Modeling Renal Maturation of Drug Clearance

Why Renal Maturation Is Not Always GFR Maturation

Renal function changes during development, so a drug cleared by the kidneys may have lower clearance in a young infant than in a mature individual. A useful starting point is glomerular filtration rate (GFR), because filtration contributes to the renal elimination of many drugs. But renal drug clearance is not synonymous with GFR: the amount filtered depends on how much drug is unbound in plasma, and renal secretion and reabsorption can also contribute.

The modeling question is therefore two-part: how does the relevant renal process change with age, and which renal process is relevant for this drug? A GFR-based maturation function can represent filtration development. Treating that same function as a universal model for every drug can misrepresent clearance when secretion, reabsorption, or age-dependent protein binding matters.

For illustration, use a sigmoidal Hill function with postmenstrual age (PMA) \(A\), measured in weeks:

Here \(A_{50}\) is the PMA at which this curve reaches one-half of its mature value, and \(h\) controls its steepness. \(A\) and \(A_{50}\) must use the same units. This is a population representation: an individual's renal function may differ from the curve. The formula also does not, by itself, account for body size or for transitions associated with birth.

From GFR to Filtration Clearance

Under linear filtration, the filtered amount per unit time equals the unbound plasma concentration multiplied by GFR. If \(C_p\) is the total plasma concentration and \(f_u\) is the unbound fraction, the unbound concentration is \(f_u C_p\). Therefore, when the drug is freely filtered and \(f_u\) is constant over the concentration range of interest, filtration clearance is \(f_u\) times GFR.

The units follow directly: a dimensionless \(f_u\) multiplied by GFR in volume per time gives clearance in volume per time. If protein binding changes with age, use \(f_u(A)\), rather than silently assuming a constant unbound fraction. The equation also assumes that the drug is not materially restricted from filtration by size or other properties and that filtration is not saturated.

When Renal Clearance Includes Other Processes

In a simplified linear description, renal excretion can include filtered drug, drug added to the tubule by secretion, and drug removed from the tubule by reabsorption. Expressing each process as an equivalent clearance relative to total plasma concentration gives:

This is a bookkeeping representation for linear processes, not a claim that secretion and reabsorption are always independent or constant. The terms are effective contributions to net renal clearance under the conditions being modeled. If a transporter is saturated, inhibited, or developmentally changing, a single constant secretion term may be inadequate. If reabsorption depends on urine flow or other conditions, a fixed reabsorption term may also be inadequate.

A filtration-only model is most defensible when filtration dominates renal elimination, the unbound fraction is known or approximately stable, and there is no strong evidence that secretion or reabsorption substantially changes the age pattern. For a drug with important non-filtration pathways, the maturation of those pathways is drug-specific; it should not be presumed to equal GFR maturation.

Checking Whether a GFR-Based Model Is Adequate

A filtration-based prediction can be compared with renal clearance estimated from drug concentrations and urinary excretion, when such data are available. In a linear setting, renal clearance is the urinary excretion rate divided by plasma concentration; for data collected over an interval, the corresponding estimate is the amount excreted unchanged divided by the plasma area under the concentration-time curve over that interval. The model and the observation must refer to the same drug, time interval, and clearance scale.

Useful checks include plotting observed renal clearance against the filtration prediction, examining residuals across developmental age, and checking whether the unbound fraction and GFR inputs are expressed on compatible scales. A consistent positive difference may suggest missing net secretion or another model discrepancy; a consistent negative difference may reflect net reabsorption or an overestimated filtration component. A trend in the discrepancy with age warns that a constant non-filtration contribution may also be inadequate.

A common modeling error is to multiply total adult renal clearance by the GFR maturation fraction and call the result a renal maturation model. That assumes every component of adult renal clearance changes in proportion to GFR. It is justified only when GFR-linked filtration is the relevant dominant component, or when proportional maturation of all relevant components has independent support. Otherwise, the model should distinguish filtration from drug-specific processes.

Modeling Choices and Limitations

The main alternatives answer different questions. A filtration-only model is parsimonious and physiologically interpretable when filtration dominates. A whole-clearance empirical maturation curve may describe observed age variation without assigning it to a particular renal mechanism, but it does not identify the process responsible. A pathway model is more mechanistic, but requires sufficient information to support its additional components. These choices are not interchangeable: a good fit from an empirical curve does not demonstrate that GFR alone explains the pattern.

A further practical limitation is scale. GFR may be reported indexed to a standard body surface area, while drug clearance may be an absolute volume per time. Those quantities cannot be substituted directly without a compatible conversion and modeling scale. Keep maturation, size, and reference units explicit; a dimensionless age function does not resolve a mismatch in the underlying units.