The opinion of the court was delivered by: Wolfson, United States District Judge
Glossary of Abbreviations
AHI Amersham Health Inc. (U.S.-based Counterclaim Plaintiff)
ASD GEH Area Sales Director
AWC Adequate and well-controlled study
BDI Bracco Diagnostics Inc.
[witness] D Designated deposition testimony
[witness] Dec Designated declaration
CE Continuing Education for doctors, nurses and technicians
CIN Contrast Induced Nephropathy or renal damage caused by x-ray contrast medium
CM Contrast Medium or Contrast Media
CME Continuing Medical Education for doctors
CMS Centers for Medicare and Medicaid Services
CT Computer Tomography. A type of x-ray procedure where the CM is given by i.v. administration
CT DCAM Novation's DCAM for CT (i.e., x-ray) contrast media
CT DCAM Novation's DCAM for both CT (i.e., x-ray) and MR contrast media
C x Bracco's Proposed Post-Trial Conclusion Of Law at paragraph "x"
Dx : y Defendant's Trial Exhibit "x" at page "y" (where y is the last three numbers of a Bates number, if applicable)
DCAM Decision Criteria Award Matrix
DHRxns Delayed Hypersensitivity Reactions
Dual DCAM Novation's DCAM for a dual source award for both CT (i.e., x-ray) and MR contrast media
FDA United States Food and Drug Administration
GEH GEH Healthcare, which acquired the three named defendants, who in turn acquired Amersham and Nycomed
GPO Group Purchasing Organization
HOCM High Osmolar Contrast Medium
i.a. intra-arterial (form of administration directly into an artery)
i.v. Intra-venous (form of administration directly into a vein)
IOCM GEH's trademarked term, Isosmolar Contrast Medium
ITB Novation's June 14, 2004 "Invitation To Bid"
JACC Journal of the American College of Cardiology
LBB "Low Best Bid" or "Low Best Bidder"
LOCM Low Osmolar Contrast Medium
MA Meta-Analysis, a type of clinical study analysis
MACE Major Adverse Cardiac Events or Major Adverse Clinical Events, depending on the study design
MR DCAM Novation's DCAM for MR contrast media
MRI magnetic resonance imaging contrast media
NEJM New England Journal of Medicine
NFC Non-Financial Criteria
NQWMI Non-Q-wave Miocardial Infarction
RFA GEH's responses to Bracco's requests for admissions
RFP GPO Request For Proposal
OTSheet Omnipaque Toss Sheet
Px : y Plaintiff's Trial Exhibit "x" at page "y" (where y is the last three numbers of a Bates number, if applicable)
PCI percutaneous cardiac intervention
POA Plan of Attack or Plan of Action
PTCA Percutaneous Transluminal Coronary Angioplasty
SR Systematic Review (type of clinical study analysis)
TCT Transcatheter Cardiovascular Therapeutics (TCT) Scientific Symposium
TF Novation's ICM Task Force
URTBrochure Unchallenged Renal Tolerability Brochure
x T y Trial Transcript Volume "x" at page "y"
VVAT Visipaque Value Analyis Tool
Isovue Bracco x-ray contrast agent
ProHance Bracco MRI contrast agent
MultiHance Bracco MRI contrast agent
Visipaque GEH x-ray contrast agent
Omnipaque GEH x-ray contrast agent
Omniscan GEH MRI contrast agent
Optiray Tyco/Mallinckrodt x-ray contrast agent
Hexabrix Tyco/Mallinckrodt x-ray contrast agent
This matter comes before the Court upon a Complaint brought by Plaintiff Bracco Diagnostics Inc. (referred to herein as "Bracco") against Defendants Amersham Health Inc., Amersham Health AS, and Amersham PLC (collectively referred to herein as "GEH") for alleged false advertising in violation of the Lanham Act. In response, GEH filed a Counterclaim against Bracco for alleged false advertising of its own line of products. Bracco and GEH have competing product lines in the contrast medium healthcare industry. The crux of Bracco's case is that GEH has falsely advertised the superiority of its product, Visipaque, over Bracco's product, Isovue. The Court conducted a thirty-nine day bench trial with numerous experts*fn1 and witnesses testifying as to each party's product lines and the underlying clinical studies upon which GEH and Bracco have based their advertising campaigns.
In light of the evidence presented at trial, the Court concludes that GEH did promote false messages which were sufficient in number to constitute actionable commercial advertisements or promotions under the Lanham Act, however the Court finds that Bracco has failed to establish a causal nexus between GEH's false advertisements and Bracco's alleged lost profit damages. In that regard, the Court determines that the greater number of GEH's advertisements were in fact true and based on reliable scientific studies. The messages that the Court finds false are those that extrapolate beyond the studies' results. In connection with Bracco's claim, the Court finds that an injunction and damages for post and future corrective advertising are appropriate remedies to prevent future violations of the Lanham Act. As to GEH's counterclaim, GEH dismissed its claim for damages and Bracco has stipulated that it no longer uses the offending advertisements. Thus, although the Court finds that certain of Bracco's ads were false, nonetheless, an injunction is not appropriate in this case. In addition, the Court imposes an alternative dispute mechanism applicable to both parties for safeguarding against any future false advertisements.
