The study was conducted using an adapted two-round Delphi methodology, with the aim of reaching consensus on the standardization of key aspects regarding the definition, diagnosis, and management of RSS.

The Delphi method is a social research tool whose purpose is to achieve a reliable collective opinion from a group of experts.11 Participants must be consulted at least twice on the same topic, allowing them to revise their responses after learning the opinions of the other panelists. Anonymity of responses is guaranteed, and there is no temporal or in-person interaction among participants. Communication of information is channeled through the study’s scientific committee, which determines the data shared with the panelists.11

An advisory scientific committee including 5 experts (4 specialists in intensive care medicine and 1 specialist in anesthesiology and resuscitation), active members of their respective scientific societies, with extensive clinical, teaching, and research experience in the field of sepsis and hemodynamics (with an H-index > 20 in these fields), was responsible for designing and validating the content of the Delphi questionnaire, validating the aggregated analysis results, and drafting and validating the final document.

The expert panel consisted of 30 specialists with expertise in sepsis and hemodynamics.

The preparation and writing of this manuscript followed the EQUATOR Network guidelines with the recommendations of the RIGHT-Statement (Reporting Items for Practice Guidelines in Healthcare) and the Appraisal of Guidelines for Research & Evaluation (AGREE-II) tool (Tables 1 and 2 of the supplementary data).

The questionnaire was designed based on a targeted literature search conducted across the Medline and PubMed databases (https://pubmed.ncbi.nlm.nih.gov/), to identify relevant aspects of the study. MeSH and automatic filters were applied to limit the results to publications from the last 10 years, studies in humans and adults (over 19 years of age) to obtain an updated and relevant overview of the available literature (Table 3 of the supplementary data). Based on this information, a questionnaire with 28 statements was developed. For those items where no consensus was reached, the statements were redefined, and 5 questions were removed due to lack of scientific evidence for the 2nd round, although they had been included in the 1st one to assess the clinical perception of the experts. The questionnaire was validated through a semantic review to ensure terminological consistency, as well as an evaluation of the relevance of each item.

Each statement had to be rated according to the level of agreement using the following options: “Strongly agree,” “Agree,” “Slightly agree,” and “Disagree.” The statements were organized into 3 sections: block #1: Definition of RSS (7 statements); block #2: Diagnostic criteria for RSS (7 statements); and block #3: Therapeutic strategy for RSS (14 statements). Due to technical issues, the results of question 12 in block #3 were not obtained. The questionnaire was programmed using the Google® Forms tool (https://docs.google.com/forms/u/0/).

The expert panel consisted of 30 professionals involved in the management of RSS in Spain, including 20 intensivists and 10 anesthesiologists. The Delphi study was double-blind for the panelists. In the Delphi context, “double-blind” indicates that panelists are unaware of the identity and individual responses of the other participants, ensuring anonymity among them. The coordinating team knows the identity of the panelists only for logistical management purposes, without influencing their responses or individual analysis.

Panelists were selected by the advisory scientific committee. For their selection, rigorous criteria were considered, such as their professional trajectory, the relevance and volume of their scientific publications, their participation in research projects, and their affiliation with highly prestigious hospitals (Table 4 and Fig. 1 of the supplementary data).

An invitation describing the objective of the study and the Delphi process was sent. After agreeing to participate, the experts received a link to access the Delphi questionnaire, with the option to withdraw at any time.

The 1st round took place on December 10th, 2024, with responses collected from December 12 through January 8, and the 2nd round on February 26th, 2025, from March 3 through March 10th. The questionnaire was completed virtually and anonymously.

After analyzing the data obtained in the 1st round, items that had reached consensus were removed. For the 2nd round, the updated Delphi questionnaire was sent to all respondents, along with a summary of the 1st-round results, allowing them to modify their responses based on the general judgment expressed by the expert panel.

The advisory committee reviewed the results of both the 1st and 2nd rounds after their completion (Fig. 2, supplementary data).

Consensus was considered achieved when at least 75% of the panelists provided the same response. In the absence of an absolute majority for a single category, it was agreed to group responses into “Strongly agree” and “Agree” on one hand, and “Slightly agree” and “Disagree” on the other.

