Quantifying Risk: The Role of Absolute and Relative Measures in Interpreting Risk of Adverse Reactions from Product Labels of Antipsychotic Medications

Abstract

Pharmaceutical product labeling as approved by regulatory agencies include statements of adverse event risk. Product labels include descriptive statements such as whether events are uncommon or rare, as well as percentage occurrence for more common events. In addition tables are provided with the frequencies of the latter events for both product and placebo as observed in clinical trials. Competing products are not mentioned in a specific drug's product labeling but indirect comparisons can be made using the corresponding label information for the alternate product. Two types of tools are easily used for this purpose: absolute measures such as number needed to harm (NNH), and relative measures such as relative risk increase (RRI). The calculations for both of these types of quantitative measures are presented using as examples the oral first-line second-generation antipsychotic medications. Among three sample outcomes selected a priori, akathisia, weight gain, and discontinuation from a clinical trial because of an adverse reaction, there appears to be differences among the different antipsychotics versus placebo. Aripiprazole was associated with the highest risk for akathisia, particularly when used as adjunctive treatment of major depressive disorder (NNH 5, 95% CI 4-7; RRI 525%, 95% CI 267%-964%). Although insufficient information was available in product labeling to calculate the CI, olanzapine was associated with the highest risk for weight gain of at least 7% from baseline (NNH 6, RRI 640% for adults; NNH 4, RRI 314% for adolescents), and quetiapine for the indication of bipolar depression was associated with the highest risk of discontinuation from a clinical trial because of an adverse reaction (NNH 8, RRI 265% for 600 mg/d; NNH 15, RRI 137% for 300 mg/d). In conclusion, with certain limitations, it is possible for the clinician to extract information from medication product labeling regarding the frequency with which certain adverse reactions can be expected. This supplements, but does not replace, information reported directly in clinical trial reports.