Prematurity and Gestational-Age Effects on Drug-Metabolizing Capacity

Why Age at Birth Matters to Capacity Predictions

A preterm and a term infant can have the same postnatal age while having very different developmental histories. For example, two infants observed 12 weeks after birth might have been born at 30 and 40 weeks of gestational age. Their postmenstrual ages are therefore 42 and 52 weeks, respectively. A model that uses only time since birth assigns them the same age; a model that also accounts for gestational age at birth does not.

For drug-metabolizing enzymes, this distinction matters because capacity changes during development. A PBPK model can represent an infant’s enzyme capacity as a size-scaled adult reference capacity multiplied by an enzyme-specific maturation fraction. The comparison below focuses on the age-dependent part of that representation. It does not assume that every preterm infant has lower capacity than every term infant, or that the difference in predicted capacity is caused by prematurity itself.

Definition: Gestational age at birth, $G_b$, is the completed gestational age when the infant is born. Postnatal age, $t$, is the time elapsed since birth. Postmenstrual age, $\mathrm{PMA}$, is the sum of gestational age at birth and postnatal age: $$ \mathrm{PMA}=G_b+t. $$ In this tutorial, ages are expressed in weeks. Preterm birth is before 37 completed weeks of gestation; a term birth is at or after 37 completed weeks.

Representing Enzyme Maturation

Let $M_e(\mathrm{PMA})$ denote the maturation fraction for enzyme $e$, relative to its mature reference activity. It is dimensionless and is usually represented between 0 and 1: a value of 0.40 means 40% of the reference activity under the model’s assumptions. The appropriate maturation function and its parameters are enzyme-specific; there is no single age curve that should be applied to every enzyme.

An illustrative Hill function is:

M_e(\mathrm{PMA})= \frac{\mathrm{PMA}^{h_e}} {K_e^{h_e}+\mathrm{PMA}^{h_e}}.

Here $K_e$ is the postmenstrual age at which the function reaches one-half of its asymptote, and $h_e$ controls the steepness of the transition. This form is bounded between 0 and 1 for nonnegative age and positive $K_e$ and $h_e$. It is an illustrative parameterization, not a claim that a particular enzyme follows this curve or that its parameters are known without data.

For an infant $i$, write a simplified capacity representation as:

C_{e,i}=C_{e,\mathrm{adult}}\,s_i\,M_e(\mathrm{PMA}_i).

$C_{e,\mathrm{adult}}$ is the mature reference capacity, in the chosen capacity units, and $s_i$ is a dimensionless size-scaling factor. The product $s_iM_e$ gives capacity relative to the adult reference under this representation. If the goal is to isolate maturation, hold $s_i$ fixed across the comparison; if infant sizes differ, size scaling must be handled separately. A matched postnatal age does not imply matched size.

The basic PMA-only assumption is that, for a given enzyme, two infants at the same PMA have the same maturation fraction, all else equal. The model does not independently represent whether development occurred in utero or after birth. That is a substantive simplification, not an automatic biological fact.

Worked Example: Converting Birth and Postnatal Ages

Consider two hypothetical infants observed 12 weeks after birth. Infant A was born at 30 weeks of gestational age; Infant B was born at 40 weeks. Calculate their PMAs and explain what a PMA-based model distinguishes.

For Infant A: $$ \mathrm{PMA}_A=G_{b,A}+t_A=30+12=42\ \text{weeks}. $$ For Infant B: $$ \mathrm{PMA}_B=G_{b,B}+t_B=40+12=52\ \text{weeks}. $$

Both infants have a postnatal age of 12 weeks, but their PMAs differ by $52-42=10$ weeks. A PMA-based maturation function therefore evaluates them at 42 and 52 weeks, rather than assigning both the same maturation fraction. This calculation establishes the age inputs; it does not by itself establish the enzyme capacities, which also depend on the enzyme’s maturation function and, for absolute capacity, size scaling.

Matched-Postnatal-Age Capacity Comparison

The following worked comparison uses hypothetical parameters to show the calculation. Suppose the enzyme-specific function is $M_e(\mathrm{PMA})=\mathrm{PMA}^3/(48^3+\mathrm{PMA}^3)$, with age in weeks. The example assumes the same size-scaling factor, $s_i=0.060$, for both infants so that the contrast isolates the modeled maturation difference. The factor 0.060 is not a general value for infants.

Worked Example: Comparing Capacity at 12 Weeks After Birth

Use the two infants from the previous example. Infant A, born at 30 weeks, is at 42 weeks PMA; Infant B, born at 40 weeks, is at 52 weeks PMA. Calculate each maturation fraction and relative total capacity, $s_iM_e$.

For Infant A, $42^3=74{,}088$ and $48^3=110{,}592$. Thus: $$ M_e(42)=\frac{74{,}088}{110{,}592+74{,}088} =\frac{74{,}088}{184{,}680} \approx 0.4012. $$ The relative total capacity is: $$ s_A M_e(42)=0.060\times 0.4011696\approx 0.02407. $$

For Infant B, $52^3=140{,}608$, so: $$ M_e(52)=\frac{140{,}608}{110{,}592+140{,}608} =\frac{140{,}608}{251{,}200} \approx 0.5597452. $$ Its relative total capacity is: $$ s_B M_e(52)=0.060\times 0.5597452\approx 0.0335847. $$

Under these assumptions, the modeled maturation fraction at 42 weeks PMA is about 0.4012, compared with 0.5597452 at 52 weeks PMA. With the same size factor, Infant A’s relative capacity is about $0.02407/0.0335847\approx0.71670$ times Infant B’s, or about 28.33% lower relative to Infant B’s predicted value. These are outputs of the illustrative function and assumptions, not clinical estimates for a real enzyme or a treatment recommendation.

