R.F. v. Abbott Laboratories

745 A.2d 1174 (2000) 162 N.J. 596

R.F. and R.F., Plaintiffs-Appellants and Cross-Respondents,
v.
ABBOTT LABORATORIES, Defendant-Respondent and Cross-Appellant.

Supreme Court of New Jersey.

Argued September 14, 1999. Reargued November 29, 1999. Decided February 29, 2000.

Brian Wolfman, Washington, DC, a member of the District of Columbia and Arkansas bars, and George T. Baxter, New York City, for plaintiffs-appellants and cross-respondents (Mr. Baxter, attorney).

Kimball R. Anderson, Chicago, IL, a member of the Illinois bar, and Anne M. Patterson, Pensacola FL, for defendant-respondent and cross-appellant (Riker, Danzig, Scherer, Hyland & Perretti, Morristown, attorneys).

E. Drew Britcher, Morristown, submitted a brief on behalf of amicus curiae, The Association of Trial Lawyers of America-New Jersey (Leonard & Butler, attorneys).

The opinion of the Court was delivered by GARIBALDI, J.

In September 1986, plaintiff, R.F., received a transfusion of blood incident to surgery, that was infected with the human immunodeficiency virus ("HIV"), causing her to subsequently test positive for the presence of that virus. The blood which was transfused into R.F. had been previously screened for HIV infection at the Bergen Community Blood Center ("BCBC") with the first commercially-available HIV blood screening test, manufactured by defendant, Abbott Laboratories ("Abbott"), available on the commercial market between March 1985 and January 1987.

R.F. and her husband (collectively, "plaintiffs"), claim that the HIV blood test used by the BCBC was defective under N.J.S.A. 2A:58C-2,^[1] because its package insert failed to provide adequate instructions or warnings regarding the sensitivity limitations allegedly inherent in Abbott's test. Specifically, plaintiffs contend that in light of its knowledge that the test was not 100% sensitive, Abbott should have instructed blood banks to retest samples that were negative yet "borderline," meaning samples that had yielded results close to the test's "cutoff value." The cutoff value was a value defined by the federal Food and Drug Administration ("FDA") in the test's instructional pamphlet, to be used by blood bank technicians to measure whether the HIV antibody was present in a donated blood sample.

After a five-week trial, a jury found that Abbott had provided adequate warnings, and that the test was not defective. In an unpublished decision, the Appellate Division affirmed on other grounds holding plaintiffs' state law claims were preempted by the "FDA's extensive scrutiny and monitoring of the defendant's test and dictates regarding the warning inserts." The primary issue presented in this appeal is whether federal regulation of Abbott's HIV blood screening test preempts plaintiffs' cause of action for defective design and failure to warn under N.J.S.A. 2A:58C-2.

I.

A. Facts

In 1981, five cases of a rare pneumonia, were reported in healthy young men in Los Angeles. However, it was not until 1982 that the Centers for Disease Control ("CDC") identified the infection as the autoimmune deficiency virus ("AIDS"). In the ensuing years, the number of newly-reported cases of AIDS grew exponentially:

1982:     593
1983:   3,000
1984:   6,900
1985:   8,000
1986:  14,000

By 1983, scientists understood that AIDS was caused by a virus that could be transmitted through intra-venous drug use, blood transfusions, or sexual contact, and had identified certain high-risk groups including: homosexual men, IV drug users, Haitian immigrants, and hemophiliacs (persons suffering from a blood clotting disorder requiring transfusions of large amounts of blood proteins to prevent bleeding). The discovery that the HIV virus could be transmitted through blood transfusions created a national health crisis with respect to the availability of safe blood, because donated blood was a high-risk source of infection, and the pool of qualified blood donors was reduced by both persons with AIDS, and members of the identified high-risk groups. As a result, as early as 1983, the United States Public Health Service recommended that physicians use transfusions sparingly, and encourage autologous (or self) blood donation.

As the agency responsible for the safety of the blood supply, the FDA took action quickly to contain the AIDS epidemic by seeking help from both public and private sources. Initially, the FDA turned to the National Cancer Institute ("NCI"), where in May 1984, Dr. Robert Gallo had published the first description of a screening test that could detect the HIV infection in blood samples. Dr. Gallo's prototype screening test was an "ELISA" or "EIA" test, a generic term that describes "Enzyme Linked Immunoabsorbant Assays." ELISA tests, which had been used for the detection of other blood-borne viruses such as hepatitis, function by detecting the body's immune response (or antibodies) to a virus or its antigen components, rather than the virus itself.^[2] Dr. Gallo's prototype test used a bead coated with an inactive form of the HIV virus, designed to detect the presence of HIV antibodies in donated blood samples. If HIV antibodies were detected, then it could be concluded that the patient's immune system had been exposed to (and had responded to) the HIV virus. The prototype was approximately 85% effective in detecting infected samples. The government applied for, and ultimately received, a patent for Dr. Gallo's HIV ELISA test.

On May 3, 1984, soon after Dr. Gallo's publication, the United States government published a solicitation in the Federal Register ("Notice") seeking private manufacturers who could further develop, manufacture, and mass-distribute Dr. Gallo's prototype HIV screening test to the nation's blood banks. See Request for Applications to Produce a Virus and an Assay System for Detection of Antibodies to the Virus Associated with Acquired Immune Deficiency Syndrome (AIDS), 49 Fed.Reg. 18899 (May 3, 1984). The Notice stated in part:

The Department of Health and Human Services solicits applications to produce the HTLV-III^[3] virus and to develop and distribute [Dr. Gallo's prototype] assay system for the detection of antibodies to HTLV-III, the newly discovered virus associated with the AIDS Syndrome, under non-exclusive, royalty-bearing licenses from the Department....

[Id. at 18899.]

The Notice sought private manufacturers who could: (1) grow the virus "on a large scale;" (2) develop an ELISA test kit for distribution to blood banks, phasmapherisis centers, hematology and disease laboratories, and medical research institutions; (3) provide nationwide distribution of the HIV test kit; and (4) monitor the field performance of the test in blood banks and research laboratories. Id. at 18899-18900.

According to the terms of the Notice, a manufacturer was required to obtain a license from the FDA prior to releasing its test for commercial distribution. Id. at 18900. In "view of the important public health considerations involved," responses were requested within 10 days from the date of the Notice. Ibid. Dr. Harry Meyer, Director of the FDA's Center for Drugs and Biologics from 1982 until September 1986, was involved in the selection of manufacturers for the development of Dr. Gallo's test. Dr. Meyer testified at trial that the "government wanted to see the most efficient and rapid development not only [of] the initial test but for the research for future generations of the test[;] ... [g]etting a good test out was about as high a priority as the government had at the time."

On May 9, 1984, Abbott filed a response to the Notice demonstrating its capacity to develop and manufacture Dr. Gallo's prototype. In July 1984, after reviewing the application and meeting with a group of Abbott scientists and representatives to discuss their proposal, the FDA informed Abbott that it would be granted a license for the government's patent so that it could develop and manufacture the prototype, and cultivate the virus. According to Abbott's lead scientist in the development of the HIV assay, Dr. John Heller, Abbott's primary purpose in the test's development stage "was to further improve th[e] assay regarding sensitivity,^[4] and also to scale up all of the procedures so that we could manufacture on the order of millions of tests a month." Dr. Heller testified that there was "clearly a sense of urgency" in Abbott's development of the test since the company "knew that people were becoming infected through blood transfusion, [and] ... that people were dying of AIDS as a result of blood transfusion."

Four other manufacturers also were selected to develop Dr. Gallo's prototype. Genetic Systems, Inc., headed by plaintiffs' expert, Dr. Robert Nowinski, applied for, but did not receive, a license to develop and manufacture the government's prototype. However, Genetic Systems did ultimately manufacture an HIV blood screening test which was licensed by the FDA in 1986 (one year after Abbott's test was licensed), yet competed with Abbott's version of the test somewhat unsuccessfully in the commercial market.

Although the FDA considered the blood screening test as both a "device" under the Medical Device Amendments of 1976, 21 U.S.C. § 360c to § 360ee ("MDA") to the federal Food, Drug and Cosmetic Act, 21 U.S.C. § 301 to § 395 ("FDCA"), and a "biologic" under the Public Health Service Act, 42 U.S.C. § 201 to § 300aaa (the "PHSA"), the development, manufacturing, and field performance of the HIV test, was overseen by the FDA's Office of Biologics Research and Review ("OBRR"). That is consistent with the FDA's 1982 designation of the Bureau of Biologics as "the lead Bureau for regulating certain medical devices used in the processing or administration of biological products," such as the test kit. Working Relationship Agreement Among FDA's Bureaus of Medical Devices, Radiological Health, and Biologics; Availability of Document, 47 Fed.Reg. 15412, 15412 (April 9, 1982) ("Working Relationship Agreement"). The FDA clarified in the Working Relationship Agreement that "biological medical devices," such as Abbott's HIV blood screening test, are subject to the provisions of both the FDCA and the PHSA. Ibid. Therefore, although the test kit was largely regulated by the OBRR as a biologic (because the virus was the main component of the test) the OBRR required that the test be listed as a medical device, and its package insert drafted pursuant to Labeling for In Vitro Diagnostic Products, 21 C.F.R. § 809.10(b) (1985), a medical device regulation.

It was undisputed at trial that the FDA was intimately and proactively involved in the development, clinical trials, and licensure process of Abbott's proposed HIV blood screening test. Dr. Meyer described the relationship between the FDA and the manufacturers as follows:

It was an intense relationship, I mean, it was a very high pressured time.... There was a big stake in getting that test out, so we all recognized it was everybody's first priority. In all of the years I was in FDA and talking about priorities on hot drug development or hot biologic development, I can't think of any time there was more priority than that intense period of getting an effective AIDS test out.

Dr. Meyer testified that during the development and licensure of the test, there was a "continual dialogue" between the FDA's scientific staff and the manufacturers regarding their progress and clinical results. Similarly after licensure, the FDA continued this dialogue with respect to the monitoring of the test, and the development of a Second Generation Test.

Throughout the fall of 1984, Abbott developed the HIV test kit and submitted proposals to the FDA to conduct clinical trials. Abbott's completed test had the same basic components used by Dr. Gallo.^[5] During December 1984, Abbott conducted clinical trials of its proposed assay that included approximately 10,000 random donors, including a sample group of AIDS and ARC patients (ARC patients are infected with the virus yet do not exhibit the symptoms of full-blown AIDS). The results of the trials using AIDS and ARC patients showed that the test was 98.3% effective in detecting the antibody in AIDS patients, and 65% effective in detecting the antibody in ARC patients. In addition, the FDA collected 30,000 samples from across the United States, distributing 6,000 samples to each of the five manufacturers developing Dr. Gallo's prototype. Each manufacturer tested the 6,000 samples with the proposed test, returned the data to the FDA, and then the FDA published the results.

On December 19, 1984, Abbott sent all of its clinical data to Dr. Esber, Director of the OBRR Center for Drugs & Biologics, with the product license application for the proposed test. According to Marijane Sidote, Abbott's Director of Regulatory Affairs throughout the relevant time period, the FDA "took the unprecedented step of asking for raw data points [for the clinical trials], and printouts from the machines, so they could independently examine and analyze the data." Moreover, Sidote testified that:

[d]uring the next two-and-a-half months there was a continual communication between me and the FDA groups that were working on this product. They asked me for additional information in some cases. One of the things they asked us to do was to test samples that they had in their laboratory, so they sent us several thousand samples for us to test. The results of that were also reported, so there was considerable paperwork that I filed following the initial product license application.

As part of the product license application, Abbott submitted a draft of the test's package insert. In the December 19, 1984 draft, Abbott suggested that the package insert state: "Specimens with absorbance values within a ± 10% range of the Cutoff Value should be retested to confirm the initial results." (emphasis added). However, in their January 29, 1985 response, according to Sidote, the FDA "mandated that [this provision] be deleted." Dr. Heller testified that the FDA decided not to instruct blood banks to retest "borderline" negative samples because: (1) there was no scientific basis for the belief that samples close to the borderline were more likely to be false-negative than negative samples with results well-below the cutoff; and (2) such a provision would effectively redefine the cutoff to the lower value for which there would be associated a new set of "borderline samples."

Dr. Heller testified that in defining the test's cutoff value, the FDA "specifically indicated [to Abbott] that they had reentered all of the raw data and redid the statistical analysis to determine that the cutoff was in the appropriate place." During this process, the FDA and Abbott balanced the goal of maximized sensitivity with the problem of having an abundance of false-positive results. The appearance of false-positive results was a major concern throughout the development of the First and Second Generation Tests, since false-positive results not only caused a blood bank to destroy that particular sample, but also disqualified the donor from future blood donations, thereby further limiting the qualified source of donated blood.

The FDA's active role in setting the test's cutoff was confirmed by Dr. Meyer who testified that, although Abbott contributed to the discussion, the FDA ultimately determined the appropriate cutoff value.^[6] Dr. Meyer also testified more generally that the "FDA virtually wrote parts of th[e] package [insert], and one way you could see that, if you look at each of the manufacturer's package [inserts] for that test at that time, there are sections of them ... that are virtually identical."

In addition to rejecting the provision requiring negative samples to be retested, according to Sidote the FDA "made very extensive changes to [Abbott's package insert] draft[s], giving [them] specific language that they wanted us to put in the package insert." Similarly, Dr. Heller testified that "the FDA dictated ... a lot of the language in the package insert, and also dictated how the tables [reflecting the clinical trial data] would be arranged and the order in which material would appear within the package insert." Dr. Meyer confirmed this testimony stating the FDA was "intimately involved with Abbott in their packag[e] labeling exactly indicating what the uses and limitations of the test [we]re."

