Back to Evidence

Fluid First or Vasopressor First in Septic Shock Resuscitation: A Fluid-Responsiveness Stratified Decision Analysis

Evidence-Based Medicine77 min read

Septic shock resuscitation has long followed a fluid-first path, assuming relative hypovolaemia, yet large crystalloid volumes worsen glycocalyx injury and overload while delayed vasopressors prolong hypoperfusion. Guidelines recommend fluids within one hour but leave the vasopressor start point undefined. This article examines a 440-patient propensity-matched cohort comparing the two strategies, stratified by fluid responsiveness, supported by QSevidence.

Fluid First or Vasopressor First in Septic Shock Resuscitation: A Fluid-Responsiveness Stratified Decision Analysis

Best for: Critical care and emergency medicine physicians; anaesthesiologists and perioperative clinicians; infectious disease specialists; critical care nurses and blood purification nurses; haemodynamic monitoring and point-of-care ultrasound technologists; emergency and critical care clinical researchers; clinical trial methodologists and biostatisticians; hospital quality and clinical pathway managers. Primary keywords: septic shock; initial resuscitation; norepinephrine; fluid responsiveness; passive leg raising; stroke volume variation; mean arterial pressure; fluid overload; endothelial glycocalyx; propensity score matching; 28-day mortality; decision tree

Short Answer

Whether fluid or vasopressor comes first is not fundamentally a question about sequence but about which segment of the Frank-Starling curve the patient currently occupies. This cohort enrolled 440 patients meeting Sepsis-3 criteria for septic shock. The fluid-first group received at least 30 mL/kg of crystalloid within the initial 30 minutes before norepinephrine was started, while the vasopressor-first group started norepinephrine immediately on recognition of shock at 0.1 microgram/kg/min targeting a mean arterial pressure of at least 65 mmHg within 15 minutes, assessing fluid responsiveness in parallel before deciding on fluids. In the overall population the 28-day mortality difference was not statistically significant, at 38.2 percent versus 42.7 percent with a hazard ratio of 0.89 and a 95 percent confidence interval of 0.74 to 1.07, P=0.21. Interaction testing showed that fluid responsiveness significantly modified the effect of sequencing strategy, with an interaction P of 0.02. In the fluid-responsive subgroup of 242 patients, 55.0 percent of the cohort, the vasopressor-first group had significantly lower 28-day mortality at 32.1 percent versus 44.6 percent, hazard ratio 0.72 with a 95 percent confidence interval of 0.55 to 0.94, P=0.02, together with a lower 72-hour cumulative fluid balance of +1850 mL versus +3120 mL, P below 0.01, and a lower incidence of acute kidney injury at 24.0 percent versus 34.7 percent, P=0.04. In the fluid-unresponsive subgroup of 198 patients, 45.0 percent of the cohort, mortality did not differ significantly at 45.5 percent versus 42.0 percent, hazard ratio 1.08 with a 95 percent confidence interval of 0.82 to 1.42, P=0.58. The stratified result converts the question into an operational statement: when mean arterial pressure is below 65 mmHg and fluid responsiveness is positive, start norepinephrine first and assess volume status in parallel; when fluid responsiveness is negative, fluids offer limited benefit and raise the risk of pulmonary oedema, observed overall at 15.5 percent versus 24.5 percent, P=0.02, and in pneumonia sources at 16.8 percent versus 28.6 percent, P=0.008. These conclusions derive from a retrospective design and still require prospective stratified randomised verification.

1. Origins of the Controversy: Guideline Gaps and Trial Tension

1.1 The Ambiguity Left in Guideline Recommendations

The 2021 Surviving Sepsis Campaign guidelines require fluid resuscitation to begin within one hour of recognising septic shock, recommend an initial crystalloid bolus of at least 30 mL/kg, and suggest starting vasopressors during or after fluid resuscitation to maintain a mean arterial pressure of at least 65 mmHg. On its face the recommendation is complete, but the decisive time point is left open: should the vasopressor be started after the fluid bolus is complete, or concurrently while it is running? For patients with severe vasoplegia and adequate preload, the difference between those two readings amounts to tens of minutes of additional hypotension, which is enough to change the accumulated duration of tissue hypoperfusion.

