<note> sample="quota" bates="2060549643" isource="pm" decade="1990" class="ui" date="19970304" </note>

Philip Morris USA Confidential PHILIP MORRIS U.S.A. INTEROFFICE CORRESPONDENCE
Richmond, Virginia To: SA Worldwide Date: March 4, 1997 From: R. A.
Carchman Subject: 1997 Cancer/Swiss Study 

I am sending out for your review and comments my preliminary thoughts
on a recently published paper "Lung Carcinoma Trends by Histologic
Type in Vaud and Neuchatel, Switzerland, 1974-1994." Cancer 79(5):
906-914, 1997 by F. Levi, et al. In this paper, the authors utilized
cancer registries from Vaud and Neuchatel Switzerland to gather data
from 1974 - 1994 on lung cancer incidence by histologic type, age
and gender. The authors conclude that "adenocarcinoma is sustaining
a new lung carcinoma epidemic, chiefly attributable to the switch
to low tar, filtered cigarettes." Table 1 of this publication represents
the ~ 20-year period of observation for males and females by number
and percent by histologic type of lung tumor. Table 2 presents the
same data by quinquennium and age standardized by gender. Since the
focus of the authors is on adenocarcinoma, it is important to note
that in the "Materials and Methods" section they have combined all
adenocarcinomas including brochioloalveolar carcinoma. I raise this
for the following reasons: In a recent review (not cited in the cancer
paper) by Barkley and Green (J. Clin. Oncol 14: 2377-2386, 1996) and
what is considered to be a landmark study by Auerbach and Garfinkel
(also not cited, Cancer 68: 197301977, 1991) (same journal as the
Swiss study) relevant points are raised concerning changing patterns
of cancer with emphasis on adenocarcinoma and bronchioloalveolar carcinoma
(BAC). Both of the previous reports indicate that BAC (ie., peripheral)
is increasing in young nonsmoking females and that bronchiogenic (i.e.,
central) cancers in nonsmokers and former smokes showed a decreased
incidence. Furthermore these authors (Barkley and Green) state "BAC
appears to be responsible for most, if not all the increase in adenocarcinoma."
Auerbach and Garfinkel state that "adenocarcinomas are associated
less directly with tobacco use or any other carcinogens." In addition,
the similar rise reported in males and females by Levi, et. al. are
to be expected based on BAC alone and are inconsistent with smoking
as a cause (see Barkley and Green review). Increased risk for BAC
development is seen in the following occupations: construction, motor
freight, wood, paper mills, and sugar cane and farmers. Fibrosis or
sclerosis appears to be common predisposing features associated with
this adenocarcinoma. Viral etiology has also been postulated. Pulmonary
infections (e.g. TB) may lead to BAC and reports of drug induced damage
predisposing to BAC have also been reported (see Barkley and Green
review). Table 2 indicates that small cell carcinoma (i.e., central)
in females has increased to the same extent as adenocarcinoma. Though
the authors do not discuss this point, this observation is also inconsistent
with their hypothesis. Finally, the Levi et al. paper utilized a log-linear
Poisson model with arbitrary constraints on the parameters in order
to separate the effects of age, period of diagnosis and birth cohort
on incidence rates. The authors indicate that "given the arbitrariness
of the constraints imposed those models should be interpreted with
caution." Well, these are some of my initial comments. I look forward
to hearing your comments. 

cc: W. Reininghaus C. Ellis R. Cox H-J Haussmann