Critical Care Deep Pathophysiology

Sepsis: The Complete Pathophysiology

A comprehensive exploration of the molecular, cellular, and systemic mechanisms that transform infection into life-threatening organ dysfunction—and why understanding these mechanisms transforms clinical care.

Section I

Foundations: What Sepsis Really Is

Sepsis is not simply infection. It is a dysregulated host response to infection that causes life-threatening organ dysfunction. Understanding this distinction is the foundation of everything that follows.

The Evolution of Understanding

From SIRS to Sepsis-3: Why Definitions Matter

The definition of sepsis has evolved as our understanding of pathophysiology has deepened. Each redefinition reflects new mechanistic insights.

1991 — Sepsis-1 (SIRS-based)
Sepsis = Infection + ≥2 SIRS criteria (temp, HR, RR, WBC). Problem: SIRS criteria too sensitive—pancreatitis, burns, post-surgical patients meet criteria without infection. The definition captured inflammation, not dysregulation.
2001 — Sepsis-2
Expanded criteria but kept SIRS foundation. Added clinical signs of organ dysfunction. Still lacked specificity—patients could have "sepsis" without meaningful risk of death.
2016 — Sepsis-3 (Current)
Sepsis = Life-threatening organ dysfunction caused by dysregulated host response to infection. Operationalized as infection + SOFA score increase ≥2. This definition captures what actually kills patients: not infection alone, but the body's catastrophic response to it.

The Paradigm Shift

Sepsis-3 recognizes that sepsis is fundamentally a problem of HOST RESPONSE, not just pathogen presence. Two patients with identical infections can have vastly different outcomes based on how their immune and physiological systems respond. This is why sepsis is so heterogeneous—and why single-target therapies have repeatedly failed.

Core Concept

The Central Question: Why Does Infection Become Sepsis?

Every day, humans successfully fight off countless infections. Why do some infections trigger the catastrophic dysregulation we call sepsis?

Why does infection sometimes cause organ dysfunction?

Because the immune response, designed to be local and contained, becomes systemic and dysregulated. Mediators meant to fight infection at the site of invasion instead circulate throughout the body, activating inflammatory cascades in organs far from the original infection.

Why does the immune response become systemic?

When local containment fails—either because pathogen burden is overwhelming, host defenses are compromised, or the infection reaches the bloodstream—pattern recognition receptors throughout the body detect danger signals and activate simultaneously. The response is no longer proportional to the threat.

Why can't the body "turn off" the response once it starts?

The inflammatory response has positive feedback loops designed to amplify the signal for effective pathogen killing. Cytokines induce more cytokine production; complement activates more complement; coagulation activates more coagulation. These amplification loops, essential for local infection control, become destructive when activated systemically.

Why does inflammation cause organ dysfunction?

The inflammatory response causes endothelial activation and dysfunction throughout the body. This leads to: (1) Vasodilation → hypotension, (2) Increased permeability → edema and hypovolemia, (3) Procoagulant state → microvascular thrombosis, (4) Impaired oxygen extraction → cellular hypoxia despite adequate delivery. Every organ depends on its microcirculation; when the endothelium fails globally, every organ fails.

Why do patients die from sepsis?

Death results from refractory shock (cardiovascular collapse), multi-organ failure (accumulated dysfunction exceeds physiological reserve), or late immunosuppression (inability to clear initial infection or secondary infections). The same immune system that over-responded early often under-responds late—sepsis is not one disease but a dynamic trajectory through multiple pathophysiological states.

Epidemiology

The Scope of the Problem

49M
Cases per year (global)
11M
Deaths per year (global)
20-30%
Hospital mortality
#1
Cause of ICU death

Why Mortality Remains High Despite Advances

Heterogeneity: "Sepsis" encompasses vastly different host responses, pathogen types, and infection sources. No single treatment can address this diversity.

Time-Dependence: Pathophysiology evolves rapidly. The hyperinflammatory patient at hour 6 may be immunosuppressed at hour 72. Treatments appropriate for one phase may harm in another.

Complexity: Multiple simultaneous pathways (immune, coagulation, endothelial, metabolic) interact in ways that single-target therapies cannot address.

Late Recognition: By the time patients meet clinical criteria, pathophysiology is often well-established. Earlier intervention might prevent the cascade, but early sepsis is difficult to distinguish from uncomplicated infection.

Evidence Base

Singer M, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016 • Rudd KE, et al. Global, regional, and national sepsis incidence and mortality. Lancet 2020 • Hotchkiss RS, et al. Sepsis and septic shock. Nat Rev Dis Primers 2016

Section II

Pathogen Recognition: How the Body Detects Danger

The immune response begins with detection. Pattern recognition receptors (PRRs) identify molecular signatures of pathogens and tissue damage, initiating the cascade that can save—or destroy—the host.

The Detection System

PAMPs, DAMPs, and PRRs: The Molecular Alarm System

How does the body know infection is present?

Pattern recognition receptors (PRRs) on immune cells detect conserved molecular patterns unique to pathogens—Pathogen-Associated Molecular Patterns (PAMPs). These patterns are essential to microbial survival and cannot be easily mutated away, making them reliable danger signals.

What are the major PAMPs?