The affidavit of P. Ann Hoppe, who was at all relevant times the Deputy Director/Acting Director of the FDA's Division of Blood and Blood Products of the Center for Biologics Evaluation and Research, confirmed that the warning and limitation language of the package insert was "to a significant extent, dictated by the FDA":

Specifically, with respect to Abbott's package insert, the warnings and statements relating to the limitations of the test were not only appropriate, but also were supported by extensive clinical data that the FDA reviewed. In addition, the language contained in the insert was approved by the FDA and, to a significant extent, dictated by the FDA.... Not only did the FDA approve the data in the insert, but the FDA also went to great lengths to assure consistency among manufacturers in the type of information that was available to users in the package insert. All of these licensed products have very common elements in their package inserts, if not identical language in some respects, because the FDA participated in the drafting of what should be contained therein.

[(emphasis added).]

In particular, the FDA essentially authored three critical sections of the test's package insert entitled: "INTERPRETATION OF RESULTS," "LIMITATIONS OF PROCEDURE," and "SENSITIVITY AND SPECIFICITY." Those sections appeared in identical form in each of the five manufacturers' package inserts; none of them mentioned "borderlines," and none of them instructed users to retest initially negative samples. The INTERPRETATION OF RESULTS section of the completed package insert instructed blood bank technicians:

1. Specimens with absorbance values less than that of the Cutoff Value are negative by the criteria of ABBOTT HTLV III EIA.

2. Specimens with absorbance values greater than or equal to the Cutoff Value are considered reactive by the criteria of the ABBOTT HTLV III EIA and should also be retested before interpretation using the original sample source.

3. Specimens which have been found to be repeatably reactive [or repeatably above the Cutoff Value] are interpreted to be positive for antibody to HTLV III ...

[(emphasis added).]

Both the LIMITATIONS OF PROCEDURE and the SENSITIVITY AND SPECIFICITY sections indicated that the Test was not 100% sensitive in detecting the HIV antibody in truly infected blood samples. Specifically, the LIMITATIONS OF PROCEDURE section states in part: "A negative test result does not exclude the possibility of exposure to or infection with HTLV III" (emphasis supplied). Similarly, the SENSITIVITY AND SPECIFICITY section sets forth a textual explanation of Abbott's clinical trials which revealed:

1. Sensitivity based on an assumed 100% prevalence of HTLV III antibody in AIDS patients is estimated to be 98.3%.

2. Specificity^[7] based on an assumed zero prevalence of HTLV III antibody in random donors is estimated to be 99.8%.

Moreover, Table III of the package insert, which summarized Abbott's clinical trial data, concluded that Abbott's test was positive in "57 (98.3%) of 58 AIDS patients and 72 (67.3%) of 107 ARC patients."

On March 1, 1985, after the package insert was completed, the OBRR issued a license that "authorized Abbott to manufacture and sell in interstate and foreign commerce HTLV III in an in vitro ELISA test," hereinafter referred to as the "Test" or the "First Generation Test." In an accompanying letter to Abbott, the FDA set forth several conditions of the license, including that (1) Abbott submit "ongoing stability studies" of the Test for review by the OBRR; (2) Abbott submit a sample of every lot of the Test and await FDA approval before that lot is commercially-distributed; and that (3) Abbott report all significant product defects or product complaints concerning the use of the Test.

In addition, the FDA's letter accompanying the license specifically indicated that if Abbott sought to amend the labeling or package insert of the Test, "it w[ould] be necessary ... to submit an amendment to either [the] product or establishment license application for review and approval prior to implementation." Moreover, the Test's labeling was subject to a similar, promulgated FDA regulation for biologics, 21 C.F.R. § 601.12, which stated:

(a) General Important proposed changes in location, equipment, management and responsible personnel, or in manufacturing methods and labeling, of any product for which a license is in effect or for which an application for license is pending shall be reported to the Director, Office of Biologics Research and Review, by the manufacturer, and unless in case of emergency, not less that 30 days in advance of the time such changes are intended to be made.

(b) Manufacturing methods and labeling Proposed changes in the manufacturing methods and labeling may not become effective until notification of acceptance is received from the Director, Office of Biologics Research and Review.

[Changes to be reported, 21 C.F.R. § 601.12 (1985) (emphasis added).]^[8]

As indicated by its drafting of the package insert warnings, the FDA was aware of the limitations of Abbott's First Generation Test (and ELISA tests in general) at the time of its development. Specifically, the FDA understood that since all ELISA tests detected the presence of antibodies to virus, and not the presence of viruses themselves, the tests were by design subject to a "window period" during which an infected individual does not have detectable levels of antibodies, causing the test to produce false-negative results. That "window period" is the length of time required for the body to produce enough antibodies to be detected. In 1985-86, the scientific community could not predict precisely how long the window period was for HIV. The FDA understood that it was possible that there were individuals infected with the HIV virus who had not yet produced HIV antibodies and, therefore, could donate infected blood (fully-capable of spreading the disease) that could not be screened by any of the manufacturers' ELISA tests.

Accordingly, prior to the introduction of the Test onto the market, the FDA conducted a mass-mailing campaign to blood banks and physicians regarding the use and limitations of the first HIV blood screening tests. On February 19, 1985, the FDA sent a memorandum to all registered blood banks in the United States, including the BCBC, regarding the impending licensure of the first HIV antibody blood screening tests. The FDA's memorandum, summarizing an attached copy of the CDC's January 11, 1985 newsletter, explicitly described the limitations of the tests as follows:

... [A] negative antibody test result does not necessarily mean that one is free from virus. Antibody may not have developed, or be undetectable, if infection was recent. There is at least one report that 4 of 96 individuals carried the virus for 6 months without developing detectable antibodies.

See Provisional Public Health Service Inter-Agency Recommendations for Screening Donated Blood and Plasma for Antibody to the Virus Causing Acquired Immunodeficiency Syndrome, Morbidity and Mortality Weekly Report, Vol. 34, No. 1 (Centers for Disease Control) (Jan. 11, 1985). Soon thereafter, the FDA sent out a similar "Dear Doctor" letter to all licensed physicians in the United States, including R.F.'s physician, expressly warning of the above limitations.^[9] Dr. Meyer testified that the FDA's circulation of these letters was an unusual and "exceptional" measure, which demonstrated "the enormous effort that was made not only by [the FDA] but also by industry and others to fully inform users of [the new test of] what could be expected of it."

In view of the recognized limitations of the First Generation Test, the development of a Second Generation Test began as soon as the Test was licensed in March 1985. In developing the Second Generation Test, Abbott focused on isolating an antibody that could function as an "early marker" of the infection, and therefore improve the Test's sensitivity in the window period. During this post-licensure period, the FDA continued to monitor both the field performance of the First Generation Test and Abbott's development of a Second Generation Test. Dr. Meyer testified that once the Test was marketed, there was a "whole host of monitoring efforts" by the FDA, and that there was "a lot of discussion on the monitoring of the test and its performance in practice." In fact, by FDA order, Abbott tested a portion of every manufactured lot of the Test with a control panel of samples provided by the FDA. Abbott then submitted its data from those tests, and a sample of each lot, to the FDA. After reviewing the data, and possibly testing the sample lot themselves, the FDA would notify Abbott by letter or fax whether the lot was approved for sale. Abbott was prohibited from distributing any portion of a lot before a formal release was received from the FDA.

According to Dr. Meyer, the FDA monitored the Test's performance not only through information collected from Abbott and the other manufacturers, but also from blood banks, other government agencies, and scientific publications. Moreover, throughout the relevant time period, Dr. Meyer represented the FDA on the "AIDS Executive Task Force," a group of representatives from government agencies (such as the NIH and CDC) who met to discuss AIDS-related issues. With respect to the blood banks, Dr. Meyer explained that:

all of the users [of the Test] ... are blood banks that are under [FDA] regulation, so the major blood bank organizations were doing a great deal of tracking of performance, and those blood bank organizations reported to us. You might even have major blood banks that had the results and report directly to us. Then in addition[,] ....[in] those early months after [the Test] came out, we asked for and promptly got reporting I think on an every two week interval from essentially, I think it was well over 50 percent of the blood and plasma phoresis facilities in the country monitoring their experience repeat reactive rates.

Dr. Meyer testified that "[t]he other big avenue of information ... [was] the Centers for Disease Control." Dr. Meyer explained that the CDC was:

supporting all the testing activity by all the state health networks throughout the country, and they were all using this type of test, and they had an enormous amount of experience in using tests of this sort ... I and people on my staff [we]re in more or less continual dialogue with the CDC.

...

The CDC ... was doing a lot of direct research on the performance of the [T]est after the [T]est came out. [At a meeting of the AIDS Executive Task Force,] one of the critical articles that I heard the preliminary report [of was] the big public workshop we had at the NIH on performance of tests in July 1985, ... that was a report largely done by the CDC ... [T]hat gave us a great deal of information.

Similarly, in her affidavit P. Anne Hoppe stated that:

[o]nce Abbott's test was in wide use in the field, there were ongoing communications between Abbott and the FDA relating to the test's performance.... As a result of the ongoing communication between the FDA, the blood banking industry and manufacturers such as Abbott, the FDA was well-informed about test performance in the field.... It is significant to note that had the FDA felt that Abbott's test was defective or that Abbott's warnings were inadequate based upon reliable and significant new information, the FDA would have taken action to protect the public's interest.

According to the trial testimony, the field performance data, as well as independent scientific research, reinforced the FDA and Abbott's initial belief that the incidence of false-negative results were not clustered around the cutoff, but rather were distributed randomly below the cutoff value. In that regard, both Dr. Thomas Zuck, the Director of the FDA's Division of Blood and Blood Products, and Dr. Meyer testified that throughout the relevant 1985-86 time period the Test was state-of-the-art. In fact, Dr. Meyer testified: "I can say without qualification whatsoever that all those [F]irst [G]eneration [T]ests were equivalent, there weren't any that were significantly better than any others." Similarly, Dr. Meyer testified that "there [was] never any serious question that Abbott's test or anyone else's test was not within the specs that were described in the package circular." That testimony was confirmed by Dr. Zuck, who testified that the FDA had no concerns about the sensitivity of the Test during the post-licensure period.

On July 28, 1986, approximately one month prior to R.F.'s transfusion, Abbott submitted a product license amendment for the Second Generation Test, which indicated that the p41 enriched test would be better able to detect positive samples that were "borderline or negative" by the First Generation Test. (emphasis added). In other words, the improved test was intended to be better able to detect all positive samples that falsely registered below the cutoff by the First Generation Test.

Although Abbott requested immediate permission to market the improved test, the FDA first required that Abbott conduct clinical trials with the proposed Second Generation Test. During the clinical trials, Abbott found a higher false-positive rate than expected and therefore, reformed the test in the laboratory and began the clinical trials again. After submitting the second round of clinical trials to the FDA, the Second Generation Test was approved for the market in January 1987. Abbott immediately withdrew the First Generation Test, and sent a Western Union mailgram to customers stating that the test improved the sensitivity and specificity of the First Generation Test.

On September 4, 1986, prior to the approval of the Second Generation Test, R.F. received a unit of blood that was infected with the HIV virus during orthopedic surgery in Valley Hospital in Ridgewood, New Jersey. The unit of blood was previously tested for HIV infection at the BCBC on August 26, 1986, with the First Generation Test. When tested, the unit of blood was "borderline negative" with a result of 0.121, slightly below the cutoff value for a positive reading of 0.128.^[10] On November 4, 1986, the same blood donor again donated blood at BCBC. That donation, also tested with the First Generation Test, was found to be above the 0.128 reading and therefore, HIV positive. A review of the BCBC's records determined that blood from the earlier donation had been transfused into R.F. On September 4, 1987, R.F. was notified that she was infected with HIV.

B. Procedural History

On June 20, 1989, plaintiffs filed a complaint against Abbott and others, including R.F.'s treating physician, BCBC, and Valley Hospital. Prior to trial, plaintiffs settled their claims against the all defendants other than Abbott for $1,500,000. With respect to Abbott, plaintiffs alleged: (1) breach of implied and express warranties; (2) inadequate warnings; (3) strict products liability; (4) negligence; and (5) willful, intentional, and wanton disregard for the rights and safety of others.

Plaintiffs' main contention focused on whether Abbott provided adequate warnings within the package insert. Specifically, plaintiffs assert that Abbott was aware that the Test was producing false-negative results in the early part of 1986 for samples that produced results near the "borderline" (or close to the cutoff value) of the Test. As a result, plaintiffs argue Abbott was required under New Jersey law to warn of the incidence and the inherent dangerousness of borderline samples in a supplemental package insert, and/or instruct blood banks to retest such "borderline" samples.^[11]

On September 23, 1993 Abbott filed a motion for summary judgment on the grounds that the express preemption provision of the MDA, specifically 21 U.S.C. § 360k(a) ("Section 360k(a)"),^[12] preempted plaintiffs' state law claims. Relying on Feldman v. Lederle Laboratories, 97 N.J. 429, 479 A.2d 374 (1984) ("Feldman I") and Feldman v. Lederle Laboratories, 125 N.J. 117, 592 A.2d 1176 (1991) ("Feldman II"), the trial court denied Abbott's motion, holding that it would be inconsistent with congressional intent to deny private citizens a tort remedy based on the FDCA, since it is a statute designed to protect the public's health and welfare; and moreover, there were material issues of fact regarding whether Abbott complied with federal warning requirements. Abbott's motion for leave to seek an interlocutory appeal of that decision was denied. In 1996, Abbott again filed a motion for summary judgment alleging federal preemption of the state law claims. On September 12, 1996, the trial court denied that motion largely based on the United States Supreme Court decision in Medtronic v. Lohr, 518 U.S. 470, 116 S.Ct. 2240, 135 L.Ed.2d 700 (1996).

Beginning in September 1996, the trial court conducted a five-week jury trial to determine Abbott's liability. All the evidence before this Court was before that jury. The jury returned a verdict in favor of Abbott on all counts, concluding that Abbott had provided sufficient and adequate warnings, and that the Test was not defective. The trial court entered judgment for Abbott on November 29, 1996.

