Class Effects of Pharmaceuticals and Therapeutic Interchangeability
Class substitution means that an agent in a given class can be substituted for an alternative agent prescribed from the same class. Class approval means that a new agent can be indicated for a condition despite the fact that it was not studied for that indication. The drug may be approved because it demonstrated a similar class effect to the approved drug. It is probably more useful not to be abstract about these considerations, but to discuss these concepts in terms of specific classes of pharmaceuticals. Two examples are the statin, lipid-lowering agents and the converting enzyme inhibitors, both well established classes of drugs.
The statins have very similar mechanisms of action, and every statin that has been studied in large natural history endpoint trials has been shown to reduce mortality. Although there are some differences in effects on lipid components, it appears that these differences are not clinically significant. Given this information, physicians who base their therapeutic treatments on evidence-based trials should be confident in their ability to substitute one statin for another and should expect similar benefits at comparable doses. That does not mean that there are not important differences. Drug interactions are one area in which, given the other therapies a patient is taking, one agent may be superior to another. Liver toxicity may be another area in which one statin in an individual patient may cause a greater rise in liver enzymes than another statin. For these reasons, the physician should be the only one allowed to interchange agents. Still, we cannot deny a therapeutic class effect. Additionally, after four or five statins have been shown in different settings to reduce ischemic-related morbidity and mortality, how may resource-intensive, large natural history endpoint studies should be undertaken? Although there is no ready answer, most physicians and regulators are willing to accept the surrogate endpoint of cholesterol (specifically LDL cholesterol) reduction instead of myocardial infarction and death as an appropriate endpoint—an endpoint on which to base clinical treatment decisions as well as drug approval. Some would want a separate efficacy study for each agent, going so far as to demand a dose–response curve for each agent established—not for the surrogate endpoint but for the natural history endpoints of death or ischemic morbidity (stroke, myocardial infarction, etc.). Such a position would be extreme, wasteful of resources, and dangerous to patients partaking in these studies (if the study were placebo-controlled, patients on placebos would not receive appropriate therapy for prolonged periods). Using a putative placebo for comparative studies is a workable alternative but is still costly and utilizes too many clinical resources. Why not halt further drug approvals if two drugs are established as effective (ie, no more so-called “me-too” agents)? It turns out that me-too agents, because of fewer drug interactions, reduced side effects, better kinetics, and thus improved patient compliance, are often improved agents. The first drug in a field is not always the most useful one.