A small effect that turns up in every study
Reddy and colleagues, working with NIH data, found that women in the tallest height quartile had roughly 1.4 times the dizygotic twin rate of women in the shortest quartile, and women with BMI ≥ 30 had roughly 1.5 times the rate of women with BMI < 25 [1]. Basso and colleagues, using a Danish cohort, found very similar effect sizes — and noted that the height effect was independent of the BMI effect [4].
The signal is robust but small compared to age, ART, family history or ethnicity. For the calculator we use:
- Height: ×0.9 (under 160 cm), ×1.0 (160–170), ×1.2 (170–175), ×1.5 (over 175 cm)
- BMI: ×0.95 (under 25), ×1.1 (25–30), ×1.3 (over 30)
These two multipliers compose independently in the model.
The IGF-1 hypothesis
The leading mechanistic explanation is insulin-like growth factor 1. IGF-1 sensitises ovarian follicles to FSH, making it more likely that more than one will reach the threshold for ovulation in a single cycle. Both height and BMI correlate positively with circulating IGF-1: taller women have higher IGF-1 because they were exposed to higher growth-axis activity during development; women with higher BMI have somewhat higher IGF-1 because of metabolic factors.
Steinman has argued that the same axis explains the dietary signal observed for dairy and twinning [2]: dairy consumption raises IGF-1 modestly. The hypothesis remains plausible but not definitively proven, and other mechanisms (including direct gonadotropin effects in obesity) likely contribute.
What this changes in practice
For most readers, height and BMI together move twin probability by a factor of roughly 1 to 2 — meaningful but small relative to the headline factors. The clinical implication is essentially nil: nobody recommends gaining or losing weight or selecting a partner by height to influence twin probability.
Where it does matter is in epidemiological modelling. When researchers analyse twin rates over time, they have to control for population-level shifts in BMI, which would otherwise confound trend estimates. The recent rise in BMI in many high-income countries has likely contributed a small upward push to twin rates, alongside the much larger ART and age effects.
Two important caveats
- Height and BMI affect dizygotic twins only. Monozygotic twin rates are essentially flat across body composition, age and ethnicity.
- Self-reported BMI is noisy. If you are using the calculator for a personal estimate, your BMI is one of the smaller dials and should not be over-weighted in interpretation.