With the epidemic of childhood obesity, it is not uncommon for prescribers to puzzle over an appropriate drug dose for an obese child. Defining the optimum therapeutic dose of a drug relies on an understanding of pharmacokinetics and pharmacodynamics. Both these processes can be affected by body composition and the physiological changes that occur in obese children. As a rule of thumb, 75% of excess weight in obese subjects is fat mass, and the remainder lean mass. Although it is reasonable to assume that increases in fat mass alter the distribution of lipophilic drugs and increases in lean mass alter drug clearance, good quality and consistent clinical data supporting these assumptions are lacking for the majority of drugs. The relatively few clinical studies that have evaluated the impact of obesity have often been limited by poor design and insufficient sample size. Moreover, clinical studies conducted during drug development rarely include (or are required to include) obese subjects. Guidance on dosing obese children ought to be provided by drug manufacturers. This could be achieved by including obese patients in studies where possible, enabling the effect of body size on pharmacotherapy to be evaluated. This approach could be further augmented by the use of physiologically based-pharmacokinetic models during early (preclinical) development to predict the impact of obesity on drug disposition, and subsequent clinical studies later in development to provide confirmatory proof. In the meantime, for the majority of drugs already prescribed in children, particularly those where the therapeutic range is narrow or there is significant toxicity, the lack of a validated body size descriptor to use at the bedside means the choice of dose will rely on empirical experience and application of the precautionary principle.
Statistics from Altmetric.com
Should drug doses be adjusted for obese children?
Childhood obesity continues to increase rapidly in the West and in developing countries undergoing nutritional transition. Prevalence rates across Europe vary, with the UK having one of the highest rates. At present, childhood obesity is commonly defined on the basis of weight relative to height, using body mass index (BMI). Children are defined as overweight and obese using the 85th and 95th percentiles of the UK reference curves (known as the National BMI Percentile Classification). According to the Health Survey for England, the proportion of obese boys and girls aged 2–15 years increased from 11% and 12% in 1995 to 17% and 15% in 2006, respectively.1 A more recent evaluation suggests that the rate of increase may be starting to slow in the UK, but nevertheless the prevalence of obesity remains unacceptably high.2 Furthermore, as the rate has increased so has the frequency of comorbidities associated with obesity.3 4
Hence, it is not uncommon for clinicians in primary or secondary care to treat illness in an obese child. Often this will require treatment with a medicine for which there is no guide regarding which body size metric to use in determining the optimum dose for the obese child.
Defining an optimum dose for obese children
Defining optimum therapeutic doses relies on an understanding of the relationship between systemic exposure to the drug (pharmacokinetics) and pharmacological response (pharmacodynamics).5 Whether obesity affects either of these processes has been extensively reviewed, but there is significant uncertainty and few practical recommendations for optimising doses have emerged.6,–,8 This is not surprising as there is a dearth of good quality clinical studies evaluating the impact of obesity on pharmacokinetics and pharmacodynamics. Perhaps a major reason is that clinical studies, the majority of which involve adults, rarely include obese patients and where studies have attempted to evaluate the effects of obesity, the results have been inconclusive.7 With the growing proportion of obese children, this gap in knowledge generates variable dosing approaches with the potential to confound the clinicians' ability to properly interpret the relationship between treatment dose and clinical outcomes.
