New Year’s Resolution: Let’s Dose Vancomycin Correctly in the ED

New Year’s Resolution: Let’s Dose Vancomycin Correctly in the ED


VancoBagVancomycin remains one of our workhorse antimicrobials for treating infections caused by methicillin-resistant S. aureus
(MRSA). As the incidence of MRSA infections continues to rise AND we are starting to see increasing minimum inhibitory concentrations (MIC) with vancomycin, it is paramount that we optimize its use, starting in the Emergency Department (ED).

The American Society of Health System Pharmacists (ASHP) teamed up with the Infectious Diseases Society of America (IDSA) in 2009 to publish a consensus statement on Vancomycin Therapeutic Monitoring. 1 To prevent the development of bacterial resistance, they recommend using higher doses (15-20 mg/kg) to achieve proper serum trough concentrations.

For patients who begin their care in the ED, it is our responsibility to calculate and order the correct vancomycin dose. Yet all too commonly, a default fixed dose of 1 gm is prescribed instead.

“This flat dosing regimen might result in a delay in achieving therapeutic concentrations and thereby impact eventual outcomes.” 2

Proper Vancomycin Dosing

  • 15-20 mg/kg every 8-12 hours in patients with normal renal function 1
  • In seriously ill patients (eg, sepsis, meningitis, infective endocarditis) with suspected MRSA infection, a loading dose of 25-30 mg/kg may be considered 3
  • Actual body weight should be used
  • IDSA recommends a max dose of 2 gm 3
  • In adults, we round to the nearest 250 mg increment
Example 1
126 kg, MRSA cellulitis → 126 kg * 20 mg/kg = 2,520 mg → Give 2 gm (max dose)
Example 2
70 kg, septic shock → 70 kg * 25 mg/kg = 1,750 mg

How are we doing?

Short answer: Not great!

Two ED studies were published in 2013 that addressed this very topic.

  • A retrospective cohort of 240 random vancomycin doses in the ED found the mean vancomycin dose was 14.6 mg/kg. On the surface, this doesn’t look too bad when comparing it to the recommended 15-20 mg/kg standard. However, only 20% of patients actually received the correct weight-based dose. When broken down by patient weight, the only patients that received the correct dose were those where a 1 gm dose was correct! 2
  • A second retrospective cohort study of 4,441 vancomycin doses in the ED found only 22% were correct based on weight using the 15-20 mg/kg standard. Over 70% of patients were underdosed! Not surprisingly, patients underdosed were more often found to have subtherapeutic trough levels as an inpatient. 4

Interestingly, the second study found longer hospital length of stay and higher mortality in patients who received more than 20 mg/kg of vancomycin in the ED. The etiology of this association is unclear, however this may have been due to the phenomenon of sicker patients already being more likely to die. It should be noted that vancomycin guidelines are intended for patients with known or highly suspected MRSA infections. While it is known that patients with methicillin-sensitive staph aureus (MSSA) infections actually do worse when receiving vancomycin compared to beta-lactams, it is challenging to differentiate early in a patient’s ED course with the laboratory testing currently available.

Further, what we do in the ED matters! Most doses of vancomycin in this study were continued unchanged by the inpatient team.

“The fact that most vancomycin dosing was continued unchanged after admission to the hospital, despite being administered outside of recommended range in the majority of patients, again highlights the fact that ED treatment is highly influential on subsequent inpatient care.” 4

How can we do better? [Section updated 11/3/14]

  1. In the age of Computerized Provider Order Entry (CPOE), we must take advantage of the clinical decision support these tools offer. Instead of having order sentences for 1 gm, replace it with weight-based order sentences. One group found a significant reduction in 1 gm doses ordered after employing this strategy. 5
  2. A new randomized trial compared ED patients receiving 15 mg/kg initial doses vs. 30 mg/kg. 6 There was a significantly greater proportion of patients reaching target trough levels of 15 mg/L at 12 hours among the patients who received a 30 mg/kg loading dose as compared with a traditional 15 mg/kg dose (34% vs 3%, P < 0.01). This study did not use a max dose of 2 gm. They included patients up to 120 kg who received 3.6 gm loading doses! There was no difference in incidence of nephrotoxicity between the groups.

