Why Clearance and Distribution Must Be Considered Separately
Obesity can change body composition, organ blood flow, and the activity or capacity of drug-elimination pathways. These changes do not imply a single, uniform increase in pharmacokinetic parameters. Clearance determines the rate at which drug is eliminated relative to its concentration, while volume of distribution describes the relationship between the amount of drug in the body and its measured concentration. Either parameter may change with body size, but they need not change by the same proportion.
This distinction matters for dosing. Under linear pharmacokinetics, maintenance exposure depends directly on clearance. A loading dose, by contrast, is related primarily to volume of distribution. Thus, a body-size rule that happens to predict a reasonable maintenance dose does not automatically predict a reasonable loading dose, and vice versa.
The discussion below uses simple models to make these relationships explicit. They are ways to compare hypotheses about size effects, not universal descriptions of people with obesity. The appropriate model depends on the drug, the relevant elimination pathways, the population in which the model was developed, and the data available to evaluate it.
Modeling Size Effects Without Assuming a Universal Rule
A common way to represent a size relationship is to scale a parameter relative to a reference body weight \(W_{\mathrm{ref}}\). For clearance, one possible model is
where \(CL_{\mathrm{ref}}\) is clearance at the reference weight and \(a\) is the clearance size exponent. An exponent of zero represents no weight relationship in this model; an exponent of one represents proportional scaling with weight; an exponent between zero and one represents an increase that is less than proportional. The exponent is an estimated or hypothesized relationship, not a biological constant that applies to every drug.
The same general form can be used for volume, with its own exponent \(b\):
There is no requirement that \(a=b\). For example, a drug that distributes mainly in aqueous spaces may not show the same relationship between \(V\) and total body weight as a drug that partitions extensively into adipose tissue. Likewise, obesity-related changes in renal or hepatic elimination do not necessarily track changes in distribution. Earlier material on allometric scaling introduces size-based power models; here the practical question is what separate clearance and volume predictions imply for dosing.
For an oral dose, bioavailability \(F\) is the fraction of the dose reaching the systemic circulation. Under linear, time-invariant pharmacokinetics, the AUC over a dosing interval \(\tau\) at steady state is \(AUC_{\tau}=F D/CL\). For an intravenous dose, \(F=1\). The following comparison assumes the same \(F\) across the size groups unless stated otherwise. If obesity also changes bioavailability, a clearance-only comparison will not capture the full exposure difference.
Why Dose per Kilogram Can Mislead
Suppose a dose is scaled to body weight according to \(D(W)=D_{\mathrm{ref}}(W/W_{\mathrm{ref}})^q\). Fixed dosing has \(q=0\); dosing directly in proportion to weight has \(q=1\). If clearance follows the power model above and bioavailability is unchanged, the ratio of AUCs at weight \(W\) and reference weight \(W_{\mathrm{ref}}\) is
This follows because AUC is proportional to dose divided by clearance: the dose contributes the exponent \(q\), while clearance contributes the exponent \(a\) in the denominator. Therefore, weight-proportional dosing preserves AUC only when \(q=a=1\). If clearance rises less than proportionally with weight, a dose that rises directly with weight predicts higher AUC. If clearance rises more than proportionally, it predicts lower AUC.
This result is conditional on the model and its assumptions; it is not a recommendation to use a particular exponent. It also illustrates why “milligrams per kilogram” is not itself a pharmacokinetic justification. The dose rule encodes an assumed relationship between dose requirement and size. That assumption needs to be compared with the drug’s clearance relationship and the clinical exposure target.
Worked Example: AUC Under Weight-Proportional Dosing
Consider a hypothetical drug given intravenously to a reference patient weighing 70 kg. Clearance is 5.00 L/h at that weight. Compare this patient with a hypothetical 140 kg patient under a model with \(a=0.75\). Assume linear pharmacokinetics and unchanged bioavailability.