In children, doses are typically individualised using either a direct measure of body size, such as bodyweight or body surface area, or a surrogate for ‘normal size’ or ‘normal function’ such as age. Using such measures to adjust doses for obese children assumes that body composition and function are similar in the obese and non-obese child. Although this assumption is likely to hold for normal weight children of varying size, it is unlikely to do so for obese children since excess weight is not composed of similar proportions of adipose tissue and lean body mass.7 9
Body composition and physiological changes in obese children
Obese children have highly significant excesses in whole body fat mass, lean mass and bone mineral content as measured by the ‘four component model’ (incorporating air-displacement plethysmography and dual energy x-ray absorptiometry).9 However, increases in fat mass are substantially greater than increases in lean mass and it is suggested that, as a rule of thumb, 75% of excess weight is fat mass and the remainder lean mass. Indeed, fat mass in obese children tends to be between 30% and 50% of weight, but approaches 60% fat in some children. This also serves to highlight the variability in fatness between individuals for a given BMI value.9 The majority of excess fat is located in the abdominal region, but substantial excess fat is also in the leg. For lean mass, less of the excess is in the trunk and more in the leg, with some also in the arm.9
Young obese children tend to be taller than their non-obese peers, which may reflect earlier maturation. To enable precise interpretation of the effect of obesity on body composition, both fat mass and lean mass therefore require adjustment for height.9 However, the greater fat mass and lean mass of obese children are evident even after adjusting for height. Obese children also have increased hydration of lean mass which has been attributed to an expanded extracellular water space. This over-hydration will inflate lean mass values, and this factor needs to be taken into account when assessing body composition changes over time.9
In normal weight individuals, adipose tissue receives approximately 5% and 8.5% of cardiac output in males and females, respectively.10 11 As fat mass increases, regional blood flow per gram of adipose tissue reduces, and this may limit the distribution of lipophilic drugs.12 There may also be alterations in levels of plasma proteins, in particular α1-acid glycoprotein, where a doubling has been reported in obese patients.13 Although such a change can affect the pharmacokinetic parameters of drugs that bind to these proteins, in most cases this will not influence the clinical exposure of a patient to a drug.14
Obese individuals show fatty degeneration of the liver and a rise in the prevalence of non-alcoholic fatty liver disease has been reported in obese youth.15 Case series reports suggest a male predominance and a higher relative increase in serum alanine transaminase than in serum aspartate transaminase.15 Such changes can be indicative of altered hepatic metabolic capacity, although the evidence to date suggests that hepatic metabolism of drugs is not modified.6 Although animal studies suggest that renal clearance increases in obesity as a consequence of the increase in kidney weight, renal blood flow and glomerular filtration rate, conflicting clinical data have been reported.16,–,19 The outcomes of the clinical studies, however, may have been confounded by comorbidities such as diabetes and hypertension that are known to affect renal function.
Effect of obesity on the pharmacokinetic–pharmacodynamic relationship
Pharmacokinetic considerations depend on whether the dose is being selected for acute usage (eg, anaesthesia) or for chronic conditions. Alongside drug related factors such as lipophilicity and binding to plasma proteins, the distribution of drugs to and from blood and other tissues depends on body composition and regional blood flow (table 1).
The increased fat mass and lean mass in obese children should conceivably affect the distribution of drugs into tissues and ought to be reflected in the pharmacokinetic parameter, volume of distribution (Vd). Loading doses are based on Vd and thus for a relatively hydrophilic drug, where distribution is restricted to lean tissue, calculation of doses should be based on ideal body weight (IBW). For drugs that partially distribute into fat tissue, doses should be based on IBW plus a percentage of IBW (table 2). Lipophilic drugs that distribute predominantly into fat tissue or freely between lean and fat tissue should be based on total body weight (TBW).6
So far, the clinical evidence to support this approach is mixed. There certainly seem to be sufficient data to suggest that hydrophilic drugs (eg, rocuronium, aminoglycosides) whose Vd has been determined in normal weight subjects to be small, should be administered according to IBW or a percentage in excess of IBW rather than TBW.20,–,22 However, the distribution of highly lipophilic drugs in obese patients may not be as predictable and obese patients may not necessarily display a Vd that is larger than that found in normal bodyweight individuals, that is, lipophilic drugs in vivo do not necessarily distribute into adipose tissue extensively.23 24 The complex interplay of drug physico-chemical properties and physiological factors that influence distribution in vivo is perhaps not as well understood. For practical purposes, where a drug has a known large Vd in normal weight patients, the choice of an initial loading dose ought to be balanced between the rapid need to attain therapeutic concentrations and the safety profile of the drug. Rather than administering a single large loading dose (calculated using TBW), one option would be to split the loading dose into smaller portions and administer them as a series of mini-loading doses, assessing response after each dose.25
