Current and emerging therapeutic options for refractory septic shock: A systematic review

Abstract

BACKGROUND

Refractory septic shock is a critical and multifaceted condition that continues to pose significant challenges in critical care.

AIM

To systematically review randomized trials on emerging interventions for refractory septic shock, assessing mortality, vasopressor use, intensive care unit (ICU) length of stay, and organ dysfunction.

METHODS

A systematic search was conducted in PubMed, EMBASE, Cochrane CENTRAL Library, and Web of Science for studies published between 2000 and 2024. Inclusion criteria encompassed randomized controlled trials (RCT) evaluating innovative therapies for refractory septic shock. Variables of interest: The primary outcome was all-cause mortality among patients treated with novel interventions. Secondary outcomes included length of stay in the ICU, total hospital length of stay, and use of vasoactive drugs. Methodological rigor was assessed using the Cochrane Risk of Bias tool.

RESULTS

From 850 records, 24 RCTs met the inclusion criteria, evaluating therapies such as methylene blue, vasopressin, terlipressin, and combinations of hydrocortisone, vitamin C, and thiamine. Mortality rates ranged from 28.6% to 56.8%. Methylene blue reduced vasopressor dependency in patients requiring high norepinephrine doses by 1.0 vasopressor-free day, and terlipressin improved renal perfusion by 13.1%. Combination therapies enhanced secondary outcomes, including reductions in Sequential Organ Failure Assessment score. However, no single intervention consistently demonstrated significant survival benefits.

CONCLUSION

Adjunctive therapies for refractory septic shock may improve hemodynamics and organ function, however, they have not been shown to consistently reduce mortality. Larger trials are needed to confirm these findings. Multimodal approaches targeting inflammation are critical.

Key Words: Refractory septic shock; Vasopressors; Adjunctive therapies; Sepsis; Mortality outcome

Core Tip: This systematic review of 24 randomized controlled trials evaluates novel therapies for refractory septic shock, including methylene blue, vasopressin, terlipressin, and hydrocortisone-vitamin C-thiamine combinations. While these interventions improve hemodynamics and organ function, no consistent mortality reduction was observed. Methylene blue reduced vasopressor dependency, and terlipressin enhanced renal perfusion. Larger, standardized trials are needed to validate findings and guide multimodal treatment strategies.

INTRODUCTION

Refractory septic shock is a critical and multifaceted condition that continues to pose significant challenges in critical care, with mortality rates ranging from 28.6% to 56.8%[1-3], despite advancements in medical understanding and therapeutic strategies. Although definitions vary across studies and expert consensus[4,5], septic shock is broadly characterized by hypotension (systolic blood pressure ≤ 90 mmHg or mean arterial pressure (MAP) ≤ 65 mmHg) and elevated serum lactate levels (> 2 mmol/L) persisting despite adequate fluid resuscitation. Patients often require high doses of vasopressors - such as norepinephrine (NE) at rates exceeding 0.5-1.0 μg/kg/minute, or equivalent agents - to achieve hemodynamic stability[6,7].

The pathophysiology of refractory septic shock is characterized by profound vasomotor dysfunction and vascular hyporesponsiveness to catecholamines, driven by a complex interplay of mechanisms, including excessive inflammatory responses, endothelial dysfunction, activation of coagulation pathways, and dysregulation of vasodilatory mediators[2,8-12]. These interconnected processes result in severe and sustained hypotension, complicating management and rendering standard therapeutic approaches insufficient in many cases[12]. Amid persistent challenges in improving patient outcomes with current protocols, the exploration of novel therapeutic interventions remains critical.

MATERIALS AND METHODS

This systematic review aims to synthesize evidence on innovative adjunctive treatments for refractory septic shock, including methylene blue, vasopressin, and combinations of hydrocortisone, vitamin C, and thiamine. By assessing their effects on mortality and secondary outcomes—such as vasopressor dependency, intensive care unit (ICU) length of stay, and organ dysfunction—this review provides a comprehensive evaluation of current data and highlights potential directions for future research.

This systematic review was conducted in accordance with the PRISMA 2020 guidelines to ensure rigor, transparency, and reproducibility. Its primary objective was to synthesize the existing literature on the effectiveness of novel therapeutic interventions for refractory septic shock, with a specific focus on their impact on mortality and other critical clinical outcomes. Although the review protocol was not registered, it adhered to predefined inclusion and exclusion criteria to ensure methodological transparency (Figure 1).

Figure 1  Flow chart: Study selection.

Search strategy

A comprehensive search strategy was developed to identify relevant studies. The search terms included: (“septic” OR “sepsis” OR “infection”) AND (“shock” OR “Norepinephrine” OR “high dose” OR “refractory”). The strategy targeted literature published from 2000 onwards to encompass the latest advancements in septic shock management. Searches were conducted across the following databases: PubMed, EMBASE, Cochrane CENTRAL Library, and Web of Science.

Each database was searched using tailored queries, and results were compiled and updated through 2024. The search strategy aimed to ensure inclusivity while maintaining a focus on clinically relevant evidence. These databases were selected for their comprehensive coverage of clinical trials, systematic reviews, and high-impact medical literature relevant to critical care and septic shock.

