Why combine body size and renal function?
A maintenance regimen is intended to replace drug eliminated over time while keeping exposure near a therapeutic target. In a patient with obesity and impaired renal function, both body size and renal function may contribute to clearance. If the model represents either covariate poorly, a dose calculated from it can miss the target even when the arithmetic is correct. The calculation below makes each modeling choice explicit and checks the resulting regimen against the exposure target.
The worked calculations use a wholly hypothetical drug and a hypothetical clearance model. They illustrate a model-informed calculation, not a dosing recommendation. A clinical regimen would require a drug-specific model, an appropriate target, validated covariates, patient-specific clinical judgment, and monitoring.
State the model before calculating
For the examples, assume a validated model for this hypothetical drug predicts total clearance as
Here \(CL_{\mathrm{ref}}=8.0\ \mathrm{L/h}\) is clearance at the reference descriptor weight of \(60\ \mathrm{kg}\) and reference GFR of \(100\ \mathrm{mL/min}\); \(W_{\mathrm{desc}}\) is the chosen body-size descriptor; and \(f_R=0.75\) is the model’s renal fraction at the reference condition. The exponent 0.75 and the form of the renal adjustment are assumed to have been estimated and validated for this drug. They are not universal pharmacokinetic rules. The bracket is dimensionless, so the predicted clearance retains units of \(\mathrm{L/h}\).
This model assumes that the descriptor and renal covariates combine multiplicatively in the stated way, that clearance is stable over the dosing interval, and that the drug’s exposure is linear in dose over the range considered. It also assumes that the GFR supplied to the model is an absolute value in \(\mathrm{mL/min}\), not a value indexed to body surface area. Changing the renal measure, covariate form, or descriptor would require using a model validated for that choice.
For the hypothetical patient below, use adjusted body weight as the descriptor. This is a choice stipulated by the drug model for this example, not a universal rule for obesity dosing. The earlier material on choosing body-weight descriptors addresses why descriptor selection is drug- and model-dependent.
Worked clearance and maintenance dose
Worked Example: Combining Adjusted Body Weight and Renal Function
Consider a fictional 58-year-old woman who is 168.0 cm tall and weighs 106.5585 kg. Suppose the model specifies adjusted body weight with a 0.4 adjustment factor. For this hypothetical calculation, the ideal body weight (IBW) convention gives 59.622 kg at this height. Then
Suppose the patient’s absolute GFR entered into the model is \(49.14\ \mathrm{mL/min}\). The renal multiplier is
The body-size multiplier is calculated using the unrounded descriptor ratio and exponent:
Therefore, retaining the unrounded multiplier in the clearance calculation,
The predicted clearance is about \(6.047\ \mathrm{L/h}\) to three decimal places. Keeping the additional digits in intermediate calculations avoids a rounding discrepancy in subsequent dose calculations.
Now suppose the hypothetical target is an average steady-state concentration of \(8.0\ \mathrm{mg/L}\), the dosing interval is \(\tau=12\ \mathrm{h}\), and oral bioavailability is \(F=0.80\). The target AUC over one interval is
The dose that exactly gives this target under the model is
The units reduce to milligrams. If the hypothetical formulation permits doses in 25 mg increments, choose 725 mg every 12 hours for the exposure check. Its predicted steady-state AUC per interval is
Thus the predicted average concentration is \(95.91/12\approx7.992\ \mathrm{mg/L}\), close to the \(8.0\ \mathrm{mg/L}\) target. This is a model-based average-exposure check, not evidence that the patient will attain that concentration or that every concentration in the interval is suitable.
Why the descriptor and renal term both matter
In the stated model, the size multiplier and renal multiplier have distinct roles. The descriptor determines the size scaling, while GFR changes the renal adjustment. A descriptor is not interchangeable with actual body weight simply because actual weight is readily available. The appropriate descriptor is the one supported by the model and its intended population. Likewise, the renal covariate must be supplied in the form used to develop the model.
The following comparison holds GFR fixed at \(49.14\ \mathrm{mL/min}\) and changes only the weight descriptor. It illustrates the consequences of changing an input; it does not establish that any alternative descriptor is clinically preferable.
| Descriptor in the same model | Approximate size multiplier | Approximate predicted CL |
|---|---|---|
| IBW, 59.622 kg | 0.9953 | 4.925 L/h |
| Adjusted body weight, 78.3966 kg | 1.2221 | 6.047 L/h |
| Actual body weight, 106.5585 kg | 1.5384 | 7.613 L/h |
Because the target dose is proportional to predicted clearance in this linear calculation, the alternative descriptors would also produce different calculated doses. Choosing whichever descriptor happens to generate a desired dose would reverse the reasoning: model structure and evidence should determine the descriptor, and the exposure target should then be checked against the resulting prediction.