The plaintiffs appealed. In an unpublished decision, the Appellate Division affirmed the judgment in favor of Abbott based on the conclusion that plaintiffs' claims were preempted by the MDA. The panel reasoned that the FDA's inaction with respect to the recall of the Test or modification of the Test's package insert should be construed as approval in light of: (1) the agency's extensive role in the development of the Test and the inserts; (2) the agency's continued monitoring of the Test; and (3) the agency's retention of jurisdiction to recall the Tests and to direct the amendment of the package inserts. The panel emphasized that Abbott was prohibited from altering the Test or the package inserts prior to the FDA's completing its review of Abbott's data, a task that was not accomplished until early 1987.

On July 15, 1998, this Court granted the plaintiffs' petition for certification, and subsequently ordered the Appellate Division to reconsider the plaintiffs' preemption arguments in light of Baird v. Am. Med. Optics, 155 N.J. 54, 713 A.2d 1019 (1998), and Medical Devices; Preemption of State Product Liability Claims, 62 Fed. Reg. 65387 (Dec. 12, 1997) (to be codified at 21 C.F.R. pt. 808);^[13] and to consider the plaintiffs' non-preemption issues raised on appeal. 154 N.J. 606, 713 A.2d 497 (1998). In September 1998, in an unpublished decision, the Appellate Division again held that plaintiffs' failure to warn claims under N.J.S.A. 2A:58C-2(b) were preempted by the MDA. However, that court held that if the claims were not preempted, plaintiffs were entitled to a new trial because the jury instructions were prejudicially confusing.

We again granted plaintiffs' petition for certification. 158 N.J. 72, 726 A.2d 936 (1999). In September 1999, the case was presented to the Court at oral argument. After the case was argued, we informed the parties by letter that the case was to be set down for further argument, and that they were to file supplemental briefs to answer four questions: (a) What new evidence did Abbott acquire after the licensing of the Test regarding its safety? (b) Did Abbott have a duty, or was it permissible, to change the Test's package insert or issue a warning after receiving the alleged after-acquired evidence? (c) If the FDA possessed the alleged after-acquired evidence, did this alter Abbott's duties with respect to the Test's warnings? (d) Was there a federal regulatory scheme in place during the summer of 1985 that was in conflict with the state product liability law requiring manufacturers to inform users of after-acquired safety evidence?

In response to that letter, the parties submitted supplemental briefs and appendices. In plaintiffs' supplemental briefs they raised for the first time two new arguments: (1) that the Test was not a "medical device," but solely a "biologic," and therefore principles of implied rather than express preemption under Section 360k, should apply; and (2) that even if Abbott was restricted from disseminating a new package insert, it should have issued additional warnings through the use of a "Dear Doctor" letter, telegram, or other similar means.

II.

The United States Constitution provides that the Constitution and the laws of the United States, "shall be the supreme law of the land." U.S. Const., art. VI, cl. 2. The tests for determining whether state laws are preempted by federal law are well-established:

Pre-emption may be either expressed or implied, and "is compelled whether Congress' command is explicitly stated in the statute's language or implicitly contained in its structure and purpose." Absent explicit pre-emptive language, we have recognized at least two types of implied pre-emption: field pre-emption, where the scheme of federal regulation is "``so pervasive as to make reasonable the inference that Congress left no room for the States to supplement it,'" and conflict preemption, where "compliance with both federal and state regulations is a physical impossibility," or where state law "stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress[.]"

[Gade v. Nat'l Solid Wastes Management Assoc., 505 U.S. 88, 98, 112 S.Ct. 2374, 2383, 120 L.Ed.2d 73, 84 (1992) (citations omitted).]

Whether a state law stands as an obstacle to the accomplishment of a federal objective, requires a court to consider "the relationship between state and federal laws as they are interpreted and applied, not merely as they are written." Jones v. Rath Packing Co., 430 U.S. 519, 526, 97 S.Ct. 1305, 1310, 51 L.Ed.2d 604, 614 (1977) (citations omitted).

The three categories of preemption are usually defined as express preemption, implied preemption, and conflict preemption. Laurence H. Tribe, American Constitutional Law, Vol.I, § 6-28 (3d ed.2000). However, the Supreme Court and leading constitutional scholars agree that "preemption categories are not ``rigidly distinct.'" Gade, supra, 505 U.S. at 103 n. 2, 112 S.Ct. at 2386 n. 2, 120 L.Ed.2d at 88 n. 2 (citation omitted).

These three categories of preemption are anything but analytically air-tight. For example, even when Congress declares its preemptive intent in express language, deciding exactly what it meant to preempt often resembles an exercise in implied preemption analysis. So too, implied preemption analysis is inescapably tied to the presumption that Congress did not intend to allow state obstructions of federal policy, the existence of which is a central inquiry in conflict preemption analysis.

[Tribe, supra § 6-28 (citations omitted) (emphasis in original).]

Determining whether federal law preempts state law is a fact-sensitive endeavor, based on a court's review of "fragments of statutory language, random statements in the legislative history, and the degree of detail of the federal regulation." Erwin Chemerinsky, Constitutional Law: Principles and Policies, § 5.2 (1997). However, preemption "``is not to be lightly presumed.'" Turner v. First Union Nat'l Bank, 162 N.J. 75, 87, 740 A.2d 1081 (1999) (quoting Franklin Tower One, L.L.C. v. N.M., 157 N.J. 602, 615, 725 A.2d 1104 (1999)).

When a federal regulation conflicts with state law, the same preemption rules apply. Fidelity Fed. Sav. & Loan Ass'n v. de la Cuesta, 458 U.S. 141, 153-54, 102 S.Ct. 3014, 3022-23, 73 L.Ed.2d 664, 675 (1982). Despite the dissent's implication to the contrary, see post at 649-50, 745 A.2d at 1205, the United States Supreme Court has held "repeatedly that state laws can be pre-empted by federal regulations as well as federal statutes." Hillsborough Cty. Fla. v. Automated Med. Laboratories, Inc., 471 U.S. 707, 713, 105 S.Ct. 2371, 2375, 85 L.Ed.2d 714, 721 (1985) (citations omitted). As long as the agency (1) intended to preempt state law; and (2) acted within the scope of its delegated authority, federal regulations will displace conflicting state laws. Fidelity Fed., supra, 458 U.S. at 153-5, 102 S.Ct. at 3022-23, 73 L.Ed.2d at 675.

III.

Before the Appellate Division, the parties focused on whether Section 360k was applicable to this matter. We conclude, however, that based solely on principles of implied preemption, the plaintiffs cannot maintain their state failure to warn claim against Abbott.^[14]

There are three types of implied preemption: (1) field preemption, "where the scheme of federal law and regulation is ``so pervasive as to make reasonable the inference that Congress left no room for the states to supplement it;'" (2) conflict preemption, where there is a conflict between federal and state law, rendering "``compliance with both federal and state regulations ... a physical impossibility;' " and (3) preemption where "state law impedes the achievement of a federal objective;" in this case, even if federal and state law are "not mutually exclusive ... preemption will be found if state law ``stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress.'" Chemerinsky, supra § 5.2 (citations omitted). As with the three general types of preemption, these categories of implied preemption are not perfectly distinct, and in practice often overlap. Ibid.

A. Facts That Support Implied Preemption

In this case, both (1) the FDA's exercise of control and initiative over the Test's development, packaging, and field performance monitoring, and (2) the unique circumstances under which the Test arose (a national health crisis surrounding the emergence of the AIDS epidemic and the loss of a safe national blood supply), give rise to implied preemption of the plaintiffs' state law claims under each of the three categories of implied preemption.

As discussed in detail above, after scientists discovered the virus that caused the AIDS infection, the federal government turned to private industry to assist in expediently developing a blood screening test to curb the spread of the HIV virus. Once selected, Abbott and the other four manufacturers were directed to use the government's prototype in developing a test. It is undisputed that: (1) the FDA took an active role in, and then later scrutinized, the development and clinical trials of Abbott's proposed HIV blood test; (2) the FDA dictated the Test's warning labels, and explicitly rejected the additional warnings and instructions regarding borderline negative samples that the plaintiffs claim were required by state law; (3) the FDA continued to closely monitor the Test's field performance both through information received from Abbott, and information received directly from the nation's blood banks and scientific publications; (4) based on that information, the FDA believed throughout the relevant time period (1985-86) that the Test's field performance was consistent with the FDA's expectations and also with the warnings, limitations, and clinical results published in the Test's package insert; and (5) Abbott's product license, and 21 C.F.R. § 601.12, specifically prohibited it from unilaterally altering the Test's package insert or disseminating additional warnings through "Dear Doctor" letters or otherwise.^[15]

Although the plaintiffs' contended at oral argument and in their briefs to this Court that Abbott was permitted to unilaterally change the package insert under 21 C.F.R. § 601.12(f)(2)^[16] and 21 C.F.R. § 814.39(d),^[17] that argument is without merit. 21 C.F.R. § 601.12(f)(2), proposed on January 29, 1996, see Changes to Approved Application, 61 Fed.Reg. 2739-01, 2740-44 (January 29, 1996), was not effective until October 7, 1997, see Changes to Approved Application, 62 Fed.Reg. 39890-01,39901 (July 24, 1997), and therefore is inapplicable to this case. Similarly, 21 C.F.R. § 814.39(d), proposed on December 12, 1980, see Medical Devices; Premarket Approval of Medical Devices; Proposed Establishment of Regulations, 45 Fed.Reg. 81769, 81776 (Dec. 12, 1980), was not effective until November 19, 1986, see Premarket Approval of Medical Devices, 51 Fed.Reg. 26342, 26352-53 (July 22, 1986), one month after R.F.'s transfusion, and therefore is also inapplicable to this case. Moreover, as confirmed by the testimony of FDA officials who were intimately involved with the development and monitoring of both the First and Second Generation Tests, those regulations—even if timely—could not have overridden the express and specific instructions of the FDA requiring approval for any changes to the labeling of the Test.

The FDA's extensive control and pre-licensure and post-licensure scrutiny over the Test, is best summarized in the 1996 affidavit of P. Ann Hoppe, the Deputy Director/Acting Director of the FDA's Division of Blood and Blood Products:

...

5. I have personal knowledge of the regulatory approval process and monitoring application of Abbott's HTLV III ELISA test in blood banks and related donor deferral issues. As such I am aware that the FDA approved and regulated Abbott's HTLV III ELISA test as a biologic medical device. Although most medical devices are not required to be licensed, the FDA's licensing requirements provided the FDA with an additional level of control for this test. For example, the licensing regulations require approval of every change in labeling and manufacturing. In addition, the FDA has the ability to suspend the license without going through other processes. The pre-approval process included review of all of the clinical trials (including plans and protocol), all of the data in the license application, as well as the development of the statements in the labeling. The FDA was involved in the ongoing lot release testing and pre-licensing inspection which provided a level of assurance for product performance that was not available for other medical devices. In my twenty years of experience at the FDA I believe that this test was the one of most highly scrutinized with respect to initial licensing and monitoring of ongoing performance.

6. As referenced above, the HTLV III ELISA test is a medical device and licensed biological product. Abbott's HTLV III ELISA test was intensely reviewed and scrutinized not only in the time frame leading up to the March approval, but from thereon. Unlike drugs and devices, the FDA requires that biologic products be licensed as required by Section 351 of the Public Health Service Act so that the agency can maintain more control over the product. Throughout the life of a licensed product the FDA is more involved in every aspect of its manufacture. In addition, as a licensed product the FDA also has several remedial measure[s] available to it. For example, the FDA is able to immediately suspend the license and remove the product from the market if it feels that the product is unsafe.

7. Specifically, with respect to Abbott's package insert, the warnings and statements relating to the limitations of the test were not only appropriate, but also were supported by extensive clinical data that the FDA reviewed. In addition, the language contained in the insert was approved by the FDA and, to a significant extent, dictated by the FDA. In particular, the FDA specifically dictated the language contained in the package insert sections entitled "PERFORMANCE CHARACTERISTICS," "LIMITATIONS OF PROCEDURE," and "INTERPRETATION OF RESULTS." Also, the specific performance characteristics of the package insert are based directly upon the clinical data submitted in the license application. Not only did the FDA approve the data in the insert, but the FDA also went to great lengths to assure consistency among manufacturers in the type of information that was available to users in the package insert. All of these licensed products have very common elements in their package inserts, if not identical language in some respects, because the FDA participated in the drafting of what should be contained therein.

8. Once Abbott's test was in wide use in the field, there were ongoing communications between Abbott and the FDA relating to the test's performance.... As a result of the ongoing communication between the FDA, the blood banking industry and manufacturers such as Abbott, the FDA was well-informed about test performance in the field.

9. There also were no data that demonstrated that Abbott's HTLV III ELISA test did not meet the sensitivity limitations as set forth in the package insert. It is significant to note that had the FDA felt that Abbott's test was defective or that Abbott's warnings were inadequate based upon reliable and significant new information, the FDA would have taken action to protect the public's interest. For example, if the FDA had believed that the content of Abbott's package insert was no longer valid after the test was in use in the field, the FDA could have required Abbott to change the insert. Moreover, if the FDA later came to believe that it was necessary to add the term "borderline" to the insert it could have required Abbott to do so. However, the record reflects that Abbott already warned in its package insert that its test was not 100% sensitive. The FDA also could have recalled Abbott's test or independently warned customers had the FDA felt this was necessary. Again, the FDA did not feel that these types of actions were necessary and, as such, was satisfied with the performance of Abbott's test.

