Some capabilities and limitations of multiple regression analysis: application to canine coronary blood flow

Changes in inspired oxygen and carbon dioxide concentration produced concurrent changes in systemic and coronary haemodynamics in 39 anaesthetized dogs. These changes provided the data base for an exercise in multiple regression analysis using SPSS packages. The physiological problem was to identify influences on coronary blood flow additional to those associated with cardiac work or oxygen consumption. Such influences were identifiable to an extent which was probably limited as much by experimental design as by the underlying physiology. Multiple regression relationships described the integrated response of coronary blood flow to the experimental challenge more completely than any univariate approach. The exercise provided useful insights into a multifactorial problem and the need for deliberate experimental design in attempting to answer specific physiological questions.