According to the Second Amended Complaint, at all times relevant to the allegations in the complaint, Abbott developed, manufactured, labeled, distributed, and marketed Humira, a drug approved by the United States Food and Drug Administration (“FDA”) in December 2002 to treat moderately to severely active rheumatoid arthritis. In September 2004, Elnora Jones was infused with Humira to treat her rheumatoid arthritis. She was infused with Humira on a regular basis thereafter through December 2005. In June 2006, Jones was diagnosed with lymphoma, for which she subsequently underwent chemotherapy treatment. Thereafter, she was diagnosed with esophageal cancer, which was the ultimate cause of her death at the age of 81. Plaintiffs allege that the cause of Jones's lymphoma and esophageal cancer was her use of Humira, and that she was unaware of the “serious side effects and dangerous properties” of the drug. Plaintiffs claim that Abbott failed to warn consumers, including Jones, of potential adverse side effects caused by Humira and that Abbott downplayed the risk of the drug's adverse side effects. Plaintiffs assert six causes of action: (1) strict liability—design defect; (2) strict liability—failure to warn; (3) breach of implied warranty; (4) negligence; (5) violation of the Tennessee Consumer Protection Act, Tenn. Code Ann. § 47-18-101 et seq.; and (6) wrongful death, Tenn. Code Ann. § 20-5-113 et seq.

The present cross-motions relate to the production of data stored in Abbott's Adverse Event Global Information System (“AEGIS”) database. The AEGIS database is an electronic database which contains records of adverse event reports involving all of Abbott's drugs, including Humira. Abbott collects adverse event data involving its drugs from worldwide sources (including spontaneous notifications, regulatory authorities, literature, and clinical trials) and maintains this information centrally in AEGIS. The data maintained in AEGIS is used to generate the adverse event reports to the Center for Drug Evaluation and Research, which maintains the adverse event reporting system for the FDA, as well as the adverse event reports to other foreign regulatory agencies. The data is also used to generate the Periodic Safety Update Reports for Humira. The AEGIS database operates based on the Adverse Reaction Information System (“ARISg”) software. ARISg is a graphical user interface based system that offers the ability to store and process information related to adverse event reports. According to Abbott, ARISg is a confidential and proprietary software system, and Abbott's ARISg software license does not permit the company to provide a copy of the software to any other user.

*2 Each adverse event report is coded in the AEGIS database by Abbott's Medical Safety Surveillance department using the Medical Dictionary for Regulatory Activities (“MedDRA”). Abbott uses a “Malignancies—Lymphoma” Company MedDRA Query to identify adverse event reports involving lymphomas, and a “Malignancies—narrow” Standardized MedDRA Query to identify adverse event reports involving other malignancies, excluding lymphomas and skin malignancies. Each record in the AEGIS database contains the medically-relevant data concerning the adverse event report, including, among other things, patient and healthcare provider identifiers, patient demographics, concomitant medications, event descriptions, dates of procedures, and dates of notification. Many of the fields in the database contain individually identifiable and confidential patient health information protected from disclosure under the Health Insurance Portability and Accountability Act of 1996 (“HIPAA”) and related regulations. As of March 31, 2011, Abbott's AEGIS database contained over 144,000 unique adverse events reports involving Humira and approximately 240,000 non-unique reports. This total includes approximately 45,000 medically confirmed reports completed since December 31, 2002. Of these 144,000 unique adverse event reports, approximately 490 involve lymphomas and 4,400 involve malignancies (including the 490 lymphomas).