LPS (lipopolysaccharide) from Gram-negative bacteria; lipoteichoic acid and peptidoglycan from Gram-positive bacteria; flagellin from motile bacteria; viral RNA and DNA; fungal β-glucan and mannan. Each activates specific PRRs, but the downstream inflammatory response is remarkably similar.

Why does sterile inflammation (trauma, pancreatitis) look like sepsis?

Because dying cells release Damage-Associated Molecular Patterns (DAMPs)—endogenous molecules normally hidden inside cells: HMGB1, mitochondrial DNA, ATP, uric acid, heat shock proteins. These DAMPs activate the SAME PRRs as PAMPs. The immune system cannot distinguish "danger from infection" from "danger from tissue damage."

Why does this dual recognition system exist?

Evolutionary advantage: the inflammatory response that kills bacteria also promotes wound healing. Using the same receptors for both pathogen detection and tissue damage detection allows a unified emergency response. The cost: excessive tissue damage can trigger sepsis-like physiology even without infection.

Why does this matter clinically?

It explains why SIRS criteria fail to distinguish sepsis from other inflammatory states—because at the molecular level, the response IS similar. It also explains why trauma patients, burn patients, and pancreatitis patients can develop "sepsis-like" multi-organ failure without infection, and why anti-inflammatory therapies that work in sterile inflammation might not work in sepsis (where you need SOME inflammation to clear pathogens).

The Key Receptors

Toll-Like Receptors: The Master Switches

Toll-like receptors (TLRs) are the best-characterized PRRs. Understanding which TLR detects what explains pathogen-specific responses.

TLR4: The LPS Receptor (Gram-Negatives)

Ligand: Lipopolysaccharide (LPS, endotoxin) from Gram-negative bacterial outer membrane.

Mechanism: LPS binds LPS-binding protein (LBP) → transfers to CD14 → presented to TLR4/MD-2 complex. This triggers both MyD88-dependent (rapid NF-κB activation) and TRIF-dependent (interferon response) pathways.

Clinical Relevance: TLR4 is why Gram-negative sepsis is classically more fulminant—LPS is an extraordinarily potent immune activator. As little as 1 ng/kg can cause fever in humans. Gram-negative bacteremia releases massive LPS loads.

Why Anti-LPS Therapies Failed: By the time patients present, LPS has already triggered the cascade. Blocking LPS doesn't reverse already-activated inflammation. Also, some TLR4 signaling is needed for bacterial clearance.

TLR2: Gram-Positive Recognition

Ligands: Lipoteichoic acid, peptidoglycan, lipoproteins from Gram-positive bacteria. Also recognizes some fungal and mycobacterial components.

Mechanism: Forms heterodimers with TLR1 or TLR6, expanding recognition range. Signals via MyD88.

Clinical Note: Gram-positive sepsis may have slightly different kinetics but ultimately triggers similar downstream pathways. Staph aureus toxins can act as superantigens (see below), adding another layer of immune activation.

Intracellular PRRs: The Second Line

NOD-like Receptors (NLRs): Cytoplasmic sensors. NOD1/NOD2 detect bacterial peptidoglycan fragments. Crucial for detecting intracellular bacteria.

NLRP3 Inflammasome: Assembles in response to diverse danger signals (ATP, uric acid, bacterial toxins). Activates caspase-1, which cleaves pro-IL-1β and pro-IL-18 into active forms. This is why IL-1β requires TWO signals: NF-κB activation to make pro-IL-1β, then inflammasome activation to release it.

RIG-I and MDA5: Detect viral RNA in cytoplasm. Trigger type I interferon response. Relevant in viral sepsis (influenza, COVID-19).

cGAS-STING: Detects cytoplasmic DNA (bacterial, viral, or self-DNA from damaged mitochondria). Potent inducer of type I interferons and inflammation.

The NF-κB Pathway

Most PRR signaling converges on NF-κB, the master transcription factor of inflammation. In resting cells, NF-κB is sequestered in cytoplasm by IκB. PRR activation triggers IκB kinase (IKK) → IκB phosphorylation and degradation → NF-κB translocates to nucleus → transcription of hundreds of inflammatory genes: cytokines (TNF-α, IL-1β, IL-6), chemokines, adhesion molecules, and more. This single pathway explains why diverse triggers produce similar inflammatory responses.

Special Case

Superantigens: Bypassing Normal Immune Regulation

Some pathogens produce toxins that cause massive, non-specific T-cell activation—superantigens. These cause particularly explosive septic shock.

The Superantigen Mechanism

Normal antigen presentation: APC presents specific peptide in MHC groove → only T-cells with matching TCR respond (~0.001% of T-cells).

Superantigens: Bridge MHC Class II and TCR Vβ region OUTSIDE the normal binding sites → activate any T-cell with that Vβ type → up to 20% of T-cells activated simultaneously.

Result: Massive cytokine release ("cytokine storm") far exceeding normal infection. Examples: Staphylococcal TSST-1 (toxic shock syndrome), Streptococcal pyrogenic exotoxins (streptococcal toxic shock).

Clinical Clue: Superantigen-mediated shock often occurs with relatively low bacterial burden. Blood cultures may be negative. Look for localizing signs (tampon use, wound infection, pharyngitis) and characteristic rash (diffuse erythroderma → desquamation).

