OBJECTIVES: To replicate a widely cited Ford et al. study (NEJM 2007) that concluded that the decline in CHD mortality in the U.S. between 1980-2000 could be equally attributed to: (1) medical therapies; and (2) changes in CHD risk factors outside of medical therapies. Our replication corrected for one element that could affect (2) above: changes in CHD-related medication rates from 1980-2000. METHODS: Ford et al. used average changes in blood pressure and cholesterol in untreated populations to estimate the total deaths prevented or postponed (DPP) from these risk factor reductions outside of medical therapy. However, treatment rates increased from 1980-2000, so it is possible that reductions in average blood pressure or cholesterol among untreated individuals were (at least partially) an artifact of this methodological approach. We adjusted for changing treatment rates by adding back systolic blood pressure (18.8 mmHg) and LDL cholesterol (1.09 mmol/L) to some (for blood pressure) or all (for cholesterol) treated individuals (treatment rates of 18.3% and 7.7%, for blood pressure and cholesterol, respectively) in 2000 to approximate treatment rates in 1980 (6.0% and 0%, respectively). Otherwise, we used the same methods and data sources, such as individual-level data from the National Health and Nutrition Examination Study, as Ford et al. RESULTS: In our replication, we estimated total DPP of 236,728 between 1980-2000, with 159,330 (66.3%) attributable to medical therapies and 81,092 (33.7%) from risk factor changes outside of medical therapies. Results were sensitive to assumptions around the levels of blood pressure and cholesterol added back to treated individuals in 2000. CONCLUSIONS: After correcting for changes in treatment rates, approximately two thirds of the decline CHD mortality in the U.S. from 1980-2000 could be attributed to medical therapies, with changes in risk factors outside of medical therapies accounting for one third of the CHD mortality decline.