In Re Sabin Oral Polio Vaccine Prod. Liability Lit., 763 F. Supp. 811 (D. Md. 1991)

*812 Stanley P. Kops, Adler & Kops, Philadelphia, Pa., and Marc S. Moller, Kreindler & Kreindler, New York City, for plaintiffs.

*813 Rupert M. Mitsch, Julie Zatz and Ann Wion, Torts Branch, Civil Div., U.S. Dept. of Justice, Washington, D.C., for defendant.

MOTZ, District Judge.

These seven cases, which have been transferred to the District of Maryland pursuant to 28 U.S.C. § 1407 (1988), are brought against the United States under the Federal Tort Claims Act, 28 U.S.C. §§ 2671-80 (1988). Plaintiffs (or persons whom they represent) contracted poliomyelitis, allegedly as a result of the administration of doses of oral polio vaccine ("OPV") manufactured by Lederle Laboratories, Inc. Plaintiffs contend that the Division of Biologic Standards of the National Institutes of Health ("DBS") was negligent in various respects in approving release to the public of the vaccine allegedly giving rise to their injuries.^[1]

On July 12, 1990, I issued an opinion denying a motion for summary judgment filed by plaintiffs and granting in part and denying in part a similar motion filed by the United States. 743 F. Supp. 410. Broadly stated, I sustained the Government's "discretionary function" defense as to plaintiffs' claims relating to the establishment of the OPV program but rejected it as to actions taken by DBS in implementing the program. The parties then consented to my trying in a consolidated proceeding the liability issues left open by my earlier opinion which were common to all of the cases. A fifteen-day trial ended on March 27, 1991. This opinion constitutes my findings of fact and conclusions of law under Fed.R.Civ.P. 52(a).

I have concluded that over the years DBS has, in violation of 42 C.F.R. § 73.114(b) (1) (iii), approved for release type 3 polio vaccine, knowing that it is more neurovirulent than type 1 polio vaccine, which is mandated by regulation as the reference standard.^[2] I have also concluded that, in violation of 42 C.F.R. § 73.113(b), DBS has released lots of vaccine, including the lots involved in two of the pending cases, which were more than five tissue culture passages from the original strain from which they were derived. There are, however, two observations which I would like to make at the outset.

First, although I find DBS to have been negligent in the eyes of the law, I do not find DBS officials and consultants to have been derelict in the performance of their duties. To the contrary, the scientists who established and implemented the OPV program were (and are) extraordinarily able professionals who consistently acted in the public interest as they reasonably perceived it to be. They made judgments on extremely difficult questions which, strictly from the standpoint of public health, appear to have been entirely proper. We are fortunate to have persons like Bennett Elisberg, Ruth Kirschstein and Joseph Melnick dedicate themselves to public service, and it was a privilege to have them testify before me.

Second, my finding that regulatory violations occurred does not imply that the public health is or has been endangered in any respect. According to the undisputed record, the OPV used in the United States has always been "state of the art" vaccine and the OPV program has resulted in the virtual eradication of wild poliovirus in the Western Hemisphere. I might add that the irony that I am finding what is perhaps the most successful public health program in history to have been implemented "negligently" is not lost upon me. However, that finding is one to which my understanding of the law and my perception of the facts have inexorably led.

Poliomyelitis is a disease of the central nervous system. It is of three types. Type 1 can be contracted only from a type 1 virus, type 2 only from a type 2 virus and type 3 only from a type 3 virus. Each of the plaintiffs here suffers from the third type of the disease.

In 1955 a vaccine against each type of polio, developed by Dr. Jonas Salk, was licensed for production and use in the United States. This vaccine was an inactivated polio vaccine ("IPV") derived from a virus which had been killed. The use of the Salk vaccine dramatically decreased the incidence of polio but did not eradicate it. Between 1958 and 1961, there were nearly 19,000 cases of the disease in the United States. Thirteen thousand people were paralyzed and over 1,000 people died.

While Dr. Salk had been developing his inactive vaccine, other scientists, including Dr. Albert Sabin, had been working on an oral poliovirus vaccine. OPV is produced from a live virus attenuated but not killed during the production process. Scientists perceived that OPV might have several advantages over IPV.^[4] It was relatively less costly and, as its name reflects, could be taken orally whereas IPV required three initial inoculations followed by a least one subsequent booster shot. Moreover, OPV is preferable to IPV because persons who are immunized with IPV can still be linked in the chain of infection; although immunized themselves against the disease, they can serve as carriers of the wild polio virus and pass it on to others. Of course, despite these advantages, there is an inherent risk in the use of OPV. Like other live virus vaccines (such as those used for smallpox and yellow fever), OPV stimulates immunity by inducing a mild infection in vaccinees. Thus, a person vaccinated with OPV or a person who comes into close contact with the vaccine's virus (usually by exposure to a vaccinated person) may develop polio.^[5]

On June 30, 1958, the Surgeon General of the United States appointed an ad hoc committee to consider whether, in light of the continued incidence of polio after the institution of the IPV program, an effective OPV program could and should be developed. There were six members of the committee. The director of DBS chaired the committee; the five other members were academicians who had been actively involved in research on polio and polio vaccines. The purposes of the committee were to (1) assist in the selection of the virus strains to use in the manufacture of an OPV, (2) evaluate data pertaining to the efficacy and safety of vaccines produced from the selected strains and (3) develop regulations for governing the licensing and manufacturing of OPV.

During 1958 and 1959 twenty field trials involving the use of OPV were conducted throughout the world. The most extensive trials, in which vaccines derived from the strains developed by Dr. Sabin were used, were conducted in Eastern Europe and the Soviet Union. By the end of 1959 over 15 million persons had been vaccinated in the Soviet Union alone. Dr. Dorothy Horstmann, an internationally respected epidemiologist, spent two months in the Soviet Union and Eastern Europe under the auspices of the World Health Organization ("WHO"), conducting her own review of the methodology and results of the Soviet tests. After completing her studies, she *815 reported to the ad hoc committee her conclusion that the tests had been properly conducted and that vaccines derived from the Sabin strains were safe and effective. The same conclusion was confirmed by other laboratory and field tests, which were subjected to critical scrutiny at a series of international conferences in which members of the ad hoc committee participated. After reviewing all of the information available to it, the committee recommended to the Surgeon General without qualification that the Sabin type 1 and type 2 strains be used. Although the committee noted a concern that the Sabin type 3 strain had "less than optimum immunogenic capacity and shows a tendency to change its neurovirulence characteristics after passage in serum," it stated that the strain was "satisfactory from the point of view of neurovirulence" and recommended its use.

On August 24, 1960, the Surgeon General, acting upon the recommendation of the ad hoc committee, declared that OPV was suitable for use in the United States and released the committee's recommendations concerning the use of the Sabin strain. In March 1961 DBS adopted regulations, along lines recommended by the ad hoc committee, governing the issuance of manufacturing licenses and the approval and release of vaccine.

Lederle was an early pioneer in the development of OPV. A Lederle employee, Dr. Harold Cox, had developed one of the strains which competed with the Sabin strain in the clinical field trials. Lederle also lent support to Dr. Hilary Koprowski, the developer of the other major strain used in the clinical field trials. When the Surgeon General's ad hoc committee selected the Sabin strain over the Cox and Koprowski strains for use in the United States, Lederle quickly purchased Sabin strain material from which to derive its own seeds and manufacture its own vaccine. It tested its vaccine lots in the summer of 1961 and submitted the test results to DBS in late December of that year.

