Well, thank you Dr. Lewis. I'm Carrie Hamilton, one of the medicine chiefs. And I'm here today to introduce our speaker for us. So it's an honor for us to have a Dr. Taison Bell, from the Division of Infectious Disease and Critical Care Medicine, here to speak with us today. 

Dr. Bell grew up actually just about an hour from Charlottesville. Went to college here at UVA before venturing north to further his education and training. He completed his residency and chief residency at Mass General. And his ID fellowship through Mass General and Brigham and Women's Hospital furthered his unique skillset through an additional fellowship in critical care medicine at the NIH. 

Through his early career, Dr. Bell has established himself as a leader and exemplary teacher at his institutions and nationally. And Dr. Bell's realm of expertise is broad, as he can speak eloquently on topics ranging from principles of effective didactics, approaches to diagnostic mysteries antimicrobial stewardship, and cost benefit analysis in ICU setting, just to name a few. 

Dr. Bell has quickly become one of the most beloved attendings at UVA with this range of teaching expertise along with his calm and engaging demeanor, and his humanistic approach to daily care. And for these efforts, he has earned several honors including most recently UVA internal medicine residency and patient attending of the year award last year. So with that background, we're very fortunate to have him here to speak with us today about the intricacies of antimicrobial utilization and a few new frontiers on antibiotic use. So without further ado, please join me in welcoming Dr. Bell. 

[APPLAUSE] 

Thank you for that. All right. It's good to be here everyone. How you doing? Kind of a dead crowd. 

[LAUGHTER] 

Come on. You got to do better than that. It's Friday. How we doing? 

Good. 

All right. Where's my mouse? Back here? No. It's right there. 

All right. So we are going to talk about the intricacies of antimicrobial localization, as Carrie said. Continuous and extended infusion beta-lactams. This is a topic that I certainly get excited about. Can't seem to get a lot of other people excited about. But hopefully, my task today is to make it a little more exciting. And first and foremost, give you a framework to think about when something like this might be reasonable to do. 

So first, conflicts of interest. I have none. But in the market, please let me know if you are aware of anything I might entangle myself with. So objectives here are threefold. So first, we want to learn the pharmakinetic pharmacodynamic characteristics of commonly prescribed antibiotics. 

Number two, we want to identify clinical situations where it might be reasonable to use continuous or extended infusion. And I'll note these as CI and EI beta-lactams where they might be useful. And then three, a goal whenever I came to a grand rounds, want to make sure you can make yourself look smart on rounds. 

So with that, let's jump in to a case. My toolbar up here is going to help us keep track of where we are in the talk. So it'll change color. And that will help you keep track of where we are. 

So first, we'll start with a case presentation. Patient PK-- now this is when I was a Fellow at the NIH, a critical care Fellow-- a 16-year-old boy who was referred to us from a hospital in Bangalore, India. That's a real outside hospital. 

He was transferred for consideration for transplant aplastic anemia. He was originally diagnosed a year ago in Kenya. His family, recognizing poor medical care in India, [INAUDIBLE] into Kenya for this condition, moved him to India. 

He'd been hospitalized for seven months. So he tried the standard ATG cyclosporine eltrombopag. It was refractory. He received zero transfusions. And he was neutropenic for pretty much this whole duration. 

When they talked to us over the phone, they said he had a few pneumonias but he was stable, per their report. So he left the outside hospital without having been prescribed antibiotics. So luckily for me, I was not on call this night, my friend was. 

But we made a plan that he would admit the patient, do a lot of the primary workup. I would come in in the morning early because at the NIH there were no residents. The Fellows did everything from top to bottom. 

So I was going to come in early, review the records, and we'd be ready to present them the next morning. So upon transfer, fever to 102. He has a new distributive shock, which hadn't been reported before. He gets resuscitated, he gets vasopressors. 

Unfortunately, the [INAUDIBLE] respiratory failure gets intubated. And then had a bedside echo that showed a new cardiomyopathy, which wasn't known to us. And then I get there in the morning and discover that things aren't really as they were reported to us. 

