close

Вход

Забыли?

вход по аккаунту

?

389

код для вставкиСкачать
1310
Breast Carcinoma Survival Analysis for African
American and White Women in an Equal-Access
Health Care System
Barbara E. Wojcik, Ph.D.
Martha K. Spinks, M.A., M.S.W.
Scott A. Optenberg, Dr. P.H.
BACKGROUND. This retrospective review of breast carcinoma cases in the Department of Defense (DoD) Central Tumor Registry evaluated differences in survival
patterns between African American and white women treated in U.S. military
Analysis Branch, Center for Healthcare Education and Studies, Army Medical Department
Center and School, Fort Sam Houston, Texas.
health care facilities. The study examined the effects of age, stage of cancer, tumor
size, grade, lymph node involvement, waiting time between diagnosis and first
treatment, marital status, military dependent status, alcohol usage, tobacco usage,
and family history of cancer.
METHODS. Researchers reviewed the tumor registry records of 6577 women (5879
whites and 698 African Americans) diagnosed with breast carcinoma. The patients,
ages 19–97 years, were diagnosed between 1975 and 1994. A hazard ratio (relative
risk of mortality) model compared African American and white patients, adjusting
for various combinations of covariates; impact of independent variables on the
risk of death; prognostic factors significantly associated with survival; disease free
and overall survival times; effects of ethnicity, stage, and age on survival; and
trends in stage at diagnosis. A P value (2-sided) of less than 0.05 was considered
statistically significant.
RESULTS. After adjustment for age, the risk of death was 1.45 (95% confidence
interval [CI], 1.20–1.76) times greater for African American women than for white
women. Adjustment for stage reduced the risk to 1.41 (95% CI, 1.16–1.70); further
adjustment for demographic variables and most clinical variables had no effect.
Presented at the 124th Annual Meeting of the
American Public Health Association, New York,
New York, November 19, 1996.
The authors are grateful to Ruth Byers, Staff
Officer, U.S. Army Medical Command, for providing many reports and substantial guidance
in the use and management of the Department
of Defense (DoD) Central Tumor Registry
(ACTUR).
Address for reprints: Barbara Wojcik, Ph.D.,
Center for Healthcare Education and Studies,
AMEDDC&S, 1608 Stanley Road, Building 2268,
Fort Sam Houston, TX 78234-6125.
The opinions expressed herein are those of the
authors and do not necessarily reflect those of
the Department of the Army or the Department
of Defense.
Received March 25, 1997; revisions received
July 1, 1997, and October 9, 1997; accepted
October 9, 1997.
Still, African American women treated in the military health care facilities had a
better survival rate than African American women represented in the Surveillance,
Epidemiology, and End Results (SEER) Program of the National Cancer Institute.
In our study, the 5-year risk of death, from any cause, was 1.37 for African American
women with breast carcinoma; in other words, the mortality rate for African American women was 24.77% compared with 18.08% for white women. In the latest
SEER data, the 5-year relative risk of death for African American women compared
with white women is 1.86. The mortality rate in SEER is 34.2% for African American
women and 18.4% for white women. The survival rate for white DoD beneficiaries
is comparable to that for white women in SEER.
CONCLUSIONS. These observations suggest that ready access to medical facilities
and the full complement of treatment options that are standard for all DoD patients
improve survival rates for African American women. However, a significant unexplained difference in survival still exists between African American and white military beneficiaries. Cancer 1998;82:1310–8. q 1998 American Cancer Society.
KEYWORDS: breast neoplasms, African Americans, whites, survival analysis, risk
factors, age factors.
P
rior studies of breast carcinoma conducted in the United States
have shown that African American women experience shorter sur-
q 1998 American Cancer Society
/ 7bbe$$0446
03-10-98 12:41:29
cana
W: Cancer
Breast Carcinoma Survival Analysis/Wojcik et al.
vival after diagnosis than their white counterparts.1 – 7
In 1994, Eley et al.3 reported that after controlling for
geographic site and age, the death rate was 2.2 times
higher for African American women than for white
women. However, after adjustment for stage, treatment, comorbidity, and pathologic and sociodemographic factors, the morbidity for African American
women in that study was reduced to 1.3 times that for
whites (95% confidence interval [CI], 1.0 – 1.8), and not
statistically significant. Garfinkel et al. found that African American women diagnosed with breast carcinoma had lower 1-year and 5-year relative survival
rates than white women at any age and stage category
at the time of diagnosis.8
Dayal et al.6 suggested that socioeconomic status
may explain differential survival rates. According to
Vernon et al.,1 delay in seeking treatment had no direct
effect on survival when ethnicity and other variables,
such as age, socioeconomic status, and stage of disease, were controlled.
