[CANCER RESEARCH 43, 6077-6079,
December 1983]
Hepatitis, Alcohol Consumption, Cigarette Smoking, and Hepatocellular
Carcinoma in Los Angeles1
Mimi C. Yu,2 Thomas Mack, Rosemarie Hanisch, Robert L. Peters, Brian E. Henderson, and Malcolm C. Pike
Departments of Preventive Medicine and Pathology, University of Southern California School of Medicine, Los Angeles, California 90033
the sequence who met our matching criteria. We secured a control for
78 patients; 9 were blacks, 5 were Spanish-surnamed whites, and 64
ABSTRACT
Seventy-eight black patients and white patients with primary
hepatocellular carcinoma (PHC), 70 years or younger at diagno
sis, and 78 age-, sex-, and race-matched neighborhood controls
were interviewed. Information sought included usual dietary and
drinking habits, cigarette smoking habits, prior medical condi
tions including a history of hepatitis, prior exposure to blood
products, and occupational history. Cigarette smoking was a
risk factor for PHC; the relative risk (RR) for current smokers of
more than one pack/day compared to nonsmokers was 2.6.
Alcohol consumption was also a significant risk factor for PHC;
individuals who drank 80 g or more of ethanol per day had a RR
of 4.2 compared to those drinking less than 10 g/day. In addition,
a history of hepatitis (RR = 13.0) and a history of blood trans
fusions (RR = 7.0) were significant risk factors for PHC. Each of
these factors remained significant after adjustment was made
for the others.
INTRODUCTION
As part of a large multisite case-control study that we are
conducting in Los Angeles County to investigate a wide range
of risk factors, we have interviewed 78 PHC3 cases and matched
controls. This paper reports our findings from this PHC question
naire study.
MATERIALS AND METHODS
We interviewed 92 black patients and white patients with PHC, whom
we identified from the files of the Los Angeles County Cancer Surveillance
Program (3); Asian patients were excluded from this study. The patients
were incident cases who were diagnosed in the 5 years from 1975 to
1979 and who were 70 years of age or younger at the time of diagnosis.
There were 11 black patients, and 14 of the remaining 81 white patients
had Spanish surnames. One of the authors, Dr. Robert L. Peters,
reviewed pathology slides from 65 of these 92 PHC patients and judged
that all were compatible with a diagnosis of PHC.
For each of the 92 patients, we sought to interview a control who was
matched to the patient on sex, birth date (within 5 years), race (black,
Spanish-surnamed white, and other white), and neighborhood of resi
dence at diagnosis. To search for the "neighborhood" control, we fol
lowed a procedure that defines a sequence of houses on specified
neighborhood blocks. We attempted to identify the sex, age, and race
of all inhabitants of each housing unit; not-at-home units were revisited
to complete the census. Our goal was to interview the first resident in
1Supported by Grants CA 19171, CA 19496, CA 14089, and CA 17054 from
the National Cancer Institute of the NIH.
2 To whom requests for reprints should be addressed, at Department of Preven
tive Medicine, University of Southern California School of Medicine, 2025 Zonal
Avenue, Los Angeles, Calif. 90033.
3 The abbreviations used are: PHC, primary hepatocellular carcinoma; RR,
relative risk; HBSAG, hepatitis B surface antigen; HBV, hepatitis B virus.
Received April 11,1983; accepted August 31,1983.
DECEMBER
1983
were other whites.
All interviews were conducted by R. Hanisch. She interviewed 20
patients in person; the remaining 72 patients were dead when she first
contacted the families. For the dead patients, she conducted the inter
view through a close family member; in 40 instances, she interviewed
the spouse of the dead patient.
The questionnaire requested information on occupational history,
usual dietary and drinking habits, cigarette smoking habits, prior medical
conditions including a history of hepatitis, exposure to blood products,
use of drugs, and family history of certain diseases.
Questionnaire data from the 78 patients and their matched controls
were analyzed by standard matched-pair methods (2). Pairs in which
either the case or the control failed to answer the relevant question were
eliminated from the analysis. All statistical significance levels quoted (p
values) are one-sided.
RESULTS
There were 50 male patients and 28 female patients. The
mean age at diagnosis of the cases was 55.7 years. The mean
age of the controls (at date of diagnosis of the index case) was
55.1 years. Cases and controls were similar with respect to
religion, marital status, and education.
Our questions on cigarette smoking habits first asked whether
the individual had ever smoked at least 100 cigarettes up to the
date of diagnosis (for controls, date of diagnosis of the index
case). For those who had smoked 100 or more cigarettes, we
asked the age that they started smoking, the amount that they
were smoking at the time of diagnosis and, for exsmokers, the
year that they stopped smoking and the amount that they were
smoking just before they stopped.
