Re-examination of some of the Framingham blood-pressure data.

Mathematical smoothing of data for the Framingham study leads to the conclusion that there is no threshold of normality for blood-pressure. A reasonable therapeutic goal might be "the lower the better". However, examination of the unsmoothed Framingham data indicates that while systolic pressures fit this no-threshold model reasonably well, changes in diastolic pressure below about 90 mm Hg have little or no prognostic significance. The poor fit of the diastolic points to the no-threshold model may also help to explain why at Framingham raised systolic pressure seemed to have a greater impact than raised diastolic pressure.