Updated Sepsis Guidelines

Although I posted this on my old blog, it’s worth repeating. Sepsis is too common!

Dr. Alo’s Sepsis Guidelines!

(yes we need them, people seem to forget the guidelines, so this is a brief, sometimes funny, look at sepsis)

I was going to write only about pressors, but I decided to discuss sepsis in general and go in depth into pressors. Pressors are a favorite topic of mine, and I feel we don’t know enough about them.

What is sepsis?
Sepsis is very common. It’s an overwhelming infection. It’s also a topic that isn’t discussed enough in the world of medicine. Sure, we all hear a few things here and there, but we need an in depth discussion of this topic. Too many times you see holes in the management and a lack of understanding of what is going on. Sometimes you even see mismanagement and you hope that that doesn’t happen. Early identification and treatment improves outcomes big time! Septic patients can’t maintain their blood pressure, their blood vessels get leaky for various reasons, and it’s hard to maintain a good pressure. They will usually have blood pressures like 70/30 or 65/28 or lower.

Bugs are always the culprits. Draw blood cultures, get urine and sputum cultures, and start broad spectrum antibiotics. In cardiology we say time is myocardium, in sepsis, time is tissue! Start whatever broad spectrum antibiotics you have available to you. If they are a nursing home resident (or recently hospitalized), add vancomycin! Don’t be thrifty or try and save money! Start them quick! The sooner the better!

Fluids, fluids, fluids!
The next best thing is…. FLUIDS! They key with sepsis is fluids. No one ever seems to remember fluids in the sepsis algorithm and sepsis protocols. Fluids should be administered first and foremost. The definition of septic shock (no longer being able to maintain a good blood pressure with sepsis) requires that fluid challenges were given and that the blood pressure no longer responds to fluids. So you keep giving fluids, until the blood pressure stops responding to fluids. This may be two or three liters of fluid, or maybe eight or nine liters. It depends on the patient. Of course, if you give them a small 500mL bolus and their pressure goes from 68 to 76…. then it is responding to fluids. Keep giving fluids. And give them some more! Some hospitals have their own algorithm that calls for 8-10 liters right off the bat!

Sometimes you see doctors give a 250mL bolus of fluid. That is the equivalent of spitting in some one’s mouth. That is not even a spit worth of fluid. Only 20-22% of 0.9 normal saline fluid actually stays in your veins, so a “bolus” of 250mL is a spit’s worth of fluid. Please bolus 500 or more. In sepsis, you have to use liters! If they are responding to fluids, keep giving them. Keep giving them fluids until they stop improving with fluids. If their heart rate is elevated and their pressure is low, they are begging you for fluid, please give them fluids. There seems to be this unfounded fear that giving too much fluid will “overload” a patients heart and cause back up into their lungs and there by compromising their circulation and lung’s ability to oxygenate. That is never the case in septic patients. Even if a heart had an ejection fraction of 25% (normal is 55), it’d take 3 liters of fluid in under an hour to overwhelm their heart. That’s hard to do. Most of the time, one liter fluid boluses are given over an hour, so you aren’t going to overwhelm any hearts. (except the sickest of hearts, and that’s not often)

Ok, so finally after giving the patient five or six liters of fluid (or more) and noticing that their blood pressure no longer responds to fluids, you can add an IV pressor agent. So what are pressors? Medications that increase your blood pressure and help perfuse your vital organs while your body and antibiotics work to fight off the infection that is causing this. Unfortunately, there isn’t a topic in medical school on pressors. Yes, you go over these drugs in pharmacology class, but that was years ago, and no one ever taught you how to realistically use them in real practice. Sure, you used them during residency, but you just did what your attending did and they never really taught you more than a few buzzwords about each drug. And most residents don’t read, so they really don’t know what they are doing other than, “Sure, just start dopamine.” In fact, when you talk to most attendings (graduated, fully licensed physicians), you’ll find that they are clueless as well, with the exception of some very good critical care physicians (and maybe a few others).

Pressing on!

It’s important to know these drugs inside out and learn as much as possible about them, so you can use them appropriately. When it comes down to it, if a patient needs a pressor, just pick one and go with it. There are no large scale studies that say that one is better than the other for initial blood pressure management. However, there are a lot of smaller studies that help identify which drug should go where.


This is the first agent that needs to be understood. It does everything and is very potent. It is used when patients die (code blue situations) and is quite useful. Epinephrine activates A1, A2, B1, B2 receptors equally and increases cardiac output, systemic vascular resistance and increases heart rate. This drug is rarely used as a pressor unless the others haven’t been working. Don’t be afraid to use this as a drip medication if nothing else seems to be working. This should not be a first line agent in sepsis (or most other conditions), but don’t be afraid to use it if all others are failing.


