Hi, everyone. My name is Gideon Stitt. I'm a critical care pharmacist and T32 postdoctoral research fellow in clinical pharmacology at the Children's Hospital of Philadelphia. I'm excited to talk to you all today about when does the pediatric patient become an adult. A relevant question and a clinical challenge as we see more and more large teenagers within pediatric ICUs and need to figure out at what point do I think about this patient still as a child, do I think about them as an adult, and how do I dose their various medications optimally? My only disclosure today is that I do receive grant funding from the NIH to support my postdoctoral fellowship. I have three main objectives for this talk. At the completion of the activity, I hope that you'll all be able to, number one, explain the basic pharmacokinetic, or PK, changes that occur as children grow older. Number two, describe some of the mathematical techniques that seek to quantify these maturational PK changes that occur in patients. And finally, to apply principles of developmental pharmacology, as well as allometry, which we're going to get into a little later, to develop optimal dosing recommendations in larger children. We'll start with a figure that many of you may have seen before that nicely summarizes the four main processes that reside under the umbrella of pharmacokinetics, absorption, distribution, metabolism, and elimination. We start on the left side by administering a dose, and there may be some absorption process, particularly if given enterally, topically, subcutaneously, et cetera. Or we can sort of bypass this by giving something intravenously. Then enters the central compartment where we think about the distribution of the drug. It may be going off to tissue reservoirs. It may be going to a site of action. It may be going places that we don't necessarily want it to go. Following that, we'll have some biotransformation through the metabolism of most drugs. Some drugs may be activated if they're a prodrug, or they may be activated from a parent drug into an inactive metabolite. Finally, we have excretion where we clear the drug and eliminate it from the body. Starting with absorption, we'll define it as the rate and extent at which a drug leaves the site of administration and moves into circulation. So it's really this process of getting a drug into the body. We'll contrast that to bioavailability, which is a fraction of an administered dose that reaches the systemic circulation as intact drug. So we have to think about not only what dose am I giving, but how much of that dose is actually reaching the circulation of this patient as intact drug. It may be not absorbed through the intestinal tract. It may be partly metabolized by intestinal SIP enzymes, for example, but we're really focusing on intact drug when we think about bioavailability. As our standard intravenous administration is 100% bioavailable because we are physically delivering intact drug into the systemic circulation. And finally, bioavailability considers the extent, but not the rate of absorption overall. There are a number of factors that will affect enteral absorption of medications. And we'll start on the left side, thinking about developmental factors. In neonates, achlorhydria and low intraduodenal levels of bile salts may slow the absorption of different medications. Younger infants have decreased gastric emptying, meaning that if a drug is absorbed in the intestines, it will take longer for that drug to reach its site of absorption. And you may see a slower onset or slower peak as that drug is absorbed. And finally, in children less than five years, there are changes in first pass metabolism, particularly as various SIP enzymes that reside within the intestines and the liver develop further. When we then think about critically ill patients, they often have decreased GI perfusion, perhaps due to shunting of blood elsewhere. They'll often have reduced GI motility and gastric emptying, perhaps from things like opioids or barbiturates that we're giving them. And then physical things like enteral tubes and nutritional formula interactions. Some drugs may stick to an enteral tube. Some drugs may particularly stick to the various enteral nutrition products. And so all of these can affect enteral absorption. The next pharmacokinetic parameter we're gonna think about is volume of distribution. And all volume of distribution is, is a relationship that relates the amount of drug administered to the concentration achieved in the blood or plasma. It may not always be an identifiable physiologic volume. For example, you could calculate a volume of distribution that is literally a larger volume than your patient. All that's telling you is that if your patient were one homogenous bucket, that that is the volume of that bucket that would be necessary in order to relate the dose administered to the concentration in the blood. It's simply telling you that that drug largely leaves the blood and goes out into the periphery. Whereas a drug with very small volume of distribution of small bucket will relate that dose and that concentration. And therefore that drug is largely staying within the vasculature. You can see how it's defined here. Volume of distribution or VDN liters is simply dose over concentration. There are a number of developmental factors that can alter the volume of distribution for a drug as patients get older. The largest among them being changes in total body water and body fat. With total body water decreasing as patients age from newborns to infants, children, teenagers, and adults, and alternatively a relative increase in body fat over time. Depending on whether a drug is hydrophilic or lipophilic, this can alter this drug's volume of distribution as a patient ages. Moving next to changes in metabolism over time. Broadly speaking, metabolism prepares drugs for elimination. All it's really trying to do is make a compound as hydrophilic