Page 165 · Estimating maternal-to-infant drug transfer via the milk/plasma (M/P) ratio and relative infant dose
When a breastfeeding mother takes a medication, a fraction of the circulating drug equilibrates into breast milk. The milk-to-plasma (M/P) ratio — the drug concentration in milk divided by the concentration in maternal plasma at the same timepoint — is the most widely reported single number summarizing this transfer. It is not a measure of infant risk by itself, but it is the foundation on which more clinically meaningful estimates, such as the relative infant dose, are built.
The M/P ratio is a concentration ratio, not a dose. An M/P ratio of 2 does not mean the infant receives twice the maternal dose — it means the drug is roughly twice as concentrated in milk as in plasma, at whatever concentration the mother's plasma happens to be. A drug with a high M/P ratio but very low maternal plasma levels (because of a small maternal dose or extensive first-pass metabolism) can still deliver only a trivial dose to the infant.
Conversely, a drug with a "reassuring" low M/P ratio can still pose meaningful exposure if maternal plasma concentrations are very high (large maternal dose, poor maternal clearance) or if milk intake volume is large relative to infant body weight (as with an exclusively breastfed neonate).
For this reason, the M/P ratio should always be interpreted alongside maternal dose, milk volume, and the resulting estimated and relative infant dose — never in isolation.
Two general approaches are used to determine M/P ratio in lactation pharmacokinetic studies:
• Single time-point ratio: milk and plasma samples are drawn simultaneously, typically near the expected peak (Cmax) of the dosing interval. Simple to obtain, but sensitive to the exact timing relative to the last maternal dose, since milk and plasma concentrations do not necessarily peak at the same moment.
• AUC-based ratio (preferred): serial milk and plasma samples are collected across a full dosing interval, and the area-under-the-curve (AUC) for each compartment is calculated. The AUC(milk)/AUC(plasma) ratio is considered the more robust and reproducible estimate, since it averages out timing mismatches between the two compartments.
Because milk composition changes across a feed (foremilk is more aqueous, hindmilk more lipid-rich) and across lactation stage (colostrum vs. mature milk), reported M/P ratios for the same drug can vary meaningfully between studies.
M/P ratio alone cannot capture: (1) how much milk the infant actually ingests, (2) the infant's own capacity to absorb, metabolize, and clear the drug — which is markedly reduced in neonates and premature infants, and (3) whether the drug has active or toxic metabolites formed independently by the infant. A full risk assessment therefore layers M/P ratio together with maternal dosing, milk volume, and infant-specific pharmacokinetic considerations — the subject of the following stages.
A high M/P ratio is often treated as an automatic red flag, but it is only one input. Some drugs with M/P ratios near or above 1 are still considered compatible with breastfeeding because the absolute maternal plasma concentration — and therefore the absolute infant exposure — is very low.
Most drugs cross from maternal plasma into milk by passive diffusion across the mammary alveolar epithelium — a lipid bilayer barrier. The same physicochemical properties that govern placental drug transfer during pregnancy — lipophilicity, molecular weight, plasma protein binding, and ionization state — also govern how readily a drug partitions into breast milk.
The mammary alveolar epithelium behaves like most biological membranes: small, uncharged, lipophilic molecules cross most readily, while large, charged, or highly polar molecules are excluded. Only the free (unbound) fraction of drug in maternal plasma is available to diffuse — bound drug cannot cross. Diffusion continues until the free-drug concentration on both sides of the membrane approaches equilibrium, modulated by the additional factors below.
Molecular weight is a coarse but useful filter: most drugs under roughly 500 Da diffuse relatively freely, while very large molecules (e.g., heparins, insulin, many biologics and monoclonal antibodies) transfer into milk poorly or not at all, and are frequently considered compatible with breastfeeding on this basis alone.
Breast milk is slightly more acidic (pH ≈ 7.0) than maternal plasma (pH ≈ 7.4). For weakly basic drugs, this pH gradient favors "ion trapping": the un-ionized (diffusible) form crosses into milk, but a larger fraction becomes ionized once in the slightly more acidic milk environment, and ionized drug cannot easily diffuse back out. This traps a disproportionate amount of basic drug in milk relative to plasma, producing M/P ratios above 1.
Weakly acidic drugs show the opposite tendency: they are more ionized in plasma (pH 7.4) than would be favored in milk, which tends to limit their accumulation in milk and produce lower M/P ratios.
