Why Age and Dose Must Be Considered Together
A child with lower drug clearance may have higher exposure than an older child given the same dose. That pattern can reflect developmental maturation of elimination pathways. But high exposure can also arise because elimination is saturable: as concentration rises, elimination no longer increases in direct proportion to concentration. These mechanisms can look similar in a single comparison, yet imply different exposure changes when the dose changes.
The distinction matters for pediatric dose selection and extrapolation. A model that attributes saturation to immature clearance may mispredict exposure at a higher dose; a model that attributes maturation to saturation may mispredict exposure in a younger age group. The key evidence is not age-related exposure alone, but whether the data support a developmental change in clearance, a concentration-dependent change in elimination, or both.
The earlier material on body-size scaling, maturation functions, and identifiability of size and maturation effects provides the basis for describing age-related clearance. Here, the emphasis is on how to contrast that developmental mechanism with nonlinear, saturable elimination.
Competing Elimination Models
Begin with a one-compartment model in which \(C\) is plasma concentration, \(V\) is the apparent volume of distribution, and drug enters the compartment at a known rate \(R_{\mathrm{in}}(t)\). For an intravenous input, a linear elimination model is
Here, \(CL(a)\) is clearance at age \(a\). In a developmental model, clearance may depend on body size and an age-related maturation function, for example \(CL(a)=CL_{\mathrm{ref}}(WT/WT_{\mathrm{ref}})^{0.75}M(a)\), where \(M(a)\) is the maturation factor relative to the reference. Under linear elimination, clearance does not depend on concentration: doubling concentration doubles the elimination rate.
A saturable pathway is often represented by Michaelis–Menten elimination. Its elimination rate approaches a maximum capacity \(V_{\max}\) as concentration increases, and \(K_m\) is the concentration at which that pathway’s elimination rate is half its maximum.
The age dependence of \(V_{\max}\), \(K_m\), or both can be considered if supported by the data and biological rationale. It should not be assumed that all saturable parameters are age-invariant. A combined model allows a linear pathway and a saturable pathway to contribute simultaneously:
These equations describe elimination rate in amount per time before division by \(V\). The linear component is \(CL(a)C\); the saturable component is \(V_{\max}(a)C/(K_m+C)\). The models assume a well-mixed compartment, appropriately specified drug input, and parameters that are constant over the modeled interval. They are simplifications: distribution phases, absorption, time-varying physiology, or unmodeled elimination routes can make concentration data depart from these predictions.
Why Saturation Produces Dose-Dependent Exposure
At low concentrations relative to \(K_m\), \(K_m+C\) is approximately \(K_m\), so the saturable pathway behaves nearly linearly:
At concentrations much greater than \(K_m\), the same pathway approaches a constant elimination rate:
Thus, dose proportionality is expected under linear PK with unchanged bioavailability and parameters, but can fail under saturable elimination. In the latter case, a proportional increase in input may cause a more-than-proportional increase in concentration. This pattern alone does not prove saturation: age, body size, dose assignment, adherence, input rate, and other model errors can also create apparent dose dependence.
Worked Example: Age-Related Clearance With Dose-Proportional Exposure
Suppose a hypothetical linear model has already accounted for body size and estimates clearance of 2.00 L/h in a younger age group and 3.60 L/h in an older age group. Assume an intravenous dose, complete observation of the concentration-time profile, and linear elimination. What are the expected AUCs after 20 mg and 40 mg?
For intravenous dosing under linear elimination, \(AUC=D/CL\). At 20 mg, the younger group has
The older group has \(20/3.60=5.56\ \mathrm{mg\cdot h/L}\). At 40 mg, the corresponding AUCs are \(40/2.00=20.0\ \mathrm{mg\cdot h/L}\) and \(40/3.60=11.1\ \mathrm{mg\cdot h/L}\). Within each age group, doubling the dose doubles AUC. The difference between age groups is consistent with a clearance difference, but this comparison by itself does not establish the cause of that difference or rule out saturation at concentrations not observed.
Worked Example: Detecting Dose Dependence Under Saturable Elimination
Consider a hypothetical one-compartment model with Michaelis–Menten elimination only. The younger group has \(V_{\max}=12\) mg/h and \(K_m=2.00\) mg/L; the older group has \(V_{\max}=24\) mg/h and the same \(K_m\). Each group receives constant intravenous infusions at 6 and 9 mg/h. Assume steady state is reached, the parameters do not change during infusion, and the infusion rate is below \(V_{\max}\).
At steady state, input equals elimination:
For the younger group at 6 mg/h, \(C_{\mathrm{ss}}=(6)(2)/(12-6)=2.00\) mg/L. At 9 mg/h, it is \((9)(2)/(12-9)=6.00\) mg/L. Increasing the infusion rate by a factor of \(9/6=1.50\) increases predicted concentration threefold.
For the older group, the same calculations give \((6)(2)/(24-6)=0.667\) mg/L and \((9)(2)/(24-9)=1.20\) mg/L. The concentration ratio is \(1.20/0.667=1.80\), greater than the input-rate ratio of 1.50. Both groups show dose-dependent departure from proportionality, but the size of that departure differs because their capacities differ. The formula also makes a boundary clear: for this model, no finite steady-state concentration exists when \(R_{\mathrm{in}}\geq V_{\max}\).
Combining Mechanisms and Comparing Predictions
A real drug can have both a concentration-independent elimination route and a saturable route. At steady state under constant intravenous infusion, the combined model satisfies
This equation can be solved numerically in general. Its two elimination terms are useful for interpretation: one grows linearly with concentration, while the other approaches a maximum. A fitted combined model is not automatically preferable; extra parameters can fit noise, and the data must contain enough information to estimate them.