Inclusion and exclusion criteria

Studies were eligible for inclusion if they were RCTs published in peer-reviewed journals between January 2000 and December 2024, available in English with full-text accessibility. The target population comprised adult patients (≥ 18 years) with septic shock requiring vasopressor support despite adequate fluid resuscitation, defined by MAP ≤ 65 mmHg or systolic blood pressure ≤ 90 mmHg, with evidence of tissue hypoperfusion such as elevated lactate > 2 mmol/L or other markers of organ dysfunction. Eligible interventions included novel or adjunctive therapeutic approaches for refractory septic shock, encompassing methylene blue, vasopressin, terlipressin, corticosteroids, vitamin C, thiamine, or combination therapies administered in addition to standard care consisting of fluid resuscitation, antibiotics, source control, and conventional vasopressors. Studies required appropriate comparator groups receiving placebo, standard care, or alternative active interventions. They must have reported primary outcomes of all-cause mortality at any time point (28-day, 90-day, hospital, or ICU mortality) or secondary outcomes including vasopressor-free days, ICU length of stay, hospital length of stay, Sequential Organ Failure Assessment (SOFA) score, shock reversal, or hemodynamic parameters.

Studies were excluded if they employed non-randomized designs such as observational studies, case series, case reports, systematic reviews, meta-analyses, narrative reviews, conference abstracts, editorials, letters to the editor, or lacked appropriate control groups. Population exclusions comprised pediatric patients (< 18 years), cardiogenic, hypovolemic, or obstructive shock without septic component, normotensive patients with sepsis without shock, and animal or in vitro studies. Intervention exclusions included preventive measures administered before sepsis onset, interventions focused solely on infection treatment without hemodynamic support, and studies evaluating only standard care interventions such as conventional antibiotics, fluids, or first-line vasopressors as monotherapy. Studies were also excluded if they failed to report mortality or relevant clinical outcomes, focused exclusively on laboratory parameters without clinical endpoints, had follow-up periods < 24 hours, represented duplicate publications or multiple reports of the same study population, contained insufficient data for extraction despite author contact, were retracted publications, published in languages other than English, constituted gray literature, theses, or unpublished manuscripts, or demonstrated significant methodological flaws that could not be adequately assessed using standard risk of bias tools.

While recognizing the variability in definitions across studies, we included investigations that defined refractory septic shock as persistent hypotension despite adequate fluid resuscitation (typically ≥ 30 mL/kg crystalloids), vasopressor requirements (normally NE ≥ 0.1 μg/kg/minute, though thresholds varied among studies), and ongoing evidence of tissue hypoperfusion. When multiple publications reported on the same patient cohort, we included the publication with the most comprehensive outcome data or the most extended follow-up period. We also extracted supplementary information from related publications as appropriate to ensure thorough data capture without duplication.

Screening and selection process

Following the database searches, all retrieved articles were imported into the Rayyan^© platform (https://www.rayyan.ai), a software designed to facilitate systematic review workflows. Two independent reviewers (JMSJ and FBH) screened titles and abstracts while remaining blinded to each other's decisions. Full texts of potentially eligible studies were subsequently reviewed to determine final inclusion. Discrepancies were resolved through consensus meetings, ensuring that all inclusion and exclusion decisions were jointly validated. After consensus was reached on included studies, a final list of studies was compiled and independently verified for adherence to the inclusion criteria. Duplicate records across databases were identified and removed using the Rayyan^© deduplication feature, followed by manual verification.

Quality assessment and risk of bias evaluation

The methodological quality of included studies was evaluated using the Cochrane Risk of Bias 2 (RoB 2) tool[13-15]. This assessment covered the following domains: Randomization process: Most studies ensured proper randomization and allocation concealment. Deviations from Intended Interventions: Blinding minimized performance bias in many studies, although open-label trials raised some concerns. Missing outcome data: Studies with incomplete outcome data were flagged; however, most demonstrated low attrition bias. Measurement of outcomes: Objective outcomes, such as mortality, were reliably reported, whereas subjective outcomes raised concerns in non-blinded studies. Selection of Reported Results: Pre-registration ensured transparency; however, some studies exhibited selective outcome reporting. Overall, most studies were rated as "low risk" or "some concerns", supporting the robustness of the synthesized evidence (Table 1)[15-29]. The heterogeneity among studies was assessed using qualitative synthesis and, where appropriate, quantitative metrics such as the I² statistic.

Table 1 Risk of bias assessment.

Ref.	Randomization process	Deviations from intended interventions	Missing outcome data	Measurement of outcomes	Selection of the reported results	Overall risk of bias
Lyu et al[15], 2022	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Ibarra-Estrada et al[16], 2023	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Hajjar et al[17], 2019	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Moskowitz et al[18], 2020	Low risk	Some concerns	Low risk	Low risk	Some concerns	Some concerns
Fujii et al[19], 2020	Low risk	Low risk	Low risk	Some concerns	Some concerns	Some concerns
Sevransky et al[20], 2021	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Wang et al[21], 2022	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns
Cherukuri et al[22], 2019	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Douglas et al[23], 2020	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Morelli et al[24], 2008	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns
Albanèse et al[25], 2005	Low risk	High Risk	Some concerns	High Risk	High Risk	High Risk
Hyvernat et al[26], 2016	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Annane et al[27], 2002	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Liu et al[28], 2018	Some concerns	Low risk	Some concerns	Some concerns	Some concerns	Some concerns
Xiao et al[29], 2016	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns	Some concerns

Data extraction

Data extraction was performed independently by two reviewers to enhance accuracy and reliability. A third researcher resolved any discrepancies through discussion. Key data extracted included study characteristics (design, population, and interventions) as well as primary and secondary outcomes (mortality, ICU length of stay, total hospital stay, and vasopressor-free-days). All data were recorded using a standardized form containing fields for study details, interventions, outcomes, and risk of bias ratings.

