Gestation increases the biomechanical loading of lower extremities. Gestational loading may influence anthropometrics of articular surfaces in similar means as bone diaphyseal properties. This study aimed to investigate whether gravidity (i.e. number of pregnancies) and parity (i.e. number of deliveries) is associated with knee breadth among middle-aged women. The study sample comprised 815 women from the Northern Finland Birth Cohort 1966. The median parity count of our sample was 2 and the median gravidity count 3. At the age of 46, questionnaires were used to enquire gravidity and parity, and posteroanterior knee radiographs were used to obtain two knee breadth parameters (tibial plateau mediolateral breadth (TPML) and femoral condylar mediolateral breadth (FCML)) as representatives of articular size. The associations of gravidity and parity with knee breadth were analyzed using general linear models with adjustments for height, weight, leisure-time physical activity, smoking, and education years. Individuals with osteoarthritic changes were excluded from our sample. The mean TPML in our sample was 70.3 mm and the mean FCML 71.6 mm respectively. In the fully adjusted models, gravidity and parity showed positive associations with knee breadth. Each pregnancy was associated with 0.11–0.14% larger knee breath (p < 0.05), and each delivery accounted for an increase of 0.20% in knee breadth (p < 0.01). Between-group comparisons showed that multiparous women had 0.68–1.01% larger knee breath than nulli- and primiparous women (p < 0.05). Pregnancies and deliveries seem to increase the mediolateral breadth of the knee. This increase is potentially associated with increased biomechanical loadings during gestation.