During discovery, plaintiffs' prior counsel requested that Abbott produce its “complete files relating to all adverse reaction reports associated with the use of Humira, alone and/or in combination with Methotrexate” in plaintiffs' First Set of Requests for Production (served on November 20, 2007). In response, Abbott agreed to produce “documents obtained from its AEGIS database that identify, from the period following December 31, 2002, the adverse event reports associated with Humira that involve lymphoma or malignancy and, for the period prior to December 31, 2002, serious adverse event reports that include lymphomas and other malignancies.” On July 11, 2008, Abbott produced “copies of Suspect Adverse Reaction Reports [CIOMS Reports] and MedWatch Reports associated with Humira that involve lymphomas or other malignancies, with personal patient and physician identifying information redacted.” In total, approximately 4,400 adverse event reports (comprising over 8,800 pages) involving lymphomas or other malignancies were produced to the plaintiffs. In preparing the documents for production, Abbott conducted line-by-line redactions of individually identifiable patient information. It is undisputed that plaintiffs' prior counsel raised no objections to the scope or form of Abbott's adverse event reports production. Subsequently, plaintiffs' prior counsel was allowed to withdraw as counsel of record. Plaintiffs later retained new counsel, who filed their notice of appearance in February 2011.[1] At plaintiffs' new counsel's request, Abbott reproduced the 4,400 adverse event reports in text-searchable PDF format.

Plaintiffs now seek production of Abbott's entire AEGIS database as it relates to Humira, without limiting the adverse event reports to those only involving lymphomas and other malignancies. Given the licensing restrictions on Abbott's ARISg software, plaintiffs ask that Abbott extract the Humira data from AEGIS and produce the information in Microsoft Access format.[2]Alternatively, plaintiffs request that Abbott reproduce the 4,400 adverse events reports in Microsoft Access format.

Rule 26 of the Federal Rules of Civil Procedure provides that “[p]arties may obtain discovery regarding any nonprivileged matter that is relevant to any party's claim or defense,” and that “[r]elevant information need not be admissible at the trial if the discovery appears reasonably calculated to lead to the discovery of admissible evidence.” Fed. R. Civ. P. 26(b)(1). To establish relevance, “[t]he party seeking discovery must be able to articulate the possible linkage between the discovery sought and admissible evidence.” Allen v. Howmedica Leibinger, GmhH, 190 F.R.D. 518, 522 (W.D. Tenn. 1999). Here, the claims at issue relate to Jones developing lymphoma and esophageal cancer allegedly as a result of receiving Humira treatments and Abbott's failure to warn regarding these conditions. Thus, plaintiffs' discovery request seeking adverse event reports for Humira involving other adverse effects that have no association with lymphoma or malignancies does not meet the broad relevancy standards of Rule 26.

*3 At the motion hearings and in their briefs, plaintiffs argued that they need all adverse event reports for Humira (1) in order for their expert, Keith Altman, Ph.D., to perform a proportional reporting rate/ratio (“PRR”) analysis, and (2) to determine whether there were “safety signals” related to the use of Humira, such that Abbott should have engaged in further investigation into potential safety risks of the drug and that Abbott should have taken further action to warn based on those signals.[3]

First, as to plaintiffs' argument that they need certain critical data contained in all adverse event reports in order to conduct a PRR analysis, the court is aware of only two reported cases that have specifically addressed the relevancy and admissibility of PRR analyses in drug products liability cases.[4] In In re Meridia Prods. Liab. Litig., 328 F. Supp. 2d 791 (N.D. Ohio 2004), the plaintiffs alleged that their use of the diet drug Meridia (manufactured and distributed by Abbott) caused cardiovascular and cerebrovascular injuries, including heart attack, stroke, tachycardia, heart palpitations, chest pain, high blood pressure, hypertension, and death. Id. at 795-96. In their opposition to Abbott's motion for summary judgment, the plaintiffs offered a PRR analysis conducted by Altman. Altman looked at the number of all adverse event reports received for Meridia and calculated the percentage of those reports that were associated with cardiovascular events. Id. at 807. Altman then compared the percentages of reports from adverse cardiovascular events for Meridia and Xenical (a competing diet drug), and found that the incidences of cardiovascular injuries were more than twice as likely to be reported for Meridiathan Xenical. Id. Plaintiffs attempted to offer Altman's PRR analysis as evidence of general causation. Id. The court rejected the PRR analysis, observing that “proportional reporting rate analyses are incomplete and often misleading because they do not show the total distribution of reports.” Id. The court further stated that Altman's PRR analysis was “irrelevant” and “insufficient to create a genuine issue of material fact regarding general causation.” Id. at 808.