Lederle's vaccine had to pass a monkey neurovirulence test in order to qualify for licensure. 42 C.F.R. § 73.114(b) (1) (iii) provides, in part, that a "virus pool under test is satisfactory for poliovirus vaccine only ... if a comparative analysis of the test results demonstrate that the neurovirulence of the test virus pool does not exceed that of the Reference Attenuated Poliovirus." Since type 1 vaccine had proven to be the safest in the clinical field trials, type 1 was selected as the reference virus. See 42 C.F.R. § 73.111. Thus, in order to obtain a license for its type 3 vaccine, Lederle had to demonstrate, inter alia, that the results of the monkey neurovirulence tests which it performed on its licensing lots of type 3 vaccine did not exceed the results of similar tests which it performed on the type 1 reference standard. DBS required that this demonstration be made as to five consecutive lots manufactured under the same conditions and derived from the same virus seed.

Briefly summarized, monkey neurovirulence tests are conducted as follows. A given number of monkeys is inoculated with the material under test. Some of the monkeys are injected intrathalamically, i.e. at the stem of the brain, and others are injected intraspinally, i.e. at the base of the spine. Different groups of monkeys are injected with the tested material at different levels of dilution. (Levels of dilution range from 100 through 104 with 100 (undiluted) being the most concentrated level). After the monkeys have been inoculated, they are kept under observation for 17 to 21 days. At the end of the observation period, they are sacrificed and a pathological examination is performed on sections of histologic material taken from their brains, their spinal cords and their lumbar spines.

The results of the tests are then recorded and the data arrayed so that a comparative analysis can be made. The regulations require that five factors be taken into account in this analysis: "(a) the number of animals showing lesions characteristic of poliovirus infection, (b) the number of animals *816 showing lesions other than those characteristic of poliovirus infection, (c) the severity of the lesions, (d) the degree of dissemination of the lesions, and (e) the rate of occurrence of paralysis not attributable to the mechanical injury resulting from inoculation trauma." 42 C.F.R. § 73.114(b) (1) (iii).

Dr. Ruth Kirschstein, the DBS official in charge of evaluating neurovirulence data in the 1960s and early 1970s, devised a grading system for comparing test results. Under her system, the sections of histologic material taken from a monkey were assigned a number, ranging from "0," indicating no polio lesions, to "4," reflecting severe polio lesions. If the section being tested was at the site of the inoculation, the number assigned to it was referred to as the "severity score." If, on the other hand, the section being tested was the one most distant from the site of the inoculation, it was referred to as the "spread score." Thus, in an intrathalamic test the section of histologic material taken from the brain would provide a severity score and the section of material taken from the lumbar spine area would provide the spread score. Conversely, in an intraspinal test the section of histologic material from the lumbar spine area would provide the severity score and the section from the brain would provide the spread score. Dr. Kirschstein assigned to each monkey a "final neurovirulence grade" which was the higher of the severity or spread score.

The tests results on lot 317, one of Lederle's five licensing lots, raised problems at DBS. On Lederle's first intrathalamic test of lot 317, conducted on August 29, 1961, one of the monkeys had a severity score of 4.^[6] Lederle had never had a grade 3 or a grade 4 monkey on its tests of the type 1 reference vaccine. Although it was not required by regulation to do so, DBS also conducted its own tests on Lederle's licensing lots. On an intrathalamic test conducted on October 21, 1961, DBS experienced two grade 4 monkeys. Like Lederle, it had not had a grade 4 monkey on its intrathalamic reference tests. DBS ordered that retests be conducted. On Lederle's second test, performed on November 21, 1961, results showed a grade 3 monkey, and on DBS's second test, performed on December 17, 1961, results showed a grade 4 monkey. These results were still troubling, and DBS ordered that a third round of tests be conducted. On this occasion, because Lederle and DBS were showing different results, Lederle and DBS exchanged some monkeys and each performed one half of its tests using the other laboratory's monkeys. Lederle's third test, using its own monkeys, showed no problems but it experienced a grade 3 monkey at both the 100 level and at the 10-1 level using DBS monkeys. Arguably, DBS's third test results were satisfactory. Although the results showed a total of three grade 3 monkeys, DBS had experienced grade 3 monkeys on its reference tests. However, two of the grade 3 monkeys on DBS's third lot 317 test were Lederle's monkeys, and Lederle had not experienced grade 3 monkeys on its reference tests.

On March 1, 1962, the Surgeon General's ad hoc committee held a meeting at which it discussed Lederle's product license application. After reviewing the testing history of lot 317, the committee concluded that lot *817 317 satisfied the regulatory neurovirulence standards. As recorded in the committee's minutes, "It was deduced that, despite the findings observed on two occasions, no difference could be demonstrated between this lot [317] and the other four in the series.... It was concluded that all of the Lederle type 3 lots met the requirements." On the basis of this finding, DBS granted Lederle's type 3 license.

Although DBS granted Lederle's product license application, the experience with the type 3 licensing lots had taught DBS that type 3 vaccine is more neurovirulent than type 1 vaccine.^[7] This teaching was soon confirmed when subsequent Lederle type 3 lots (manufactured from the same seed which had been used to produce the licensing lots) proved to be excessively neurovirulent in comparison with the reference vaccine. Moreover, as the OPV program got underway, a concern arose that type 3 vaccine was the cause of an unacceptably high number of cases of vaccine-induced paralysis.

The neurovirulence test results and suspect performance by type 3 vaccines in the field led one member of the ad hoc committee, Dr. David Bodian, to question whether type 3 vaccine should be withdrawn until a more satisfactory seed was found. After discussion of the issue, however, the committee recommended and DBS decided not to discontinue the use of type 3 vaccine. According to the trial testimony of Dr. Kirschstein,^[8] the ad hoc committee and DBS personnel also discussed whether, in light of the fact that type 3 vaccine was proving to be more neurovirulent than the type 1 reference, an amendment to the regulations should be sought which would change the reference for type 3 vaccine. The decision was made not to seek an amendment because of a concern about uncertainties inherent in the amendment process. DBS and its consultants feared that, at the least, a proposed amendment might result in an interruption of the OPV program. At the worst, a proposed amendment *818 changing the reference standard might erode public confidence in the OPV program and deter parents from having their children vaccinated. Were that to occur, the OPV program would be doomed.

Instead of seeking an amendment to the regulations, DBS decided to attempt to live within the existing regulations. To that end it adopted a statistical model known as the "Maloney Table" for the purpose of conducting comparative analysis of neurovirulence test results. In effect, that table which was derived in part from type 3 data was substituted for the actual results of the tests on the reference vaccine. The table, unlike the reference history, permitted the occasional occurrence of a grade 4 monkey and generally liberalized the release criteria.

In 1972 DBS was reorganized and transferred from the National Institutes of Health to the Food & Drug Administration. The agency's responsibilities and functions remained the same. However, the reorganization reflected a change of orientation from research to regulation. Shortly thereafter, Dr. Elisberg replaced Dr. Kirschstein as the DBS official responsible for evaluating neurovirulence data.

