What the Evidence Actually Shows for the Sepsis Hour-1 Bundle

Sepsis is life-threatening organ dysfunction caused by a dysregulated response to infection. Septic shock is a high-risk subset with persistent circulatory and metabolic abnormalities. Delay in effective treatment can be deadly, especially when shock is present. At the same time, many noninfectious emergencies resemble sepsis, and unnecessary broad-spectrum antibiotics and fluid can cause harm.

The “Hour-1 Bundle” was created to make early actions occur together: measure lactate, obtain blood cultures before antibiotics when that does not delay treatment, give broad-spectrum antimicrobials, begin fluid for hypotension or marked lactate elevation, and use vasopressors when hypotension persists. Evidence strength is not equal across every item on that list, or across every patient you will see.

Where the one-hour idea came from#

Early sepsis campaigns responded to delayed recognition, late antibiotics, inadequate resuscitation, and fragmented care. Bundles simplified a complex emergency into observable actions and deadlines. Hospitals built alerts, sepsis teams, order sets, and audit systems around them.

The one-hour target was supported largely by biological plausibility and observational associations between antibiotic delay and mortality, and in shock, each interval without active treatment allows uncontrolled infection and organ injury. Randomly assigning people with septic shock to deliberate delay would be unethical.

Observational evidence is still vulnerable to bias. Recognition time can be documented after clinicians have already acted. Sicker patients may be recognized sooner but die more often, creating confounding in one direction; difficult diagnoses may receive late therapy and have worse outcomes, creating it in another. “Time zero” is not a natural timestamp.

What the New York mandate study found#

After New York required protocolized sepsis care, a large observational study examined more than 49,000 patients. Longer time to completion of the three-hour bundle and longer time to antibiotics were associated with higher in-hospital mortality, but time to completion of an initial fluid bolus was not significantly associated with mortality.

This supports urgency, particularly for antibiotics, but it does not prove every bundle element causes the observed difference. Hospitals that deliver care faster may differ in staffing, recognition, source control, and overall quality, and residual confounding remains after adjustment. Note also what the study measured: a three-hour bundle. Do not read it as randomized proof of a universal one-hour threshold, because risk likely changes continuously and differs by shock, organism, source, and adequacy of the selected drug.

The 2026 guideline is more explicit about probability#

The 2026 Surviving Sepsis Campaign separates clinical groups. For possible, probable, or definite septic shock, it strongly recommends antimicrobials immediately, ideally within one hour. It gives the same timing for probable or definite sepsis without shock. Certainty is rated very low.

For possible sepsis without shock, the current recommendation table suggests a time-limited rapid investigation. If concern for infection persists, antimicrobials should be administered within three hours from first suspicion. If infection likelihood is low and there is no shock, deferral with close monitoring is suggested.

This is not permission to ignore early deterioration. It is recognition that a stable patient with one nonspecific alert differs from a person with mottling, hypotension, altered mentation, and a likely infection; diagnostic work and treatment can proceed in parallel.

Antibiotic timing is only useful if the drug is active#

Fast administration of an antibiotic that does not cover the pathogen is not effective early therapy. Your empiric selection depends on suspected source, community versus health-care acquisition, prior cultures, recent antibiotics, immune status, devices, allergies, local resistance, and organ function.

Broader is not always safer. Unnecessary duplicate gram-negative, anaerobic, MRSA, or antifungal coverage adds toxicity and selection pressure, and the 2026 guideline recommends risk-based coverage for multidrug-resistant organisms and advises against extra coverage when risk is low.

Cultures should be obtained as soon as possible and ideally before antibiotics, provided that collection does not meaningfully delay urgent treatment. Adequate blood volume and multiple sets improve yield. Cultures after treatment can be negative despite infection.

Reassessment completes the antibiotic decision#

Your empiric decision is provisional. At 24 to 72 hours, microbiology, imaging, clinical course, operative findings, biomarkers, and alternative diagnoses can support narrowing, stopping, or changing therapy.

The 2026 guideline states that clinicians should continuously reevaluate, seek noninfectious causes, and discontinue empiric therapy when another cause is demonstrated or strongly suspected; it recommends de-escalation when a confirmed organism and susceptibility profile are available. This makes stewardship part of the emergency pathway, not a later objection to it. A system should measure prompt appropriate initiation and prompt correction.

Lactate is a risk and perfusion marker#

Lactate can rise when oxygen delivery is inadequate, adrenergic metabolism accelerates, clearance falls, or other metabolic processes change; elevated values in sepsis correlate with worse outcomes, and serial decline can contribute to resuscitation assessment.

It is not specific. Seizures, severe asthma treatment, epinephrine, liver dysfunction, metformin toxicity, ischemia, intense exertion, and other conditions can raise lactate. A normal lactate does not exclude serious infection or organ dysfunction. The current guideline suggests measuring it in possible, probable, or definite sepsis or shock. Drawing it should not become a reason to delay airway support, antibiotics, vasopressors, or source control.

