# Petitioners Reply Brief — Industrial Union Dept., AFL-CIO v. American Petroleum Institute

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/brief%3Amicro_IA40385006_0035%3A15

## Record

- **Collection:** Supreme Court brief
- **Document type:** Petitioners Reply Brief
- **Published:** January 1, 1980
- **Citation:** 448 U.S. 607

## Text

FOR ARGUMENT

October Term, 1979

IN THE
Supreme Court of the United States

OCT & 99

ICHAEL RODAK, JR., CLERK

No. 78-911

InpustRIAL Union Department, AFL-CIO,

Petitioner,
v.

AMERICAN PetroLeuM Institute, et al.,

Respondents.

No. 78-1036

Ray MarsHAui, SECRETARY OF LABOR,

Petitioner,
v.

AMERICAN PETROLEUM INstiTuUTE, et al.,

Respondents.

On Writ of Certiorari to the United States
Court of Appeals for the Fifth Circuit

REPLY BRIEF FOR PETITIONER
INDUSTRIAL UNION DEPARTMENT, AFL-CIO

J ALBERT WOLL
Woll & Mayer

815 Fifteenth Street, N.W.

Washington, D.C. 20005

LAURENCE GOLD
American Federation of
Labor and Congress of
Industrial Organizations
815 Sixteenth Street, N.W.
Washington, D.C. 20006

GEORGE KAUFMANN
2101 L Street, N.W.
Washington, D.C. 20036

Grorce H, COHEN
Rospert M. WEINBERG
JEREMIAH A, CoLLIns
Bredhoff, Gottesman, Cohen
& Weinberg
1000 Connecticut Ave., N.W.
Washington, D.C. 20036

ELLIot BREDHOFF
General Counsel,
Industrial Union Department,
AFL-CIO
1000 Connecticut Ave., N.W.
Washiagton, D.C. 20036

Attorneys for Petitioner
Industrial Union Department,
AFL-CIO

ee ee ooo

>:

TABLE OF CONTENTS

Page
SPCR PAT Er ENOEN | 6 ies caiealsb ils WeeeR Nay eN SaeeRD 1
PE Lay Uh ae Vaas ee ped eae SESE N RAC aA Rae R ees 2
EE PR okie kde En he) aha een Rae os eee 7
CURIE dea i eave leh Tee eN ceva ube kee es 18
ADDENDUM A... . A Survey of ‘‘Non-Absolut-

ist’? Statutory Approaches to
Protecting Health and Safety la

ADDENDUM B.. . An Analysis of the Basis for
the Secretary’s Rejection of
Industry’s Contention That
Epidemiologic Studies Have
Shown Benzene to be Safe at
Low Levels of Exposure .... 1b

ADDENDUM C... . An Additional Analysis of the
Evidence Supporting the Seec-
retary’s Finding that it is
Impossible to Derive a Mean-
ingful Risk Assessment for
Benzene, and in Particuiar, His
Rejection of Dr. Wilson’s At-
tempted Risk Assessment .... le

AUTHORITIES
Cases: Page

American Iron and Steel Institute v. OSHA, 577 F.2d
825 (3rd Cir. 1978), petitions for cert. pending, Nos.

TB-91B 'TOGIO 6 isccietesiccssboeueeeee 6b
E. I. duPont de Nemours & Co. v. Train, 430 U.S. 112
(1977) ..n.ncc's vow eiceie ppinisyiw pune eaen an 3, 2a
Environmental Defense Fund v. EPA, 548 F.2d 998
(D.G, Cir. 1996). cis chdiauakacneneialeaeeenene 4a
Forester v. CPSC, 559 F.2d 774 (D.C. Cir. 1977) ..... 3a
FPC v. Florida Power & Light Co., 404 U.S. 453
|: >) a nPaperieg ryote Cs ee 9
IUD v. Hodgson, 499 F.2d 467 (D.C. Cir. 1974) ...... 17

Labor Board v. Fruit Packers, 377 U.S. 58 (1964) ... 5
Permian Basin Area Rates Cases, 390 U.S. 747 (1968) 15
Radio Corp. v. United States, 341 U.S. 412 (1951)... 15

Society of Plasties Indus., Inc. v. OSHA, 509 F.2d 1301
(2d Cir. 1974), cert. denied, 421 U.S. 992 (1975) .. 17, 6b

United States v. Bacto-Unidisk, 394 U.S. 784 (1969) .. 15
‘United States ex rel. Chapman v. FPC, 191 F.2d 796

(4th Cir. 1951), aff’d, 345 U.S. 153 (1958) ......... 9
United States v. General Motors Corp., 518 F.2d 420
(DC: Cie.) EG}. uicebiis «eke ae cee ead eee 3a
United States v. Rutherford, —— U.S. ——, 47
US.L.W. 4726 (June G, 1878) 3. cccccicsccaseuenees 5a
Vermont Yankee Nuclear Power Corp. v. NRDC, 435
U.K S89 (108B), oc avnaacaeweanea eet 3
Weyerhaeuser Corp. v. Costle, 590 F.2d 1011 (D.C. Cir. - —
CEBFB) ~ 0.055 <0 view 0's cieid iat tinnw tnx Wniaeu aban ae ane 2a

Statutes: Page

Child Protection and Toy Safety Act of 1969, as
amended, P.L. 91-113, 83 Stat. 187-189, 15 U.S.C.

hie eck a whilue pha dhik- bee% vaxwe 3a
aes alia a Wh dua aie! bin'hida44') ale oe alone s 3a
Clean Air Act Amendments of 1970, as amended, P.L.
91-604, 84 Stat. 1676, 42 U.S.C. § 7401 et seg. ....... 3a
I di kv ou ob wl ves seebc kev 3a
TE REDO os sas cjcdive cidvehech ols deus 3a
Clean Water Act of 1977, P.L. 92-500, 86 Stat. 816,
ee NE OP OR ik wb ca Wid cece viele caceecs 2a, 3a
I set, yy, 4s a sa’ cle a'elne ch wine 3a
Consumer Product Safety Act of 1972, P.L. 92-573, 86
Stat, 1207, 15 U.S.C. § 2051 ef seg. 006s cece ccc ees 4a
The Energy Policy and Conservation Act of 1975, P.L.
94-163, 89 Stat. 923, 42 U.S.C, § 6201 et seq. ........ la
ee icin sb gil by 4h oe ov iow vo 0.000 la

Federal Environmental Pesticide Control Act of 1972,
as amended, P.L. 92-516, 86 Stat. 975, 7 U.S.C. § 136

ea Dee ksh Ui eda bw Keb da v's avalos 4a
ENO ESE Stat is nr a 4a
ee ga Wwe 6d 6's, s 0-01 bd. 0 weed cs 4a
EE ECR Tel LPR on a 4a
ee eT ik Gas ook sdk da'vineteslencn ck, 4a
SE eo ee ete yl a ee bas 4a
Sera bi eos Us 1 WAS o Chk Cee bawec cts 4a
a EA he SEUSS ype Ag an 4a
Federal Food, Drug, and Cosmetic Act of 1938, 52 Stat.
Ry BE WAIL. 4 GOL OF BOO. vec cece cecncervvosas 4a, 5a
GU he ys yg deep ad wo Sv won dene 4a
Re dL occas ai vvep ahalp ds acd tic places Da
EE eS cis wo bia bd coca. 64-0 Olle Kaos 5a
EE erg eis Ue oe tS 5a
ira is uaian SGN Uls war sa deere seas es 5a
I Pe Ng iz Cu anf staale-p'e-aiegalicen's Poa e cay 4a
ili

Page

Federal Water Pollution Control Act Amendments of

1972, as amended, P.L. 92-500, 86 Stat. 816, 33 U.S.C.
CBR 6b SOG. an ivensies cv eceverassenasnevens’s la, 2a
Pe RRICMECED! isc dh LA ee qek saw hiemeunep ys os 000d oi 2a
© TSEECRPIS EA) ob ivvslncecnsedavatedavtvend ese ilelss 2a
OS RURAUMPC ED cc tece sys ctbeasivivnetuepetsageakuwe 2a
CRURMEMICS) hos secede Re eese es aces eqs he Upp epee’ 2a
GC RGRROITAPUIED occ cn rene geet sn cee etasdme esse ys 2a
Flood Control Act of 1936, 49 Stat. 1570, 33 U.S.C.
Mi Fe OE DON 5 aie oe on co ile sale Con aeehelees dese nays la
ht 1 Sr eee eres I ar owen la

National Traffic and Motor Vehicle Safety Act of 1966,
as amended, P.L. 85-563, 80 Stat. 718, 15 U.S.C.

2) 8 Fo Peer ere er rer rere Tere rh ay Tee ee te 3a
CCE yi Ns Ni wep aves Taha taeepapeecnny ear paws 3a
Occupational Safety and Health Act of 1970, 84 Stat.
RO Se Us BOE OE OO icc avcdevacensesés passim
RC ie eee CRED E AER GASSER ADRES R ENE AR ED 3
i, getNth e Gruner Ena ie recy ginempe me mre maar ey | 2 3
BOE GD. bonis 8065 & oan we Roe Wad ea] Cue 3, 4, 5, 6, 17, 4e
SREP ASK oie sve aee ehh bid ape tan ee eee ee Mae ka 6
RES 6 VE ve be py were aw POLE RE ER ee he See 8

The Outer Continental Shelf Lands Act Amendments
of 1978, P.L. 95-372, 92 Stat. 632, 43 U.S.C. § 1331

DORE Fos otins thos PEN y OAS 55 UTED RET SRA CA SE la
GRRE) has whihe ads Rhee LOS Ca ERE SNoeae WEES paaM la
Toxie Substances Control Act of 1976, P.L. 94-469, 90
Stet. 2008, 15 U.S.C. § QG01 Cf 909. pice vcvcciceses 4a
ED eens tha pas Vue cu meee Ree he RD Seales 4a
SNEED cand nkons tae sa cee sed eu rash ewes ta

Legislative Materials:
Legislative History of the Occupational Safety and
FORT Ot OE TOTG fos i ecarvekanition keds 3, 4, 5, 7, 10, 11

Page
Senate Report No. 91-1282, 91st Cong., 2d Sess.
CEREUD leicwecvy badaldulaeeelinnelsu ee umee een 7,10
House Report No. 91-1765, 91st Cong., 2d Sess.
CHOU OE inva WN iecdne Le vaak ieee eel pele eee eee 10
House Report No. 91-2191, 91st Cong., 2d Sess.
CTOTED. Sb'5% db ud goed uk eneSuee se nea eee 11
House Report No. 92-1153, 92d Cong., 2d Sess.
CUED 0.6.0 kina nade bode tea ee eee ee 4a
Senate Report No. 89-1301, 89th Cong., 2d Sess.
(SOS HS he canln sees ekmal tiene phe ke Ghanian tere 3a
Senate Report No. 91-237, 91st Cong., 1st Sess.
CRPOD Sie dssncicncstdeecorsenescene ieee 3a
Senate Report No. 92-749, 92d Cong., 2d Sess.
CRPEME cay meld bane e cee eda beleeed ei eee nl eee 4a

Miscellaneous:

Baram, Regulation of Health, Safety and Environ-
mental Quality and the Use of Cost-Benefit Analy-
sis, Final Report to the Administrative Conference

of the United States, March 1, 1979 ............... 14
Cornfield, Carcinogenic Risk Assessment, 198 Science
Ge ieee ed Ce ie hance eeu ween treme ine 14

Decision of the Commissioner in the Matter of Diethyl-
stilbestrol: Withdrawal of Approval of New Animal
Drug Applications, 44 Fed. Reg. 54852 (Sept. 21,
1979)

House Committee on Interstate and Foreign Com-
merce, Subcommittee on Oversight and Investiga-
tions, Federal Regulation and Regulatory Reform,
95th Cong., let Sess. 515 (1976) ... ccc. cccccvccun 14

Lovins, Cost-Risk-Benefit Assessment in Energy
45 Geo. Wash. L. Rev. 911 (1977) ............... 14, 9e

McGarity, Substantive and Procedural Discretion in
Administrative Resolution of Science Policy Ques-
tions: Regulating Carcinogens in EPA and OSHA,

CF Gam tid, | Cae (ASTON a ek Sehr dabtan on dectabek 13

IN THE

Supreme Court of the United States

October Term, 1979

No. 78-911
InpustriaL Union Department, AFL-CIO,
Petitioner,
v.
AMERICAN PETROLEUM InstiTuTEs, et al.,
Respondents.
No. 78-1036
Ray MarsHai, SecreTARY oF Lasor,
Petitioner,
v.
AMERICAN PetroLeuM Institute, ef al.,
Respondents.