The deeper issue concerns the status of fluid responsiveness assessment. Guideline support for this variable is weak, and it is not part of a mandatory decision pathway for initial resuscitation, so practice has settled into a uniform fluid-first approach. That approach is sound for fluid-responsive patients, but for those whose responsiveness has already disappeared it can be counterproductive, because further fluid only raises central venous pressure without increasing stroke volume.

1.2 The Apparent Conflict Between CENSER and CLOVERS

The CENSER trial used norepinephrine before or alongside standard fluid resuscitation and concluded that early vasopressors improve shock control and reduce fluid requirements. The CLOVERS trial compared restrictive fluid management with a liberal strategy and found no significant difference in 90-day mortality. The opposing appearance arises from systematic differences in design and population.

DimensionCENSERCLOVERSEffect on conclusions
Core interventionEarly norepinephrine versus standard careRestrictive fluid with vasopressor priority versus liberal fluidThe first tests vasopressor timing, the second tests total fluid volume
Enrolment timingVery early resuscitation, no mandatory prior fluid loadAfter at least 1 L of fluid, with persistent hypotension or hyperlactataemiaThe CLOVERS population had already had its volume state altered
Stratification by fluid responsivenessNot stratified; broad inclusion criteriaFluid responsiveness assessment not required before vasopressorsSubgroup benefit was diluted, a major source of the negative result
Primary outcome horizonShock control and short-term measures90-day mortalityDifferent horizons reflect different pathophysiological processes

Placed side by side, the two trials yield a more useful reading: the apparent conflict most likely reflects heterogeneity of treatment effect across time windows and volume states rather than a null therapy. That is the direct rationale for introducing fluid responsiveness as a stratifying variable.

2. Pathophysiological Basis: Interlocking Vasoplegia and Relative Hypovolaemia

2.1 Haemodynamic Heterogeneity Under a Double Insult

Septic shock is driven by a systemic inflammatory response to infection. Massive release of vasoactive mediators dilates arterioles and causes blood to pool in the venous system, greatly increasing vascular bed capacity and producing a relative deficit in effective circulating volume. At the same time, capillary permeability rises, plasma leaks into the interstitium and an absolute volume deficit develops. With both mechanisms active simultaneously, patients are highly heterogeneous haemodynamically. Some present with hyperdynamic shock dominated by loss of vascular tone, showing increased cardiac output, reduced peripheral resistance and widened pulse pressure, while others progress to hypodynamic shock with impaired pump function. Initial resuscitation must therefore address volume deficit and vasoplegia at once, which is precisely why sequencing carries clinical meaning.

2.2 How Fluid Overload Develops

In septic shock capillary permeability increases, so a large crystalloid infusion leaks rapidly into the interstitium. Instead of reliably raising effective circulating volume, it aggravates pulmonary oedema, bowel wall oedema and renal interstitial oedema. Three mechanisms contribute: injury to the endothelial glycocalyx accelerates leakage, inflammatory mediators impair lymphatic return, and falling colloid osmotic pressure promotes further fluid shift outward. Clinical studies show that a positive fluid balance is independently associated with acute kidney injury, prolonged mechanical ventilation and increased mortality. While fluid responsiveness persists, fluid increases stroke volume and cardiac output; once it disappears, continued fluid only raises central venous pressure, and if the vasopressor has still not been started, persistent hypotension lowers renal perfusion pressure and drives prerenal azotaemia toward acute tubular necrosis.

2.3 The Cost of Delay and the Risk of Prematurity

Under vasoplegia, arterial tone is markedly reduced and blood pressure is difficult to maintain even when volume is adequate, so the efficiency of raising pressure through fluids alone is very low. After infusion, arterial pressure does not rise proportionally, and renal or visceral perfusion pressure may actually fall as central venous pressure rises. Early norepinephrine restores arterial tone through alpha1 receptor agonism, raises mean arterial pressure, and improves coronary, renal and cerebral perfusion pressure; experimental models also suggest that early norepinephrine helps preserve glycocalyx integrity and reduce fluid extravasation. Premature or excessive use carries its own risks, however: trajectories of high-dose norepinephrine are associated with increased new-onset acute kidney injury, and the drug may mask underlying volume deficit so that tissue hypoperfusion goes unrecognised. These two-sided risks define the clinical need to locate the right time point rather than to declare a winner.