10. All labeling changes for the HTLV III ELISA test must be approved by the FDA under 21 C.F.R. [ § ] 601.12(b). Abbott could not have independently modified or otherwise changed its package insert without FDA approval since the HTLV III ELISA test was a licensed product. Once the cutoff is established, neither the manufacturer nor the blood banks are authorized to instruct technicians to deviate from the insert. Blood banks must run the test according to instructions authorized by the FDA. Under the applicable regulations, Abbott would have needed pre-approval from the FDA to change its labeling in any significant way. The FDA also considered advertising and promotion to be part of labeling as long as it related to the performance of the product. As such, Abbott could not have independently issued Dear Doctor or similar letters without FDA approval. In addition, recommended retesting of samples near the cutoff value effectively would constitute a change in the cutoff value and the FDA would consider this a major change in labeling. As such, the FDA would have required extensive additional data from clinical studies to demonstrate that such a change in the insert was supported. In the evaluation of how the test should be applied, the FDA considered and rejected the concept of retesting within +/-10% of the cutoff.

11. The FDA knew from the data in its possession that Abbott's test was not 100% sensitive and it very carefully considered the warnings that should be in the insert. The FDA was aware that there could be false negative results when setting the cutoff value and this was a known risk or limitation of the test. This limitation was described in the data in the insert under the section "PERFORMANCE CHARACTERISTICS." If sufficient antibodies were present in an individual, Abbott's test was very dependable in detecting those antibodies as expressed by the sensitivity and specificity data. The FDA considered the warnings in Abbott's package insert to be appropriate at all times leading up to the FDA's approval of Abbott's subsequent test in January, 1987. The data in the insert gave users more precise information about the performance of the test and the FDA considered this information to be sufficient and appropriate for public safety.

12. The FDA was in continual communication with Abbott and the FDA had access to ongoing lot release information. All this time the FDA allowed the test to be on the market without any changes in labeling. If the FDA felt there were any problems that affected the safety of the test, or that legitimately invalidated the contents of the insert, or there were significant new developments requiring a change to the insert, the FDA had numerous opportunities available to it to take action. The fact that the FDA never took any such action demonstrates that the FDA felt the labeling for Abbott's test was appropriate and adequate.

...

[ (emphasis added).]

The critical facts set forth in P. Ann Hoppe's affidavit were never contested by the plaintiffs and, indeed, were confirmed by the trial testimony of Dr. Meyer, the Director of the FDA's Center for Drugs and Biologics, Dr. Zuck, the Director of the FDA's Division of Blood and Blood Products, and Sidote. Accordingly, the dissent's assertion, see post at 652-53, 745 A.2d at 1207, that Abbott relied solely on its own analysis and interpretation in concluding that the Test's labeling could not have been changed, is without merit.

B. Application of Three Categories of Implied Preemption

First, the extensive control and continuous scrutiny of the Test by the FDA was so pervasive as to make reasonable the inference "that [the FDA] left no room for the state[s] to supplement it." Secondly, requiring blood banks to retest the "borderline" samples, and to warn that borderline results were inherently dangerous, would have been in direct conflict with the specific mandates of the FDA. The FDA dealt directly with the issue of how to define the cutoff, and concluded that false-negative results were not clustered around the cutoff, but were randomly spread over the entire scale of negative results. Additionally, the FDA did not only scrutinize the language required in the package insert warning, but also in large part, dictated the wording of the insert. Moreover, the FDA's regulations and its specific mandates restricted Abbott's ability to amend the package insert.^[18]

The plaintiffs' state law cause of action based on a claim of inadequate warnings, where those warnings are dictated by the FDA, is "inconsistent" with federal regulation, and is therefore preempted. Public policy justified the FDA's intense efforts to develop testing and appropriate warnings in order to combat the spread of HIV. The FDA's active involvement at every step of the test's development, approval, and use in the field, reflected the risk-utility analysis undertaken by the FDA to address significant public policy considerations.

The record establishes that the FDA believed that the Test met all specifications (including sensitivity) as set forth in the package insert, and that retesting samples near the cutoff was not scientifically supported at any time. Thus, the FDA's mandate directing Abbott not to provide for retesting of samples near the cutoff— the concept on which plaintiffs base their warnings claim—remained in force as part of a conscious ongoing risk-benefit analysis by the FDA in managing a public health crisis. Plaintiffs' state law claims would directly contradict the FDA's requirements and interfere with the FDA's objectives, and therefore are preempted.

IV.

The Court's conclusion, that plaintiffs' state law claim is preempted by the FDA's unique regulation of the Test based on principles of implied preemption rather than express preemption under the MDA, is not in conflict with the holdings of Medtronic, supra, 518 U.S. at 492-502, 116 S.Ct. at 2254-58, 135 L.Ed.2d at 720-26 (holding the MDA did not preempt plaintiffs' state common law claims for defective design, defective manufacture, and failure to warn), or Baird, supra, 155 N.J. 54, 713 A.2d 1019 (holding plaintiff's state law claims for design defect, manufacturing defect, breach of warranty, and fraudulent misrepresentation were not preempted by the MDA). This case is distinguishable because it does not merely involve the FDA's minimal ``substantial equivalent' notification process at issue in Medtronic, supra, 518 U.S. at 477-79, 116 S.Ct. at 2247, 135 L.Ed.2d at 711, or the Investigational Device Exemption process at issue in Baird, supra, 155 N.J. at 58-9, 713 A.2d 1019, or even the more rigorous Premarket Approval ("PMA") process, at issue in Mitchell v. Collagen Corp., 126 F.3d 902, 905 (1997), cert. denied, 523 U.S. 1020, 118 S.Ct. 1300, 140 L.Ed.2d 467 (Mar. 23, 1998).

Rather, Abbott's test was subject to specific requirements, unique to the HIV antibody test, that exceeded even the rigorous PMA process that a majority of courts have concluded is a specific federal requirement that triggers preemption.^[19] An examination of those cases discloses that much of the reasoning of the courts is consistent with an implied preemption analysis. See, e.g., Worthy v. Collagen Corp., 967 S.W.2d 360, 376 (Tex.1998), cert. denied, 524 U.S. 954, 118 S.Ct. 2372, 141 L.Ed.2d 740 (1998) (holding that it was the degree of FDA's involvement in the approval process of a particular "device," not the PMA process itself, that resulted in preemption of state cause of action); see also Gade, supra, 505 U.S. at 103 n. 2, 112 S.Ct. at 2386 n. 2, 120 L.Ed.2d at 88 n. 2; Tribe, supra, § 6-28 at 1177.

Nor is our holding inconsistent with the policy considerations enunciated by this Court in Feldman I, supra, 97 N.J. 429, 479 A.2d 374; Feldman II, supra, 125 N.J. 117, 592 A.2d 1176; and Feldman v. Lederle Laboratories, 132 N.J. 339, 625 A.2d 1066 (1993). Those cases involved claims of implied preemption under the FDCA in the context of the routine FDA approval of a prescription drug. In Feldman II, the Court concluded that the FDCA was intended to protect consumers, and thus preemption would be contrary to that intention because it would leave plaintiffs without a remedy. The court stated that:

immunizing a drug manufacturer against liability for marketing a product without a warning of a known or knowable risk is in conflict with Congress' well-recognized purpose in enacting the FDCA, particularly as amended in 1962. We will not recognize such an anomalous claim of immunity based solely on a manufacturer's representations of the industry's belief.

[Feldman II, supra, 125 N.J. at 154, 592 A.2d 1176.]

Although the language is compelling, it has little relevance to this case. It was the FDA's concern for the public's health that led it to specifically dictate the limitation and warning language of the Test's package insert that are the basis of plaintiffs' claims. Moreover, the Court's finding in Feldman II—that even if the drug manufacturer could not have provided a warning, it could have suspended production of the drug—demonstrates the difference between this case and that matter. Feldman II, supra, 125 N.J. at 153, 592 A.2d 1176. Although not perfect, the first ELISA HIV blood screening tests were the best devices available to the blood banks to prevent the spread of the HIV virus through transfusions from the nation's blood supply. This is not a case where a manufacturer is shielding itself from a claim by use of the general mandates of the FDA.

As the United States Supreme Court has observed, "prior cases on pre-emption are not precise ``guidelines' [since]... each case turns on the peculiarities and special features of the federal regulatory scheme in question." City of Burbank v. Lockheed Air Terminal Inc., 411 U.S. 624, 638, 93 S.Ct. 1854, 1862, 36 L.Ed.2d 547, 556 (1973). Preemption determinations are very fact-sensitive. They are "very much based on the record and context of a particular case." Chemerinsky, supra, § 5.2.5.

An examination of the cases relied on by our dissenting brethren reveals that they have forgotten that well-recognized principle. In Hillsborough, the Supreme Court found that the FDA did not intend its regulations to be exclusive. Hillsborough Cty. Fla., supra, 471 U.S. at 714-15, 105 S.Ct. at 2375-76, 85 L.Ed.2d at 721. The FDA expressly explained in a statement accompanying the original issuance of the regulations in question that "[t]hese regulations are not intended to usurp the powers of State or local authorities to regulate plasmapheresis procedures in their locality." Id. 471 U.S. at 715, 105 S.Ct. at 2375, 85 L.Ed.2d at 722. Faced with that clear statement of the FDA's intent, the Supreme Court found that neither Congress nor the FDA either expressly or impliedly preempted state and local regulations of plasamerpheresis.

Likewise, Hurley v. Lederle Laboratories Div. of Am. Cyanamid Co., 863 F.2d 1173, 1179-80 (5 th Cir.1988) and Roberson v. E.I. Dupont De Nemours & Co., 863 F.Supp. 929, 933 (W.D.Ark.1994), are distinguishable from this case because in those cases the manufacturer had withheld material information from the appropriate agency that might have affected the agency's decision regarding the content of the warning. Similarly, in Silkwood v. Kerr-McGee Corp., 464 U.S. 238, 257, 104 S.Ct. 615, 626, 78 L.Ed.2d 443, 458 (1984), the Court found that when "federal standards have been violated," paying both federal administrative fines and state tort damages does not "frustrate any purpose of the federal remedial scheme." There is no evidence that Abbott withheld any material evidence from the FDA regarding the field performance of the Test. Indeed, the record undisputedly establishes that the FDA was involved in the development and the active monitoring of the First Generation Test, as well as in the development of its Second Generation Test.

As the agency in charge of the nation's blood, the FDA determined that it was unwise, and a threat to the blood supply, to retest for false borderline negatives. The FDA weighed the harm and danger to unfortunate people, like R.F., against the greater necessity for the public's need for a safer blood supply. The record establishes that the scientific community knew that the Test had to be improved and continuously worked towards that goal. The FDA constantly monitored Abbott's Test and was always aware not only of Abbott's research, but also of the research of the scientific community to improve the Test. This was the FDA's decision; we should not second guess it.

V.

Because we conclude that the FDA's initial and continued control and monitoring of the Test preempts plaintiffs' N.J.S.A. 2A:58C-2 claim, we need not address plaintiffs' other claims of trial error. We observe, however, that a review of the record establishes that Abbott did not receive any after-acquired information that indicated that the Test's field performance was inconsistent with the FDA's pre-licensure expectations; the Test's clinical trial data; or the terms of the Test's package insert. Therefore, the state law duty to warn product users of after-acquired evidence under Feldman I, supra, 97 N.J. 429, 479 A.2d 374, is inapplicable.

In arguing to the contrary, that the plaintiffs created a material issue of fact regarding whether the Test "failed to contain adequate warnings or instructions concerning [its] known propensity to record false negative results," the dissent relies on testimony of the plaintiffs' expert (and former Abbott competitor), Dr. Nowinski, that was discredited at trial. Post at 641, 745 A.2d at 1200. Specifically, the dissent focuses on Dr. Nowinski's testimony that during the 1985-86 time period, there were four scientific reports that suggested the Test was rendering false-negative results at a rate significantly higher than its competitors' HIV blood tests, specifically: (1) a September 15, 1985 report by the Canadian Centre for Disease Control, see O'Shaughnessy, M.V., et als., Detection of Antibody to HTLV-III By Commercial Kits, Can. Med. Assoc. J., Vol. 133 (Sept. 15, 1985) ("Canadian Study"); (2) a study by Dr. Michael Ascher presented at an NIH conference in July 1986 which was never published ("Ascher Study"); (3) a study by Dr. Alfred Saah also presented at the NIH conference in July 1986, that was published in September 1987, one year after R.F.'s transfusion ("Saah Study"); and (4) a May 20, 1986 letter report from the Syracuse Red Cross to Abbott. See post at 640 - 41, 745 A.2d at 1199 - 1200.

As discussed briefly above, during the 1985-86 time period, scientists studying the HIV virus and the HIV blood tests in particular, began to "tease the virus apart" to determine which of the virus' proteins stimulated the first production of antibodies by the body. The HIV virus consists of an outer membrane (envelope) and a core. The envelope is partially comprised of a protein called "gp41" or "p41," and the core is comprised of a protein called "p24." Although there was considerable confusion and debate in the scientific community regarding the significance of a sample that contained only p24 in late 1985 and early 1986, there was considerable support for the proposition that these "p24-only" samples were positive for the HIV infection. However, in early 1986, Abbott and other AIDS researchers were finding that, in fact, p41 was the best early marker of the HIV infection, and moreover, that the appearance of p24 alone in a blood sample was not indicative of HIV infection. The problem with p24 was that it produced non-specific reactions, that is, it would react with other proteins or antibodies in the body which were not associated with the HIV virus, thereby producing false-positive results.

Dr. Meyer testified that during the experiments to determine the early marker of HIV, "it looked like p24 ... might be the antigen to which the virus might make the first antibody response." However, Dr. Meyer explained that by the summer of 1986, the data revealed that "p41 was actually the one that came up first, not p24, and you were also discovering that [the] antibody to p24 as shown by the Western Blot was frequently non-specific [, meaning "false-positive"] ..." Dr. Meyer admitted that prior to the summer of 1986, the FDA thought that Abbott was wrong in their hypothesis that gp41, rather than p24, was the best indicator of the virus, but by the summer of 1986 they "realized that Abbott was definitely right." Similarly, Dr. Zuck testified that through early 1986, the "dogma" of the scientific community was that p24 was the best marker of the infection; however, after Abbott published its experiments and results regarding the appearance of gp41 as the early marker, the scientific community recognized that the dogma "was wrong and [Abbott's] interpretation was right, and so, in fact, Abbott wasn't on the wrong track, I was wrong, we were all wrong ... so when Abbott said they were going to improve their sensitivity by adding 41, they, in fact, were doing the right thing."