Because only unbound drug is available to diffuse, drugs that are extensively bound to maternal plasma proteins (albumin, alpha-1-acid glycoprotein) tend to show lower M/P ratios — there is simply less free drug available to cross into milk. This is one of the most consistent predictors used in early screening of a new medication's lactation compatibility: a drug with >90% protein binding is often presumed to have limited milk transfer, all else being equal.
Lipophilicity acts in the opposite direction: milk contains a fat fraction (particularly hindmilk), and highly lipophilic drugs can partition preferentially into milk fat globules, elevating the effective M/P ratio beyond what aqueous-phase diffusion alone would predict.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| High lipophilicity (logP) | Small lipophilic drugs (e.g., sedatives, some antidepressants) | Partitions into milk fat globules; favors membrane diffusion | Tends to raise M/P ratio |
| High plasma protein binding | Warfarin-like, highly bound acidic/neutral drugs | Reduces free fraction available to diffuse | Tends to lower M/P ratio |
| Weak base, favorable pKa | Many CNS-active and antihistamine-class drugs | Ion trapping in slightly acidic milk | Tends to raise M/P ratio |
| High molecular weight (>800 Da) | Heparins, insulin, most biologics/mAbs | Poor passive membrane permeability | Minimal transfer regardless of other properties |
The M/P ratio only becomes clinically actionable once it is combined with two additional pieces of information: the maternal plasma drug concentration (driven by maternal dose and pharmacokinetics) and the volume of milk the infant actually consumes. Together these allow calculation of the estimated infant dose (EID) — an approximation of how much drug, in absolute terms, the infant receives per day through breastfeeding.
The estimated infant dose is calculated as:
EID (mg/kg/day) = [drug]milk (mg/mL) × milk intake (mL/kg/day)
where [drug]milk is derived from the M/P ratio applied to the maternal average or peak plasma concentration:
[drug]milk = M/P ratio × [drug]plasma
A representative milk intake value of ~150 mL/kg/day is commonly used for an exclusively breastfed infant, though actual intake varies with infant age, feeding frequency, and whether feeds are exclusive or supplemented.
This EID converts an abstract concentration ratio into an absolute daily dose the infant is estimated to receive — a number that can then be directly compared against known safe or therapeutic dosing ranges for that drug in infants, where such data exist.
Several assumptions embedded in the EID calculation should be treated as approximations rather than precise measurements:
• Maternal plasma concentration used (peak vs. average vs. trough) substantially changes the result — using Cmax gives a conservative (higher) worst-case estimate, while using the average steady-state concentration gives a more typical-case estimate.
• Milk intake volume is assumed rather than measured for an individual infant-mother pair, and true intake can vary two- to three-fold between infants of the same age.
• The calculation assumes the M/P ratio is constant across the dosing interval, when in reality it can fluctuate as maternal and milk concentrations rise and fall at different rates.
Because of these assumptions, EID is best treated as an order-of-magnitude estimate suitable for clinical decision-making, not as a precise pharmacokinetic measurement.
Consider a maternal plasma concentration of 80 relative units and an M/P ratio of 0.5 (milk concentration = 40 relative units, matching the default slider values above). If a representative milk intake of 150 mL/kg/day is assumed, the infant's estimated intake scales directly with the milk concentration: a doubling of the M/P ratio (with plasma concentration held constant) doubles the milk concentration and, proportionally, the estimated infant dose.
This proportional relationship is precisely why the M/P ratio remains a useful screening number even though it is not itself a dose: for a given maternal dosing regimen, EID scales linearly with M/P ratio.
EID gives an absolute quantity of drug, but "is that amount concerning?" cannot be answered without a reference point — which is exactly what the relative infant dose provides in the next stage.
The relative infant dose (RID) expresses the estimated infant dose as a percentage of the maternal weight-adjusted dose, providing a normalized, drug-independent number that clinicians can use as a rule-of-thumb screening threshold. It answers a more intuitive question than the M/P ratio alone: "what fraction of a full therapeutic dose is the infant effectively receiving?"
Relative infant dose is calculated as:
RID (%) = [EID (mg/kg/day) ÷ maternal dose (mg/kg/day)] × 100
By normalizing to the maternal weight-adjusted dose, RID allows comparison across drugs of very different potency and dosing ranges. A widely referenced rule of thumb, popularized by Thomas Hale's lactation pharmacology work, treats an RID below 10% as generally reassuring for most medications and most healthy term infants — though this threshold is a heuristic, not an absolute safety cutoff, and some drugs remain a concern even below it (for example, those with active toxic metabolites or a narrow therapeutic index).
