Why the Age Scale Matters
A newborn’s age can be described in more than one way. A baby born early may have spent several weeks outside the womb while still being at an earlier developmental stage than a baby born at term. For pharmacometric analysis, this distinction matters: time since birth and developmental age answer different questions. Choosing the wrong age scale can make a maturation covariate misleading even when the arithmetic and model fitting are otherwise correct.
The key quantities are gestational age, postnatal age, and postmenstrual age. Gestational age describes age relative to the obstetric dating origin; postnatal age measures elapsed time since birth; postmenstrual age combines gestational age at birth with time since birth. They are related, but they are not interchangeable.
Gestational age is an obstetric dating convention, not a direct measurement of time since fertilization. The two are often roughly two weeks apart, but that approximation is not a substitute for the recorded gestational age or for a specified dating method. At birth, PNA is zero, so PMA equals GA at birth. During pregnancy, PMA is not needed as a separate postbirth measure: gestational age describes the pregnancy. PNA does not apply before birth.
Converting Between Weeks and Days
One week contains seven days. A notation such as 32 weeks 5 days means \(32+5/7\) weeks; it does not mean 32.5 weeks. When adding or subtracting ages in weeks and days, carry or borrow groups of seven days. Decimal weeks are useful for calculations, but weeks-and-days notation is often easier to read in clinical descriptions.
Worked Example: Calculating Postmenstrual Age
Suppose a hypothetical infant was born at a gestational age of 29 weeks 5 days. A sample is collected 6 weeks 2 days after birth. What is the infant’s PMA at sampling?
The postnatal age is \(6\) weeks \(2\) days. Add it to the gestational age at birth:
The seven days carry to one additional week. Thus, the infant’s PMA at sampling is 36 weeks 0 days. As a decimal-week check, \(29+5/7+6+2/7=36\) weeks.
This conversion follows from elapsed time. If the obstetric dating origin is assigned gestational age \(GA_{\mathrm{birth}}\) at birth, then after \(PNA\) units of elapsed time the age measured from that same origin is \(GA_{\mathrm{birth}}+PNA\). The equation does not claim that all developmental processes progress identically in every infant; it defines how the age scale is calculated.
Worked Example: Recovering Postnatal Age
A hypothetical infant was born at 33 weeks 4 days. At a later assessment, the recorded PMA is 41 weeks 1 day. Calculate the PNA.
Rearrange the age relation: \(PNA=PMA-GA_{\mathrm{birth}}\). Subtract the birth gestational age, borrowing one week (seven days) because one day is less than four days:
The calculated postnatal age is 7 weeks 4 days. Checking the addition confirms the result: 33 weeks 4 days plus 7 weeks 4 days is 40 weeks 8 days, or 41 weeks 1 day.
Selecting an Age Scale for Maturation
The appropriate age variable depends on the biological question and on the period over which the process develops. If the purpose is to describe an interval since birth—for example, time after a dose, duration of hospitalization, or time since a clinical event—PNA is the natural scale. If the purpose is to describe a developmental process that begins before birth and continues after birth, PMA can provide a continuous age scale across the birth boundary. Gestational age is the corresponding age measure before birth.
This is a modeling choice that should be supported by the scientific question and the data. PMA can be a useful developmental-age covariate for a maturation function, but its use does not establish that the process is determined by age alone. Nor does the age conversion demonstrate that two infants at the same PMA have identical physiology. Other sources of variation and uncertainty remain.
Worked Example: Putting the Age Scale Into a Maturation Function
Consider a hypothetical infant born at 28 weeks 0 days and assessed 7 weeks 0 days after birth. The resulting PMA is 35 weeks. Suppose, only to illustrate the consequence of the age input, a previously specified Hill maturation function has \(TM_{50}=40\) weeks and \(h=4\):
If maturation is being indexed by PMA, substitute \(t=35\) weeks. The units match: both \(t\) and \(TM_{50}\) are in weeks.
Equivalently, \(35^4=1{,}500{,}625\), \(40^4=2{,}560{,}000\), and their ratio in the function is \(1{,}500{,}625/4{,}060{,}625=2401/6497\), which rounds to 0.3696. The result is a dimensionless fraction of the function’s maximum under this illustrative model.
