Adverse drug reactions are of particular interest when they have been clandestine for many years. We were admitted access to unpublished data through the Civil Ombudsman’s unusual decision to advise the Norwegian Ministry of Health to let us examine files in the archives of the Norwegian Medicinal Agency. We found that reports of a series of randomized controlled studies evaluating the efficacy and safety of paroxetine already existed at the time of the first application for marketing authorization. Common to these studies were that they were small, were of short duration and were mostly published individually. Some of them have remained unpublished and no thorough meta-analysis has been carried out either for efficacy data or for adverse event data. The main published conclusions from these studies are that paroxetine is a safe and effective medication for the treatment of major depression.2–7
During the widely clinical use of paroxetine, more adverse drug reactions have become evident than were evident when the drug was first licensed in 1989. Our hypothesis was that some of these adverse drug reactions were already documented in the initial clinical trials forming part of the application for marketing authorization, but that no adequate pooling or meta-analysis of the data was carried out. We therefore wanted to explore the usefulness of gathering information about the occurrence of adverse reactions to a drug by means of randomized controlled trials at the time of first application for marketing authorization and to analyse these data by a meta analysis.
We wanted to determine to what extent adverse drug effects associated with a selective serotonin reuptake inhibitor (SSRI) were known but not assessed before application for registration of paroxetine.
Methods With a special permit from the Norwegian Ministry of Health, we obtained reports from 82 clinical trials presented by the paroxetine license holder in 1989. There were 17 double blind, placebo controlled clinical trials with parallel design with 903 patients on paroxetine and 592 on placebo. Altogether 32 adverse effects showed a risk difference (RD) between paroxetine and control groups of more than 0.5%.We did a meta-analysis for each of these adverse effects. We then compared the outcome with the frequencies stated in the Summary of Product Characteristics (SPC) at the time of registration and those reported in the current SPC.
At the time of registration 19 of the adverse effects were statistically significant. Only eight of these adverse effects were listed as being common in the first SPC from 1989. Five out of the nineteen adverse effects are not mentioned in the current SPC. Among them are headache with RD 5.4%, decreased libido RD 2.6%, nervousness RD 2.0% and paresthesia (brain zaps) RD 1.7%.
Frequently occurring adverse reactions that are included in today’s SPC for paroxetine were evident and documented already in the early studies accompanying the application for marketing authorization in 1989. Some other adverse effects observed then are still not mentioned in the SPC of today. Meta-analyses of adverse effects should be mandatory at the stage of first registration of a drug.
I see from my webstats that GlaxoSmithKline from Hertford, UK, have been looking in today.
I have a message for you:
You manufactured and marketed a drug that I took. Because you never reported these adverse events at the time I took it, I put my faith in you, I trusted you. I have since learned that you not only duped me but our regulator too, and now it seems, the Norwegian regulator!
It's my mission to make sure you don't dupe others, be they regulators, doctors or more importantly, patients.
Do you want me to go away Glaxo?
See me as your confessional box, although I won't be telling you to go away and say three Hail Mary's for your sins.
I'm here to stay.
GET USED TO IT.
Read the new book, The Evidence, However, Is Clear...The Seroxat Scandal
By Bob Fiddaman
AVAILABLE FOR DOWNLOAD HERE
PAPERBACK COMING SOON