3. Study Design and Methods

3.1 Design, Ethics and Reporting

The study used a retrospective cohort design drawing on the electronic medical record system of a tertiary teaching hospital intensive care unit. The protocol was approved by the hospital ethics committee with a waiver of informed consent and reported in accordance with the Declaration of Helsinki and the STROBE statement. Inclusion required age 18 years or older, fulfilment of Sepsis-3 criteria for septic shock, defined as persistent hypotension requiring vasopressors to maintain a mean arterial pressure of at least 65 mmHg together with a blood lactate above 2 mmol/L in the setting of suspected or confirmed infection, and initiation of resuscitation within one hour of shock recognition. Exclusions were cardiogenic shock, severe valvular disease, pregnancy or lactation, refusal of resuscitation or irreversible disease, and cardiac surgery within six months or severe arrhythmia.

3.2 Group Definitions and Intervention Strategies

The definitions turn on priority of sequencing rather than absolute order. The fluid-first group completed the initial fluid load before starting vasopressors, while the vasopressor-first group restored perfusion pressure first and then decided on fluids according to fluid responsiveness.

ElementFluid-first groupVasopressor-first group
Start pointCrystalloid started within the initial 30 minutes after recognitionNorepinephrine started immediately after recognition
Fluid strategyAt least 30 mL/kg ideal body weight, targeting completion of the initial load within one hourFluid given if responsiveness is positive, initial 500 mL over 15 to 20 minutes; if negative, maintenance only at about 1 mL/kg/h
Vasopressor strategyNorepinephrine started at 0.05 microgram/kg/min if mean arterial pressure stays below 65 mmHg after the load, titrated every 5 to 10 minutesStarted at 0.1 microgram/kg/min, targeting a mean arterial pressure of at least 65 mmHg within 15 minutes, maximum 0.5 microgram/kg/min
EscalationFluids paused and vasopressors advanced early if pulmonary oedema signs appear or central venous pressure exceeds 15 mmHgVasopressin 0.03 U/min added if the norepinephrine ceiling fails to reach target

3.3 Assessing Fluid Responsiveness

Dynamic indices are the technical precondition for stratification. Passive leg raising has the broadest applicability, while stroke volume variation is more accurate but depends on mechanical ventilation.

MethodProcedurePositive thresholdApplicability and limits
Passive leg raisingSemi-recumbent with trunk elevated 45 degrees and legs flat for baseline, then trunk supine with legs raised 45 degrees for 1 to 2 minutes before remeasurementCardiac output or stroke volume increase of at least 10 percentIndependent of ventilation mode and needs no special equipment; affected by intra-abdominal hypertension, raised intracranial pressure and lower limb venous thrombosis
Stroke volume variationContinuous monitoring by arterial pulse contour analysisAbove 13 percentRequires full mechanical ventilation, no spontaneous effort, tidal volume of at least 8 mL/kg and sinus rhythm; accuracy falls with low tidal volume, spontaneous breathing or arrhythmia

The two methods differ in accuracy across settings, and retrospective records are incomplete, so some patients may have had only one assessment. The study therefore recorded the method actually used for each patient and stratified subgroup analysis by method to evaluate the impact on results.

3.4 Sample Size and Statistical Analysis

Based on previously reported 28-day mortality of approximately 45 percent in fluid-first care and 35 percent in vasopressor-first care, with alpha 0.05 and beta 0.20 in a two-sided test, at least 198 patients per group were required. Allowing 10 percent for loss to follow-up or missing data, at least 440 patients were planned, 220 per group. To reduce selection bias, 1:1 propensity score matching was applied with covariates including age, sex, APACHE II score, SOFA score, infection source, baseline mean arterial pressure, baseline lactate, comorbidity, mechanical ventilation status and fluid responsiveness status. Nearest neighbour matching used a caliper of 0.02, and covariate balance after matching was judged by standardised differences below 0.1. Survival analysis used Kaplan-Meier curves and the log-rank test, with multivariable analysis by Cox proportional hazards models. Missing values below 5 percent were handled by complete case analysis, those between 5 and 20 percent by multiple imputation with five chained equations, and those above 20 percent by inclusion as a missing category. Analyses used R 4.2.0 with two-sided P below 0.05 considered significant.