A. Parties and Product Lines
GEH and Bracco market and sell x-ray contrast media ("CM") in the United States. CM are classified by osmolality. HOCM (high osmolar CM) have osmolalities of greater than 1500 mOsm/kg. LOCM (low osmolar CM) have osmolalities between 600 and 850 and include Omnipaque (iohexol), Isovue (iopamidol), Hexabrix (ioxaglate), Ultravist (iopromide), Iomeron (iomeprol), and Optiray (ioversol). The osmolality of blood is approximately 290 mOsm/kg. Both GEH and Bracco market LOCM; GEH sells Omnipaque and Bracco sells Isovue. In addition, GEH also markets a product called Visipaque (iodixanol) which it classifies as iso-osmolar or isotonic, (i.e. - its osmolality equals blood). Visipaque is referred to in various medical literature as an IOCM (iso-osmolar CM). Part of GEH's advertising campaign is that its iso-osmolar CM performs better than LOCM. Visipaque was introduced in 1996, ten years after Omnipaque and Isovue were marketed and is the only "IOCM" available in the U.S.
On December 16, 2003, Bracco filed a four count Complaint in the District of New Jersey against GEH alleging: (1) dissemination of false and misleading advertisements in violation of Section 43(a) of the Lanham Act; and (2) N.J.S.A. 56:4-1, et seq.; (3) violations of the common law of unfair competition; and (4) negligent misrepresentations. GEH filed an Answer and two counterclaims against Bracco alleging: (1) dissemination into commerce of allegedly false and misleading statements concerning the relative safety of Omnipaque, Visipaque, and Isovue in violation of Section 43(a) of the Lanham Act; and (2) N.J.S.A. 56:4-1, et seq. GEH's counterclaim was filed against Bracco and its foreign affiliates, Bracco S.p.A. and Bracco Imaging S.p.A. However, pursuant to an Order entered on September 7, 2004, GEH's counterclaim against Bracco's foreign affiliates was dismissed for lack of personal jurisdiction. Motions for Summary Judgment were denied by the Court, after which, a thirty-nine day bench trial was conducted between the period of May 7, 2007 and December 2007, followed by further written submissions. The Court held a hearing on May 15, 2008, wherein the Court resolved evidentiary objections regarding the admission of disputed exhibits. Subsequently, the parties submitted proposed Findings of Fact and Conclusions of Law, which were supplemented by Reply briefs and additional Daubert briefs to exclude expert testimony proffered by both sides.
A. Bracco's Case in Chief
As set forth below, the Court finds that GEH advertises and promotes Visipaque with establishment claims asserting that studies show it is superior in several ways, including renal and cardiovascular safety, pain, heat and discomfort. The Court further finds that: the spike in Visipaque sales that started in 2003 was primarily due to the publication of the NEPHRIC study; GEH's advertising of NEPHRIC through true renal ads and promotions also contributed significantly to GEH's success with Visipaque; only a fraction of GEH's ads were false; while these false ads were sufficient in number to constitute actionable promotions under the Lanham Act, they were not the cause of GPO contracts being awarded to GEH. In addition, the Court finds that the limited false ads disseminated by GEH were not willfully false because GEH relied on scientific studies, which have not been disproved, and that GEH had a protocol in place for approving advertisements that attempted to ensure against falsity.
1. GEH's Establishment Claims Of Renal Superiority
In late 2002 to early 2003, GEH focused its ads and promotions on renal establishment claims based on the NEPHRIC study; GEH claimed that Visipaque had superior renal safety over competitor drugs or LOCM. (See, e.g., P1672 ("[w]e will begin to shift our focus from a Excellent Patient Comfort/Cardiac Safety message to the prime message being Excellent Renal Tolerance"), P106, P849:932, P1269:219, P1147:261, P1265:203, P1266:208; 13 T 72, 16 T 95, 6 T 58-79,102-104, 17 T 69-70). These claims are of two types: the data (from NEPHRIC or other studies) show (a) Visipaque is superior to a LOCM or all LOCM or (b) Visipaque is as good as or better than LOCM with pretreatments.*fn2 GEH uses the term LOCM to obscure the fact that its own drug, Omnipaque, was the comparator in the NEPHRIC study, and to thereby lessen any impact on Omnipaque, specifically, and to generalize the results to all LOCM, including Isovue. (P1534, P1535, P1519, P1523; 13 T 62-63).
GEH designed and then planned to disseminate the claims through multiple promotional channels (print media, websites, GEH representatives, medical doctors and CME's). (8 T 4-15,81- 129, 135-37, 17 T 53,64-68; P869 ('03), P849 ('04), P2098 ('05)).*fn3 GEH determined that the claims were the most effective way to convert sales based on its experience, (e.g., 7 T 178), and marketing research (P196, P696, P1400, P1700, P1716:739, P1742, P2038; 13 T 27-28).
The GEH representatives were instructed in Plans of Action ("POA") (e.g., P2101:559 (e.g., "less incidence of CIN")), memoranda (e.g., P102 ("top 3 messages"), P104, P353, P398:394, P639:552 ("(CIN) in high-risk patients was 11X less likely . than with LOCM"), P640, P661, P662, P696, P772, P1832:194, P2027, P4249; D790), training (e.g., P651 (e.g., "Visipaque is clinically proven to be . safer for high risk patients"), P1136:370 ("safer")) and Medical Bulletins (e.g., P402:563 (e.g., "NEPHRIC data clearly demonstrate . a significantly better renal safety profile than a traditional [LOCM], such as iohexol, in at-risk patients"), P538:888, P798:078) to disseminate the claims. (E.g., P85, P632, P774, P1008, P1012, P1021, P1080, P1082, P1136, P1178, P1373:182, P1561, P1572:893, P1681, P1699, P1721, P2099, P2100, P2101, P3708; 8 T 68-81,129-135,148-80, 9 T 5-53,65-80,88-102).*fn4
The claims were then disseminated nationwide using print media, GEH representatives and continuing medical education ("CME") presentations. The print media (e.g., brochures, websites, presentations, articles) with extracted messages that were identified in Bracco's pretrial brief and discovery responses and addressed at trial include:*fn5
· Press releases on its website: "The NEPHRIC data clearly demonstrate that VisipaqueTM offers a significantly better renal safety profile than traditional low osmolar non-ionic contrast media in at-risk patients . . . . We believe that the data strongly support VisipaqueTM as the agent of choice for these patient groups." P2449:379, P69:915, P254:863, P772:340, P1448:898, P4149:p2; 7 T 68-69.