Outcomes

The primary outcome was all-cause mortality among septic shock patients treated with novel therapies. Secondary outcomes included length of stay in the ICU, total hospital stays, and vasoactive drug use.

These outcomes were selected to provide a comprehensive assessment of the clinical efficacy of the interventions and their impact on patient management. For studies with incomplete outcome data, sensitivity analyses were planned to assess the impact of missing data on the overall conclusions.

RESULTS

Study selection

A comprehensive search across multiple databases initially identified 850 records. After removing 50 duplicates, 750 articles were screened based on titles and abstracts. Of these, 50 full-text articles were assessed for eligibility, with 26 excluded for not meeting the predefined inclusion criteria. Ultimately, 24 studies were included in the final analysis. The study selection process is illustrated in the PRISMA flow diagram (Figure 1).

Characteristics of included studies

The 24 studies included in this meta-analysis comprised a combination of large multicenter RCTs and smaller-scale investigations (Table 2).

Table 2 Study description.

Ref.	Type of study	Type of patients	Primary outcomes	Number of patients	Mean age	Mortality
Lyu et al[15], 2022	RCT, single-center, double-blind	Adult patients with septic shock	Mortality at 90 days	Intervention Group: 213 patients received a combination of hydrocortisone, vitamin C, and thiamine. Control Group: 213 patients received 0.9% saline	64	90-day Mortality: 40.4% in the intervention group vs 39.0% in the placebo group. 28-day Mortality: 37.1% in the intervention group vs 36.2% in the placebo group
Ibarra-Estrada et al[16], 2023	RCT, double-blind	Adults Diagnosed with septic shock due to suspected or confirmed infection, requiring NE to maintain a MAP ≥ 65 mmHg and serum lactate > 2 mmol/L after adequate fluid resuscitation	28-day mortality	Methylene blue group: 45 patients received methylene blue as an adjunctive therapy. Placebo group: 46	46-47	28-mortality: MB 33% and P 46%
Hajjar et al[17], 2019	RCT, double-blind	Adults (≥ 18 years) with cancer admitted to the ICU, showing documented or strong clinical suspicion of infection and meeting at least two criteria of the systemic inflammatory response syndrome	28-day all-cause mortality	Vasopressin group: 125 patients NE group: 125 patients	NR	28-day mortality: Vasopressin group: 56.8%. NE group: 52.8%. 90-day mortality: Vasopressin group: 72.0%. NE group: 75.2%
Moskowitz et al[18], 2020	RCT, multicenter, double-blind	Adult patients with septic shock with infusion vasopressors	Change in SOFA score	Intervention: 100 patients received a combination of vitamin C, hydrocortisone, and thiamine. Control: 100 patients received a placebo	68	30-day mortality: Intervention group: 34.7%, placebo group: 29.3%
Fujii et al[19], 2020	RCT, multicenter, open label	Adults with documented infection, increase of at least 2 points in the SOFA score. Lactate levels greater than 2 mmol/L. Vasopressor dependency for at least 2 hours prior to enrollment	Time alive and free from vasopressors	Intervention Group: 109 patients received intravenous vitamin C, hydrocortisone, and thiamine. Control group: 107 patients received intravenous hydrocortisone alone	61.7	Intervention group: 28.6% mortality. Control group: 24.5% mortality
Sevransky et al[20], 2021	RCT, double-blind, adaptive	Adults with sepsis requiring mechanical ventilation or high-flow oxygen support due to respiratory failure. Continuous vasopressor support to maintain adequate blood pressure	days free of vasopressors and mechanical ventilation in the first 30 days	Treatment group: 1000 patients received combination therapy with vitamin C, thiamine, and corticosteroids. Control group: 1000 patients received matching placebos for each active agent	65	Interrupted before ending
Wang et al[21], 2022	Pilot RCT	Adults (≥ 18 years) with septic shock. NE dose ≥ 15 µg/minute	Renal perfusion: Assessed through changes in urine output and renal function parameters	22 patients: TP group: 10 patients. Usual care group: 12 patients	62.5	27.3% for the TP group and 41.7% for the usual care group
Cherukuri et al[22], 2019	RCT, double-blind	Adult patients with sepsis or septic shock	Length of ICU stay: Days on ventilator: Days on intravenous blood pressure support: 28-day mortality rate	Intervention: 32 patients received 100000 IU of vitamin A intramuscularly daily for 7 days. Control: 32 patients received a blinded placebo	51	28-Day Mortality Rates. Vitamin A group: 34%. Placebo group: 28%
Douglas et al[23], 2020	RCT, multicenter	Adults who presented with signs of sepsis and hypotension (MAP ≤ 65 mmHg) after receiving between 1 L and 3 L of fluids	Positive fluid balance at 72 hours or at ICU discharge, whichever occurred first	83 patients in the intervention group and 41 patients in the usual care group	62	Intervention group: 20.5%. Control group: 24.4%
Morelli et al[24], 2008	RCT	Adult patients with septic shock	The mortality rate is 28 days after treatment initiation. The outcome aimed to assess the effectiveness of combined dobutamine and TP treatment in improving survival rates in patients with septic shock	60	65	TG (dobutamine and TP): What is the primary and secondary outcome. 40% mortality. CG (standard treatment): 60% mortality
Albanèse et al[25], 2005	RCT, open-label study	Diagnosed with hyperdynamic septic shock after fluid resuscitation-Exhibited hemodynamic instability (MAP ≤ 60 mmHg). Had two or more organ dysfunctions	MAP achieved in patients with hyperdynamic septic shock after treatment with either NE or TP	20 patients, NE group: 10 patients, TP group: 10 patients	65	30% among the patients with hyperdynamic septic shock