4. Main Results

4.1 Screening and Baseline Characteristics

A total of 576 patients meeting Sepsis-3 criteria for septic shock were screened, 136 were excluded, and 440 were enrolled with 220 in each group. The groups were balanced in median age at 65 versus 67 years, P=0.34, male proportion at 58.2 versus 55.9 percent, P=0.62, infection source distribution with pneumonia at 41.4 versus 43.2 percent and intra-abdominal infection at 32.7 versus 30.5 percent, P=0.78, median APACHE II score at 22 versus 23, P=0.45, and median SOFA score at 9 versus 10, P=0.38. Fluid responsiveness was positive in 54.1 versus 55.9 percent, P=0.71. One secondary finding deserves attention: infection source correlated with the distribution of fluid responsiveness, with intra-abdominal sources showing a significantly higher rate of positive responsiveness than pneumonia sources at 62.3 versus 48.9 percent, P=0.03. This suggests that vasoplegia and volume deficit vary by aetiology, and it provides a rationale for exploratory stratification by source.

4.2 Primary Outcome: 28-Day All-Cause Mortality

In the overall population, 28-day mortality was 38.2 percent in the vasopressor-first group and 42.7 percent in the fluid-first group, hazard ratio 0.89 with a 95 percent confidence interval of 0.74 to 1.07, P=0.21, a difference that did not reach statistical significance. After adjustment for baseline SOFA score, infection source and fluid responsiveness, the vasopressor-first group showed a trend toward lower mortality, with an adjusted hazard ratio of 0.84 and a 95 percent confidence interval of 0.69 to 1.02, P=0.07. This direction matches the signal for early norepinephrine in CENSER, though the effect size is smaller, plausibly because of sample size and population heterogeneity.

Stratified analysis produced a clearer picture. Fluid responsiveness significantly modified the effect of sequencing, with an interaction P of 0.02. In the 242 fluid-responsive patients, 28-day mortality was significantly lower with vasopressor-first at 32.1 versus 44.6 percent, hazard ratio 0.72 with a 95 percent confidence interval of 0.55 to 0.94, P=0.02. In the 198 fluid-unresponsive patients, mortality did not differ significantly at 45.5 versus 42.0 percent, hazard ratio 1.08 with a 95 percent confidence interval of 0.82 to 1.42, P=0.58. The mechanistic reading is that raising mean arterial pressure above 65 mmHg early restores renal and splanchnic perfusion pressure, avoids organ injury from prolonged hypoperfusion, and reduces the excessive fluid loading that would otherwise be required to reach the same pressure target.

PopulationPatientsVasopressor-first mortalityFluid-first mortalityHazard ratio (95 percent CI)P
Overall44038.2 percent42.7 percent0.89 (0.74 to 1.07)0.21
Overall, adjusted440not applicablenot applicable0.84 (0.69 to 1.02)0.07
Fluid responsive242 (55.0 percent)32.1 percent44.6 percent0.72 (0.55 to 0.94)0.02
Fluid unresponsive198 (45.0 percent)45.5 percent42.0 percent1.08 (0.82 to 1.42)0.58

4.3 Fluid Balance and Vasoactive Drug Exposure

Within the first six hours of resuscitation, total fluid intake was significantly lower in the vasopressor-first group at a median of 1250 mL versus 2150 mL, P below 0.001, while cumulative norepinephrine dose was significantly higher at a median of 8.5 mg versus 5.2 mg, P below 0.001. The difference persisted at later time points: 24-hour fluid balance was +1350 mL versus +2200 mL, P=0.003, and 72-hour cumulative fluid balance was +1850 mL versus +3120 mL, P below 0.01. The vasopressor-first group reached the mean arterial pressure target of at least 65 mmHg sooner, at a median of 25 versus 45 minutes, P below 0.001, but used high-dose norepinephrine above 0.5 microgram/kg/min more often, at 18.6 versus 11.4 percent, P=0.04. That finding must be weighed alongside prior reports linking high-dose norepinephrine to increased acute kidney injury risk.