Prior to assuming this position, Dr. Elisberg had not had extensive personal experience with the OPV program generally or with neurovirulence testing specifically. Therefore, when he first assumed his new responsibilities, he continued to use the same methods of comparative analysis which DBS had previously followed. However, in the latter part of 1973 he directed that DBS discontinue use of the Maloney Table. He took this action after the district court criticized the Maloney Table as being inconsistent with the OPV regulations in Griffin v. United States, 351 F. Supp. 10, 29 n. 26 (E.D.Pa.1972), aff'd in relevant part, 500 F.2d 1059 (3d Cir. 1974). Soon thereafter, Dr. Elisberg further tightened the release criteria by disapproving the then-existing practice of permitting interlaboratory comparisons of data. Instead, he ordered that Lederle's test results could be compared only against Lederle's own reference history and that DBS's test results could be compared only against DBS's reference history.

Dr. Elisberg also developed new methods for conducting comparative analysis of neurovirulence data. He found one of these, the "mean lesion score" approach, to be quite useful. It involved scoring monkeys not by separate severity and spread scores but by a figure constituting the mean of the scores recorded for all histologic sections which were examined. It is noteworthy that while he used type 1 reference data for analyzing intrathalamic tests under this approach, he used as his reference for intraspinal tests type 3 data (derived from Seed 45 B 85 lots which had previously been released). He did so because (1) he believed that, in light of the susceptibility of the lumbar spine area to the growth of polio lesions and the occurrence of paralysis, high intraspinal neurovirulence scores did not raise serious safety concerns, and (2) he knew that if he used the type 1 test results as his reference on intraspinal tests, approximately fifty percent of the tested lots would fail.

In 1977 Dr. Elisberg used the mean lesion score approach experimentally on two lots submitted by Lederle for release: lot 217 (a type 1 lot) and lot 450 (a type 3 lot). The results of the mean lesion score analysis suggested that these two lots were more neurovirulent than the references against which they were being measured. This caused him to examine more critically the neurovirulence grades of the monkeys used on the lot 217 and lot 450 tests. He discovered that on the intrathalamic tests conducted by DBS both on lot 217 and lot 450 there was a monkey with a severity score of 3 and a spread score of 3. There had never been a 3/3 monkey in DBS's tests of the reference. Nevertheless, lots 217 and 450 would have passed under the traditional "final neurovirulence grade" method of analysis because the 3/3 monkeys which appeared would have been assigned a final neurovirulence grade of 3, which had been seen on the reference. To remedy this problem, Dr. Elisberg proposed to institute a dual severity/spread *819 scoring system (at least as to intrathalamic tests) which would have resulted in the rejection of both lots.

This proposal raised serious concerns at Lederle and led to a series of meetings between Lederle and DBS personnel.^[9] At the first such meeting, on September 13, 1977, DBS advised Lederle of the proposed failure of the two lots under the dual severity/spread scoring method. Lederle noted that if the two lots were rejected, Lederle would require at least twelve months to reestablish consistency under § 73.116(c). Lederle stressed that this would raise a severe supply problem: "the nation would be without vaccine for the latter part of 1978 and on a limited supply basis beginning in January 1979." The discussion then focused on the possibility of eliminating the consistency requirement. Dr. Ronald J. Vallancourt, manager of Lederle's Biological Section, "left the meeting with a distinct impression that Dr. Meyer [Harry M. Meyer, Jr., the Director of DBS and Dr. Elisberg's superior] will strongly consider a change in the regulations to allow for the rejection of a monopool without the resulting break in consecutiveness." Notes taken by Dr. Paul Parkman, then the Deputy Director of DBS, reveal that Dr. Meyer planned to "change regulations rather than go to [zero] vaccine status." According to Dr. Parkman's notes, the meeting ended with the view that DBS "cannot stop production in polio area because of demand."

In the wake of the September 13 meeting, Lederle officials became concerned about the broader implications of the dual scoring system. On September 22, Dr. Vallancourt met with Dr. Elisberg and obtained DBS's testing records. Dr. Vallancourt then reviewed the data and concluded that if the dual scoring system had been applied to all the lots produced from Seed 45 B 85 (the seed Lederle had been using since approximately 1965), fifty-three percent of those lots would have failed. The result would have been less dramatic if the dual scoring system had been used only for intrathalamic results. If that had been done, only nine percent of the lots derived from Seed 45 B 85 would have failed. Nevertheless, the possibility of the continued use of the dual scoring system presented a crisis for Lederle. One of its internal memoranda warned that if that scoring system were adopted, "it could have very grave consequences."

On October 4, Dr. Vallancourt discussed his growing concerns in a meeting with Dr. Meyer. Dr. Vallancourt reiterated the likelihood of a supply crisis in 1978 and 1979 if lots 217 and 450 were rejected. He pleaded with Dr. Meyer to resolve the "uncertainty" surrounding DBS's criteria for evaluating OPV lots: "Question How can a manufacturer commit time, money and effort into a program and product when forced [to deal] with all these uncertainties"? Dr. Vallancourt also stated: "We are comparing the neurovirulence of a type I reference monopool to all production types (I, II, III). Historically type III always shows more neurological activity." Dr. Meyer was equivocal in his response. He noted that *820 an advisory committee was then reviewing the polio vaccine program and was due to make a formal report in February 1978. He stated his intention "to review both this problem and the long standing potency testing problem with his staff." On October 19, Dr. Vallancourt and Mr. D. Carroll (who was on the business side at Lederle and subsequently became its president) again met with DBS officials. Mr. Carroll began the meeting by noting that the "sudden change in interpretation of polio," i.e., the shift to a dual scoring system, would "cause Lederle to be out of business since [DBS] would say 50% would fail using new criteria." He stated, according to notes taken by Dr. Parkman, that Lederle would be unwilling to invest its resources to reestablish consistency "if method of evaluation was to be changed so that 50% would fail."

Dr. Meyer responded by reviewing the history of OPV testing at DBS. He noted that he was saddled with "[b]asically unsatisfactory regulations" and that "since 1964 about 50% of Lederle T. III would fail" using the dual scoring system, with the "50% number com[ing] from evaluation of [intraspinal] tests of T. I vaccine." He indicated that he was aware that "T. III is hotter than T. I reference." He stated that "[w]here [DBS] would like to be" was comparing "T. III vs. T. III reference [with] 95% confidence limits (or 98%)." The issue of vaccine supply also arose again at the October 19 meeting. Dr. Meyer opined that the "Benefit/risk decision says you need to have polio vaccine." Later in the meeting, Dr. Vallancourt emphasized that "supply situation will be tight." The problem grew worse in November when problems arose with a third lot, lot 451. In a telephone conversation, Dr. Vallancourt informed Drs. Meyer, Parkman and Elisberg that delays in reestablishing consistency "could cause supply & inventory problems." In response, Dr. Meyer stated that he "felt it was perhaps time to move toward amending the regulations" and that he did "not regard the possibility of no polio vaccine on the market as a `feasible alternative.'"