Evidence Base

Takeuchi O, Akira S. Pattern Recognition Receptors and Inflammation. Cell 2010 • Chan JK, et al. Alarmins: awaiting a clinical response. J Clin Invest 2012 • Hotchkiss RS, Moldawer LL. Parallels in Immunology. Nat Med 2014

Section III

The Immune Response: From Protection to Destruction

The immune response to infection is meant to be fierce, focused, and self-limited. In sepsis, it becomes systemic, prolonged, and ultimately self-destructive. Understanding this transformation is key to understanding sepsis pathophysiology.

The Cytokine Cascade

Mediators of Inflammation: The Usual Suspects

1

TNF-α: The Initiator

Released within minutes of PRR activation, primarily by macrophages. TNF-α activates endothelium, increases vascular permeability, promotes coagulation, induces fever, and triggers production of other cytokines. In animal models, anti-TNF given BEFORE endotoxin is protective; given AFTER, it's ineffective or harmful. This timing explains why anti-TNF trials in human sepsis failed—patients present too late.

2

IL-1β: The Amplifier

Requires two signals: NF-κB activation (to make pro-IL-1β) + inflammasome activation (to cleave and release). Causes fever, hypotension, promotes neutrophil infiltration. Works synergistically with TNF-α—together they're far more potent than either alone. IL-1 receptor antagonist (anakinra) has shown promise in specific sepsis subgroups.

3

IL-6: The Messenger

Peak levels correlate with sepsis severity and mortality. Induces hepatic acute phase response (CRP, fibrinogen, ferritin, hepcidin). Promotes B-cell differentiation. Unlike TNF-α and IL-1β, IL-6 remains elevated for days, making it a useful biomarker. Tocilizumab (anti-IL-6R) showed benefit in COVID-19 cytokine storm.

4

Chemokines: The Recruiters

IL-8 (CXCL8) recruits neutrophils; MCP-1 (CCL2) recruits monocytes. Create concentration gradients that guide immune cells to infection site—beneficial locally, but systemic chemokine release causes diffuse immune cell infiltration into organs, contributing to ARDS, AKI, and hepatic dysfunction.

5

Anti-inflammatory Mediators

IL-10, TGF-β, IL-1RA, soluble TNF receptors attempt to counterbalance. In sepsis, this anti-inflammatory response can become excessive, leading to immunoparalysis. The balance between pro- and anti-inflammatory forces determines trajectory: resolution, chronic critical illness, or death.

The Cellular Response

Neutrophils: Heroes and Villains

What is the neutrophil's role in sepsis?

Neutrophils are first-line defenders—they phagocytose bacteria, release antimicrobial peptides, generate reactive oxygen species, and form NETs (neutrophil extracellular traps). In localized infection, these functions are essential for bacterial clearance.

Why do neutrophils cause harm in sepsis?

Systemic activation means neutrophils release their toxic payload in organs far from infection. Neutrophil elastase damages endothelium, degrades extracellular matrix, and inactivates surfactant. NETs trap platelets (contributing to microthrombosis), damage endothelial cells, and promote organ injury. The weapons designed to kill bacteria also kill host tissue.

What are NETs and why do they matter?

NETosis: neutrophils extrude their DNA decorated with histones, elastase, and myeloperoxidase—creating a web that traps bacteria. But NETs also trap platelets (promoting thrombosis), damage endothelium, and act as DAMPs (amplifying inflammation). NET components correlate with sepsis severity and mortality. DNase (which degrades NETs) is being explored therapeutically.

Why is neutrophil function impaired in late sepsis?

Neutrophils become "exhausted"—reduced phagocytosis, impaired oxidative burst, defective chemotaxis. This contributes to susceptibility to secondary infections. The same cells that were hyperactive early become hypoactive late. Sepsis-induced immunosuppression is as deadly as the initial hyperinflammation.

The Paradox

SIRS and CARS: The Two Faces of Sepsis

Sepsis is not simply hyperinflammation. It is a dynamic state where pro-inflammatory and anti-inflammatory forces compete, evolve, and can coexist in different tissue compartments.

SIRS (Pro-inflammatory)
  • Dominates early sepsis (hours to days)
  • TNF-α, IL-1β, IL-6 elevated
  • Fever, vasodilation, capillary leak
  • Neutrophil activation
  • Risk: Shock, early organ failure
  • Treatment: Source control, support
CARS (Anti-inflammatory)
  • Develops over days, may persist weeks
  • IL-10 elevated, HLA-DR decreased
  • Hypothermia, lymphopenia, anergy
  • Monocyte deactivation
  • Risk: Secondary infections, late death
  • Treatment: ? Immune stimulation

The Modern View: PICS

Many patients survive initial sepsis but develop Persistent Inflammation, Immunosuppression, and Catabolism Syndrome (PICS). They remain in ICU for weeks, are vulnerable to secondary infections, and experience profound muscle wasting. SIRS and CARS are not sequential—they coexist, creating "immunologic dissonance." This is why immunomodulatory therapies must be precisely targeted to the patient's current immune state.

Evidence of Immunosuppression in Sepsis

Lymphocyte Apoptosis: Autopsy studies show profound lymphocyte depletion in spleen and lymph nodes. Surviving lymphocytes show exhaustion phenotype (↑PD-1, ↓cytokine production).