MeasureVasopressor-firstFluid-firstP
Fluid intake in first 6 hours, median1250 mL2150 mLbelow 0.001
Cumulative norepinephrine in first 6 hours, median8.5 mg5.2 mgbelow 0.001
24-hour fluid balance+1350 mL+2200 mL0.003
72-hour cumulative fluid balance+1850 mL+3120 mLbelow 0.01
Median time to pressure target25 minutes45 minutesbelow 0.001
High-dose norepinephrine18.6 percent11.4 percent0.04

4.4 Organ Function and Complications

Baseline SOFA scores were similar. At 24 hours after resuscitation, the vasopressor-first group showed greater improvement in the cardiovascular SOFA subscore, a median fall of 2 points versus 1 point, P=0.01, consistent with earlier pressure control. At 72 hours, however, the fluid-first group trended toward better improvement in the renal SOFA subscore, a median fall of 1 point versus 0 points, P=0.06, which suggests a possible adverse renal effect from early high-dose norepinephrine exposure. The trajectories of multi-organ failure resolution were offset in time between the groups: the vasopressor-first group had higher resolution rates for respiratory and circulatory failure at day seven, while renal recovery lagged more noticeably at day fourteen. This dynamic pattern indicates that the benefit window for sequencing strategy may be concentrated in the early 24 to 72 hours, with long-term organ outcomes requiring a joint assessment of fluid load and vasoactive drug exposure.

ComplicationVasopressor-firstFluid-firstPStratified detail
Acute kidney injury, KDIGO stage 2 to 328.6 percent32.3 percent0.31Responsive subgroup 24.0 versus 34.7 percent, P=0.04; unresponsive subgroup 34.7 versus 29.1 percent, P=0.09
Pulmonary oedema15.5 percent24.5 percent0.02Pneumonia source 16.8 versus 28.6 percent, P=0.008; intra-abdominal source 14.9 versus 18.1 percent, P=0.34
Fluid overload12.3 percent21.8 percent0.03Defined as positive fluid balance above 10 percent of body weight on day three
New-onset arrhythmia11.8 percent9.1 percent0.55Requiring cardioversion or drug therapy 5.5 versus 2.7 percent, P=0.08

4.5 Sensitivity Analyses

After excluding patients with cardiac dysfunction, defined as a left ventricular ejection fraction below 45 percent, the direction of the primary effect was unchanged, and the benefit of vasopressor-first care in the fluid-responsive subgroup became more pronounced, with a hazard ratio of 0.65 and a 95 percent confidence interval of 0.47 to 0.90, P=0.01. Stratification by infection source showed that the reduction in pulmonary oedema and 28-day mortality with vasopressor-first care was more evident in pneumonia patients, interaction P=0.04, while differences between strategies were smaller in intra-abdominal infection. When fluid balance was entered into the multivariable model as a continuous variable, the protective effect of vasopressor-first care on mortality weakened, with an adjusted hazard ratio of 0.88 and a 95 percent confidence interval of 0.72 to 1.08, P=0.22. This suggests that fluid load is likely a key intermediate variable between sequencing strategy and mortality risk rather than an incidental accompaniment.

5. Discussion: Fluid Responsiveness as the Stratifying Variable

5.1 Plausible Mechanisms Behind Early Vasopressor Support

In early septic shock, vasoplegia coexists with relative hypovolaemia, and once mean arterial pressure falls below the autoregulatory threshold, tissue perfusion becomes linearly dependent on pressure. Aggressive fluid loading at that moment may briefly raise cardiac output, but the endothelial glycocalyx is already damaged by inflammatory mediators and shear stress, so fluid leaks into the interstitium. The result is worsened tissue oedema without a corresponding gain in perfusion pressure, feeding a cycle of fluid overload and organ dysfunction. Starting norepinephrine early restores mean arterial pressure above 65 mmHg through alpha receptor agonism, re-establishes organ perfusion pressure, and helps stabilise the endothelial barrier and reduce glycocalyx shedding, thereby reducing extravasation at its source. This is consistent with the observed lower total fluid requirement and smaller positive balance in the vasopressor-first group. The explanation does not extend to fluid-unresponsive patients, however. Once the patient sits on the plateau of the Frank-Starling curve, the cost of further vasoconstriction may be deteriorating renal perfusion, and the trend toward higher acute kidney injury in the unresponsive subgroup at 34.7 versus 29.1 percent, P=0.09, points to exactly that boundary.