as possible so that it can eliminate it from the body. We really think about it in two phases. Phase one metabolism, which are generally functionalization reactions largely driven by the CYP enzymes or the cytochrome P450 enzymes, often to inactive metabolites or can be used to give prodrugs their activity. Phase two metabolism are conjugation reactions. So we think of things like sulfation, glucuronidation, and acetylation. And we're adding additional moieties onto drugs in order to make them more water soluble and help to excrete these compounds. This slide demonstrate changes in metabolic capacity as children get older. So if we focus on the blue, red, and green bars, we have age along the X-axis and percentage of adult activity along the Y-axis. And you can see that there's quite a bit of development over the first year or so of life. Slide focuses on the phase two processes of sulfation, acetylation, and glucuronidation. And again, you can see that over time when these processes sort of develop is different. Elimination of drugs can take place as either unchanged drug or as metabolites. Renal elimination is the one that most people probably think of most often. And it's really comprised of glomerular filtration and active tubular secretion. And some drugs are also reabsorbed through passive tubular reabsorption. There's also biliary and fecal drug elimination can be active secretion into bile and from enterocytes. And again, there may be some reabsorptive processes through enterohepatic recycling. And finally, there's other sites where we actually continue to eliminate drugs. Things like lungs, sweat, fliva, breast milk, skin, et cetera, depending on the drug and to different amounts all can contribute additionally to drug elimination. Here we'll look at changes in glomerular filtration rate over time. Again, with age along the X-axis, we have glomerular filtration in the blue with the right-hand Y-axis showing glomerular filtration rate. And again, you can see that rapid and significant changes over the first anywhere from three, four, five years of life peaking around age six. What we know is that premature neonates have lower GFR than term neonates. There's a very rapid increase in the first two weeks of life with significant increases after that in the next several years, with adult values being reached around 12 months and actually having some overshoot. Children in sort of the kindergarten years is how I think about it, tend to actually have better GFRs than adults and then things sort of settle back down towards adult values. Glomerular secretion, again, an important clearance mechanisms for some drugs also reaches adult capacity at about 12 months of age. So where are we so far? Well, basic PK or the classic description is what the body does to a drug can be summarized by the ADME processes, the absorption, distribution, metabolism, and elimination of a drug. All of these ADME processes undergo some maturational changes as patients grow. And then as they grow older. Most of what we've focused on so far really looks closely at the early years of childhood development. So we're looking at newborns and infants and young children, but the title of this presentation is all about when do we think about teenagers as adults? So that's what we're gonna focus on next. And the question is always, how do we address these adult size children? Do we just use additional weight-based dosing up to whatever the adult dose is? Sort of a classically used technique. Do we use adult doses for patients that are over 12 years old? Also kind of another clinical rule of thumb. Or maybe it's for 40 kilograms, one of those other sort of bedside rules that help you think about when is this child sort of pharmacokinetically an adult? And that's what we're gonna focus in on for the rest of this talk. We're first gonna drill down a little bit more on the concept of clearance, because it's something that we're gonna talk about quite a bit for the rest of the hour. Clearance represents the volume of blood or plasma from which a given drug is completely removed per unit of time. It can be described in terms of hours or minutes or seconds if you want to, but more classically hours or minutes. Clearance can occur through either metabolism and or elimination. The concept really applies to both processes, but it's simply how much volume are we removing this drug from? And calculate it as follows. There's several different ways of describing it, but this is probably the easiest one. So we have clearance in liters per hour is dose in milligrams over area under the curve, or that's how we often describe total drug exposure in milligrams times hours over liters. If we simplify that, we get liters per hour. Now that we've defined clearance, we need to think about scaling and how does clearance change as a patient gets older and gets bigger? And the most simple way to scale something is linearly. This assumes a one-to-one direct relationship between patient size and clearance. It is a straight line. Probably thought of as probably the poorest model when scaling metabolic processes as changes in things like cytochrome P450, or a renal function don't change at the exact same pace as a patient gains weight or gets older or grows taller. It is, however, probably the most commonly used because it's the simplest and is really what drives most weight-based dosing. Another way to think about scaling is through the use of something called allometry, which really can be defined as how characteristics of living creatures change with size. And it was actually developed, you're probably wondering why there's a picture of a sea creature on the screen. It was developed through research on fiddler crabs and trying to figure out why their claws grow at a different rate than the rest of their body in the 1940s. It has since been much further developed and applied through pharmacokinetics quite widely. Allometric scaling can