Tuesday, May 29, 2018

The plague that is vancomycin troughs

How important are vancomycin levels? Not very.... at least as a marker of efficacy. True, higher levels probably are associated with increased nephrotoxicity (above 15 mcg/mL) - Antimicrob Agents Chemother. 2013 ;57:734-44
But the notion that troughs of 15-20 mcg/mL are the holy grail of therapeutic drug monitoring targets is simply not supported by data.
What we know: vancomycin AUC/MIC in the mid 300 to 400s (let's just say > 400) or so range is likely the best PK/PD parameter that predicts therapeutic success (pic 1).
No automatic alt text available.
 Much of these data come from the early 2000's when typical vancomycin MICs were less than 1 (Clin Infect Dis. 2007;15;44:1536-42 - pic 2). 
No automatic alt text available.
It turns out that when the MIC is 1 (generally most common these days), the probability of achieving an AUC/MIC of 400 is pretty much equally poor whether the trough is 10-15 or 15-20 (pic 3) but if it is 0.5 or less, the PTA (probability of target attainment is essentially 100% regardless of trough.
No automatic alt text available.


the problem with using troughs to predict efficacy is assuming that they're a good surrogate for AUC (which is kind of reason we use troughs since AUC is more difficult to measure). In reality they are not (pic 4).
No automatic alt text available.

How about some clinical data? Pic 5 (Clin Infect Dis. 2012 Mar 1;54:621-9.) is results from the 2012 vancomycin vs linezolid nosocomial PNA trial: ignore the superiority of linezolid for clinical success and note that the success rate for vancomycin based on day 3 troughs (0-7.9: 48%, 8-12.3: 46%, 12.4-17.4: 45.5%, > 17.4: 45.5%), shocking? Shouldn't be considering the vast majority of MICs was 1 mcg/mL.
No automatic alt text available.

What about post-hoc data from the ATTAIN trial with telavancin vs vancomycin? (pic 6), no difference in cure rate regardless of troughs but more nephrotoxicity with higher troughs.
Image may contain: text
Take home point? patient receiving vancomycin for "severe" MRSA PNA and improving, trough comes back at 13 mcg/mL, why increase the dose?

Friday, May 25, 2018

Changing Paradigms in Stress Ulcer Prophylaxis

Obviously any patient that requires mechanical ventilation for 48 hours or has "coagulopathy" should get stress ulcer prophylaxis (SUP) right? maybe not....
when the above risk factors where identified in 1994 by Deborah Cook (N Engl J Med. 1994;330(6):377-81.) and studies validating SUP as an effective means of preventing GI bleeding in critically ill patients were published (Crit Care Med.1993;21:1844-9. and Crit Care Med. 1993;21:19-30 - btw, cimetidine continuous infusion, really?) critical care was quite different than today. Particularly, the emphasis on early enteral nutrition did not really come to fruition until more recently.
In fact, even the 2010 meta-analysis by Paul Marik that supported SUP only found benefit in patients NOT receiving enteral nutrition 
No automatic alt text available.


Recently, several small RCTs have been published challenging the notion that SUP is necessary in modern practice.
The POP-UP trial (CCM 2016) showed no difference in "clinically relevant bleeding" in 214 patients randomized to PPI or placebo (100% received enteral nutrition)
No automatic alt text available.

A 2017 exploratory study in the Journal of Critical Care compared "enteral nutrition vs PPI" in 102 patients and again found no difference in bleeding 
Image may contain: text



Finally, another pilot study published last year in CCM, again had similar findings as above in 91 total patients (89% received enteral nutrition.

No automatic alt text available.
the exclusion criteria from these 3 papers are rather mixed but include things like recent GI bleed, dual anti-platelet use and in one case 100 mg prednisone equivalent.
So are these enough data to stop using pharmacological SUP, especially with the possible increased risk of pneumonia and C.diff (see 2018 ICM meta-analysis) or are these data sets just too small? The SUP-ICU trial should give us a better idea, they are expecting completion by August of this year. In the mean time,I know of at least 1 ICU that has already implemented a no SUP for patients being fed policy.




Monday, May 21, 2018

"What is the INR of FFP?"