· Computer Tomography (CT) brochures: "Nonionic Dimer Provides Lower Osmolality, Reduced . CIN" "CIN": "Nonionic Dimer": "9" ; "Nonionic Monomer": "8". P410:965, 3649:408, 3649A:408; D2324:117.
· Novation presentation: "Isosmolar VISIPAQUE. . . Demonstrated to significantly reduce incidence of Contrast-induced Nephropathy (CIN)." P2161:391. GEH rep efforts included the delivery of the claims and print media in face-to-face detailing of administrators, technicians, nurses and doctors, for which records were presented at trial from GEH's sales call record system, emails and memoranda, e.g."*fn6
· Sales Calls Records: "Discussed patient types that would benefit from Visipaque usage over Isovue. Re-affirmed with Nephric study." P2312:A637284, P4049:A637284. "Discussed having hospital start using Visipaque for high-risk patients in CT. Detailed Nephric study and core visual aid to support benefits of isosmolar Visipaque v. Isovue." P2312:A637355, P4049:A637355. "[C]linicaI studies, nephric etc show less risk nephrotox vs ... Isovue for [high risk] pts ...." P2312:A659673, P4049:A659673. "reviewed why Visi. is the best for kidneys." P2312:A670058, P4049:A670058. "Approached dr. with nephric focus and differentiating vis from locm class with regards to osmoality. Reminded dr. that patients are 11 times likely to have CIN with the locm class than visi." P3682:Omni/3727, 4049:Omni/3727.*fn7 *fn8
· Consorta detailing: "Baluchi [from Consorta] asked about Isovue with respect to Omnipaque/LOCM as positioned in NEPHRIC. We made the point that Omnipaque represents a LOCM (gold standard) and confirmed his key take away that it is isosmolar versus low-osmolar that was studied, not necessarily Visipaque v. Omnipaque." P682:286. "I.discussed the attributes of Isosmolar Visipaque including it's impact on CIN -- a clinical issue just coming to light; it's elimination of costly drug therapies (fenladopan) to prevent CIN with std LOCM". P793:514.
· HPG detailing: "Ami presented the Nephric data to Lew and he was very interested in the info. He told her that one criticism of this paper was that it was not a head to head with Isovue. Ami showed him the list of references that prove the incidence of CIN with Isovue is equal to that of Omnipaque so it is reasonable to draw a correlation that the results of Nephric would be the same if Isovue had been used." P663:942 (emphasis added).
The Court also finds that GEH-sponsored CME presentations for doctors (e.g., P849:946) delivered IOCM versus LOCM claims:*fn9
· 2004 CME On CIN: "[R]ecent controlled trials have shown that non-ionic Isosmolar contrast agents are superior to low-osmolar agents in preventing CIN." P4251:210 (emphasis added). "The use of iodixanol in at-risk patients appears to minimize the risk of CIN even without additional pharmacological prophylaxis." P4251:212.
At trial, GEH's sales rep, Mr. Joseph Murray, confirmed delivering these claims through the print media (e.g., press releases and articles) and CME-type presentations to customers in order to convert sales to Visipaque. (E.g., 16 T 31-49, 56-58, 81-88, 97-114; 17 T 49-51,64-132).
2. The Falsity Of GEH's Renal Establishment/Superiority Claims
The Court finds that certain assertions made by GEH were supported by the studies' (NEPHRIC and Chalmers) conclusions (e.g. - Visipaque may be renally superior over a LOCM - -Omnipaque), while others were not (i.e. - Visipaque may perform better than LOCM with prophylactics and Visipaque is renally superior over all LOCM). Bracco asserts that GEH's representations are false and misleading because: (a) the NEPHRIC study omitted results and has flaws that contradict GEH's claims (Pl.'s FOF ¶¶ 11-15,17); (b) the studies (NEPHRIC and Chalmers) cannot reliably (Pl.'s FOF ¶¶ 11-13) support a conclusion of Visipaque superiority over all LOCM; and (c) the weight of the clinical evidence is that Visipaque is not superior to all LOCM as a group or to Isovue individually (Pl.'s FOF ¶¶ 11-15,17). For example, Bracco asserts that results of studies done with intra-arterial ("i.a.") use are not reliable enough to predict with reasonable certainty intravenous ("i.v.") results, (see P1733)*fn10 *fn11 ; furthermore, Bracco points out that no studies compare Visipaque to a LOCM combined with pretreatments or even head-to-head with multiple LOCM.
The Court finds: (1) while there were flaws in NEPHRIC, those flaws do not vitiate its results; (2) the NEPHRIC and Chalmers studies are not unreliable in their conclusions; and (3) it has not been established by the weight of clinical evidence that Visipaque is superior to all LOCM as a group or to Isovue individually. Although there has never been one adequate and well-controlled ("AWC") clinical study (let alone two, done the same way with the same drug) showing that Visipaque is superior to any LOCM (even Omnipaque), with or without pretreatments,*fn12 *fn13 the Lanham Act does not demand such a rigorous finding. Nonetheless, although not dispositive, the FDA agrees with the Court's findings in numerous letters sent to GEH, including one as recent as March 21, 2005, where it states that the results of the NEPHRIC study cannot be extrapolated to CM other than Omnipaque in GEH advertising. (P1894.)
