Authorities are not in full agreement in regard to risk of surgery in the elderly. One hundred eight studies of surgery in the elderly over the past 40 years were reviewed. The purpose was not merely to tabulate results, but to identify differences existing between reports with regard to data reported that could affect results independent of the surgical management itself. Sources of variance that need to be taken into account in comparing mortality rates between studies, such as whether mortality was computed by number of patients or operations, differing lengths of follow-up for recording mortality, proportions of emergency versus elective operations, and types of surgical procedures, were documented. A nine-item confidence in results scale was used to classify studies into high and low confidence groups. Surgical specialties scored substantially higher than general surgical studies. More recent studies received higher scores than earlier studies. Although mortality rates varied widely depending on methods of their calculation, there appeared to be a trend toward increases in elective, but not emergency, mortality rates in general surgery since 1941 that should be examined more closely. One thing that cannot be answered clearly from these studies is the relative risk of surgery with age. Some control of variations between studies and standardization of reporting surgical deaths are required before risk of surgery in the elderly can be assessed more accurately.