So he's actually being treated for three different infections, two of which are multidrug resistant. CRE Klebsiella, a VRE bacteremia, and pulmonary aspergillus. This is the list of antibiotics that he was on up until the day of transfer. They stopped them because they knew that if they had reported this to us, that we may not have accepted the patient. 

But he had been treated with this list of antimicrobials. You can see, there's a lot going on here. So literally, he is on more antimicrobial agents than the lines in the CBC. We're in bad shape here. 

Hospitals days one through four, we DC the meropenem. We started ceptaz AV and gentamicin. Everything else we continue until we can sort things out. His fevers went away. He essentially stabilized. 

He got a whole body CT that showed extensive pulmonary nodular opacities with some cavitary changes, a complex peri-rectal abscess, which becomes [INAUDIBLE] IR. They placed the drain. Looked like on repeat imaging it wasn't completely drained. 

Then he got a right heart cath that showed he was in distributive shock, with a cardiomyopathy. Eventually got him to low dose levophed and vasopressin. Then hospital day four, new fever to 102. And his blood cultures are positive for gram negative rods. 

And he's on all this stuff. He has positive blood cultures. By this point, my friend has rotated off, and I'm presenting him in the morning. And I got to come up with the plan. Got to think about something. 

Luckily for us, he has no new organ dysfunction at this point. His estimated GFR is above 120. And so the only thing I could really come up with was to do some extended infusion or continuous infusion and antibiotic. Literally, there was nothing else that we could add at that point. 

So I have a question for you guys. How much can extended or continuous infusion beta-lactams, what can they do here? A, go all in with me. This will fix everything, all of our problems. B, it's something to do but I think we've got bigger fish to fry here. C, Jerry, Chris, what in the world is going on here? Or D, I'm not sure. 

So text UVAGME, the 22333 to join and then give me your answer here. We should update live. Someone's got my back. All right. This is neat. So it's looking like the majority of folks think that it's something we could do but we've got bigger fish to fry here. And then there is some-- one lone supporter of me right there. 

So for a little bit of background, beta-lactams, we prescribe these all the time. They are the commonly prescribed antibiotics in the ICU in the critically ill. Properties here, they're usually bacteriacidal. And they have a low volume of distribution. They're hydrophilic and primarily renal excreted. 

So ideally, if we had good pharmacokinetics, we would have our concentration of drug, the free concentration over the minimum inhibitory concentration, or the MIC, what was called the PK/PD efficacy over MIC. And we would have this for a good portion of the interval of the antibiotic dose, ideally maintained for 90% to 100% of the dosing interval, although there's some debate here. Some people say we should be 100%, some people say between 40% and 60%. And there is some class discrepancy here. 

Cephalosporins, you're looking at 60% to 70% in a bowl for peak bacteriacidal effect. Penicillin's 50% and carbapenem's 40%. This is exactly the kind of slide that has good information but it's not going to be retained. So let's look at this a different way. 

This is a graph of drug going in and drug going out. So here is time on the x-axis and concentration of the drug on the y-axis. Now pointed to a few parameters here. So C is the maximum concentration of antibiotic that gets in the bloodstream. C min down here is the minimum concentration. 

And then over here, the area under the curve is the integral of this curve right here. It gives you the total area under that. And then MIC is right here, the dotted line. And then this represents your time above MIC. And then if you have a post-antibiotic effect, that's represented by the yellow bar right here. 

This is how it works when antibiotics go into the bloodstream. You're going to have certain antibiotics that are dependent on the C max right here. And these are your aminoglycosides, daptomycin, and to some extent fluoroquinolones. There's some messiness with the fluoroquinolones right there. 

But these drugs you want the maximum to be as high as efficacious. And then you want them to go away. There's a post-antibiotic effect, so you don't need it to hang around that much longer. 

Your next group is your AUC, or Area Under the Curve, over MIC, which means that there is some way it's kind of dependent on both of these, both the C max and the time over MIC. This is vancomyocin, which we all know and love. Linezolid, macrolides, tetracyclines, and then fluoroquinolones again. So they have both of these properties going on right there. 

