|
Stil.n.i eI ,1.
<br />TABLE 3—Rlsk of Mortality Associated with Current Cigarette, Alcohol, and Mari)uana Use: Kaiser Permanente Medical
<br />Care Program Members (n = 65 171), Oakland and San Francisco, June 1979 through December 1985
<br />Current Marijuana Use,
<br />Current
<br />Consumption of Three or
<br />Cigarette Smokmga
<br />More Alcoholic Drinks per Day° At Least Once a Week Daily
<br />RR (95% CI)
<br />RR (95% Cl) RR (959, Cl) RR (959. CI)
<br />Men
<br />AIDS 1 64 (1 15, 2 34)
<br />0 94 (0.60, 1 47) 2 09 (1 42. 3 06) 1.65 (0 97, 2 82)
<br />Non -AIDS 1.76 (1.40, 2.20)
<br />1.21 (0 94, 1 56) 1.17 (0 91, 1 51) 1 31 (0.93, 1.84)
<br />Total mortality 1.75 (1 45, 2.11)
<br />1 13 (0.91, 1 40) 1.46 (1 19, 1 79) 1 43 (1.08, 1 90)
<br />Women. Total mortality 1.58 (1.25, 2 01)
<br />1.90 (1 31, 2 76) 1.23 (0.84, 1.80) 1 44 (0 80, 2.56)
<br />Note The model was adjusted loi age, race, education, mautal status obesity, cigarette smoking and alcohol use, RR = relative risk, CI =
<br />confidence Interval.
<br />-Relative to nonsmoking
<br />°Relative to occasional alcohol use
<br />°Relative to nonuser/experimental user status
<br />tory disease, injury or poisoning. "other
<br />causes" of mortality, and total non -AIDS
<br />mortality (inch only) (Table 2).
<br />A4arijuana List, in Relation
<br />m Morlah1v
<br />For men, ever use of marijuana was
<br />associated with a sigmficand increased
<br />nsk of local mortality (28%) and AIDS
<br />mortality (80%) and a nonsignificant (P>
<br />.05) increase (I I%) in risk of non -AIDS
<br />mortality. Relative risks associated with
<br />ever marijuana use for these mortality
<br />catcgones were similar or higher in
<br />nonsmokers/occasional drinkers (a group
<br />in which marijuana use could be evalu-
<br />ated without uncontrolled confounding by
<br />cigarette and substantial alcohol use). Of
<br />note was the nearly significant 47% in-
<br />crease in "other causes" of mortality,
<br />exanunatton of which revealed higher
<br />proportions of deaths from infectious
<br />diseases and from alcohol and drug abuse
<br />in ever users than in never users/
<br />expcnnlcnters. Current marijuana use was
<br />also associated wilt a significantly in-
<br />creased risk in men of total mortality
<br />(33%1) and AIDS mortality (9017().
<br />ht women, there were no significant
<br />increases or decreases in mortality risk
<br />associated with ever or current marijuana
<br />use. Current use was associated with a
<br />nearly significant 8617, increase in mortal-
<br />ityfrom injury or poisoning. which could
<br />not be attributed to any specific category
<br />of Injury
<br />Relative risks associated with man-
<br />juana use among nunsrnokers/occaslonal
<br />drinkers were generally similar to those
<br />for the complete cohort, suggesting that
<br />increased risks in the complete cohort
<br />were not an artifact resulting from incom-
<br />plete control of the effects of cigarette
<br />smoking or alcohol use. The results of an
<br />analysts of mortality excluding subjects
<br />who died wWtm the first 5 yeah of
<br />follow-up (data not shown) were similar
<br />to the overall results shown to Table 2.
<br />suggesting that the overall results were
<br />uncompromised by the possibility that
<br />serious illness occumng before the multi-
<br />phasic health checkup affected subjects'
<br />decision to use marijuana.
<br />Duration of use in current marijuana
<br />users was not consistently related to the
<br />risk of AIDS mortality in men or to total
<br />mortality to women, and had an inverse
<br />tendency in relationship to total and
<br />non -AIDS mortally 1n men (data not
<br />shown). A continuous duration -of -use
<br />variable was not significant when added
<br />to the full models for each mortality
<br />outcome.
<br />Nfanjuana use at Icast once a week
<br />was associated with slightly higher rela-
<br />tive risks of moilahty than less frequent
<br />use. The addition of frequency of use
<br />improved the fit of the model (P<.05)
<br />only for total mortality u1 men (RR = 1.25,
<br />9517, Cl = 0.97. I.62, for total mortality
<br />among those who used less than once a
<br />week and RR = 1.46, 957 Cl = 1.19.
<br />1.79, among Wosc whu used at (cast once
<br />a week, relative to nonuscrs/expennlcnt-
<br />ers).
<br />AIDS Morialin
<br />The vast nlajonty of AIDS deaths
<br />(1721207 = 83%) occurred among never
<br />manned inch. Current marijuana use was
<br />nearly twice as high in never manned asin
<br />manned rnen jlablc 1). raising the ques-
<br />tion of whether analytic control for
<br />mantal status was insufficient to adjust for
<br />confounding hfcstyle factors, particularly
<br />male homosexual behavior.
<br />To address dos question, the study
<br />cohort was linked to the Northern Califor-
<br />nia Kaiser Pennanente Medical Care
<br />Program AIDS Database, which revealed
<br />214 men with a diagnosis of AIDS after
<br />deiemtinanon of tier marijuana use
<br />status The prevalence of current mari-
<br />juana use at the unlc of the checkup (5(, i, )
<br />in these AIDS pauenLs was substanna0%
<br />higher than the prevalence to unmarried
<br />men in the total studs cohort (3871.
<br />For these _214 AIDS patients, current
<br />marijuana use was associated with a
<br />nonsignificant decrease to relative risk for
<br />total mortality IRK = 0.78. 95'n Cl =
<br />0.47, 1.30) and for AIDS mortality
<br />(RR = 0.71, 9511, Cl = 0.41, 123). As-
<br />suming that most of the unmanned nlcn
<br />who developed AIDS were homosexual
<br />or bisexual. these ftndmgs supponexl the
<br />hypoflbcsis that the presalence of man-
<br />juana use was higher in homosexual and
<br />bisexual men In the cuhon. it group at high
<br />risk for AIDS nurnehty. Therefore, male
<br />homosexual (chas Ior. a critical confound
<br />ing variable. could not he controlled lot In
<br />complete cohort nlomahn analyses.
<br />Comparative Ri.;O of lobarco,
<br />Alcohol, and Afar ijuana Use
<br />The relative risks of total monaluy In
<br />men and women. and of AIDS and
<br />non -AIDS mortality in men, associated
<br />with current cigarette smoking, consump-
<br />tion of three or more dnnks per day. and
<br />current marijuana use are shown in Table
<br />3 Except for AIDS mortality. the risks
<br />588 Amencan Journal of Public Health Apnl I9'/7. Vol 87. No 4
<br />
|