Some alternative hypotheses focus on biologic explanations for the differences in the survival profiles
of African Americans and whites. One argument for
biologic differences is that the two groups may present
differential breast tumor characteristics that do not
respond identically to the same treatment.9 – 12 Some
investigators believe that younger women have a
worse prognosis than older ones due to biologic differences in breast carcinoma. They hypothesize that
differences in survival are not a reflection of more
advanced disease, but may be due to a different molecular structure that does not present significant histopathologic differences.13
Recent meta-analyses14 – 17 have found a higher relative risk of breast carcinoma of 3.3 (95% CI, 1.2 – 9.3)
for women who consume alcohol. Palmer and Rosenberg’s18 meta-analysis of published epidemiologic
studies sought relationships between smoking and
breast carcinoma risk. None of the numerous control
studies conducted previously18 – 21 established a conclusive link between cigarette smoking and higher risk
of breast carcinoma.
Several studies examined access to health care as
a possible predictor of survival.22,23 Optenberg et al.,23
studying survival rates of African American and white
military men with prostate carcinoma, concluded that
African American military men fared better than their
civilian cohorts because of equal access to the military
health care facilities.
We examined differences in survival between African American and white women whose breast carcinoma was diagnosed and treated in the U.S. military
equal-access medical care system. We hypothesize
that equal access to health care facilities, including the
/ 7bbe$$0446
03-10-98 12:41:29
cana
1311
opportunity to receive similar medications and methods of treatment within similar time frames, should
diminish differences in survival rates among racial
groups. An important control variable in this study is
the fact that cancer treatments are largely standardized in Department of Defense medical treatment
facilities. Virtually every patient receives adjuvant
therapy in conjunction with surgery.
MATERIALS AND METHODS
The individual medical records of all women diagnosed with breast carcinoma during the period 1975 –
1994 were retrieved from the Department of Defense
(DoD) Central Tumor Registry (ACTUR) data base.24
All subjects in the study were beneficiaries of the DoD
health care system, entitled to equal access to definitive medical care, either free of charge or at a nominal
fee, at military medical facilities.
Clinical Variables
Clinical variables selected for this study were date of
diagnosis, date of treatment, date of last contact, status
at last contact, type of cancer, stage of tumor, pathologic primary tumor size, number of positive axillary
lymph nodes, overall grade, and type of treatment.
Treatment options available to patients in the ACTUR
data base are surgery, radiotherapy, chemotherapy,
and hormone therapy. All treatments combined with
surgery (i.e., surgery/radiotherapy, surgery/hormone
therapy, and surgery/radiotherapy/hormone therapy)
are grouped into the surgery category. A combination
of radiation and hormone treatment is categorized as
radiotherapy. Hormone therapy and chemotherapy
are categorized separately if they are not combined
with any other methods of treatment. The ‘‘diagnostic
procedures’’ category includes women who underwent a biopsy or non-cancer-related surgery. The ‘‘unknown treatment’’ category includes women who receive no treatment or who may receive care outside
the military health care system and therefore have incomplete records in ACTUR. Since no specific agents,
dosages, or durations of treatment are annotated in
ACTUR, there is no differentiation by type of chemotherapy or hormone therapy in this study.
Demographic Variables
Demographic variables — age, ethnicity, marital status,
military dependent status, family history of disease,
and tobacco and alcohol usage — were included. (Data
on alcohol and tobacco usage and family history of
cancer are limited by gross categorization in ACTUR
as ‘‘yes,’’ ‘‘no,’’ or ‘‘unknown.’’ Family history refers
to history of any cancer in the family, rather than specifically a history of breast carcinoma.)