On Writ of Certiorari to the United States
Court of Appeals for the Fifth Circuit

REPLY BRIEF FOR PETITIONER
INDUSTRIAL UNION DEPARTMENT, AFL-CIO

Introduction

The briefs of the industry respondents and the amici
supporting their position (these respondents and amici
shall hereinafter be referred to collectively as ‘‘industry’’)
pay lip service to sustaining the lower conrt’s statutory
analysis. In Part I, infra, we respond to such efforts as
industry makes in this direction.

The burden of industry’s submissions is, however, that
this case involves a regulation designed to eliminate a ‘‘de
minimis health risk.’’;? and that ‘*benzene exposures of

‘Brief for Respondents (hereinafter cited as ‘‘API Br.’’)
at 29.

2

10 ppm or less are unlikely to lead to any leukemia at all.
...°? In Part LI, infra, we show that this depiction is con-
trary to findings of fact made by the Secretary based on
substantial evidence in the record before him. The Secre-
tary found, based on substantial record evidence, that,
absent the standard at issue here, over 30,000 workers will
be exposed to unsafe levels of a carcinogenic substance,
benzene, every day they go to work. In the proceedings
before the Secretary, industry offered the data and other
evidence it now cites to this Court to show that such ex-
posures—between 10 ppm and 1 ppm of benzene—are safe
or of de minimis consequence; the Secretary found, how-
ever, that industry’s evidence did not support that con-
clusion.

I.

Industry's argument in support of the court of appeals’
holding that ‘‘the Secretary cannot promulgste an oceu-
pational health standard without finding a ‘reasonable[]

. relationship between expected costs and benefits’ *’ is
a set of variations on the theme that ‘‘[u]nlike the En-
dangered Species Act of 1973, or the Delaney Clause of the
Food, Drug, and Cosmetie Act, the OSH Act aims at rela-
tive rather than absolute protection.’’ (APT Br. 28.) We
agree that OSHA does not require ‘‘absolute protection.’’
But that undisputed proposition does not justify the lower
court’s holding. For Congress in different statutes has
adopted a variety of non-absolutist approaches to protect-
ing health and safety, some of which require a balancing
of costs and benefits and others of which<do not.* The
protection that the Secretary is to provide to workers
against toxie substances, and the factors which the See-
retary must consider in promulgating a standard for expo-
sure to such substances, must be ascertained from the lan-
gauge of OSHA and its legislative history, ‘‘The question
...is not what a court thinks is generally appropriate to
the regulatory process; it is what Congress intended for

27d. at 16 ( emphasis in original).
3 We survey these statutes in Addendum A.

3

these regulations.’’ E. I. duPont de Nemours & Co. v. Train,
430 U.S. 112, 188 (1977). Cf. Vermont Yankee Nuclear
Power Corp. v. NRDC, 435 U.S. 519 (1978).

1. Industry emphasizes that the ‘‘key’’ to our case is
‘*the first sentence of Section 6(b)(5).’’ (API Br. 11). We
do indeed contend that this case is governed by that sen-
tence:

The Secretary, in promulgating standards dealing
with toxic materials or harmful physical agents under
this subsection, shall set the standard which most ade-
quately assures, to the extent feasible, on the basis of
the best available evidence, that no employee will suffer
material impairment of health or functional capacity
even if such employee has regular exposure to the
hazard dealt with by such standard for the period of
his working life.
We do so because the language and history of that sentence
demonstrate that its office is to establish the extent and
limit of the Secretary’s authority to regulate exposure to
toxic substances.

The Senate’s final act prior to passing 8.2193 was to
rewrite that sentence. The language of § 2(b) and § 3(8)—
the other provisions cited by industry to support the argu-
ment that the Act is ‘‘relative’’ not ‘‘absolute’’—had
already been settled. Yet, Senator Dominick, the Nixon
Administration's spokesman on the Act, continued to insist
that the first sentence of § 6(b)(5) as reported out by Com-
mittee—** [t]he Secretary, in promulgating standards under
this subsection, shall set the standard which most adequately
and feasibly assures, on the basis of the best available evi-
dence, that no employee will suffer any impairment of health
or functional capacity, cr diminished life expectancy even if
such employee has regular exposure to the hazard dealt
with by such standard for the period of his working life’’
(Leg. Hist. 242)—was defective in that:

*See IUD Br. at 27-33; Fed. Parties Br. at 42-48; and Fed.
Parties Reply Br. at 19-21 for the arguments showing that these
provisions considered alone do not support Industry’s position.

4

Any administrator responsible for enforcing the
statute will be faced with an impossible choice. Kither
he must forbid employment in all occupations where
there is any risk of injury, even if the technical state
of the art could not remove the hazard, or he must
ignore the mandate of Congress. .. .

The confusion inherent in this provision will, in my
opinion, render it no more than a hortatory admonition
to the Secretary to fulfill his responsibilities under the
atti

In trying to make this bill more palatable there is
no point in putting in criteria which cannot be met.
[ Leg. Hist. 481-482].

The debate that ensued makes it plain that all concerned
understood that the language of the first sentence of § 6(b)
(5) would set the extent of the Secretary’s authority to
regulate toxic substances.

The modification of that sentence ‘‘agreeable to all’’
adopted at the conclusion of that debate solved the problem
Sens: sominick had identified, as he stated:

1 ‘ortunately, we had language providing that ahy-
on. vould be assured that no one would have a hazard,
...no one would have any problem for the rest of his
working life.

It was an unrealistic standard. [Leg. Hist. 502].

The statutory words chosen, on their face, state that the
Secretary is to set levels of protection possible to achieve,
not those that are perfect in theory but impossible of
achievement.5 And, that reading of the language is con-

5 For a more complete discussion of the legislative history of
the first sentence of §6(b) (5), see our opening brief at pp. 13-23.
Industry, rather than focus on this critical point in the genesis of
§$6(b) (5), contends that Congress’ intention in using the word
‘*feasible’’ should instead be discerned from a statement which
accompanied the introduction of the Administration’s original
proposed occupational safety and health bill in a prior Congress.
API Br. at 42-43, 9b. Industry mistakenly attributes that state-
ment to Senator Javits, misconstrues its import, and ignores
Senator Javits’ reservations as to the merit of the Administration
bill. See Fed. Parties Reply Br. at 22-23, n. 17.

5

sistent with Senator Dominick’s earlier statement of his
conception of the proper role of the Secretary in setting
standards:

The duty to set standards and produce criteria for
standards implies that the criteria and the standards
will be as effective as possible; not that they will
produce a perfect environment. No human effort can
produce a perfect environment no matter how hard
we try. [Leg. Hist. 376.]

There is not a word in the debate suggesting that the Senate
intended to require the Secretary to balance costs and bene-
fits.° Given that the term ‘‘feasibility’’ does not in common
parlance connote a cost-benefit evaluation, and that Con-
gress has in numerous instances treated feasibility and cost-
benefit evaluation as discrete concepts, ‘‘the silence of
[Congress] ... is pregnant with significance.’’ Cf. Labor
Board v. Fruit Packers, 377 U.S. 58, 66-67 (1964).7

2. It is inherently unlikely that the second and third
sentences of § 6(b)(5), which were adopted in Committee,
inform a proper understanding of the amendment to the
first sentence, whch was adopted on the Senate floor to
clarify what the Committee bill meant. And in fact, the
latter two sentences do not alter the substantive criteria
for toxie substance standards contained in the first sen-
tence; those sentences merely set forth the types of evi-
dence the Secretary is to employ in carrying out the man-
date of the first sentence—types of evidence which include
not only ‘‘research, demonstrations, [and] experiments,’’
but also ‘‘such other information as may be appropriate.’’

The House, whose bill did not set forth any criteria
governing the Secretary’s standard-setting authority, ac-

* See Fed. Parties Reply Br. at 24-26 for a demonstration of the
irrelevance to the issue here of the substitution of the word
‘‘material’’ for ‘‘any’’ which occurred in the same amendment to
§$6(b)(5) described in the text.

*For examples of statutes in which Congress has treated feasi-

bility and cost-benefit evaluation as discrete concepts see Addendum
A to this brief.

6

cepted §6(b)(5) as passed by the Senate. To be sure as
industry points out, the Conferees altered the Senate bill by
adopting §6(f) which provides that the Secretary’s findings
are to be supported by ‘‘substantial evidence on the record
as a whole.’’ The Conferees did not suggest that §6(f)
modifies the Act’s substantive provisions and we cannot
discern the logie of industry’s implication that it does so.

3. It is decisive that industry’s argument that the Secre-
tary is required to balance costs and benefits lacks a statu-
tory predicate. But out of caution we pause to note the
defects in industry’s policy justification for its position—
viz. that such an inquiry must be undertaken so that the

finite * * * * societal resources that can be devoted

to these health and safety risks * * * [are] allocate[d]
.. Where they can be expected to do the most good.®

The ‘‘logic’’ of this justification requires that the Seere-
tary, before he can promulgate a standard to protect em-
ployees against any health hazard, must inventory all other
hazards possibly faced by those employees and then com-
pare the relative costs and benefits of the proposed stand-
ard with the costs and benefits of regulating each of the
other possible hazards. And industry would read the Act
to require the Secretary to embark on that task despite
the recognition by the sponsors that:

In the field of occupational health * * * [there is a]
lack of information and records, * * *

** * (Mleterials long in industrial use are only now
being discovered to have toxie cffects. * * * [T]echno-
logical advances and new processes in American indus-
try have brought numerous new hazards to the work-
place. * * * It is estimated that every 20 minutes a new
and potentially toxic chemical is introduced into indus-
try. New processes and new sources of energy present
occupational health problems of unprecedented com-
plexity.

8 API Br. at 29, 27, 73; see also id. at 5, 18-20, 25, 32-33, 35,
58-59, 63, 66, 72, 78, 80; and Addendum C at 2¢, 7e. See also
Brief for Amicus Chamber of Commerce (hereinafter ‘‘Chamber
Br.’’).

