Factors

How previous pregnancies and twin probability affects twin probability

Parity has a modest population association with fraternal twinning. Model v2 uses a small weight and excludes breastfeeding.

Effect size: ×1.00–1.15 vs. baseline

Last updated:

Previous pregnancies

Parity—the number of previous pregnancies—has been associated with dizygotic twinning in several population datasets, but the evidence is mixed and maternal age is a stronger predictor [1]. Parity can also correlate with age, fertility and family characteristics, making a large independent coefficient difficult to defend.

Model v2 therefore uses compressed weights on the dizygotic component:

  • No previous pregnancy: ×1.00
  • One: ×1.05
  • Two: ×1.10
  • Three or more: ×1.15

These are heuristic model choices, not clinical coefficients.

Why breastfeeding was removed

The former model used a ×1.4 breastfeeding multiplier based mainly on a small observational report [2]. We did not find sufficiently robust replication to support that number. Breastfeeding is no longer asked or weighted; legacy values are neutral.

Pregnancy spacing and breastfeeding decisions involve recovery, feeding goals, contraception and individual health. They should not be changed to influence twin probability.

Place parity alongside the stronger documented factors in the combined twin-probability estimate .

Source

How we calculated this

See the multiplier and how this factor combines with the rest of the model.