Slightly weaker than epinephrine, norepinephrine activates the A1 receptors and thereby mainly increase systemic vascular resistance (your arteries constrict). At much higher doses it can activate B1 receptors, but usually not. Norepinephrine does not increase heart rate or cardiac output, don’t let anyone fool you into not starting norepinephrine because the “heart rate is too high”. There is also evidence that norepinephrine improves perfusion pressure to vital organs, especially splanchnic, renal, and glomerular blood flow and has shown to improve GFR and urine output in septic patients. Also does not raise ICP for head trauma patients, whereas dopamine raises intracranial pressure.


This is the most targeted one towards A1 that there is. Phenylephrine activates A1 only and very potently. Causes a large increase in systemic vascular resistance and thereby raises blood pressure. The single advantage of phenylephrine is that it has been approved to be given through a peripheral IV and does not have to go through a central line at it’s lowest doses. (only if mixed in the correct concentration). But be careful if your dose starts to get higher and higher. You can use this agent as a temporary agent while waiting for a central line. The problem is that this drug also causes constriction of your carotid arteries, your coronary arteries, your splanchnic arteries, your renal arteries. So you end up cutting off circulation to your brain, heart, intestines, kidneys and that is not good. Many hospitals have removed PE from their campii and no one is allowed to use it. It should probably never be used except as a last resort. If someone has bad atherosclerosis (plaques in their arteries), make sure they are never on this drug. Bottomline: Don’t ever use!


Is the precursor to norepinephrine (which is the precursor to epinephrine). It works as a catecholamine in and of itself (like epi and norepi) and by causing release of catecholamines in nerve terminals and acts on it’s own dopamine receptors. It’s the one that does it all! And that’s usually why people use it first line if they aren’t sure why someone’s blood pressure is not improving. At low doses, dopamine primarily works on dopamine receptors in the renal, mesenteric, and coronary arteries augmenting blood flow to those arteries. But there is no evidence that augmenting this blood flow improves outcomes. At medium doses, you start to see some A1 and a lot of B1 activity. This is usually why you see an increase in heart rate and a lot of ectopy (extra, unusual beats). At higher doses, it basically becomes norepinephrine. It’s main effects are by stimulating the A1 receptors. But it still stimulates B1 and the dopamine receptors, so you get an increase in heart rate as well as systemic resistance. Dopamine is usually started because it treats almost everything. No matter what is causing the hypotension, start dopamine and work your way up till it works. Of course, if your heart rate starts to skyrocket or you see a lot of unusual beats (PVCs, etc), you will probably have to change to norepinephrine. Most crash carts have dopamine in them and most ER docs start patients on dopamine if they are unsure. Of course, septic patients (that are really septic) will require much more A1 stimulation than what dopamine can provide, so you will need to switch them eventually if dopamine is not getting the job done.

Dopamine can not be run through a peripheral IV. Don’t let anyone fool you! I have heard this misnomer many times. Dopamine is toxic to tissues if it leaks into tissues or the IV site infiltrates. The only one that can be used in a peripheral IV is phenylephrine and ONLY if it is mixed correctly. Otherwise, stop being lazy and start a central line!


Dobutamine is the best when it comes to getting the heart to squeeze harder. That’s dobutamine’s main job. It is the most purely B1 agent we have. However, it can also increase heart rate. This makes it ideal to use in the heart failure setting. If a patient is hypotensive because their heart isn’t pumping strong enough, dobutamine should be your agent of choice. Pure B1! Dobutamine does have some A1 effect, but this is minimal and unpredicatble. Sometimes it raises your systemic resistance, sometimes lowers it, sometimes has no effect at all. Dobutamine does not increase heart rate as much as dopamine does. Dopamine is the king of the hill when it comes to increasing heart rate. So if it is heart failure with weak contractility, use dobutamine. If it is heart failure due to mainly a slow rate (rare), think dopamine. If the patients has severe hypotension, dobutamine’s effects on blood pressure intially are very unpredictable. If the patient has severe hypotension, use dobutamine with another pressor agent like norepinephrine so they can work together. If a patient has ischemic heart disease or some other reason why their heart may not be receiving as much oxygen as it should, speeding up the heart rate will only make this worse. Try not to use agents that increase heart rate if they have ischemic heart disease.