Hyvernat et al[26], 2016	RCT, double-blind study	Persistent hypoperfusion despite adequate fluid resuscitation and NE administration. Defined as severe sepsis with arterial hypotension (SBP < 90 mmHg or MAP < 70 mmHg) despite adequate fluid resuscitation	Mortality in 28-day	59 in the 200 mg group and 63 in the 300 mg group	64.8	200 mg group: 52.5%. 300 mg group: 44.4%
Annane et al[27], 2002	RCT double-blind, parallel group	Adults (> 18 years). Required NE to maintain MAP. Urinary output less than 0.5 mL/kg/hour for at least 1 hour. Arterial lactate levels higher than 2 mmol/L	28-day survival	Corticosteroid group: Number of Patients: 151. Treatment: Received hydrocortisone (50 mg IV every 6 hours) and fludrocortisone (50 µg orally once daily) for 7 days. Placebo group: Number of patients: 149. Treatment: Received matching placebos for 7 days	63	Corticosteroid group: 43.0% (65 out of 151 patients). Placebo group: 49.7% (74 out of 149 patients)
Morelli et al[35], 2009	Pilot RCT	Patients diagnosed with septic shock. MAP below 65 mmHg despite adequate volume resuscitation	The primary outcome was the total NE dose required to maintain a MAP of 65-75 mmHg over the 48-hour treatment period	TP group: 15 patients. Vasopressin group: 15 patients. NE gsroup: 15 patients	65	Zero during 48 hours
Lv et al[36], 2017	RCT, double-blind	Adults (≥ 18 years). Onset of septic shock within 6 hours	28-day all-cause mortality	Hydrocortisone group: 58 patients. Placebo group: 60 patients	Hydrocortisone group: 65.4 years. Placebo group: 66.1 years	28-day mortality: Hydrocortisone group: 23.1%. Placebo group: 32.2%. In-Hospital mortality: Hydrocortisone group: 31.0%. Placebo group: 40.0%
Russell et al[32], 2008	RCT, multicenter, double-blind	Adults diagnosed with septic shock requiring vasopressor support, NE	28-day all-cause mortality	Vasopressin group: 396 patients. NE group: 382 patients	64	28-day mortality: Vasopressin group: 35.4%. NE group: 39.3%. 90-day mortality: Vasopressin group: 43.9%. NE group: 49.6%
Russell et al[31], 2009	Post hoc, multicenter, blinded RCT substudy	(≥ 16 years). Diagnosed with septic shock, ongoing hypotension requiring at least 5 µg/minute of NE infusion for a minimum of 6 hours	28-day mortality	Corticosteroid treatment group: Vasopressin + corticosteroids: 296 patients. NE + corticosteroids: 293 patients. Corticosteroid treatment group: Vasopressin: 93 patients. NE: 97 patients	60	28-day mortality overall: Corticosteroids + vasopressin: 35.9% corticosteroids + NE: 44.7%. No corticosteroids + vasopressin: 66.7%. No corticosteroids + NE: 62.9%
Yildiz et al[37], 2011	RCT, double-blind	Adults (≥ 18 years). Confirmed sepsis, characterized by infection and systemic inflammatory response syndrome. Patients were often classified based on severity scores, such as APACHE II or SOFA, indicating varying degrees of organ dysfunction	28-day mortality	Steroid group: 100 patients. Placebo group: 100 patients	60	Steroid group: 16 (59.3) and placebo group: 15 (53.6)
Keh et al[30], 2003	RCT, double-blind crossover	Proven or strongly suspected infection. Presence of three or more of the following conditions: Mechanical ventilation. Heart rate > 90 beats per minute. Temperature > 38 °C or < 36 °C. White blood cell count > 12000 cells/µL or < 4000 cells/µL, or > 10% immature cells. Sepsis-induced hypotension: SBP < 90 mmHg or a reduction of > 40 mmHg from baseline	Change in MAP and systemic vascular resistance after treatment with low-dose hydrocortisone compared to placebo	Hydrocortisone group: 20. Placebo group: 20	Hydrocortisone group: 54. Placebo group: 50	Hydrocortisone group: 30%. Placebo group: 30%
Keh et al[38], 2016	RCT, double-blind	Adults (≥ 18 years). Evidence of infection Systemic response to infection (at least 2 systemic inflammatory response syndrome criteria). Organ dysfunction present for no longer than 48 hours. Not in septic shock at the time of randomization	Development of septic Shock: Within 14 days of treatment	Hydrocortisone group: 190. Placebo group: 190	65	28-day mortality: Hydrocortisone group: 21.1%. Placebo group: 23.7%. 90-day mortality: Hydrocortisone group: 28.4%. Placebo group: 30.5%. 180-day mortality: Hydrocortisone group: 34.2%. Placebo group: 36.8%
Annane et al[34], 2018	RCT, multicenter, double-blind	Adults (≥ 18 years). Diagnosed with septic shock requiring vasopressor therapy	90-day all-cause mortality	Hydrocortisone plus fludrocortisone: 614. Placebo: 626	63	28-day mortality: Hydrocortisone plus Fludrocortisone: 20%. Placebo: 25%. 90-day all-cause mortality: Hydrocortisone plus fludrocortisone group: 30%. Placebo group: 36%
Venkatesh et al[33], 2019	RCT, multicenter, double-blind	Adults undergoing mechanical ventilation. Documented or strongly suspected infection. Treated with vasopressors or inotropic agents at least 4 hours prior to randomization	90-day all-cause mortality	Hydrocortisone: 1988. Placebo: 1902	63	28-day mortality: Hydrocortisone group: 197. Placebo group: 199. 90-day all-cause mortality: Hydrocortisone group: 511; placebo group: 526
Liu et al[28], 2018	RCT, multicenter, double-blind	Adults with hypotension despite adequate fluid resuscitation, and at least two diagnostic criteria for systemic inflammatory response syndrome	28-day mortality	TP group: 312 patients. NE group: 305 patients	61	28-day mortality: TP group: 40%; NE group: 38%
Xiao et al[29], 2016	RCT	Adults with SBP < 90 mmHg and MAP < 65 mmHg are required to use vasoactive agents	7-day mortality	NE group: 17, NE + TP group: 15	63.2	7-day mortality: NE group 76.5% and NE + TP 33.3%