W: Cancer
1312
CANCER April 1, 1998 / Volume 82 / Number 7
Equal Access
Two variables were designated as primary indicators
of whether African Americans and whites receive equal
access to treatment in military institutions: 1) method
of treatment, and 2) waiting time between diagnosis
and treatment. Waiting time, disease free intervals,
and survival times were derived from ACTUR as measures of the effectiveness of treatment. Waiting time
was calculated as the difference between the date of
diagnosis and the date of first treatment. A disease
free interval was defined as the difference between the
date of first treatment and the date of first recurrence
of breast carcinoma. Survival time was calculated from
the time of treatment until death or the last follow-up
date annotated in the medical record.
Stage
Subjects were grouped by ethnicity and stage of disease to determine whether survival rates differed as a
result of the stage at which cancer was diagnosed.
The results were compared with literature describing
recent trends in breast carcinoma staging at the time
of diagnosis.25,26 If initial diagnoses frequently are
made during late stages of disease, two explanations
could be hypothesized: 1) physicians are less accurate
in diagnosing breast carcinoma in early stages, or 2)
patients are not seeking medical consultation during
the early stages of disease. If initial diagnoses in early
stages are increasing, further analysis is warranted to
determine what factors, e.g., mammograms, self exams, or public education and awareness, account for
this improvement. Unfortunately, the data base used
for this study does not contain such information.
Stages were assigned in accordance with the TNM
system for carcinoma of the breast, as recommended
by the American Joint Committee on Cancer in the
Manual for Staging of Cancer.27 To ensure the statistical validity of further analyses, patients were stratified
by stages of breast carcinoma: Stage 0 (in situ), Stage
I, Stage II (to include Stages IIA and IIB), Stage III (to
include Stages IIIA and IIIB), and Stage IV.
Recurrence
ACTUR is formatted to record up to five breast tumors
recurring over time. Our analysis is limited to analysis
of the first recurrence only. Such a format allows the
researcher to investigate whether any categories of patients were significantly more likely to suffer tumor
progression. The progression of the disease was measured by three variables: tumor size, the number of
positive axillary lymph nodes, and the recurrence of a
primary tumor.
/ 7bbe$$0446
03-10-98 12:41:29
cana
Statistical Analysis
Characteristics of African American and white women
were compared by Student’s t-test and chi-square
analysis. The bivariate relationships between ethnicity
and the independent variables helped in the evaluation of differences between groups at the time of diagnosis. Before multivariate techniques were utilized,
the validity of the proportional hazards assumption
was checked by means of examination of ln(0ln) survival curves. Assumptions about the stability of the
hazard ratio held with time. For each covariate variable that might influence survival, the Cox proportional hazards model was applied, which allowed assessment of ethnicity-adjusted hazard ratios for overall survival. Ratios then were compared with a chosen
baseline category. Several multiple regression models
were examined to find the most significant covariates
and their joint impact on survival.
RESULTS
Of the 6577 patients included in this analysis, 5879
(89.4% of the total) were white and 698 (10.6%) were
African American. The majority of women in this study
(nearly 90%) were civilian spouses of active-duty service members or military retirees. Active-duty or retired military women constituted 4.9% of the study
population. African American patients were more than
twice as likely to be active-duty military members as
whites (10.3% vs. 4.3%).
All DoD cancer patients receive follow-up in accordance with the American College of Surgeons’ recommendations of 3 months, then 6 months after treatment, and annually thereafter. Furthermore, the vast
majority of patients who receive treatment for any reason other than cancer, including dental care, receive
that treatment in a military medical facility, where records of all treatments are consolidated. As a result,
each time a patient is seen, for whatever purpose, that
date is entered in the tumor registry as the date of last
contact.
Although our study included diagnoses from the
period 1975 – 1994, the majority of women (96.0%)
were diagnosed between 1980 and 1994. Distribution
of selected characteristics by ethnicity (Table 1) revealed statistically significant differences between the
two groups at the time of diagnosis.
A Kaplan – Meier estimate of the survival curves
revealed a statistically significant (P Å 0.0001) difference in overall survival by ethnicity (Fig. 1). Life-table
survival for African American and white breast carcinoma patients for periods of 1 month, 1 year, 5 years,
and 10 years is summarized in Table 2. Estimated survival curves for each stage show no significant differences in survival between African Americans and
W: Cancer
Breast Carcinoma Survival Analysis/Wojcik et al.