Table 1 presents the relative risks of PHC by various smoker
categories. A nonsmoker was defined as one who either had
never smoked 100 cigarettes or a smoker who stopped 10 or
more years prior to diagnosis of PHC. We grouped exsmokers
who stopped 10 or more years ago with nonsmokers to avoid
bias that could arise from proxy respondents misclassifying such
exsmokers as nonsmokers. All other smokers who stopped prior
to diagnosis were classified as exsmokers. Cigarette smoking
was positively associated with PHC [p (linear trend) = 0.03]. The
RR for current heavy (>1 pack/day) smokers compared to
nonsmokers was 2.6 (95% confidence limits = 1.0, 6.7).
Our questions on alcohol consumption referred to usual (recent
past but before diagnosis with PHC) drinking habits. We asked
the subject about intake frequency and amount for each of the
4 kinds of alcoholic beverages (beer, sweet wine, table wine,
and hard liquor). We calculated the average daily ethanol intake
for each subject by assuming that one fluid oz of beer, wine, and
hard liquor contains 1.1, 2.9, and 9.4 g of ethanol, respectively
(1). Alcohol consumption was positively associated with PHC
(see Table 1). Individuals who drank 80 g or more of ethanol per
6077
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 1983 American Association for Cancer Research.
M. C. Yuétal.
Table 1
Alcohol consumption and cigarette smoking among cases and controls
Matched
useCigarette
Cigarette/alcohol
smokingNonsmokerExsmokerCurrent3«1
pack/day>1
pack/dayAlcohol
consumption0-9
g/day10-79
g/day80+
g/dayAlcohol
Table 2
History ot hepatitis or blood transfusions among cases and controls
95% conti95% confiFactor*
interval(0.3,(0.6,(1.0,(0.4,(1.3,(0.6,(0.7,(0.1,(1.7,4.0)2.5)6.7)1.9)13.8)3.4)5.0)4.6)113.9)
Cases
Controls
Matched RR dence interval
History of hepatitis
Yes
No
14
60
2
72
13.0
(2.2,272.4)
Received blood transfu
sion6
Yes
16
4
7.0
No
40
52
" Total pairs do not add to 78 due to missing values.
0 Only 62 case-control pairs were asked this question.
(1.7,
43.1}
study, the relative risk in current heavy smokers was 1.5. In an
earlier publication, the relative risk of HBSAG positivity for this
group of PHC patients had been reported to be 10.4 (6). These
relative risks suggest that cigarette smoking and HBV infection
pack/day>1
may behave more like additive than multiplicative risk factors for
pack/day80+
PHC. Their results were later confirmed by a case-control study
g/dayNon/exsmokerCurrent«1
conducted among Hong Kong Chinese (4). The Hong Kong study
reported an odds ratio of 3.3 in current smokers (at least 1 pack/
pack/day>1
day) compared to non- and exsmokers among HBSAG-negative
pack/dayCases32516232832183214125211Controls398171231434471411031RR1.01.11.22.61.00.9421.01.41.8CO0.814.0dence
•
We eliminated from the analysis 2 pairs in which the amount smoked by the
individuals; the risk ratio was 1.2 among HBSAG-positive individ
patient was not known.
uals. For this group of Chinese patients, the RR of PHC associ
ated with HBSAG positivity was 21.3. These risk estimates are
day had a RR of 4.2 compared to light (0 to 9 g/day) drinkers again suggestive of cigarette smoking and HBV infection acting
(95% confidence limits = 1.3,13.8).
as additive risk factors for PHC. Cigarette smoking could account
Table 1 also shows the relative risks of PHC by cigarette for a large proportion of the non-HBV related cases in both
smoking and alcohol consumption simultaneously. Exsmokers studies. In the Greek study, 78% of HBSAG-negative patients
were grouped together with nonsmokers due to their small were current smokers, one-half of whom smoked more than 1
numbers and their similarity in risk of PHC. Cigarette smoking pack/day; among control patients, 53% were current smokers
and alcohol consumption were both significantly related to PHC with 25% of those smokers smoking more than 1 pack/day. In
Hong Kong, 84% of HBSAG-negative patients were current
after allowing for the effect of the other (2).
A history of hepatitis was a third significant risk factor for PHC smokers, and almost all of them smoked 1 pack or more/day;
(RR = 13.0; see Table 2). A history of blood transfusion (RR = among HBSAG-negative controls, 65% were current smokers,
7.0; see Table 2) was also significantly associated with PHC. and two-thirds of the smokers smoked 1 pack or more/day.
Only 3 patients reported both a history of hepatitis and blood
Our present findings on cigarette smoking are consistent with
transfusion. At least 21 of the 27 patients who had a history of those of the 2 previous studies (4, 5). Our reported risks repre
hepatitis and/or blood transfusion had had the last episode more sent the overall effect of cigarette smoking in both HBSAGpositive and HBSAG-negative individuals. These estimated risks
than 10 years before the diagnosis of PHC.