APACHE II: Acute Physiology and Chronic Health Evaluation II; ICU: Intensive care unit; MAP: Mean arterial pressure; NE: Norepinephrine; RCT: Randomized clinical trial; SOFA: Sequential Organ Failure Assessment; SBP: Systolic blood pressure; TP: Terlipressin.

All studies assessed the efficacy of therapeutic interventions in patients diagnosed with refractory septic shock. Most focused on adjunctive therapies used alongside standard vasopressor treatments, particularly NE.

The interventions evaluated in the studies were varied and included corticosteroids, vitamin C, thiamine, methylene blue, vasopressin, and terlipressin. Patient populations were diverse in terms of age, severity of septic shock, and comorbid conditions. Study designs ranged from single-center to multicenter trials, contributing to the observed variability in outcomes.

Primary outcome: Mortality

Mortality was the primary outcome in most of the studies included, with rates ranging from 28.6% to 56.8%. This wide variation was influenced by factors such as the specific intervention used, patient characteristics, and study design.

For instance, the combination of hydrocortisone, vitamin C, and thiamine did not result in a significant reduction in 90-day mortality between the treatment and placebo groups. In contrast, methylene blue showed a potential reduction in 28-day mortality, with 33% in the treatment group compared to 46% in the placebo group, although the small sample size in this study suggests that these results should be interpreted with caution[16].

Secondary outcomes

Secondary outcomes assessed across the studies included shock reversal, vasopressor-free days, ICU length of stay, and organ function as measured by the SOFA score.

In terms of shock reversal, the combination of hydrocortisone, vitamin C, and thiamine was associated with faster hemodynamic recovery compared to the placebo group. Additionally, another study found that patients in the intervention group experienced a greater number of vasopressor-free days, indicating that adjunctive therapies may help reduce reliance on vasopressors in critically ill patients.

Regarding the use of vasopressors, the addition of vasopressin significantly reduced the dependence on NE. Another study found improved renal perfusion in patients receiving terlipressin, although this did not lead to a reduction in overall mortality.

Several studies also measured the impact of interventions on organ function, as indicated by changes in SOFA scores. One study reported significant reductions in SOFA scores in patients treated with vitamin C, hydrocortisone, and thiamine, suggesting that these therapies may offer broader benefits for organ function, beyond simply improving hemodynamic status (Table 3).

Table 3 Secondary endpoints of the studies.