On November 28 DBS officials met with their Canadian and British counterparts to discuss the impending supply crisis. One participant walked away with the impression that "there is the very real possibility of a worldwide shortage of (at least type III) oral polio vaccine in the very near future." During the meeting, Dr. Meyer once again stated that "T. III [is greater than] T. I [in neurovirulence], but T. I is the reference. Regs. suggest that I, II, III [are] homogeneous, but they are not." Later, Dr. Meyer suggested that "T. III ref. [is] needed since there is a diff. in virus." The participants discussed the need to "soften regs. in actuality" and debated the most opportune time to do so.

On December 12, Dr. T.R. Ubertini of Lederle had a five-hour meeting with Dr. Elisberg. At the meeting, Dr. Elisberg mentioned "his concern for allowing only safe vaccine to be marketed in this country." He expressed concern over "the pressure brought on the DBS by consumer groups, Dr. Salk's crusade and the occurrence of vaccine associated cases." Dr. Elisberg insisted that "[t]he regulations are to be interpreted as they read, i.e. any lot of poliovirus vaccine that exceeds (intrathalamically and intraspinally) the neurovirulence of the reference should not be submitted to the DBS," although he indicated that there was uncertainty over exactly how to define the "level of acceptance."

A final meeting was held December 21. According to a Lederle memorandum, all the participants agreed "that [Lederle's] type 3 is more neurovirulent than type 1 reference." The memorandum further stated:

Dr. Meyer asked us to bear with them in their efforts to change regulations which may be accomplished in mid 1978. In interim we are advised to release our protocols as we have been i.e.: avoid critical review of intra-spinal results; consider grade 3 lesion on type 3 test as acceptable on intra-thalamic test. Where we have a question Dr. Ubertini will call Dr. Elisberg before signing and releasing a neurovirulence protocol. Dr. Meyer stated that during this period [DBS] *821 and Lederle must continue with guidelines that may be technically outside the regulations as written but we must do this if we are to continue to offer vaccine.

(Emphasis added).

In fact, the regulations were never amended.^[10] Further, although the dual scoring system was thereafter followed by DBS (causing lots 217, 450 and 451 not to be released), it was applied only to intrathalamic test results.

Seed 45 B 85, in use from approximately 1965 to 1983, produced vaccine which DBS found to be generally acceptable. However, the progeny of another material known as Sabin Original Rederived ("SOR"), which had originally been produced by Pfizer Laboratories, Ltd., proved to be even more satisfactory. As early as January 1978 Dr. Elisberg suggested to Lederle that it explore the possibility of obtaining a portion of the seed material which Pfizer had produced from SOR. Acting upon this suggestion, Lederle contacted Institut Merieux, which had acquired Pfizer's seeds when Pfizer had gone out of the OPV business. In February 1981 Institut Merieux shipped a portion of the Pfizer material to Lederle, and Lederle then produced its own seed from the material. This seed, denominated Seed 45 B 165, was approved for use by DBS in 1983, and it has since been the source for all of the OPV used in the United States. The vaccine involved in two of the pending cases was produced from it.

SOR was cloned from a Sabin seed through a process known as plaque purification. The genetic qualities of SOR are said to be even more stable than those of the original Sabin material, and the vaccine produced from it has been the safest and most efficacious ever manufactured. Nevertheless, because the vaccine manufactured from Seed 45 B 165 is more than five tissue culture passages from the original Sabin material, its use is in violation of § 73.113(b). I will defer a more detailed statement of facts concerning this issue until Section III, infra.

I will first address one of the two fundamental issues on which I find plaintiffs are entitled to prevail: DBS's continuous failure from 1962 to the present to comply with the mandate of 42 C.F.R. § 73.114(b) (1) (iii) that a lot of vaccine be approved for release "only ... if a comparative analysis of the test results demonstrate that the neurovirulence of the test virus pool does not exceed that of the Reference Attenuated Poliovirus."

At the summary judgment stage plaintiffs argued, in effect, that § 73.114(b) (1) (iii) deprives DBS officials of the authority to exercise any professional judgment in determining neurovirulence, converting that determination into the essentially clerical function of comparing the numerical test results of the vaccine under scrutiny with the numerical test results of the reference virus. I rejected that argument. As stated in my earlier opinion,

Neurovirulence testing is not merely a numbers game but requires the exercise of sound scientific judgment. Section 73.114(b) (1) (iii) calls for a `comparative evaluation,' not simply a comparison of numerical test scores. Moreover, as experience accumulates and knowledge increases, approaches to determining neurovirulence change.... Thus, although the decisions made by DBS officials under § 73.114(b) (1) (iii) do not constitute the exercise of a discretionary function immune from judicial review, upon such review they are to be judged in their fullness rather than dissected by an auditor's pen.

*822 I reaffirm these views. The power conferred upon an agency or other delegated authority must be commensurate with the responsibilities assigned to it. However, to the extent that the adoption of a regulatory standard has resolved underlying safety and policy issues, regulators must work within the environment which that standard creates. They may not ignore what the regulation says because they believe that it is unwise or has become outdated. Nor may they circumvent it by merely paying lip-service to its terms.

That is precisely what happened in this case. When the OPV regulations were first adopted, type 1 vaccine was chosen as the reference vaccine for the monkey neurovirulence tests. Unfortunately, the early test results quickly demonstrated that the neurovirulence of type 3 vaccine is higher than that of type 1 vaccine. At this juncture DBS committed the critical error which it never remedied. Instead of seeking to have the regulations amended to change the reference standard, it compromised itself by accepting test results as comparable when, in fact, they were not comparable at all. To that end it devised and put into use the Maloney Table which substituted statistical projections (based, in part, upon type 3 data) for the actual history of the type 1 vaccine.

It is true that when Dr. Elisberg joined DBS and became responsible for analyzing neurovirulence data, he tightened the release criteria. After learning of the district court's decision in Griffin v. United States, 351 F. Supp. 10, 29 n. 26 (E.D.Pa. 1972), aff'd in relevant part, 500 F.2d 1059 (3d Cir.1974), he directed that use of the Maloney Table be discontinued; he ordered that test results be compared only on an intra-laboratory, rather than an inter-laboratory basis; and he instituted a dual severity/spread scoring system in conducting comparative analysis of the results of intrathalamic tests. However, Dr. Elisberg knew that type 3 vaccine is more neurovirulent than type 1 vaccine; that had been part of DBS's institutional knowledge from the time that Lederle had submitted its initial licensing lots. It was also confirmed by Dr. Elisberg's dual scoring system and his experimental mean lesion score approach to comparing evaluation of neurovirulence data. As a result, in 1977 when it was perceived that the application of the dual scoring system to intraspinal tests would result in rejection of approximately fifty percent of the type 3 lots DBS decided to evaluate the intraspinal test results only by the final neurovirulence grade scoring system.

Solely as a matter of public health administration, there was a sound basis for DBS's decision. By 1977 DBS had had fifteen years of experience with the OPV program. This experience had demonstrated that, regardless of its failure to perform satisfactorily (in comparison with the type 1 reference vaccine) on monkey neurovirulence tests, type 3 vaccine was safe (to the extent that a live vaccine can ever be said to be safe) when administered to humans. Further, DBS officials necessarily were concerned about the detrimental, if not devastating, impact which rejection of a substantial number of type 3 vaccine lots would have had on the vaccine supply. Because of the consistency of manufacture requirement of § 73.116(c), the distribution of all type 3 vaccine might well have been curtailed.