Monocyte Deactivation: HLA-DR expression falls (<30% of normal)—monocytes cannot present antigen effectively. This predicts nosocomial infection and mortality.

Reactivation of Latent Viruses: CMV, HSV, EBV frequently reactivate in ICU sepsis patients—a marker of immunosuppression.

Cause of Late Deaths: Most sepsis deaths beyond day 3 involve secondary infections or failure to clear the initial infection—not hyperinflammation.

Evidence Base

Hotchkiss RS, et al. Immunosuppression in Sepsis. Nat Rev Immunol 2013 • Boomer JS, et al. Immunosuppression in patients who die of sepsis. JAMA 2011 • Mira JC, et al. Sepsis Pathophysiology, Chronic Critical Illness, and PICS. Crit Care Med 2017

Section IV

Endothelial Dysfunction: The Unifying Mechanism

The endothelium is arguably the most important organ in sepsis. Its dysfunction explains vasodilation, capillary leak, coagulopathy, and impaired tissue oxygenation. Every organ fails because its endothelium fails.

The Healthy Endothelium

What the Endothelium Does (And Why It Matters)

The endothelium is not a passive barrier—it is an active organ covering 4,000-7,000 m² that regulates vascular tone, permeability, coagulation, and immune cell trafficking. Its dysfunction is the common pathway of sepsis pathophysiology.

1

Vascular Tone Regulation

Healthy endothelium produces nitric oxide (NO) via endothelial NO synthase (eNOS), causing vasodilation. Also produces endothelin-1 (vasoconstrictor). The balance maintains appropriate perfusion pressure. In sepsis, inducible NOS (iNOS) is massively upregulated → excessive NO → pathological vasodilation that doesn't respond normally to vasopressors.

2

Barrier Function

Endothelial cells are connected by tight junctions and adherens junctions (VE-cadherin). The glycocalyx—a carbohydrate-rich layer covering the luminal surface—contributes to barrier function. In sepsis, inflammatory mediators disrupt junctions, glycocalyx is shed → plasma leaks into interstitium → edema, hypovolemia.

3

Anticoagulant Surface

Healthy endothelium actively prevents clotting: expresses thrombomodulin (activates protein C), heparan sulfate (potentiates antithrombin), tissue factor pathway inhibitor (TFPI), and produces prostacyclin and NO (inhibit platelets). Blood remains liquid because endothelium is actively anticoagulant.

4

Immune Cell Regulation

Resting endothelium expresses minimal adhesion molecules. Activation induces E-selectin, P-selectin, ICAM-1, VCAM-1—allowing leukocyte rolling, adhesion, and transmigration. This is essential for directing immune cells to infection; but systemic activation causes diffuse leukocyte infiltration into organs.

Endothelial Injury

How Sepsis Destroys the Endothelium

What activates the endothelium in sepsis?

Multiple simultaneous insults: direct pathogen effects (LPS binds endothelial TLR4), circulating cytokines (TNF-α, IL-1β), activated complement (C5a), hypoxia, and damage products from dying cells. The endothelium is attacked from every angle.

What happens to activated endothelium?

Phenotypic switch: anti-inflammatory → pro-inflammatory; anticoagulant → procoagulant; barrier → leaky. Endothelial cells express tissue factor (initiating coagulation), downregulate thrombomodulin, upregulate adhesion molecules, produce more NO (via iNOS), and lose glycocalyx.

What is the glycocalyx and why does its loss matter?

The glycocalyx is a gel-like layer (0.5-5 μm thick) of proteoglycans and glycoproteins on the endothelial surface. It acts as: a molecular sieve (preventing albumin leak), a mechanosensor (transducing shear stress to NO production), an anticoagulant surface (binding antithrombin), and an anti-adhesive layer (preventing leukocyte/platelet adhesion). Sepsis strips the glycocalyx within hours—syndecan-1 levels in blood reflect this shedding and correlate with mortality.

Why can't fluid resuscitation fix capillary leak?

Crystalloid passes freely through damaged endothelium—only ~25% stays intravascular when glycocalyx is gone. More fluid causes more edema, increasing diffusion distance for oxygen without restoring intravascular volume. This is why "too much fluid" is harmful—it doesn't stay where you put it.

Can the endothelium recover?

Yes, if the patient survives. Glycocalyx regeneration takes 5-7 days. Endothelial repair involves both proliferation of surviving endothelial cells and contribution from circulating endothelial progenitor cells. But ongoing insults (persistent inflammation, repeated fluid boluses, ongoing infection) impair recovery. Protecting the endothelium may be as important as targeting the infection.

The Coagulation Catastrophe

Sepsis-Induced Coagulopathy and DIC

Sepsis almost universally activates coagulation. In its most severe form, this becomes disseminated intravascular coagulation (DIC)—simultaneous microvascular thrombosis AND bleeding due to consumption.

The Three Hits of Sepsis Coagulopathy

Hit 1 - Tissue Factor Activation: Monocytes and endothelial cells express tissue factor in response to LPS and cytokines. TF initiates the extrinsic pathway, generating thrombin. Normally localized to injury sites, TF in sepsis is expressed systemically.

Hit 2 - Impaired Anticoagulation: Antithrombin is consumed and downregulated. Thrombomodulin is shed from endothelial surface → less protein C activation. TFPI is overwhelmed. The three main anticoagulant systems all fail simultaneously.

