Surviving Sepsis Campaign "golden hour" — lactate, cultures, antibiotics, fluids, and vasopressors raced against a 60-minute clock
The Hour-1 bundle begins the instant sepsis is recognized — not when a bed opens up or a physician is free. A serum lactate is the fastest, cheapest signal of tissue hypoperfusion available: it can be run on a point-of-care blood gas analyzer in minutes and requires no imaging, no culture incubation, and no waiting. It anchors two things at once — a severity threshold that helps classify septic shock, and a resuscitation endpoint that can be re-checked to see whether treatment is working.
Lactate rises when tissue oxygen delivery cannot meet demand — from hypoperfusion, from mitochondrial dysfunction in sepsis itself, or, less commonly, from adrenergic-driven aerobic glycolysis. Whatever the exact mechanism in a given patient, an elevated lactate is a robust, reproducible marker that correlates with illness severity and mortality risk across multiple sepsis studies.
Because the assay is fast and requires only a blood draw, it is the ideal first move: it does not compete for time against anything else in the bundle, and its result actively shapes the next steps — a lactate ≥4 mmol/L is itself the trigger for the 30 mL/kg fluid bolus, tying element 1 directly to element 4.
A single lactate value is a snapshot; the trend is the therapeutic signal. Surviving Sepsis Campaign guidance recommends re-measuring lactate within 2–4 hours in patients with an initial elevated value, using clearance (a falling trend, typically ≥10–20% per interval) as one marker that resuscitation is on the right track.
Lactate can also be falsely reassuring or falsely alarming — liver dysfunction slows clearance independent of perfusion, and some medications (e.g., epinephrine) raise lactate directly — so it is interpreted alongside mean arterial pressure, urine output, capillary refill, and mental status, never in isolation.
A normal initial lactate does not rule out septic shock, and a single elevated value does not by itself confirm it — lactate is one input into a clinical gestalt that also includes blood pressure, perfusion exam, and organ dysfunction scoring (SOFA/qSOFA).
Antibiotics sterilize blood cultures within hours of the first dose, so if cultures are going to identify a causative organism at all, they need to be drawn before that first dose is infused. The Hour-1 bundle asks for exactly that — two sets, from two separate venipuncture sites, ideally including at least one peripheral draw — obtained in the same short window as everything else, without ever becoming the reason antibiotics are delayed.
The instruction is "before antibiotics," not "instead of timely antibiotics." In practice this means a clinician or nurse draws two culture sets in the same few minutes that the antibiotic order is being verified and the first bag is being spiked — parallel workflow, not serial delay. If cultures cannot be obtained without a meaningful delay (for example, difficult IV access), guidance is explicit: give the antibiotics anyway. A missed culture is a diagnostic setback; a delayed antibiotic in septic shock is a mortality risk.
Even a single dose of antibiotics can drop blood culture positivity substantially, so the sequencing genuinely matters when it can be achieved quickly — it is worth the extra two minutes, just never worth the extra thirty.
Cultures drawn at recognition pay off 24–72 hours later, when growth and susceptibility results allow de-escalation from broad empiric coverage to a narrow, targeted regimen — reducing resistance pressure, drug toxicity, and cost, without compromising the patient who already received timely empiric treatment.
Other cultures (urine, sputum, wound, line tips) are obtained as clinically indicated from the same infection-source workup, but blood cultures are the only ones explicitly time-gated against antibiotic administration in the Hour-1 bundle itself.
Two sets from two sites is the standard specifically because a single positive bottle cannot distinguish a true bloodstream pathogen from a skin-flora contaminant (e.g., coagulase-negative staphylococci) — concordant growth in both sets substantially raises confidence that the organism is real.
Of all five bundle elements, antibiotic timing carries the most direct, most frequently cited mortality signal. Broad-spectrum empiric antibiotics — chosen to cover the most likely pathogens and sources given the patient's presentation — are infused as soon as possible, targeted at or under 60 minutes from recognition. This is the element the "golden hour" framing was built around.
The landmark association between antibiotic delay and mortality in septic shock comes from a large retrospective cohort (Kumar et al., Critical Care Medicine 2006), which found survival fell in roughly stepwise fashion for each hour that effective antimicrobial administration was delayed after the onset of septic-shock-associated hypotension. Later studies have refined and nuanced this relationship — the effect is clearest in the sickest patients (frank shock) and less pronounced in less severe presentations — but the directional message has held up across the sepsis literature: sooner is better, and the first hour matters most.
This is why antibiotic timing, more than any other single bundle element, gets referenced by name in mortality and quality-improvement literature, and why it is the element most tightly coupled to the "golden hour" metaphor itself.
