It would be hard to find a psychiatrist who cannot tell you a story about a patient who appeared to do worse after switching from a brand-name medication to a generic version. But what do such stories mean?
One of cardinal rules of scientific evidence is that correlation does not imply causation. When a change in symptoms or side effects occurs after a medication switch, we are naturally tempted to blame the switch – or praise it in cases of improvement. However, given the high rates of positive and negative placebo responses in psychiatry, we need to be cautious about ascribing symptomatic changes to changes in pills.
Unfortunately, most of the evidence of failure of generic substitutions in psychiatry comes from either single cases or very small case series, virtually all written by authors who are also paid consultants for pharmaceutical companies. A representative example is a recent case series of seven patients treated at a tertiary care psychiatric clinic in Toronto, Canada (Rosenthal J et al., J Fam Practice 2008;57:109-114). In a retrospective chart review, the authors noted that between 2004 and 2006, seven of their patients who had been stable on medication had relapsed for no apparent reason. Each of these patients had been switched either from a brand-name antidepressant to a generic or from one generic to a different generic. The authors then assumed that the drug substitutions caused the relapses. The problem is that they made no attempt to systematically identify all the patients in their practice who had undergone a generic substitution over that three-year period. Presumably, hundreds of patients had switched to generics with no untoward effects. We are more likely to remember those patients who do poorly than those who do just fine – an example of a common pitfall of clinical research called “recall bias.”
One of the few examples of robust research on generics in psychiatry was related to the antipsychotic clozapine. In 1998, Novartis’ patent for brand name Clozaril expired and generic formulations of clozapine entered the market. A battle of research papers ensued. Novartis started by funding a study in which 45 patients were randomly assigned to either Clozaril (21 patients) or generic clozapine (24 patients). The patients were not told about the different formulations but the investigators knew (i.e., it was a single blind study). After eight weeks, the patients were crossed over to the opposite formulation. Patients were more likely to relapse when switched to generic clozapine than when they were switched to branded Clozaril (Kluznik JC et al., J Clin Psychiatry 2001;62(suppl 5):14-17).
This paper caused a stir in the field and prompted the FDA to request that the generic makers repeat their bioequivalence studies, which they did, to the FDA’s satisfaction. Meanwhile, the generic companies sponsored their own studies, both of them retrospective reviews of health systems that had mandated that all patients be switched from brand to generic clozapine. Both analyses concluded that there was no clinical deterioration after generic substitution (Makela EH et al., Ann Pharmacother 2003;37(3):350-353; Stoner SC et al., Pharmacotherapy 2003;23(6):806-810). Later, three more studies appeared, sponsored by neither Novartis nor the generic companies. The largest was a study in the United Kingdom of 337 patients who were followed by researchers from a month before the brand-to-generic switch to three months after the switch. They found no evidence of clinical deterioration or of the need to increase dosages (Paton C, Br J Psychiatry 2006 Aug;189:184-185).
A similar current controversy involves the generic formulation of GlaxoSmithKline’s Wellbutrin XL. In 2007, the FDA received reports about 78 patients who were switched from Wellbutrin XL 300 mg to Teva Pharmaceutical’s generic Budeprion XL 300 mg. The complaints ranged from lack of efficacy to new onset side effects. A private lab called Consumer Lab (http://tinyurl.com/ce636m) ran its own tests of the generic and reportedly found that 34% of the bupropion contained in the pill had been released after two hours, as opposed to only 8% of the bupropion in Wellbutrin XL. The FDA conducted its own review (http://tinyurl.com/d255gm) in which they reanalyzed the original bioequivalence data submitted by Teva. The FDA’s report re-affirmed that the concentrations of the generic were within the acceptable limits (80% to 125% of the original), but they acknowledged that the generic did indeed release bupropion into the bloodstream more quickly, with a T max (time to maximum plasma concentration) of 2-3 hours vs. Wellbutrin XL’s T-max of 5-6 hours. However, the FDA concluded that this more rapid pill dissolution was not “clinically significant,” and said that the various complaints about the generic were consistent with the recurrent nature of depressive symptoms and with the known side effects of bupropion products.
So what to make of this assertion? It’s hard to say, because we have no way of knowing if a more rapid dissolution of bupropion would change its efficacy (this seems unlikely) or worsen its side effects (more plausible). It would be nice to see rigorous comparative trials of generic vs. branded Wellbutrin XL, but that’s unlikely to happen any time soon. In the meantime, we should more closely monitor patients taking generic versions of extended release medications, because the FDA seems to allow more variation in release parameters than they do in final plasma concentration.