Tobacco Documents Online

PHILIP MORRIS INCORPORATED 120 PARK AVE NU E. NEW YORK, N.Y. 10007 Jersox E Ltxcoux VICE PRE3IDENTi..$7RATEGIC RESEARCM June 9, 1986 Dear Dr. Jarvik: My reaction to the galley proofs you enclosed with your letter of May 30th is particularly to the first sentence in the summary paragraph and depends in part on the construction the reader can be expected to place on the word linking, in the context of the whole paragraph. As a matter of fact, I can cite data that seems to me to be discordant with any definition of "linking." Note the following: 1962 cigarette consumption (per capita age 15 and over, per annum) 1961-65 animal fat consumption (average grams per capita, per day) Argentina Venezuela 1577 2123 67.5 27.4 6.9 2.8 42.3 12.8 93.1 41.2 162.1 88.4 Lung Cancer Death Rate per 100,0001 Ages 35-44 45-54 55-64 65-74 11980 data for Venezuela, and simple average of 1979 and 1981 for Argentina.

Although the pair of countries I have selected presents one of the most striking contradictions, it is not a sport. Multiple regression analysis for a broad list of countries with lung cancer as the dependent variable and several dietary variables plus smoking, GNP, alcohol, etc. yields a significant statistical relationship with the dependent variable only for animal fat. I believe this was one of the topics I covered when I visited you in Cal ifornia . Note also that two investigators, Szepsenwol and Kroes, have produced lung cancer in experimental animals by high fat diets while attempts to produce it from cigarette smoke inhalation have been unsuccessful. There are other points: e.g. the failure of the "intervention" studies to indicate a mortality dis- advantage for cigarette smoking when smoking less versus smoking more was randomly determined by whether the respondent was selected into the test group or the control group. Also, there is Dr. Vaillant's analysis of the Grant study in which no statistically significant association of cigarette smoking with premature mortality was found when a summary psychosocial variable was in- cluded in the regression analysis. Also, as Dr. Myddleton recently pointed out, the detailed data from the largest prospective study of smoking and mortality does not show a consistant dose response relationship. The foregoing is an initial and personal reaction. If anything further develops, I will let you know. Sincerely, Murray E. Jarvik, M.D. Chief, Psychopharmacology Unit Veterans Administration Medical Center Wilshire and Sawtelle Boulevards N © Los An CA eles 90073 g , N ~ bcc: Messrs. G. Weissman W .Wr<!Ci2-1man> - A. Holtzman