Hit 3 - Impaired Fibrinolysis: PAI-1 (plasminogen activator inhibitor-1) is massively upregulated, blocking tPA. Clots form but cannot be dissolved. Fibrin deposition occludes microvessels throughout the body.

The DIC Spectrum

Non-overt DIC: Subclinical coagulation activation. Mildly elevated D-dimer, slight platelet decline. May not be clinically apparent but indicates endothelial injury.

Overt DIC: Platelet count <100K (or >50% drop), elevated PT/INR, elevated D-dimer, low fibrinogen. Microvascular thrombosis causes organ dysfunction (AKI, hepatic ischemia, digital ischemia). Consumption causes bleeding.

Treatment Principle: Treat the underlying cause (source control). Support with platelets/FFP only if actively bleeding. Heparin is controversial—may help in thrombotic-predominant DIC but worsen bleeding. The coagulopathy resolves when sepsis resolves.

Microcirculation

The Hidden Pathology: Microcirculatory Dysfunction

Macrovascular parameters (blood pressure, cardiac output) may be "normal" while the microcirculation—where oxygen exchange actually occurs—is severely deranged. This explains why patients can die with "adequate" MAP and cardiac output.

The Microcirculatory Unit

Arterioles (100 μm) → capillaries (5-10 μm) → venules. Capillary density is ~1,000-3,000/mm³ depending on tissue. EVERY cell is within 20 μm of a capillary. When capillaries fail, cells die regardless of what's happening in larger vessels.

What happens to the microcirculation in sepsis?

Heterogeneity: Some capillaries are over-perfused (shunting), others are stopped or intermittently flowing. Overall capillary density (functional capillary density) decreases. Blood can flow from arteriole to venule without perfusing capillaries—"oxygen shunting."

Why does microcirculatory flow become heterogeneous?

Endothelial swelling narrows capillary lumen. Glycocalyx loss disturbs flow dynamics. Rigid activated leukocytes plug capillaries. Microthrombi obstruct flow. Local NO production varies, creating patches of vasodilation and vasoconstriction. The coordinated flow of healthy microcirculation is replaced by chaos.

Why doesn't increasing blood pressure fix this?

Vasopressors constrict arterioles, which may actually reduce microcirculatory flow even as MAP rises. "Hemodynamic coherence"—the coupling of macro and microcirculation—is lost. High MAP does not guarantee tissue perfusion. This is why lactate can remain elevated despite normalized blood pressure.

Can we monitor the microcirculation?

Sublingual videomicroscopy can visualize capillary flow directly. Research shows microcirculatory parameters predict outcome better than macrovascular parameters. But this technology is not yet standard. Clinically, we use surrogates: lactate (suggests inadequate tissue oxygenation), mottling (visible microcirculatory failure in skin), capillary refill time.

Evidence Base

Ince C. The Microcirculation is the Motor of Sepsis. Crit Care 2005 • Iba T, Levy JH. Sepsis-induced Coagulopathy and DIC. Anesthesiology 2020 • Johansson PI, et al. A high admission syndecan-1 level is associated with mortality. Crit Care 2011 • De Backer D, et al. Microcirculatory alterations in sepsis. ICM 2002

Section V

Cardiovascular Dysfunction: Septic Shock Physiology

Septic shock is defined by circulatory failure requiring vasopressors to maintain MAP ≥65 with lactate >2 despite adequate fluid resuscitation. Understanding the cardiovascular pathophysiology explains why septic shock behaves differently from other shock states.

The Hemodynamic Signature

Why Septic Shock is "Warm Shock"

Why do septic patients have warm extremities and bounding pulses initially?

Massive vasodilation (↓SVR) is the hallmark. iNOS produces excessive NO; prostacyclin levels rise; vasopressin stores deplete. The vasculature "relaxes." Cardiac output increases compensatorily, but can't overcome the profound ↓SVR. This is distributive shock—flow is high but pressure is low.

Why does vasodilation occur?

Inflammatory cytokines (TNF-α, IL-1β) induce iNOS in vascular smooth muscle and endothelium. NO activates guanylate cyclase → ↑cGMP → smooth muscle relaxation. Unlike eNOS (which produces small, localized NO amounts), iNOS produces massive, sustained NO. Additionally, smooth muscle K-ATP channels open, hyperpolarizing cells and preventing contraction.

Why don't vasopressors work well?

"Vasopressor resistance" has multiple causes: receptor downregulation (prolonged catecholamine exposure), downstream signaling disruption (NO blocks calcium influx), depleted vasopressin stores, relative adrenal insufficiency, metabolic acidosis (reduces catecholamine receptor sensitivity). This is why escalating vasopressor doses have diminishing returns.

Why is lactate elevated if cardiac output is high?

Lactate in sepsis reflects multiple problems: (1) Microcirculatory shunting (flow bypasses capillaries → tissue hypoxia despite high CO), (2) Mitochondrial dysfunction (cells can't use delivered O₂), (3) Accelerated glycolysis driven by catecholamines, (4) Impaired hepatic lactate clearance. High CO doesn't guarantee tissue oxygenation.

When does septic shock become "cold"?

Late/decompensated septic shock: myocardial depression reduces CO, exhausted compensatory mechanisms, severe hypovolemia from capillary leak. Now it's low CO AND low SVR—combined distributive and cardiogenic shock. Cold, mottled extremities, weak pulses. Worse prognosis than hyperdynamic state.