The situation is even more uncertain when attempting to select maintenance or chronic doses for obese individuals. Maintenance doses are based on clearance, a function of intrinsic metabolic capacity and perfusion of organs such as the liver and kidney, which constitute lean mass. Although absolute lean mass increases in obese children, it is a smaller proportion of the excess weight and therefore the percentage of lean mass per kilogram of TBW is reduced.26 As a consequence, clearance of drugs does not increase proportionally with bodyweight in obese individuals.26 This leads to the conclusion that maintenance doses in obese children should be adjusted to a body size metric closely related to lean mass such as lean body weight (LBW) and not TBW (table 2). However, at present there is no simple bedside method for calculating LBW in children. Moreover, although this approach is supported by theoretical pharmacokinetic considerations, the few clinical studies that have been reported have provided inconclusive data. The studies have been limited by poor design (to evaluate the obesity effect) and insufficient sample sizes. A larger than anticipated between-subject variability may be related to the fact that individuals with the same BMI differ in their body composition of fat mass and lean mass, particularly between ethnic groups.6 9 Recent debate for and against the use of LBW as the body size metric for calculating doses of drugs for use in obese adults has shown a close relationship between LBW and body surface area (calculated on actual body weight).27 28 However, neither of these size metrics explained more than 23% of the total between-individual variability in absolute clearance for eight different anticancer drugs in obese subjects.28 Furthermore, in cancer chemotherapy, the concern has been expressed that dose reductions using various adjustment methods such as LBW or dose capping may actually negatively impact on patient outcome and survival.29 30 There is clearly a need for alternative approaches to dose calculation in obesity that reach beyond body size.
Table 3 summarises the published evidence from pharmacokinetic studies for the optimal body size metric on which to base doses in obese adults. No equivalent information is available from studies undertaken in children, but the drugs represented are commonly used in paediatrics. For a number of drugs, even where research studies have been completed, consensus has not been reached on the optimal size descriptor on whichto base loading and maintenance doses.
Very little has been reported on the impact of obesity on pharmacodynamics, although the few reports that have emerged suggest that obese patients may exhibit an altered response to pharmacotherapy. Obese adults with type II diabetes have been shown to be more sensitive to the blood glucose regulating agent, glipizide and less sensitive to the antihypertensive effect of verapamil.31 32 This is of interest since medical comorbidities typically associated with adult obesity (eg, type II diabetes and cardiovascular disease) are becoming increasingly prevalent in paediatric obesity.33 There is also evidence that obesity may influence the natural history of asthma control and the response to different agents.34 Nevertheless, as with pharmacokinetic studies, the numbers of subjects are limited and the findings need further investigation. The challenge will be to design studies able to distinguish between pharmacokinetic and pharmacodynamic changes in obese children.
Role of modelling and simulation in predicting the effects of obesity on pharmacokinetics
To properly study the influence of obesity on pharmacokinetics, a dual ‘bottom up’ and ‘top down’ modelling approach has been advocated.35
A ‘bottom-up’ approach that involves the development of physiologically based-pharmacokinetic (PB-PK) models is being increasingly used in drug development to determine a candidate drug's likely pharmacokinetics and drug–drug interaction potential prior to selection for first time dosing in humans. PB-PK models can reliably predict systemic drug exposure in both healthy and diseased population such as those with liver cirrhosis.36
PB-PK models draw together the physiological and biochemical data that determine drug absorption, distribution, metabolism and excretion and then link them in a physiologically realistic ‘systems’ model (figure 1). A representation of a PB-PK model alongside a classical two-compartmental pharmacokinetic model is shown in figure 1. In contrast to the classic compartmental models, PB-PK models require a large amount of systems information on physiology and biochemistry, such as organ size, organ blood flow, organ tissue composition and drug elimination pathways including the enzymology of hepatic enzyme systems. The building of these ‘bottom up’ models and their validation is thus very time consuming. The models also require information on the drug (chemical and physicochemical data such as fat solubility) and on study design.35 However, since these models are able to account for changes in body composition and also drug characteristics such as lipophilicity, modelling the pharmacokinetics of drugs in obese and morbidly obese subjects is possible. An adult PB-PK model for lean (BMI 18.5–24.5), obese (BMI >30) and morbidly obese (BMI >40)) adults has been developed and evaluated for nine drugs where data are available clinically. The clearance ratio (obese:lean) was less than 2-fold in all cases and between 0.8- and 1.25-fold in six out of nine cases.37 A recently published PB-PK study has predicted the concentration–time profile of propofol in morbidly obese and normal weight adult subjects following a single dose and continuous infusion. For a single dose, the model predicted a faster washout in morbidly obese compared to normal weight subjects in agreement with a clinical study.38 The model also predicted an accumulation of propofol during infusion and a longer washout on stopping the drug in morbidly obese compared to normal weight subjects, although this was not backed up by the clinical data.