7

Recent scientific knowledge points to hitherto un-
suspected cause-and-effect relationships between oc-
cupational exposures and many of the so-called chronic
diseases—cancer, respiratory ailments, allergies, heart
disease, and others.”

Nevertheless under industry’s construct, to do anything
the Secretary would have to investigate and evaluate
everything. This in the name of*‘common sense’? and
**pragmatism.’’ See APT Br, 5, 29.

II.

In our opening brief, at 8-11, we set forth the Secretary’s
nine critical findings which form the predicate for the
promulgation of the henzene standard. None of those find-
ings was upset by the court below. Those findings negate
any contention that the health risk sought to be reduced by
the standard is either non-existent or de minimis: exposure
to airborne benzene causes levkemia and other serious and
sometimes fatal diseases; hundreds of employees in this
country have died of leukemia from occupational exposure
to airborne benzene; without the standard, over 30,000 em-
ployees in this country will regularly be exposed to air-
borne henzene at their workplaces at levels in excess of 1
ppm; there is no level of exposure to benzene which can be
considered safe in terms of henzene’s leukemogenic effect :
a decrease in the exposure level to benzene will result in a
decreased risk of leukemia; and, there is no way to quantify
the number of leukemia deaths that would result if the
standard is not put into effect, let alone to conclude that
that number would be zero or de minimis.

In making these findings, the Secretary considered evi-
dence submitted by industry in support of the contention
that there is no health risk posed by regular exposure to
benzene at levels of 10 ppm and below. The Secretary, after
reviewing the entire reord, including the cross-examination
of industry's witnesses, found that industry’s evidence does

°S. Rep. No. 91-1282, 91st Cong., 2d Sess. 2-3 (1970), Leg. Hist.
142-143.

8

not support its contention. In this Court, industry in effect
resubmits that same evidence. First, industry refuses to
accept that the Secretary found, based on ample record evi-
dence, that there is no safe level of exposure to airborne
benzene. Second, industry ignores that the Secretary found,
again based on ample record evidence, that there is no way
to quantify the number of lives that will be saved by the
standard, Given the weight industry places on the assertion
that the standard does not add to the protection afforded
the employees covered, we turn now to the substantial evi-
dence that supports the Secretary's contrary conclusion.

A. The Secretary’s finding that there is no level of
exposure to benzene which can be considered safe is
supported by substantial evidence in the record.

In the proceeding before the Secretary, it was undis-
puted that exposure to airborne benzene causes leukemia;
studies had reported the incidence of leukemia among work-
ers exposed to high levels of benzene for varying lengths
of time to be as much as a twenty times greater than the
incidence among the general population. Pet. App. 100a.
There was considerable dispute, however, on the question
whether a safe level of exposure to benzene could be identi-
fied. A number of eminent medical scientists—inter alia, Dr.
Muzaffer Aksoy of Instanbul University (Ex. 60 at 12-13;
Tr. 148-149); Dr. Marvin Sakol (Tr. 328); Dr. Bernard
Goldstein of the New York University Medical Center (Fx.
75 at 2; Tr. 351-352); Dr. Peter Infante of the National
Institute for Occupational Safety and Health (‘*‘NIOSH’’)
(Tr. 895); and Dr. Robert Laigo of NIOSH (Tr. 1229)—
testified that, on the basis of the current state of medical
and scientific knowledge relating to bene exposure to
benzene must be considered unsafe at all ¥vels. NIOSH,
the research arin created by the Act to provide the Secre-
tary expert assistance in standard setting,’® also went on
record in support of the proposition that no level of expo-
sure to benzene can be considered safe. Ex. 2.5 at 6; Tr.

1 Section 22(a) of the Act, 29 U.S.C. § 671(a).

9

745-746." Indeed, on the basis of its analysis of the scien-
tific evidence concerning the risks posed by benzene, NIOSH
had ‘‘strongly recommend[ed]”’ in 1976 that the Secretary
immediately promulgate an emergency temporary standard
reducing permissible exposures to 1 ppm. Pet. App. 72a;
Ex. 2-6 at 1; see also Ex. 2-5 at 1, 6.

In view of this testimony, industry’s assertion that the
Secretary’s no-safe-level finding rested solely on a ‘‘regu-
latory ‘policy assumption,’ ’’ rather than on a ‘factual
hasis’’ is wrong.’? While the issue could not be resolved by
empirical evidence, the expert scientific testimony upon
which the Secretary relied constitutes substantial evidence

in support of his determination. As this Court has repeat-
edly held,

well-reasoned expert testimony—based on what is
known and uncontradicted by empirical evidence—may
in and of itself be ‘‘substantial evidence’? when first-
hand evidence on the question . . . is unavailable. This
proposition has been so long accepted, .. . that we do
not consider it fairly in dispute.

FPC v, Florida Power & Light Co., 404 U.S. 453, 464-466
(1972). The Court in that case went on to quote from an
opinion by Judge Parker in United States ex rel. Chapman
v. FPC, 191 F.2d 796, 808 (4th Cir. 1951), aff’d, 345 U.S, 153
(1953), the following passage which has application here as
well:

The [substantial-evidence] rule is no different because

'' The Secretary’s no-safe-level determination was also supported
by extensive record evidence regarding susdkiodnad in general, See,
e.g., the Report to the Surgeon General by the Ad Hoc Committee
on the Evaluation of Low Levels of Environmental Chemieal Car-
cinogens of the National Cancer Institute (Ex. 272 at 1); the
views of the Committee on Public Information in the Prevention of
Occupational Cancer of the National Academy of Sciences (re-
ported in Ex. 76 at 4); and the testimony of Dr. Herman Kraybill
of the National Cancer Institute (Tr. 1210) and of Dr. Charles
Shaw, a cancer researcher at the M. D, Anderson Hospital and
Tumor Institute (Tr. 140). See also kx. 97 at 55.
12 API Br. at 16, and n. 38.

10

the questions involve matters of scientifie knowledge
and the evidence consists largely of the opinion of ex-
perts. The court may not, for that reason, ignore the
conclusions of the experts and the Commission and put
itself in the absurd position of substituting its judg-
ment for theirs on controverted matters of [science].
It is in just such matters that the findings of the Com-
mission, because of its experience and the assistance
of its technical staff, should be accorded the greatest

- weight and the courts should be hesitant to substitute
their judgment for that of the Commission."

B. The Secretary's finding that there is no way to
estimate the number of lives that will be saved by the
standard is supported by substantial evidence in the
record.

There is no direct empirical data documenting the extent
of the risk of leukemia for employees exposed to airborne
"Industry appears to suggest that by eliminating language
which would have allowed standards to be issued on the basis of
“views and arguments,’’ Congress meant to preclude the Secretary
from relying on expert scientific opinions. But industry misappre-
hends the nature of Congress’ concern regarding that language.
The ‘views and arguments’’ provision appeared in the Daniels
bill as reported out of the House Committee (Leg. Hist. 728-730),
and in the Williams bill (S. 2193) as originally introduced (Leg.
Hist. 6-7). It stated: ‘' Upon the entire reeord before him, ineluding
. any evidence, data views and arguments submitted in connec-
tion with the hearing, the Secretary shall... issue a rule... .”’
(Leg. Hist. 7, 780). The provision was deleted from the Williams
bill by the Senate Committee, without diseussion, see S. Rep. No.
91-1282, 91st Cong., 2d Sess. (1970), Leg. Hist. 141 ef seq. In lien
of the Daniels bill, the House passed the Steiger-Sikes substitute
bill, which did not specify the basis upon which standards were
to be promulgated, but whieh listed ‘‘ views or arguments’? among
the matters which interested parties could submit to the standard-
setting agency. See Leg. Hist. 1120-1122. The Conference Com-
mittee adopted the Senate provision without any discussion of the
“views and arguments’’ language contained in the House provi-
sion. See H.R. Rep. No. 91-1765, 91st Cong., 2d Sess. 35 (1970),
Leg. Hist. 1188. On!» one statement in the legislative history cited
by industry actually dears on the ‘‘views and arguments’’ provi-
sion; it is found in the ‘‘ Additional Minority Views’’ expressed in

11

benzene at or below the level of 10 ppm. The absence of such
direct data is not attributable to any failure on the Secre-
tary’s part. The latency period for benzene-induced leu-
kemia can span a period of twenty years or more. Pet. App.
182a. Consequently, a valid study of actual case histories
(an epidemiologic study) must examine benzene exposures
which oeeurred many years in the past. As the Secretary
explained, ‘‘it is extvemely difficult to reconstruct and de-
fine employee exposures retrospectively.’’ Pet. App. 182a.
Moreover, exposures from that long ago were in general
‘‘most certainly ... at higher concentrations than present
levels.’’ [bid. See Tr. 2556-2557, 2667. Thus, the Secretary
found that there was an absence ‘‘of an adequate size cohort
of workers exposed to benzene less than 10 ppm and who
have been followed for an adequate amount of time.’’ Pet.
App. 182a-183a."

Lacking any direct data showing actual experience with

the House Report. See H.R. Rep. No. 91-2191, 91st Cong., 2d Sess.
53, 54 (1970), Leg. Tist. 883, 884. That statement refutes indus-
try’s theory as to why the provision was revised.

Under the heading ‘Full Protection of the Administrative Pro-
cedures Act Denied,’’ the authors of the minority views stated that
by allowing standards to be based ‘‘upon ‘views 1d arguments’
rather than solely on probative evidence,’’? the Committee bill
‘*departfed] from the provisions of the APA’’ and ‘‘deviat{ed]
from regular and fair procedures.’’ Ibid. They protested that
‘(t]he requirements that administrative decisions be based upon
facts and sound reasons mean little when a loophole of this type is
included in a statute.’ Ibid.

As we have shown in the text, it has long been settled that an
agency subject to a substantial evidence test may properly base its
regulations on expert scientific opinions. It is therefore plain that
the authors of the minority views were not suggesting that the
Secretary could not consider evidence of this sort. Rather, the fair
reading is that the authors of the minority views feared that the
unusual reference to ‘views and arguments’’ in the Committee bill
would permit the Secretary to base a standard on nothing more than
the mere arguments of interested parties.

™ See Addendum B to this brief for a discussion of the evidence
supporting the Secretary’s rejection of industry’s contention that
certain epidemiologic studies demonstrate an absence of leukemia
risk from benzene exposures of 10 ppm or below.

12

benzene exposures at 10 ppm and below, the Secretary con-
sidered whether he could utilize the results of studies in the
record involving exposure to higher concentrations of ben-
zene to estimate how many lives would be at risk at ex-
posure levels of 10 ppm and below. The Secretary con-
cluded that ‘‘there does not exist an adequate scientific
basis for establishing the quantitative dose-response rela-
tionship between exposure to benzene and the induction of
leukemia and other blood diseases.’’ Pet. App. 160a. He
reached that conclusion for two reasons.

First, the predicate for estimating unknown quantities
from known quantities is that there must first be a known
quantity. But the Secretary found ‘‘uncertainty in... the
actual magnitude of expected deaths”? attributable to any
level of exposure. Pet. App. 160a. The available epidemio-
logic data showing the leukemogenie effects of benzene ex-
posures at levels thought to be in excess of 10 ppm do not
show the level and duration of exposure of the cases
studied, and the data understate, to a degree which cannot
be determined, the number of leukemia deaths resulting
from such exposure.© As Dr. Herman Kraybill of the Na-
tional Cancer Institute testified, he along with other scien-
tists had attempted to develop a dose-response curve for
benzene, but had abandoned the effort due to the lack of
necessary data. Tr. 760-761, 776. In Dr. Kraybill’s words
(Tr. 760-761) :

... [We like to estimate risk factors. This has been
done, as many of you recall, with vinyl chloride several
years ago.