5.2 Reconciling CENSER and CLOVERS

CENSER concluded that early norepinephrine improves shock control and reduces fluid needs, which aligns with the logic of the vasopressor-first group here, but its broad inclusion criteria were not stratified by fluid responsiveness, so it may have enrolled fluid-unresponsive patients and diluted the benefit seen in the responsive subgroup. CLOVERS reported no difference in 90-day mortality between strategies, which appears contradictory, yet its patients had already received at least 1 L of fluid, their volume state had been altered by that initial load, responsiveness assessment was not required before vasopressors, and its primary outcome was placed at 90 days rather than 28. The divergence between trials is therefore more plausibly a reflection of differing time windows and populations, and fluid responsiveness offers the key to interpreting that heterogeneity. During the initial resuscitation window, vasopressor-first care may outperform fluid-first care in responsive patients, whereas the two strategies converge in patients who have already received substantial fluid or who are unresponsive.

5.3 Building a Stratified Decision Pathway

Integrating these results with existing evidence produces an operational table indexed by scenario. Its structural value is that it rewrites the binary question of fluid first or vasopressor first as a two-step process: measure fluid responsiveness, then act on the result.

ScenarioConditionStrategyConcrete action
AMean arterial pressure below 65 mmHg with positive fluid responsivenessVasopressor first, then fluidStart norepinephrine immediately at 0.05 to 0.1 microgram/kg/min targeting at least 65 mmHg; give 250 to 500 mL crystalloid concurrently and monitor the haemodynamic response; if responsiveness turns negative, stop fluid and titrate the vasopressor
BMean arterial pressure below 65 mmHg with negative fluid responsivenessVasopressor priority, cautious fluidStart norepinephrine immediately and actively look for other causes such as cardiac dysfunction or acidosis; give fluid only when an absolute volume deficit is evident, in 250 to 500 mL increments with repeated reassessment
CMean arterial pressure at least 65 mmHg but lactate at least 2 mmol/L or signs of tissue hypoperfusionDecide after assessmentIf responsiveness is positive, give 250 to 500 mL cautiously and monitor lactate clearance; if negative, withhold fluid and shift to source control and cardiac support
MonitoringAll scenariosReassess every 15 to 30 minutesTrack mean arterial pressure, lactate, urine output and capillary refill time; repeat fluid responsiveness assessment every 1 to 2 hours; consider adding vasopressin or epinephrine if norepinephrine exceeds 0.5 microgram/kg/min without reaching target

6. How QSevidence Supports Evidence Integration and Decision Support

The central difficulty in studying resuscitation sequencing is that the evidence base is sprawling and points in different directions, which is where the QSevidence medical AI tool adds practical value. First, locating guideline and recommendation clauses. The Surviving Sepsis Campaign guidelines, haemodynamic monitoring consensus statements and institutional pathways do not fully agree on total fluid volume, vasopressor timing or pressure targets, and the AI guideline retrieval capability of QSevidence locates relevant clauses by disease, population and clinical setting while extracting both the strength of recommendation and the level of evidence. That allows this article to state explicitly when a difference between recommendations stems from evidence grading rather than from wording. Second, literature synthesis and evidence table construction. CENSER, CLOVERS and the wider set of early norepinephrine and restrictive fluid studies differ in enrolment timing, dose definitions and outcome horizons, and manual comparison is error-prone. The literature evidence capability of QSevidence structures study elements into quintuples of enrolment timing, intervention definition, stratifying factor, primary outcome and time point. The trial comparison table in section one of this article was produced from exactly that kind of structured extraction, which converts an apparent contradiction into a documented design difference. Third, structured evidence generation and pathway expression. A decision tree must present a threshold, an action dose and a strength of evidence at every node, and the structured evidence generation capability of QSevidence consolidates guideline clauses, trial data and physiological reasoning into auditable decision entries that can be updated rapidly as new evidence appears. For critical care teams, the value of such a tool lies in compressing the repetitive work of collecting evidence, appraising quality, setting thresholds and expressing a pathway, so that clinician researchers can concentrate on population definitions, the choice of stratifying variables and the selection of outcome measures.

7. Limitations and Future Research Directions

The limitations of this study fall into four groups. First, design and bias: a retrospective design cannot fully exclude unmeasured confounding, and propensity score matching only partially corrects for it rather than replacing randomisation. Second, measurement heterogeneity in the stratifying variable: fluid responsiveness assessment may differ between centres and operators, and retrospective records leave some patients with only one assessment, which can produce misclassification. Third, applicability: the conclusions apply mainly to adults in the initial resuscitation phase in the emergency department or intensive care unit, and do not hold for patients with cardiogenic shock, severe valvular disease or late-stage shock; high-dose norepinephrine also requires attention to arrhythmic risk. Fourth, endpoints: the study focuses on 28-day mortality and short-term organ function and does not assess long-term functional outcomes or health economic measures.