To lay the foundation for Bracco's claims, and GEH's defenses, the parties first presented background clinical evidence at trial. The following pertains to such evidence: randomization in a clinical trial increases its reliability;*fn14 a primary endpoint is a clinically relevant endpoint around which a study is designed;*fn15 studies may also have secondary endpoints, which are of interest but are deemed to be of less importance to the study investigators;*fn16 a MA is a statistical combination of results from multiple studies;*fn17 a p-value is a statistical measure that provides a general estimate of the probability that two tested clinical strategies are different;*fn18 furthermore, the probability that two treatments are different can be roughly estimated as 1 minus the p-value.*fn19
After laying a foundation for generalized information regarding the interpretation of medical studies, the parties presented specific clinical evidence in connection with GEH's claim that Visipaque is less nephrotoxic than other LOCM. Changes in renal function are commonly measured by serum creatinine ("SCr").*fn20 Dr. Peter Aspelin, an M.D, Ph.D., a professor of medicine in Stockholm,*fn21 and the author of NEPHRIC, testified that CIN is commonly defined as an increase in SCr up to 3 days of 0.5 mg/dL, 25%, or both,*fn22 and that rises in SCr after 3 days may be due to factors other than administration of CM.*fn23 Bracco disputes this definition of CIN; it contends that rises in SCr after three days are significant. The Court need not determine the clinical significance of CIN after three days because while I find that such data is relevant to the weight given to a study's conclusions, here I find that the use of either definition would not make the underlying study unreliable.
Nonetheless, it is undisputed that patients with both renal insufficiency (RI) and diabetes are at a higher risk for developing CIN, than patients with only RI or only diabetes.*fn24 In addition, Dr. Harold I. Feldman, an expert in internal medicine and nephrology, proffered by GEH, testified that patients with only diabetes have a lower risk than patients with only RI*fn25 and that greater contrast volume increases a patient's risk of CIN,*fn26 while N-acetylcysteine (NAC) or sodium bicarbonate may reduce CIN.*fn27 Furthermore, it was established at trial, through expert testimony and exhibits, that there is a scientifically reasonable and widely held belief in the medical community that LOCM are less nephrotoxic than HOCM due to the reduced osmolality of LOCM.*fn28 This belief is also shared by Bracco.*fn29 As to LOCM, Dr. Feldman testified at trial, that as of February 2003, there was little evidence of differences in nephrotoxicity between Omnipaque and Isovue.*fn30 Bracco's Dr. Spinazzi testified, and published to his peers, that as of the date NEPHRIC was published, it was believed that all nonionic LOCM performed similarly even though he qualified the testimony as not being supported by "evidence in the field."*fn31 Notwithstanding the prevailing belief in the field, the FDA found it to be misleading for GEH to advertise, based on the NEPHRIC results (comparing Visipaque to one LOCM, Omnipaque), that "Visipaque is safer than other conventional non-ionic contrast media." (P1894). This implies that the FDA did not believe that there was sufficient support to conclude that all LOCM perform similarly. Thus, the FDA questioned, based on the NEPHRIC study, claims of Visipaque superiority over all LOCM as opposed to merely the LOCM tested in NEPHRIC.
GEH relies on several scientific studies to support its claim that Visipaque is less nephrotoxic than other LOCM, and hence has a better renal safety profile, but primarily, GEH relies on the Chalmers and NEPHRIC studies. Chalmers, first published in 1999, was a randomized head-to-head trial of Visipaque (iodixanol) and Omnipaque (iohexol) adminstered to patients with RI.*fn32 It showed Visipaque to be less nephrotoxic than Omnipaque.*fn33 NEPHRIC was a double-blind, randomized, multi-center, head-to-head trial comparing the nephrotoxicity of Visipaque and Omnipaque in patients with RI and diabetes.*fn34 Dr. Aspelin was the principal investigator ("PI") for NEPHRIC.*fn35 He has nearly 200 published papers and is a peer-reviewer for several journals.*fn36 Dr. Aspelin was not a consultant for GEH and was not paid for his work on the NEPHRIC study, however, he did receive input from GEH regarding the formulation of the language used in his conclusions in NEPHRIC and indeed, GEH was the financial sponsor for the study.*fn37 NEPHRIC reported that Visipaque was less nephrotoxic and caused 11 times less CIN than the studied CM, Omnipaque.*fn38
Dr. Aspelin had overall responsibility for, and final authority over, the content of NEPHRIC.*fn39 The other NEPHRIC authors, including Dr. Berg (a renal physiology expert) contributed to and approved the contents of the article.*fn40 While the results of the NEPHRIC study, which was a head-to-head comparison of Omnipaque and Visipaque, provide reasonable scientific support for the claim that Visipaque performs better than Omnipaque in high risk patients, it does not support the claim that Visipaque performs better than all LOCM for that patient group.*fn41 See infra pp.132-39. Other studies in the field and referred to at trial will be reviewed below.
