Dear Editor,

Leung and French have stated that changing the goalposts in the definition of peri-procedural MI or bleeding does not necessarily make such events more or less useful in the overall risk assessment of a patient. They have also raised the issue of whether cardiac markers should be treated as continuous rather than as discrete all-or-none variables in data analysis (1).

The relationship between a variable and a particular outcome should determine if a variable is treated as a continuous or an all-or-none parameter. If there is a convenient point on the association curve for a variable and an outcome beyond which adverse events increase significantly, then it would be prudent to have that as the cut-off value for the variable.

The relation between periprocedural CK-MB elevations and late mortality is approximately linear, and myocardial necrosis has similar prognostic significance whether it is caused by a "spontaneous" ischemic event or by PCI (2). This relation between a "surrogate" and “hard clinical outcome” would be a straight line, where any cut-off that is determined is arbitrary. A cut-off of three times the upper limit of normal has no special significance beyond which adverse outcomes increase significantly.

Therefore, moving from the all-or-none analysis for cardiac biomarkers to a more continuous model would be more valid and provide better utility for overall risk assessment.

References

1. Leung DY, French JK. End points in clinical trials: are they moving the goalposts? Heart 2006;92(7):870-2.

2. Akkerhuis KM, Alexander JH, Tardiff BE, et al. Minor myocardial damage and prognosis: are spontaneous and percutaneous coronary intervention- related events different? Circulation 2002;105(5):554-6.