Very few people understand this drug. It has many actions on the body. It is a hormone called anti-diuretic hormone. It acts via the V1 receptors. No one is quite sure how or why vasopressin improves blood pressure but it works. Septic patients usually lose sensitivity to their endogenous vasopressin, and their posteror pituitary can’t make it possibly due to hypoperfusion, so they may need exogenous vasopressin. It seems like it only works in patients that have a deficiency in vasopressin (duh!). In normal people, vasopressin does nothing to blood pressure. However, in septic patients, it helps improve their blood pressure and allows us to taper doses of the other agent being used (normally norepinephrine). You normally don’t want patients on the highest dose of norepinephrine or phenylephrine for too long as it causes vasoconstriction of their veins and arteries and can lead to fingers, ears, penises, and noses necrosing and falling off. Adding vasopressin allows us to lower the dose of the other agent. Vasopressin has never been shown to improve mortality and morbidity in patients. However, it has been shown to reduce the number of days on a pressor and the length of stay in the ICU. That’s good. Vasopressin doesn’t work by itself. It has to be administered with norepinephrine to actually do anything. It also seems that only some patients benefit from vasopressin, probably the ones that aren’t making it anymore due to their sepsis. Those are probably the same ones that require stress dose steriods (hrydrocortisone IV).

One problem with vasopressin is that it can constrict blood flow to the splanchnic (mesenteric) circulation as well as the myocardium. If a patient has problems with ischemia and or poor circulation to begin with, they may not be a good candidate for vasopressin. If you think a patient has dead gut (ischemic bowel), find another drug to use, vasopressin will exacerbate it. One of the ways to treat a bad GI bleed is with vasopressin, it cuts off mesenteric and splanchnic circulation. Of course, the GI bleed dose is much higher, but you still have to remember this when a patient has ischemic gut issues. The pressor dose of vasopressin is much lower than the GI bleed dose, but don’t let that fool you. If you start a patient on vasopressin and their bicarb on their next bmp starts to fall and they have a metabolic acidosis… go feel their belly and stop the vasopressin. Check an arterial lactic acid level to be sure, but the vasopressin isn’t helping!

Stress dose steroids:

The new guidelines regarding sepsis clearly state that if a patient is sick enough to be on a pressor (assuming you fully fluid resuscitated them), they should be on stress dose steriods. Don’t screw this up! Be aggressive about it! Put them on the hydrocortisone! Normally, when you are under stress your adrenal glands release hormones which lead to the rise in cortisol in your blood stream. Cortisol is a stress hormone that helps us better deal with stress. Sepsis is a hell of a lot of stress! A really sick septic patient can not produce enough cortisol on their own and will need extra cortisol. So if a patient is sick enough to be on iv pressors, the new guidelines say they should be on stress dose steriods. Hydrocortisone 100mg IV Q6. Do not forget this! No one was ever harmed by a little extra hydrocortisone!

What about the ACTH stimulation test?

The new guidelines clearly stand against ACTH stimulation testing and say that it is not necessary. It’s an outdated test that we no longer need. Stop doing it! Just start the darn steriods! Back in the old days we could not measure ACTH. So we would measure cortisol, use cosyntropin IV to stimulate the adrenal gland to secrete more cortisol, and measure corstisol again to see how much more was released. They had to do this because there wasn’t a way to measure ACTH. This could help determine if a patient’s adrenal glands were working or not. Now we have a highly accurate ACTH assay. We can measure ACTH directly. There is no need to fiddle around with an unreliable, difficult to perform, stimulation test. If you think there is a problem draw a random cortisol and ACTH level together. You can use these two values like we use TSH and Free T4. There used to be a thyroxine (Free T4) stimualtion test 20 years ago too, but now we can actually measure TSH… so we don’t need that either.

So how do you interpret the cortisol and ACTH levels? If the ACTH is high and the cortisol is low, the problem is with the adrenal glands. If the ACTH is low and the cortisol is high, they are responding appropriately to the stress and don’t need anything.

What if the cortisol level is less than 25? According to the newly revised critical care guidelines and the protocols at Rush Medical Center and Dr. Arcot Dwarakanathan (the one who trained all the Endocrine department heads at Rush, Cook County, Loyola)… if the cortisol level is less than 25, they need to be on hydrocortisone. No question about it. They are not responding to stress appropriately.

Practically speaking, if you have a sick, septic patient…. draw the cortisol and acth now. Once the blood is drawn, start them on hydrocortisone 100mg IV Q6 right away until you get results back. This should take no more than 5 minutes. The nurse draws the blood, then gives them the steriods. If their cortisol comes back 30 or 60, you know that they are responding appropriately and can stop the hydrocortisone. If it is less than 25, keep it going and taper it as the patient improves. If the patient has been on the steroid for less than 2 weeks, you can stop it abruptly, but you probably don’t want to do that because it can affect their blood pressure. Stress dose steriods help improve blood pressure in septic patients as well as patients whose adrenal glands don’t work for other reasons, (usually being on steroids for a long time (central suppression).