... [T]o estimate the risk factors on experimental
data, ... presupposes [that] you have good toxicity
data. When I say toxicity data, I mean good dose-

Further, the Secretary found that genetic, environmental and
other factors produce variations in the susceptibility of individuals
to benzene-induced leukemia. Pet. App. 184a. Consequently, there
is no assurance that the incidence of benzene-leukemia found in any
study of a particular group of employees accurately represents the
incidence which would oceur in the general population.

13

response data[. For] vinyl chloride .. . indeed we did
have that.

But with benzene, it appeared that we didn’t have
this situation, so therefore, most of us gave up....

© & * *
With benezene, we sort of struck out."®

Second, even if the existing studies could have provided
a starting point, the Secretary found that there is no satis-
factory theory of extrapolation from known carcinogenic
effects of a substance at high exposure levels to projected
carcinogenic effects of the substance at lower exposure
levels. Pet. App. 160a. As Dr. Richard Wilson, upon whose
testimony industry places such great reliance, grudgingly
acknowledged (App. 93) :

It is conceivable, I suppose, that any dose-response

curve is possible. ... It would be a rather unusual,

almost I think, revolutionary contribution to the litera-

ture on cancer to have a dose-response curve on cancer.
See also, McGarity, Substantive and Procedural Discretion
in Administrative Resolution of Science Policy Questions:
Regulating Carcinogens in EPA and OSHA, 67 Geo.L.J.
729, 736 n, 27 (1979); Addendum C at 8¢e-9e.""

In the face of these findings, industry continues to cite

16Dr. Bernard Goldstein of the New York University Medical
Center testified to the same effect (Tr. 360) :

We don’t have information that I would say would be good
to give us a relation between any levels of benzene and leu-
kemia. [ just den’t think that one can take this type of data
and say that, ‘‘ Well, let’s see, we got nothing at 50; we got
something at 40; we got nothing at 60.’’ The literature just
is not conducive to that type of analysis.

The testimony of Dr. Charles Shaw of the M. D. Anderson Hospital
and Tumor Institute was in accord (Tr. 415). See also the testi-
mony of Dr. Infante, Tr. 1055-1056, 1156.

17 Industry nevertheless maintains that a meaningful ‘‘risk as-
sessment’’ could be prepared by making ‘‘conservative assump-
tions’’ at each point of uncertainty, so as to arrive at an ‘‘upper-
bound’’ estimate of the risk. See API Br. at 20; Chamber Br. at

14

as if it were gospel an extrapolation submitted by Dr.
Wilson based on selected epidemiologic data in the record.
It is upon this extrapolation that industry largely bases its
claim that the health risk regulated by the Secretary is non-
existent or de minimis. In addition to the Secretary’s gen-
eral findings that such extrapolations are invalid, the Sec-
retary specifically considered and rejected Dr. Wilson’s
extrapolation. Pet. App. 16la-162a. We show in Addendum
C to this brief the particular factors which support the
Secretary’s rejection of Dr. Wilson’s analysis.

10-12. But as we show more fully in Addendum C, given the gaps
in the existing data, the uncertainties which confront any effort to
estimate the extent of the risks posed by low-level benzene ex-
posures are so great that the range of possible assumptions is lit-
erally infinite; ‘‘[t]he choice of assumptions under these condi-
tions of uncertainty transcends the domain of science.’’ Lovins,
Cost-Risk-Benefit Assessments in Energy Policy, 45 Geo. Wash. I.
Rev. 911, 924 (1977). Even if the range of uncertainty were nar-
rower, industry’s notion that truly ‘‘conservative’’ assumptions
ean be identified is mistaken. ‘‘Conservative assessments embody
value judgments in a not clearly identifiable way * * * * [The]
problem with the conservatism argument is that there is no place
at which it can stop * * * * In practice of course, people do stop,
but then it becomes difficult to understand what they mean by
being conservative.’’ Cornfield, Carcinogenic Risk Assessment, 198
Science 693, 695-696 (1977).

Industry contends that various public and private bodies have
endorsed the use of risk assessments based on ‘‘conservative as-
sumptions’’ in a number of regulatory contexts. But that contention
sheds no light on how this statute should be construed and adminis-
tered. However, lest the Court be left with a false impression that
unanimity exists with respect to the role of risk assessments in
regulatory policy, we note simply that this matter is presently a
subject of substantial political debate. See generally Baram, Reqgu-
lation of Health, Safety and Environmental Quality and the Use
of Cost-Benefit Analysis, Final Report to the Administrative Con-
ference of the United States, March 1, 1979 at 19-45. For example,
a Congressional committee has stated that ‘‘[t]he limitations on
the usefulness of benefit/cost analysis in the context of health,
safety, and environmental regulatory decision-making are so severe
that they militate against its use altogether.’’ House Committee on
Interstate and Foreign Commerce, Subcommittee on Oversight and
Investigations, Federal Regulation and Regulatory Reform, 95th
Cong., Ist Sess. 515 (1976).

15

C. This case involves a regulation to protect more
than 30,000 workers from a risk of potentially great;
albeit unknowable, dimensions.

The upshot of the Secretary’s findings, based on the best
available evidence, is that the leukemia risk posed by em-
ployee exposure to benzene at levels of 10 ppm or below
is potentially great, albeit unknowable. The Secretary re-
jected industry’s contention that the risk is non-existent or
de minimis; and the Secretary’s assessment of the nature
of the risk to which the benzene standard is directed, being
firmly grounded in the record evidence, is conclusive.
United States v. Bacto-Unidisk, 394 U.S. 784, 791 (1969).'8

Industry asserts nevertheless that the court below
adopted its characterization of the health risk at issue here.
The Secretary in his Statement said: ‘‘While the actual
estimation of the number of cancers to be prevented [by the
standard] is highly uncertain, the evidence shows that the
number may be appreciable.’’ Pet. App. 160a. The court of
appeals, in considering whether the Secretary had demon-
strated sufficient benefits from the standard to satisfy a
cost-benefit evaluation, stated as follows (Pet. App. 17a,
18a):

OSHA merely assumed that benefits from the reduc-
tion ‘‘may be appreciable.’’

This justification is deficient in one crucial way:
substantial evidence does not support OSHA’s con-
clusion that benefits are likely to be appreciable. With-
out an estimate of benefits supported by substantial
evidence, OSHA is unable to justify a finding that the
benefits to be realized from the standard bear a rea-
sonable relationship to its one-half biliion dollar price
tag.

(Emphasis added).

18 Cf. also Radio Corp. v. United States, 314 U.S. 412, 419-420
(1951); Permian Basin Area Rate Cases, 390 U.S. 747, 766-767
(1968).

16

Industry implies that the court of appeals thus found
that the benefits of the standard are non-existent or de
minimis. But in context the court’s words bear a quite
different meaning. The court’s conclusion was not that the
benefits of the standard are in fact non-existent or de mini-
mis, but that the Secretary had not demonstrated by what
the court deemed appropriate evidence that the benefits of
the standard would justify its costs. As the court went on
to explain (Pet. App. 18a-19a) :

Th[e] finding and deduction [that the standard will
result in some benefit], however, does not yield the con-
clusion that measurable benefits will result, and OSHA
is unable to point to any studies or projections support-
ing such a finding. ... The lack of substantial evidence
of discernible benefits is highlighted when one consid-
ers that OSHA is unable to point to any empirical
evidence documenting a leukemia risk at 10 ppm even
though that has been the permissible exposure limit
since 1971. OSHA’s assertion that benefits from re-
ducing the permissible exposure limit from 10 ppm to
1 ppm are likely to be appreciable, an assumption based
only on inferences drawn from studies involving much
higher exposure levels rather than on studies involving
these levels or sound statistical projections from the
high-level studies, does not satisfy the reasonably
necessary requirement limiting OSHA’s action.

(Emphasis added). See also Pet. App. 33a.

The court, then, did not overturn the Secretary’s conclu-
sion that the health risk addressed by the standard is po-
tentially great, albeit unknowable—i.e., that the ‘‘number
of cancers to be prevented”’ though ‘‘uncertain . . . may
be appreciable.’’ Pet. App. 160a. Rather, the court con-
strued the Act to foreclose the Secretary from issuing a
regulation to protect employees from that risk until empiri-
eal ‘‘studies or projections’’ establish that the risk is in
fact of sufficient magnitude to justify the costs of the stand-
ard. In the court’s words, ‘‘[u]ntil such estimates are pos-

17

sible, OSHA does not have sufficient information”’ to regu-
late. Pet. App. 21a.

As we have shown, the Secretary found that the existing
epidemiologic data are inadequate to serve as a basis for
an extrapolation of the leukemia risk at benzene exposure
levels of 10 ppm or below. There is no basis in the record
for supposing that this inadequacy may be overcome by any
means short of awaiting the experience of long-term ex-
posures to benzene at the levels—i.e., 10 ppm and below—
permitted by the standard which the Secretary sought to
replace. Thus, notwithstanding the court’s disclaimer, it
may well be that the consequence of its ruling is that
‘“‘OSHA must wait until deaths occur as a result of exposure
at levels below 10 ppm before it may validly promulgate a
standard reducing the permissible exposure limit.’’ Pet.
App. 20a. See our opening brief at 27-28, nn. 27-28.

The ccurt said that the Secretary ‘‘must regulate on the
basis of what is known rather than on the basis of what is
unknown.’’ Pet. App. 20a. So, indeed, he must. But as Soe-
rates taught long ago, the highest knowledge is knowing
the limits of our own knowledge. That lesson has special
force where, as here, the inquiry is at the ‘‘frontiers of
scientific knowledge.’? JUD v. Hodgson, 499 F.2d 467, 474
(D.C. Cir. 1974). The Secretary’s instant determination
was based not, as the court of appeals would have it, on
arbitrary ‘‘assumptions,’’ but, in precise accord with the
third sentence of §6(b)(5), on the best scientific evidence
available—evidence concerning both what is known and
what is not known about benzene as a carcinogen. ‘‘ Under
the command of OSHA, it remains the duty of the Sec-
retary to act to protect the workingman, and to act even
in circumstances where existing methodology or research
is deficient.’? Society of the Plastics Indus., Inc. v. OSHA,
509 F.2d 1301, 1308 (2d Cir. 1974), cert. denied, 421 U.S.
992 (1975).

18

CONCLUSION

For the reasons stated herein and in our opening brief,
the decision of the Fifth Cireuit should be reversed insofar
as it invalidates the reduction in the permissible exposure

limit and the provisions related thereto.