Multivariate analysis of the joint effect of cigarette smoking, are very similar to those computed from the Greek data. Com
alcohol consumption, and a history of hepatitis and/or blood bining their HBSAG-negative PHC patients (50% of all cases)
transfusion showed that each risk factor remained significant with their HBSAG-positive patients, we obtain RRs of 1.8 and
after adjustment was made for the others. There were only 9 3.1 for current 1- to 20-cigarettes/day smokers and current 21+
black patients and 9 black matched controls in our study. All risk cigarettes/day smokers, respectively. Yarrish ef al. (7) recently
factors reported above show an excess of black cases and white reported that one-half of a group of PHC patients from the
cases separately in the heavy-exposure categories, strongly Philadelphia area (all but one were either black or white) were
suggesting that these risk factors apply to both black and white either HBSAG-positive or hepatitis B core antigen antibodypositive in the absence of hepatitis B surface antigen antibody
PHC patients.
so that, even though we did not test our patients for serological
There was no significant difference between cases and con
trols in their dietary habits, use of drugs, or family history of markers of HBV infection, there is some evidence that about the
same proportion of PHC patients are HBV-related in the United
cancer.
States as in Greece.
Our data show excessive alcohol intake to be strongly asso
DISCUSSION
ciated with PHC independent either of dietary and cigarette
Cigarette smoking as an independent risk factor for PHC was smoking patterns or of a history of hepatitis and/or blood trans
reported first in 1980 by Trichopoulos ef al. (5) in a study fusion. In our group of current smokers of more than 1 pack/
conducted in Greece. They reported that, among individuals who day, heavy (80 g of ethanol per day) drinkers had a 7.8-fold
were negative for HBSAG, current heavy (more than 1 pack/day) (14.0/1.8) risk of PHC compared to more moderate drinkers with
smokers had a 5.5-fold increased risk compared to non- plus similar smoking habits.
exsmokers. Among HBSAG-positive individuals in the same
HBV infection is uncommon among blacks and whites in Los
smokingstatus0-79
consumption by
g/dayNon/exsmokerCurrent«1
6078
CANCER
RESEARCH
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 1983 American Association for Cancer Research.
VOL. 43
Hepatitis, Alcohol, Cigarettes, and Liver Cancer
Angeles County, but our data suggest that it is still a very
significant risk factor for PHC even in this low-risk population. If
we consider a history of blood transfusions as a surrogate
variable for possible transmission of hepatitis viruses via contam
inated blood products, then 29% of our patients were infected
by HBV compared to 7% of the matched controls.
Although population-attributable
risks cannot be calculated
without possibility of bias from matched case-control studies (2),
some idea of the magnitude can still be thereby attained. From
the data in Tables 1 and 2, we estimate that cigarette smoking,
heavy alcohol consumption, and a history of hepatitis and/or
blood transfusion can account for 20,15, and 43%, respectively,
of all PHC cases in blacks and whites occurring in Los Angeles.
ACKNOWLEDGMENTS
We are grateful to Neeltje M. Mack for her technical assistance, and to Beth
Woodin for preparation of this manuscript.
DECEMBER
REFERENCES
1. Adams. D. F. Nutritive value of American foods in common units. USDA
Agriculture Handbook 456. Washington, D. C.: United States Government
Printing Office, 1975.
2. Brestow, N. E., and Day, N. E. Statistical methods in cancer research: the
analysis of case-control studies. IARC Sci. Pubi., 32:73, 248-279,1980.
3. Hisserich, J. C., Martin, S. P., and Henderson, B. E. An areawide cancer
reporting network. Public Health Rep., 90:15-17,1975.
4. Lam, K. C., Yu, M. C., Leung, J. W. C., and Henderson, B. E. Hepatitis B virus
and cigarette smoking: risk factors for hepatocellular carcinoma in Hong Kong.
Cancer Res., 42: 5246-5248,1982.
5. Trichopoulos, D., MacMahon, B., Sparros, L, and Merikas, G. Smoking and
hepatitis B-negative primary hepatocellular carcinoma. J. Nati. Cancer Inst.,
65:111-114,1980.
6. Trichopoulos, 0., Tabor, E., Gerety, R. J., Xirouchaki, E., Sparros, L., Muñoz,
N., and Linsen. C. A. Hepatitis B and primary hepatocellular carcinoma in a
European population. Lancet, 2:1217-1219,1978.
7. Yarrish, R. L., Werner, B. G., and Blumberg, B. S. Association of hepatitis B
virus infection with hepatocellular carcinoma in American patients. Int. J.
Cancer, 26: 711-715,1980.
1983
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 1983 American Association for Cancer Research.
6079
Hepatitis, Alcohol Consumption, Cigarette Smoking, and
Hepatocellular Carcinoma in Los Angeles
Mimi C. Yu, Thomas Mack, Rosemarie Hanisch, et al.
Cancer Res 1983;43:6077-6079.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/43/12_Part_1/6077
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 1983 American Association for Cancer Research.