be distilled down to a relatively simple mathematical equation Here we have Y as whatever biologic characteristic we want to estimate, could be clearance, for example, it could be volume of distribution or it could be something else, things like lung volumes or respiratory rate. Any of these can be scaled. W is body math, and A and B are empirically derived constants. Now, what these constants are may be different depending on what biologic process you're trying to scale. As allometric scaling was further studied and developed, a singular relationship was found to apply to quite a few different biologic processes. So in this graph, we're looking at body mass along the X-axis and metabolic rate along the Y-axis. And we're looking at homeotherms in sort of that yellow color, pikelotherms in the blue color, and metabolic rate along the Y-axis. And pikelotherms in the blue color and unicells in the red color. And really it's warm-blooded and cold-blooded and single cellular organisms. And what has been shown that really regardless of across species here is that when you take the log of the body mass and the log of the metabolic rate and plot them, you get a straight line with a slope of three over four or 0.75. This is a relationship that we're gonna come back to over and over again. On this slide, we have a table of different allometric exponents for various biologic processes. And we have both what are they predicted to be mathematically as well as what do empiric data show that they are. And largely speaking for things volume related, things like blood volume and tidal volume, we have an allometric exponent of one. So they scale on essentially a one-to-one relationship. Things like metabolic rate, cardiac output, glucose turnover has that exponent of 0.75 or three over four that we discussed on the previous slide. Whereas things relating to time and frequency, things like cardiac frequency or respiratory frequency tend to have an allometric exponent of 0.25. We're gonna bring it back and figure out how do we apply all this to pharmacokinetics and what does it mean for my 90 kilo teenager? Well, when we think about how to scale clearance as patients get bigger, it can be simplified to this equation. So clearance of our ith individual equals some standardized clearance times the weight of our individual over a standardized weight to the 0.75. And that describes the relationship of how this patient's clearance changes over time. The volume of distribution is even simpler. So the volume of the ith patient is equal to a standardized volume times that patient's weight over a standardized weight. That's it. This isn't to suggest that allometric scaling is the only way to describe changes in clearance as patients get older. Linear scaling has certainly been used. There are body surface area techniques that we'll just briefly look at in a minute. And the whole world of physiologically based pharmacokinetic modeling, which in many cases has been shown even superior to allometric scaling. So allometry is really sort of a core principle of population TK modeling, but there are other techniques. Additionally, there are modeling techniques where you may not use these standardized exponents and actually estimate them for whatever drug you're looking at specifically. However, these are sort of the classic ways that allometry is applied to pharmacokinetics. So it's important to think about how much of an effect does this really have. Taking something to the 0.75 power seems like it probably doesn't change it very much. And I think this slide and the next slide nicely demonstrate just why it's so important. Along our x-axis, we have weight. And along the y-axis, we have relative clearance. And we're going to focus on the yellow and the blue lines. The yellow line uses allometric scaling to 3 4ths power. The blue line uses linear per kilo scaling of a dose. And what we can see is that over time, the lower weights, linear scaling or per kilo, tends to underestimate relative clearance. And at high weights, it tends to overestimate it. I mentioned that there are body surface area methods of scaling as well. They are to a different exponent. But for most weights, can be made to perform very similar to sort of the classic allometric scaling that we're talking about. Overall, it doesn't really look that different. So then the question is, what's the big deal? Well, if we look here and compare, again, the blue line, which is per kilo or linear scaling, to allometric scaling, which is the reference, we'll ignore the red line for now. We can look at, as weights change along the x-axis, what is the percent difference in clearance between them along the y-axis? And if we have a cutoff of, let's say, we want them to be within 20% of each other, well, for any patient less than 30 kilos, there's a greater 30% difference in estimated clearance between these methods. If we want to tighten that up to something like 10%, well, now we're pretty close to 50 kilos before these different ways of scaling drug clearance sort of merge to within an acceptable limit. So they may seem relatively similar, but they can result in quite different answers when you're trying to figure out how to scale clearance. Again, most of this is really going to affect those smaller patients, but now we're getting up to those 40, 50, 60 kilo patients, and now we have to start thinking about, well, wait a minute, these are kind of getting to be those adult-sized patients. So this stuff actually matters for those kids. I think the biggest point that we need to make sure we get out of all of this is thinking about at what point does the curve flatten, and I can now think about, and I can now think about this pediatric patient as an adult in terms of how I dose their drugs. We've looked at various ways of scaling how clearance changes, and the biggest thing is that linear clearance, when it's scaled linearly, that's a straight line that never