IMPS (intact months of patient survival): an analysis of the results of carotid endarterectomy.

The literature on carotid surgery for lesions appropriate to prior episodes of ischemia has been reviewed. Only one randomized study and six non-controlled reports give useful data (this despite more than thirty years of surgical activity in this field). When analyzed by the IMPS (intact months of patient survival) criterion, the randomized study failed to show benefit from surgery. This failure can be attributed to a high (35%) operative stroke and death rate. That sufficiently low operative stroke and death rates are readily achievable is not clear, however, only two of six relevant non-controlled series reported in the literature had operative stroke and death rates below the 10.4% level calculated as necessary for a "break-even" situation. Three of the six non-controlled series contain sufficient follow-up data to permit IMPS comparison against the "standard" of the control group of the randomized study. Against this "standard" only one of the three non-controlled studies would have "shown benefit" from surgery. Barnett, Plum, and Walton have called for audits of endarterectomy results in institutions in which such surgery is performed. It is suggested that such audits be done by the IMPS method, which gives appropriate weight to the effects of operative, as well as of long-term follow-up, strokes and deaths.