Ref.	Secondary outcomes	Adverse events
Lyu et al[15], 2022	28day mortality; ICU mortality; Hospital mortality; reversal of shock; time to shock reversal; 12 hours delta SOFA score; ICU free days; vasopressor free days; ventilator support free days; length of stay in ICU; length of stay in hospital	Hyperglycemia; hypernatremia; fluid overload
Ibarra-Estrada et al[16], 2023	Hemodynamic parameters; organ dysfunction; length of stay; adverse events	Hypotension; serotonin syndrome; there was a risk of serotonin syndrome, particularly in patients taking selective serotonin reuptake inhibitors. Allergic Reactions to methylene blue. Other minor events; nausea, vomiting, and local infusion site reactions
Hajjar et al[17], 2019	90 days all cause mortality; days free from advanced support; adverse effects	Arrhythmia
Moskowitz et al[18], 2020	Kidney failure; 30 day mortality; ventilator free days; shock free; ICU free days	Hyperglycemia; hypernatremia; new hospital acquired infections
Fujii et al[19], 2020	90 days mortality; duration of ICU stay; duration of hospital stay; The total length of time patients remained hospitalized. Organ dysfunction scores; adverse events	Intervention, 10 (10.2%) vs usual care, 7 (15.6%). Not related
Sevransky et al[20], 2021	Mortality at 30 days, ICU mortality, mortality at 180 days, length of ICU stay, length of hospital stay, and longterm emotional and cognitive outcomes at 180 days
Wang et al[21], 2022	Mortality rate at 28 days. Hemodynamic parameters; Including MAP and NE requirements. Adverse events related to TP use. Changes in serum creatinine levels and other renal function indicators	Cardiovascular events; increased heart rate or arrhythmias. Renal events; worsening renal function or acute kidney injury. Gastrointestinal events; ischemic colitis or other gastrointestinal complications. New infections or worsening of existing infections. Injection Site Reactions; Localized reactions at the site of TP administration
Cherukuri et al[22], 2019	Serum vitamin A levels; cortisol response; incidence of adverse effects	Nausea; injection site reactions
Douglas et al[23], 2020	RRT; mechanical ventilation; discharge rates; mortality rates	Arrhythmias; NE extravasation
Morelli et al[24], 2008	NE requirements; systemic and regional hemodynamic; organ function; SvO2	TP; cardiac ischemia; decreased cardiac output; gastrointestinal ischemia; skin reactions as pallor due to vasoconstriction. Dobutamine; tachyarrhythmias; increased myocardial oxygen demand; hypotension
Albanèse et al[25], 2005	Cardiac index; oxygen delivery index; oxygen consumption index; renal function; blood lactate levels	NE; tissue ischemia due to excessive vasoconstriction; arrhythmias or increased heart rate; increased oxygen demand of tissues; decreased mesenteric blood flow; TP; decreased cardiac index and potential for reduced oxygen delivery; bradycardia; fluid overload is used with caution in patients with renal impairment
Hyvernat et al[26], 2016	Vasopressor requirements, infectious and digestive complications, hemodynamic responses, days free from mechanical ventilation, vasopressors, and renal replacement therapy, adverse events	Haemorrhagic events; superinfections
Annane et al[27], 2002	Mortality rates; overall mortality rates at various time points (e.g., 7 days, 14 days). Time to vasopressor withdrawal; adverse events; clinical improvement	Infections; increased incidence of new infections in both groups. Gastrointestinal issues; stress ulcers or gastrointestinal bleeding. Hyperglycemia; observed in the corticosteroid group. Electrolyte Imbalances; Including hyponatremia and hyperkalemia. Neurological Events; Such as delirium or altered mental status
Morelli et al[35], 2009	Hemodynamic parameters; organ function; adverse events; mortality	Tachyarrhythmias; 1 in the vasopressin group. 4 in the NE group. Changes in bilirubin levels; higher total and direct bilirubin concentrations in the vasopressin and NE groups; platelet count; a decrease in platelet count was observed only in the TP group over time. Renal function; RRT was noted in the NE group
Lv et al[36], 2017	In hospital mortality; reversal of shock; duration of ICU stay; duration of hospital stay	Hyperglycemia; hydrocortisone group; 90.9%; placebo group; 81.5%
Russell et al[32], 2008	90 days mortality; organ dysfunction; adverse events; vasopressor requirements	Cardiac events (e.g., arrhythmias); Ischemic events (e.g., limb ischemia); other complications related to the use of vasopressors
Russell et al[31], 2009	Duration of vasopressor support; sock reversal; organ dysfunction; length of ICU stay; adverse events	Arrhythmias; limb ischemia; gastrointestinal ischemia
Venkatesh et al[33], 2018	28 days all cause mortality; Evaluated to assess shorter term effects. Duration of vasopressor use; measured to determine the impact on hemodynamic support. Duration of mechanical ventilation; assessed to evaluate respiratory support needs. Organ dysfunction; quality of life	Hyperglycemia; gastrointestinal bleeding; secondary infections
Yildiz et al[37], 2011	Adverse events; hormonal levels; incidence of adrenal insufficiency (AI) and Relative Adrenal Insufficiency (RAI); Assessment of adrenal function in patients; APACHE II and SOFA scores; clinical characteristics	No serious adverse events were reported in either the steroid or placebo groups
Keh et al[30], 2003	Inflammatory markers; NE requirements; immune function; mortality rates	Infections; increased incidence of secondary infections in both groups, but no significant difference between hydrocortisone and placebo. Gastrointestinal bleeding, though it was not significantly higher in the hydrocortisone group. Hyperglycemia; the hydrocortisone group required insulin management for some patients
Keh et al[38], 2016	Time until septic shock; mortality 28 days, 90 days, 180 days; secondary infections; weaning failure; muscle weakness; hyperglycemia	Hyperglycemia; secondary infections; muscle weakness; weaning failure
Annane et al[34], 2018	Mortality rates; days alive and free of vasopressors; organ failure; mechanical ventilation	Gastrointestinal bleeding; superinfection; hyperglycemia; neurologic sequelae
Liu et al[28], 2018	change in sofa score; days alive and free of vasopressors; incidence of serious adverse events	Digital ischemia; severe diarrhea; arrhythmias; intestinal ischemia
Xiao et al[29], 2016	hemodynamic parameters; complications; success rate for 6 hours resuscitation goals	Cardiac arrhythmias; skin necrosis at infusion sites; other organ dysfunctions

APACHE II: Acute Physiology and Chronic Health Evaluation II; ICU: Intensive care unit; NE: Norepinephrine; SOFA: Sequential Organ Failure Assessment; TP: Terlipressin; SvO₂: Oxygen venous saturation.

Adverse events

Adverse events associated with the therapeutic interventions varied across studies. Commonly reported adverse effects included hyperglycemia with corticosteroid use, which was noted in several studies, highlighting the need for careful blood glucose monitoring in patients receiving corticosteroids.

Cardiac arrhythmias were identified as a significant adverse event associated with vasopressin and terlipressin use. These arrhythmias were particularly concerning in critically ill patients, who may already be at high risk of cardiovascular instability.

Injection site reactions and other minor adverse events were more commonly observed with methylene blue. While these events were generally transient and not considered severe, they nonetheless warrant attention, especially in critically ill patients receiving multiple interventions.

Subgroup and sensitivity analyses

Subgroup analyses provided valuable insights into the effectiveness of the interventions in specific patient populations. Methylene blue appeared to be particularly effective in patients requiring higher doses of NE (greater than 0.5 μg/kg/minute), suggesting that it may be especially beneficial for patients with more severe shock (Table 4).

Table 4 Vasoactive drugs.