And like you see, they're yellow. So there is a post-antibiotic effect as well, too, with these. And then finally, our beta-lactams down here, which are mostly dependent on time over MIC. So this is how your basic breakdown is of the major classes. 

So with that, I'm going to expose myself here and ask a question, and ask you to recall something I said 30 seconds ago. What is the purpose of a gentamicin trough? A, does it help determine efficacy of therapy? B, does it determine potential toxicity? C, inflict undue pain of suffering on interns. Both A and B, or none of the above. 

I can guess where C is coming from. So the majority of you think it's B, to help with potential toxicity. And you're exactly right. 

So to go back to the slide here, we're at aminoglycosides. Our C max is right here. We want the drug to get as high as it needs to be. And then in the aminoglycosides case, for UTIs, systemic infection, pulmonary, and sepsis, these are our gold peak levels. 

But after we have our peak we want our C min, we want our trough to be as low as possible. So we want the drug to go in, we want it to go out. Which is different from beta-lactams in the area under the curve [INAUDIBLE] drugs. Good job. 

We're going to focus our talk, the rest of the talk on the beta-lactams and time over MIC. So for ideal dosing, like I said before, we want to maintain free drug over MIC. Therefore, if your renal function's stable, your volume distribution is normal, you're MIC is low you should be fine. How often does those three line up in our patient population? 

We know that critically ill patients are different. They have increased volume distribution because they have extra volume expansion, capillary leak, AKI is unpredictable, fluctuations in renal function. ECMO is a complete black box here. And then we have a higher likelihood of multidrug resistant organisms, so higher MICs. 

So what does that translate to for us? Well, if we look at a schematic right here, where we have red, which is below our MIC, green is above. And this is our drug dosing right here. 

If we had a patient who all of a sudden, we give volume expansion to, their BD goes up, this essentially what happens. Your drug levels will go down. You risk going into the red right here. So that's an effective drug, that's bacteria is able to grow back, increase chance of resistance. 

This is what happens with your MIC rising up. Same effect. It's harder to get your drug levels where they need to be in order to be efficacious. 

So I want to talk about the kidney function in a little more detail. Because at least in my mind, kidney function, when it's really, really good actually makes a problem for me. And Vineet and Dennis, I know you guys probably don't like that statement. But my two specialties are not [INAUDIBLE] and organ systems so I can be objective here. 

This is a small study, 52 patients who got drug troughed. They're critically ill. They're getting either pip-tazo or a carbapenem. 

Just to orient you here, on the x-axis here is creatinine clearance. So this is poor renal function. This is great renal function. Y-axis is the trough of the concentration of the drug over the MIC. So if your trough was equal to your MIC it would be one. And that's this line right here. 

If you're down here, your drug is ineffective. It's below the MIC. At least for beta-lactams, you're not getting effective therapy. So let's call this the rectangle of regrets and sorrows. 

In contrast, up here, if you're above your MIC, these drugs right here, you do have efficacious drug. You are above your MIC. Let's call this the triangle of truth and triumph. 

Green is good, red is bad. Red is a government shut down, green is NIH employees able to take care of patients and get paid for it. Not that I'm bitter about that. 

What this group found, and you can see here, as your GFR goes up your likelihood of being able to have drug that's at the level you need them to be significantly goes down. Their crossover point here was 150 for estimated GFR. In effect, 42% of the patients included in this trial had a drug trough below the MIC threshold. 

And they found that the likelihood increased as your GFR went up. So if your GFR was above 130, and let's just cut off the graphic here, 82% of those had a trough below the MIC. So look at how much red you have here compared to the green. 

Now the way they did this, their ideal was to have a trough above the MIC right before the next dose. So that would essentially mean, if you were above the one line here, that you were 100% of the time above the MIC. So there is some debate over whether we need to be above MIC 100% of the time, like I said before. But you can't argue with the fact that this does get harder to do whatever your goal is when your renal function is really, really good. 

We're going to come back to this. This actually translates to clinical outcomes as well, too. So this was a trial done by Roberts, who publishes a lot in this area. 

384 patients who were treated with beta-lactams. About a third were getting continuous infusion. 93 of the patients receiving continuous infusion had a FT of MIC of 50% of dose interval. 