In In re Baycol Prods. Litig., 532 F. Supp. 2d 1029 (D. Minn. 2007), plaintiffs brought actions against Bayer Corporation relating to the drug Baycol, which was used to lower the lipid levels of individuals with high cholesterol with the goal of decreasing the risk of cardiac diseases. Baycolwas withdrawn from the market after thirty-one deaths in the United States were linked to its use. Id. at 1035. The court granted Bayer's Daubert motion to exclude plaintiffs' expert opinions based on adverse event reports, including the opinions of Dr. John Farquhar and Dr. Harland Austin, both of whom conducted a PRR analysis to support their expert opinions. Id. at 1039. The court reviewed the studies cited by the experts and found that “none stand for the proposition that [adverse event report] data can be used to conduct a comparative analysis of drug risks, or more specifically that Baycol is the most toxic statin.”[5] Id. at 1041.

*4 The court notes that in the FDA publication Guidance for Industry: Good Pharmacovigilance Practices and Pharmacoepidemiolgic Assessment (March 2005), the FDA identifies PRR as a “data mining method” that “may be worth considering.” Id. at 9. However, in that same guidance document, the FDA emphasizes that its guidance documents “do not establish legally enforceable responsibilities” and “should be viewed only as recommendations,” that the FDA “exercises caution when making such comparisons [between products], because voluntary adverse event reporting systems ... are subject to a variety of reporting biases,” and that “[u]se of data mining techniques is not a required part of signal identification or evaluation.” Id. at 1, 9. In sum, the court finds that the PRR analysis does not provide a basis for the plaintiffs to obtain discovery of the reports from the entire AEGIS database for Humira.[6]

Second, the plaintiffs assert that they should be permitted to obtain discovery on all adverse event reports for Humira because that discovery is relevant to determining whether there were safety signals related to the use of Humira, such that Abbott should have engaged in further investigation into potential safety risks of the drug and that Abbott should have taken further action to warn based on those signals. Plaintiffs rely on Rix v. Sanchez et al., a birth defect case in Alabama state court in which a Special Master recommended that Abbott be ordered to produce all adverse event reports from its AEGIS database with respect to the drug Depakote, regardless of whether the adverse event related to birth defects. (ECF No. 153-1, Ex. A.) The Special Master, however, gave no reason for why he believed non-birth defect adverse event reports were relevant and discoverable, other than stating “[a]fter much consideration, the undersigned recommends that Abbott be required to produce” all of the reports from the database. Moreover, the production in Rix involved approximately 24,000 adverse event reports, which is less than one-fifth of the number of reports at issue in the present case.[7]

Plaintiffs also rely on In re Neurontin, 612 F. Supp. 2d 116 (D. Mass. 2009), a case which involved allegations that Pfizer and Warner-Lambert unlawfully marketed its antiepileptic drug Neurontin for purposes for which it was not approved (so-called “off-label” uses), even though these drug companies knew that the drug caused behavioral disturbances, depression, and ultimately suicidal actions. Id. at 121-22. The district court ordered the defendants to produce its adverse event databases without limiting the reports to those involving psychiatric adverse events. (ECF No. 153-2, Ex. B.) The claims at issue in In re Neurontin and in the present case are materially different. In In re Neurontin, the drug at issue allegedly caused several different types of adverse events that could have been associated with suicide or other self-injurious behavior, such as “depression,” “confusion,” “anxiety,” “nervousness,” “thinking abnormally,” “hostility,” “overdose,” “personality disorder,” “agitation,” “hallucinations,” “psychosis,” “delirium,” “neurosis,” “paranoia,” “aggression,” “mood swings,” “apathy,” etc. Thus, under those circumstances a broader analysis of defendants' entire adverse event database was warranted. In the present case, however, the plaintiffs have not identified any other adverse events that could be related to or associated with lymphoma or other cancers, so as to justify a more expansive database production. Nor have the plaintiffs demonstrated with any supporting case authority how adverse events that are not associated with a given condition can give rise to a safety signal for that condition.