J. ALBERT WOLL
Woll & Mayer
815 Fifteenth Street, N.W.
Washington, D.“. 20005

LAURENCE GoOLp
American Federation of
Labor and Congress of
Industrial Organizations
815 Sixteenth Street, N.W.
Washington, D.C. 20006

GrorGE KAUFMANN
2101 I. Street, N.W.
Washington, D.C. 20037

Respectfully submitted,

Grorae H. Conen
Ropert M. WEINBERG
JEREMIAH A. COLLINS

Bredhoff, Gottesman, Cohen
& Weinberg

1000 Connecticut Avenue, N.W.

Washington, D.C. 20036

ELLIOT BREDHOFF

General Counsel
Industrial Union Department,
AFL-CIO

1000 Connecticut Avenue, N.W.

Washington, D.C. 20036

Attorneys for Petitioner
Industrial Union Department,

AFL-CIO

Pek

ADDENDUM A

A Survey of ‘‘Non-Absolutist’’ Statutory Ap-
proaches to Protecting Health and Safety

Balancing of costs and benefits is expressly required by
several regulatory statutes. An early example is the Flood
Control Act of 1936, 33 U.S.C. § 701(a): ‘*[T]he Federal
Government should improve or participate in the improve-
ment of navigable waters .. . for flood-eontrol purposes if
the benefits to whomsoever they accrue are in excess of the
estimated costs and if the lives and social security of people
are otherwise adversely affected.’’ More recent examples
include two statutes in which Congress has provided that
in addition to being ‘‘feasible,’’ regulations promulgated
under the statute should reflect a balancing of costs and
benefits. The Outer Continental Shelf Lands Act Amend-
ments of 1978, 43 U.S.C. § 1347(b), provide that offshore
drilling operations shall use ‘‘the best available and safest
technologies which the Secretary determines to be eco-
nomically feasible, . . . except where the Secretary deter-
mines that the incremental benefits are clearly insufficient
to justify the incremental costs of utilizing such technol-
ogies.’’ The Energy Policy and Conservation Act of 1975,
42 U.S.C. § 6295 (a) (4) (D), provides that energy efficiency
standards promulgated for certain products must not only
be ‘‘economically feasible,’’ but must, in addition, be such
that ‘‘the savings in operating costs . . . outweigh’’ the
adverse effects of the standard.

In several instances, Congress has required some form of
cost-benefit balancing with respect to the promulgation of
certain regulations under a statute, while requiring other
regulations under the same statute to be developed without
regard to any such balancing, albeit with some considera-
tion of costs. A good example is the Federal Water Pollu-
tion Control Act Amendments of 1972, as amended, 33
U.S.C. § 1251 et seq.

That statute specifies that in setting effluent limitations
to be achieved by 1977, the Environmental Protection

la

.
|

Agency (‘‘EPA’’) is to require the use of the ‘‘best practi-
cable control technology currently available (§ 1311 (b)
(1)), and the statute further provides that EPA’s ‘‘assess-
ment of [that] technology shall include consideration of the
total cost of application of technology in relation to the
effluent reduction benefits to be achieved from that applica-
tion... .’’ (§ 1314 (b)(1)(B)). The latter requirement en-
tails a ‘‘limited cost-benefit analysis’’— a ‘‘balancing test
between total cost and effluent reduction benefits [which]
is intended to limit the application of technology only where
the additional degree of effluent reduction is wholly out of
proportion to the costs of achieving such marginal level of
reduction. .. .”’ 118 Cong. Ree. 33696 (1972), discussed in
E. I. duPont de Nemours €& Co. v. Train, 430 U.S, 112, 129-
130 (1977).

On the other hand, the 1972 Amendments provide that
effluent limitations to be attained by 1983 should be based
on ‘‘the best available technology economically achieveable”’
(as amended, § 1311 (b)(2)(A)), and under that provision,
while costs are to be considered, ‘‘no balancing test will be
required.’’ 118 Cong. Rec. 33696 (1972). See Weyerhaeuser
Corp. v. Costle, 590 F.2d 1011, 1044-1046 (D.C. Cir. 1978).

The amendments also authorize EPA to set more strin-
gent effluent limitations for particular sources of pollution
where the generally mandated levels will not achieve the
required standard of water quality in a particular setting;
but before so doing the Administrator must conduct a pub-
lic hearing to assess the ‘‘economic and social’’ costs and
benefits involved (33 U.S.C. § 1312 (b)(1)), and the more
stringent limitations cannot be applied against any person
who demonstrates at that hearing that ‘‘there is no reason-
able relationship between the economic and social costs and
the benefits to be obtained. . . .’’ 33 U.S.C. ¢ 1312 (b) (2).

Similarly, the Clean Water Act of 1977, 33 U.S.C. § 1251
et seq., which established new criteria for the regulation of
*‘eonventional pollutants,’’ requires that in setting effluent
limitations for such substances EPA shall consider ‘‘the
reasonableness of the relationship between the costs of at-

2a

taining a reduction in effluents and the effluent reduction
benefits derived. .. .’’ 33 U.S.C. § 1314 (b)(4)(B).

Another statute which requires cost-benefit evaluations
for some regulations but not others is the Clean Air Act
Amendments of 1970, as amended, 42 U.S.C. § 7401 et seq.,
which were passed the same week as OSHA. In that statute,
the provision which authorizes EPA to prohibit the manu-
facture or sale of a fuel or fuel additive that significantly
impairs an emission control device expressly requires EPA
to perform ‘‘a cost benefit analysis’? (42 U.S.C. §7545f-6¢(c)
(2)(B), while the provision which authorizes the prohibition
of a fuel or additive that would ‘‘endanger the public health
or welfare’’ contains no such requirement (42 U.S.C.
§ 7545f-6e (¢)(2)(A)).

Another ‘‘non-absolutist’’ approach to health and safety
regulation, requiring a particular form of cost-benefit
balancing, has been adopted by Congress in the several
statutes which are based on the concept of ‘‘unreasonable
risk.’’

One of the first ‘‘unreasonable risk’’ provisions is con-
tained in the National Traffic and Motor Vehicle Safety
Act of 1966, as amended, 15 U.S.C. §1391(1). ‘*As the
Senate hearings make clear, the word ‘unreasonable’ was
placed in the bill deliberately, to signify a ‘common-sense’
balancing of safety benefits and economic cost.’’? United
States v. General Motors Corp., 518 F.2d 420, 435 (D.C.
Cir. 1975). See S.Rep. No. 89-1301, 89th Cong., 2d Sess. 6
(1966).

The year before OSHA was enacted, Congress incorpo-
rated an ‘‘unreasonable risk’’ requirement in the Child Pro-
tection and Toy Safety Act of 1969, as amended, 15 U.S.C.
§ 1261(s). The provision was included so as to require con-
sideration of ‘‘such factors as the utility of the object, the
degree of danger it presents, and the feasibility of design-
ing out that danger.’’ S.Rep. No. 91-237, 91st Cong., 1st
Sess. 7 (1969). See generally Forester v. CPSC, 559 F.2d
774, 789 (D.C. Cir. 1977).

3a

Two years after OSHA was enacted, Congress adopted
an ‘‘unreasonable risk’’ requirement in two more statutes.
In incorporating that concept in the Consumer Product
Safety Act of 1972, Congress stated:

Protection against unreasonable risks is central to
many Federal and State safety statutes and the courts
have had broad experience in interpreting the term’s
meaning and application, It is generally expected that
the determination of unreasonable hazard will involve
the Commission in balancing the probability that risk
will result in harm and the gravity of such harm
against the effect on the product’s utility, cost and
availability to the consumer.

H.R. Rep. No. 92-1153, 92d Cong., 2d Sess. 33 (1972).
See also S. Rep. No. 92-749, 92d Cong., 2d Sess. 6,
14-15 (1972). We discussed the pertinent provisions of the
Consumer Product Safety Act in our opening brief at pp.
30-33. Similarly, in the Federal Environmental Pesticide
Control Act of 1972, as amended, 7 U.S.C. § 136(bb), Con-
gress defined the term ‘‘unreasonable adverse effects on the
environment’’ to mean ‘‘an unreasonable risk to man or the
environment, taking into account the economic, social, and
environmental costs and benefits of the use of any pesti-
cide.’’ The EPA may regulate a pesticide only to the extent
that the pesticide entails such ‘‘unreasonable adverse ef-
fects.’ See 7 U.S.C. §§ 136a(¢e)(5)(C), (D), 136a(e)(7),
(8), 136a(d), 136d(b). This explicitly requires a balancing
of costs and benefits. See Environmental Defense Fund v.
EPA, 548 F.2d 998, 1005 (D.C. Cir. 1976).

Recently, Congress enacted yet another statute based on
the concept of ‘‘unreasonable risk’’ and explicitly requiring
a balancing of costs and benefits: the Toxic Substances
Control Act of 1976, 15 U.S.C. § 2605(a), (e¢) (1).

Finally, the provisions of the Federal Food, Drug, and
Cosmetic Act concerning human and animal drugs, 21
U.S.C. §§ 321(p) (1), 360b, illustrate an altogether different
**non-absolutist’’ approach. In determining whether a hu-
man drug is ‘‘safe’’ the Food and Drug Administration

4a

(‘‘FDA’’) must assess whether ‘‘the expected therapeutic
gain [from the drug] justifies the risk entailed by its use.’’
United States v. Rutherford, —— U.S. ——, 47 U.S.L.W.
4724, 4727 (June 8, 1979). That is to say, a drug is ‘‘safe’’
only if its curative benefits outweigh its side effects. But
while the statute requires a balancing of therapeutic costs
and benefits, it does not contemplate a balancing of economic
costs and bevefits. See Decision of the Commissioner in the
Matter of Diethylstilbestrol: Withdrawal of Approval of
New Animal Drug Applications, 44 Fed. Reg. 54852, 54881-
54883 (Sept. 21, 1979). On the other hand, the provisions
of the Food, Drug, and Cosmetic Act which authorize FDA
to set ‘‘tolerances’’ for certain hazardous substances (e.¢.,
21 U.S.C. §§ 346, 346a, 348(c)(4)(A), (B)) may contem-
plate some form of economic cost-benefit balancing. Id. at
54881-54882.

Sa

Be Pe eae fa bone eng cred (en ga eae, Flt th De ea ene ee eS ee ee ee ee ee

ADDENDUM B

An Analysis of the Basis for the Secretary’s
Rejection of Industry’s Contention That Epidemo-
logic Studies Have Shown Benzene to Be Safe at
Low Levels of Exposure

In its attempt to portray the risks arising from exposure
to benzene at concentrations of 10 ppm and below as non-
existent or de minimis, industry relies dn a group of epi-
demiologic studies which the industry respondents cited
first to the Secretary and then to the court below in their
effort to demonstrate that a safe ‘‘threshold’’ level of ex-
posure exists for benzene. API Br. at 14-17. See pp. 8-10.
of the reply brief, supra. The Secretary expregsly rejected
industry’s contentions regarding the sonitenlons which
can be drawn from ithe studies, and the court of appeals did
not upset the Secretary’s determinations. Undaunted, in-
dustry now offers its theories regarding those studies to
this Court, as if those theories were fact.

Industry advances two arguments regarding the epide-
miologic data. First, industry claims ‘‘it was shown that
when benzene exposures were reduced below 100 ppm—but
not eliminated—the excess incidence of leukemia found in
the principal studies cited by OSHA (the reports of Drs.
Infante, Aksoy and Vigliani) ceased to occur.’’ API Br. at
14 (emphasis in original). Second, industry asserts that
‘this [purported] evidence that Jow-level benzene expo-
sures present no significant risk of leukemia was bolstered
by a series of epidemiologic studies of hundreds of thou-
sands of workers exposed to low concentrations of ben-
zene’’ which, according to industry, ‘‘all failed to uncover
any excess incidence of leukemia... .’’ API Br. at 14-15.
| (These Iaiter studies are referred to as the ‘‘negative
studies. '’)

In this Addendum, we show that industry’s representa-
tions regarding the epidemiologic studies—which are ‘‘sup- ’
ported’’ by little more than citations to the arguments
presented in the briefs filed by the industry respondents

1b

il

|

before the Secretary—seriously distort the facts as found
by the Secretary and as established by the record. As we
show, the Secretary correctly found that the epidemiologic
studies lend no support to industry’s contention that the
risks posed by worker exposure to benzene at concentra-
tions of 10 ppm and below are nonexistent or de minimis.