What the Comparison Assumes

The numerical contrast rests on several assumptions. First, gestational age and postnatal age are measured consistently, so their sum is a suitable PMA input. Second, the selected maturation function and parameters are appropriate for the enzyme and population being modeled. Third, the PMA-only formulation is adequate: conditional on PMA, it does not need a separate effect for birth status or extrauterine exposure. Fourth, the example holds size scaling equal. If infants have different organ size or other size-related capacity factors, the difference in $s_i$ also affects total capacity.

A short derivation clarifies what a matched-postnatal-age contrast represents. For infants A and B observed at the same postnatal age $t$, their PMA difference is:

\mathrm{PMA}_B-\mathrm{PMA}_A =(G_{b,B}+t)-(G_{b,A}+t) =G_{b,B}-G_{b,A}.

The common postnatal age cancels. Under the PMA-only model with equal size factors, the capacity ratio is therefore:

\frac{C_{e,A}}{C_{e,B}} = \frac{M_e(G_{b,A}+t)}{M_e(G_{b,B}+t)}.

This ratio compares the model’s predicted maturation at two PMAs. It does not separately estimate an effect of being born preterm, because gestational age at birth enters only through PMA. If a model includes separate gestational and postnatal age effects, those effects need an explicit mathematical form and suitable data to estimate them; they cannot be inferred merely by relabeling a PMA comparison.

Worked Example: Same PMA, Different Age Histories

Compare two hypothetical infants at 42 weeks PMA. Infant C was born at 30 weeks and is 12 weeks postnatal; Infant D was born at 34 weeks and is 8 weeks postnatal. Use the same Hill function and assume equal size factors.

For Infant C, $30+12=42$ weeks PMA. For Infant D, $34+8=42$ weeks PMA. Therefore the PMA-only model gives both the same maturation fraction: $$ M_e(42)=\frac{42^3}{48^3+42^3} =\frac{74{,}088}{184{,}680} \approx 0.4012. $$ With $s_C=s_D=0.060$, each has relative total capacity $0.060\times0.4011696\approx0.02407$.

The prediction is equal because the model uses PMA and equal size factors, not because the infants have identical histories. If observations at the same PMA consistently differed by birth history after accounting for relevant factors, that pattern would challenge the PMA-only assumption and motivate assessment of a richer model.

Diagnostics, Alternatives, and Limitations

A PMA-only curve is a useful parsimonious starting point, but its adequacy should be evaluated rather than presumed. When suitable enzyme activity or clearance data are available, examine whether residuals show a systematic pattern by gestational age at birth or by postnatal age after accounting for PMA and relevant size factors. A persistent pattern may indicate a missing birth-history effect, an inadequate maturation curve, or confounding among age, size, illness, and treatment. Sparse observations at very early ages can make these explanations difficult to distinguish.

Parameter behavior also matters. A fitted $K_e$ or $h_e$ may be poorly determined if observations do not span the period of rapid maturation or the mature plateau. Predictions far outside the ages represented in the data are then extrapolations. Check that predicted fractions remain biologically interpretable and that fitted values are consistent with the data and model bounds. A Hill curve’s mathematical bounds do not ensure that its shape is biologically appropriate.

The principal alternative to PMA-only maturation is a model that separately represents gestational age at birth and time since birth, or includes a specified postnatal-environment or birth-history component. Such a model can express departures from a shared PMA trajectory, but it has more parameters and requires data that can distinguish those effects. Another practical alternative is stratified analysis or separate curves for selected birth-history groups; this can reveal heterogeneity but may be unstable with small samples and does not automatically identify a mechanism.

These approaches answer different questions. A PMA-only model gives a compact developmental clock. A model with separate terms asks whether infants with the same PMA differ according to how that age was reached or the conditions after birth. Neither choice is universally correct. Extrauterine physiology, clinical care, illness, and enzyme-specific biology can all matter, but a matched-age comparison alone cannot isolate their contributions. In particular, a predicted capacity difference is not evidence that prematurity caused the difference.

Common pitfall: “Both infants are 12 weeks old after birth, so they have the same enzyme maturation.” This reasoning ignores gestational age at birth. The correction is to calculate $\mathrm{PMA}=G_b+t$ for each infant, then apply the enzyme-specific maturation function. The reverse mistake is to assume that any difference predicted by PMA proves a causal effect of prematurity; the model has only assigned different PMAs, and its assumptions and data support must still be assessed.

Key takeaway: Matched postnatal age is not matched developmental age. In a PMA-based model, birth gestational age and postnatal age jointly determine the maturation input; size scaling and possible birth-history effects remain separate modeling questions.

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

An infant is born at 29 weeks and is observed 15 weeks later. What is the infant’s PMA?
Two infants have equal postnatal age and equal size factors, but one was born at a later gestational age. Under a strictly increasing PMA-only maturation function, which infant has the larger predicted maturation fraction?
For $M_e(42)=0.4011696$ and $s=0.060$, calculate relative total capacity to five significant figures.
Two infants have the same PMA but different gestational ages at birth. What does a PMA-only model predict about their maturation fractions, and what does that prediction not establish?
Name one diagnostic pattern that could suggest a PMA-only model is inadequate.

Solutions. (1) $\mathrm{PMA}=29+15=44$ weeks. (2) The infant born later has the larger PMA at the shared postnatal age; under a strictly increasing function, that infant has the larger predicted maturation fraction. (3) $0.060\times0.4011696=0.024070176$, which rounds to 0.024070. (4) The model predicts equal maturation fractions because their PMAs match. This does not establish that their actual enzyme activities are equal or that birth history has no effect. (5) For example, residuals that consistently differ by gestational age at birth among infants with similar PMA may indicate a missing birth-history effect or other model inadequacy.