“…Unfortunately, the autopsy rate was usually low in the patients who died. Indeed, recent autopsy data continue to show that the majority of the patients who die of pulmonary embolism, this diagnosis was not considered to be the cause of death prior to autopsy (Havig 1977, Sandler and Martin 1989, Karwinski 1995. Thus, it is evident that meta-analytic techniques which have recently been widely criticized (Greenland 1994, Olkon 1994, Khan et al 1996, Sacks et al 1997, cannot be recommended.…”