Background:  Obesity rates around the world are increasing. The effect of BMI on the Pearl Index of COC is unknown. Objectives:  To determine if the Pearl Index of COC differs with BMI. Design and methods: Meta-analysis of five prospective, observational cohort studies with primary endpoints of venous thromboembolism (VTE) in women using COCs. Studies were conducted between 2007 and 2017 using a similar methodology. More than 240,000 women contributing approx. 400,00 women-years, were included. Women were followed for 3-5 years. Inclusion criterion for all studies was the prescription of a new COC, with no specific exclusion criteria.  Studies were conducted across Europe (EU) and the United States (US). All women were followed for 3-5 years. Results were analysed within four age cohorts = 40 years. BMI was defined dichotomously as = 35 kg/m2 (US) and = 30 kg/m2 (EU). The Pearl Index was calculated within each age and BMI category stratified by region. Significance of factors was tested in a stratified Cox regression model; age and BMI where included as continuous variables. Results:  In the US the PI ranges from 0.15 (age 40+, BMI = 35) with higher values observed in women with BMI > = 35 kg/m2 within each age-group. Significance was obtained for both factors when simultaneously included in a Cox regression model. In the European sample PI ranges from 0.06 (age 40+, BMI = 30). Cox regression show independent effects of age and BMI on the occurrence of an unintended pregnancy. Conclusions: BMI has a significant effect on the Pearl index of COC. Increasing BMI decreases the efficacy of COC in EU and US.