This, however, is the heart of the matter. However reasonable DBS's decisions might have been dehors the regulations, it was the regulations which provided the standard by which DBS officials were required to make their neurovirulence determinations. This standard constituted one of the critical elements of the definition of safety when public approval for the implementation of the OPV program was implicitly given through the adoption of the regulations. Therefore, when DBS officials concluded that type 3 vaccine lots should be measured against a type 3 (rather than type 1) reference, it was incumbent upon them to seek an amendment to the regulation to change the reference. This was a responsibility of which they were fully aware. Their failure to live up to it and their continued violation of the existing *823 regulations subjects the United States to liability in these actions.^[11]

The second issue presented is whether vaccine lots manufactured from Seed 45 B 165 are more than five tissue culture passages from the original Sabin strain in violation of 42 C.F.R. § 73.113(b).^[12] As indicated above, Seed 45 B 165 has been developed from Sabin Original Rederived (SOR), the material from which Pfizer Laboratories, Ltd. was producing its vaccine in the United Kingdom before it went out of the OPV business. Lederle obtained seed material produced from SOR from Institut Merieux which, in turn, had acquired it from Pfizer.

The actual lineage of Seed 45 B 165 is not in dispute. The seed had the following phases of development: (1) Sabin Original-MERCK (SOM) was produced from the original Sabin material (SO); (2) Seed 127 B-III was produced from SOM; (3) a material known as PS 438/3/C/a/1 was produced from Seed 127 B-III; (4) PS 440/1 (now known as SOR) was produced from PS 438/3/C/a/1; (5) Seed 457-III was produced from SOR; and (6) Seed 45 B 165 was produced from Seed 457-III. A seventh passage occurs when vaccine lots are manufactured from Seed 45 B 165.

The first question which this lineage raises is whether the third of the phases just outlined (PS 438/3/C/a/1 from Seed 127 B-III) constitutes a tissue culture passage. The uncontradicted evidence is that it does not. The process by which the PS 438/3/C/a/1 material was derived, known as plaque purification, involves the cloning of virus cells and simply is not the same as a tissue culture passage.

This finding aside, however, the evidence is clear that vaccine lots manufactured from Seed 45 B 165 are more than five tissue culture passages from the original Sabin material. The Government does not dispute that, in fact, six tissue culture passages are involved: (1) SOM from SO; (2) Seed 127 B-III from SOM; (3) SOR from PS 438/3/C/a/1; (4) Seed 457-III from SOR; (5) Seed 45 B 165 from Seed 457-III; and (6) vaccine lots from Seed 45 B 165. However, the Government contends that the first of these passages from SO to SOM is not a "tissue culture passage" within the meaning of § 73.113(b). Alternatively, it argues that tissue culture passages should be counted from SOR, not SO or SOM. Neither of these contentions is persuasive.

The specific regulatory language governing this question is as follows: "Virus in the final vaccine shall represent no more than five tissue culture passages from the original strain, each of which shall have met the criteria of acceptability prescribed [elsewhere in the regulations]." 42 C.F.R. § 73.113(b). The Government's contention that this language is intended to apply only to post-SOM tissue culture passages is based upon the fact that SOM had been used in some of the early clinical field trials proving the acceptability of the Sabin strain. This proffered rationale is inadequate for three separate reasons.

First, the Government's reading of § 73.113(b) is directly contrary to the section's express language. Section 73.113(b) is the only place in the regulations where the word "original" appears before the word "strain." Precisely because (as the Government argues) SOM was used in the early clinical field trials, the use of the word "original" in § 73.113(b) in and of itself implies that a distinction is being drawn between SO and SOM for the purpose of setting the starting point for measuring tissue culture passages.^[13]

Second, the interpretation of § 73.113(b) which the Government now urges is contrary to what was for years the accepted understanding of the section. A Lederle memorandum dated June 22, 1979 states that "Dr. Elisberg suggested that we [Lederle] contact Merieux and request some of [the Pfizer] seed.... A regulation change would be required, however, before it could be used as it exceeds the total number of passages beyond the original Sabin seed, as allowed by law." Another memorandum of a meeting between DBS and Lederle officials on November 6, 1980, states that Dr. Elisberg indicated that DBS would have to make a decision to "(1) do something with the regulations to allow the use of the seed, i.e., a change in the regulations themselves; (2) interpret what Pfizer did as being not outside the regulations; or (3) redefine the regulations in some manner to indicate this is a rederivation of Sabin's original seed." Still another Lederle memorandum, dated May 8, 1981, stated that "qualification of the type 3 Pfizer seed will necessitate a change in the regulations regarding number of allowable passages from the Sabin original."

Third, the Government's interpretation of § 73.113(b) is lacking in a rational basis. The implied premise of the section is that vaccine manufactured within five tissue culture passages from the strain material will be fully acceptable in actual experience. SOM, of course, falls in this category. Therefore, the fact that SOM proved to be satisfactory in the original clinical field trials is wholly immaterial to the problem at which the section is directed: potential genetic instability of vaccine which is more than five tissue culture passages from the strain material.

The Government's alternative argument is that SOR, not SOM, should be the starting point for counting tissue culture passages under § 73.113(b). This argument is based upon the fact that SOR has been demonstrated by scientific tests to be at least as genetically stable as SO and, since it is derived from SO, should be considered to be the equivalent of SO.

The Government's argument on this point is somewhat disturbing. It is tantamount to saying that something is so because *825 we say it is so; that is, DBS scientists believe SOR to be genetically stable and safe, ergo it is "an original strain" within the meaning of § 73.113(b). What the regulations say, however, is that strains must be subjected to extensive clinical field trials, not simply genetic testing, in order to establish their acceptability. See § 42 C.F.R. 73.110(b). The Government cannot by interpretative fiat circumvent this requirement.^[14]

In effect, what has been played out in this litigation is the third of the options attributed to Dr. Elisberg in Lederle's memorandum of the November 6, 1980 meeting: "redefin[ing] the regulations in some manner to indicate this is a rederivation of Sabin's original seed." Indeed, subsequent Lederle memoranda reflect that this possibility was specifically discussed between DBS and Lederle officials. At a June 23, 1981 meeting, DBS officials reportedly "said that there was some international precedent for viewing the cloning process used to make the Pfizer seed as rederivation of the Sabin original. If this were so, vaccine would meet CFR guidelines. Dr. Meyer promised to discuss this matter with [DBS's] General Council [sic]." In a December 4, 1981 telephone conversation, Dr. Elisberg informed Dr. Ubertini that the consensus of WHO "is to call the Pfizer seed RSO (RNA extracted Sabin original). Any subsequent passage would be called RSO + 1, 2 etc. For production purposes five passages beyond RSO would be allowed. Dr. Elisberg likes this approach, and will try to apply it to the U.S. requirement." And on May 12, 1982, Dr. Elisberg is recorded as having definitively told Lederle that "the Bureau will follow WHO recommendations that the Pfizer seed be viewed as a rederivation of the Sabin original strain."

Again, as in the case of the neurovirulence test reference criteria, the critical error which DBS made was not to seek an amendment to the regulations to authorize the use of SOR as the starting point for counting tissue culture passages.^[15] The evidence before me demonstrates that such an amendment would clearly be proper and in the public interest. However, DBS erred in simply disregarding the regulatory mandate and doing what it thought best. Its release of vaccine lots manufactured from Seed 45 B 165 directly violated § 73.113(b).