The clearance prediction is
If both patients receive 500 mg, their AUCs are \(500/5.00=100.0\ \mathrm{mg\cdot h/L}\) and \(500/8.409=59.46\ \mathrm{mg\cdot h/L}\), respectively. Under fixed dosing, the heavier patient has 0.5946 times the reference AUC.
If the dose is instead doubled to 1,000 mg for the 140 kg patient, the predicted AUC is \(1{,}000/8.409=118.9\ \mathrm{mg\cdot h/L}\). The AUC ratio is \(118.9/100.0=1.189\). Equivalently, the exponent result gives \(2^{1-0.75}=2^{0.25}=1.189\). Thus, weight-proportional dosing predicts about 18.9% higher AUC in this example, not equal AUC. This is a model-based comparison, not evidence that this particular dose is clinically appropriate.
Worked Comparisons of Clearance and Volume
Worked Example: Comparing Clearance Models Against Hypothetical Data
Suppose a hypothetical dataset has a reference observation at 70 kg with clearance 5.00 L/h and an observation at 140 kg with clearance 7.50 L/h. Compare three candidate predictions for the 140 kg observation: no size effect, proportional scaling, and an allometric power model with exponent \(0.75\). Treat the reference clearance as fixed for this arithmetic comparison.
The no-size-effect prediction is \(5.00\ \mathrm{L/h}\). Proportional scaling predicts \(5.00(140/70)^1=10.00\ \mathrm{L/h}\). The allometric prediction is \(5.00(140/70)^{0.75}=8.409\ \mathrm{L/h}\).
Subtracting the observed 7.50 L/h from each prediction gives prediction errors of \(-2.50\), \(+2.50\), and \(+0.909\ \mathrm{L/h}\), respectively. For these two observations, the allometric model is closest at 140 kg. This small comparison does not establish that the allometric model is generally correct: the exponent was specified rather than estimated here, uncertainty is not represented, and performance must be assessed across the relevant population.
A model comparison should consider more than whether a weight term improves fit in the data used to estimate it. Check whether the relationship is supported across the observed size range, whether prediction errors show a remaining trend with weight, and whether uncertainty becomes large at the extremes. A model can fit the average pattern yet predict poorly for an individual. Its application beyond the studied range is especially uncertain.
Worked Example: A Loading Dose Depends on Volume
For an idealized one-compartment intravenous loading dose, assume an immediate target concentration of 10.0 mg/L, complete delivery, and \(C=A/V\). A hypothetical reference patient has \(V_{\mathrm{ref}}=40.0\) L at 70 kg. For a hypothetical 140 kg patient, suppose a drug-specific assessment supports \(V=52.0\) L. Compare the loading dose based on that volume with the dose from proportional scaling of volume with total weight.
The amount required for the target concentration is \(A=C V\). Thus the loading dose under the stated assumptions is \(10.0\ \mathrm{mg/L}\times52.0\ \mathrm{L}=520\ \mathrm{mg}\). Proportional scaling would instead predict \(V=40.0(140/70)=80.0\) L and a loading dose of \(10.0\times80.0=800\ \mathrm{mg}\).
The difference arises from the volume assumption, not from the clearance calculation. A loading dose based on an unsuitable volume prediction could produce an initial concentration above or below the target. In practice, route, infusion duration, distribution kinetics, safety limits, and the definition of the target concentration also matter; the calculation isolates the basic relationship.
Maintenance Exposure, Concentration-Time Shape, and Diagnostics
For a constant intravenous infusion at rate \(R_0\), the steady-state concentration in a one-compartment linear model is \(C_{\mathrm{ss}}=R_0/CL\). For repeated intermittent dosing, \(AUC_{\tau}=F D/CL\) gives average exposure over an interval as \(AUC_{\tau}/\tau\). These relationships show why clearance is central to maintenance dosing. Volume still matters for the time course: in a one-compartment model, the elimination rate constant is \(k=CL/V\), and the half-life is \(t_{1/2}=\ln(2)V/CL\). If clearance and volume change in different proportions, the rate of accumulation and decline may change even when a dose adjustment matches average exposure.