In the 1st round, consensus was reached for 5 out of 7 statements. The experts agreed that RSS is a clinical entity requiring a differential treatment strategy (100% of experts), as well as on the need to consider persistence of the condition (76.7%), together with the presence of hypoperfusion (96.6%), hypotension (80%), and failure of ≥1 organs (80%) within the definition of the profile of a patient with RSS.

No consensus was reached regarding whether a patient with RSS is one who does not respond to 1st-line therapy but may respond to 2nd- and 3rd-line therapy (66.6%), or one who does not respond to 1st- or 2nd-line therapy but may respond to 3rd-line therapy (63.4%).

The statement regarding lack of response to 1st- and 2nd-line therapy but possible response to 3rd-line therapy was eliminated, as such a level of refractoriness was considered unnecessary. The statement regarding whether a patient who does not respond to 1st-line therapy may respond to 2nd- and 3rd-line therapy was reformulated and included in the 2nd round (A patient with RSS is one who does not respond to 1st-line therapy (optimized treatment: adequate preload, appropriate antibiotic therapy, source control, and treatment with norepinephrine), but may respond to other non-catecholaminergic vasopressors (vasopressin, angiotensin II, etc.)), reaching consensus with 93.3% of experts.

Table 5 and Fig. 3 of the supplementary data show the level of agreement for each statement in block 1 according to the round.

Regarding the 7 statements included in block 2, no consensus was reached for any of them after the 1st round.

All statements were reformulated and included in the 2nd round. Additionally, the statement regarding ruling out cardiac dysfunction to optimize RSS treatment through echocardiography or advanced hemodynamic monitoring was moved to block #3, as it referred to treatment.

In the following round, consensus was reached regarding the need to receive hydrocortisone to be considered RSS (80% of experts), as well as considering baseline norepinephrine dose as a defining criterion (90%), that receiving a 2nd vasopressor is not necessary (86.7%), as well as considering lactate level for the diagnosis of RSS (80%), and the absence of a decrease in lactate levels as part of the diagnostic criteria (80%).

No consensus was ever reached regarding whether prolonged capillary refill time (>3 s) should be considered a diagnostic criterion for RSS (46.4% of experts).

Table 6 and Fig. 4 of the supplementary data show the level of agreement for each statement in block 2 according to the round. Table 3 presents the agreed diagnostic criteria.

Among the statements in block #3, consensus was reached after the 1st round for 8 out of 14 statements. All the experts (100%) agreed that multimodal treatment is an optimal therapeutic strategy for the management of RSS, as well as ICU admission, placement of central venous access for vasopressor administration, and the addition of a 2nd and/or 3rd vasopressor. Moreover, consensus was reached on the need for continuous advanced monitoring (76.6% of experts), echocardiography as a fundamental tool (96.7%), vasopressin as the 2nd vasopressor of choice in patients with renal failure (83.3%), and considering concomitant treatments with angiotensin II receptor blockers (ARBs) or angiotensin-converting enzyme inhibitors (ACE inhibitors) when selecting the vasopressor (90%). Regarding the use of adjunctive therapies, the expert panel reached consensus on not using immunoglobulins routinely in the treatment of RSS (80%).

No consensus was ever reached for the remaining statements. The statements regarding angiotensin II as the treatment of choice in renal failure and renin as a biomarker for angiotensin II treatment were reformulated and included in the 2nd round. The remaining statements were eliminated due to insufficient published scientific evidence to address these questions.

In the 2nd round, 90% of experts agreed on ruling out the presence of cardiac dysfunction to optimize treatment and that angiotensin II could have a renal protective effect vs baseline norepinephrine in patients with renal failure, while 96.7% of experts also considered that, if urgent renin measurement were available, they would use it to identify patients who would benefit from vasopressor treatment with angiotensin II.

Table 7 and Fig. 5 of the supplementary data show the level of agreement for each statement in block #3 according to the round. Fig. 1 shows the agreed treatment algorithm.

Figure 1.

Accepted treatment algorithm for RSS.