*5 Finally, plaintiffs request in the alternative that Abbott reproduce the 4,400 adverse events reports in Microsoft Access format. The court denies this request. Plaintiffs' prior counsel accepted Abbott's database production without objecting to the scope or form of the production.[8] After plaintiffs' new counsel appeared in the case, Abbott reproduced the adverse event reports in text-searchable PDF format at new counsel's request. Abbott has provided evidence that it would require considerable time and expense to reproduce the same information in Microsoft Access format. The court is not inclined to require Abbott to reproduce the same adverse event reports in a different format for a third time.

For the above reasons, the plaintiffs' motion to compel is DENIED and Abbott's motion for protective order is GRANTED.

IT IS SO ORDERED.

Plaintiffs' new counsel represents other plaintiffs in at least four cases filed against Abbott concerning Humira. Murthy v. Abbott (S.D. Tex. filed Jan. 1, 2011); Calisi v. Abbott (D. Mass. filed April 18, 2011); Pletan v. Abbott (Ill. Cir. Ct. filed Apr. 26, 2011); and Bixby v. Abbott (N.D. Ill. filed May 20, 2011).

Abbott produced its AEGIS data in Microsoft Access format in another case involving Depakote, a different Abbott drug. Rix v. Sanchez (Ala. Cir. Ct.)

As defined by the FDA in a guidance document, a “safety signal”:

In a third case, Smith v. Pfizer, Inc., No. 3:05-0444, 2010 WL 1963379 (M.D. Tenn. May 14, 2010), the court excluded the expert testimony of Dr. Cheryl Blume to the extent her testimony relied upon a chart created by plaintiff's counsel which used a PRR analysis. The court excluded Dr. Blume's testimony based on her admission that she had not verified plaintiff's counsel's PRR analysis and thus was not able to validate the analysis. Id. at *12-13. Altman was one of the attorneys who represented the plaintiff and presumably was involved in the PRR analysis.

The In re Baycol court did find, however, that “the AER data relevant to this case presented a very strong signal concerning Baycol and its association with rhabdomyolysis, and such evidence may be relevant at trial. It thus follows that Plaintiffs' expert may testify as to the existence of this signal.” Id. at 1043. There is no indication from the court's opinion that plaintiffs' experts' signaling testimony was based on other adverse event reports that were unrelated to rhabdomyolysis.

The court further notes that in order to conduct a PRR analysis, Altman must compare the reporting ratios of Humira to other drugs, which necessarily requires that Altman have access to the adverse event reports of these other drugs (which are not contained in Abbott's database). The FDA's Adverse Event Reporting System collects adverse event reports for all drugs from all manufacturers and other sources, and the plaintiffs have access to this database. Thus, to the extent the plaintiffs wish to pursue its PRR analysis, the FDA's database should provide the plaintiffs with the necessary data to conduct the analysis.

According to Abbott's Assistant Director of Global Medical Services—Data Management who was involved in the AEGIS data production in Rix, it took approximately ten weeks to complete the production of the database, which included the time for Abbott employees to extract the data from AEGIS and export it to a Microsoft Access format, and for the manual redaction of patient data to conform with HIPAA and other medical privacy laws and regulations.

Abbott's reliance on this court's Local Rule 26.1(e) is misplaced. Local Rule 26.1(e), which governs electronic discovery, went into effect on March 1, 2011. Thus, Local Rule 26.1(e)(6) does not govern the dispute surrounding the document production in July 2008.