RID values between roughly 10% and 25% are used more selectively, often reserved for situations with strong maternal treatment need, close infant monitoring, or lack of a suitable therapeutic alternative.
In addition to RID, Hale's Medications & Mothers' Milk reference assigns a qualitative lactation risk category to many drugs, synthesizing RID, published infant outcome data, and pharmacologic reasoning:
• L1 (safest): extensive controlled studies show no demonstrated risk to the infant • L2 (safer): limited studies without increased adverse infant effects • L3 (moderately safe): no controlled studies, or studies show minimal non-threatening effects; used when benefit justifies potential risk • L4 (possibly hazardous): positive evidence of risk, but benefit may outweigh risk in life-threatening maternal situations • L5 (contraindicated): significant documented risk to the infant; use is contraindicated in breastfeeding
These categories are a useful adjunct to a quantitative RID calculation, particularly for drugs with limited pharmacokinetic data.
RID thresholds assume a healthy term infant with mature hepatic and renal clearance. Neonates — and especially preterm infants — have substantially reduced drug metabolism and elimination capacity, meaning the same RID can represent meaningfully greater relative exposure in a young or premature infant than in an older, exclusively formula-supplemented infant. Drugs with long half-lives can also accumulate in the infant across repeated feeds even at a seemingly low RID per dose.
For these reasons, RID should be interpreted alongside infant age, gestational maturity at birth, feeding pattern, and the specific pharmacologic behavior of the drug — not applied as a single universal cutoff.
RID <10% is a widely used screening heuristic, not a guarantee of safety — and RID >10% does not automatically mean breastfeeding must stop. Both directions require clinical judgment layered on top of the number.
Every lactation medication decision is ultimately a balance: the documented or estimated risk of infant drug exposure, weighed against the clear and well-established risks of unnecessarily interrupting or discontinuing breastfeeding, and against the risk of leaving the mother's own medical condition inadequately treated. Quantitative tools like M/P ratio, EID, and RID inform this balance — they do not replace clinical judgment.
Discontinuing a needed maternal medication is not a risk-free default choice. Untreated maternal depression, uncontrolled seizures, poorly managed autoimmune disease, or inadequately treated infection all carry direct risks to maternal health, and indirectly to the infant through impaired caregiving capacity, disrupted bonding, and in some cases the need to discontinue breastfeeding entirely (with the loss of its own well-documented immunologic, nutritional, and developmental benefits).
A structured approach typically considers: (1) is the maternal condition itself high-risk if left undertreated, (2) is there a therapeutically comparable alternative drug with a more favorable lactation profile, (3) what does the quantitative RID and qualitative risk category suggest about likely infant exposure, and (4) is infant monitoring (e.g., for sedation, feeding pattern, growth) practical and sufficient to detect a problem early if one arises.
Certain infant populations warrant more conservative thresholds than the general RID heuristic:
• Preterm and low-birth-weight infants: immature hepatic and renal clearance mean drug accumulation is more likely even at a low RID • Neonates in the first weeks of life: similarly reduced metabolic capacity, particularly for drugs cleared via cytochrome P450 pathways that mature over the first months of life • Infants with comorbidities: hepatic or renal impairment further reduces already-limited clearance capacity • Drugs with active/toxic metabolites: the parent drug's RID may understate true risk if the infant's own metabolism generates a harmful metabolite
In these situations, clinicians often favor drugs with the most extensive safety data (Hale L1–L2), the lowest RID, or defer dosing timing relative to feeds where feasible.
When a maternal medication with some degree of milk transfer is clinically necessary, several practical strategies can reduce effective infant exposure without requiring discontinuation of breastfeeding:
• Timing maternal dosing immediately after a feed, so maternal plasma (and therefore milk) concentrations are declining by the time of the next feed • Selecting, where therapeutically equivalent options exist, the specific drug within a class with the lowest reported M/P ratio, RID, or most favorable lactation risk category • Using the lowest effective maternal dose for the shortest necessary duration • Monitoring the infant directly for expected class-specific effects (e.g., sedation, feeding changes, rash) rather than relying on the RID estimate alone • Involving a lactation pharmacology reference (e.g., LactMed, Hale's Medications & Mothers' Milk) and, where appropriate, a lactation consultant or clinical pharmacist in shared decision-making with the mother
The overwhelming majority of maternal medications are compatible with continued breastfeeding when M/P ratio, estimated infant dose, and relative infant dose are used to inform — rather than automatically override — a shared clinical decision between mother and care team.