So that means that 50% of the time we were sure that their drug levels above the MIC for their isolate. Whereas, intermittent infusion, 80% of those patients. So 20% versus 7% were getting adequate dosing, or half the interval. And the patients who failed to reach this threshold were 32% less likely to have a positive clinical outcome. 

Now if you look at the graphic here, I've plotted out on a line, the one threshold right here. So what they're saying is, people who were FT MIC 50% below that threshold. So these are these people here, amoxicillin and ampicillin, they had less favorable clinical outcomes. And this is at 50%. And over here we're at 100%. 

So one could say, well, if we're looking at ICU-- DALI stands for Defining Antibiotic Levels in the ICU-- and we're using amoxicillin and ampicillin, maybe that's defines the poor clinical outcome in the first place. That's a fair criticism. 

But nevertheless, you can see that we're doing relatively OK if we're looking at half the dose interval for these antibiotics that we commonly use. But once we go to 100% we're starting to dance a little bit with that line right there. So we're starting to dip above, especially with [INAUDIBLE] plot. [INAUDIBLE] pip-tazo right here. Look at that. 

We do have a clinical basis for wanting to have our drug levels adequate. So next question. What are my options for GFR that's way high, and I want to give beta-lactam, and I want to make sure that levels are good? 

A, I'll give you that one. That's an option. B, keep the frequency but increase the antibiotic dose. C, decrease the frequency. D, this will do it. 

[LAUGHTER] 

This will do it but you may not want to go that way. And then combination. So a little more mixed here. So I'm glad nobody wants to end up on the first page of The Washington Post. That's good. 

[INAUDIBLE] infusion will definitely do it. Keeping the frequency but increasing the antibiotic dose. Now that would work too. I'll show you why in a little bit. 

Decreasing the frequency. You actually want to increase the frequency. Give the dose more often and that will keep you above MIC. So the answer is A and B. 

And I'll show this on the graphic here. So this is rectangular regrets and sorrows, this our triangle of truth. This is antibiotic that's really dancing with above and below the MIC. So option one, let's just increase the dose. That's what this would look like. 

And what common antibiotic do we do this for? The tazo. 3.375 standard dose. If we're not going after the pseudomonas, we increase that. Worried about MIC creep. 

If we do have pseudomonas and we're worried about it, so we increased from 3.375 to 4.5. And this is the effect. We drag the whole curve up. And then as a result, we spend less time in the rectangle of regrets and sorrows. So that's option number one. 

Option number two, increase the dose frequency. So can you think of non beta-lactam where we may do this for that's commonly used, like a fifth of patients? Vancomycin. So sometimes we'll either increase the dose of vancomycin. We do that plenty of times. But also, we've gone from Q12 dosing to QA dosing. 

Lower the dose sometimes. So your peaks are lower but your troughs are also lower. And you spend more time compared to here above the MIC. And if you work in a pediatric ward-- I did a PICU rotation-- they start their vancomycin standard at q six hour dosing just to take advantage of this effect so you can stay above the MIC. So that works too. 

An extended infusion. Now I couldn't figure out how to do this in PowerPoint. So you'll have to forgive me. There's a limit to the magic. 

This is going from-- and in the case the tazo, from a 0.5 hours to a 30 minute infusion. Extend it out to four hours and you're just dragging that curve out and staying longer above the MIC. And then your last option right here is continuous effusion, where you're keeping those drug levels relatively stable throughout the infusion of a 24-hour period. So that's what we can do. 

I want to talk a little bit about the data here. I'll just cut to the chase and say there's not a big gold standard slam dunk here. And there's a lot of data to sift through. But what I'm going to do-- I've been select about where we're going to talk about. 

I want to demonstrate why it's been so hard to get really good, high quality randomized trials demonstrating mortality benefit for doing this strategy. I'll tell the story of Dr. Joel Dulhunty, who's an intensivist in Queensland University in Australia, where he did two trials. One that looked at that pharmacokinetic levels, and another looking at mortality. One was positive one was negative. We'll talk about why. Then I'll talk about a Cochrane review briefly that show another issue with trying to study this. And then a meta analysis. 