The Septic Heart

Septic Cardiomyopathy: The Hidden Dysfunction

Cardiac dysfunction occurs in 40-60% of septic patients but is often masked by the low afterload (vasodilation) and high catecholamine state. When the ventricle is challenged (fluid loading, increased afterload with vasopressors), dysfunction becomes apparent.

Mechanisms of Septic Cardiomyopathy

Circulating Myocardial Depressant Substances: TNF-α, IL-1β, IL-6, NO, and other mediators directly impair contractility. LPS itself causes myocyte dysfunction. Serum from septic patients depresses normal myocytes in vitro.

Mitochondrial Dysfunction: Cardiomyocyte mitochondria are damaged by sepsis (ROS, NO). Energy production fails, limiting contractile function.

Calcium Handling Abnormalities: Sepsis impairs calcium cycling—reduced SR calcium uptake and release impairs contraction-relaxation coupling.

Coronary Microcirculation: Sepsis affects coronary microcirculation like all vascular beds, potentially causing patchy myocardial ischemia even without epicardial coronary disease.

Reversibility: Unlike ischemic cardiomyopathy, septic cardiomyopathy is typically reversible in survivors (7-10 days). Initial biventricular dilation may actually be protective—allows maintenance of stroke volume despite reduced EF (preload recruitment).

Clinical Recognition

Echo Findings: Reduced EF (may be masked by low afterload—use load-independent measures), LV dilation, RV dysfunction (common and prognostically important), sometimes global hypokinesis with preserved LV size ("stress cardiomyopathy" pattern).

Biomarkers: Troponin elevation common (myocardial injury, not necessarily ACS), BNP elevation reflects wall stress.

Management: Avoid excessive fluids (worsens biventricular dilation). Dobutamine may help if evidence of inadequate CO despite fluids, but increases oxygen demand. Mechanical support (ECMO) in refractory cases.

Clinical Application

Why Fluid Resuscitation Has Limits

Why do we give fluids in sepsis?

Absolute hypovolemia (from capillary leak, fever, poor intake) and relative hypovolemia (vasodilation expands vascular capacitance). Fluids restore preload → increase stroke volume (Frank-Starling mechanism) → increase cardiac output → improve tissue perfusion.

Why do fluids stop working?

The Frank-Starling curve flattens—beyond a certain preload, more volume doesn't increase output. In septic cardiomyopathy, the curve is depressed AND shifted. Additionally, with damaged glycocalyx, crystalloid redistributes rapidly to interstitium. A 1L bolus may add only 200-300 mL to intravascular volume after 30 minutes.

Why is fluid overload harmful?

Interstitial edema increases diffusion distance for oxygen → tissue hypoxia. Pulmonary edema impairs gas exchange. Renal congestion (↑venous pressure) reduces GFR. Gut edema promotes bacterial translocation. Myocardial edema worsens diastolic function. Every organ is harmed by edema.

How do we know when to stop fluids?

Static markers (CVP) are unreliable. Dynamic assessments: pulse pressure variation, stroke volume variation, passive leg raise. Response to fluid challenge: if MAP/CO doesn't improve with 250-500 mL, patient is unlikely to be fluid responsive. The question isn't "does the patient need fluid?" but "will fluid improve perfusion without causing harm?"

The Modern Approach

Initial resuscitation: 30 mL/kg crystalloid within first 3 hours (Surviving Sepsis Campaign). After initial bolus: frequent reassessment. Stop fluids when patient is no longer fluid-responsive OR signs of fluid overload appear. Many patients need only 2-3 L total, not the 6-10 L given in past practice. De-resuscitation (active fluid removal) once shock resolves may improve outcomes.

Evidence Base

Hollenberg SM. Vasoactive drugs in circulatory shock. AJRCCM 2011 • Landesberg G, et al. Sepsis-induced Myocardial Dysfunction. Crit Care 2012 • Marik PE, et al. Fluid responsiveness: An evolution. Br J Anaesth 2020 • CLOVERS Trial. NEJM 2023

Section VI

Metabolic Dysfunction: The Cellular Energy Crisis

Sepsis fundamentally disrupts cellular metabolism. Even when oxygen delivery appears adequate, cells cannot produce ATP efficiently.

Mitochondrial Dysfunction

When the Powerhouse Fails

Cytopathic hypoxia—cells cannot utilize oxygen even when delivered. Tissue pO₂ may be normal, but ATP production is impaired because mitochondria are dysfunctional from NO inhibition of cytochrome c oxidase, peroxynitrite damage, and inflammatory mediators.

The Lactate Story

Lactate reflects multiple mechanisms: tissue hypoxia (Type A), aerobic glycolysis from activated immune cells and catecholamines, impaired hepatic clearance, and mitochondrial dysfunction. Target: normalization OR >10-20% decrease per 2 hours. Persistent elevation despite resuscitation suggests ongoing source or cytopathic hypoxia.

Adrenal Axis

Relative Adrenal Insufficiency

The stress of sepsis demands supraphysiologic cortisol. In some patients, the adrenal response is inadequate. Hydrocortisone 200 mg/day recommended for septic shock not responding to fluids and vasopressors—restores vascular reactivity and allows vasopressor weaning.