RECOVER was a randomized blinded head-to-head clinical trial, published in 2006, comparing the nephrotoxicity of Visipaque and Hexabrix (ionic low osmolar CM) in patients with RI.*fn42 It showed that Visipaque was less nephrotoxic and caused less CIN than Hexabrix.*fn43 Neither party, Bracco nor GEH, was involved in the study or publication of RECOVER.*fn44 Bracco alleges that RECOVER is unreliable due to a discrepancy with an earlier published abstract. However, the RECOVER authors explained in a published letter to the editors that the published results in the Journal of the American College of Cardiology (herein "JACC") were accurate, and that the results reported in the earlier abstract were based on preliminary data.*fn45 Accordingly, the Court finds this study to support the contention that Visipaque is less nephrotoxic than Hexabrix.
Jingwei was a head-to-head clinical trial, also published in 2006, comparing the nephrotoxicity of Visipaque and Isovue in patients undergoing percutaneous cardiac intervention ("PCI").*fn46 It showed that Visipaque caused smaller SCr elevations.*fn47 Nonetheless, there was no clinically significant difference in the occurrence of CIN. The record does not indicate any involvement by GEH or Bracco in the Jingwei study.
c. The McCullough Meta-Analysis ("MA")
The McCullough Meta-Analysis ("MA"), published in 2006, used patient level data from head-to-head randomized intra-arterial clinical trials gathered from a GEH patient database.*fn48 Dr. Peter McCullough, a consultant for GEH, and co-authors had control over the MA.*fn49 McCullough found that Visipaque was less nephrotoxic than LOCM in: (i) all risk level patients; (ii) patients with RI; and (iii) patients with RI and diabetes.*fn50 Bracco's expert, Dr. Lee Jen Wei, re-analyzed the MA and confirmed that Visipaque causes less CIN than the LOCM analyzed in the study using the CIN definition chosen by McCullough.*fn51 However, as Wei cogently and significantly pointed out during his testimony, the McCullough MA was comprised of 16 studies, 9 of which were Omnipaque and 7 of which were non-Omnipaque LOCM. Isovue only represented 1 of the 16 studies. Dr. Wei concluded through statistical analysis, and the Court finds his testimony credible and persuasive, that when the non-Omnipaque studies were compared to Visipaque there was no statistically significant difference in CIN and that the nine Omnipaque studies skewed the results. Therefore, the Court does not find the McCullough Meta Analysis Study reliable for the claim that Visipaque ("IOCM") causes less CIN than all LOCM.
The VALOR study, sponsored by GEH, was an head-to-head clinical trial comparing the nephrotoxicity of Visipaque and Optiray, and allowed for the discretionary administration of Nacetylcysteine ("NAC").*fn52 Following a protocol specified interim analysis, it was determined that patients receiving NAC had more CIN.*fn53 Thus, enrollment was suspended and then terminated.*fn54 A manuscript reporting on VALOR was submitted for publication in 2007.*fn55 The incidence of CIN was lower with Visipaque than Optiray, and Visipaque caused a lower maximum percentage change in SCr from the baseline.*fn56 However, the study concluded that there was no statistically significant difference in the incidence of CIN between the two CM tested; therefore no reliable conclusions can be drawn from this study as to Visipaque's renal superiority.
IMPACT was a study sponsored by Bracco and completed in 2006.*fn57 It was not a prospective study, but combined secondary data from two previously completed Bracco studies, INVICTA and VIRPACT, that were designed to study image quality, not CIN.*fn58 The post-hoc combination of data from two studies was not disclosed in the manuscript and is not an accepted practice in the scientific community.*fn59 Although Dr. Feldman testified that IMPACT does not contradict the conclusions of NEPHRIC because of the different patient sample groups, IMPACT does come to the conclusion that Visipaque and Isovue performed similarly and had similar renal safety profiles in patients at elevated risk for CIN.*fn60
CARE was another Bracco sponsored study*fn61 comparing Visipaque and Isovue. CARE was published in May 2007, and, prior to that, was not available to GEH.*fn62 All patients received sodium bicarbonate according to a protocol from the Merten study, which showed that sodium bicarbonate reduced CIN when used with Isovue.*fn63 Merten concluded that sodium bicarbonate inhibited the negative effects of hyperosmolar stress caused by LOCM such as Isovue.*fn64 At the time the CARE protocol was finalized, there was no significant evidence that bicarbonate was beneficial when used with an iso-osmolar agent like Visipaque.*fn65 Dr. Feldman testified that CARE does not speak to the relative nephrotoxicity of Visipaque and Isovue without use of bicarbonate and does not contradict Chalmers, NEPHRIC, RECOVER, Jingwei, or the McCullough MA.*fn66 That is true, however, the Court finds this study's findings probative because its results indicate no statistically signifcant difference in CIN between Isovue with sodium bicarbonate and Visipaque with sodium bicarbonate.
g. The Sharma Pooled Analysis, Solomon Systematic Review and Solomon/DuMouchel Articles are Biased and Methodologically Flawed
The Sharma Pooled Analysis (D 262A) was drafted in-house by Bracco and was based upon a prior article by Dr. Alberto Spinazzi, Bracco's senior vice-president responsible for medical and regulatory programs.*fn67 Bracco performed the statistical analysis*fn68 and paid Dr. Samin K. Sharma, a doctor at Mount Sinai School of Medicine, $50,000 for his costs associated with the article.*fn69
Bracco also drafted the Solomon Systematic Review (D 107) and paid Dr. Richard J. Solomon, a specialist in internal medicine and nephrology and an expert proffered by Bracco, $30,000 for his involvement.*fn70 Together with Dr. Solomon, Bracco published an abstract of its review, but without data from Chalmers to "strengthen the argument" of equivalency between Isovue and Visipaque.*fn71
Bracco and Dr. Spinazzi were intimately involved in drafting the Solomon/DuMouchel article (D 222).*fn72 Because of methodological flaws, the Court finds that no reasonable conclusions on the relative nephrotoxicity of Visipaque, Omnipaque and Isovue can be drawn from the Sharma, Solomon Systematic Review or Solomon/DuMouchel articles.*fn73 Bracco's expert, Dr. Isabel Elaine Allen, attempted to validate the Solomon Systematic Review, but her analysis was plagued by errors.*fn74 The reported CIN rates in both Solomon and Sharma were in fact lower for Visipaque than for both Isovue*fn75 and Omnipaque,*fn76 although the difference in the rate of CIN between Visipaque and Isovue was not statistically significant. There was a statistically significant difference in the rate of CIN between Visipaque and Omnipaque, and furthermore a statistically significant difference between Isovue and Omnipaque; Visipaque and Isovue performed better than Omnipaque overall.