So first, 2013, continuous infusion of beta-lactams on multicenter double-blind randomized control trial. These were five ICUs that were distributed among Australia and the special administrative region of Hong Kong. Participants had severe sepsis, planned to start these antibiotics. This is tamenzin right here. The tazo or meropenem expected to last over 48 hours. 

Their primary outcome was looking at the concentration of antibiotic over MIC at days three and four. The secondary outcomes you can see here, clinical cure, ICU-free days, survival, et cetera. Got 60 patients, a relatively small study, 30 in each group. 

So this is what they found. Continuous effusion led to higher rates of clinical cure. Now a little confusing the way they display the data here. But the box and whiskers plot, the white here is intermittent bolus. And here is continuous infusions. So these are your three antibiotics with your two different dosing strategies. 

They found that compared to intermittent bolus your continuous effusion no surprise. You had higher levels at a ratio of around 2.8 to 1 or so. So they were reliably able to have their concentration of antibiotic over the MIC, 82% versus 29%. 

And then when they looked at clinical cure, this was assessed by an independent board, who determined whether the patient was cured of their disease. For either mcirobiologically or by clinical parameters, they found a favorable outcome for the patients who were getting continuous infusion antibiotics by 70% to 43%. 

So they saw this trial and said, well, this is great. How about we expand it and look at these outcomes here, ICU-free days, mortality? We'll need a bigger trial, we'll need more time. But let's go for it. 

So 2015. This time 25 ICUs. They were going in much bigger now, Australia, New Zealand, Hong Kong still. Severe sepsis, same antibiotics. Primary outcome now is the number of alive ICU-free days at day 28. And then secondary survival, clinical cure, organ-free failure, the rates of bacteremia, et cetera. 

So we're looking at harder clinical outcomes here. 433 patients randomized. They had to screen around 3,000 patients to get to this. 219 versus 224 in effusion and bolus, respectively. 

These were the microbiological isolates. The first thing I want to point out to you is you saw 443 participants, or 400 or so that they enrolled. This is the number of isolates they got. So 40 and the continuous group in 43. So you're looking at about 20% of their enrollees had positive cultures. We'll get back to that. 

But you have about a fourth of gram positives. The rest are made up by gram negatives. They didn't have a lot of intermicrobial resistance, at least looking at the absolute numbers. 

And this is what they found. ICU-alive days, no change. 90-day survival, no change. Clinical cure, no change. Length of stay, somehow a significant value that they waved off that they couldn't really explain. Adverse events were similar between the two groups. 

So negative across the board. So you might ask, why is that? They did this trial two years ago and found at the very least that clinical cure was efficacious with continuous. So what gives right here? This is their Kaplan-Meier curve that again shows no difference 

So let's look at them side by side. The thing I want to point you to is right here. The 2013 study, for whatever reason, had no patients who were on continuous renal replacement therapy, or intermittent therapy. Whereas, a quarter of their patients in 2015 were on some form of renal replacement therapy. And I remind you of the triangle of truth, rectangle of regrets and sorrows. 

So remember, if you have renal failure, you're more likely be around here. And that means it's going to be much more likely that your drug with intermittent bolus is going to be the triangle of truth. So that means if you're trying to study the difference in antibiotic delivery strategy, you're not going to be able to tease out much of a difference if your drug at control is probably going to work just as well in the first place. 

There are some other things too. Like I said before, only 19% of the cases had documented infection. So those 81% of patients who did not have a documented infection by blood culture, some of them probably did not have an infection. So you're not going to find a difference between your antibiotics if there's no infection. 

And then there's a low likelihood of MDR organisms in Australia and New Zealand. It's something that they're very proud of, but didn't help them in this trial right here. It is a good thing. 

So let's talk about this Cochrane review from 2013. The thing I want to point you out to is the Cochrane reviews do these cool little tables where they show the different sorts of biases that your different studies can include. They include random sequence, generation, blinding, incomplete data, and outcomes. 

This was a negative trial that did not find any benefit to continuous or intermittent infusion. But they pointed to a lot of heterogeneity in the trials right here. So you're seeing a lot of red in both the plot right here and in the diagram right here. 

