Combining these diagnostic tools with AI-driven decision sup port systems holds the promise of guiding timely, individualized therapeutic strategies, ultimately improving patient outcomes [26].
3.5. Implementation science The translation of research advances to bedside practice remains challenging. ML tools will need rigorous external validation, trans parency, and workflow integration to avoid alert fatigue and disparities in care [22]. Preventing and treating the long-term consequences of septic shock in survivors is becoming ever more important as acute mortality de clines.
Approximately half of patients survive septic shock, yet many develop lasting complications including cardiovascular disease, chronic kidney disease, neurocognitive impairment, and functional disability. For instance, Shankar-Hari et al. followed nearly 95,000 sepsis survivors across England and found that age, comorbid burden, and organ dysfunction known at index hospitalization were strongly associated with elevated mortality for up to 5–6 years post-discharge [27].
More recently, Malomo et al. demonstrated that structural cardiac changes (left ventricular dilation, elevated myocardial T1) persist weeks after sepsis in UK survivors, suggesting ongoing risk to cardiovascular health [28].
Addressing these long-term outcomes requires systematic follow-up, better rehabilitation pathways, and integration of chronic disease pre vention into sepsis survivorship care plans. International collaboration and high-quality, longitudinal datasets will be essential to characterise the trajectories of survivorship, identify modifiable risk factors, and evaluate interventions spanning from acute care through extended recovery.
Ultimately, a comprehensive strategy is needed to reduce the full burden of sepsis, both immediately and in the long term.
4. Conclusions: Septic shock management has progressed from inflammation- centered models to evidence-based, protocol-driven approaches that incorporate hemodynamic optimization and early interventions.
Despite these advances, considerable uncertainty persists due to patient het erogeneity and variability in treatment responses. The future of care lies in precision medicine, with artificial intelligence and machine learning offering the potential to individualize treatment, enhance early detec tion and prognostication, and refine therapeutic strategies.
Successfully bridging innovation with clinical implementation will be essential to improving outcomes in septic shock.

References
[1] Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for Sepsis and septic shock (Sepsis-3). JAMA 2016 Feb 23;315(8): 801–10. [2] Bone RC, Balk RA, Cerra FB, et al. Definitions for Sepsis and organ failure and guidelines for the use of innovative therapies in Sepsis. The ACCP/SCCM consensus conference committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992 Jun;101(6):1644–55.

[3] Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis campaign: international guidelines for management of Sepsis and septic shock 2021. Intensive Care Med 2021 Nov;47(11):1181–247.

[4] Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe Sepsis and septic shock. N Engl J Med 2001 Nov 8;345(19):1368–77.

[5] Yealy DM, Kellum JA, Huang DT, et al. A randomized trial of protocol-based care for early septic shock. N Engl J Med 2014 May 1;370(18):1683–93.

[6] Annane D, Vignon P, Renault A, et al. Norepinephrine plus dobutamine versus epinephrine alone for management of septic shock: a randomised trial. Lancet 2007 Aug 25;370(9588):676–84.

[7] Rhee C, Dantes R, Epstein L, et al. Incidence and trends of Sepsis in US hospitals using clinical vs claims data, 2009-2014. JAMA 2017 Oct 3;318(13):1241–9.

[8] Ince C, Mayeux PR, Nguyen T, et al. The endothelium in Sepsis. Shock 2016 Mar;45 (3):259–70.

[9] Marik PE. Lactate guided resuscitation-nothing is more dangerous than conscientious foolishness. J Thorac Dis 2019 Sep;11(Suppl. 15):S1969–s1972.
[10] Martí-Carvajal AJ, Sol`a I, Lathyris D, Cardona AF. Human recombinant activated protein C for severe Sepsis. Cochrane Database Syst Rev 2012 Mar 14;(3): Cd004388. [11] Maitland K, Kiguli S, Opoka RO, et al. Mortality after fluid bolus in African children with severe infection. N Engl J Med 2011 Jun 30;364(26):2483–95.
[12] Monnet X, Lai C, Teboul JL. How I personalize fluid therapy in septic shock? Crit Care 2023 Mar 24;27(1):123.
[13] Hamzaoui O, Shi R. Early norepinephrine use in septic shock. J Thorac Dis 2020 Feb;12(Suppl. 1):S72–s77.

[14] Ham SY, Kim SO, Kim WJ, et al. Early vasopressor use following hypovolemia in septic shock: a systematic review and meta-analysis. J Crit Care 2023;75:154234.

[15] Pirracchio R, Annane D, Waschka AK, et al. Patient-level meta-analysis of low-dose hydrocortisone in adults with septic shock. NEJM Evid 2023 Jun;2(6): EVIDoa2300034. [16] Martin-Loeches I, Singer M, Leone M. Sepsis: key insights, future directions, and immediate goals. A review and expert opinion. Intensive Care Med 2024 Dec;50 (12):2043–9.

[17] Seymour CW, Kennedy JN, Wang S, et al. Derivation, validation, and potential treatment implications of novel clinical phenotypes for Sepsis. JAMA 2019 May 28; 321(20):2003–17.

[18] Kattan E, Castro R, Miralles-Aguiar F, Hern´andez G, Rola P. The emerging concept of fluid tolerance: a position paper. J Crit Care 2022 Oct;71:154070.
[19] Duranteau J, De Backer D, Donadello K, et al. The future of intensive care: the study of the microcirculation will help to guide our therapies. Crit Care 2023 May 16;27(1):190.

[20] Kattan E, Bakker J, Estenssoro E, et al. Hemodynamic phenotype-based, capillary refill time-targeted resuscitation in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial study protocol. Rev Bras Ter Intensiva 2022 Jan-Mar;34 (1):96–106.
[21] Komorowski M, Celi LA, Badawi O, Gordon AC, Faisal AA. The artificial intelligence clinician learns optimal treatment strategies for sepsis in intensive care. Nat Med 2018 Nov;24(11):1716–20.

[22] Shashikumar SP, Stanley MD, Sadiq I, et al. Early sepsis detection in critical care patients using multiscale blood pressure and heart rate dynamics. J Electrocardiol 2017 Nov-Dec;50(6):739–43.

[23] Kalimouttou A, Kennedy JN, Feng J, et al. Optimal vasopressin initiation in septic shock: the OVISS reinforcement learning study. JAMA 2025 May 20;333(19): 1688–98. [24] Calfee CS, Delucchi K, Parsons PE, Thompson BT, Ware LB, Matthay MA. Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials. Lancet Respir Med 2014 Aug;2(8):611–20.

[25] François B, Levy M, Ferrer R, Laterre PF, Angus DC. A mechanism-based prognostic enrichment strategy for the development of the TREM-1 inhibitor nangibotide in septic shock. Intensive Care Med 2025 May;51(5):965–7.

[26] Plata-Menchaca EP, Ruiz-Rodríguez JC, Ferrer R. Early diagnosis of Sepsis: the role of biomarkers and rapid microbiological tests. Semin Respir Crit Care Med 2024 Aug;45(4):479–90.

[27] Shankar-Hari M, Harrison DA, Ferrando-Vivas P, Rubenfeld GD, Rowan K. Risk factors at index hospitalization associated with longer-term mortality in adult Sepsis survivors. JAMA Netw Open 2019 May 3;2(5):e194900.

[28] Malomo S, Oswald T, Stephenson E, et al. Characterisation of post-sepsis cardiomyopathy using cardiovascular magnetic resonance. Diagnostics (Basel) 2025 Apr 14;15(8)