Ref.	NE dosage	TP dosage	Vasopressin dosage	Dobutamine dosage
Lyu et al[15], 2022	Not specified
Ibarra-Estrada et al[16], 2023	Not specified		0.03 IU/minute if NE dose reached ≥ 0.25 μg/kg/minute
Hajjar et al[17], 2019	Infusion was started at 5 mL/hours and increased by 2.5 mL/hours every 10 minutes to reach a maximum target rate of 30 mL/hours		Ranged from 0.01 to 0.06, IU/minute
Moskowitz et al[18], 2020	Not specified
Fujii et al[19], 2020	Initial dose: 0.05 to 0.5 μg/kg/minute, Titration: Dose adjusted based on the patient's blood pressure response, aiming to maintain a MAP ≥ 65 mmHg
Sevransky et al[20], 2021	Initial dose: Start at 0.05 to 0.1 μg/kg/minute. Titration: Adjust as necessary to achieve a MAP ≥ 65 mmHg
Wang et al[21], 2022	NE greater than or equal to 15 μg/minute (Norepinephrine ≥ 15 μg/minute)
Cherukuri et al[22], 2019	Not specified
Douglas et al[23], 2020	Initial dose: 0.05 to 0.1 μg/kg/minute; Titration: Dose increased based on the patient's response, often up to 0.5 μg/kg/minute or higher if necessary to achieve and maintain a MAP of ≥ 65 mmHg
Morelli et al[24], 2008	Continuous infusion: 0.9 mg/kg/minute to maintain MAP (MAP) at 70 mmHg	Single dose: 1 mg administered as a bolus infusion		Initial dose: 3 mg/kg/minute. Titration: The dose was progressively increased in increments of 1 to 3 mg/kg/minute to reverse the anticipated decrease in mixed SvO2 caused by the infusion of TP
Albanèse et al[25], 2005	Initial Dose: 0.3 μg/kg/minute. Titration: Increased by increments of 0.3 μg/kg/minute at 4-minute intervals to achieve a target MAP of 65 to 75 mmHg	Initial bolus: 1 mg (equivalent to 0.03-0.04 UI/minute). Second bolus: An additional 1 mg was given if the MAP remained below 65 mmHg after 20 minutes
Hyvernat et al[26], 2016	Initial dosage: The starting dose typically ranged from 0.05 to 0.5 μg/kg/minute and was titrated based on the patient's response. Titration: Doses were adjusted as needed to achieve the target MAP
Annane et al[27], 2002	Systolic arterial pressure lower than 90 mmHg for at least 1 hours despite adequate fluid replacement and administration of more than 5 μg/kg of dopamine, or current treatment with epinephrine or NE
Morelli et al[35], 2009	NE 15 μg/minute	Terlipressin 1.3 μg/kg/hour	Vasopressina 0.03 U/minute
Lv et al[36], 2017	Not specified
Russell et al[32], 2008	5 μg of NE per minute
Russell et al[31], 2009	5 μg of NE per minute
Venkatesh et al[33], 2019	Not specified
Yildiz et al[37], 2011	Not specified
Keh et al[30], 2003	Not specified
Keh et al[38], 2016	Not specified
Annane et al[34], 2018	0.5 to 1.0 µg/kg/minute to maintain MAP
Liu et al[28], 2018	starting dose of 4 µg/min, which could be titrated up to a maximum of 30 µg/minute	20 µg/hour, with the option to titrate up to a maximum of 160 µg/hour
Xiao et al[29], 2016	From 0.5 to 2.22 µg/kg/minute to maintain MAP	1.3 mg/kg/hour

MAP: Mean arterial pressure; NE: Norepinephrine; SvO₂: Oxygen venous saturation; TP: Terlipressin.

Similarly, the combination of vitamin C, hydrocortisone, and thiamine led to significant improvements in SOFA scores, particularly in patients with more severe organ dysfunction. This suggests adjunctive therapies may provide greater benefits to patients experiencing multiple organ failure.

Sensitivity analyses confirmed the robustness of these findings, although limitations such as small sample sizes and variability in study designs were acknowledged. These factors may have introduced some degree of bias, which should be considered when interpreting the results.

The substantial heterogeneity observed across studies precluded the performance of quantitative meta-analysis. The I² statistic for mortality outcomes exceeded 75%, indicating considerable statistical heterogeneity. Additionally, the diversity in interventions, patient populations, and outcome definitions made pooling of results clinically inappropriate. Therefore, we present a qualitative synthesis of the evidence with detailed tabular summaries.

DISCUSSION

This systematic review synthesizes findings from 24 randomized clinical trials investigating adjunctive therapies for refractory septic shock, encompassing interventions such as vasopressin, terlipressin, methylene blue, and combinations of hydrocortisone, vitamin C, and thiamine. The results highlight the persistent challenge of improving survival rates, with mortality ranging from 28.6% to 56.8%[15-19,21-33]. While some interventions demonstrated benefits in hemodynamic stabilization and increased vasopressor-free days, their impact on overall survival remains inconclusive[15-19,21].

The definition of refractory septic shock varied among studies, typically involving sustained hypotension (MAP ≤ 65 mmHg) despite adequate fluid resuscitation and high-dose vasopressors[15-17,21,25,27,34]. However, the lack of standardized inclusion criteria and endpoints—such as varying definitions of mortality (28-day vs 90-day) —changes in the SOFA score, and ICU length of stay, highlights significant heterogeneity among the studies[15-19,21,27,34]. This variability complicates direct comparisons and emphasizes the need for a consensus on the definition and measurement of refractory septic shock and its outcomes.