1313
TABLE 1
Distributions of Study Variables for Breast Carcinoma Patients,
by Ethnicity
Characteristic
Mean patient age, yrs (no. of cases)
Patient age (yrs), % (no. of cases)
õ40
40–49
50–65
ú65
Mean wait time, days (no. of cases)
Mean tumor size, cm (no. of cases)
Stage, % (no. of cases)b
0
I
II
III
IV
Lymph nodes, % (no. of cases)
0
1–3
4–10
ú10
Unknown
Family history of cancer, % (no. of cases)
Yes
No
Unknown
Tobacco usage, % (no. of cases)
Yes
No
Unknown
Alcohol usage, % (no. of cases)
Yes
No
Unknown
Status, % (no. of cases)
Military member
Spouse/child
Mother/other
Marital status, % (no. of cases)
Single
Married
Separated
Divorced
Widowed
Unknown
African
American
White
P valuea
50.0 (698)
57.1 (5879)
õ0.001
22.8 (159)
23.9 (167)
41.0 (286)
12.3 (86)
21.2 (698)
2.5 (571)
9.2 (538)
19.4 (1140)
45.2 (2658)
26.3 (1543)
29.5 (5879)
2.3 (4889)
8.2 (57)
37.7 (263)
39.0 (272)
9.9 (69)
5.3 (37)
9.3 (548)
46.6 (2740)
32.6 (1917)
6.0 (355)
5.4 (319)
õ0.001
52.7 (368)
18.6 (130)
9.6 (67)
5.2 (36)
13.9 (97)
60.3 (3543)
14.7 (864)
7.3 (428)
4.2 (248)
13.5 (796)
0.001
25.5 (178)
39.4 (275)
35.1 (245)
32.3 (1900)
31.3 (1839)
36.4 (2140)
õ0.001
25.6 (179)
40.8 (285)
33.5 (234)
31.7 (1862)
36.7 (2157)
31.6 (1860)
0.004
25.5 (178)
38.0 (265)
36.5 (255)
31.4 (1846)
31.9 (1876)
36.7 (2157)
0.001
10.3 (72)
84.5 (590)
5.2 (36)
4.29 (252)
91.60 (5385)
4.12 (242)
õ0.001
3.6 (25)
73.6 (514)
0.6 (4)
2.2 (15)
9.7 (68)
10.3 (72)
2.4 (141)
74.4 (4374)
0.2 (14)
1.4 (80)
12.0 (704)
9.6 (566)
õ0.001
0.002
0.013
FIGURE 1. Long term Kaplan–Meier survival is shown for African American and white breast carcinoma patients (P ° 0.001).
vival of breast carcinoma patients. The unadjusted estimate of ethnicity specific risk ratio for African Americans was 1.41 (95% CI, 1.17 – 1.70).
0.039
a
All P values are 2-sided. For mean age, tumor size, and wait time, P values were derived from
Student’s t tests. All other P values were derived from chi-square tests.
b
Stage was assigned according to the TNM system for breast carcinoma.27
whites except for Stage II, mainly because of large variations in data and a paucity of cases with long survival
times.
Covariates
Table 3 gives values of ethnicity-adjusted hazard ratios
indicating the effect of selected covariates on the sur-
/ 7bbe$$0446
03-10-98 12:41:29
cana
Equal Access
The two variables designated as primary indicators of
equal access to treatment in military institutions —
method of treatment and waiting time between diagnosis and treatment — showed that African American
and white women receive equal access and have the
opportunity for equal treatment in military medical
facilities. In fact, waiting time from diagnosis to treatment was significantly shorter for African American
women, who waited an average of 21.15 days, whereas
white women averaged 29.51 days (P Å 0.0024) (Table
1). All women who waited for treatment longer than
6 months after diagnosis (or received no treatment at
all) had a risk of death 2.88 times greater (95% CI, 2.05 –
4.05) than women who received treatment within 1
month of diagnosis (Table 3). Most patients in the
ACTUR data base received treatment within 2 weeks
of diagnosis.