Take Home Points

  • Forget that you ever learned 1 gm of vancomycin is the default dose.
  • When vancomycin is indicated in an ED patient, utilize a weight-based dosing strategy of at least 15-20 mg/kg actual body weight to a max of 2 gm (though reference 6 did not use a max dose).
  • What we do in the ED matters for long-term outcomes because many doses are continued unchanged by the inpatient team.
  • Utilize CPOE tools to guide providers to order vancomycin based on weight.

Original: January 1, 2014
Updated: November 3, 2014

Rybak M, Lomaestro B, Rotschafer J, et al. Therapeutic monitoring of vancomycin in adult patients: a consensus review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. Am J Health Syst Pharm. 2009;66(1):82-98. [PubMed]
Rosini J, Grovola M, Levine B, Jasani N. Prescribing habits of vancomycin in the Emergency Department: are we dosing appropriately? J Emerg Med. 2013;44(5):979-984. [PubMed]
Liu C, Bayer A, Cosgrove S, et al. Clinical practice guidelines by the infectious diseases society of america for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clin Infect Dis. 2011;52(3):e18-55. [PubMed]
Fuller B, Mohr N, Skrupky L, Mueller K, McCammon C. Emergency Department vancomycin use: dosing practices and associated outcomes. J Emerg Med. 2013;44(5):910-918. [PubMed]
Frankel K, Rosini J, Levine B, Papas M, Jasani N. Computerized provider order entry improves compliance of vancomycin dosing guidelines in the emergency department. Am J Emerg Med. 2013;31(12):1715-1716. [PubMed]
Rosini J, Laughner J, Levine B, Papas M, Reinhardt J, Jasani N. A randomized trial of loading vancomycin in the emergency department. Ann Pharmacother. 2015;49(1):6-13. [PubMed]

Bryan D. Hayes, PharmD, FAACT, FASHP

Bryan D. Hayes, PharmD, FAACT, FASHP

Chief Science Officer, ALiEM
Creator and Lead Editor, Capsules series, ALiEMU
Attending Pharmacist, EM and Toxicology, MGH
Assistant Professor of EM, Harvard Medical School
  • HH

    Great point: Vancomycin is frequently dosed incorrectly in the ED and in house.

    However, I would much rather see — at least in critically-ill (ie septic shock/severe sepsis) patients — vancomycin given without delay and be somewhat insufficiently dosed than see 20mg/kg delayed due to non-STAT pathways (even with CPOE), “calculation” time, “approval from pharmacy” time, etc.

    More antibiotic can be added later or more frequently, but early antibiotics (and source control) are of the upmost importance and — I would argue — more important than the first dose being 20mg/kg.


  • patrick

    Great article and food for thought…

    How much of the increased LOS and increased mortality is from a delay in abx tx because proper dosing was ordered from pharmacy and had to wait an hr or longer?

    • Bryan D. Hayes

      This study (#4 on the reference list above) was a retrospective one, and none of those endpoints were measured. It is, at best, an interesting side note, not one that can lead to concrete conclusions. If the ED (or other hospital unit) stocks 1 gm bags of vancomycin, it should not be an issue to stock 1.5 gm and 2 gm bags in addition (or in place of). This would minimize any potential delay. Further, given vancomycin’s prolonged administration time and the likelihood of a gram negative sepsis causing clinical deterioration quicker, the gram negative antibiotic is generally administered first. This provides time to receive vancomycin from the pharmacy.

      • PB

        Go hard…and go to ICU. It is very important the source control and the minimal short time from admision to begining of antimicrobials. It is logical to do a big loading dose. My question is about the type of antibiotic that vancomycin is: “time dependent”, so, not in USA but in other countries vancomycin is used by continued infusion, not bolus, a rational approach thinking in pharmacokynetics of this antibiotic. What is your oppinion? big bolus dose anyway and follow with bolus or continuous? Best wihes. PB

        • Bryan D. Hayes

          This is a great question. There are several studies looking at continuous infusion vancomycin. I think the data is still insufficient to conclude if it is better and leads to less nephrotoxicity than intermittent dosing ( From an ED standpoint, it’s tough to justify taking up an IV line with vancomycin when it may be the only one available for the first part of a patient’s course. We don’t use continuous infusion antibiotics at our institution.

