A friendly note from the FDA

Via Pharma Manufacturing, Reuters has some tough news about GSK's API manufacturing facility (emphasis mine):

The U.S. FDA issued a warning letter to GlaxoSmithKline after determining that GSK did not take sufficient action to resolve problems after contaminated APIs were found at the drugmaker's Ireland manufacturing plant. 

The contaminated ingredient, paroxetine, is used to make antidepressant drugs Paxil and Seroxat. During an October inspection, an FDA investigator reported that paroxetine was contaminated with material from the Cork plant's pharmaceutical waste tank. According to the FDA, some batches of drugs using the contaminated ingredient were later shipped and GSK failed to notify its customers about the lapse. 

The most strongly worded portion of the March 18 FDA warning letter stated, "We are concerned that your firm does not consider the entry of pharmaceutical waste streams into your manufacturing process a significant deviation with a potential quality impact. In your response to the Form FDA-483, you acknowledged that you should have informed your customers of this incident; however, you did not describe any recent or future communication with your customers regarding the incident to rectify the prior lapse." 

GSK has responded by stating its plans to recall certain batches of Paxil from wholesalers.

From the FDA warning letter itself [I've changed the "(b)(4)" from the letter to "[redacted]"]:

...Your firm discovered that the [redacted] used to manufacture [redacted] batches of [redacted] and [redacted] batches of [redacted] was contaminated with material from your pharmaceutical waste tank, which contained APIs, intermediates, and solvents. (emphasis CJ's) Examples of chemicals that are collected in the waste tank include [redacted]. Your firm became aware of this contamination in January 2012 and completed risk assessments to determine the impact on the quality of [redacted] manufactured using the contaminated solvents on April 19, 2013. Your firm distributed [redacted] shipments of [redacted] potentially contaminated [redacted] batches after becoming aware of this significant deviation. In contrast, [redacted] batches made with the contaminated [redacted] were rejected. 

Quality impact assessments were made for both [redacted] and [redacted], but we note that the approach taken in the two assessments was different. For instance, the [redacted] assessment noted that the standard release testing did not detect significant quantities of contaminants in the potentially impacted [redacted] batches, but that additional testing on [redacted] from [redacted] showed the impacted batches were exposed to significant amounts of [redacted]. The assessment states that the sample preparation used in the [redacted] sample release testing appears to be incapable of complete extraction of the potential contaminants, and it therefore relied on results obtained from the additional testing from the [redacted] of [redacted] product to demonstrate that the [redacted] batches were impacted by the pharmaceutical waste contamination event. Your firm’s assessment for [redacted] included no such additional testing and relied on the [redacted] samples’ passing test results, concluding that there was no quality impact to the [redacted] batches....

...Please also describe why the quality assessments appear to assume uniform distribution of contaminants following addition of [redacted] to the waste stream and before the backflow of contaminants into the [redacted] tank. Provide a revised quality assessment for [redacted] that clearly describes all calculations used to support the conclusions, and clearly describe any altered conclusions after addressing the issues described in this letter. For each analytical method used to support your conclusions, provide method qualification information, including the limit of detection for each potential contaminant. Also, provide a quality impact assessment for your [redacted] product, which was also manufactured using [redacted] around the time of the initial contamination in the [redacted] tank. Describe any contact you have had with the customers of the potentially affected products and your plans with respect to the disposition of any potentially affected batches that remain within expiry.  

(You have to love a letter that is addressed directly to the CEO. Wow.)

It's difficult to tell exactly what happened from this warning letter, but I am wondering if there was a tank that was somehow connected to a waste tank and something burped over? Or the wrong valve got opened for just a second? I'm confused. There's a hell of a story around here somewhere.

Ultimately, what is really bad is that rather than trashing the batches, they decided that they could continue processing. While it may have saved them a batch or two of product, it sure seems like it has made much worse headaches later.