I. The A&R. Vigliani and Infante Studies

Industry maintains that the data assembled by Dr. Aksoy
and Dr. Vigliani, which demonstrate a risk of leukemia at
certain exposure levels, show that ‘‘when benzene expo-
sures were reduced below 100 ppm—but not eliminated—the
excess of leukemia found in the ... studies . . . ceased to
occur’? API Br at 14 (emphasis in original). Although
industry makes this representation as if it were an undis-
puted fact, in reality industry’s view of the data is directly
contrary not only to the findings of the Secretary regard-
ing this data, but to the statements of Dr. Aksoy and Dr.
Vigliani as well. ©

Industry’s view of the Aksoy and Vigliani data is
premised on the assumption that when the use of high-
benzene solvents ceased in the workplaces studied by Drs.
Aksoy and Vigliani, the solvents which were substituted
continued to contain substantial amounts of benzene. But
Dr. Aksoy stated that ‘‘[t]he substitutes for benzene in the
workplaces studied were analyzed by gas chromatography
... The results showed that the substitutes contained no
benzene.’’ Ex. 60 at 3 (emphasis added). See also Tr. 2567
(testimony of Dr. Irving Tabershaw). Similarly, Dr. Vigli-
ani wrote that no new cases of leukemia were reported in the
workplaces he studied after ‘‘the replacement of benzene
with toulene’’ (Ex. 2.15 at 125); and the Secretary found
that Dr. Vigliani’s data demonstrate that ‘‘toulene and
benzene are ‘. . . quite different chemical entities from
toxicological and pharmacological points of view’ and that,
‘toulene free from benzene can be safely used in industry
as a suitable solvent in replacement of benzene.’’’ Pet. App.
101a (emphasis added).

2b

et

Thus, as the Secretary found, the Aksoy and Vigliani
data show that the excess leukemia incidence in the popu-
lations in question ended when exposures to benzene had
been eliminated—not, as industry would have it, when
exposures had been reduced to low levels, Pet. App. 100a-
102a. The court of appeals did not disagree with that con-
clusion. See Pet. App. 8a, n. 13.

Industry’s interpretation of the Infante study is as far-
fetched as its view of the Aksoy and Vigliani studies. In-
dustry makes much of the fact that while Dr. Infante’s
study found a large excess incidence of leukemia among a
group of workers who had been exposed to benzene in the
1940’s, a study by University of North Carolina (‘‘UNC’’)
researchers involving workers employed at a later date in
one of the Pliofilm operations which had been studied by Dr.
Infante did not uncover any excess incidence of leukemia.
API Br. at 14, citing Ex. 187.B-13. Industry contends that
this divergence results from the fact that exposures in the
plant were probably lower during the time period studied
by these researchers than during tke time period studied by
Dr. Infante. But the number of Pliofilm workers studied
by UNC was so small that the study involved only 18
deaths from any cause among those workers. Ex. 187.B-13,
Table 1, p. 9.1 In the general population, only about .75
percent of deaths are from leukemia. See Tr. 2791-2792.
Consequently, out of 18 deaths in the general population,
the predicted number of leukemia deaths would be 0.13.
Thus, one would not expect to find a leukemia death among
the workers studied by UNC even if those workers were sub-
jected to a risk factor much higher than that of the general
population. As Dr. Infante testified with respect to a similar
study involving the same plants:

‘‘[W]hen you are expecting less than one case, and...
you observe none, ... to make any kind of scientific

1 able 1 shows that 18 deaths due to all causes were observed in
“*Oecupational Title Group 19.’’ Group 19 is the Pliofilm area, see
Ex. 187.B-13 at 8, 40.

3b

statement ... even about absence of risk on those num-
bers would, to me, be entirely misleading. (Tr. 1079-
1080) .?

II. The ‘‘Negative Studies’’

Industry asserts that the absence of any ‘‘significant’’
risk of leukemia at low-level benzene exposures was demon-
strated at the hearing by ‘‘a series of epidemiologic studies
of hundreds of thousands of workers exposed to low con-
centrations of benzene.’’ API Br. at 14-15. But the Secre-
tary analyzed each of those studies and found that because
of gross deficiencies in the design of the studies, their re-
sults are meaningless. See Pet. App. 107a-112a, 123a-124a,
182a. As we now show, the Secretary’s appraisal of the
‘‘negative studies’’ was well-founded.

a) The Thorpe Study. Industry places its primary re-
liance on a study conducted by Dr. John Thorpe, Associate
Medical Director of Exxon Corporation. See API Br. at
15-16. But the Secretary found the Thorpe study to be
invalid (Pet. App. 110a-111la); and that judgment was sup-
ported by the testimony of NIOSH (Tr. 745) and several
expert witnesses. See, e.g., Tr. 914, 1200-1201 (Dr. Infante) ;
Tr. 914-915 (Dr. Smith).*

2 In addition, the UNC study used an extremely inexact method
for determining the department in which an employee was deemed
to have worked. See Ex. 187.B-13 at 25. As a result, it is entirely
possible that many of the workers whom the UNC study assumed
had worked in Pliofilm operations had actually spent most of their
time in other departments where benzene was not present. For
these and other reasons, the UNC researchers certainly did not
regard their study as demonstrating that low levels of benzene
exposure pose no significant health risks. To the contrary, the study
recommended that ‘‘[e]xposures to benzene should be maintained
as far below the Threshold Limit Value [i.e., 10 ppm] as is pos-
sible.’’ Id. at 35.

3 See also the responses of Dr. Lamm and Dr. Tabershaw, testify-
ing for respondent Manufacturing Chemists Association, to the
questioning of Dr. Lloyd. Tr. 2661-2664.

4b

Dr. Thorpe had almost no idea what proportion of the
workers studied actually had been exposed to benzene, let
alone how long they had been exposed (Tr. 2021-2022, 2025-
2029). As was conceded by Dr. Steven Lamm, a witness
for respondent Manufacturing Chemists Association
(‘‘MCA”’), ‘there may well have been, within Dr. Thorpe’s
group, some subgroups which had exceedingly high rates of
leukemia that he didn’t identify in his analysis.’’ Tr. 2664.4

An epidemiologic study which, like Dr. Thorpe’s, is ‘‘di-
luted’’ by the inclusion of large numbers of unexposed
workers can easily mask severe cancer risks. This is illus-
trated by a series of studies concerning carcinogenic coke
oven emissions which demonstrate dramatically the danger
of relying on a ‘‘diluted’’ study. The original coke oven
studies were not limited to the employees who were actually
exposed to the emissions from the ovens and at first did not
show a significant excess incidence of cancer among the
employees studied. But when the researchers went on to
focus on a particular subgroup, top-side coke oven workers,
who were so exposed, an enormous increase in the incidence

* Industry quotes the statement of Dr. Robert Snyder, a witness
for respondent American Iron and Steel Institute, that ‘‘using the

most powerful statistical tools available and concentrating on the
most highly exposed population of workers, it was not possible to
detect a greater incidence of leukemia [in the Thorpe study] than
in the population at large.’’ API Br. at 15-16, n. 35, quoting Ex.
156-2 at 17. Industry implies that Dr. Snyder meant to indicate
that the overall results of the Thorpe Study are such as to rule out
the possibility that the workers included in the study who were
actually exposed to benzene may have suffered a significant excess

incidence of leukemia. But to the extent Dr. Snyder’s statement
could be so interpreted, he retracted it on cross-examination. When

questioned regarding his statement, Dr. Snyder explained that in
referring to ‘‘the most highly exposed population of workers,’’ he
was not referring to the particular subgroup of workers in the
Thorpe study who were actually exposed to benzene; rather, he
explained, ‘‘[p]eople who are working . . . in the petroleum in-
dustry [a]re exposed to a variety of chemicals and that is all it
means.”’ Tr. 3227-3228. Dr. Snyder further testified as follows:

5b

of lung cancer was detected.’ As was noted repeatedly
during the benzene hearing, the coke oven emissions data
present a classic example of how easily a severe risk to an
exposed subgroup can be obscured by data concerning a
group containing large numbers of unexposed workers.
See, eg., Tr. 1487-1438, 2329.°

In short, as the Secretary stated, ‘‘[t]he fundamental
premise of any epidemiological investigation of the poten-
tial relationship of occupational exposure to an agent and
observed health effects is... a careful definition of an
exposed cohort.’’ Pet. App. 124a. See also Pet. App. 182a.

~———e-

Dr. Lloyd: But IT wonder what your point is about the use of
‘powerful statistical tools,’’ what that might add to the
analysis.

Dr. Snyder; His data was incomplete, but within the frame-
work of the data that he had here, trying to draw some con-
clusions based on the statistics he could apply, given that these
are imperfect, and given the fact that we know he didn’t have
complete data, he tried to do the best he could with what he
had, and that was the only point that I was trying to make
here. Tr. 3292-3294.

Similarly, the statement in a paper by the National Academy of
Sciences that the Thorpe study ‘‘seem[s] to indicate that there is
not an outstandingly high incidence of leukemia in the workers
studied’? (Ex. 2-4 at ii, cited in APT Br. at 16, n. 36) obviously
refers to the entire group studied by Dr. Thorpe, not to the sub-
group of workers who were actually exposed to benzene.

SSee Ex. 113; Ex. 217.27.D. Largely as a result of these later
studies, employee exposure to coke oven emissions is now regulated
by an OSHA standard, which has been upheld in pertinent part by
the Court of Appeals for the Third Circuit, American Iron and
Steel Institute v. OSHA, 577 F.2d 825 (1978), petitions for cert.
pending, Nos. 78-918, 78-919.

® Indeed, the concessions made by industry witnesses such as
Dr. Lamm and Dr. Snyder regarding the ‘‘dilution’’ inherent in
the Thorpe study came in response to the questions of Dr, William
Lloyd, who was the principal author of the coke oven studies. See,
e.g., Tr. 2664.

ee

6b

PA tithe cB es 1 kc nomen tn

The Thorpe study, as the Secretary found, failed to comply
with this threshold requirement. See Pet. App. 110a-111a.

Nor was the ‘‘dilution’’ problem the only defect in the
Thorpe study. The Secretary cited several others. Pet. App.
110a, For example, Dr. Thorpe’s method of determining
the number of leukemia deaths suffered by the workers
was highly inaccurate (Tr. 2015-2029, 2071, 2074-2075).
Moreover, the study excluded altogether the workers who
had terminated their employment early, and the medical
history of retirees was not investigated with any accuracy.
(Tr. 2023-2025). Workers who terminate(their employment
early, and also retirees, are known tc exhibit a particularly
~ age of disease, and therefore any-study which
excludes-those groups may severely understate the cancer
risk which is actually present. In this regard the classic ex-
ample is a series of British studies concerning vinyl chlo-
ride, in which the very great cancer hazard posed by that
substance 7 was not evident until workers ‘who had termi-
nated their employment early were included in the studies.
See Tr. 1200, 1438-1440. As Dr. Infante testified, by exclud-
ing such workers the Thorpe study is guilty of the very flaw
which was highlighted by the British vinyl chloride studies.
Tr. 1200.