  • James McCormack

    Hi Bryan – just to follow up with my twitter post.

    The point I was trying to make is that almost all of our dosing recommendations for ABX are pulled out of our rear ends – given that, I don’t disagree with your recommendation.

    What I do know is higher doses increase cost and the potential for side effects. I worked in infectious disease for 15 years and have looked at dosing evidence for even longer. I don’t think there is anyway increasing an antibiotic dose by 20-30% can make an important difference in outcome or reduce the chance of resistance. I get the “logic” AND I would LOVE to be proven wrong.

    I just wish that ID as a group did a few dose studies and had a look to see if it makes any difference. Have you seen any study ever that showed better outcomes with higher doses. I remember the trend to do serum bactericidal titers – does anyone do that anymore?

    I’ve published two articles on AMG levels showing that there is little if any evidence for a therapeutic range and NOBODY has ever published anything that contradicts what I’ve said.

    I’m also just now working on my 15th lawsuit (as an expert witness) for patients who have developed ototoxicity from AMG – most of the time levels were measured but no other monitoring and the AMG was used way too long

    We used to say that you had to finish the ABX until it was all gone. There isn’t any evidence to support that and for most infections far shorter durations are now recommended

    The best dosing recommendation I was ever given was by one of the leading East Coast ID docs when I asked him how he figured out what dose of ceftazidime to use for an osteomyelitis of the jaw caused by Pseudomonas – with a southern accent he stated “1 gram feels good, but 2 grams feels real good”

    My approach to ABX dosing is if they are big and sick give lot’s and then stop ASAP. But that is also opinion and pulled out of my rear.

    Just wish we had better (or even any) evidence.


    • Bryan D. Hayes

      Thanks Dr. McCormack. I appreciate the follow up comment to our Twitter discussion. It seems that you agree with my recommendations (which are based completely on guidelines and primary literature), yet you feel the evidence behind those published recommendations are not sufficient. The example you provided from the ID physician actually confirms what the post says: bigger patients need more drug.

      I think our opinion differs in that I do believe that increasing the dose can have an effect on outcome. Aminoglycosides are a different story regarding risk for toxicity, and I do not discuss that in the post. There have been a few recent studies linking vancomycin troughs less than 15 to worse outcomes ( and In fact, the ED-based studies that I present in this post also reflect sub-therapeutic trough concentrations from lower doses of vancomycin. With increasing MICs, it is important to target the correct MIC/AUC ratio (although also a level 3 recommendation). The whole point is that if we believe in using vancomycin, and believe in targeting trough concentrations of 15-20 for most infections, then we need to make sure we attain those concentrations adequately… and that comes directly from dosing appropriately.

      • Emily Heil

        If only there were a way to accurately and quickly determine patient specific AUCs for vancomycin! I would invest in that software (USC has a nice program, would be amazing if they could drill it down to a quick and easy app We have definitely settled for troughs as the next best thing, but having an initial vancomycin trough of <15 has been linked with increasing vancomycin failure. That being said, depending on the bug and the patient, there are probably many patients that don't need those aggressive troughs to have success ( In an Emergency department setting where things such as the pathogen and susceptibilities are not known, I think you have to err on the aggressive side.

    • James McCormack

      Hi Bryan – very much agree that we likely agree on most of this. If you would ever like to chat about it I would love to do so over Skype.

      My reason for bringing up aminoglycosides is that for years we focussed on levels and paid little if any attention to ototoxicity. We would debate whether we should choose 80 or 100 mg of gentamicin when there is no possible way that could have an important change (or actually any change) in outcome.

      I agree that in a sick person we should give doses of antibiotics in the upper range and give multiple ABX but that is an emotional response not an evidence based one. AS an example numerous studies show that adding aminoglycosides to other ABX does nothing to outcome other than increase cost and toxicity. In my opinion these drugs should have been taken off the market years ago or at least severely restricted.