Let me finally consider various ancillary contentions made by plaintiffs which I find to be substantially without merit.

In October 1962 DBS began to require manufacturers to use test regimens different from those prescribed by the regulations then in effect. At that time §§ 73.114(b) (1) (i) and (ii) required manufacturers to perform intrathalamic tests with ten monkeys inoculated each at 100 and 10-1, and intraspinal tests with five monkeys inoculated each at 100, 10-1, 10-2, 10-3 and 10-4. Under the regimens as modified by DBS, intrathalamic tests were to be given to thirty monkeys inoculated at 100 and intraspinal tests were to be given to five monkeys each inoculated at 100, 10-3 and 10-4. It was not until 1968 that *826 the regulations were amended to incorporate the modified regimens.

I previously denied motions for partial summary judgment which both sides filed on this issue. On the one hand, I ruled that a letter dated October 23, 1962 from the Director of DBS notifying manufacturers of the new test regimens did not constitute an official finding of equivalency authorizing the modification under the "equivalent methods" provision of § 73.118. On the other hand, I ruled that if, in fact, the change in test regimens yielded more complete and reliable information on which to base determinations of neurovirulence, to hold that DBS's failure to make a formal finding of equivalency under § 73.118 constituted actionable negligence would be to elevate form over substance. The evidence at trial was undisputed that the test regimens as modified did yield more complete and reliable information than the regimens originally prescribed. In light of this evidence, I find that DBS's technical violation of the regulations does not establish the government's liability. See Jayvee Brand, Inc. v. United States, 721 F.2d 385, 391-92 (D.C.Cir.1983).

Section 73.114(b) (1) (i) provides, in part, as follows:

Each of at least 30 monkeys shall be injected [intrathalamically].... Only monkeys that show evidence of inoculation into the thalamus shall be considered as having been injected satisfactorily. If on examination there is evidence of failure to inoculate virus pool material into the thalamus, additional monkeys may be inoculated in order to reestablish the minimum number of 30 monkeys for the test.

Section 73.114(b) (1) (ii) contains a similar provision relating to intraspinal tests.

It is undisputed that from the inception of the neurovirulence testing program through the present DBS has accepted test results if, in the case of intrathalamic tests, twenty-four of the surviving monkeys are shown to have been inoculated with the tested material; in the case of intraspinal tests, the same is true if four of the surviving monkeys at each dilution level are shown to have been inoculated with the tested material. Plaintiffs contend that DBS's acceptance of these test results is in direct violation of the regulatory mandate and that virtually all of its neurovirulence testing has therefore been invalid.

I do not agree. The practice which DBS adopted is based upon the clause in § 73.114(b) (1) (iii) providing that "the virus pool under test is satisfactory for poliovirus vaccine only if at least 80 percent of the animals in each group survive the observation period and if a comparative analysis of the test results demonstrate that the neurovirulence of the test virus pool does not exceed that of the Reference Attenuated Poliovirus" (emphasis added). Since this provision clearly authorizes the acceptance of test results if twenty-four monkeys tested intrathalamically and four monkeys tested intraspinally at each dilution level survive the observation period, DBS has read §§ 73.114(b) (1) (i)-(iii) together as permitting the acceptance of test results if at least eighty percent of the tested monkeys were demonstrated to have been effectively inoculated and to have survived. This interpretation is entirely reasonable and must be upheld. See generally Young v. Community Nutrition Inst., 476 U.S. 974, 981-82, 106 S. Ct. 2360, 2364-65, 90 L. Ed. 2d 959 (1986); Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 413-14, 65 S. Ct. 1215, 1217, 89 L. Ed. 1700 (1945).^[16]

Section 73.110(b) (5) provides that "[s]ubsequent and identical neurovirulence tests shall be performed in monkeys whenever there is evidence of a change in the neurovirulence of the production virus." Plaintiffs contend that this section requires that a seed be retested whenever a lot manufactured from that seed fails the neurovirulence test of § 73.114(b) (1) (iii). Alternatively, they argue that, regardless of what the regulations themselves require, DBS had an established practice of retesting a seed after one of its lots proved to be excessively neurovirulent. Cf. Berkovitz v. United States, 486 U.S. 531, 547, 108 S. Ct. 1954, 1964, 100 L. Ed. 2d 531 (1988).

Plaintiffs' interpretation of § 73.110(b) (5) is plainly wrong. The consistency of manufacture provision of § 73.116(c) specifically addresses the failure of individual OPV lots. It merely requires that if a single lot does not pass the neurovirulence test, five new consecutive lots must pass the neurovirulence test before any one of them can be released. I likewise reject plaintiffs' alternative argument, which is based upon the deposition testimony of Dr. Elisberg in a prior case. That case was settled before Dr. Elisberg had the opportunity to correct the record, and he testified at trial here that his deposition testimony was simply in error. I fully credit his trial testimony in that regard. The uncontradicted evidence before me establishes that since the neurovirulence of a seed is affected only if it is not kept properly refrigerated, there is no need to retest a seed absent evidence that it has been allowed to warm. Since seed materials are kept in separate vials, this is a rare occurrence. Further, the evidence establishes that the most effective way to determine the safety of a seed is to test its progeny. Even if the seed itself were to meet the neurovirulence criteria on a retest, its use would have to be discontinued if lots manufactured from it were not meeting those criteria.

Plaintiffs contend that DBS violated the regulations in three respects by permitting Lederle's type 3 vaccine to be originally licensed. First, they assert that DBS should have required SOM to be tested as a seed. DBS takes the position that this was unnecessary because SOM had been used in clinical field trials leading to the adoption of the OPV program and was deemed to be a strain which had already been adequately tested. In my earlier opinion I did not decide this issue, leaving it to be further addressed by the parties. Nothing which occurred at trial persuaded me that plaintiffs met their burden of demonstrating that DBS's interpretation of the regulations on this point was unreasonable. To the contrary, since Lederle itself did not produce the SOM material and since the seeds and lots which were SOM's progeny were subjected to neurovirulence testing, DBS's interpretation of the regulations was entirely proper.

Second, plaintiffs fault DBS for allegedly not having required testing to demonstrate that the seed from which the licensing lots were produced was free of SV 40, an extraneous microbial agent, in violation of § 73.110(b) (3). They point to the absence of records regarding testing for SV 40 in the documents produced by the government in an attempt to prove that the tests were not performed. Assuming that the negative inference which plaintiffs seek to draw is appropriate, their argument is insubstantial since the record is undisputed that the failure to conduct an SV 40 test would not have affected the neurovirulence of the seed and its progeny or prevented the extent of their neurovirulence from being properly measured.