The Canadian, Ascher, and Saah studies relied upon by Dr. Nowinski and the dissent were conducted during the time period when the significance of a p24-only sample was unclear to AIDS researchers. According to Dr. Nowinski, the results of those studies indicated that the Test was defective because it would not render a positive result for p-24 only samples. However, his cross-examination, and the testimony of Drs. Heller, Zuck, and Meyer, clearly indicated that Dr. Nowinski's conclusions were unreliable as they were based on a generally rejected principle of AIDS research: that HIV infection could be detected by the appearance of p24 antibody alone. In fact, during his cross-examination, Dr. Nowinski was impeached by reference to several articles and treatises, including a textbook written by Dr. Nowinski's colleague at Genetic Systems, that stated: "[w]hen Western Blot was initiated in 1985 a band indicating antibody to p-24 core protein was regarded as evidence of infection. It has now been shown that the apparent bands in this region represent artifacts and are not specific for HIV infection." Dr. Nowinski's reliance on the outdated principle was partially explained by his testimony that he had stopped following HIV literature during the mid-1980s.

Both Dr. Zuck and Dr. Heller testified regarding the indeterminacy of the Canadian, Saah, and Ascher Studies based on their use of p24-only samples. Specifically, Dr. Zuck testified that at the time Saah and Ascher presented their data at the NIH conference, the FDA was aware of new data suggesting the non-specificity of p24-only samples, and therefore was unsure how to interpret their results. Dr. Zuck testified that:

[t]he problem with the Saah presentation was that he defined an infected person as a person who had p24 only if it was a strong staining band, and neither CDC nor FDA defined ["]truly infected["] by that criteria, and so we really weren't sure whether these people were infected or not, we didn't have any more sophisticated tests at that time, and so it was unclear when he said well, Abbott or any one of those manufacturer[s] picked up X percentage, X percentage of what[?], do we know if these men are really infected[?] He also went back and reclassified a sample as positive if it later became positive but was indeterminate at the time it was studied, so we were unsure of what those percentages meant.

According to Dr. Zuck, Dr. Ascher's experiments had similar problems. Not only did Dr. Ascher include p24-only samples in his experiments, but he used an unlicensed confirmatory test (an immunofluorescence test) to determine if the samples were truly positive. In addition, Dr. Ascher's results were problematic because they were never published and his samples were unavailable to Abbott or the FDA for retesting. Dr. Zuck concluded that after Saah and Ascher presented their studies at the NIH conference, the FDA did not conclude that the HIV blood tests (Abbott or the others) were defective because, based on the developing knowledge regarding the significance of p24-only samples, the FDA was not "quite sure what either study really meant." As a general matter, Dr. Zuck stated that at no time were there any data that suggested that the "[T]est wasn't performing in full compliance with the package insert."

Similarly, Dr. Heller testified that although Abbott was concerned about Dr. Saah and Dr. Ascher's results: (1) their findings that there would be false-negatives using the Test were consistent with the package insert; and (2) it was unclear whether the samples used were truly positive because many of them were p24-only samples. For the same reasons, Dr. Heller discredited the significance of the 1985 Canadian Study that had concluded, in part, that the Test failed to detect as positive p24-only samples. Dr. Heller even pointed out that the same group of Canadian scientists published another article in July of 1986 recognizing that p24 alone was a false indication of HIV infection. See O'Shaughnessy, M.V., et als., HTLV-III Infection in Canada in 1985, Can. Med. Assoc. J., Vol. 135 (Sept. 1, 1986).

Setting aside the unreliability of the studies cited by the dissent, that evidence even if credited does not establish that the Test was performing in a manner inconsistent with the package insert—which explicitly recognized the possibility of false-negative results—or the FDA's pre-licensure expectations. In fact, as discussed above, Drs. Zuck and Meyer of the FDA both testified that there was never any field performance information that suggested that the Test was either performing worse than expected, or performing inconsistently with the warnings in the package insert. Even Dr. Nowinski did not dispute that there was no field data that was inconsistent with Abbott's representation in the package insert that the test revealed a 98.3% sensitivity rate in detecting positive samples in AIDS patients. Moreover, Dr. Nowinski testified that the results of the Canadian study, and the Saah and Ascher studies were well-known by the FDA:

[T]here was tremendous discussion amongst all the manufacturers, amongst the blood banks, amongst the sales reps who are calling on the blood banks, all of the various parties had [this data]....this type of information was very broadly disseminated ... everyone who was involved in the scientific activities knew about this, every manufacturer was aware of it and certainly the FDA was aware of it.

However, as indicated by Dr. Zuck and Dr. Meyer's trial testimony, there was nothing in the Saah, Ascher, or Canadian Studies that compelled the FDA to take any action with respect to the Test or its package insert.

Additionally, the dissent relies on a May 20, 1986 letter from the Syracuse Red Cross to Abbott that reported that in conducting a quality control experiment, the Test recorded false-negative results with respect to blood samples that had been previously confirmed as positive by Abbott's own laboratory. See post at 640, 745 A.2d at 1200. However, the record compels the conclusion that this isolated finding of the Syracuse Red Cross indicated a problem with certain lots of the Test, rather than a defect in the Test's design or general performance. First, the Red Cross' May 20, 1986 letter itself states that based on the appearance of the false negative results, it was concerned about "certain lot numbers of the current Abbott assay." Second, the blood samples used by the Syracuse Red Cross in conducting this experiment were chosen because they had initially tested positive with the Test. Third, according to Abbott's summary of the Syracuse Red Cross' data to the FDA in its May 27, 1986 submission of field data, the Red Cross' complaint with respect to the Test was described as a product quality issue regarding certain product lots, as opposed to a concern about the Test's sensitivity in general. Additionally, an Abbott Product Complaint Analysis dated April 29, 1986, regarding the Syracuse Red Cross, indicates that Abbott replaced 4,000 of the Red Cross' Test kits because the Red Cross believed that certain lots of the Test were defective. That remedial measure is consistent with an interpretation that the problems with the Test were based on the quality of specific lots of the Test, and not the Test generally.

The dissent also relies on Abbott's July 28, 1986 product licensing amendment and January 19, 1987 mailgram regarding approval and manufacturing of the Second Generation Test. See post at 641 - 42, 745 A.2d at 1200-01. However, those documents are evidence only that Abbott believed that the Second Generation Test would have an improved sensitivity over the First Generation Test—a result contemplated by Abbott and the FDA even before the First Generation Test was marketed.

Finally, in support of its argument that there was after-acquired evidence suggesting the Test was defective, the dissent relies on: (1) a June 28, 1993 Wall Street Journal article that was specifically excluded by the trial court, and discredited during the voir dire testimony of Dr. Zuck; and (2) the findings of a federal district court in another action, Genetic Sys. Corp. v. Abbott Laboratories, 691 F.Supp. 407 (D.D.C.1988), neither of which are part of the record on appeal to this Court. See post at 642 - 43, 745 A.2d at 1201. However, we observe that the dissent's assertion, post at 643, 745 A.2d at 1201 - 02, based on Genetic Sys., that Abbott had knowledge not known to FDA that its First Generation Test was not as sensitive as the tests of its competitors, is totally baseless. First the dissent relies on an anti-trust action instituted by Dr. Nowinsky, plaintiffs' main expert and the CEO of Genetic Systems, a competitor of Abbott, against the American Red Cross and Abbott alleging violations of the Sherman and Clayton Acts; tortious interference against Abbott for prospective economic advantage; and unfair competition in that Abbott engaged intentionally to deprive Genetic Systems of actual and prospective business relationships, advantages, and opportunities. Moreover, unlike this case, where a jury in a five week trial heard all the evidence that is now before this Court, in Genetic Sys., the evidence the dissent relies on was accepted as true by the trial court on the motion to dismiss the complaint, and therefore was never heard by a jury.

VI.

This case merits preemption because of its unique facts. In addition to the Test being developed expediently in the face of a national crisis, the FDA: (1) actively directed Abbott concerning the design format of the test; (2) determined the appropriate cutoff; (3) dictated the exact wording of Abbott's package insert; (4) directed Abbott not to instruct blood banks to retest samples below the cutoff; and (5) actively monitored both the First Generation Test's field performance, and the development of the Second Generation Test. As modified, we affirm the judgment of the Appellate Division.

STEIN, J., dissenting.

Underlying this appeal are the allegations of plaintiffs concerning R.F.'s receipt in September 1986 of a transfusion of a unit of blood that was infected with the human immunodeficiency virus (HIV) causing her subsequently to test positive for the presence of that virus. Prior to R.F.'s hip-replacement surgery in 1986 she expressed concern to her surgeon about the safety of donated blood, but was assured by him that blood donated to the Bergen County Blood Center (Blood Center) would be tested and could safely be used in the event she required a blood transfusion during surgery. The blood unit received by plaintiff was supplied by a donor to the Blood Center and was tested for HIV antibodies on August 26, 1986 using defendant Abbott Laboratories' (Abbott) HTLV III test kit. When that unit of blood was tested, it tested negative for the HIV virus with a test result of 0.121. The Abbott Test Kit instructed users that the cut-off value for a negative unit was 0.128, and that units of blood with readings below that cut-off level should be used and not retested.

The essence of plaintiffs' contentions in this litigation is that Abbott knew, on the date the infected unit of blood was tested at the Blood Center and for several months prior thereto, that its test tended to report HIV positive samples of blood as false negatives to a substantially greater extent than did the tests of its two primary competitors; that prior to the date the infected unit of blood was tested Abbott had submitted to the FDA a product license amendment to obtain approval for a revised test that "would be better able to detect positive samples that were ``borderline or negative' by the First Generation Test," ante at 613, 745 A.2d at 1184; and that Abbott failed to warn users of its test concerning its tendency to record false negative readings and failed to request authorization from the FDA to issue such a warning to users.

Notwithstanding those contentions, the Court today holds that plaintiffs' state law failure-to-warn claim is impliedly preempted—not expressly preempted pursuant to the Medical Device Amendments of 1976, 21 U.S.C.A. § 360k, to the Federal Food, Drug, and Cosmetic Act, 21 U.S.C.A. § 301 to § 397 (FDCA), as the Appellate Division held—because of the comprehensiveness of the regulatory control over the Abbott Test exercised by the FDA. Because in my view the Court's preemption analysis exaggerates the legal significance of the FDA's regulatory role, understates the materiality of the factual dispute concerning information about the test's deficiencies allegedly acquired by Abbott after licensing of its test in March 1985, and diminishes the precedential effect of this Court's federal preemption jurisprudence, I am compelled to dissent.

I

A

Because this case was tried to a jury on the basis of the trial court's conclusion that plaintiff's failure-to-warn claim was not preempted, and the jury returned a verdict for Abbott, the relevance of the majority's holding that federal preemption bars plaintiffs' claim is understood only in the context of the Appellate Division's determination that—absent preemption—plaintiffs would be entitled to a new trial because of reversible error in the trial court's charge and in the jury verdict sheet. As the Appellate Division noted, both that portion of the trial court's instruction to the jury and the first jury interrogatory that focused on the statutory defense to product liability actions, N.J.S.A. 2A:58C-3a(3), precluding liability if the harm "was caused by an unavoidably unsafe aspect of the product and the product was accompanied by an adequate warning," misdirected the jury's attention to whether the "blood supply," rather than Abbott's HTLV Test Kit, was unavoidably unsafe. The language of the first interrogatory on the jury Verdict Form illuminates the issue:

Has Abbott Laboratories demonstrated by a preponderance of the evidence that the blood supply was unavoidably unsafe at the time of plaintiff R.F.'s transfusion and that the risk of HIV infection could not be completely eliminated through testing?

[emphasis added.]

As the Appellate Division observed, "the issues before the jury included whether the defendant's product was ``unavoidably unsafe' ... and the adequacy of the warning..., not whether the blood supply was unsafe irrespective of defendant's test." (emphasis added). The interrogatories on the verdict sheet addressed to the two other statutory defenses, N.J.S.A. 2A:58C-3a(1) and (2), also incorrectly directed the jury to consider whether the blood supply was unavoidably unsafe, rather than requiring the jury to focus on the alleged deficiency in Abbott's test.

Moreover, interrogatory number six on the verdict sheet, addressing the failure-to-warn claim, incorrectly enhances the plaintiffs' burden by requiring the jury to find not only that the warnings accompanying Abbott's test were inadequate to warn the Blood Center of the limitations of the test, but also that the test was not reasonably fit, suitable or safe when it left Abbott's control. That interrogatory stated:

Taking into consideration your answer to question no. 5, have the plaintiffs demonstrated by a preponderance of the evidence that the warnings and instructions that accompanied the Abbott Laboratories test were inadequate to warn the Bergen Community Blood Center of the limitations of the test, and that the test was not reasonably fit, suitable or safe when it left the control of Abbott Laboratories?

The appropriate burden of proof for a plaintiff to sustain a failure-to-warn claim is that "the product causing the harm was not reasonably fit, suitable or safe for its intended purpose because it ... failed to contain adequate warnings or instructions." (emphasis added). N.J.S.A. 2A:58C-2. Interrogatory number six, however, required the plaintiffs to prove not only that Abbott's warnings were inadequate, but in addition, that the Abbott test also was not reasonably fit, suitable or safe. See Feldman v. Lederle Labs., 125 N.J. 117, 144, 592 A.2d 1176 (1991) (Feldman II); Freund v. Cellofilm Properties, Inc., 87 N.J. 229, 242, 432 A.2d 925 (1981). I am fully in accord with the Appellate Division's conclusion that "the lengthy charge as a whole and the [jury] interrogatory were prejudicially confusing in light of the focus of the inquiry on the blood supply, not the defendant's test."