      As you may know I have an issue with the doses of medications we use, and that we way overdose most people (with ABX maybe being the exception)

      In addition, the whole concept of finishing an ABX course (at least for most infections) is a made up concept.

      Having spent 15 years doing infectious disease I think it is about time that the community needs to start doing studies that look at outcomes and NOT levels and MICs anymore. Those are surrogate markers that may have little to do with anything – unless one can show with well-designed studies that it does make a difference. Whatever happened to doing serum bactericidal titers? In theory they should have worked great.

      We need studies of vancomycin 1000 mg BID versus 15mg/kg not you and I debating the importance of lack thereof because neither of us knows because we have no reasonable evidence. These studies are not difficult or expensive to do and if I was a betting person I would bet it would make no difference at all in outcome. We also need studies of different doses of ABX, or different combinations etc.

      It is VERY unusual that doubling the dose of ANY drug changes ANY outcomes in ANYTHING -in fact I can’t think of any – one had to increase the dose of statins 8 x to show a difference in outcome, the same applies to PPIs, beta agonists etc. I’m not sure why ABX would be much if any different.

      Until then, we have a study like the ones you quoted (the secnd reference didn’t link properly) that has so many potential biases that it makes it nothing but hypothesis generating at best.

      Thanks for reading my rant 🙂

      • Bryan D. Hayes

        Great discussion points. I updated the link for the second reference. Thanks for catching that.

        My thought is that there is no universal practice that will work for all conditions. I agree with your premise in the article about starting much lower with many drug doses, but even that does not hold true in all circumstances and can have detrimental patient outcomes. For example, if we started a nitroglycerin infusion at 5 mcg/min for flash pulmonary edema (as is recommended by the package insert), the patient would fail therapy and most likely require intubation. Only by starting at higher doses (for this indication) will the patient benefit, in some cases saving a life. Given I am a practicing emergency medicine/toxicology pharmacist, but I can cite many examples of doubling doses of medications having life saving benefit. The examples you referenced are mainly for chronic conditions.

        • James McCormack

          Agree and thanks – if you have any studies showing that doubling of a dose leads to life saving benefit (over the lower dose) I would love to see them. I’m not aware of any but I haven’t looked in detail at the emerg literature and I would like to be complete in my evidence. Thanks.

          • Bryan D. Hayes

            It’s unfortunate, but you probably won’t find much of this type of data in a study. Emergency practice is not always laid out clearly in a randomized, controlled trial. I just want to make sure the point you made earlier (“It is VERY unusual that doubling the dose of ANY drug changes ANY outcomes in ANYTHING -in fact I can’t think of any”) is not taken as fact. I list below several examples where doubling the dose produces the desired, potentially life-saving effect.
            – nitroglycerin for flash pulmonary edema
            – adenosine for SVT
            – metoprolol, diltiazem, or verapamil for rate control in afib
            – vasopressors for shock
            – epinephrine for anaphylaxis
            – albuterol for asthma
            – naloxone for opioid overdose

          • James McCormack

            Great discussion – I think this would make an excellent point counterpoint discussion.

            So if your data is not in a study then it must be anecdotal – which, while it could be true, it could also very easily be a biased evaluation of cause and effect. It is likely that some of your examples may be legitimate, but we all know the problems in medicine caused by clinicians believing what they did was what made the difference.

            Here is what I think is factual – which is the point I was trying to make

            1) there are no RCTs (at least none that I have been able to find) showing that doubling the dose of a medication improves outcomes – there is a very good reason why – almost all dose response curves are NOT effect vs concentration but effect vs LOG concentration

            2) I am aware of a number of studies showing that doubling the dose of medications either does absolutely nothing or the difference is so small it is clinically unimportant – almost all the effect of a drug typically comes from the lowest available dose (especially new drugs)

            3) to get a doubling of the effect (if in fact that is even possible) you have to increase doses 4-8 fold at least

            That is why I don’t think it is possible that dosing 1000 mg BID versus 15mg/kg BID can make a clinically important difference – the increase in dose is not big enough.