According to the results, RSS would constitute a distinct clinical entity requiring a specific therapeutic strategy. This reinforces the need to identify non-responder patients who may benefit from personalized therapeutic approaches.12

Consensus was reached that RSS should be persistent (>1 h), which is consistent with the literature emphasizing duration despite adequate resuscitation.14 Although there is no universal threshold, a duration of “>1 h” would be useful for early identification of patients with RSS.

In addition, the definition of RSS should include the presence of hypoperfusion data, such as elevated lactate, oliguria, or metabolic acidosis. This finding is supported by international clinical practice guidelines13,15 and reinforces the inclusion of these markers in its definition.

Organ dysfunction, beyond hypotension, should also be included as a diagnostic criterion, in line with the Sepsis-3 definition,16 which highlights that organ dysfunction, especially cardiovascular, indicates greater severity in defining the refractory nature.17–19 It was agreed to include hypotension ([MAP] < 65 mmHg) in the definition, facilitating the identification of severe patients and the indication for vasopressor therapy.2,16,20

The expert panel reached consensus in considering RSS in patients who do not respond to 1st-line treatment but may respond to other non-catecholaminergic vasopressors (vasopressin, angiotensin II). No consensus was reached on considering other treatment lines beyond vasopressor support as a criterion for RSS, suggesting that experts defined refractoriness based on response to or need for 2nd- or 3rd-line vasopressor support rather than other treatments.13,15,21

Agreement also increased regarding the concept that a patient with RSS may not respond to 1st-line therapy but to 2nd- or 3rd-line therapy. This supports a stepwise therapeutic approach, in which additional interventions such as corticosteroids, alternative vasopressors, vitamin C, or even ECMO are introduced after failure of initial therapy with fluids, antibiotics, or baseline norepinephrine.1,6,15

Following expert consensus, RSS is considered to be present in patients who, after optimized initial management, maintain hypotension and/or signs of global hypoperfusion for more than one hour, requiring the use of one or more vasopressors and advanced monitoring, in a setting of multiorgan failure.

The lack of a threshold dose of baseline norepinephrine to define RSS reflects clinical variability and the absence of a universal definition of refractoriness to vasopressor treatment.13,22,23

Although consensus was not reached on a specific norepinephrine dose to define RSS, most of the expert panel rejected considering doses < 0.25 μg/kg/min, suggesting that the threshold dose should be at least above this level. In the 2nd round, consensus was reached to consider baseline norepinephrine dose alone as a criterion to define RSS, recognizing that the need for increasing doses reflects refractoriness to vasopressor support.23

In this regard, there was no consensus on considering the use of a second vasopressor as a mandatory diagnostic criterion to define RSS. Although baseline norepinephrine as 1st -line therapy is widely accepted, guidelines do not specify a 2nd agent to define refractoriness,13 and the decision to add vasopressin or epinephrine depends on the clinical context.24 In the 2nd round, it was agreed that a second vasopressor is not necessary to diagnose RSS, as its absence does not rule out refractoriness if the response to treatment is inadequate.13

Regarding corticosteroid use, experts agreed that all patients with RSS should have received hydrocortisone, unless contraindicated, in full compliance with international clinical practice guidelines. This recommendation is consistent with former studies which, in addition to demonstrating a mortality benefit, have shown reduced vasopressor requirements and faster resolution of shock in patients treated with corticosteroids.25–27

Although elevated lactate levels are associated with worse prognosis and their reduction with better outcomes, it is recommended to evaluate lactate within a dynamic assessment of hemodynamic status,22,23 as it may be elevated due to causes unrelated to hypoperfusion. More than 75% of experts considered that levels > 3 mmol/L could indicate RSS, and the majority rejected considering a clearance < 10% as a diagnostic criterion. After reformulation, consensus was reached to include lactate level and lack of decrease in its levels in the diagnostic definition of RSS, given its usefulness as a marker of hypoperfusion and an indicator of refractoriness.16

Although capillary refill time is a useful and non-invasive tool to assess peripheral perfusion and microcirculatory status, its value as an isolated diagnostic criterion for RSS is limited.13 No consensus was ever breached on considering capillary refill time > 3 s as pathological, due to its sensitivity to external factors and the lack of solid evidence as a single marker of refractoriness, recommending instead a multimodal monitoring approach.28