The NEPHRIC study, reported in the NEJM (P2467), compared Visipaque and Omnipaque head-to-head, but stated in its conclusion that "[n]ephropathy induced by contrast medium may be less likely to develop in high-risk patients when iodixanol [(an iso-osmolar contrast medium)] is used rather than a low-osmolar, nonionic contrast medium." Bracco assails the reliability of the NEPHRIC study by contending: (1) it was not designed to test whether osmolality is responsible for CIN (e.g., 20 T 6) and therefore cannot support the conclusion that Visipaque performs better than all LOCM in connection with renal function and CIN; (2) it has never been repeated in an AWC study; (3) it does not provide any support for the conclusion that Visipaque is as good as or better than LOCM with prophylactics;*fn77 and (4) it does not represent the weight of scientific evidence. (P2467; 3 T 89-90).
In addition, Bracco avers, through the testimony of Dr. Solomon, that Table IV of the NEPHRIC article, which purports to present results from other studies, is inaccurate and misleading because it incorrectly reports the results of those studies. (3 T 126-31; P3148, 37, 2053, 2386, 2390). Bracco also alleges that Table IV is inaccurate and misleading because it does not report the allegedly contradictory results of GEH's NEPHRIC II study;*fn78 but NEPHRIC II was not completed prior to the publication of the original NEPHRIC article and therefore could not have an impact on the reliability of the Table IV charts when published. Additionally, any such allegations as to the results of the NEPHRIC II study are speculative and it is improper for the Court to draw any inferences in the absence of its production.
According to Bracco there are several additional flaws inherent in NEPHRIC which make it unreliable:
· Primary outcome flaw. The record indicates that Nephric's primary endpoint -- mean peak change in serum creatinine -- is not a reliable metric, although it is used in the article and GEH's ads (e.g., the 11 times better assertion). (33 T 207) (Feldman). P4288:437 ("unknown clinical significance"), P200 (p3, FDA rejects Nephric's mean peak change endpoint); see also P1540 (definition of clinically significant)
· Omitting of key results. The NEJM article does not report the 25% rise in serum-creatinine results (P44, P4144, P1887; D2039T), which GEH added to make the study more comparable to the Chalmers study and to provide a more rigorous test for CIN. Id.; 3 T 90-99. Instead, the article falsely states that the secondary endpoints were significantly better. (P2467:913). GEH's marketing director was aware of this unreported data (e.g., P1951:131 (declining to provide the 25% results in Spain)) and he permitted the article to falsely report that all of the secondary outcomes showed a statistically significant (defined as p<0.05) difference="" between="" visipaque="" and="">0.05)>*fn79
· Hydration flaws. Inadequate hydration was described by a GEH doctor as one of the "greatest weaknesses of the study" but it was not acknowledged in the article or any ad. (P530:148, 20 T 25; 3 T 101-105).
· Baseline and other population flaws. The patients in the Omnipaque group had worse baseline values (P207-08, P979, P1887, P48, P49; 3 T 119-120; 20 T 35), which greatly increases the chance of getting CIN. Id. The patient groups also had other differences that were never analyzed together. Id.; 3 T 105-109,112-117. Furthermore, since there was no standard hydration, and hydration changes baseline values to an unknown extent, there is no way to know the correct values and thus there was no way to accurately calculate mean peak change or CIN. P823, P4250; 20 T 39.
· Improper manipulation. While the study was ongoing, and in violation of the protocol and proper practice, GEH took secret and forbidden peeks at the data looking for trends, and even changed the study endpoints and stopped the study early in response.*fn80 E.g., P562:821 (found "Mean of max day 2 and 3: 20% (25% for Omnipaque and 15% for Visipaque)" to compare treatment arms before study completion), 1882:966 ("The statistician thought the data [from the two treatment arms] looked equal in both CrCl groups."); 11 T 20-26, 43-44, 71-75; D2039T ("pretend", "plausibility will not increase"), 2339, 2440. These types of unplanned interim analyses (defined by D2340:42 and D2339:34 as any comparison of treatment arms prior to completion) and secret attempt to "fix" the study midstream makes the trial non-prospective. P830; 20 T 28-33, 11 T 14-44, 70-71; P130. When a NEJM reviewer asked whether there was an interim analysis (P1869, 46), GEH and the authors replied there was none, and then amended the article to falsely say there was none. (P68; 20 T 17-18).
· Hidden duration of diabetes flaw. Duration of diabetes may be a predictor of CIN (P 4377:29K; 34 T 57-62). GEH found that the statistically significant higher duration of diabetes in the Omnipaque patients may explain the results, independent of the CM, making the conclusion of the article unreliable. GEH did not reveal these results to the public. (P49, P967, P4364, P4364T, P4365, P4365T; 20 T 33-34).