African American and white patients received
comparable types of medical treatment. The most
common treatment for both groups at all stages was
surgery, with African Americans being treated somewhat more often with surgery than their white counterparts (93.8% African Americans compared with
90.3% whites). African Americans were treated more
often with hormones (0.7% African Americans compared with 0.4% whites), whereas whites were more
likely to receive radiotherapy (1.9% of African Americans compared with 2.3% whites), chemotherapy
(0.6% African Americans compared with 1.3% whites),
or only diagnostic procedures (2.6% African Americans
compared with 4.6% whites). Due to low input cell
frequencies, differences in stage-adjusted treatment
between African Americans and whites could not be
determined by chi-square analysis.
W: Cancer
1314
CANCER April 1, 1998 / Volume 82 / Number 7
TABLE 2
Life-Table Survival of Breast Carcinoma Patients in the ACTUR
Data Base
% Survival (standard error)
Period after
treatment
Overall
1 mo
1 yr
5 yrs
10 yrs
Stage 0
1 mo
1 yr
5 yrs
10 yrs
Stage I
1 mo
1 yr
5 yrs
10 yrs
Stage II
1 mo
1 yr
5 yrs
10 yrs
Stage III
1 mo
1 yr
5 yrs
10 yrs
Stage IV
1 mo
1 yr
5 yrs
10 yrs
African American
White
99.3 (0.32)
96.8 (0.70)
75.2 (2.31)
58.5 (3.92)
99.3 (0.11)
97.0 (0.24)
81.9 (0.68)
66.5 (1.19)
100.0 (0.00)
97.6 (2.35)
97.6 (2.35)
48.8 (34.53)b
100.0 (0.00)
99.5 (0.35)
93.0 (1.72)
80.7 (4.51)
100.0 (0.00)
99.2 (0.59)
92.2 (2.39)
76.6 (6.46)
99.8 (0.09)
99.1 (0.20)
91.1 (0.78)
77.4 (1.76)
100.0 (0.00)
98.0 (0.89)
71.9 (3.83)
59.3 (5.50)
99.7 (0.12)
98.4 (0.30)
81.5 (1.18)
64.9 (1.95)
100.0 (0.00)
92.8 (3.50)
52.1 (8.18)
32.9 (8.61)
98.6 (0.63)
93.9 (1.32)
56.9 (3.46)
43.8 (4.05)
89.2 (5.10)
77.3 (7.09)
15.1 (7.56)
7.6 (6.54)
92.4 (1.50)
71.0 (2.65)
29.5 (3.10)
12.8 (2.97)
Log Rank
P valuea
0.0001
ns
ns
0.017
ns
Grade
Histologic grade was a strong indicator of survival,
with hazard ratios for poorly differentiated and undifferentiated cases of 3.71 (95% CI, 2.58 – 5.33) and 6.71
(95% CI, 3.31 – 13.63), respectively, compared with
well-differentiated cases (Table 3). Moderately differentiated cases had a 1.75 hazard ratio (95% CI, 1.17 –
2.61).
ns
ACUTR: Department of Defense Central Tumor Registry; ns: not significant.
a
All P values are 2-sided.
b
At 10 years, only 2 patients were alive; the death of 1 created a 48% survival rate.
Stage and Age
In this study, stage and age were the major risk factors
for breast carcinoma. A regression model using only
age, stage, and ethnicity as independent variables produced the same ethnicity-adjusted risk ratio of dying
from cancer as a full model containing the whole spectrum of demographic and clinical variables.
Tumor stage appeared to be one of the strongest
predictors of death from all causes (Table 3). Women
with Stage III tumors had a risk of death 6.55 times
greater (95% CI, 4.40 – 9.77) than women with Stage 0
tumors. For women with Stage IV tumors, this risk was
elevated to a ratio of 18.60 (95% CI, 12.27 – 26.59).
African American women’s tumors were significantly larger (2.53 cm compared with 2.28 cm, P Å
0.013) when first diagnosed (Table 1). More African
/ 7bbe$$0446
American patients were diagnosed at Stages II and
III — 48.86% African Americans compared with 38.65%
whites.
Our study found younger women (age õ40 years)
at a higher risk of dying than patients ages 40 – 49 years
or patients ages 50 – 65 years (Table 3). The hazard
ratio for women ages 40 – 49 years was 0.74 (95% CI,
0.59 – 0.93) when compared with the baseline category
of women younger than 40 years. For women ages 50 –
65 years, the hazard ratio was 0.80 (95% CI, 0.65 – 0.99).