In contrast, this is a trial recently looking at oral direct thrombin inhibitors versus oral factor Xa inhibitors for a pulmonary embolism. You can see much less bias right here. So this would be ideal. But what happens with a lot of these antibody trials is that you have a lot of red. 

Let's look at a meta analysis that had a positive outcome for continuous infusion. This was 2014. These are the trials you can see here on this side. There's a couple of ones that we saw earlier. 

And over here, favors infusion. On the other side, favors bolus. And we can see here that we have positive effect right here. 0.66 risk reduction ratio for continuous infusion. 

The thing I want to point to is, when we looked at whether randomized controlled trials and non-randomized controlled trials supported the data, there was a big difference here. So if we looked at randomized controlled trials you can see here, you cross one. Whereas the non-randomized controlled trial, 0.57 and we stay pretty tight. Not a lot of heterogenicity. This was for mortality. 

If we look at clinical success, RCT is much more patients. We cross one right here, and we're at one. But we see a benefit with non-RCTs. So this demonstrates that with the high quality trials, we're not seeing a lot of those outcomes for a lot of these reasons that I outlined before. A lot of these randomized or non-randomized observational studies is where you do see a lot of the positive outcome. 

And that's because finding the right population of patients to study is going to be hard. So you'd want to have ideally people you suspect infection with. That's an easy one. 

You want to make sure that they have the infection that you suspect them of. You want the organism to be susceptible. You don't want outright resistance. But they can't be too susceptible because if your MIC is super low it's going to be easy to get your time above MIC with standard dosing anyway. So you need an elevated MIC. 

And then for the reasons of the triangle of truth, we might want to exclude people who have renal disease. So how many of these people, who don't have renal disease, have proven infection with elevated but not resistant MIC with the organism that we can document? And then we can split those groups into two and then compare whether continuous or extended is going to work for those people. 

And that's going to take a lot of time and a lot of money. And there's just no interest in developing a trial that's going to be high quality. So even these meta analysis point to the fact that we need randomized trials, we need randomized trials. I'm not sure we're actually going to get them. I think we're going to have to go what we got right now and use clinical rationale and what we have. 

So what about the cost? This is interesting. If we're looking at standard dose pep-tazo, again, 3.375 to 4.5 q six hours, that means we would give in the course of a day four doses, or four bags of this. And that's 13 and a half to 18 grams per day. 

Extended infusion, 3.375 q eight hours. Now the dose, 3.375 doesn't change according to renal function. It's just an interval. So at the very least, you're going to be giving three doses per day, and then sometimes two doses per day. That's 10.125 five grams per day. That's a cost or a savings at least of the drug of 3.375 to 7.8 or so. You're saving at least one to two bags of this stuff per day if go to extend infusion strategy. 

There was an interesting little trial out of Albany that looked at changing their antibiotic strategy to all extended infusion for pip-tazo. And they looked at a host of outcomes. This was done between 2000, 2004. Midway in between they started doing extended infusion for all their pip-tazo. We had about 200 patients. 

These were their findings. So mortality, 12% versus 31%. [INAUDIBLE] in hospital stays, 21 versus 38, favoring extended infusion. Single center, [INAUDIBLE] cohort study outcomes that we don't expect, I got you. We're not going to look at that. 

But look here. Their total daily dose was reduced by 25% to 50%. So one in three doses per day. And this saved them 68,000 ranging up to 135,000 in annual direct drug acquisition costs. 

I got a few nervous responses, but I managed to pull the pip-tazo expense at UVA. Someone sent me this spreadsheet. Thank you, Danielle Griggs. 

This is our annual acquisition costs for pip-tazo. We buy 30,000 vials at a cost of $228,000, or $329,000 per year. So what would it look like if we were to do a strategy like this? 

I spend a lot of my time in the business school. So let's do a little financial model here. If I put down here our cost of acquisition, $329,000. These are the different prices of the different vials. 

And let's say we reduce our purchase by 35% or so. For some reason, it's not working. It worked earlier. 