A ‘top down’ approach will involve improved paediatric clinical study designs that, whenever possible, endeavour to explore obesity related changes in the exposure–response relationships through incorporation of pharmacokinetic–pharmacodynamic (PK–PD) assessments. The adoption of population PK–PD modelling methods that utilise sparse data and advances in microanalytical methodologies for drug assays overcomes some of the ethical and practical challenges of conducting such studies.39 40 Another challenge will be to create sensitive clinical tools to assess the relationship between drug exposure and response in children of all sizes.
Discussion and conclusion
As is the case with adult obesity, the impact of childhood obesity on drug doses is not yet well understood. Although it is reasonably well known that excess weight consists of a disproportional increase in fat mass compared to lean mass, what impact obesity has on physiology relevant to drug disposition, such as protein binding, liver and renal function, remains unclear. Moreover, although it is reasonable to assume that the increase in fat mass has the potential to alter the distribution of lipophilic drugs and the increase in lean mass to alter drug clearance, it is not possible to reliably predict the magnitude of dose change, if any, for a specific drug based simply on its physico-chemical properties or known pharmacokinetics in normal weight patients.
For the prescribing paediatrician, there is usually little guidance forthcoming from the marketing company. This is not altogether surprising since more often than not they have no supporting data upon which to base recommendations. Clinical studies conducted during drug development rarely include (or are required to include) obese subjects. This may be historical, since there does not appear to be any good reason why healthy obese individuals are not enrolled in early phase studies. Indeed by recruiting obese patients into studies where possible throughout a drug's development cycle, the effect of body size on the PK–PD relationship could be evaluated using modelling approaches. This approach could be further augmented by the use of PB-PK models during early (preclinical) development to predict the impact of obesity on drug disposition, and subsequent clinical studies later in development to provide confirmatory proof.
In the meantime, for those drugs already prescribed in children and where the therapeutic range is narrow or there is potential for significant toxicity, urgent clinical studies are required to determine how pharmacokinetic parameters change with a relevant measure of body size (table 2). Although clearance of drugs is reported to be most closely correlated with LBW, the debate is still open in terms of a universal ‘size’ predictor of steady state drug clearance and thus maintenance doses. One limitation of LBW is that at present there is no validated estimator of LBW in obese children based on total weight and height.
In conclusion, drug doses may indeed need to be adjusted for obese children, but good quality and consistent clinical data to support dosing changes in obese children for the majority of drugs are lacking. Even more worryingly, there appears to be little effort from industry, regulatory authorities or academics to address the lack of robust data for individual drugs. Studies of drugs in development need to include obese subjects to reflect the demography of target populations, so that optimum doses are defined for people of all shapes and sizes in society. For the clinician, the lack of a validated body size descriptor to use at the bedside means that for the majority of drugs the choice of dose will rely on empirical experience and application of the precautionary principle.
Competing interests TNJ is a part-time employee of Simcyp Limited.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.