In short, while Dr. Thorpe may not be ‘‘ashamed’’ of his
study (Tr. 2021), the Secretary was right in regarding the
study as casting no light on the question of the leukemia
risk posed by low levels of exposure to benzene.

b) The Tabershaw/Cooper Study. The Tabershaw/
Cooper Study, relied upon by industry (API Br. at 15)
suifered from the same basic flaws as the Thorpe study, but
to an even greater degree. As the Secretary emphasized,
this study, involving 17 selected refineries, was designed
to investigate the effects of exposure to hydrocarbons in
general, not benzene. Pet. App. 108a. Dr. Williatn Gaffey,

a Employee exposure to vinyl chloride is regulated by an OSHA
standard which was upheld in Society of the Plastics Industry, Inc.
v. OSHA, 509 F.2d 1301 (2d Cir.), cert. denied, 421 U.S. 992
(1975).

7b

one of the authors of the report, testified that for all he
knew, fewer than 10 percent of the workers studied may
have been exposed to benzene. Tr. 1411; see also Tr. 1351,
1358, 1484. The Secretary properly concluded that ‘‘since
no group of workers exposed to benzene was clearly identi-
fied, this study provides no relevant information as to the
leukemogenic risk of exposure to betizene.’’ Pet. App. 108a,
Dr. Gaffey in effect admitted the correctness of the Secre-
tary’s conclusion when he was asked a question comparing
his study to the ‘‘diluted’’ coke oven studies discussed
above:

Mr. Wagoner: What I would like to ask you is given
the similarities between the two studies, could you
\.possibly have within your study population, a smaller
subsegment who had been exposed to benzene who are
experiencing a risk factor of the order of magnitude
similar to that found by Dr. Lloyd in the steel industry,
and yet by Atudying it in the method that you have,
you would have obscured or overlooked this risk?
Dr. Gaffey: IT cannot rule out that possibility. (Tr.
1438). ;

Tn addition, just as the Thorpe study had exeluded work-
ers who terminated their employment early, so did the
Tabershaw/Cooper Study. As just noted, this method of
proceeding obscures the leukemia risk posed by benzene to
the workers studied by Dr. Gaffey and his associates. See
Tr. 1488-1440. For these reasons, the Seeretary’s refusal
to rely on the Tabershaw/Cooper study (Pet. App. 107a-
108a) was well-founded.

ec) The Stallones Study. The study by Dr. Reuel Stallones
of Shell Oil Company workers (see API Br. at 15, referring
to Ex. 115.C-2) rests on no better footing. As the Secretary
noted ‘‘[a]gain, failure to correctly identify or define a
benzene-exposed cohort limits the usefulness of this study.”’
Pet. App. llla. The study ineluded all Shell employees,
including office workers (see Tr. 2324-2325), and Dr. Stal-
lones had to concede that ‘perhaps quite a large propor-
tion [of the group studied] are not exposed to benzene.”’

8b

a rt ates.

VS Rtrmaetnin st

Tr. 2302. Nor did Dr. Stallones have any idea as to the
lengths of exposures involved. Tr. 2300. Moreover, 80%
of the exposures monitored were below 1 ppm. Tr. 2296. Like
the other ‘‘negative studies,’’ the Stallones study excluded
workers who left their employment early. Tr. 2298. Con-
sequently, as was made clear by Dr. Stallones’ responses to
questions posed by Dr. Infante, no conclusions can be
drawn from this study. See Tr. 2326-2331.

d) The Steel Industry Studies. 1t is remarkable that
industry should claim to find significance in ‘‘a long-term
mortality study of the steel industry,’’ which, according to
industry, ‘‘showed that coke oven and by-products facili-
ties workers, exposed over the years to low concentrations
of benzene, had not experienced any excess incidence of
leukemia.’? API Br. at 15. For these are the very coke oven
studies which demonstrate the error of drawing any con-
clusions regarding the risk posed by a chemical from a
study which does not focus specifically on workers who are
actually exposed to the chemical. (See pp. 5b-6b, supra.)

The coke oven studies, designed to determine the inci-
dence of cancer (particularly lung cancer) resulting from
coke oven emissions, did not seek to identify a group of
workers exposed to benzene. Consequently, as Dr. Lloyd, the
principal author of the studies, established by his ques-
tioning of the witnesses for the American Iron and Steel
Institute, there is no way to tell from the coke oven studies
what the incidence of leukemia may be among those steel-
workers who are actually exposed to benzene. See Tr. 3315.
And see generally Tr. 3300-3306, 3313-3315. In addition,
the number of employees in the work area suggested by
industry as most likely to include some benzene-exposed
workers—the coke plant by-products area—was too small
to permit any conclusions regarding the relationship be-
tween benzene and leukemia to be drawn in any event. See
Tr. 3314. In short, as is apparent from Dr. Lloyd’s exami-
nation of the AIST witnesses, industry’s assertion that his
studies demonstrate anything with respect to benzene is a
gross misuse of the studies. The Secretary properly con-

9b

cluded that ‘‘[g]iven the small numbers of observed and
expected rates of leukemia, and in the absence of a study
designed to assess coke oven worker exposure to benzene . i,
it [is] inappropriate to rely on these studies... .’’ Pet. App.
11lla.

Summary
As this review has shown, industry’s contention that low
levels of exposure to benzene involve risks which are non-
existent or de minimis finds no support whatsoever in the
epidemiologic studies upon which industry relies, and was
properly rejected by the Secretary.

10b

ADDENDUM C

An Additional Analysis of the Evidence Support-
ing the Secretary’s Finding That It is Impossible
to Derive a Meaningful Risk Assessment for
Benzene, and in Particular, His Rejection of Dr.
Wilson’s Attempted Risk Assessment.

As shown in the reply brief, supra at 10-14, the Secretary
found that there is no way to extrapolate from the risks
attributable to high concentrations of benzene to predicted
risks of benzene exposure at levels of 10 ppm and below;
and the court of appeals did not upset that finding. The
Secretary based his finding on inadequacies in the known
data and in the hypothesized theories of extrapolation.
Further, the Secretary specifically considered and rejected
the handwritten extrapolation prepared by Dr. Richard
Wilson, a physicist, which industry submitted in the rule-
making proceedings, and which, notwithstanding the Sec-
retary’s rejection, industry presents to this Court.

We show herein that an examination of the Wilson testi-
mony illustrates the correctness of the Secretary’s finding
that such extrapolations regarding the risks of exposure
to benzene are invalid. As we show, Dr. Wilson’s method
was to ignore the gaps in the existing data or, where he
acknowledged that such a gap existed, to fill it in with as-
sumptions which were designed to understate the magnitude
of the projected risk.

I.

To construct an extrapolation from known points to
unknown points, there must first be known points—certain
dose-response points for benzene-leukemia must be identi-
fied as a predicate to any extrapolation of the response at
other doses. A given ‘‘dose’’ of benzene is essentially a
function of two factors: the level of benzene exposure and
the duration of benzene exposure. Regular exposure to 1
ppm of benzene over 40 years is not the same dose of
benzene as exposure to 1 ppm of benzene over a five year

le

period. The ‘‘response’’ to a given dose of benzene is the
incidence of leukemia associated with that dose.

To select a starting point for his extrapolation, Dr. Wil-
son referred to four of the epidemiologic studies in the
record: those of Drs. Infante, Vigliani, Aksoy and Thorpe.
On a hand-written graph, reproduced in the margin,’ Dr.
Wilson recorded his understanding of the results of the
four studies. As the graph shows, for each study Dr. Wilson
recorded the exposure level he believed to have been in-
volved, and the multiple of the leukemia incidence in the
general population (i.e., the ‘‘risk factor’’) he believed to
have been reported.”

But what Dr. Wilson recorded does not correspond to the

1 Dr, Wilson’s graph, which appears at App. 170, is as follows:

Via@linnt
ITALY
Rovan Prior

oF
BENzENE INDUCED

LEvKE MIA, V3.

é CoN CENTRATION

vee” Fi'quee 4.

Aksey (Teaxev) Av. Bewren€
Concentration

; th
. 200 400 wiles Riko. roe
. 1 i l | baila

i Cppesewe st: NDARD

2 On the graph, the horizontal axis plots benzene exposure levels
and the vertical axis plots the risk factor. Each of the four studies
is represented by a point on the graph, identified by the names of
the four researchers. Vertical and horizontal lines extending out
from each of the four points represent the range of uncertainty
which Dr. Wilson believed to exist with respect to the risk factor
and exposure levels involved in each study.

2c

ee) ee ee ee Ce ee ee

actual data produced by the studies. The data in fact, as
the Secretary found, do not provide sufficient information
to plot a dose-response curve. Dr. Wilson filled in the gaps
in the data with assumptions which are not only unfounded
but tend to greatly understate the risk of leukemia resulting
from benzene exposure. We illustrate these points below
by reference to Dr. Wilson’s treatment of each of the
factors involved in a dose-response analysis.

DOSE

A. Duration. Dr. Wilson did not even discuss the question
of the duration of the benzene exposures involved in the
four studies. In utilizing the studies for his extrapolation,
he apparently assumed, erroneously, that the leukemia
incidence reported in each study represents the risk re-
sulting from benzene exposure spanning an entire working
life at the levels involved in the study. See App. 89. In
actuality, however, the duration of the exposures involved
in each of the studies is highly uncertain, and may well
have been quite short.

For example, the group of employees studied by Dr.
Infante included every white male employee who worked
at any time during the years 1940 to 1949 on a job which
was thought to entail exposure to benzene at the plants in
question. Ex. 40 at 5. It is impossible to estimate what the
actual durations of exposure may have been. A later study
involving the same plants suggests that workers may have
moved frequently from jobs involving benzene exposure to
jobs involving no exposure, and vice-versa. See Ex. 187.B-
13 at 25. Hence the aggregate exposures reflected in the
Infante study may have been short. Similarly, in both the
Vigliani and Aksoy studies it is likely that the duration of
exposure varied greatly from employee to employee, with
many employees exposed for only a few years, and few if
any employees exposed for more than fifteen or twenty
years. See Pet. App. 100a-101a.°

3 As for the Thorpe study, the Secretary properly found that
study to be meaningless (see Addendum B, pp. 4b-5b), and we
therefore do not consider it here.

8c

A §6(b)(5) standard must be geared to protecting an
employee who ‘‘has regular exposure to the hazard dealt
with by such standard for the period of his working life.’’
To provide the information necessary under § 6(b)(5), an
extrapolation must show the risk of leukemia associated
with a given level of exposure over the course of an em-
ployee’s entire working life: some thirty-five to fifty years.
The data here do not allow for any such extrapolation. Dr.
Wilson simply willed away that problem by falsely as-
suming that the epidemiologic studies represent the effects
of lifetime exposures. His ‘‘risk assessment’’ is therefore
not only meaningless but necessarily understated to a
significant, albeit unknowable degree.

