Background subtraction for contrast echocardiography is considered a prerequisite for an appropriate assessment of flow from the time-intensity curves. We evaluated the influence of different background intensities on flow determinations with contrast echocardiography. Experiments were performed with a mixing chamber as a phantom for myocardial perfusion. Three regions of interest with different background intensities but equal flow were examined. Time-intensity curves were obtained and, after background subtraction, a number of curve parameters were determined: peak intensity, area under the curve, curve width and decay after peak intensity. The values of these parameters differed significantly between a region with a low and a region with a high background. In addition, a model is presented to describe these findings. It is concluded that subtraction of the background intensity from a time-intensity curve induces errors that make a quantitative interpretation of the curves ambiguous. These drawbacks are avoided if regions of interest with equal background intensities are compared.