flattens out. Whereas if we scale something allometrically, and particularly if we then add in things like maturational constants to account for changes in SIP metabolites, SIP metabolism, for example, we can figure out where does that line flatten. This table is from a paper that looked at age-dependent enzyme activity as a fraction of adult values for a number of different SIP enzymes. We have SIP3A4, 1A2, and 2E1, as well as a couple of different UGT pathways within phase II glucuronidation. And I think the biggest thing, if you want to take a 10,000-foot view from this table, is that at 10 years of age, all of these processes had reached adult values. So in terms of these specific enzymes, a 10-year-old will metabolize a drug as quickly as an adult. That previous slide only looked at specific enzymes, but we can also look at things like liver and kidney development as a percentage of total body weight. Here we have age on the X-axis, liver as a percentage of total body weight along the left-hand Y-axis, and kidney as a percentage of total body weight along the right-hand Y-axis. And you can see that, relative to the size of the patient, they flatten out at a certain point. Really, by about 15 years, in terms of relative size to the rest of the body, the liver and kidney have sort of flattened out. The curves have flattened here. And so these patients are gonna be quite similar to an adult patient. To this point, we've really focused on maturational changes. And at what point do things like metabolism and clearance transition to adult values? And it really is kind of somewhere around that, you know, 12 to 15-year-old range, and maybe in the 40 to 50-kilo range. Classic bedside values in your head are reasonable checkpoints. I think the other population that we really need to think about, though, are the obese pediatric patients, because obesity doesn't necessarily mean that that patient is developmentally the same as an adult. So we're gonna first look at obesity and allometry and see how they kind of work together. So obesity is often defined by BMI in most studies that we're looking at. It mainly describes body size and only has sort of an approximate relationship with excess body fat. Remember, you could find, you know, Mr. Universe or a bodybuilder would probably be obese by BMI. Dose scaling based on classical allometry works reasonably well for what I'm gonna call large individuals. So in this case, I'm talking about the 15-year-old who's six foot six and the center on his basketball team. However, clearance organ function may vary less than expected when weight is driven by fat accumulation. So in this case, while it may be a patient who's the same weight as that, quote, large individual, our six foot six teenager, this patient may weigh the same, but they may be quite a bit shorter, they may be quite a bit younger, because more of their weight is made up from adipose tissue. And this is where it gets a little bit more complicated. There's a really nice paper that was published by Ryan Kras, looking at vancomycin dosing and what they term super obese adult patients. And I think it has applications to what we're talking about. So what they did was look at, they created a POPPK model in 346 patients, and they had a large range of body weight from about 70 kilos to just under 300 kilograms, as well as a wide range of BMIs from about 30 up to about 80. And here's their final POPPK model. You can see that they described drug clearance as a constant, they included covariates of patient age, patient serum creatinine, patient sex, as well as, there it is, allometrically scaled total body weight. And what they found is TDD, or total daily doses, above 4,500 milligrams were not required for the average age of the patient. And above 4,500 milligrams were not required to achieve target AUCs at clinically relevant values of clearance in this population. So even in these very large patients, they didn't need linearly scaled doses. Essentially, the take home is that our patients who are very large, driven more by adipose tissue, they may need a slightly higher dose than a patient who is relatively similar otherwise, without that adipose tissue, but probably not as much as we might expect. And this is sort of the output of that paper when they performed modeling and simulation. So what you're able to essentially do with their equation is put in your patient covariates, and you come up with estimated clearance along the left-hand side here, in liters per hour. Then you work your way across and you get what your recommended loading dose would be, the recommended maintenance dose, and then the probability of a 24-hour AUC greater than 400, probability of 48-hour efficacy with an AUC, again, greater than 400, as well as a probability of toxicity. So they've used this model to account for relevant covariates that we would think about in a large teenager, things like age, renal function, and total body weight, translated that into a scaled clearance, and from there have come up with dosing recommendations that optimize drug exposure and minimize toxicity. So now the rubber is really gonna meet the road. We have these adult-sized teenagers and you're asking the pharmacist or you need to come up with a dose. And I think there's a few key questions to work through in order to come up with the best answer. The first one is what is being dosed, because that matters. Is it a very hydrophilic drug where changes in adipose tissue may not affect it quite as much versus a lipophilic drug? How old is my patient? As we saw before, maturational changes tend to sort of flatten out somewhere in the early teen years. Are they what I'd previously described as sort of a large child with relatively balanced height and weight? Or is it a patient that has more significant adipose tissue? All of these are things to think about in terms of if I give a standard weight-based dose to this patient whose lean body