I frequently encounter people touting that "The INR of FFP is like 1.6 or something". Indeed I learned something similar myself at some point. It turns out however that the mean INR of FFP is actually 1.1 (Transfusion 2005;45:1234-5.). So why does it seem like no matter how much FFP you tend to give, it is very difficult to get the INR much lower than 1.6 or so?
As can be seen from the image below from (Transfusion. 2012;52 Suppl 1:45S-55S.), The relationship between INR and circulating coagulation factors (I am talking about warfarin here) is non-linear. While an INR of 1.7 still considered to be within the zone of "normal hemostasis", it also corresponds to almost 70% of circulating factors being depleted. In contrast, beyond that point, the INR increases more dramatically with small reductions in factor concentrations

Image may contain: text



FFP replaces a relatively fixed percentage of coagulation factors (vertical arrows), the difference is, at higher INR values(below 30% coagulation factors), this corresponds to a large change in INR, however when the INR is lower than 1.7 or so, the same fixed increase in factor concentrations translates to a minor/negligible change in INR. This means that to get an INR back to "normal" an unrealistic amount of FFP would be needed.


Tuesday, May 15, 2018

"Balanced Crystalloids" and Hyperkalemia

With the recent publication of the SMART and SALT-ED trials, the chloride rich vs poor (balanced/buffered) crystalloid debate is back in the spotlight after a break post the orginal Yunos JAMA 2012 and SPLIT JAMA 2015. Something I commonly hear from internal medicine residents however is that LR and plasma-lyte should DEFINITELY be avoided in the setting of hyperkalemia, since they contain 4 and 5 meq/L of potassium respectively.
It turns out however that normal saline ( strong ion difference of 0, pH of ~ 5.5) increases potassium more than either of the chloride poor solutions do.
1) Serum potassium level should not be able to exceed 4meq/L (LR) or 5meq/L(PL) since that is the concentration of K in the solution. If a patient's K is greater than 5, PL will actually lower potassium

2) given how acidic NS is (above), the increased hydrogen ion concentration exchanges with intracellular potassium, actually worsening hyperkalemia despite not having any K in the solution

3) several trials in the peri-op setting have demonstrated increased potassium with NS compared to LR or PL


Image may contain: text

No automatic alt text available.

Jerry Altshuler, Pharm.D., BCPS, BCCCP
Critical Care Pharmacy Coordinator
Director, PGY-2 Critical Care Pharmacy Residency Program 
The Mount Sinai Hospital
1 Gustave L. Levy Place
New York, New York 10029
@hockeypharmd

Friday, May 11, 2018

Does Acidosis Decrease the Response to Catecholamine Vasopressors?

One question I commonly encounter on rounds is "does acidosis reduce the activity of catecholamines?" While I think that at this point it is relatively clear that acidosis likely reduces LV contractility due to reduced intracellular calcium entry as well as H+ competition for calcium binding sites on myocardial troponin, I think that the notion that catecholamines "don't work as well in acidosis" is less clear. The bulk of the data that comprises this idea can be seen in these three pictures and goes as follows:
 
1) A report from 1958 describing the effects of catecholamines on dogs over varying pH values suggested less response at a pH of 6.76 than 7.27 and 7.34 (they also note that they gave 6 human patients sodium lactate, increased their pH and decreased their pressor requirement - that's 1950's publication standards for you).
Calif Med. 1958 Jun; 88(6): 437–440.
No automatic alt text available.
2) A 1988 paper in chickens showed reduced beta receptor expression when the acidotic chicken's pH fell. Isoproteronol responsiveness was concordantly reduced.
Am J Physiol. 1988 Jan;254(1 Pt 2):H20-7.
Image may contain: text
3) The most recent paper from 2014, looking at contractile force in Human mammary arteries showing INCREASED catecholamine response as the pH decreased.
J Thorac Cardiovasc Surg. 2014 May;147(5):1698-705.
No automatic alt text available.
In real life this is concept is obviously difficult to pinpoint as usually the degree of critical illness tends to correlate with the degree of acidosis and hypotension and can easily be misidentified as non-causative correlate.
 
What is everyone's thoughts on this question?

Jerry Altshuler, Pharm.D., BCPS, BCCCP
Critical Care Pharmacy Coordinator
Director, PGY-2 Critical Care Pharmacy Residency Program 
The Mount Sinai Hospital
1 Gustave L. Levy Place
New York, New York 10029
@hockeypharmd




Featured Post

Giapreza - A Closer Look at the Pharmacology of Ang2

Angiotensin-II (Ang2) is now an FDA approved vasopressor. With this new addition to the available options, experts are combing over the ...