B

The Court's opinion, which meticulously recites the facts concerning the FDA's supervisory role in reviewing the Abbott blood test prior to licensing in March 1985, ante at 600 - 18, 745 A.2d at 1176-86, attempts to diminish the significance of the trial testimony elicited by plaintiffs to demonstrate that after licensure, and specifically between September 1985 and July 1986,—the crucial period before plaintiff's transfusion—Abbott acquired information suggesting that its test frequently failed to detect blood that was confirmed to be HIV positive, and that in two scientific studies its test was shown to be substantially less capable of detecting HIV positive blood than were the blood tests of its primary competitors.

Dr. Robert Nowinski, plaintiffs' expert witness (who had been employed by a competitor of Abbott), offered the following detailed evidence of reports and studies suggestive of deficiencies in the Abbott blood test that were known to Abbott subsequent to licensure and prior to R.F.'s transfusion:

(a) A September 15, 1985 report by the Canadian Red Cross concerning seven blood samples that were verified as positive by the Western Blot test but were not detected by the Abbott test;

(b) The Syracuse Red Cross correspondence of May 20, 1986, in which blood confirmed to be HIV positive by the Western Blot test was tested three separate times by the Abbott test, and that test recorded negative results on all three occasions;

(c) The Ascher study conducted at the University of California Laboratory at Berkeley involved twelve blood samples of blood known to be HIV positive. The Abbott test recorded negative results on all twelve samples. Two other commercial tests using the same blood sample reported positive results for all twelve samples;

(d) The study conducted by Dr. Saah, which Dr. Nowinski described as the "Mac Study" involving 5,000 homosexual individuals monitored for HIV antibodies. Within that study entitled "The Multi-Center AIDS Cohort Study" presented at the National Institute of Health Consensus Conference on July 7, 1986, the data focused on thirty homosexual males who had been tested every six months in order to determine when "sero" conversion occurs, described by plaintiff's expert as the stage in the development of disease when the blood sample converted from HIV negative to HIV positive. The thirty blood samples were tested by test kits of five manufacturers all approved by the FDA. The manufacturers were Litton, Electro-nucleonic, Abbott, DuPont, and Genetic. The rate of detection of various manufacturers in the Saah study was as follows: Litton 2/30; Electro-nucleonic 4/30, Abbott 13/30, DuPont 25/30, Genetic Systems 25/30. Abbott's detection rate was 43% whereas DuPont and Genetic Systems detection rates were 83%.

Dr. Nowinski also supported his conclusion that Abbott knew that its original test recorded false negative results close to the cutoff by referring to Abbott's letter to the FDA dated July 28, 1986 proposing a modification of Abbott's blood test. That letter stated that "the addition of affinity purified p41 derived from HTLV III viral lysate into the current bead coating antigen solution enhanced the detection of samples containing anti-HTLV III that are borderline or negative by the current assay." (emphasis added). To the same effect, Dr. Nowinski quoted from the mailgram sent by Abbott to its customers on January 19, 1987 to announce the approval to its modified blood test:

The improved assay significantly enhances the detection of early sero conversion and low level antibody positive specimens. In clinical comparisons between the improved assay and other licensed assays, it demonstrated significantly higher signals on borderline low-level positive specimens. This increased signal pushes low-level samples farther above the cutoff and assures their detection.

[emphasis added.]

Plaintiffs' allegations concerning Abbott's knowledge that its blood test frequently produced false negative results and was less sensitive in detecting HIV contaminated blood than were the tests of its primary competitors was corroborated by other sources. The trial court, in twice denying Abbott's motions for summary judgment, referred to a June 28, 1993 Wall Street Journal article included in plaintiffs' certifications that discussed a pending congressional investigation concerning the "sensitivity problem with the original Abbott test." Similarly in Genetic Systems Corporation v. Abbott Laboratories, 691 F.Supp. 407 (D.D.C.1988), the United States District Court for the District of Columbia, in an opinion granting summary judgment to the American National Red Cross (Red Cross) and partial summary judgment to Abbott on plaintiffs' anti-trust claims, made factual findings that took note of the substantial dissatisfaction within the Red Cross in the summer and fall of 1986 over the lack of sensitivity of the Abbott blood test.

In July 1986, studies presented at the National Institutes of Health Conference on AIDS indicated that the Abbott AIDS test then in use by the Red Cross was less sensitive to detecting contaminated blood than the test of Genetic Systems. That same month, Victor W. Schmitt, Vice President of Medical Operations of the Red Cross, inquired of Genetic Systems how long it would take for it to supply all Red Cross Regions with their requirements for AIDS tests. Genetic Systems advised the Red Cross that it could do so within four weeks. In the meantime, the Red Cross continued to use the Abbott test and Abbott sought FDA approval of a new, "improved" AIDS test.

By November 1986, the Red Cross was aware that the new Abbott test would not be available until December 1986; that availability date was later delayed to January 1987. The Red Cross became frustrated at the regulatory delays, however, and developed various strategies in response to the problem. Initially, the Red Cross adopted a "contingency plan" under which it would become a "co-investigator" with Abbott to facilitate licensure by the Food and Drug Administration. This contingency plan was never put into effect, however, and the Red Cross adopted a second contingency plan to purchase blood test equipment from alternative suppliers. This second plan was also never implemented because Abbott's new test received FDA approval in January 1987.

One month after Abbott received approval for its new test, the Red Cross received proposals from Genetic Systems, Abbott, and DuPont to supply the Red Cross' requirements for AIDS and the two hepatitis tests for a multi-year period commencing July 1, 1987.

The record clearly reflects, and defendants do not deny, that a variety of views, some "strongly held and sharply conflicting," emerged within the Red Cross as to the desirability of contracting with the various test equipment suppliers for the next contract period. For example, in March 1987, the Red Cross Task Force found that the test kits of Genetic Systems, DuPont, and Abbott "all appear to perform with comparable sensitivity and specificity at this time" and would be "acceptable" but recommended that Genetic Systems continue to have a "significant part" of the Red Cross' AIDS test business in light of the dissatisfaction with the Abbott test in 1986.

[Id. at 411-12 (footnotes omitted) (emphasis added).]

The Court responds to plaintiffs' trial testimony indicating that the Abbott blood test had a tendency to produce false negative results by asserting that the testimony was "discredited at trial." The Court supports that assertion by citations to the direct testimony of three scientists, Harry Meyer, Director of the FDA's Center for Drugs and Biologics until September 1986 and the official with primary responsibility in the FDA for licensing and implementation of the Abbott test; John Heller, an executive at Abbott since 1983 who was Abbott's "lead scientist" in the development of the Abbott test; and Thomas Zuck, Director of the FDA Division of Blood and Blood Products during the period of licensure of the Abbott test and, after leaving the FDA in 1987, a consultant for Abbott Laboratories for the past "four or five years" prior to 1996 when he testified. Putting to one side the fact that defendant's experts were no more impartial than was plaintiffs' expert, the thrust of their testimony was to suggest, but not conclude, that the blood samples used in the Canadian, Saah and Ascher studies may not have been HIV positive after all. The witnesses testified that at the time the Canadian, Saah and Ascher studies were conducted many scientists believed that a protein called "p24" was the earliest and clearest indicator of infected blood, although subsequently Abbott and other scientists concluded that a different protein, "p41," found in the outer membrane of the HIV virus was a more reliable indicator.

Accordingly, Meyer, Heller, and Zuck implied, without ever challenging the statistical results of the Canadian, Saah and Ascher studies, that some of the Abbott test's "false-negative" results in those studies may have been correct in that all of the blood samples may not have been positive. Ante at 630-35, 745 A.2d at 1194-97.

Far from repudiating the results of those studies, Abbott's witnesses instead acknowledged that the studies created doubt and uncertainty. Dr. Zuck, responding to questions concerning the Saah study, observed that "we weren't exactly sure what p24 only meant at that point ... and so we were having growing concern about how the knowledge was evolving," and concerning the Ascher study he noted that "we were again left with the quandary what does this really mean, who among these people is infected and who isn't." Dr. Meyer, the FDA official with direct responsibility for the quality of licensed blood tests, states that "[m]y reaction [to the studies] was one of interest, but recognizing some of the potential flaws and [not] racing off to over interpret the data.... Those were very preliminary data. I was interested in knowing more about it, and we did generate more data about it." Even Dr. Heller, the Abbott scientist in charge of developing the Abbott test, grudgingly acknowledged concern and the need for further inquiry about the Saah and Ascher studies: "Oh, exactly, that's the responsible thing to do scientifically, to try and investigate and understand what is being presented."

The obvious point is that, however persuasive or unpersuasive a jury might find the testimony of each side's experts on the issue of the Abbott test's propensity to record false negatives, a material question of fact on that issue is presented by the proofs as the Law Division judge twice concluded. The appropriate mechanism for resolving that factual issue is by a remand for a new trial. The Court, however, forecloses plaintiffs' opportunity to prove their case by implying preemption of their cause of action even in the absence of any statutory foundation or promulgated regulatory basis for that conclusion.

II

Preemption

The Court's implied preemption analysis and rationale differs markedly from that of the Appellate Division. That court assumed, consistent with the parties' presentations, that the Abbott blood test was licensed as a medical device pursuant to the Medical Device Amendments of 1976, 21 U.S.C.A. §§ 360c-360ee (MDA), to the Federal Food, Drug, and Cosmetic Act, 21 U.S.C.A. §§ 301-397. Before this Court, however, plaintiffs persuasively asserted that although Abbott's blood test kit met the definition of a "device" pursuant to the MDA, the FDA regarded the test kit's virus as its primary component. Plaintiffs submitted incontrovertible documentary evidence to the Court that the product license authorizing the use of that virus in Abbott's test kit was issued not under the MDA, but rather by the Office of Biologics Research and Review pursuant to the Public Health Service Act, 42 U.S.C.A. § 262. The Court offhandedly observes, without any explanation, that plaintiffs' assertion that the test kit was approved as a "biologic," not as a "device," is "unsupported by the record" and "irrelevant" to the Court's conclusion. Ante at 619 n. 14, 745 A.2d at 1187 n. 14. Nevertheless, the Court abandons all reliance on the MDA's express preemption provision, 21 U.S.C.A. § 360k, and rests its preemption holding on the doctrine of implied preemption. One can reasonably infer that the Court would not have eschewed all reliance on the MDA's express preemption provision unless it concluded that it was compelled to do so.

The broad basic principles that govern federal preemption of state law are well settled, but their proper application to specific cases can be elusive. We recently restated the controlling general principles in Franklin Tower One, L.L.C. v. N.M., 157 N.J. 602, 615-16, 725 A.2d 1104 (1999):

There are several theories under which federal law will preempt a state statute. We begin by noting that "pre-emption is not to be lightly presumed," and that "the historic police powers of the States [are] not to be superseded by [federal law] unless that was the clear and manifest purpose of Congress." The party claiming preemption bears the burden of supporting that claim by "clear and manifest evidence."

Congress explicitly may express its intent to preempt state law. Alternatively, preemption may be inferred where the federal legislation is so comprehensive that it creates the inference that Congress intended to leave no room for state regulation in the area.

Preemption also may be found where state law actually conflicts with federal law. Conflict preemption occurs in two instances: where "compliance with both federal and state regulations is a physical impossibility," or where a state law "stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress." Under conflict preemption analysis, a court first must consider the purposes of the federal law, and then evaluate the effect of the state law on those purposes.

[ (citations omitted.) ]

The Court primarily rests its implied preemption analysis on "field" preemption, relying on the comprehensive FDA regulation of the Abbott HTLV III Test Kit, ante at 625, 745 A.2d at 1191. But the Court also finds both variants of implied conflict preemption to exist, apparently concluding that Abbott could not provide any additional warnings pursuant to state law without violating the FDA's labeling requirements, and that added warnings would serve as an obstacle to the FDA's regulatory goals. Ibid.

Field Preemption

The Court's preemption discussion relies most heavily on the doctrine of "field" preemption, and assumes that the FDA's unique regulation and supervision of all aspects of Abbott's blood test during the licensing process necessarily excludes any exposure of Abbott to state product liability claims based on inadequate warnings.

The United States Supreme Court precedent applying field preemption appears to emphasize a Congressional purpose to occupy a specific area of regulation to the exclusion of state law. In Schneidewind v. ANR Pipeline Company, 485 U.S. 293, 108 S.Ct. 1145, 99 L.Ed.2d 316 (1988), two natural gas companies challenged the authority of the Michigan Public Service Commission to regulate their issuance of long-term securities, contending that the Federal Energy Regulatory Commission (FERC), the regulatory body charged with implementation of the Natural Gas Act of 1938 (NGA), 15 U.S.C.A. § 717, possessed exclusive jurisdiction to regulate all aspects of operations and financing of natural gas companies. The Court observed that the NGA consistently has been recognized as a comprehensive scheme of federal regulation over natural gas in interstate commerce, id. at 300, 108 S.Ct. at 1151, 99 L.Ed.2d at 325, that in exercising its rate-making authority FERC possesses the power to control indirectly the capital structure of gas companies, id. at 302, 108 S.Ct. at 1151, 99 L.Ed.2d at 326, and that FERC authority over the issuance of certificates of public convenience and necessity empowers FERC to approve the manner in which capital construction undertaken by gas companies will be financed. Id. at 302-03, 108 S.Ct. at 1152, 99 L.Ed.2d at 326-27. The Court observed that the Michigan statute addressed issues concerning rates and capitalization of gas companies that Congress delegated exclusively to FERC:

In this case we are presented with a state law whose central purpose is to regulate matters that Congress intended FERC to regulate. Not only is such regulation the function of the federal regulatory scheme, but the NGA has equipped FERC adequately to address the precise concerns Act 144 purports to manage.