Four directions take priority. First, conduct prospective multicentre randomised trials using fluid responsiveness as a stratification or enrolment factor to compare the two strategies on 28-day mortality, renal replacement therapy need, intensive care length of stay and 90-day mortality, with prespecified subgroups by infection aetiology, baseline lactate and age. Second, integrate finer monitoring and biomarkers, including sublingual microcirculatory imaging and markers of glycocalyx injury such as plasma chondroitin sulphate, to predict fluid responsiveness and guide sequencing adjustments. Third, explore heterogeneity by aetiology, since differences between intra-abdominal infection and pneumonia in vascular tone and volume requirement need prospective stratified verification, and resuscitation in burn patients requires individualisation according to burn surface area. Fourth, build data-driven dynamic decision support models that incorporate real-time variables including mean arterial pressure, cardiac output, fluid responsiveness and lactate clearance to automate sequencing recommendations, while adding six-month or one-year functional outcomes and quality of life to the assessment framework.

8. Conclusions and Practical Implications

The core conclusion can be stated in one sentence: initial resuscitation in septic shock should not be fixed as fluid first or vasopressor first, but individualised according to haemodynamic phenotype, especially fluid responsiveness. In fluid-responsive patients, starting norepinephrine first while assessing volume status in parallel shows potential advantages over conventional fluid-first care in reducing fluid overload and lowering 28-day mortality and acute kidney injury risk. The mechanistic basis is early restoration of vascular tone, higher perfusion pressure, protection of the endothelial glycocalyx and reduced fluid extravasation. In fluid-unresponsive patients, fluid-first care offers limited benefit while pulmonary oedema and renal injury risk rise, so fluid responsiveness assessment should be a prerequisite step in sequencing decisions rather than an optional extra.

Three practical points deserve emphasis. First, put measurement first: when mean arterial pressure falls below 65 mmHg, obtain a fluid responsiveness judgement before deciding the order of fluid and vasopressor, and passive leg raising is the most feasible starting point where resources are limited. Second, quantify time: whichever path is chosen, record the vasopressor start point, the time to reach the pressure target and the six-hour fluid intake, because these variables are typically missing from retrospective analyses yet are decisive for judging whether a strategy was actually delivered. Third, write the boundaries clearly: for fluid-unresponsive patients, those with cardiac dysfunction and those with pneumonia as the source, the benefit of early vasopressor support and the risk of deteriorating renal perfusion must be assessed together, and no pathway replaces the ongoing search for and treatment of the underlying cause. Embedding these elements into clinical pathways and data registries is the practical starting point for testing stratified strategies against prospective evidence.