Women older than 65 years had a worse prognosis
than any other age group, with a hazard ratio of 1.22
(95% CI, 0.98 – 1.53) compared with the baseline category of women younger than 40 years.
African American women were diagnosed at a significantly younger age than whites (at a mean age of
50.0 years, compared with 57.1 years for whites) (P
õ 0.001) (Table 1). At initial diagnosis, 22.78% of the
African American women in this study were younger
than 40 years, compared with 9.15% of white women
(Table 1). Between ages 50 – 65 years, however, whites
began to catch up with African Americans; and after
age 65 years, the percentage of white women diagnosed with breast carcinoma was more than double
that of African American women.
03-10-98 12:41:29
cana
Lymph Node Status
At the time of diagnosis, 52.72% of African American
patients were lymph node negative, compared with
60.27% whites (Table 1). The risk of death for patients
with 1 – 3 axillary lymph nodes was 1.62 times greater
(95% CI, 1.35 – 1.95) than for lymph node negative patients (Table 3). With 4 – 10 lymph nodes, the ethnicityadjusted hazard ratio increased to 2.60 (95% CI, 2.11 –
3.19). The occurrence of more than 10 positive axillary
lymph nodes was a powerful factor, with an ethnicityadjusted hazard ratio of 5.57 (95% CI, 4.54 – 6.83) when
compared with lymph node negative patients.
Alcohol and Tobacco
ACTUR categorizes patients simply as users, nonusers,
or status unknown, without reference to frequency or
history of alcohol and tobacco use. More whites than
African Americans reported alcohol and tobacco use
(Table 1). The relative risk for patients who indicated
W: Cancer
Breast Carcinoma Survival Analysis/Wojcik et al.
1315
TABLE 3
Proportional Hazards Regression Model: Ethnicity-Adjusted Hazard Ratios of Selected Covariates for Overall Survival
% (no.)
of patients
Characteristics
Ethnicity
White
African American
Age (yrs)
õ40
40–49
50–65
ú65
Grade
Well differentiated
Moderately differentiated
Poorly differentiated
Undifferentiated
Unknown
Lymph node negative
Positive lymph nodes (no.)
1–3
4–10
ú10
Unknown
Tumor size (cm)
õ1
1–2
3–5
ú5
Unknown
Tumor stagea
0
I
II
III
Hazard ratio
(95% CI)
89.4 (5879)
10.6 (692)
1.00 (Baseline)
1.41 (1.17–1.70)
10.6 (697)
19.9 (1307)
44.8 (2944)
24.8 (1629)
1.00 (Baseline)
0.74 (0.59–0.93)
0.80 (0.65–0.99)
1.22 (0.98–1.53)
8.5 (559)
19.5 (954)
17.5 (1150)
0.4 (25)
59.1 (3889)
59.5 (3911)
1.00 (Baseline)
1.75 (1.17–2.61)
3.71 (2.58–5.33)
6.71 (3.31–13.63)
2.28 (1.62–3.22)
1.00 (Baseline)
15.1 (994)
1.5 (495)
4.3 (894)
13.6 (893)
1.62 (1.35–1.95)
2.60 (2.11–3.19)
5.57 (4.54–6.83)
3.17 (2.70–3.72)
24.0 (1578)
28.1 (1848)
25.0 (1647)
5.9 (387)
17.0 (1117)
1.00 (Baseline)
0.98 (0.80–1.20)
1.65 (1.37–1.98)
1.46 (1.23–1.70)
3.62 (2.88–4.54)
9.2 (605)
45.7 (3003)
33.3 (2189)
6.4 (424)
1.00 (Baseline)
1.33 (0.90–1.95)
2.68 (1.84–3.91)
6.55 (4.40–9.77)
Characteristics
Tumor stagea
IV
Wait time from diagnosis to treatment (mos)
°1
ú1–6
ú6
Tobacco usage
No
Yes
Unknown
Alcohol usage
No
Yes
Unknown
Family history
No
Yes
Unknown
Marital status
Married
Single
Separated
Divorced
Widowed
Unknown
Dependent status
Service member, active or retired
Spouse or child
Mother/other
% (no.)