Third, plaintiffs argue that lot 317, one of the five lots which Lederle submitted in support of its application for the type 3 license, clearly exceeded the reference vaccine in neurovirulence. I completely agree. One need not play a "numbers game" to conclude that the lot should have failed *828 when a grade 4 monkey appeared on Lederle's first intrathalamic test and two grade 4 monkeys appeared on DBS's first intrathalamic test. The second tests conducted both by Lederle and DBS were also unsatisfactory. The results of the third round of tests were somewhat better but still problematic; in any event, lot 317 should have been rejected before the third tests were ever conducted.^[17]

If, however, the impropriety in the original licensure of Lederle's type 3 vaccine were the only issue on which I was ruling in favor of plaintiffs, a substantial issue of proximate cause would be presented. This is so because I further find that if Lederle had not been granted a type 3 product license in 1962 on the basis of the five lots which were then submitted, it would nevertheless have continued in its efforts to obtain a type 3 license on the basis of the five initial lots produced from Seed 45 B 85. This finding is based upon the uncontradicted evidence that (1) Lederle and DBS were committed to the development and implementation of the OPV program, and (2) the first five lots manufactured from Seed 45 B 85 were less neurovirulent than the original licensing lots and would have passed the neurovirulence test as DBS was then applying it. Since the vaccine involved in each of the pending cases was manufactured long after Seed 45 B 85 was first put into use, I have great difficulty in perceiving how the original licensure of Lederle's vaccine could be deemed to have proximately caused plaintiffs' injuries. I consider it prudential, however, to refrain from ruling on that issue. Although the parties have agreed to be bound by my finding of fact on the question of whether Lederle would have been licensed despite the failure of its initial licensing lots, the legal consequences of that finding turn upon the tort law of the states in which the various transferor courts sit. As for case of Miller v. United States (the only one of the pending cases in which Maryland law applies), it will be soon enough for me to decide the proximate cause issue if and when a reversal of my other rulings on appeal renders it of more than academic interest.

The regulations do not require DBS to conduct its own neurovirulence tests. They require only that DBS review manufacturers' neurovirulence tests. Plaintiffs contend, however, that at least until 1972, DBS had established a practice of conducting neurovirulence testing of its own and that it negligently failed to comply with that practice when permitting the release of certain lots. The evidence does not support this contention. DBS conducted tests of its own to the extent, but only to the extent, that its time and other resources permitted. DBS personnel simply could not conduct tests of their own on all occasions, and the evidence shows that they exercised sound judgment in utilizing their limited resources as they did. They are to be commended, not criticized, for the extra efforts which they made.

Plaintiffs allege that DBS failed to review the test results on Seed 45 B 85 before authorizing the release of lots manufactured from that seed. The regulations do not expressly require that DBS itself review the test results on seeds. Section 73.116(g) requires only that test results for vaccine lots be submitted to DBS. While Dr. Elisberg testified that nevertheless during the years of his tenure DBS did *829 review the test results on seeds, the record is not clear whether that practice was established in the earlier years when Seed 45 B 85 was tested and first put into use. Further, plaintiffs have not proved to my satisfaction that DBS did not in fact review the test results on Seed 45 B 85. While a date stamp on the copy of the protocol from DBS's file submitted by the government in this litigation indicates that the protocol was received after the first lots from Seed 45 B 85 were approved and released, I am not persuaded that this means that DBS had not received another copy of the protocol on an earlier occasion or had not reviewed the test results at Lederle's facilities pursuant to 42 C.F.R. § 73.30-73.32. In any event, plaintiffs' contention is without underlying substance since the test results on Seed 45 B 85 were entirely satisfactory under the neurovirulence criteria as DBS was then applying them.

In summary, while I substantially reject the many subsidiary contentions made by plaintiffs, I find that DBS violated the regulations by releasing vaccine lots which exceeded the reference vaccine in neurovirulence and (in the case of the progeny of Seed 45 B 165) by releasing vaccine lots which were more than five tissue culture passages from the original Sabin strain. I further find that these violations will subject the United States to liability under the Federal Tort Claims Act in the event that the individual plaintiffs are able to meet their burden of proof on any issues of causation particular to them. I will confer with counsel immediately concerning the manner in which now to proceed in order to obtain prompt and expeditious appellate review of my rulings.

[1] As noted in the opinion which I issued on July 12, 1990, DBS was transferred from the National Institutes of Health to the Food & Drug Administration and named the Bureau of Biologics in 1972. In 1984 the Bureau of Biologics was renamed the Office of Biologics Research and Review. I will again use "DBS" to connote the Division of Biologic Standards and its successors.

[2] The OPV regulations now appear in 21 C.F.R. § 6030.10-6030.17 (1990). I will cite to them (as I did in my earlier opinion) by their original number.

[3] This section of the opinion is (with slight modifications) the same as Section I of my earlier opinion. I am repeating it here for the convenience of any court which reviews my rulings.

[4] The so-called "Cutter incident" also led to consideration of the establishment of an OPV program. In that incident several persons who were vaccinated with Salk vaccine manufactured by Cutter Laboratories contracted polio as a consequence of the fact that the virus used in the vaccine administered to them had not been killed, as had been believed, during the manufacturing process. This incident led to the "consistency of manufacture" requirement embodied in 42 C.F.R. § 73.116(c).

[5] The extent of the risk has been demonstrated by the actual experience under the OPV program. The annual number of cases of polio in the United States has averaged less than ten per year, or one case for every 2.6 million doses of vaccine distributed.

[6] DBS traditionally has used the results of intrathalamic tests to determine neurovirulence and the results of intraspinal tests to confirm the susceptibility of the monkeys being tested to the growth of poliovirus. The Surgeon General's ad hoc committee had recommended that the two tests be used for these respective purposes. Dr. Bennett Elisberg, who replaced Dr. Kirschstein at DBS in 1972, came to believe that the intraspinal test is the most effective method of determining neurovirulence. This view is shared by public health administrators in many countries, and WHO regulations (which have not been adopted in the United States) require that only intraspinal tests be conducted. Nevertheless, DBS has continued to focus on intrathalamic test results in determining neurovirulence. I am making no finding as to whether intrathalamic or intraspinal tests provide more discriminating data on the neurovirulence question. However, as discussed infra, I find that DBS officials have purposely ignored intraspinal data in determining neurovirulence because of their concern about maintaining an adequate OPV supply.

[7] It is important to remember that the term "neurovirulence" has a technical meaning in this context: the performance of a tested vaccine on monkey neurovirulence tests. Because these tests only approximate field conditions, the term does not necessarily connote actual potential for causing harm to the human nervous system. The crux of this case lies in the fact that, early on during the regulation of OPV, DBS lost confidence in the ability of comparative analysis of monkey neurovirulence test results, in the form mandated by the regulations, to predict accurately the actual safety properties of type 3 vaccine. This belief became all the stronger over the years as accumulating epidemiological data confirmed that type 3 vaccine was safe.

It was Dr. Elisberg who most clearly explained the distinction in the definition of neurovirulence at trial, and I fully accept his testimony. However, Dr. Elisberg drew another distinction which I do not accept: that although he knew type 3 vaccine to be more neurovirulent generically than the type 1 reference, he did not believe that any of the specific type 3 lots which were released during his tenure exceeded the type 1 reference in neurovirulence. The record is overwhelming that the testing results on all type 3 lots not simply those which were rejected by DBS or withdrawn by Lederle had higher intraspinal test results than the type 1 reference. As discussed infra, this is the reason that Dr. Elisberg used type 3 vaccine as his point of reference for his mean lesion score approach in comparing intraspinal test results. Furthermore, it is also the reason that he did not apply his dual scoring system to intraspinal tests. I recognize that he testified that there is a "rationale" for discounting intraspinal test results: that the area being inoculated is the area most susceptible to paralysis. However, this rationale only supports DBS's reasoning in concluding that excessive "neurovirulence" does not translate into an unacceptable safety risk; it does not provide an excuse for disregarding the test results themselves, which § 73.114(b) (1) (iii) mandates be considered.

