Abstracts

Rationale: Generic drugs are considered bioequivalent (BE) and therefore therapeutically equivalent to innovator brands. However, there are several reports of loss of efficacy when brands are switched to generic. This suggests bio-in-equivalence of generics despite passing the bioequivalence criteria. Therapeutic response is usually a sum of drug and placebo effects. Placebo effect is due to expectation and/or learning and conditioning to desired outcome. This effect is estimated to contribute 20-50% of anticonvulsive therapeutic response. Changes in placebo and drug effect can occur when brand anticonvulsants are switched to generic. Thus in effort to determine the probable reason for the loss of efficacy after generic switch, we sought to examine the extent of complaints in literature. Specifically, the number and type of study designs reporting loss of efficacy, the suggested reasons for loss of efficacy and number of reported complaints for each of the anticonvulsants.Methods: PubMed and EMBASE journal databases were searched using a search phrase containing all anticonvulsant names combined with the phrase ‘generic drug substitution’. The search results were assessed to remove repeated and irrelevant articles.Results: We found 36 articles that linked brand to generic or generic to generic switch to adverse health or economic outcomes. Most of the studies were case control (n=12) and case reports (n=8). Other kind of studies were; prospective or retrospective crossover studies (n=4), surveys (n=3), reviews (n=3), simulations (n=3), case series (n=2), and randomized trial (n=1). The generic anticonvulsants which appeared or directly linked to adverse outcomes in the aforementioned studies were; lamotrigine (n=12), phenytoin (n=9), carbamazepine (n=8), valproic acid (n=8), gabapentin (n=6), topiramate (n=5), zonisamide (n=5) and levetiracetam (n=5) Some of the suggested explanations for therapeutic in-equivalence of generics included that; current BE metrics (Cmax, AUC) may not be predictive of therapeutic response and alternate BE metrics (e.g. partial AUC) may be necessary; current BE criteria may not ensure BE for narrow therapeutic index drugs and due to within-individual variation in therapeutic response some patients are more sensitive to minor fluctuations of bioavailability. Nevertheless, some other reasons suggested for the observed loss of efficacy were; negative attitudes towards medication, nocebo effect and possible confounding effect inherent in observational studies.Conclusions: The number of studies suggesting therapeutic in-equivalence of generic anticonvulsants is large and plausible explanations have been offered. It is important to continue spontaneous reporting of loss of efficacy and to develop innovative quantitative methods which will timely inform regulatory agencies to initiate active investigations when a predetermined threshold of reports is surpassed.