            When it comes to ABX, which is what we were initially discussing I would challenge you to show me any studies (because that is what you need as anecdotes don’t work in this situation) showing that adjusting doses on the basis of levels, or MICs etc changes any outcomes. I would even be interested in any studies showing that 500mg vs 1000mg improves outcomes or 1g versus 2 grams etc.

            Just one study. If there isn’t one then we need to be VERY cautious as to how strongly we promote our dosing recommendations.

            In my opinion (based on the best available evidence) pharmacists have wasted careers doing TDM (at least the level part).

            As another similar example, recently the Am Phar Assoc promoted the value of pharmacists by showing how pharmacists are doing important work because their interventions led to lower blood pressure,cholesterol and glucose. When you look at the study patients had to be on 50% more medications and the reduction in the surrogate numbers that occurred in that study have been shown in large long term RCTs to either show no change in clinical outcomes or even WORSEN outcomes. That bothers me and should bother the health care system and patients

            Enough of my ranting 🙂

          • Bryan D. Hayes

            Some more good points. I don’t disagree that using surrogate markers for efficacy are suboptimal (eg, MIC/AUC ratios). I still think we agree that particularly with first dose antibiotics, giving the highest dose with the least toxicity is probably beneficial in sick patients even if it isn’t producing as much increase in the trough level as many expect.

            I think we also need to be clear on what you mean by “outcome.” Your statement (“It is VERY unusual that doubling the dose of ANY drug changes ANY outcomes in ANYTHING -in fact I can’t think of any”) did not specify. An outcome does not have to be decreased mortality. For example, there are plenty of studies in the anesthesiology literature showing that doubling the dose of a neuromuscular blocker produces almost double the level of paralysis. Here are just two studies, one with succinylcholine and the other with rocuronium: and In the emergency medicine, surgery, intensive care medicine, and anesthesiology worlds, that is a very meaningful outcome. The other examples I provided in my previous comment are similar, though I don’t have all of the data handy.

          • James McCormack

            I definitely agree with the high dose ABX at the start especially if they are very sick. I also agree if there is ONE area were there is likely a scenario that doubling the dose gives doubling the effect it would be anesthesia – I always use that example as the purest form of how unbelievable and useful and instantly effective some medications are.

            Personally I think MIC/AUC are not just suboptimal but misleading as they divert attention from doing clinically important studies. By outcome I absolutely didn’t mean just mortality – the comment applies to blood pressure, pain, asthma exacerbations, depression scores, cholesterol – pretty much anything you can think of.

            A great example is ezetimibe – 1/40th of the recommended starting dose of 10 mg provided 50% of the effect of lowering low-density lipoprotein levels achieved with 10 mg – or in other words you had to increase the dose 40X to get a doubling of the effect

          • I can only speak to cardiology, but there are a number of randomized controlled trials demonstrating that increases in dose confer additional improvements in clinical outcomes. Granted, the dose:outcome relationship in these comparisons is far from 1:1, but the differences are clinically meaningful nonetheless, and serve as the basis for target doses recommended in clinical practice guidelines.

            In heart failure, higher doses of ACE inhibitors reduce hospitalizations compared to lower doses (ATLAS trial:, as do angiotensin receptor blockers (HEAAL trial: Likewise, higher doses of beta blockers reduce mortality and ejection fraction compared to lower doses (MOCHA trial:

            In coronary artery disease, higher doses of statins confer improvements in major adverse cardiovascular events compared to lower doses (TNT trial:

            Most recently, dose-related differences in both efficacy and safety have been observed with the new oral anticoagulants. In RE-LY (, a higher dose of dabigatran conferred an improvement in stroke or systemic embolism vs. warfarin, whereas a lower dose was non-inferior to warfarin but safer.

            While I certainly agree that there is often a diminishing return as one increases the dose of a medication, it is not entirely accurate to say there are no randomized controlled trials showing that increased doses fail to confer meaningful improvements in clinical outcomes.

  • Emily Heil

    In a septic patient, the question of time to antibiotic administration is certainly a priority. Like Bryan mentioned above, administering the broad-spectrum beta lactam first will allow a little more time to get the drug from pharmacy, and adding other doses to ED stocks can speed up the process.