[8] Dr. Kirschstein's testimony was in response to rather pointed questions which I asked her. Her answers were candid and forthright. Let me note, incidentally, another aspect of the evidence which demonstrates the integrity of Dr. Kirschstein, Dr. Elisberg and others at DBS. Although I find that they did attempt to square a circle in interpreting the regulations to permit the release of type 3 vaccine which did not meet the neurovirulence requirements, at no time did they make any effort to "fudge" the underlying data. Although neurovirulence scores seem clear-cut when stated in their final formulation, in fact they are the result of the exercise of extremely delicate professional judgment and could have been subject to manipulation by anyone without the highest scruples.

[9] In order to make discovery in this case manageable, the parties agreed to stipulate to the authenticity of Lederle's documents. The Government now contends that the stipulation as to authenticity was not intended to constitute a waiver of any hearsay objections which it might have. This contention would seem to violate the spirit, if not the letter, of the parties' agreement. In any event, it is wholly without merit. First, I find that the stipulation of authenticity cures any hearsay objection. As authentic business documents, the memoranda are admissible under Fed.R.Evid. 803(6). Second, the statements in the memoranda are corroborated by other evidence in the case, including the trial testimony of ex-DBS officials and DBS's own notes and memoranda. Third, little or no hearsay issue actually exists in many instances because the materiality of many of the statements in the memoranda lie in the fact that they were made as observations by Lederle personnel a fact which cannot be disputed. Fourth, and most importantly, my findings of fact would be precisely the same whether or not I were to consider the Lederle documents at all. In that connection I might mention that although I, of course, became generally aware of the contents of various memoranda both during the summary judgment proceedings and at trial, when I prepared this opinion, being aware of the Government's objection, I first analyzed the evidence other than the Lederle memoranda in drawing my conclusions. These conclusions were simply confirmed by my subsequent review of the memoranda.

[10] In 1986 DBS did seek extensive amendments for the purpose of adopting the WHO regulatory scheme in the United States. Under the WHO regulations (of which, incidentally, Dr. Elisberg is a strong proponent) only intraspinal tests need to be conducted and type 3 lots are to be compared to a type 3 reference history. Opposed by Lederle, the proposed amendments failed.

[11] In support of its contention that type 3 vaccine is not excessively neurovirulent, the Government presented summaries showing that all lots manufactured from Seeds 45 B 85 and 45 B 165 had been properly released under the criteria which had been contemporaneously applied to them and that only a few isolated lots produced from 45 B 85 would have been rejected by the stricter criteria adopted after they had been passed. As to the latter, the Government further proved that at least five consecutive satisfactory lots were subsequently manufactured and that the consistency of manufacture requirement of § 73.116(c) therefore would not have resulted in the cessation of the use of Seed 45 B 85. Further, Dr. Neal Nathanson, an extremely well-qualified expert for the Government, presented charts and summaries which impressively demonstrated that under the criteria which he deems to be proper for determining neurovirulence, the vaccine lots from Seeds 45 B 85 and 45 B 165 were completely satisfactory.

The Government's proof in this regard begs the question. Paradoxically, it is the Government which is seeking to turn neurovirulence testing into a "numbers game." The summaries and charts are persuasive only when the methods of comparative analysis which form their premise are accepted. However, it was in selecting the methods of comparative analysis which it did that DBS violated the regulations. Throughout its regulation of the OPV program (first by blinding itself to the lot 317 data, then by using the Maloney Table and finally by assiduously avoiding close examination of intraspinal test results because of a concern over vaccine availability), DBS purposely chose tools of analysis which, at best, obscured or ignored relevant data. This cannot be deemed to constitute the exercise of sound scientific judgment.

[12] Technically, of course, 21 C.F.R. § 630.13(a), not 42 C.F.R. § 73.113(b), is the regulation which has been violated by the release of Seed 45 B 165 lots because the OPV regulations were renumbered before Seed 45 B 165 came into use. Likewise, it was 21 C.F.R. § 630.16(b) (1) (iii), not 42 C.F.R. § 73.114(b) (1) (iii), which was violated when DBS released Seed 45 B 85 lots after the regulations were renumbered.

[13] The Government's only response on this point is that Dr. Joseph Melnick, who wrote the first draft of the proposed regulations, testified at trial that he understood § 73.113(b) to permit the release of vaccine five tissue culture passages away from SOM. However, the subjective intent of the author of proposed regulations certainly cannot be dispositive if the language which is disseminated for public comment implies a meaning different from that which is intended. Moreover, the language originally drafted by Dr. Melnick apparently was different than that which was finally adopted. Section 73.113(b) as initially written would have permitted vaccine representing "no more than three tissue culture passages from the seed strain." When finally adopted, the phrase "original strain" was substituted for "seed strain." This change is not only clear on its face but may also reflect a compromise reached in the redrafting since the number of permitted tissue culture passages was increased from three to five.

[14] The wisdom of the requirement of clinical field trials is demonstrated by evidence in the record that at least one strain material which had been demonstrated to have greater genetic stability than the Sabin strain proved to be unsatisfactory when subjected to experimental field tests. The Government seeks to distinguish between that strain material and SOR on the ground that SOR is itself derived from the Sabin strain. Nothing, however, in the regulation supports that distinction.

[15] When in 1986 DBS did unsuccessfully seek extensive amendments to the regulations in order to adopt the WHO regulatory scheme, it proposed a provision which would have permitted SOR to be used as the starting point for counting tissue culture passages. In its comments to this proposed section, the FDA (the parent agency of DBS) stated its understanding that the current regulation required "that virus in the final vaccine represent no more than five passages from the original strain" and that this "restriction ... is unnecessarily rigid." That comment stands in stark contrast to the contention made by the Government here as to DBS's interpretation of the regulation.

[16] When in 1962 DBS changed the test regimens, it circulated to the manufacturers a proposed revision to § 73.114(b) (1) (i) which included the following sentence: "The number of monkeys originally inoculated on a test shall be sufficient to assure that at least thirty shall be available for calculating neurovirulence response (i.e. paralysis, histopathological lesions)." If this language had been included in the regulations when they were originally adopted or when they were amended in 1968, plaintiffs' contention that thirty monkeys were required to be successfully inoculated would be stronger. However, this sentence has never appeared in the regulations. The actual language of § 73.114(b) (1) (i) permitting (but not requiring) the inoculation of additional monkeys may be read as applying to the situation where eighty percent of the originally tested monkeys which were successfully inoculated did not survive the observation period.

[17] I note that the ad hoc committee appears to have approved lot 317 by comparing it not to the reference vaccine but to the other four licensing lots which Lederle submitted. Although it concluded that "all of the Lederle Type 3 lots met the [regulatory] requirements," it prefaced this conclusion by observing that "no difference could be demonstrated between this lot [317] and the other four in the series." I further note that in retrospect, at least, DBS agreed that the initial licensing lots had been excessively neurovirulent. For example, in 1977 Dr. Parkman recorded Dr. Meyer's view that "for 1961-1964, all type III's were too neurovirulent."