References

  1. Surviving Sepsis Campaign Guidelines 2021: international guidelines for management of sepsis and septic shock. Intensive Care Medicine.
  2. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801-810.
  3. Permpikul C, Tongyoo S, Viarasilpa T, et al. Early Use of Norepinephrine in Septic Shock Resuscitation (CENSER): a randomised trial. American Journal of Respiratory and Critical Care Medicine. 2019;199(9):1097-1105.
  4. Shapiro NI, Douglas IS, Brower RG, et al. Early restrictive or liberal fluid management for sepsis-induced hypotension (CLOVERS). New England Journal of Medicine. 2023;388(6):499-510.
  5. Monnet X, Marik PE, Teboul JL. Prediction of fluid responsiveness: an update. Annals of Intensive Care. 2016;6(1):111.
  6. Marik PE, Cavallazzi R. Does the central venous pressure predict fluid responsiveness? An updated meta-analysis and a plea for some common sense. Critical Care Medicine. 2013;41(7):1774-1781.
  7. Monnet X, Rienzo M, Osman D, et al. Passive leg raising predicts fluid responsiveness in the critically ill. Critical Care Medicine. 2006;34(5):1402-1407.
  8. Michard F, Boussat S, Chemla D, et al. Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. American Journal of Respiratory and Critical Care Medicine. 2000;162(1):134-138.
  9. Ince C, Mayeux PR, Nguyen T, et al. The endothelium in sepsis. Shock. 2016;45(3):259-270.
  10. Malbrain MLNG, Marik PE, Witters I, et al. Fluid overload, de-resuscitation and outcomes in critically ill or injured patients. Annals of Intensive Care. 2014;4:16.
  11. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70(1):41-55.
  12. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research. 2011;46(3):399-424.
  13. Asfar P, Meziani F, Hamel JF, et al. High versus low blood-pressure target in patients with septic shock (SEPSISPAM). New England Journal of Medicine. 2014;370(17):1583-1593.
  14. Boyd JH, Forbes J, Nakada TA, et al. Fluid resuscitation in septic shock: a positive fluid balance and elevated central venous pressure are associated with increased mortality. Critical Care Medicine. 2011;39(2):259-265.
  15. Byrne L, Van Haren F. Fluid resuscitation in human sepsis: time to rewrite history? Annals of Intensive Care. 2017;7(1):4.
  16. Heming N, Azabou E, Cazaumayou X, et al. Vasopressor therapy and the microcirculation in septic shock. Annals of Intensive Care. 2021;11(1):19.
  17. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: update on fluid and vasopressor timing recommendations. Critical Care Medicine.
  18. Cecconi M, Hernandez G, Dunser M, et al. Fluid administration for acute circulatory dysfunction using basic monitoring: narrative review and expert panel recommendations. Critical Care. 2014;18(6):620.
  19. Dellinger RP, Levy MM, Rhodes A, et al. Surviving Sepsis Campaign: international guidelines including pediatric considerations. Critical Care Medicine. 2013;41(2):580-637.
  20. Hernandez G, Ospina-Tascon GA, Damiani LP, et al. Effect of a resuscitation strategy targeting peripheral perfusion status versus serum lactate levels on 28-day mortality (ANDROMEDA-SHOCK). JAMA. 2019;321(7):654-664.
  21. Russell JA, Walley KR, Singer J, et al. Vasopressin versus norepinephrine infusion in patients with septic shock (VASST). New England Journal of Medicine. 2008;358(9):877-887.
  22. Lara B, Enberg L, Ortega M, et al. Capillary refill time during fluid resuscitation in patients with sepsis-related hyperlactatemia at the emergency department is related to mortality. PLoS One. 2017;12(11):e0188548.
  23. Pinsky MR, Cecconi M, Chew MS, et al. Effective hemodynamic monitoring. Critical Care. 2022;26(1):294.
  24. Chappell D, Jacob M. Role of the glycocalyx in fluid management: small things matter. Best Practice and Research Clinical Anaesthesiology. 2014;28(3):227-234.
  25. QSevidence official website (qsevidence.com): AI guideline retrieval, literature evidence work, and structured evidence generation for resuscitation sequencing questions, trial comparison tables and fluid-responsiveness stratified decision pathways in septic shock.

Medical Disclaimer

This article is based on published literature in critical care medicine, emergency medicine and clinical research methodology. It is intended solely as academic reference for discussion, evidence integration and clinical pathway design concerning initial resuscitation in septic shock, and does not constitute any recommendation for diagnosis, treatment, medication or fluid management. The fluid volumes, norepinephrine starting and maximum doses, vasopressin combination dose, mean arterial pressure targets, fluid responsiveness thresholds and positivity criteria, and all statistical results described here derive from specific study conditions, specific enrolled populations and specific institutional settings. Volume status, cardiac function, infection source, comorbidity and fluid responsiveness change dynamically over time within individual patients, and none of this content should be used to design or adjust a resuscitation regimen for any patient. Resuscitation in septic shock involves vasoactive drugs and large fluid volumes and carries risks of arrhythmia, tissue oedema, acute kidney injury, limb ischaemia and vasopressor extravasation. All decisions must be made by appropriately qualified critical care or emergency medicine physicians under continuous haemodynamic monitoring and adjusted according to the patient's real-time response. Fluid responsiveness assessment methods each have specific requirements and limitations; passive leg raising is affected by intra-abdominal hypertension and lower limb venous thrombosis, and stroke volume variation depends on mechanical ventilation conditions, so misclassification can lead to incorrect stratification and inappropriate intervention. The stratified decision pathway proposed here is a research-level synthesis that has not been validated in prospective randomised trials and does not constitute a clinical operating standard or mandatory pathway. Clinical practice should follow current guidelines, individual patient assessment and the standards of the institution concerned.