[Id. at 308-09, 108 S.Ct. at 1155, 99 L.Ed.2d at 330-31.]

Similarly, in City of Burbank v. Lockheed Air Terminal, Inc., 411 U.S. 624, 93 S.Ct. 1854, 36 L.Ed.2d 547 (1973), an action was brought by the owner and operator of an airport and an interstate air carrier to invalidate a city ordinance prohibiting jet aircraft from taking off from the airport between 11:00 p.m. and 7:00 a.m. The plaintiffs contended that the ordinance was invalid because it was preempted by the provisions of the Federal Aviation Act of 1958, 49 U.S.C.A. § 1301, as amended by the Noise Control Act of 1972, 42 U.S.C.A. § 4901. By a 5-4 vote, the Court determined that although the federal legislation contained no express preemption provisions, Congress unmistakably intended to occupy the entire field of aircraft noise regulations:

Our prior cases on pre-emption are not precise "guide-lines" in the present controversy, for each case turns on the peculiarities and special features of the federal regulatory scheme in question. Control of noise is of course deepseated in the police power of the States. Yet the pervasive control vested in EPA and in FAA under the 1972 Act seems to us to leave no room for local curfews or other local controls. What the ultimate remedy may be for aircraft noise which plagues many communities and tens of thousands of people is not known. The procedures under the 1972 Act are under way. In addition, the Administrator has imposed a variety of regulations relating to takeoff and landing procedures and runway preferences. The Federal Aviation Act requires a delicate balance between safety and efficiency, and the protection of persons on the ground. Any regulations adopted by the Administrator to control noise pollution must be consistent with the "highest degree of safety." The interdependence of these factors requires a uniform and exclusive system of federal regulation if the congressional objectives underlying the Federal Aviation Act are to be fulfilled.

[Burbank, supra, 411 U.S. at 638-39, 93 S.Ct. at 1862, 36 L.Ed.2d at 556-57 (citations and footnote omitted).]

In contrast, in Silkwood v. Kerr-McGee Corporation, 464 U.S. 238, 104 S.Ct. 615, 78 L.Ed.2d 443 (1984), the issue was whether a state-authorized award of punitive damages based on the discharge of plutonium from a federally-licensed nuclear facility was preempted on the basis of the Atomic Energy Act's comprehensive authority over all safety aspects of nuclear energy. The Court noted that it recently had held, based on an examination of the legislative history of the Atomic Energy Act, that Congress intended the federal government to "occup[y] the entire field of nuclear safety concerns, except the limited powers expressly ceded to the states." Id. at 249, 104 S.Ct. at 621-22, 78 L.Ed.2d at 453 (quoting Pacific Gas & Elec. Co. v. State Energy Resources Conservation & Dev. Comm'n., 461 U.S. 190, 212, 103 S.Ct. 1713, 1726, 75 L.Ed.2d 752, 770 (1983)). Nevertheless, after examining the relevant legislative history and related federal statutes, the Court concluded that Congress did not intend to override traditional state tort law:

In sum, it is clear that in enacting and amending the Price-Anderson Act, Congress assumed that state-law remedies, in whatever form they might take, were available to those injured by nuclear incidents. This was so even though it was well aware of the NRC's exclusive authority to regulate safety matters. No doubt there is tension between the conclusion that safety regulation is the exclusive concern of the federal law and the conclusion that a state may nevertheless award damages based on its own law of liability. But as we understand what was done over the years in the legislation concerning nuclear energy, Congress intended to stand by both concepts and to tolerate whatever tension there was between them. We can do no less. It may be that the award of damages based on the state law of negligence or strict liability is regulatory in the sense that a nuclear plant will be threatened with damages liability if it does not conform to state standards, but that regulatory consequence was something that Congress was quite willing to accept.

[Id. at 256, 104 S.Ct. at 625, 78 L.Ed.2d at 457.]

The "field preemption" case most analogous to the issues presented by this appeal is Hillsborough County, Florida v. Automated Medical Laboratories, Inc., 471 U.S. 707, 105 S.Ct. 2371, 85 L.Ed.2d 714 (1985). In Hillsborough an operator of blood plasma centers throughout the country challenged on preemption grounds the validity of an ordinance adopted by Hillsborough County that regulated blood plasma centers and their donors. The ordinance was challenged on the ground that Congress, pursuant to the Public Health Service Act, 42 U.S.C.A. § 262(a), had provided for the licensing of all blood plasma vendors and, pursuant to comprehensive regulations adopted by the Food and Drug Administration, had established detailed and specific standards governing the collection of blood plasma. Id. at 709-10, 105 S.Ct. at 2373, 85 L.Ed.2d at 718-19. Plaintiff argued, relying primarily on the pervasiveness of the FDA's regulatory scheme, that all state and local regulation of plasma collection was preempted. The Court rejected plaintiff's reliance on the "field" preemption doctrine, noting its concern that to rest federal preemption on the comprehensiveness of agency regulation could jeopardize the Court's Supremacy Clause jurisprudence:

We are even more reluctant to infer pre-emption from the comprehensiveness of regulations than from the comprehensiveness of statutes. As a result of their specialized functions, agencies normally deal with problems in far more detail than does Congress. To infer pre-emption whenever an agency deals with a problem comprehensively is virtually tantamount to saying that whenever a federal agency decides to step into a field, its regulations will be exclusive. Such a rule, of course, would be inconsistent with the federal-state balance embodied in our Supremacy Clause jurisprudence.

Moreover, because agencies normally address problems in a detailed manner and can speak through a variety of means, including regulations, preambles, interpretive statements, and responses to comments, we can expect that they will make their intentions clear if they intend for their regulations to be exclusive. Thus, if an agency does not speak to the question of pre-emption, we will pause before saying that the mere volume and complexity of its regulations indicate that the agency did in fact intend to pre-empt. Given the presumption that state and local regulation related to matters of health and safety can normally coexist with federal regulations, we will seldom infer, solely from the comprehensiveness of federal regulations, an intent to pre-empt in its entirety a field related to health and safety.

[Id. at 717-18, 105 S.Ct. at 2377, 85 L.Ed.2d at 724 (emphasis added) (citations omitted).]

The majority's answer to the Supreme Court's warning in Hillsborough that the comprehensiveness of an agency's regulatory efforts is a weak reed on which to rest preemption completely misses the mark. The Court responds by noting that the Supreme Court has held "repeatedly that state laws can be preempted by federal regulations as well as by statutes." Ante at 619, 745 A.2d at 1187. That statement is accurate but inapposite, because the FDA promulgated no preemptive regulations in dealing with HIV blood tests. As the Supreme Court cautioned in Hillsborough, "if an agency does not speak to the question of preemption, we will pause before saying that the mere volume and complexity of its regulations indicate that the agency did in fact intend to preempt." Id. at 718, 105 S.Ct. at 2377, 85 L.Ed.2d at 724 (emphasis added).

This Court's willingness to rest its "field" preemption analysis solely on the basis of the comprehensiveness of the FDA's regulation of Abbott's test prior to licensure is suspect on two grounds. First, it disregards the Supreme Court's admonition in Hillsborough that "we will seldom infer, solely from the comprehensiveness of federal regulations, an intent to pre-empt in its entirety a field related to health and safety." Ibid. The Court's cautionary language in Hillsborough is consistent with the basic principle that "the historic police powers of the States were not to be superseded by [federal law] unless that was the clear and manifest purpose of Congress." Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230, 67 S.Ct. 1146, 1152, 91 L.Ed. 1447, 1459 (1947) (citations omitted). Although the Court's opinion attempts to justify federal preemption not on the basis of Congressional action but rather on the basis of a national health emergency concerning a safe blood supply, ante at 620 - 21, 745 A.2d at 1188, experience informs us that Congress knows how to mandate federal preemption in the context of a national health emergency. During the crisis that arose over the swine flu epidemic in the mid-seventies, Congress enacted federal legislation that immunized from liability the manufacturer of the swine flu vaccine and imposed liability on the federal government for state tort claims arising out of use of the vaccine. See National Swine Flu Immunization Program of 1976, Pub.L. No. 94-380, 90 Stat. 1113 (1976); Petty v. United States, 740 F.2d 1428 (8th Cir.1984). In contrast, although Congress obviously was aware of the blood supply concerns associated with the AIDS epidemic, it took no action to preempt state law tort claims. The Court skates on thin jurisprudential ice when it implies such preemption solely on the basis of an administrative agency's understandably cautious and thorough procedure for licensing blood tests.

Second, the Court either overlooks or disregards that the plaintiffs' focus is not on the FDA's regulation prior to licensure, but on Abbott's after-acquired knowledge in the summer of 1986 that its test's tendency to record false negatives around the cutoff required prompt remediation. The record demonstrates that during this post-licensure period Abbott was concerned about its test, not only because of safety concerns, but because the blood tests of competitors appeared, according to some research reports, to be more sensitive in detecting HIV contaminated blood. Abbott knew enough about its test's deficiencies to propose to the FDA in July 1986 a modified test that "would be better able to detect positive samples that were ``borderline or negative' by the First Generation Test." Thus, the focus of plaintiffs' argument shifts from the FDA's comprehensive regulation prior to licensure, to Abbott's state of knowledge in the summer of 1986 about apparent deficiencies in its own test, and its duty to warn users of those deficiencies. Accordingly, the Court's application of the doctrine of "field preemption" to Abbott's duty to warn in the summer and fall of 1986 is unsupported by federal preemption jurisprudence and, as well, by the record that suggests the FDA's regulatory efforts peaked when the Abbott blood test was licensed in 1985, and that after licensure the FDA impatiently awaited Abbott's submission of a modified and unproved test.

Conflict Preemption

The majority opinion asserts that, in addition to finding implied "field" preemption of plaintiffs' claims based on the comprehensiveness of FDA regulation, a state failure-to-warn claim would be subject to implied conflict preemption because the FDA designed the label on the Abbott test and at the time of licensure precluded any change in the label without prior FDA approval. The Court's conflict preemption analysis, authored by the sole dissenting Justice in Feldman v. Lederle Laboratories, 125 N.J. 117, 158, 592 A.2d 1176 (1991) (Garibaldi, J. dissenting), overstates the preemptive effect of the FDA-designed label and resurrects conflict preemption obstacles to state tort law that were definitively addressed and rejected in Lederle.

The federal courts clearly have recognized that the preemptive effect of a warning designed or approved by a federal agency depends on whether the regulated entity has disclosed all relevant information to the agency. See, e.g., Hurley v. Lederle Laboratories, 863 F.2d 1173, 1179-80 (5th Cir.1988) (holding that notwithstanding FDA approval of warning on DPT vaccine, preemptive effect of FDA action depends on jury's determination whether manufacturer withheld from FDA information that might have affected FDA's decision on content of warning); Roberson v. E.I. Dupont De Nemours & Co., 863 F.Supp. 929, 933 (W.D.Ark.1994) (holding that manufacturer could be estopped from asserting federal preemption of failure-to-warn claim if manufacturer withheld material facts from Environmental Protection Agency).

The Court also relies on the applicable regulation promulgated pursuant to the Public Health Service Act, 21 C.F.R. 601.12(b), that precluded changes in the labeling of biologics prior to "acceptance" of the changes from the Director, Office of Biologics Research and Review. This Court addressed and rejected precisely that contention in Lederle:

The assertion that Lederle would have been subject to punishment for "misbranding" or would have had its product barred from the market if it had added an unapproved warning is based merely on Lederle's interpretation of the regulations then in effect. Lederle presents no evidence indicating that the FDA ever took such action for adding a warning suggesting a limitation on use rather than an unapproved representation of a benefit, nor does the record reflect any statement by the FDA to the effect that it would have taken action had a manufacturer attempted to warn without prior approval.

[125 N.J. at 148, 592 A.2d 1176.]

Just as did Lederle, the Court cites to subsequent modifications of FDA regulations relating both to medical devices and biologics that, after the critical events here, would permit manufacturers to augment their labels with additional warnings without prior FDA approval, asserting that those revised regulations demonstrate that in 1986 Abbott was prohibited from revising its warning without prior approval. Ante at 621-22, 745 A.2d at 1188-89. We also considered and rejected an analogous argument in Lederle:

Although we recognize that subsequent actions of an agency are relevant to determining prior intent, we do not agree that the conclusion that Lederle asserts necessarily follows. It contends that the modification shows that independent action was not permitted under the former regulations. An equally-compelling conclusion, however, is that the FDA had not addressed itself to the issue before. ("One could argue that the changes made in § 130.9 in 1965 * * * indicate that FDA approval of precautionary information was required prior thereto. One could also argue that the change in language was meant to clear up a misconception."). Arguably, then, on squarely considering the issue, the FDA determined that warning of possible dangerous side effects "at the earliest possible time," was consistent with its primary purpose to protect the public health.

[Id. at 149, 592 A.2d 1176 (citations omitted).]

Despite persuasive evidence in this record about the deficiencies in Abbott's blood test and its propensity for recording false negatives, and the evidence revealed in the Genetic Systems litigation, supra, at 642, 745 A.2d at 1200-01, that Abbott was in jeopardy of losing Red Cross contracts because of the unreliability of its test, no evidence in the record reveals or suggests that Abbott ever attempted to persuade the FDA to modify the test's label to warn users to retest in the event of negative readings close to the cutoff. The Court concludes, however, that Abbott was helpless, and that no matter what it knew about the deficiencies in its test, it was absolutely barred by the FDA from taking any action to warn users. This Court evaluated and rebuffed similar protestations in Lederle:

We also note that Lederle is not faced with the Hobson's choice of either complying with federal regulations and continuing to be subject to damages in state tort actions or providing additional warnings and thereby violating federal law. The regulations in question have long since been changed expressly to allow pharmaceutical companies to implement necessary changes in labeling while a supplemental new-drug application is pending. Furthermore, the present regulations governing labeling of prescription drugs require that the labeling "describe serious adverse reactions and potential safety hazards, [and] limitations in use imposed by them." Interestingly, that provision states that "labeling shall be revised to include [such] a warning as soon as there is reasonable evidence of an association of a serious hazard with a drug; a causal relationship need not have been proved."