mass may or may not be sort of what I would expect for a patient their age, upon which standard dosing recommendations have been based, am I gonna achieve the drug exposure and therefore the effect that I want? Our last step of all of this is figuring out when the rubber meets the road, how do I come up with a dose for the patient who's in front of me? And I think that linearly scaled or weight-based dosing up to an adult dose continues to be the most practical solution for most patients and for most drugs. As utilization of advanced modeling and simulation grows, we may see dosing recommendations become more complex. So for instance, that vancomycin equation that we went over, it sort of mirrors some of the dosing that you may think about back for things like neonatal gentamicin dosing, where you think about things like gestational age and corrected age and go through a little algorithm and come up with the dose. Well, that may be something that we are either able or need to do for drugs that are more sensitive to changes that have a narrow therapeutic index and we wanna make sure that we're optimizing for our patients. There are some bedside tools available though, and that's what I wanna dive into now. This was a really nice paper that was published by a group out of Colorado in 2015 on the development of recommendations for dosing of commonly prescribed medications in critically ill obese children. I used it every day in the PICU and I used it for every resident that I trained. And I think every ICU practitioner that I worked with, I pointed them towards this paper. So a really useful paper and one that I think everybody can good use. This group summarized dosing recommendations in obese patients for 113 medications. And what they found was that for these medications, 52 of them should be dosed based on total body weight, 43 based on adjusted body weight and 18 based on ideal body weight. This is an important question that frequently comes up when you have an obese teenager who you're admitting to the ICU and you think, what weight should I put in that dosing weight box? Because that's gonna affect every drug order. My recommendation has always been to put total body weight in there. And then if a drug, a specific drug needs to be dosed on either adjusted or ideal body weight to go into those individual orders and change them. Because most drugs, whether it's an obese teenager or an obese adult will continue to be dosed on total body weight up to a cap dose that is generally the standard adult dose. Here we have how they calculated these different weights. So ideal body weight, there are a number of different equations that are available. This group used the 50th percentile BMI for age times the patient's height to get the ideal body weight for that patient. Total body weight is total body weight and adjusted body weight is the ideal body weight calculated above plus a pre-specified co-factor that I'll show you an example of times the difference between their total body weight and their ideal body weight. So here's a little snippet of an awesome table that they include in this paper that shows you which drug in the first column there, which body weight to dose on in the middle column there, and then some comments over on the right-hand side. For example, something like sulfamethoxazole trimethoprim, they recommend basing on total body weight because of its wide volume of distribution into the extravascular space. If we come down in the middle to tobramycin, you'll see that they recommend an adjusted body weight with a co-factor of 0.4, which is sort of a traditional aminoglycoside co-factor. This table includes a lot of other useful drugs to think about, things like x-metatomidine, which they actually recommend using ideal body weight. Propofol is total body weight. So different drugs may be optimally dosed on different body weights. Most are total body weight. And then think about the others as specific exceptions that you'll want to continue to pay extra attention to. In summary, multiple PK parameters change as children grow older, particularly those affecting drug clearance and volume of distribution. Allometric scaling is probably a better tool than linear scaling for extrapolating doses between patients of different sizes, but it's difficult to do clinically. When you're at the bedside, you need something that can be easy to use. In that case, linear scaling can be used In that case, linear scaling to adult doses combined with the use of optimal size descriptors, like we talked about in that last slide, may be the best current tool for recommending doses in larger children. Now, a key point to keep in mind is that depending on the institution where you work, particularly for continuous infusions, things can get a little extra complicated because you may only dose them in a weight-based fashion, even for your larger patients, not every institution will have flat dosing in milligrams per hour, as an example, built into their computers or their smart pumps. And so in these cases, it's extra important to pay attention and think about at what weight-based dose for this patient will I be hitting sort of my either normal or maximal infusion rates and keeping those in mind so that we don't inadvertently give too much drug to a patient. Outside of that sort of specific example, again, thinking about total body weight for the most part, special body weight descriptors for specific drugs that need it and capping at adult doses is gonna do the job for the vast majority of our larger teenage patients, because at that point, their metabolic processes are quite similar to an adult and likely have been for a few years. And so pharmacokinetically, they will behave very similar to an adult and we'll scale them linearly as best we can.

pediatric pharmacokinetics

medication dosing

allometric scaling

critical care pharmacy

obesity implications

drug metabolism

pharmacokinetic modeling