· Misrepresentative conclusion and manuscript. GEH's marketing director provided input to the NEJM article to try to make it misleading, and then celebrated the final version's obscuring of the limitation of the results of the study to Omnipaque and its overly broad and unsupportable conclusion. P1519, P1534, P1535, P4210 (admitted to only show input), P1532, P4208, P1672, P1873 (conclusion same as TCT abstract), P1876, P4208; 6 T 87-89,
13 T 62-71; see also P480.
(Pl.'s FOF ¶ 13). Taking into consideration that the study compares one LOCM (Omnipaque) with Visipaque, nonetheless, the Court does not find that the study's results are vitiated by the flaws identified by Bracco. In addition, the NEPHRIC study uses conservative language in its conclusion (e.g. - use of Visipaque "may" cause less CIN than "a" LOCM), which does not render the study unreliable merely because it only compared Visipaque and Omnipaque. As it pertains to GEH's advertising, however, the non-definitive language used in the NEPHRIC conclusion permits GEH to use it for the contention that Visipaque may be renally better than a LOCM (which in the context of the article, means better than Omnipaque), but only if GEH plainly identifies, in same size print (and not in footnoted material), that Omnipaque was the only LOCM compared and that the NEPHRIC findings are limited to the studied CM.*fn81 The NEPHRIC article cannot be used to claim that Visipaque has a superior renal safety profile to all or any other LOCM. Further, in regard to the claim that "Visipaque is as good as or better than LOCM with prophylactics," the Court finds that the results of the NEPHRIC study cannot reliably support such a conclusion because using a LOCM with prophylactics was not part of the results of the study and was only concluded through a MA.
Moreover, Bracco contends that reliance on NEPHRIC is unreasonable because all other reliable clinical trials, reviews and MAs demonstrate no basis for a superiority claim of Visipaque over Isovue. Indeed, Bracco contends that all reported AWC clinical evidence and properly conducted MAs (e.g., Dr. Wei's unrebutted MA of GEH data) show no statistically significant difference between Visipaque over either Isovue or Optiray, whether given i.a. or i.v.*fn82 (1 T -4 T,11- 12). Bracco also contends that for seven years prior to the NEPHRIC campaign, no doctors had ever observed that Visipaque caused less CIN, Id., and that GEH's internal hidden data, never mentioned in its ads, also show no renal superiority for Visipaque.*fn83 GEH reported to its representatives that the studies provide valid and reliable information and that anecdotal experiences are not reliable in making CIN comparisons. (P260:413; 13 T 72-74).
In further support of its claim for false advertising, Bracco also relies on FDA findings which declined to approve a renal superiority claim for Visipaque. The FDA has repeatedly found (e.g., in the years 1996, 2001, 2005) that there is inadequate support to make renal superiority claims for Visipaque. (E.g., P596, P457, P585, P200, P816, P1894; 7 T 54-56; 14 T 41-57; 15 T 25-37, 38-39, 43-48). Furthermore, in 2001, GEH submitted the proposed NEPHRIC study to the FDA with proposed superiority claims. (P200, P199, P4205, P816, P556, P818, P1542, P264B, P271, P1670, P1543, P810, P972; 20 T 48-57; 14 T 57-62). The FDA again rejected GEH's proposed claims, requiring AWC clinical trials and rejecting NEPHRIC as an AWC clinical trial (e.g., "the current [NEPHRIC] protocol contained a number of sources of variability, which may confound the ability to clearly determine the effects of the drug on renal function"). (P200, 1542; 20 T 48-51).
As part of GEH's rebuttal to the assertion that its representations constituted false advertising, GEH relied on four studies: Chalmers, NEPHRIC, RECOVER and the McCullough MA. Bracco contends that flaws in these studies vitiate their results as follows: (a) Chalmers was not AWC (small and unblinded), showed no significant difference (there was total agreement that the 10% test is irrelevant), and even the authors concluded it was weak (3 T); (b) NEPHRIC is unreliable; (c) RECOVER only involves Hexabrix (an ionic agent), it showed no differences in certain CIN measures and it is unreliable (3 T; Solomon Dec; D1990; P3823); and (d) the McCullough MA is of limited value as demonstrated by Dr. Wei's unrebutted testimony (11 T -12 T) that the McCullough MA results were mostly due to Omnipaque (and not Isovue).*fn84 The Court finds that these studies' conclusions do not establish the proposition that Visipaque has renal superiority to all LOCM.