The potential impaired response to catecholamines in RSS would support the consensus on multimodal treatment, combining ≥ 2 vasopressors. The sequential addition of vasopressin or angiotensin II would allow optimization of perfusion without increasing the adverse effects of a single agent.6,29 Similarly, the need for continuous advanced hemodynamic monitoring was agreed upon to guide the administration of vasoactive agents, inotropes, and fluid therapy, especially in patients with complex hemodynamic profiles or multiple organ dysfunction.30

Echocardiography was recognized as a key imaging modality to assess cardiovascular status and the dynamic follow-up of ventricular function, preload dependence, and response to treatment, due to its non-invasive nature and availability in intensive care units (ICU).20,31 This perception was reinforced by the need to rule out cardiac dysfunction to optimize RSS treatment, specifying diagnostic criteria such as a left ventricular ejection fraction (LVEF) < 45%–50%, allowing experts to agree on the importance of early identification to adjust individualized treatment.20,32

The need for advanced life support, monitoring, and complex therapies in RSS would justify that all patients should be admitted to the ICU²⁰. In addition, it was agreed that central venous access should be established for the safe administration of vasopressors, avoiding the risk of extravasation and tissue necrosis associated with peripheral administration.33

Finally, consensus was reached on the addition of a 2nd or 3rd vasopressor as part of a stepwise therapeutic strategy.1,6,15 In patients with renal failure, vasopressin is preferred as the 2nd agent due to its favorable hemodynamic profile and lower impact on renal perfusion vs baseline norepinephrine.29,31 The importance of considering chronic treatment with ACEi or ARBs when selecting the vasopressor was recognized. In patients treated with ACEi, angiotensin II is effective by enhancing the hemodynamic response and reducing catecholamine use, whereas in those treated with ARBs, AT1 receptor blockade reduces the effect of exogenous angiotensin II.34,35

No consensus was ever reached on the use of angiotensin II as a 2nd vasopressor in patients with renal failure or myocardial dysfunction under optimal inotropic treatment, likely due to limited evidence and its application restricted to specific contexts,36 nor on vasopressin as a 2nd vasopressor in patients with myocardial dysfunction, due to its potential effects on afterload and coronary perfusion.29,31 After reformulation, consensus was reached on the possible renal protective effect of angiotensin II vs baseline norepinephrine in patients with renal failure,37 as it may improve renal perfusion in subgroups of patients with vasodilatory shock.38

Regarding adjunctive therapies such as methylene blue, blood purification, vitamin C and/or thiamine, and immunoglobulins, the lack of consensus may reflect the absence of robust and consistent evidence supporting their systematic use, as results are heterogeneous and they are considered experimental or compassionate-use therapies.5,20,39

No agreement was ever reached either on the usefulness of copeptin or renin measurement to individualize vasopressor treatment. Although these markers could help identify patient phenotypes potentially responsive to treatment—low copeptin levels may predict better response to vasopressin, and high renin levels to angiotensin II—their clinical use is limited by lack of standardization, availability, and validation.20,40 Urgent renin measurement to guide angiotensin II treatment achieved consensus after reformulation. Although not widely available, its use would help identify patients with dysfunction of the renin–angiotensin–aldosterone axis6 who could potentially benefit from angiotensin II administration.

The Delphi methodology allows participants to express their opinions freely but has limitations, such as potential bias in expert selection, as the most experienced professionals may not be identified. The recommendations reflect the consensus opinion of the expert panel and not the official position of any scientific society. Additionally, the structured questionnaire may restrict responses. The number and type of modifications introduced in the 2nd round, together with the use of a 4-point scale, may have unintentionally influenced responses. In some cases, the wording of certain multiple-response questions in the first round may have posed limitations in comprehension (non-exclusive intervals or formulations open to interpretation), which may have affected the level of consensus achieved and required adjustments in the subsequent round. Similarly, relevant aspects such as the impact of respiratory distress and higher severity profiles (APACHE > 30) were not addressed in depth, constituting a limitation of the study.