10.3 Severe Sepsis, Septic Shock, and Infective Endocarditis

Sepsis kills brain-injured patients by stealing MAP, raising temperature, and seeding the CSF or heart valves. This independent OpenExamPrep section covers Sepsis-3 language, Surviving Sepsis Campaign (SSC) Hour-1 actions from the 2021 adult guidelines, why norepinephrine is first-line, how meningitis and ventriculitis function as source-control problems, and the neurologic complications of infective endocarditis (IE)—embolic stroke, mycotic aneurysm, anticoagulation decisions, and conceptual operative indications. SSC is a named source, not a claim that OpenExamPrep rewrites that campaign.

Sepsis and septic shock, stripped for the exam

Sepsis is life-threatening organ dysfunction from a dysregulated response to infection, operationalized as an acute SOFA rise of 2 or more. Septic shock is a subset with underlying circulatory and cellular abnormalities: vasopressors are required to keep MAP ≥65 mmHg and lactate is >2 mmol/L despite adequate volume. You do not need a perfect SOFA worksheet to start treatment. Fever plus hypotension plus a likely source is enough to move.

Neuro ICU modifiers: fever raises cerebral metabolic demand. Hypotension drops CPP. Hypercapnia from respiratory failure raises ICP. A “soft” MAP of 60 mmHg that a medical ICU might tolerate can infarct a recently recanalized hemisphere. You still use the SSC MAP floor of 65 mmHg as the septic-shock target unless a monitored ICP or an autoregulation discussion has set a different CPP goal—and even then you do not leave the patient at MAP 50 mmHg because “the brain likes it low.”

Hour-1 concepts (SSC 2021)

SSC groups initial resuscitation as a Hour-1 bundle. Not every checkbox will finish at minute 60, but you start them on recognition:

1. Measure lactate; repeat if the first value is >2 mmol/L.
2. Obtain blood cultures before antibiotics if that causes no substantial delay (SSC language: under about 45 minutes).
3. Give broad-spectrum antimicrobials immediately, ideally within 1 hour when septic shock or a high likelihood of sepsis is present.
4. Begin 30 mL/kg crystalloid for hypotension or lactate ≥4 mmol/L.
5. Start vasopressors during or after fluids to keep MAP ≥65 mmHg.

The culture rule is the one people fail. If the patient is crashing, do not wander looking for a second set of cultures for 20 minutes. Draw what you can on the way to the drug. If the patient is stable enough that two blood-culture sets and an EVD CSF sample will take 10 minutes, take them. CSF, urine, and sputum are source cultures, not a reason to park antibiotics at the nursing station.

Volume is the other trap. 30 mL/kg is the SSC starting crystalloid dose for hypotension or a lactate ≥4 mmol/L, using balanced crystalloid preferentially over 0.9% saline in the 2021 document. In a patient with malignant edema, hydrocephalus, or an open belly, you still give what perfusion needs, but you reassess after the first liters with lactate, capillary refill, ultrasound, and ICP. You do not infuse 3 L blindly into a tight brain because a bundle poster said so. Balanced fluid does not cancel the need for source control.

Norepinephrine and the rest of the pressor stack

SSC recommends norepinephrine as first-line over dopamine, epinephrine, or vasopressin as monotherapy. Dopamine is more arrhythmogenic. If norepinephrine is unavailable, epinephrine or dopamine are stopgaps. When MAP remains low on norepinephrine, add vasopressin (often considered around norepinephrine 0.25–0.5 µg/kg/min) rather than doubling norepinephrine indefinitely. Epinephrine is a later add. Angiotensin II appears in refractory vasodilatory shock in selected centers; it is not the opening move.

Do not delay norepinephrine for a triple-lumen insertion if a proximal peripheral or intraosseous line can start the infusion. SSC allows peripheral norepinephrine for a limited time through an appropriate vein. In a patient with ICP 30 mmHg, every minute at MAP 55 mmHg is a CPP emergency.

Steroids: for septic shock with ongoing vasopressor need, hydrocortisone is a SSC suggestion, not a substitute for source control. It is a different decision from dexamethasone for pneumococcal meningitis.

Neurologic sources: meningitis and ventriculitis

Community meningitis, postoperative meningitis, and external ventricular drain (EVD) or shunt ventriculitis are sepsis sources sitting in your unit. Hour-1 still applies, but source control means CSF sampling, imaging for abscess or empyema, and device removal or exchange when the hardware is infected. Broad abdominal coverage with piperacillin-tazobactam does not treat meningitis (see the next section on blood–brain barrier penetration). If the patient has an EVD and new fever, cloudy CSF, or unexplained lactate, culture the CSF and start meningitis-dose agents the same hour you start norepinephrine.