Worked Example: Separating Linear and Saturable Elimination
At one hypothetical age, suppose \(CL=1.00\) L/h, \(V_{\max}=8.00\) mg/h, and \(K_m=2.00\) mg/L. Find the steady-state concentration at infusion rates of 6 and 10 mg/h.
For 6 mg/h, the balance is \(6=C+8C/(2+C)\), with units of mg/h on each term. Multiplying by \(2+C\) gives \(12+6C=C^2+10C\), or \(C^2+4C-12=0\). The positive root is \(C=2.00\) mg/L; the negative root is not a physically meaningful concentration. At that concentration, the linear route eliminates \((1.00)(2.00)=2.00\) mg/h, and the saturable route eliminates \(8(2)/(2+2)=4.00\) mg/h. Together they balance the 6 mg/h input.
For 10 mg/h, \(10=C+8C/(2+C)\). Multiplication gives \(20+10C=C^2+10C\), so \(C^2=20\), and the positive solution is \(C=4.47\) mg/L. The linear pathway then eliminates 4.47 mg/h, and the saturable pathway eliminates \(8(4.47)/(2+4.47)=5.53\) mg/h, totaling 10.0 mg/h after rounding. Increasing input by a factor of \(10/6=1.67\) increases concentration by \(4.47/2.00=2.24\). This is more than proportional because the saturable contribution is approaching its 8 mg/h maximum.
Using Pediatric Data to Evaluate Mechanisms
A useful comparison requires dose and age information that are not completely confounded. If younger children receive only low doses and older children only high doses, a dose trend can be mistaken for maturation, or vice versa. Where ethically and practically appropriate, observations spanning dose levels within age ranges, together with overlap in dose or concentration across ages, help distinguish the explanations. Known input and reliable dose records are also important; for oral data, age-related absorption or bioavailability changes can mimic changes in elimination.
The model comparison should reflect the question. A linear developmental model asks whether clearance changes with age after accounting for size and relevant covariates. A saturable model asks whether concentration-dependent elimination improves the description across the observed concentration range. A combined model asks whether both are needed. Compare predicted concentration-time profiles and exposure across dose levels and ages, not only a single overall fit statistic. A model may fit the observed data yet give sharply different predictions outside the tested doses or age range.
| Evidence pattern | Consistent interpretation | Important qualification |
|---|---|---|
| Exposure differs by age, but dose proportionality holds within age groups | Developmental clearance differences may explain the age pattern | Does not rule out saturation outside the observed concentration range |
| Within-age exposure rises more than proportionally with dose | Saturation may contribute | Check dose input, adherence, absorption, and other systematic differences |
| Both age trends and dose nonproportionality remain | A combined mechanism may be needed | Parameters may be weakly identified without age and concentration coverage |
A practical warning sign is a mechanism estimate that depends strongly on a few high-concentration observations or on fixing a parameter to an assumed value. Another is a model that predicts implausible concentrations, or cannot reproduce the observed dose trend, in a group excluded from fitting. These are reasons to revisit assumptions and data coverage; they do not by themselves identify the correct alternative. Formal age-stratified diagnostic methods are a separate step.
Common Pitfall and Limitations
Wrong reasoning: “Younger children have higher concentrations at the same dose, so elimination must be saturated.” This does not follow. A lower age-related clearance can produce higher exposure under a fully linear model. Conversely, an apparent age-related clearance difference can arise when age groups receive different doses and one group enters a saturable concentration range.
Correction: Examine how exposure changes with dose within age ranges, and how concentration-time behavior changes across the observed concentration range. Evaluate size and maturation in the clearance model, then assess whether a saturable term is supported in addition to those effects. The earlier material on identifiability of size and maturation effects is particularly relevant: if age, size, and dose vary together, the available data may not uniquely distinguish their contributions.
Important limitations remain even with a good model comparison. A narrow concentration range may make a saturable pathway look linear, because at \(C\ll K_m\) its rate is approximately \((V_{\max}/K_m)C\). A narrow age range may provide little evidence about maturation. Sparse or noisy observations can obscure concentration-dependent elimination, while misspecified absorption or distribution can be misattributed to clearance. Finally, statistical fit does not establish a biological mechanism. Conclusions should state the observed range, assumptions, and uncertainty, and avoid extrapolating far beyond the ages and doses informing the model.
Check Your Understanding
Answer each question, then compare with the solutions.
- In a linear intravenous model with unchanged clearance, what happens to AUC when dose doubles?
- For a Michaelis–Menten pathway, what does the elimination rate approach when \(C\) is much greater than \(K_m\)?
- Why can an age-only comparison not establish that a concentration difference is caused by saturation?
- For a saturable-only pathway with \(V_{\max}=10\) mg/h, can a constant infusion of 10 mg/h achieve a finite steady state under the stated model?
Solutions: (1) AUC doubles because \(AUC=D/CL\) when clearance is unchanged and elimination is linear. (2) The rate approaches \(V_{\max}\), since \(C/(K_m+C)\) approaches 1. (3) Age-related clearance, body size, dose assignment, or other age-linked factors can also explain the difference; dose-dependent evidence and an appropriate model comparison are needed. (4) No. The steady-state expression \(C_{\mathrm{ss}}=R_{\mathrm{in}}K_m/(V_{\max}-R_{\mathrm{in}})\) has a zero denominator at 10 mg/h; elimination approaches but does not exceed the 10 mg/h capacity at finite concentration.