But if we reduce by 25%, then we're looking at a cost savings of around $75,000. If that extended out to 50%, it's around $150,000 or so. This is just by changing how we dose antibiotic without changing outcomes, at least from renal failure, and potentially improving outcomes. 

There is some data for this now. So last year, a meta analysis published in CCN looking specifically at extended infusion pip-tazo. This is looking at mortality here. And I'll spend a little bit of time orienting right here. 

Up here are three studies where patients had an average mortality that was above 20%, so healthier patients. All these studies required either a SAPS, or a SOFA, or APACHE score to predict mortality. And then down here are people who had a higher risk of death. 

So you can see here, at least from a clinical cure standpoint, you do have a positive effect that favors prolonged infusion. So these two different models here, the fixed effect model and the random effects, are just how the data is distributed. So fixed effects assumes that all the trials that you're studying have similar enrollment criteria or similar treatment strategies. They're pretty lined up. 

Whereas random effects allows for a little bit of distribution in how the trials are done. So you would expect to see a wider distribution with the random effects model, which you do see here. It holds for both models here. 

So this seems to favor extended infusion, at least for your critically ill patients. You do crossover right here up top. But a positive effect, as well. 

And then looking at mortality, you have to flip it now. Now the left side favors prolonged infusion and the right side does not. But same effect here. We can see with our lower acuity patients we have a positive effect. It does play with one, but lines up on the left side. 

And then down here, in both the random and fixed effects model, pooled. You see a positive effect right here. So at least for pip-tazo we have a meta analysis that demonstrates that there's some benefit for clinical cure, and may be a benefit for mortality, especially in our sicker patient populations. 

So simpler dosing strategy simplifies the supply chain, makes ordering easy for you guys. Who would not like this? The nurses. 

[LAUGHTER] 

This is Heidi [INAUDIBLE], Joe [INAUDIBLE] sister. This was funny. I didn't tell in the context of my talk. But I asked her, so what would you think if I wanted to switch all of our pip-tazo to extended infusion dosing? She said, oh, come on. 

Why would that be? The answer is simple. If you're driving home from today, and you depend on the 250 bypass. And I told you, I'm going to block one of these lanes for 30 minutes. You might say, all right, whatever. It's a little minor nuisance, or whatever. Maybe I'll finish that podcast on the way home. 

But what if I told you I was going to block it for four hours. That's a much bigger deal. So nurses, you've got to understand nurses. If you walk around with a short sleeved shirt on, nurses will look at your antecubital fossa and judge how easy it is to get access. That's how they think. They love access. 

If you're going to tie up their access for eight or nine hours a day, that's a big deal. That means more IV sticks for them, more work for them, more hassle, more IVs infiltrating. You're just tying up a line for most of the day. So they really don't like this. 

There are some other considerations as well, too. So for compatibility, you're tying up the line longer. Then you're going have to pay more attention to what can actually infuse with your drug. So that can be an issue as well, too. 

And then stability of the drug at room temperature. So for instance, meropenem, the package insert says that it stays stable at room temperature for one hour. So if you're infusing it quickly it's not a big deal. But you extend that, the data does show, at least some small studies, that you do maintain stability with the longer infusion. But it goes against the package inserts. It takes a lot of guts to go against that. 

So these are some real issues that will come up with this. But overall, I would say that the high quality RCTs to support prolonged infusion, I would say the gold standard do not exist and they probably won't for reasons we outlined. But it makes sense based on pharmacodynamic principles. 

There is evidence of benefit from observational studies and limited RCTs and meta analysis without evidence of additional toxicity. And the people you want to think about this are people with deep-seated infections, CF and structural lung disease where it's more difficult to deliver antibiotic, the severe infection plus super normal renal function. So thinking about a really good GFR is a potential problem we don't usually think about, and in pathogens that have an intrinsic high likelihood to develop resistance. I'm thinking of pseudomonas, B cepacia, and acinetobacter. These are when it might make sense to consider doing this. 

Now to go back to our patient, who has refractory anemia, multiple infections, shock, respiratory failure. He's on all the antibiotics. He's got a new fever and positive gram negative rods with super normal renal function. What did we do? 