If PNA were substituted instead, \(t=7\) weeks and the same calculation would give \(M(7)=2401/(2401+2{,}560{,}000)\approx0.0009366\). That large difference does not show that one calculation is universally correct; it shows that the age definition is consequential. If the intended function describes developmental progression across prenatal and postnatal life, using time since birth would reset the model’s age origin at delivery and could misrepresent that progression. The age scale and the interpretation of \(TM_{50}\) must agree.
Checking Data and Interpreting the Result
Before using an age covariate, check that the source data identify what each age means, its reference date, and its units. A field labeled “age” is not enough: it might refer to gestational age at birth, PNA at sampling, or PMA at sampling. In longitudinal data, the age should correspond to each observation’s time point, not simply to a single value attached to the infant.
Useful checks include:
- Verify the reference event. Confirm that PNA is measured from birth and that the gestational-age value is the value at birth when calculating PMA.
- Check units and precision. Convert days consistently; do not add a value in days directly to a value in weeks without conversion.
- Check the identity. After calculating PMA, subtract the recorded PNA. The result should recover the gestational age at birth, subject to the precision of the recorded dates.
- Look for implausible combinations. A negative computed PNA after birth or a PMA lower than the recorded GA at birth indicates a likely date, unit, or transcription problem.
- Account for dating uncertainty. Recorded gestational age may be rounded or based on an estimate. Derived PMA inherits that uncertainty; arithmetic cannot make the original age estimate more precise.
When ages are recorded only to completed weeks, an apparent discrepancy of several days may result from rounding rather than a true contradiction. Use the data’s stated precision, document any conversion convention, and avoid reporting calculated ages to more precision than the inputs support.
Common Mistakes and Limitations
| Mistake | Why it is wrong | Correction |
|---|---|---|
| Treating PNA and PMA as synonyms | They have different origins and can differ substantially, especially after preterm birth. | State which age is being used and calculate PMA from GA at birth plus PNA when appropriate. |
| Using age since conception as the recorded GA | Obstetric gestational age is conventionally dated from the last menstrual period, not fertilization. | Use the defined clinical dating convention and do not apply a rough two-week adjustment as if it were exact. |
| Entering PNA into a developmental-age function by default | That makes birth the age origin, even if the intended developmental process spans prenatal and postnatal life. | Choose the scale to match the process being represented; justify that choice scientifically. |
| Adding decimal-looking week notation | For example, 29 weeks 5 days is not 29.5 weeks. | Convert days using \(d/7\), or add in weeks and days with carries of seven. |
No single age scale is automatically best for every model. A developmental process may not follow PMA alone, and the recorded GA can be uncertain. The choice should be explicit, consistent with the modeled biology, and checked against the observation schedule and data definitions. The earlier material on maturation functions supplies the function form; this age distinction determines what quantity should be supplied as its age input.
Check Your Understanding
Answer each question, then compare with the solutions below.
- An infant is born at 31 weeks 6 days and assessed 4 weeks 3 days later. What is the PMA?
- An infant’s GA at birth is 35 weeks 2 days and the assessment PMA is 40 weeks 0 days. What is the PNA?
- Two hypothetical infants are assessed at the same PMA. One was born at 28 weeks and the other at 34 weeks. Which has had more time since birth, and by how much?
- For a model intended to represent maturation continuing across birth, explain why PNA and PMA are not interchangeable inputs.
- A record gives GA at birth as 30 weeks 4 days and PNA as 2 weeks 5 days. State the PMA and one data-quality check for the result.
Solutions. (1) \(31\) weeks \(6\) days plus \(4\) weeks \(3\) days is \(35\) weeks \(9\) days, or 36 weeks 2 days. (2) Borrow one week: \(40\) weeks \(0\) days is \(39\) weeks \(7\) days. Subtract \(35\) weeks \(2\) days to obtain 4 weeks 5 days. (3) The infant born at 28 weeks has had 6 weeks more time since birth, because the difference between birth gestational ages is \(34-28=6\) weeks while their PMA is equal. (4) PNA starts at birth and resets the age origin there; PMA adds time since birth to gestational age at birth and therefore represents a continuous age scale across birth. Use the scale aligned with the developmental process, not merely the easiest available variable. (5) \(30\) weeks \(4\) days plus \(2\) weeks \(5\) days is \(32\) weeks \(9\) days, or 33 weeks 2 days. Check that subtracting the recorded PNA from PMA recovers 30 weeks 4 days and confirm that both inputs refer to the same infant and assessment date.