The question of whether metabolic drug clearance is decreased in elderly people has been the subject of considerable debate and is very important because clearance is a determinant of dosing. Drug clearance has been shown to be consistently impaired for flow-limited (high-clearance) drugs, but there have been conflicting results for capacity-limited (low-clearance) drugs. A limitation of the studies of capacity-limited drugs is that most have estimated clearance based on total drug concentrations (protein-bound plus free). Total drug clearance reflects both the intrinsic clearance of free drug and the extent of protein binding. Total clearance is a valid measure for capacity-limited drugs with low protein binding and appears to be consistently impaired in elderly subjects. For phenazone [antipyrine] (fraction unbound [f(u)] >0.9), seven studies have demonstrated statistical reductions in clearance of 20-52%. For theophylline (f(u) 0.6), five studies have demonstrated reductions in clearance of 22-35%. For paracetamol [acetaminophen] (f(u) 0.8), the clearance of which has been quoted as unchanged, four studies have demonstrated reductions in clearance of 19-35%. For highly protein-bound drugs, total clearance is not the appropriate parameter. Free drug clearance is more appropriate since it is independent of changes in protein binding. The literature was reviewed to test the hypothesis that in elderly people, capacity-limited drugs with high protein binding will show decreased free clearance even in the absence of a decrease in total clearance. For these drugs, data for free drug clearance based on measurement of actual free drug concentrations are limited, but suggest that the intrinsic metabolic clearance is impaired in elderly subjects. Four studies of naproxen (f(u) <0.01) have shown reduced free drug clearance of 50% or more. Two studies of valproic acid (f(u) 0.1-0.2) have shown reduced free clearance of 39% and 65%. Two studies of ibuprofen (f(u) <0.01) have shown reduced free clearance of S-ibuprofen of 21% and 28%. There is some indirect evidence for reduced clearance of the highly protein-bound drugs oxaprozin, temazepam, lorazepam, diazepam, phenytoin and warfarin, although studies measuring free concentrations are lacking. Together, the above studies support the hypothesis that the intrinsic metabolic drug clearance is impaired in elderly subjects, in the order of 20-60%, and that this effect is masked if highly protein-bound drugs are assessed using total drug clearance. If the findings are confirmed in future well-designed studies of free drug clearance, there are profound and beneficial implications for dosing of drugs in elderly people. Lower doses are likely to achieve appropriate concentrations, allowing full efficacy but decreased dose-related adverse effects.