This commentary suggests that clinicians require an additional perspective to that expressed in Wernicke's review. Clinicians must consider the relative risks, because even small differences may be a reason to adjust usage of a particular drug. As early as 1993, reviews noted a 'serious' interaction potential. Also, the product information requires supplementation because it does not provide the emphasis on predictable interactions that clinicians require. An extensive basis in experimental pharmacology provides the foundations for our knowledge of cytochrome P450 interactions which predict the effects on the levels of other drugs. This work has been confirmed in human studies. Adverse outcomes related to fluoxetine interactions are relatively frequent and sometimes fatal. Patients may suffer serious and irreversible ill-effects from the increased risk of many manifestations of toxicity, and the cost of resultant medico-legal settlements is considerable. A clinician who balances risks may use fluoxetine less frequently.