Other neuro sources: sinusitis in intubated patients, hospital pneumonia, urinary catheters, surgical wounds, and infected scalp incisions. A “culture-negative” septic shock after clipping may still be a missed abdominal catastrophe or line infection; look below the neck.

Infective endocarditis: embolic stroke and mycotic aneurysm

Left-sided vegetations throw to brain, spleen, and kidney. Staphylococcus aureus, streptococci, and enterococci dominate native-valve disease; staphylococci dominate prosthetic valves and injection-drug right-sided disease (which more often embolizes to lung). Neurologic presentations include ischemic stroke, hemorrhagic transformation, meningitis, and mycotic (infectious) aneurysms from septic embolization into vasa vasorum or vessel wall.

Any IE patient with headache, a new deficit, meningismus, or unexplained coma needs brain imaging. MRI finds silent emboli often; CT is faster when you need to exclude hemorrhage before anticoagulation or surgery. Mycotic aneurysms may be small and distal; CTA, MRA, or catheter angiography is required when the story fits (delayed hemorrhage, SAH, or IE plus headache). Ruptured or enlarging mycotic aneurysms need endovascular or surgical treatment, often before valve surgery and before restarting anticoagulation, when the heart can wait.

When not to anticoagulate

IE is not itself an indication for anticoagulation. Vegetations are infected fibrin; blood thinners do not reliably prevent emboli and they do increase intracranial hemorrhage. American Heart Association adult IE guidance: for mechanical-valve IE with embolic stroke, it is reasonable to hold anticoagulation for at least 2 weeks of antibiotics to reduce hemorrhagic conversion. Intracranial hemorrhage: stop anticoagulation. Native-valve IE with ischemic stroke should not get new warfarin “for the vegetation.” If a mechanical valve absolutely requires coverage, teams often use short-acting unfractionated heparin after imaging shows a stable, small infarct—and they accept that this is a multidisciplinary, high-risk call, not an automatic restart.

Antiplatelet drugs are not a stroke-prevention strategy for IE either, unless there is a separate coronary or stent indication.

Lumbar puncture before anticoagulation if hemorrhage or meningitis is in play

If the deficit could be meningitis, SAH, or a hemorrhagic infarct, do not start heparin on the way to CT. Image first. If imaging is negative for mass effect and you still need CSF (meningitis overlap, or mycotic aneurysm not yet excluded as a sentinel leak), a lumbar puncture (LP) before therapeutic anticoagulation is the safer sequence: an LP into an occult hematoma or a fully anticoagulated patient with an undiagnosed mycotic leak is a disaster. After a traumatic LP, delay therapeutic anticoagulation until you know the fluid and the imaging. This is the same logic as “CT before LP” in possible bacterial meningitis with mass-effect risk, now applied to the IE patient you were about to heparinize for a mechanical valve.

Conceptual surgical indications

You are not being asked to quote millimeter vegetation cutoffs from memory. Know the three buckets that send IE to the operating room:

Timing versus the brain: silent infarcts and TIAs usually should not delay urgently indicated valve surgery. A large ischemic stroke often leads teams to wait on the order of days to ~4 weeks if the heart can wait, because cardiopulmonary bypass and heparinization can convert the infarct. ICH commonly delays surgery ~4 weeks when the valve is stable. If the patient is in refractory heart-failure shock from a flail leaflet, the heart wins and you operate with eyes open. That nuance is the exam, not a single universal day count.

Putting sepsis and IE together on one shift

A febrile patient with a new hemiparesis is not “just a stroke” until you have thought about IE and meningitis. Listen for a regurgitant murmur, look at nail beds and fundi, get blood cultures before antibiotics if you can do it immediately, and echo once the airway and CPP are stable. A transesophageal echocardiogram is more sensitive for vegetations and abscess than transthoracic imaging, but it does not precede Hour-1 drugs in shock.

Source control in IE is the surgeon and the duration of antibiotics, not a PICC and a prayer. Source control in ventriculitis is the drain. Source control in bowel perforation is the abdomen. SSC language without source control is incomplete care.

Exam traps

Waiting an hour for cultures in frank septic shock. Giving 30 mL/kg into an already overloaded, high-ICP brain without reassessment. Starting dopamine because “renal dose” folklore. Covering EVD infection with piperacillin-tazobactam. Anticoagulating native-valve IE to “stabilize the vegetation.” Restarting warfarin the night of an embolic stroke because the INR is 1.6. Skipping vascular imaging in IE plus delayed intracranial hemorrhage. Delaying a clearly indicated valve operation for a tiny silent MRI lesion when the patient is drowning in pulmonary edema.