When assessing a size model, inspect the data and predictions rather than relying on a dosing slogan. Useful checks include:
- Plot individual and typical-value clearance and volume estimates against body size, with uncertainty visible where possible.
- Check whether residuals or prediction errors retain a systematic trend with weight or body composition.
- Look for sparse observations or few participants at the highest weights; apparent precision there may be misleading.
- Check whether the model’s predictions remain plausible over the intended dosing range and whether the model has been evaluated in a population similar to the intended patients.
- Consider whether changes in organ function, comedications, or other covariates could be confused with a body-size association.
A size relationship estimated from observational data is not automatically a causal effect of obesity. Weight may be associated with other characteristics that influence clearance, and the data may not distinguish those effects. A plausible mechanism can motivate a model, but it does not replace evaluation of its predictions. If bioavailability, protein binding, or nonlinear elimination also changes, the simple linear exposure comparisons above may not apply as written.
Comparing Approaches and Recognizing Their Limits
| Approach | What it assumes | Potential limitation |
|---|---|---|
| Fixed dose | Dose does not rise with weight. | If clearance rises with size, exposure may fall as weight increases. |
| Dose proportional to weight | Dose requirement rises in direct proportion to weight. | If clearance rises less or more than proportionally, predicted AUC changes with weight. |
| Power model for clearance or volume | A parameter follows a power relationship over the modeled range. | A fitted exponent may be uncertain, drug-specific, or unreliable outside the observed range. |
| Mechanistic or covariate-rich model | Measured physiological or clinical covariates represent relevant processes. | More detail requires adequate data and may not resolve unmeasured or correlated effects. |
No row in this table is a universal dosing recommendation. A model may be useful for predicting population-average exposure but still be inadequate for an individual without concentration data or other relevant covariates. Conversely, a simple model can be preferable if it is better supported and validated for the intended setting. The key is to compare predictions and uncertainty under explicit assumptions, separately for clearance and volume.
Common Pitfall
Wrong reasoning: “The patient weighs twice as much, so both clearance and volume must double; therefore, doubling every dose preserves both exposure and concentration.” This treats body weight as a direct multiplier for two distinct parameters and assumes that their size relationships are identical.
Correction: Specify separate models for \(CL\) and \(V\), then use the relevant relationship for the dosing objective. Under linear pharmacokinetics, maintenance AUC depends on \(D/CL\); an idealized loading dose depends on target concentration times \(V\). Check the resulting predictions against data and acknowledge uncertainty. A correct calculation under a weakly supported model is still a weak prediction.
Check Your Understanding
Answer each question using the stated assumptions. Brief solutions follow the questions.
- A drug has \(CL\propto W^{0.60}\) and is dosed in direct proportion to weight. By what factor does predicted AUC change when weight doubles?
- At a reference weight, \(CL=4.00\) L/h. Under a fixed-clearance model, what is the predicted clearance at twice the reference weight? What does the model imply about AUC under a fixed dose?
- Why can matching average exposure with a maintenance dose fail to guarantee the same initial concentration after a loading dose?
- Name two checks that could reveal poor performance of a size model in people with obesity.
Solutions. (1) The AUC ratio is \(2^{1-0.60}=2^{0.40}=1.320\), so AUC is predicted to increase by about 32.0%. (2) The prediction remains \(4.00\) L/h because the model has no size effect. Under a fixed dose and unchanged bioavailability, AUC is also predicted to remain unchanged. (3) Maintenance exposure depends on clearance, whereas the idealized loading dose for a target concentration depends on volume; different size effects on \(CL\) and \(V\) can therefore produce different concentration-time profiles. (4) Examples include checking residuals for a trend with weight, assessing prediction error at high weights, and verifying that the model has enough observations across the intended size range.