Central lines:

If a patient requires iv pressors, please put in a central line. If the peripheral line blows, it’d suck to have norepinephrine or phenylyephrine leaking into your skin and fat. Your skin would die off. If they are on a pressor they need a line. Be aggressive about this and don’t get lazy! If you are going to do a line, put it in the subclavian. The cleanest and easiest location. You don’t need ultrasound, you don’t need anything. They last the longest and are the least infected location. Don’t worry about causing a pneumothorax (collapsed lung). If you do it correctly, you will never have that complication. The key is to actually do it correctly. The problem is that most physicians probably were never taught how to do it correctly. I see people screw this up all the time. Put the patient in head down, legs up position (Trendelenburg). This fills up the subclavian and neck veins so you can access them more easily. Turn down the tidal volume on the ventilator. If the tidal volume is over 400, turn it down to 400 or less. High Vt doesn’t allow the subclavians to fill up properly. This happens more often than you think.

I remember an ER attending and senior resident trying to get a subclavian on a little old lady with a tidal volume of 600. They tried about 30 times and couldn’t get it. They didn’t lay her with her head lower than her legs either. They tried both groins and couldn’t get the line either. I told my intern that the ER attending and the senior resident couldn’t get a line in this lady, and he was going to get it in 2 seconds. I put her down (head below legs), turned the Vt down to 350 (what it should be), and on the first stick (and my careful guidance) he got the line in. We went down to the ER to let the resident know what we did differently (it is a teaching institution) and he was very happy. Use full sterile precuations! Everyone working on the line wears a mask, hairnet, sterile gloves and sterile gown. Any helpers (nurses, students, etc) should wear a facemask and hairnet. Prep the area with chlorhexadine prep. Rub the area well for at least 2 minutes. That’s how chlorhexidine works (not by drying, that’s betadine). Chlorhexadine has improved infections and outcomes. After prepping the area well, gown and glove up, and lay the drape correctly and do not contaminate what you just cleaned. People always screw that up! Clean a wide area so you can’t screw up the drape too easily. I always prep the IJ as well as the subclavian, in case we can’t get one, we do the other without restarting.

Arterial lines:

If a patient is on a pressor, they need invasive monitoring. An arterial line sits in your artery and gives us a live blood pressure with every beat of the heart. This way we know what the pressure is all the time and can use it to titrate the pressor up or down. Be aggressive about these and make sure all patients have them!

CVP and volume monitoring:

The question is always asked, so how do you know when you have given enough fluids? If a ptient has a central line, you can hook up the distal port to a gadget that takes pressure readings and gives you an idea of how well the patient is hydrated. Unfortunately, it depends on where the thing is zeroed, where its located in elevation and relation to the right atrium, and the operator. If you have it connected correctly, you want a CVP of 8-12, this means that we have adequately hydrated the patient. There are no perfect ways to monitor fluid status in patients, including CVP. And it’s hard to know if they are overloaded or underloaded. You can listen to their lungs, but that is only 30% sensitive and varies depending on your experience. Another way to monitor fluid status is by monitoring urine output. You want a patient to pee 35-105mL of urine per hour. If they aren’t peeing this much, give them more fluid. Of course, this doesn’t work in patients with renal failure that don’t make urine at all.


If they aren’t breathing well or have altered mental status, intubate them and let the machine breath for them. You want a venous O2 saturation of at least 70%. Do not check ABGs and decide to intubate them based on the ABG. It is not a laboratory decision. It’s a clinical decision. If you aren’t sure, just intubate them! Be aggressive about this! Don;t waste time and “see if they’ll come around”. Just do it! And unless they are 7 foot 2 inches tall, they don;t need a tidal volume over 600. Most people require 300-480 at the most. It’s 6-8mL per kilogram of IDEAL body weight. Not their actual weight!

Lactic Acidosis

Tissue hypooxygenation is the fastest way to produce lactic acid. Fix the oxygenation, you will fix the acidosis. There is no other way around it. Giving someone bicarbonate either as a drip or in boluses does not treat lactic acidosis. Yes, people do it all the time. It’s wrong! They have studied this time and time again, bicarb does not treat lactic acidosis. It only worsens outcomes. It makes the blood pH “look” better while making the interior of the cells way more acidotic. Stop doing it! They took it out of ACLS protocols for a reason! Bicarb is only indicated in a few condistions; tricyclic antidepressent overdose, rhabdomyolysis, type 4 RTA, and that’s about it. Stop using it! The worst use of bicarb is in DKA patients. Yes, patients in DKA will have an extreme acidosis, even as low as 6.7 or 6.8. The treatment is tons, and tons, and tons of fluid. Fluid will fix this situation. Don’t be an idiot and screw up the body’s compensatory mechanism by giving bicarb. The acidosis is how the body copes with this extreme dehydration. Give them 4 liters of normal saline and watch them pH magically normalize.


Wow, I was going to write about pressors only, and it turned into a long post. Sheesh! Download my pressor chart HERE!