Fluids are a treatment with a dose-response curve#

The traditional bundle called for rapid 30 milliliters per kilogram of crystalloid for hypotension or lactate at least 4 millimoles per liter. The 2026 guideline suggests at least 30 milliliters per kilogram during the first three hours for sepsis-induced hypoperfusion or septic shock, not a mandatory one-hour infusion for every positive screen.

It explicitly calls for individual patient factors and frequent reassessment to avoid under- or over-resuscitation. Heart failure, kidney failure, cirrhosis, pregnancy, age, prior fluids, and respiratory failure can alter tolerance. A person with profound volume loss differs from someone with cardiogenic and septic physiology together. Dynamic assessments, capillary refill, blood pressure, urine output, mentation, skin perfusion, bedside ultrasound, passive leg raise, and response to small boluses can inform further fluid. No one measure is perfect.

Vasopressors should not wait for endless fluid#

When hypotension persists, norepinephrine is generally the first vasopressor. Current guidance supports starting vasopressors through an appropriate peripheral vein rather than waiting for central access, with monitoring for infiltration and a plan for access management.

In unstable shock, fluids and vasopressors may begin concurrently. The initial mean arterial pressure target is around 65 millimeters of mercury for most adults, with individualization. Restoring pressure does not guarantee microcirculatory recovery, and escalating to a high number can add arrhythmia or ischemia, so the bundle's clock should accelerate hemodynamic support, not force completion of a fixed fluid amount before vasopressors are allowed.

Source control is not a checkbox in the hour#

Antibiotics cannot drain an abscess, remove infected hardware, debride necrotic tissue, or relieve an obstructed infected urinary system. The current guideline recommends rapid evaluation for an anatomical source requiring intervention and suggests control ideally within six hours when required.

Timing depends on procedure feasibility, stabilization, imaging, surgical or interventional resources, and risk. Delay can be fatal, but a rushed unsafe procedure can also harm. A sepsis program needs escalation pathways beyond medication orders.

Sepsis screening creates false positives by design#

Screening tools favor sensitivity because missed sepsis is dangerous. Fever, tachycardia, rapid breathing, low blood pressure, confusion, white-cell changes, and organ dysfunction also occur with hemorrhage, pulmonary embolism, pancreatitis, adrenal crisis, medication reactions, trauma, and other conditions.

The Sepsis-3 definitions conceptualize sepsis as infection plus organ dysfunction, not infection alone and not a nonspecific inflammatory response. Definitions support research and recognition; they do not replace the diagnosis you make at the bedside. The 2026 guideline says sepsis should not be ruled in or out with a single biomarker or diagnostic test, and a positive alert should trigger a bedside huddle and rapid assessment, not automatic certainty.

SEP-1 is not the Hour-1 Bundle#

The U.S. Centers for Medicare and Medicaid Services currently reports SEP-1, a composite measure of specified management for severe sepsis and septic shock. It has detailed time windows, abstraction rules, inclusion criteria, and all-or-none scoring.

The Surviving Sepsis Campaign guideline is a clinical document updated through evidence review and grading. Its hour-one framework and later stratified recommendations are not identical to SEP-1. A hospital can improve measure compliance without proving that every required action caused better outcomes.

A joint professional-society position paper raised concerns that broad quality measures can drive antibiotic overuse and recommended focusing mandates more narrowly on septic shock, among other revisions. Quality policy must balance missed treatment against harm from false positives.

Why certainty can be very low despite a strong recommendation#

GRADE separates certainty of evidence from recommendation strength: the one-hour recommendation is strong for shock because the condition is lethal, effective antimicrobials are necessary, biological rationale is compelling, and waiting for a definitive trial is unacceptable. Evidence certainty is very low because direct randomized timing evidence is unavailable and observational studies have bias.

That combination is not hypocrisy. It is a transparent value judgment under uncertainty. It also means you should apply the boundary carefully. Extending a shock recommendation to every mildly abnormal screen changes the benefit-harm balance.

A clinically faithful early pathway#

Recognize instability and begin airway, breathing, circulation, monitoring, and access. Assess infection probability and likely source. Obtain cultures and lactate without delaying urgent care. Give active empiric antimicrobials immediately in shock or probable or definite sepsis. For possible sepsis without shock, investigate rapidly and treat within three hours if concern persists.

Use fluids in reassessed increments consistent with the person's physiology. Start vasopressors early when needed. Arrange source control. Document time zero and diagnostic certainty honestly. Reassess antibiotics daily and stop them when infection becomes unlikely.

This is more sophisticated than either “the clock is everything” or “bundles are useless.” It preserves urgency where benefit is most likely and diagnostic care where indiscriminate treatment can harm.

References#

1. Surviving Sepsis Campaign 2026 guideline
2. SCCM current recommendation table
3. New York timing study
4. Joint society quality-measure position
5. CMS current SEP-1 reporting page
6. Sepsis-3 consensus definitions