    Some would even argue that perhaps 2g should be the standard initial dose, and then dose adjustments can be made after that load. Loading doses can be divided up too (, so even if the patient only got 1 gram to start, you don’t need to wait 8-12 hours to get another dose.

    Having an initial vancomycin trough <15 has been shown to be an independent predictor of vancomycin failure (, so getting the initial dose right AND fast is the ideal!

  • Mary Shue

    Leave it to you, Bryan, to always stir up a good debate! I agree with your position. I also want to caution everyone who is dosing vanco without pharmacist input to keep an eye out for patients with, or acute kidney injury, hemodialysis patients, those with nephrotoxic agents on board, the frail elderly, etc. As you said, inpatient often continues the dosing starting in the ED, especially if there doesn’t happen to be a monitoring service. These patients may benefit from dosing on the more conservative end of the recommended range until vanco levels can be evaluated.

    • Bryan D. Hayes

      Mary, thanks for your comment. To be clear, this is the guideline position. I’m just ensuring everyone is aware of it. The guidelines do state (as does the blog post) that these recommendations are for patients with normal renal function. You are correct in that there are other factors that may affect subsequent dosing. Input from experts, whether a pharmacist or other, may be indicated in patients with challenging pharmacokinetic or pharmacodynamic issues.

      • Javier

        Great article.

        When do You use doses each eight hours?


        • Bryan D. Hayes

          It depends on renal function and clearance of the drug. If not achieving adequate troughs with every 12 hour dosing, switching to every 8 might work. We see it a lot in young, trauma patients and some Neuro ICU patients who clear the drug rapidly. Sometimes, every 6 hour dosing is even necessary.

  • Christian McClung

    Hi Bryan,
    Thanks for the thought provoking article. I am a community emergency physician and I can appreciate that the correct dosing is 15mg/kg and that if I am going to dose it to do it correctly. What I find confusing is what we are calling vancomycin failures. From my brief review of the literature it seems it is defined as an aggregate of 30d mortality, 60d MRSA recurrence, and ‘microbiologic failure’. None of these subgroups demonstrate statistical significance alone but in aggregate, only, with low Ns reported. The other problem I have is that the research leading to these considerations is often to justify drug development. For instance Dr Schentag et al receives support from Pfizer and he promotes Synercid to combine with Vancomycin. Of course Synercid is manufactured by King Pharma a division of Pfizer. He further promoters that endpoints in antibiotic studies should be shifted to Microbiologic outcomes instead of clinical in order to lower drug development costs and time to NDA approval. That I have a problem with because it takes us into the side of healthcare where financial analysts and financial epidemiologists promote and forecast drug sales. As just a community EP I am have a simple question, has anyone shown that patients dispositioned to the ICU and given vancomycin in the ER benefit from higher dosing in terms of survival to hospital discharge with good functional outcome? I frankly don’t care if I underdose the patient going to the floor. I do care if I am correctly dosing the ICU player. Furthermore does MIC level and bacterial burden coincide? As far as I know there is no current marker of bacterial burden except surrogates such has morphologic changes in neutrophils (Dohle bodies) and lactic acid levels. What we need is a marker that coincides with heavy bacterial burden and maybe this subpoulation is where the discussion of -static vs -cidal dosing would d have relevance.

    • Bryan D. Hayes

      Dr. McClung,

      You make a lot of excellent points. I can’t speak to any financial biases for recommending vancomycin, since it is off-patent. Synercid has so many deleterious side effects that we don’t even have it on formulary at my academic institution.

      There aren’t many studies that directly link vancomycin to improved outcomes as you suggest, other than clinical and biological cure rates. There have been a few recent studies linking vancomycin troughs less than 15 to worse outcomes, and In fact, the ED-based studies that I present in this post also reflect sub-therapeutic trough concentrations from lower doses of vancomycin. With increasing MICs, it is important to target the correct MIC/AUC ratio (although also a level 3 recommendation). The whole point is that if we believe in using vancomycin, and believe in targeting trough concentrations of 15-20 for most infections, then we need to make sure we attain those concentrations adequately… and that comes directly from dosing appropriately.

      I encourage you to read through the previous comments and responses. Many of your questions have been discussed and I think you might find them useful.