Evidence Base

Singer M. Mitochondrial dysfunction in sepsis. Virulence 2014 • NICE-SUGAR Investigators. NEJM 2009 • Annane D, et al. APROCCHSS Trial. NEJM 2018

Section VII

Organ Dysfunction: The Final Common Pathway

Sepsis causes dysfunction in virtually every organ system through shared mechanisms of microcirculatory failure, inflammation, and metabolic dysfunction.

Respiratory

ARDS: Lung as Target and Source

The lung receives entire cardiac output—maximum exposure to inflammatory mediators. Endothelial and epithelial injury floods alveoli with protein-rich fluid. Shunt causes refractory hypoxemia. Lung-protective ventilation (6 mL/kg, appropriate PEEP, prone positioning) limits ventilator-induced lung injury and systemic cytokine release.

Renal

Sepsis-Associated AKI

SA-AKI occurs in 40-50% of septic patients. Unlike ischemic ATN, septic AKI involves microcirculatory dysfunction and inflammatory injury rather than pure hypoperfusion—renal blood flow may be preserved. Usually reversible if patient survives.

Neurologic

Sepsis-Associated Encephalopathy

How does systemic infection affect the brain?

The brain was thought to be "immune privileged" behind the blood-brain barrier. But: (1) Circumventricular organs lack BBB, (2) Cytokines signal across BBB via receptors, (3) Vagal afferents transmit inflammatory signals, (4) BBB becomes permeable in sepsis, (5) Cerebral endothelium becomes activated. The brain is not isolated from systemic inflammation.

What damages neurons in SAE?

Microglial activation (brain's resident macrophages release neurotoxic mediators), cerebral microcirculatory dysfunction, impaired cerebral autoregulation, mitochondrial dysfunction in neurons, excitotoxicity (glutamate-mediated injury), and oxidative stress. Direct infection of CNS is NOT required.

Why is delirium so common in ICU sepsis?

SAE is compounded by: sedative/analgesic effects, sleep deprivation, metabolic derangements (hypoglycemia, uremia, hepatic encephalopathy), hypoxia, and the ICU environment itself. Distinguishing SAE from other causes of delirium is often impossible—treatment is supportive.

Long-Term Consequences

Sepsis survivors have increased rates of cognitive impairment, depression, PTSD, and functional disability. One-third have cognitive deficits equivalent to mild Alzheimer's or moderate TBI at 1 year. The acute brain injury of sepsis has lasting consequences. Early mobilization, minimizing sedation, and preventing delirium may mitigate long-term cognitive effects.

Hematologic

Thrombocytopenia and Anemia

Why Platelets Fall in Sepsis

Consumption: DIC consumes platelets in microthrombi. Hemophagocytic lymphohistiocytosis (HLH) may occur in severe sepsis—macrophages engulf platelets.

Sequestration: Spleen sequesters platelets. Activated endothelium traps platelets.

Decreased Production: Bone marrow suppression from sepsis, medications (especially linezolid, vancomycin).

Prognostic Value: Thrombocytopenia correlates with severity. Failure to recover platelets by day 4 predicts poor outcome. Platelet count is a component of SOFA score.

Anemia of Critical Illness

Mechanism: Inflammatory cytokines suppress erythropoietin, impair iron utilization (hepcidin blocks iron release), and shorten RBC survival. Phlebotomy for lab tests compounds loss.

Transfusion Threshold: Hemoglobin 7 g/dL for most patients (TRICC, TRISS trials). Higher thresholds don't improve outcomes and may cause harm. Exception: active bleeding, acute coronary syndrome.

Evidence Base

Bellani G, et al. Epidemiology, patterns of care, and mortality for ARDS. JAMA 2016 • Prowle JR. Sepsis-associated AKI. Clin J Am Soc Nephrol 2018 • Gofton TE, Young GB. Sepsis-associated encephalopathy. Nat Rev Neurol 2012 • Hébert PC, et al. TRICC Trial. NEJM 1999

Section VIII

Clinical Integration: Mechanism-Based Management

Understanding sepsis pathophysiology transforms management from protocol-following to physiologically-reasoned care. Every intervention should address a specific mechanism.

The Hour-1 Bundle

Why Early Intervention Matters

1

Measure Lactate

The Why: Lactate reflects tissue hypoperfusion, mitochondrial dysfunction, and stress metabolism. It identifies patients at risk before other signs appear. Elevated lactate (>2) with infection = sepsis by Sepsis-3 criteria even without hypotension. Guides resuscitation targets and indicates response to therapy.

2

Obtain Cultures Before Antibiotics

The Why: Source control requires knowing the pathogen. Cultures identify organism and sensitivities for de-escalation. Blood cultures positive in only 30-40% of sepsis, but when positive, dramatically alter management. Don't delay antibiotics for cultures, but draw them first.

3

Administer Broad-Spectrum Antibiotics

The Why: Each hour of delay increases mortality ~7%. The inflammatory cascade has positive feedback—early pathogen control limits amplification. Broad spectrum initially (cover likely organisms), then narrow based on cultures. Inadequate initial coverage is associated with 2-3× mortality.

