This report was never submitted to the FDA according to the testimony of Glaxo employees, Parisian said, but it should have been because it "is safety information."
In 1998, Glaxo prepared another internal report titled, "Paroxetine and Reports of Congential Abnormalities," with a cutoff date of November 1997, that showed 42 cases of congenital anomalies.
Parisian said this analysis was also never submitted to the FDA, as far as she knew, but under the rules it should have been.
Peavy then had Parisian review the draft version of the report, showing the same cutoff date and 42 congenital anomalies. On page 5, it stated: "The incidence rate of congenital abnormalities as observed in the data reported in this document is 13.3 percent."
But the draft version contained the following statements that did not appear in the final report: "A figure much higher than that reported by birth-defect monitoring bodies. Taken at face value, this represents an alarming finding."
Under FDA rules, drug companies are required to report alarming findings to the FDA, Parisian said. "They are also required to report any kind of safety analysis that they have done internally like this to address an issue about safety."
In fact, Peavy produced a 1984 letter to the company from the FDA, which specifically stated that the agency wanted to be informed about any "alarming findings" in either human or animal studies.
The 13.3 percent incident rate of congenital abnormalities was between six and seven times the 2.5 percent overall birth defect rate you would expect to see for births in the US, Parisian told the jury. The CDC uses a 2.5 percent background rate.
Next Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19
(Note: You can view every article as one long page if you sign up as an Advocate Member, or higher).