Worked Example: Recalculating After a Lower GFR
Suppose the same patient’s absolute GFR later falls to \(30.0\ \mathrm{mL/min}\). Hold the other model inputs fixed. The renal multiplier becomes
Using the same body-size multiplier, predicted clearance is
For the same target \(8.0\ \mathrm{mg/L}\), 12-hour interval, and \(F=0.80\), the model-based dose is
If the hypothetical available increments are 25 mg, 550 mg every 12 hours is the nearest increment. Its predicted average concentration is
The prediction is slightly below the target, as expected from rounding the calculated dose downward. This comparison assumes the model remains valid at the new renal function and that the change in GFR is represented appropriately. A real change in renal function may also require reassessing the input measurement, treatment context, and monitoring plan.
Calculation workflow and diagnostics
Check the specified body-size descriptor, renal-function measure and units, model reference values, and whether the patient is within the population in which the model was evaluated.
Show the descriptor calculation, each covariate multiplier, and the clearance units. Carry unrounded values through the arithmetic.
For a linear model, use \(D=C_{\mathrm{target}}CL\tau/F\), checking that the target refers to average concentration or AUC over the same interval.
After rounding to an available dose, recalculate the predicted AUC or average concentration. Consider whether other exposure measures or constraints also matter.
Diagnostics begin with input and unit checks. A GFR in \(\mathrm{mL/min/1.73\,m^2}\) cannot be substituted as though it were an absolute GFR in \(\mathrm{mL/min}\) unless it has been converted as required by the model. A weight in pounds cannot be used in a formula expecting kilograms. A dose calculation with units that do not reduce to mass per administration signals an algebra or unit error.
Next, compare each covariate with the model’s supported range and check whether the predicted clearance is plausible for the drug and patient. A model can return a precise number outside its development range, but numerical precision does not establish predictive reliability. Large changes in predicted clearance under small, clinically plausible changes in GFR or descriptor weight should prompt sensitivity analysis and review of the model form.
The main alternative is a different validated covariate model, perhaps one using another body-size descriptor or renal-function relationship. Such alternatives should be compared using their supporting data, predictive performance, and applicability to the patient—not by selecting the result that best matches the desired target. Therapeutic drug monitoring, when appropriate for the drug, can add observed concentration information and support subsequent individualized adjustment. It does not remove the need to interpret timing, adherence, assay, and model assumptions correctly.
Assumptions, limitations, and a common mistake
The dose equation assumes linear, dose-proportional pharmacokinetics over the relevant range, stable clearance, known bioavailability, and a target defined for the same dosing interval. It predicts average exposure only. If absorption is delayed, the drug has saturable elimination, renal function is changing, or the model omits important covariates, the calculated average may not adequately describe the clinically relevant exposure. A model for total clearance also does not by itself determine a loading dose, peak concentration, or trough concentration.
A common mistake is to put actual body weight into a model that specifies adjusted body weight, then interpret the resulting higher predicted clearance as proof that a larger dose is needed. The reasoning is wrong because the model’s descriptor is part of its definition, not a freely interchangeable input. The correction is to use the descriptor validated for that model, calculate the dose from the resulting clearance, and assess uncertainty and exposure with suitable clinical information. The introductory material on combined renal, hepatic, and body-size effects is relevant when multiple covariates or interactions must be represented; this worked calculation assumes the particular multiplicative form stated above.
Check Your Understanding
Use the hypothetical model and assumptions stated in this tutorial. Each problem includes a brief solution so that the arithmetic and interpretation can be checked.
- If \(W_{\mathrm{desc}}=60\ \mathrm{kg}\) and \(GFR=100\ \mathrm{mL/min}\), what clearance does the model predict? Solution: Both multipliers equal 1, so \(CL=8.0\ \mathrm{L/h}\).
- For the first worked example, what target AUC per 12-hour interval corresponds to \(8.0\ \mathrm{mg/L}\) average concentration? Solution: \(8.0\ \mathrm{mg/L}\times12\ \mathrm{h}=96\ \mathrm{mg\cdot h/L}\).
- Why is \(95.91\ \mathrm{mg\cdot h/L}\) the relevant check for the rounded 725 mg regimen rather than the unrounded calculated dose? Solution: The regimen to be used in the prediction is 725 mg, so \(AUC_{\tau,\mathrm{ss}}=FD/CL=(0.80)(725)/6.04748\approx95.91\ \mathrm{mg\cdot h/L}\). The unrounded dose would give the target by construction.
- Can the model’s predicted average concentration alone establish that peak concentrations are safe? Solution: No. It predicts average exposure over the interval; peak assessment requires an appropriate concentration-time model and relevant inputs.
- What should be checked before entering a reported GFR into this model? Solution: Confirm that it is expressed as absolute \(\mathrm{mL/min}\), or convert it appropriately if reported indexed to body surface area, and confirm that the resulting value is within the model’s supported range.