4

Fluid Resuscitation (30 mL/kg Crystalloid)

The Why: Restore preload to optimize cardiac output. Treat absolute hypovolemia (from leak and losses) and relative hypovolemia (from vasodilation). 30 mL/kg is starting point, not endpoint—reassess and individualize. Some patients need less (cardiogenic component); some need more (ongoing losses).

5

Vasopressors for Refractory Hypotension

The Why: If MAP remains <65 despite fluids, vasopressors restore perfusion pressure. Norepinephrine first-line (α-1 vasoconstriction + some β-1 inotropy). Early vasopressors may be preferable to excessive fluids—restores pressure while limiting fluid overload. Target MAP 65-70 (higher targets not beneficial in general sepsis population).

Source Control

The Most Important Intervention

Antibiotics kill bacteria; source control removes the nidus of infection and stops ongoing bacterial/toxin release. Without source control, antibiotics alone cannot overcome sepsis.

Source Control Principles

Timing: Within 6-12 hours of identification when feasible. Delay increases mortality.

Examples: Drain abscess, remove infected catheter/device, debride necrotic tissue, relieve obstruction (biliary, urinary), resect perforated viscus.

Risk/Benefit: Sometimes aggressive surgery is too risky (hemodynamic instability). Temporizing measures (percutaneous drainage) may bridge to definitive surgery.

Common Miss: Infected lines, sinusitis (in intubated patients), Clostridioides difficile (can't drain—need vancomycin/fidaxomicin, possibly surgery for fulminant colitis).

Therapeutic Reasoning

Mechanism-Based Decision Making

Case: MAP 58 Despite 4L Crystalloid and Norepinephrine 15 μg/min

Mechanistic Questions:

1. Is there a source control problem? (Re-image, reassess cultures)

2. Is the patient fluid-responsive? (PLR, dynamic assessment—if no, stop fluids)

3. Is there cardiac dysfunction? (Echo—if depressed EF, consider dobutamine)

4. Is there relative adrenal insufficiency? (Start hydrocortisone)

5. Is there vasopressin depletion? (Add vasopressin 0.03-0.04 U/min)

6. Is there severe acidosis impairing catecholamine response? (Consider bicarbonate if pH <7.15)

Case: Lactate 4.5 Despite MAP 70 and ScvO₂ 75%

Traditional View: "Perfusion looks adequate—lactate must be from somewhere else"

Mechanistic View: Macrovascular parameters don't reflect microcirculation. Options:

• Microcirculatory dysfunction (may not respond to more fluids/pressors)

• Mitochondrial dysfunction (cytopathic hypoxia—needs time, not intervention)

• Accelerated glycolysis from catecholamines (consider reducing epinephrine if on it)

• Hepatic clearance impairment (check liver function)

• Ongoing source (re-evaluate source control)

Action: Ensure source controlled, avoid excessive intervention if macro parameters adequate, follow lactate trend (20% decrease/2h target), consider mottling/cap refill as microcirculatory surrogates.

Case: Day 7, Patient Improving but Now New Fever and WBC Rise

Mechanistic Question: Is this new infection, or persistent original infection, or something else?

• By day 7, immunosuppression (CARS) may be established → vulnerable to secondary infections

• HLA-DR on monocytes, lymphocyte count may indicate immunoparalysis

• Common secondary infections: VAP, CLABSI, C. diff, fungal infections

• Drug fever, VTE, transfusion reaction are non-infectious considerations

Action: Full infectious workup, consider broadening coverage to include resistant organisms and fungi. This is different from Day 1 sepsis—different pathogens, different immune context.

The Future

Precision Sepsis Medicine

Why One-Size-Fits-All Therapies Fail

Sepsis is not one disease—it's a heterogeneous syndrome with multiple phenotypes. Trials targeting single pathways (anti-TNF, anti-endotoxin, activated protein C) showed no benefit in unselected populations. Emerging evidence suggests these therapies might work in specific subgroups:

Phenotype Classification: Machine learning identifies sepsis phenotypes (α, β, γ, δ in one study) with different mortality and treatment responses. The hyperinflammatory phenotype may respond to anti-inflammatory therapy; the immunosuppressed phenotype may respond to immune stimulation.

Biomarker-Guided Therapy: Procalcitonin to guide antibiotic duration. Ferritin for macrophage activation. HLA-DR for immunoparalysis. Future management may involve real-time immune phenotyping.

Targeted Therapies in Development: IL-1 blockade (anakinra) in hyperinflammatory subgroup. GM-CSF or IFN-γ for immunoparalysis. Anti-NET therapies. Mitochondrial-targeted antioxidants.

"Sepsis is not simply infection plus inflammation. It is a dynamic, evolving syndrome where the same patient may need anti-inflammatory therapy at hour 6 and immune stimulation at day 6. The clinician who understands mechanism can adapt to this evolution; the clinician who follows protocols cannot."

— The Principle of Sepsis Pathophysiology

Key References

Evans L, et al. Surviving Sepsis Campaign 2021 Guidelines. Crit Care Med 2021 • Seymour CW, et al. Derivation and validation of sepsis phenotypes. JAMA 2019 • Kumar A, et al. Duration of hypotension before initiation of antibiotic therapy. Crit Care Med 2006 • Rhodes A, et al. SSC 2016. Intensive Care Med 2017 • Prescott HC, Angus DC. Enhancing Recovery From Sepsis. JAMA 2018