In terms of hemodynamic targets, maintaining a MAP ≥ 65 mmHg was consistently used as the goal across the studies[15-18,21,25,27,34], in line with current clinical guidelines[1]. However, evidence suggests that personalized hemodynamic targets could lead to better outcomes, particularly in patients with preexisting conditions such as hypertension. Emerging studies advocate for tailoring MAP targets to individual patient needs, rather than relying on fixed thresholds. This approach could help balance the risks of over-resuscitation, which may lead to complications such as myocardial injury, against the dangers of under-resuscitation, which exacerbates tissue hypoperfusion[1,17,19,21,34].

The therapies evaluated in this review demonstrated varying degrees of efficacy. Methylene blue, through its inhibition of nitric oxide synthesis, exhibited potential for reducing mortality[16], though further validation in larger trials is needed. Vasopressin was found to reduce NE requirements, but its impact on mortality remained inconsistent across studies[17,21,31,32]. Similarly, combinations of hydrocortisone, vitamin C, and thiamine improved secondary outcomes, such as vasopressor-free days and organ dysfunction, without consistently reducing mortality[15,18,19]. These findings suggest that while these therapies provide hemodynamic and metabolic support, their effects may not be sufficient to address the complex pathophysiology of refractory septic shock.

Mortality remained the most commonly reported primary endpoint across studies[15-19,21-33], underscoring its importance as the benchmark for assessing therapeutic efficacy. Secondary outcomes, such as SOFA score changes, vasopressor-free days, and ICU length of stay, highlighted the capacity of certain therapies to stabilize hemodynamics and improve organ function[15-19,21]. For instance, methylene blue demonstrated significant potential in reducing vasopressor dependency[16,21], and terlipressin improved renal perfusion and urine output[21,25]. However, the long-term impact of these therapies remains unclear.

Adverse events were variably reported and included hyperglycemia, particularly with corticosteroids, as observed in Lyu et al[15] and Venkatesh et al[33]. Cardiac arrhythmias were noted with vasopressin and terlipressin, while methylene blue was associated with injection site reactions[16]. These adverse events highlight the importance of carefully balancing the therapeutic benefits of these interventions with their potential risks, particularly in critically ill patients who may already be vulnerable to complications.

Variability in dosing regimens further complicates direct comparisons among studies. For example, vasopressin dosages ranged from 0.01 to 0.06 IU/minute[17,31,32], while NE was titrated to maintain MAP ≥ 65 mmHg, often exceeding 0.5 μg/kg/minute in severe cases[15,16,19,21]. These discrepancies highlight the need for standardizing dosing protocols to optimize clinical outcomes and minimize adverse events.

Limitations

The studies included in this review exhibited several methodological limitations, although some of these limitations can be mitigated through careful interpretation and ongoing research. Small sample sizes were a common limitation, reducing the statistical power to detect significant differences in both mortality and secondary outcomes. Despite these small sample sizes, the consistency of trends across multiple studies strengthens the reliability of the findings, and these smaller trials often provide valuable preliminary data that warrant further investigation in larger, multicenter trials.

The heterogeneity of study designs, inclusion criteria, and interventions introduced variability that complicated direct comparisons. However, this heterogeneity reflects the diversity of clinical scenarios encountered in real-world settings and enhances the external validity of the findings. Additionally, subgroup analyses performed in some studies—such as those focusing on patients with higher NE requirements or specific patient populations—provided valuable insights into how tailored therapeutic approaches may improve outcomes.

Many studies were limited by short follow-up periods, typically 28 or 90 days, restricting the ability to evaluate long-term survival and quality of life. While short-term follow-ups are important for understanding immediate survival and stabilization—often primary goals in the acute phase of septic shock—future studies could build on these findings by incorporating extended follow-up periods to assess longer-term outcomes.

Inconsistent definitions of refractory septic shock and varying thresholds for vasopressor refractoriness further challenged the generalizability of the findings. Efforts to standardize these definitions are ongoing, and the use of core inclusion criteria, such as MAP thresholds and vasopressor requirements, provides reasonable consistency for interpreting results across studies.

Finally, adverse event reporting was inconsistent, limiting the comprehensive assessment of the safety profiles of the interventions. Nonetheless, key adverse events — such as hyperglycemia and cardiac arrhythmias — were consistently documented in high-quality trials. The adoption of standardized reporting guidelines, such as those recommended by CONSORT, could improve the uniformity of safety data in future studies.

Future directions

Future research should prioritize multimodal treatment strategies that integrate immune modulation, hemodynamic optimization, and organ protection. Larger, multicenter trials using harmonized definitions and standardized methodologies are essential to validate the findings of this review. Investigating cytokine-blocking therapies targeting interleukin (IL)-1, IL-6, and tumor necrosis factor alpha, as well as precision tools such as the Hypotension Prediction Index, could offer new insights into targeted interventions[1].

Additionally, long-term outcomes, including survival beyond 180 days and quality of life, should be incorporated into future studies to provide a more comprehensive evaluation of therapeutic efficacy. Standardizing dosing protocols and exploring combination therapies tailored to patient-specific conditions could further refine treatment approaches for refractory septic shock.

CONCLUSION

Refractory septic shock remains a critical clinical challenge, with high mortality rates despite the use of adjunctive therapies such as methylene blue, vasopressin, and combinations of hydrocortisone, vitamin C, and thiamine. While these interventions show potential in improving hemodynamic stability and secondary outcomes, their impact on survival remains uncertain. The variability in study designs and outcome definitions underscores the need for standardization in future trials. Larger, well-designed studies are required to confirm the long-term efficacy and safety of these therapies and to explore personalized treatment strategies for better patient outcomes. These efforts are essential to guide more effective clinical management strategies and improve outcomes for this high-risk patient population.

ACKNOWLEDGEMENTS

The authors would like to thank those who provided essential support throughout this systematic review process.