Empiric selection is guided by the suspected source (pulmonary, urinary, intra-abdominal, skin/soft tissue, line-associated, unknown), local antibiogram resistance patterns, patient risk factors for resistant organisms (recent hospitalization, prior resistant infection, immunosuppression), and allergy history — all triaged rapidly, because the goal is coverage broad enough to be very likely effective against the true pathogen, given while the identity of that pathogen is still unknown.
Once culture and susceptibility results return, therapy is de-escalated to the narrowest effective regimen — this is where element 2 (cultures before antibiotics) pays its dividend, and why the two elements are designed to work together rather than in tension.
"Broad-spectrum, fast" is a deliberate trade against "narrow, guided" — under time pressure with an unidentified organism and shock physiology, the mortality cost of getting coverage wrong or being late outweighs the stewardship cost of over-covering for 48–72 hours until de-escalation is possible.
When hypotension (MAP <65 mmHg) or a lactate ≥4 mmol/L signals shock physiology, the bundle calls for rapid administration of balanced crystalloid — a weight-based 30 mL/kg bolus, begun within the first hour — to restore circulating volume and tissue perfusion before the body's compensatory mechanisms are exhausted.
The 30 mL/kg figure is a pragmatic, weight-based starting dose intended to rapidly re-expand intravascular volume in a patient whose vasculature has become leaky and dilated under septic inflammation — not a precisely individualized prescription. It is deliberately simple so it can be started immediately, at the bedside, without waiting for invasive monitoring to first quantify exact fluid needs.
Balanced crystalloids (fluids with an electrolyte composition closer to plasma than 0.9% saline) are generally preferred over normal saline for large-volume resuscitation, based on evidence associating large saline volumes with hyperchloremic metabolic acidosis and, in some studies, worse renal outcomes. Albumin may be added in select patients requiring large crystalloid volumes, but crystalloid remains first-line.
30 mL/kg is a starting point, not a ceiling or a fixed protocol to complete regardless of response. Ongoing fluid administration beyond the initial bolus should be guided by repeated assessment of fluid responsiveness — dynamic measures such as passive leg raise, pulse pressure variation, or bedside ultrasound of the inferior vena cava — rather than by volume alone, because over-resuscitation carries real harm: pulmonary edema, worsened oxygenation, abdominal compartment syndrome, and tissue edema that impairs wound healing and organ function.
Critically, if MAP remains inadequate once the bolus is running, vasopressors are not deferred until the fluid is finished — element 5 can and should begin concurrently.
Fluid resuscitation and vasopressor initiation are not strictly sequential steps — SSC guidance explicitly allows starting vasopressors during ongoing fluid administration if MAP stays below target, rather than waiting for the full 30 mL/kg to complete.
Some patients remain hypotensive despite adequate fluid resuscitation — fluid-refractory shock. For them, the bundle calls for prompt initiation of vasopressors, with norepinephrine as first-line, targeting a mean arterial pressure of at least 65 mmHg to preserve perfusion to the brain, kidneys, and other vital organs while the underlying infection is brought under control.
Norepinephrine is favored as the first-line vasopressor in septic shock because it provides potent alpha-adrenergic vasoconstriction to restore vascular tone with comparatively less tachyarrhythmia risk than agents like dopamine, based on trial evidence comparing the two. A MAP target of 65 mmHg is used because it approximates the lower limit at which autoregulation in vital organ vascular beds (brain, kidney) can be expected to maintain adequate blood flow in most patients — targeting substantially higher has not shown consistent added benefit and can increase arrhythmia and other adverse effects.
Individual targets may be adjusted upward for patients with chronic hypertension whose autoregulatory curve is shifted, but 65 mmHg is the standard starting threshold.
A common historical pattern of care delayed vasopressors until a large fluid volume had been completely infused, treating pressors as a rescue therapy reserved for fluid failure. Current guidance pushes toward earlier initiation — vasopressors can be started peripherally for a short period while central access is obtained, and are titrated alongside ongoing fluid administration rather than after it, because prolonged untreated hypotension itself worsens organ injury regardless of its eventual cause.
If norepinephrine doses climb without achieving target MAP, vasopressin is typically added as a second agent rather than escalating a single drug indefinitely, and refractory cases may prompt evaluation for adjunctive corticosteroids and reassessment of source control.
The clinical shift from "fluids first, pressors only if that fails" to "start pressors alongside fluids when MAP is inadequate" reflects the same golden-hour logic driving the whole bundle: time spent hypotensive, regardless of which single intervention is technically "in progress," is time organs are underperfused.