The issue then is what choices, if any, did Lederle have in 1962 and 1963. Even if we assume Lederle could not have provided a warning, an assumption we find unsupported by the evidence, it could have suspended production of Declomycin for what it termed the "probably * * * short period of time" before which Lederle anticipated that that drug too would be "officially implicated with tooth staining." Alternatively, Lederle could have quite reasonably and responsibly removed from the market those forms and methods of packaging of Declomycin that were geared toward pediatric use. It could have raised the price as a form of loss-spreading insurance against claims. Lederle also could have continued to press for FDA approval of a warning for Declomycin after the administration's initial response declining to change its official position. (The record is conspicuously silent on any continuing efforts by Lederle to confirm the association of Declomycin with tooth staining or to gain FDA sanction of a warning for Declomycin. ("at least in some fields, such as those impacting on public health, a manufacturer may be expected to be informed and affirmatively to seek out information concerning the public's use of its own product").) The suggestion that Lederle was in some way compelled to continue distributing a drug without warning of the strong likelihood of serious side effects for some child users and that Congress and the FDA intended that it would be immune from liability for doing so cannot be reconciled with the primary purpose of the FDA to promote and protect the health of the citizens of the United States.

[Id. at 153-54, 592 A.2d 1176 (citations omitted).]

In finding that plaintiffs' state law claims are preempted no matter what the evidence at a retrial may disclose about Abbott's knowledge of deficiencies in its blood test—deficiencies alleged to pose a severe threat to human life—this Court's holding threatens the "federal-state balance embodied in our Supremacy Clause jurisprudence." Hillsborough, supra, 471 U.S. at 717, 105 S.Ct. at 2377, 85 L.Ed.2d at 724. The Court also disregards Justice Frankfurter's admonition that "[a]ny indulgence in construction should be in favor of the States, because Congress can speak with drastic clarity whenever it chooses to assure full federal authority." Bethlehem Steel Co. v. New York State Labor Relations Bd., 330 U.S. 767, 780, 67 S.Ct. 1026, 1033, 91 L.Ed. 1234, 1249 (1947). Finally, the majority disregards or ignores this Court's forceful holding in Lederle, supra, rejecting contentions that mirror Abbott's assertions in this appeal, and declining to imply federal preemption of plaintiff's state law tort claim:

The presence of the following factors reinforces the traditional presumption against preemption: that there was no explicit provision for preemption of state tort claims; that the subject matter infringes on the State's inherent powers to protect and promote the health and safety of its citizens; and that preemption would effectively eliminate all means of recourse for the plaintiff. If there had been a need to immunize prescription drug and antibiotic manufacturers from tort liability, the determination of that need should have been "made by Congress in an unambiguous mandate and not by the courts." To date Congress has not seen fit to express such a mandate. We will not do so in its stead.

[Lederle, supra, 125 N.J. at 156, 592 A.2d 1176 (citations omitted).]

III

In Lederle, the consequence of defendant's failure to warn was permanently stained teeth. In Abbott's case, the consequences are a matter of life and death. In the absence of express congressional action, this Court should deny Abbott's claim that it is immunized from state tort law liability for failure to warn users of its blood test of the devastating consequences that could result from a false-negative test reading of HIV contaminated blood. Because of those consequences, the Court's implied preemption analysis simply is too flimsy to support preemption of plaintiffs' state law tort claim.

I would remand this matter for retrial of plaintiffs' state law failure-to-warn claim.

O'HERN and COLEMAN, JJ., join in the dissent.

For modification and affirmance— Chief Justice PORITZ and Justices GARIBALDI, VERNIERO and KING—4.

For remandment—Justices O'HERN, STEIN and COLEMAN—3.

NOTES

[1] N.J.S.A. 2A:58C-2 provides:

A manufacturer or seller of a product shall be liable in a product liability action only if the claimant proves by a preponderance of the evidence that the product causing the harm was not reasonably fit, suitable or safe for its intended purpose because it: a. deviated from the design specifications, formulae, or performance standards of the manufacturer or from otherwise identical units manufactured to the same manufacturing specifications or formulae, or b. failed to contain adequate warnings or instructions, or c. was designed in a defective manner.

[N.J.S.A. 2A:58C-2 (emphasis added).]

[2] An antigen is "a substance that when introduced into the body stimulates the production of an antibody." Conversely, an antibody is "[a]ny of various proteins in the blood that are generated in reaction to foreign proteins or polysaccharides, neutralize them, and so produce immunity against certain microorganisms aor their toxins." Webster's II: New Riverside University Dictionary at 113 (1994).

[3] "HTLV III" is synonymous with HIV.

[4] "Sensitivity" refers to the ability of the test to detect infection in blood samples of AIDS patients.

[5] To run Abbott's test, a blood bank technician would place ¼ inch polystyrene, virus-coated beads, into a plastic tray provided with the test kit. In separate wells of the tray, positive and negative control solutions (solutions with or without the HIV antibody, respectively) are added. The technician then calculates the cutoff value for that particular test based on a mean value associated with the positive and negative control solutions. The cutoff value, which is computed individually each time the test is run, is ultimately used to determine whether the donor's blood sample is positive or negative. Then, the patient's blood (diluted with a buffer) is placed in the well with the beads. After a few cycles of incubation, and the addition of an enzyme marker, the sample is measured by a special device, and compared to the cutoff value previously calculated for that test cycle. According to the package insert, if the sample is above the cutoff, it is retested; if it is below the cutoff it is deemed negative for the HIV antibody.

[6] In her December 10, 1993 affidavit in support of Abbott's motion for reconsideration of the denial of Abbott's original motion to dismiss, Sidote explained that:

On February 28, 1985, at a meeting for the final review of Abbott's package insert, the FDA and I discussed the method for calculation of the test's cutoff value. The decision as to how to calculate the cutoff value, as reflected in Abbott's package insert, was in accordance with the requirements of the FDA.

[7] "Specificity" refers to the test's ability to correctly conclude that HIV antibody is not present in the specimen. In other words, the specificity of a test is indicated by the incidence of false-positive results, whereas sensitivity is indicated by the incidence of false-negative results.

[8] This version of 21 C.F.R. § 601.12 was adopted as revised on April 1, 1985 and remained substantively unchanged until October 7, 1997. See Changes to Approved Application, 62 Fed.Reg. 39890, 39901 (July 24, 1997).

[9] However, in spite of its limitations, the ELISA blood screening tests were effective. A paper published in the journal Transfusion in 1993 shows that in 1984, prior to the use of the ELISA tests, there were 714 new cases of AIDS caused by transfusion. However, after the first full year of the use of the ELISA test (1986), the number of AIDS cases associated with transfusion dropped to below 20 per year. See R.M. Selik, et al., Trends in Transfusion-Associated Acquired Immunodeficiency Syndrome in the United States, 1982-1991, Transfusion, Vol. 33, No. 11 (1993).

[10] Dr. Heller testified that even though the sample tested close to the cutoff, it did not have any detectable HIV antibody. According to Dr. Heller, the 0.121 value is explained by a phenomenon called "non-specific binding" where a non-HIV antibody binds to the test bead causing a measurable reading of antibody, a phenomenon probably caused by an immune response unrelated to an HIV infection.

[11] At their last oral argument before the Court, plaintiffs' counsel confirmed that it was their position that the defect for which Abbott had a duty to warn with respect to the initial insert and the after-acquired evidence was limited to the Test's production of false-negative results at the borderline, and not the production of false-negative results generally.

[12] Section 360k(a), entitled "State and local requirements respecting devices," provides:

(a) General Rule

Except as provided in subsection (b) of this section [when a state or political subdivision applies for an exception], no State or political subdivision of a State may establish or continue in effect with respect to a device intended for human use any requirement—

(1) which is different from, or in addition to, any requirement applicable under this chapter to the device, and

(2) which relates to the safety or the effectiveness of the device or to any other matter included in a requirement applicable to a device under this chapter.

[21 U.S.C. § 360k(a).]

[13] However, this regulation was subsequently withdrawn. See Medical Devices; Preemption of State Product Liability Claims, 63 Fed.Reg. 39789 (July 24, 1998).

[14] Therefore, plaintiffs' argument that the Test was regulated solely as a "biologic" rather than a "device" is not only unsupported by the record, but irrelevant to the Court's conclusion.

[15] In addition, (as recognized by the jury) the plaintiffs did not present any credible evidence that borderline negative test results were more likely to be false-negative, and therefore were inherently dangerous.

[16] 21 C.F.R. § 601.12(f)(2) provides:

(f) Labeling changes.

...

(2) Labeling changes requiring supplement submission—product with a labeling change that may be distributed before FDA approval.

(i) An applicant shall submit, at the time such change is made, a supplement for any change in the package insert, package label, or container label to accomplish any of the following:

(A) To add or strengthen a contraindication, warning, precaution, or adverse reaction;

(B) To add or strengthen a statement about abuse, dependence, psychological effect, or overdosage;

(C) To add or strengthen an instruction about dosage and administration that is intended to increase the safety of the use of the product; and

(D) To delete false, misleading, or unsupported indications for use or claims for effectiveness.

(ii) Pending approval of the supplement by FDA, the applicant may distribute a product with a package insert, package label, or container label bearing such change at the time the supplement is submitted. The supplement shall clearly identify the change being made and include necessary supporting data. The supplement and its mailing cover shall be plainly marked: "Special Labeling Supplement—Changes Being Effected."

[Changes to an Approved Application, 21 C.F.R. § 601.12(f)(2) (1999).]

[17] 21 C.F.R. § 814.39(d) provides:

(d)(1) After FDA approves a PMA, any change described in paragraph (d)(2) of this section that enhances the safety of the device or the safety in the use of the device may be placed into effect by the applicant prior to the receipt under S 814.17 of a written FDA order approving the PMA supplement ...

(2) The following changes are permitted by paragraph (d)(1) of this section:

(i) Labeling changes that add or strengthen a contraindication, warning, precaution, or information about an adverse reaction.

(ii) Labeling changes that add or strengthen an instruction that is intended to enhance the safe use of the device.

(iii) Labeling changes that delete misleading, false, or unsupported indications.

(iv) Changes in quality controls or manufacturing process that add a new specification or test method, or otherwise provide additional assurance of purity, identity, strength, or reliability of the device.

[PMA Supplements, 21 C.F.R. § 814.39(d) (1999).]

[18] We observe that plaintiffs' argument that "Dear Doctor" letters, telegrams, or other similar materials are not "labeling" changes under 21 U.S.C. § 321(m) is wrong. Those alternative methods of warning the Test's users suggested by plaintiffs clearly constitute "labeling" under the FDCA. Cases interpreting the definition of "labeling" have held that "it is clear that supplementary or explanatory information disseminated by the producer of a drug or device may constitute ``labeling,' regardless of whether it physically accompanies the product." Washington Legal Foundation v. Kessler, 880 F.Supp. 26, 28 (D.D.C. 1995); see also United States v. LeBeau, 985 F.2d 563 (7 th Cir. 1993) ("[E]ven when literature on a drug's use is sent in a different package that the drug itself, the literature may still be a ``label.'") Walls v. Armour Pharmaceutical Co., 832 F.Supp. 1467, 1482-83 (M.D.Fla.1993), aff'd, 53 F.3d 1184 (11 th Cir. 1995)) (in an action against a pharmaceutical company in part for the negligent failure to warn of the risk of HIV transmittal through use of blood plasma, recognized that for the purposes of Section 601.12 "labeling is broadly defined [by Section 321(m)] and includes not only package inserts, but also separate communications concerning the drug, such as ``Dear Doctor' letters sent by drug manufacturers to physicians ....") (each case citing Kordel v. United States, 335 U.S. 345, 350, 69 S.Ct. 106, 110, 93 L.Ed. 52, 57 (1948) (holding "[n]o physical attachment one to the other is necessary[; i]t is the textual relationship that is significant.")).

[19] See, e.g., Martello v. Ciba Vision Corp., 42 F.3d 1167, 1169 (8 th Cir.1994), cert. denied, 515 U.S. 1161, 115 S.Ct. 2614, 132 L.Ed.2d 857 (1985); Stamps v. Collagen Corp., 984 F.2d 1416, 1421-22 (5 th Cir.1993), cert. denied, 510 U.S. 824, 114 S.Ct. 86, 126 L.Ed.2d 54 (1993); Fry v. Allergan Med. Optics, 695 A.2d 511, 516 (R.I.1997), cert. denied, 522 U.S. 952, 118 S.Ct. 374, 139 L.Ed.2d 291 (1997); Easterling v. Cardiac Pacemakers, Inc., 986 F.Supp. 366, 374 (E.D.La.1997); Lake v. TPLC, 1 F.Supp.2d 84, 86-87 (D.Mass. 1998); Martin v. Telectronics Pacing Sys., Inc., 105 F.3d 1090, 1098-1100 (6 th Cir. 1997), cert. denied, 522 U.S. 1075, 118 S.Ct. 850, 139 L.Ed.2d 751 (1998); cf. Lakie v. SmithKline Beecham, 965 F.Supp. 49, 53 (D.D.C.1997) (holding that PMA process is not specific regulation triggering preemption under Section Section 360k); Sowell v. Bausch & Lomb, Inc., 230 A.D.2d 77, 84, 656 N.Y.S.2d 16 (N.Y.App.Div.1997) (holding that PMA process is not specific regulation because the requirements are not contained in formal regulation); Goodlin v. Medtronic, 167 F.3d 1367, 1376 (11 th Cir.1999) (holding that PMA process is not specific requirement because of lack of substantive prerequisites for approval).