B. Level. In constructing his graph, Dr. Wilson had to
determine the levels of exposure which had produced the
- leukemia incidence reported in each study, But as the Secre-
tary found, enormous uncertainty surrounds that question.
See, e.g., Pet. App. 182a. For example, in the case of the
Infante study the evidence indicates that at some times the
benzene concentrations at the plants in question may have
been as low as 0 to 15 pprh, while at other times they may
have been much higher. And with respect to that study the
Secretary found:

... [A]ny conelusion regarding the actual individual
time-weighted average exposure is speculative. Be-
cause of the lack of definitive exposure data, OSHA
cannot derive any conclusions linking the excess leu-
kemia risk observed with any specific exposure level.

Td. at 105a.

Consequently, as Dr. Douglas Smith of NIOSH testified,
it is impossible to determine what level of exposures actu-
ally caused the observed leukemia deaths reported by Dr.
Infante: |

A lot of emphasis has been placed upon the fact that
individuals could have been exposed to these very high
levels that were basically extremes and as has been
stated, it is possible but we don’t know any more than

4c

to speculate. By the same token, one can speculate also
that they were exposed to the low ranges as well, and
that effects could have occurred from the low ranges.

Tr. 996,

Dr. Bernard Goldstein of the New York University
Medical Center made the same point with respect to all the
epidemiologic studies of benzene exposure: ‘‘[WJe really
have no information as to the exposure level of any of the
individuals who have developed acute leukemia.’’ Ex. 75
at 2.

Nevertheless Dr. Wilson’s graph interprets each of the
epidemiologic studies as showing a particular incidence
of leukemia resulting from a particular exposure level
(or in some cases, a narrow range of exposure levels). Thus,
in this respect as well, the Wilson risk assessment rests on
an assumption contrary to fact and to the Secretary’s
findings.

RESPONSE

Dr. Wilson did not take into account the possibility that
the studies may have greatly underreported the actual
incidence of leukemia in the workers studied. Yet Dr.
Infante was unable to determine whether or not twenty-five
percent of the workers included in the group he studied had
died, and for purposes of the study he assumed that every
one of them was alive. Among the other seventy-five percent
of the workers, the additional leukemia cases discovered by
Dr. Sakol among the employees studied by Dr. Infante
indicate that Dr. Infante’s method of determining the
causes of death among the workers resulted in his missing
a significant number of cases of leukemia.* Also, many of

Industry erroneously asserts, API Br. at 11, n. 19, that the
additional leukemia deaths found by Dr. Sakol do not indicate that
Dr. Infante’s study underestimated the risk of leukemia. That
assertion is contrary to the testimony of the authors of the Infante
study. See Tr. 768-772, 903-905.

5e

the workers studied were still alive, and thus their eventual
cause of death cannot be known.®

As a result of these and other factors, Dr. Infante’s own
assessment was that his study had ‘‘tremendously under-
estimated the true risk of leukemia in this population... .”’
Tr. 768; see also Tr. 874-875. Edward Baier, then Deputy
Director of NIOSH, agreed with that assessment. Tr. 747.
And Dr. Sakol’s conclusion was that the Infante study
‘grossly underestimates the incidence’’ of leukemia in the
workers studied. Tr. 314.

Similarly, by Dr. Aksoy’s own account, the Aksoy study
substantially understates the actual incidence of benzene
leukemia among the workers he studied (Ex. 60 at 1-2). And
the Thorpe study was designed in a manner which tends so
patently to understate the risk of benzene leukemia that the
Secretary properly found that study to be meaningless. See
Addendum B, pp

Nevertheless, in performing his risk assessment Dr.
Wilson did not take into account the likelihood that all
of the epidemiologic studies may seriously underestimate
the risk of benzene leukemia. To the contrary, as we next
discuss, he arbitrarily chose to assume that the Infante and
Vigliani studies had greatly overstated the risk.®

For the reasons just stated, Dr. Wilson’s ‘‘risk assess-
ment,’’ based on the studies of Drs. Infante, Vigliani, Aksoy
and Thorpe, would contain fatal deficiencies even if the
four studies were consistent in their results. But in fact the
results of the four studies are not consistent, and Dr. Wil-

* The foregoing facts appear at Pet. App. 102a-103a, 104a-105a,
107a. See also the material cited in our opening brief at 9-10, n. 4.

* An additional flaw in Dr. Wilson’s response ‘‘caleulation’’ is
the following: in utilizing the four studies Dr. Wilson made no
allowance whatsoever for the critical fact that susceptibility to
benzene leukemia varies among different individuals and groups.
See the reply brief, supra at p. 12, n. 15. Dr. Wilson thus simply dis-
regarded a consideration which introduces uncertainties of unknown
dimensions into any attempted risk assessment for benezene.

6c

CBB OT eh ns ee dee

son therefore found it necessary ‘‘to make sense out of this
highly discorded [sic] set of data.*’ App. 88. In order to
‘‘make sense’’ out of the data, Dr. Wilson chose to assume
that the Infante data had overstated the risk to the popula-
tion Dr. Infante had studied,’ while in reality, as we have
shown, the Infante data were almost certainly understated.
Dr. Wilson also assumed that the studies of Dr. Vigliani
were overstated. As a physicist who lacked any special
ability to determine the validity of any of the epidemiologic
studies in the record, Dr. Wilson did not reach his assess-
ment of the Infante and Vigliani data on the basis of any
expert evaluation of that data. Rather, he believed that the
results reported by Drs. Infante and Vigliani should be
discounted merely because they appeared to be inconsistent
with the studies of Dr. Aksoy and Dr, Thorpe—this despite
the fact that the Aksoy data, even according to Dr. Aksoy,
underestimate the risk, and that the Thorpe study con-
tains such manifest flaws that the Secretary properly re-
jected its findings. See p. 3c, n. 3, supra.

Dr. Wilson’s arbitrary and erroneous decision to discount
the results of the Infante and Vigliani studies led him to
base his calculations on an assumed dose-response curve,
which he described as ‘‘slightly better than a guess’’ (App.
92), and which predicts a risk much smaller than the risk
which he would have predicted had he not regarded the
Infante and Vigliani data as overstated.®

What is more, while Dr. Wilson made the peculiar assumption
that for each case of benzene leukemia there will be an additional
case of some other form of cancer, and he therefore multiplied his
dose-response projections by a factor of two (App. 171), he did not
take into account any of the known benzene-related risks other than
cancer, such as aplastic anemia. Any attempt to factor in these
additional risks would involve the same problems as are presented
with respect to assessing the leukemia risk. See, e.g., Pet. App.
82a-98a, 160a.

™See App. 171, 88, 92-93.

Sid. at 88.

®The extent to which Dr. Wilson discounted the Infante and
Vigliani data can be seen by examining Dr. Wilson’s graph (see
note 1, supra).

7¢

II.

The Secretary found that apart from inadequacies in the
data, there is no satisfactory theory of extrapolation from
known carcinogenic effects of a substance at high exposure
levels to projected carcinogenic effects of the substance at
lower exposure levels. Dr. Wilson chose to estimate the
risks of benzene exposure at low levels by drawing a linear
dose-response curve, viz. he assumed that the risks are
directly proportional to the level of exposure. See note
supra, But Dr. Wilson conceded that the actual dose-re-
sponse curve could well be very different from the linear
extrapolation he employed (Tr. 2719-2720), and that ‘‘any
dose-response curve is possible.’’ App. 93.

Dr. Wilson described the linear extrapolation technique
as a ‘‘fairly prudent and slightly conservative’’ assumption
(App. 89). But Dr. Lloyd suggested that at high levels of
benzene exposure, the number of diagnosed leukemias may
be reduced by the fact that many workers may die of other
benzene-induced diseases such as aplastic anemia before
the onset of leukemia, while at low levels this phenomenon

In constructing dose-response curves on his graph, Dr. Wilson
assumed that such curves would be linear, i.e., that risk is directly
proportional to exposure. Consequently, the dose-response curve
attributable to any given study under Dr. Wilson’s system would
consist of a line between the point which represents the results
of that study and the point on the vertical axis of the graph
which represents a normal incidence of leukemia (i.e., a risk factor
of one) at zero exposure. (On the graph, the latter point is where
the three lines marked ‘‘risk’’ intersect).

The dose-response curve which Dr. Wilson selected to predict
the risk of benzene leukemia is the line marked ‘‘1.75 x 10/yr.
@ 1 ppm”’ on the graph (App. 171). On that dose-response curve,
a risk factor of 7 would oceur at approximately 550 ppm. Yet on
his graph, Dr. Wilson recorded the Infante data as showing a risk
factor as high as 7 at exposures somewhere between 10 ppm and
100 ppm. Similarly, Dr. Wilson plotted the Vigliani data as show-
ing a risk factor of approximately 15 at exposures of approximately
300 ppm (and the risk factor actually reported by Dr. Vigliani
was 20, see Pet. App. 100a) ; yet the dose-response curve selected
by Dr. Wilson shows a risk factor of only 4 at that level of exposure.

8c

may not be present to the same degree. If that were the case,
a linear extrapolation would understate the leukemia risk
at low levels of exposure, and a curve rising steeply at low
doses and flattening out at the top would be accurate. Dr.
Lloyd stated, based on his experience, that similar cireum-
stances have been found to exist with respect to beryllium
and asbestos, and he asked Dr. Wilson whether the same
might not be true of benzene, ‘‘[plarticularly in light of
Dr. Aksoy’s report with so many of the shoemakers dying
from their blood dyscrasia [a non-malignant disorder].”’
App. 93. Dr. Wilson replied that he could not rule out that
possibility. Ibid.

The possibility discussed by Dr. Lloyd is only one of
many factors which indicate that the ‘‘linear hypothesis’’
may lead to an underestimate of the risk of low-level ben-
zene exposures. As a commentator has observed, ‘* [uJntil
we have a detailed understanding of the molecular mecha-
nisms of carcinogenesis, we have no basis for saying whether
the ‘linear hypothesis’ assures an adequate margin of
safety.’’ Lovins, Cost-Risk-Benefit Assessments in Energy
Policy, 45 Geo. Wash. L. Rev. 911, 922 (1977).

ITI.

Finally, even if a meaningful dose-response curve could
be constructed, an additional problem would remain. The
actual number of deaths which the standard would prevent
cannot be determined in the absence of information as to
the number of workers who, were the permissible exposure
level not reduced below 10 ppm, would be exposed at each
level between 10 ppm and 1 ppm. The data on this point
are very imprecise, and Dr. Wilson admitted that his caleu-
lations were based on an assumption that without the stand-
ard now under review, workers would be exposed to 3 ppm
of benzene. This assumption surely understates the extent
to which the standard would reduce exposures, and there-
fore leads to an underestimate of the health benefits which
would result from the standard. As Dr. Wilson conceded
(App. 90):

9c

‘*T tried to [determine] ... who is exposed to what.
..« but I didn't find it. * * * * So I made a very rough
guess that they were three parts per million.

Dr. Wilson further conceded that this ‘‘guess’’ was based
on at least one inaccurate assumption (/bid.) :

MS. KALARIS: I believe you stated that only a few
100 people would be regularly exposed to 10 ppm of
benzene concentration. Can you explain the basis of
that statement?

DR. WILSON: That is maybe incorrect. It came out
of a page in this report here, and again, they were
vague on this particular question. Therefore, I am a
little vague on it. Probably, as I was thinking about it
last night, I think a few hundreds; it is probably closer
to 1,000 than 100, but T really don’t know.’’

10¢

eae ands me oti"

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/brief%3Amicro_IA40385006_0035%3A15. Public record. Not legal advice.