Do package inserts reflect symptoms experienced in practice?: assessment using an automated phone pharmacovigilance system with varenicline and zolpidem in a primary care setting.

BACKGROUND

While the US FDA maintains a voluntary reporting system, postmarketing adverse drug events (ADEs) are underreported, and this case report-based system does not allow accurate determination of incidence.

OBJECTIVE

The aim of the study was to assess the usefulness of an automated phone pharmacovigilance system for ambulatory patients by comparing systematically collected, patient-reported symptoms to reflect possible ADEs with those reported on the package inserts of two drugs with postmarketing safety concerns, varenicline and zolpidem.

METHODS

English-speaking adults who received a prescription for zolpidem (n = 370) or varenicline (n = 107) from a primary care physician at one of 11 participating clinics, and who participated in the pharmacovigilance system during 2008-2010, were included in the study. Patients were called approximately 4 weeks following their visit and asked to complete a standard script that asked about adherence and pre-specified symptoms.

MAIN OUTCOME MEASURES

The main outcome measures were elicited rates of pre-specified symptoms or possible ADEs.

RESULTS

Compared with the package insert, patients taking zolpidem were significantly (p < 0.001) more likely to report fatigue (9.0% vs 1.0%), itching (4.5% vs 1.0%) and muscle aches (5.6% vs 1.0%). Elicited rates of depression and hallucination were similar to those reported in the package insert. Patients taking varenicline were significantly more likely to report confusion (1.7% vs 0.1%), depression (3.4% vs 0.1%), fatigue (6.0% vs 1.0%), hallucinations (1.7% vs 0.1%), muscle aches (6.0% vs 1.0%) and sexual dysfunction (4.3% vs 0.1%).

CONCLUSIONS

Automated phone pharmacovigilance can provide estimates of possible ADEs in clinical practice. In the case of varenicline, these data support some of the safety concerns that have come to light postmarketing, while others such as depression and hallucination related to zolpidem were not detected. These data highlight the potential value of, and innovative ways of collecting, information about possible ADEs directly from patients.