Turning specifically to the NEPHRIC study, despite certain flaws, there were significant reliable aspects. The Visipaque and Omnipaque groups in NEPHRIC were demographically comparable.*fn85 The requisite number of patients pursuant to the protocol were included.*fn86 All patients had RI and met the inclusion criteria.*fn87 Also, the Court is not convinced that the use of NAC in 11 patients affected the viability of the NEPHRIC results.*fn88
Furthermore, contrary to Bracco's assertion, no interim analysis, as that term is understood and defined by the scientific community, was performed during NEPHRIC.*fn89 ICH and FDA Guidelines for clinical trials, adopted by Bracco's expert Dr. Sanford Bolton (an expert in pharmaceutics, physical pharmacy and bio-statistics, as authoritative),*fn90 define an interim analysis as the unblinded comparison of treatment results.*fn91 During NEPHRIC, the results were not unblinded and treatment results were not compared.*fn92 ICH and FDA Guidelines acknowledge that a sponsor may, without impacting a study's validity, monitor the success of planned accrual targets and the appropriateness of design assumptions.*fn93 Indeed, it is the sponsor's responsibility to do so.*fn94 Thus, GEH's monitoring of patient enrollment, sample size assumptions and overall (not separated into two treatment groups) SCr changes (e.g., Ps 562, 1882, 1883, 1884, 1885 1890, 1891), do not constitute interim analyses.*fn95
As to Bracco's claim that GEH influenced the wording of the NEPHRIC study conclusion, the Court finds that GEH did have input. Nonetheless, Dr. Aspelin's first draft dated March 6, 2002, which was authored before GEH offered comments and before presentation of an abstract at the TCT conference, also included a conclusion applying NEPHRIC results to the class of LOCM.*fn96*fn97
Further, Dr. Aspelin testified that he, his co-authors, and the New England Journal of Medicine editors believed in the scientific reasonableness of the conclusion.*fn98 All of this lends support to the reliability of the article, but combined with the chronic rejection by the FDA of its use for superiority advertising and the fact that the NEPHRIC article only compares one LOCM to Visipaque, it cannot be concluded from the study and the article that Visipaque is renally better than all LOCM. Indeed, this latter finding is also supported by the non-definitive language used in NEPHRIC's own conclusion that Visipaque MAY cause less CIN than A LOCM. Thus, the Court concludes that the NEPHRIC results do not support a claim of Visipaque renal superiority over all LOCM or any LOCM other than the one tested in that study (Omnipaque) because only one LOCM was compared, and because the NEPHRIC conclusion does not make an absolute claim of Visipaque renal superiority, hedging its findings with less than definitive language; NEPHRIC also does not support a conclusion that Visipaque has renal superiority over LOCM with prophylactics because the study did not compare any LOCM with prophylactics against Visipaque.
i. None of the Proffered Studies Demonstrate that all LOCM (including Isovue) without Prophylactics Cause the Same Rate of CIN
Bracco contends that P1937, an internal GEH document with MA results, shows differences in rates of CIN between LOCM and that as such, a head-to-head study with one LOCM cannot be extrapolated to other LOCM. GEH contends that this was not an analysis of relative CIN rates*fn99 and that published guidelines treat all LOCM, including Omnipaque and Isovue, as functionally interchangeable.*fn100 However, the Court finds that even though multiple CM are categorized together as LOCM, it does not mean that they have the same effect, or produce the same rate of CIN. The Court is not persuaded that all LOCM perform identically - multiple studies introduced in evidence show that not all LOCM perform similarly nor do they produce the same rate of CIN. (See RECOVER, Jingwei, McCullough MA, VALOR, IMPACT and CARE).
3. GEH's Establishment Claims Of Non-Renal Superiority
GEH also disseminated establishment claims of cardiovascular system, pain, warmth, discomfort and patient movement superiority, and establishment claims that Visipaque, and isoosmolar agents generally, are a superior class of drugs that lead to lower hospital, legal and patient care costs. The Court finds that these claims explicitly or implicitly assert that data from clinical studies show that Visipaque is superior to a LOCM or all LOCM and thus they are establishment claims. As with the renal establishment/superiority claims, these claims were: (a) designed to be disseminated (Pl.'s FOF ¶ 4); (b) shown to be effective by GEH's collective experience and marketing research (Pl.'s FOF ¶ 4); (c) disseminated by the GEH representatives (Pl.'s FOF ¶ 5); and (d) disseminated in various channels of communication.*fn101
a. GEH's Non-Renal Cardiovascular Superiority Claims Are Not False And Misleading
Examples of the cardiovascular system establishment/superiority claims extracted from GEH's print media, sales calls records and CME-type presentations are:*fn102
· Press releases on website even during trial: "Abstract Shows Significantly Lower Incidence of [Major Adverse Cardiac Events or Major Adverse Clinical ("MACE")] Following [PCI] Using Visipaque Compared to Isovue.." P2669:480, P3114H:857-58, P1893:940-41, P4151:p2; 7 T 69.
· CT brochures: "Nonionic Dimer Provides Reduced MACE ." P410:965, P3649:408, P3649A:408, D2324:117.
· Sales call records: "Visipaque doesn't increase heart rate or B/P like LOCM". P3682:Omni/38573, 4049:Omni/38573.
Bracco contends that GEH's cardiovascular claims are false because: (a) the studies do not support the claims (e.g., no superiority over all LOCM, i.a. results do not predict i.v. results, and any difference purportedly shown in studies was fleeting because the results for different contrast agents converge after 30 days); (b) there were allegedly omitted results and flaws that contradict the claims made in GEH's advertising; and (c) the studies GEH relies on (COURT and VICC) are unreliable. The following is a detailed analysis of Bracco's allegations beginning with the various promotional materials disseminated and moving on to the integrity of the COURT and VICC trials. Bracco's assertions, include: (1) the weight of the clinical evidence shows that Visipaque is not superior to all LOCM and definitely not superior to Isovue (Bracco concedes that ionic Hexabrix may be inferior); and (2) there has never been one AWC clinical study (let alone two, done the same way with the same drug) showing that Visipaque is superior to a LOCM with regard to MACE.
Referencing a series of letters dating back to 1996, Bracco argues that the FDA found no support for GEH to make cardiovascular superiority claims. (E.g., P596, P457, P588, P585, P82; 15 T 25-39, 43-44). Nonetheless, some of these letters pre-date the COURT trial (circa 2000) and all pre-date the VICC trial (circa 2005) and since that time, new evidence has come to light. Thus, the Court finds these FDA letters are neither dispositive, nor highly probative ...