We began extended infusion septaz, avibactam. We added tigecycline to do something else. It did not fix anything. 

We went to the OR for a partial colectomy and diversion. And then underwent a repeat trial of HD cyclosporin, high dose cyclophosphamide. You got campath. None of it worked. 

His only chance was to get a stem cell transplant. Do you want to talk to me about why he got a transplant? There are reasons. But as you may have predicted, did not go well. He developed profound organ failure afterwards. 

He hung on for about 20 days after that. Passed away. And his autopsy showed the disseminated aspergillus had involved essentially every organ that they looked at, as well as his vasculature. So he had a bad outcome. But it was a good learning case for all of us. 

These are my references. I want to do a special shout out to our pharmacists. One, because they are the ones who are reliably as nerdy as I am when it comes to things like this. So they're very helpful. 

And in particular, Becky Hockman, Heather Cox, and Allan Stillwell. I don't see her here. And Danielle Griggs, who pulled the pharmacy data for me. I love pharmacists because they improve outcomes, they decrease costs, and they make our job easier. So make sure to thank them when you see them again. 

And then some obligatory pictures of people who make this easy, possible for me. This is the Boston Symphony orchestra. This is my wife. She's a pretty good singer. 

[LAUGHTER] 

This is my daughter Ruby, who's 10 months old. And the joy of her life was knowing that she could reach the piano keys. This is my son Alain, who's five years old, who was worried about me being nervous for this talk. So he asked that we wear a pair of father-son socks. So we're wearing matching socks today. 

[LAUGHTER] 

That's it. 

[APPLAUSE] 

What about infections where you're worried about penetration [INAUDIBLE]? 

It makes sense from a theoretical standpoint. I wasn't able to find trials looking at specific organ sites for that. I may not be aware of some. If anyone knows of any please let me know. 

That does make a lot of sense to me, especially in sites where it's hard to get that drug in. You really want to push the pedal and keep your drug above MIC as long as possible. And I think personally, it's better to have a more even time above than these peaks and valleys that we typically use. 

Question [INAUDIBLE]. 

The other patient population from where the [INAUDIBLE] is shown to be helpful, [INAUDIBLE]. Are there [INAUDIBLE] papers [INAUDIBLE]? 

Great point. Thank you. Anyone else? 

[INAUDIBLE] and it's effect on [INAUDIBLE]. Just wondering [INAUDIBLE]. And then the studies really do not fan out when they were completed. I'm just wondering whether we should rethink [INAUDIBLE] in the human body when there are other factors like [INAUDIBLE] cells that also participate getting [INAUDIBLE]. 

I [INAUDIBLE] patients who had more renal failure and more [INAUDIBLE]. That might be in part explain why you see more [INAUDIBLE] when you're [INAUDIBLE]. When you're working with this set of septic patients, you also see [INAUDIBLE]. 

The question slash comment was about are there other factors at play that are contributing to these negative results in addition or outside of just the drug levels themselves? My short answer is, I would love if people had an immune system. I think that helps immensely. 

That's the main lesson from my ID fellowship is that having an immune system is pretty damn good. So I think that's definitely a contributing factor. And certainly in our case, our patient did not have an immune system for a long time, over a year did not contribute to anything. 

But I think outside of looking at factors like that, I would say there's a strong pharmacodynamic and kinetic reason to look at extending infusion antibiotics. And this is just based on how they work. I think the trials, getting to the perfect sort of trial where we get that perfect patient population where we can split it out, that's going to be very, very difficult. 

And looking at patients who have a decreased risk of mortality in the first place, so your lower acuity patients, it's going to be harder to tease that benefit out as well, too. Because they're not going to be that critically ill subset that have these dosing issues, like the bottom distribution is why their renal failure is fluctuating. And those people I think is where we need to think about doing it, not necessarily on the floor. 

That does make me think. If I encountered the IV drug abuser who is on nafcillin, and they're just chewing through other drugs, and they have a deep-seated infection, I am going to think about that. But I don't think the data necessarily supports using the strategy for that person specifically. It's much more solid in the critically ill patient population. All right. Thank you. 

[APPLAUSE]