Parity: the small uplift with each pregnancy
Across population-level studies, dizygotic twin probability rises modestly with parity — the number of previous pregnancies. The classic reference is Bortolus and colleagues’ 1999 review, which pooled European cohort data and found roughly a 10–30% increase per additional prior pregnancy [1]. Hoekstra and colleagues confirmed the pattern in a more recent meta-review [3].
For our calculator we apply:
- No previous pregnancies (P0): ×1.0
- One previous pregnancy (P1): ×1.1
- Two previous pregnancies (P2): ×1.2
- Three or more (P3+): ×1.3
The mechanism is not fully understood. The leading hypotheses are gradual hormonal recalibration (prolonged exposure to higher estrogens through pregnancy and lactation) and a selection effect (women who conceive easily for one pregnancy are more likely to conceive easily for subsequent ones, including twin pregnancies). The signal is small enough that it is not a major driver of personal twin probability — but it is a consistent statistical signal in the population.
Breastfeeding: a separate, larger effect
Steinman’s 2001 paper proposed and tested the hypothesis that conceptions occurring during ongoing breastfeeding have higher twin probability than conceptions in non-lactating women [2]. The reasoning is straightforward: breastfeeding suppresses ovulation through prolactin, and when ovulation does resume, the underlying FSH state can briefly favour multifollicular ovulation in some women.
Steinman estimated the effect at roughly 1.5× — meaningfully higher than the parity effect. We use ×1.4 in the calculator. ACOG’s guidance recognises this as a recognised but still under-studied factor in twin epidemiology [4].
How these stack with the rest of the model
Because the effects are small and overlap (multiparous women are more likely to be currently breastfeeding), we deliberately use moderate multipliers and rely on the model’s overall cap of 25% to keep extreme compositions in check.
A practical example: a 36-year-old woman, P2, currently breastfeeding, with a maternal-side family history of fraternal twins, lands at:
- 1.5% baseline × 4.0 (age) × 1.2 (P2) × 1.4 (breastfeeding) × 2.5 (family history) ≈ 25.2% — capped at 25%.
The cap is doing real work here. Without it, the model would output an unrealistically high estimate that does not reflect the way these factors interact in real populations.
What this means in practice
If you are weighing whether to plan a pregnancy while still breastfeeding, twin probability is a downstream concern compared to your own recovery, child-spacing preferences, and clinician guidance. The uplift is real but small in absolute terms — a couple of extra percentage points on top of your baseline.
If you are reading this with a third or fourth pregnancy in mind, the parity effect is real but not a reason to plan around twins. It is a reason to expect the dial to move slightly, in the direction the population data already shows.