of patients
Hazard ratio
(95% CI)
5.4 (356)
18.60 (12.27–26.59)
94.3 (6316)
3.9 (258)
1.7 (115)
1.00 (Baseline)
1.36 (0.99–1.87)
2.88 (2.05–4.05)
37.1 (2442)
31.8 (2094)
31.0 (2041)
1.00 (Baseline)
1.49 (1.28–1.74)
1.18 (1.01–1.37)
32.6 (2141)
36.7 (2412)
30.8 (2024)
1.00 (Baseline)
0.92 (0.79–1.06)
0.97 (0.82–1.14)
32.1 (2114)
36.3 (2385)
31.6 (2078)
1.00 (Baseline)
1.06 (0.90–1.25)
1.05 (0.91–1.22)
74.3 (4888)
2.5 (166)
0.3 (18)
1.4 (95)
11.7 (772)
9.7 (638)
1.00 (Baseline)
1.37 (0.84–1.96)
2.00 (0.64–6.22)
1.35 (0.81–2.24)
1.52 (1.27–1.81)
1.26 (1.06–1.48)
4.9 (328)
90.9 (5975)
4.2 (278)
1.00 (Baseline)
0.98 (0.70–1.37)
3.13 (2.16–4.54)
CI: confidence interval.
a
Stage was assigned according to the TNM system for breast carcinoma.27
that they drank some alcohol was lower than for those
who denied any alcohol use, at 0.92 (95% CI, 0.79 –
1.06) (Table 3). Among patients whose alcohol usage
was unknown, the risk was 0.97 (95% CI, 0.82 – 1.14).
Smokers were at a significantly higher risk, with an
ethnicity-adjusted hazard ratio of 1.49 (95% CI, 1.28 –
1.74) when compared with women in the study who
denied smoking. The relative risk for patients with unknown status compared with nonsmokers was 1.18
(95% CI, 1.01 – 1.37).
Family History
More whites than African Americans reported a family
history of cancer (Table 1). (The ACTUR data base does
not specify the type of cancer or the relationship of
the patient to family members with a cancer history).
Risk of cancer differed little between women who reported any relatives with cancer (1.06; 95% CI, 0.90 –
/ 7bbe$$0446
03-10-98 12:41:29
cana
1.25) and women with an unknown family history
(1.05; 95% CI, 0.91 – 1.22) (Table 3).
Marital Status
Women in ACTUR who were separated, widowed, single, or divorced had poorer survival rates than women
who were married (Table 3). The relative risk estimate
of breast carcinoma for never-married women was
1.37 (95% CI, 0.84 – 1.96), and for divorcees 1.35 (95%
CI, 0.81 – 2.24). Only for widows was the relative risk
significant, at 1.52 (95% CI, 1.27 – 1.81).
Recurrence
No statistically significant difference in tumor recurrence was found between African Americans (12.18%)
and whites (10.80%) (Table 4). Among African American and white patients with tumor progression, no
significant difference was found in tumor size, disease
W: Cancer
1316
CANCER April 1, 1998 / Volume 82 / Number 7
TABLE 4
Recurrence Characteristics by Ethnicity
Characteristic
Mean age, yrs (no. of patients)
Mean tumor size, cm (no. of tumors)
Recurrence, % (no. of cases)
Disease free interval, yrs (no. of cases)
°1
ú1–2
ú2–10
ú10
Prior stage, % (no. of cases)c
0
I
II
III
IV
Prior treatment, % (no. of cases)
Surgery
Radiotherapy
Chemotherapy
Hormone therapy
Other/no treatment
Crude mortality, % (no. of cases)
African
American
White
P valuea
45.9 (85)
3.1 (62)
12.2 (85)
54.3 (635)
3.2 (511)
10.8 (635)
0.0001
ns
ns
32.9 (28)
35.3 (30)
25.9 (22)
5.9 (5)
27.7 (176)
38.0 (241)
31.2 (198)
3.2 (20)
3.5 (3)
29.4 (25)
44.7 (38)
12.9 (11)
9.4 (8)
2.4 (15)
29.1 (185)
43.0 (273)
10.7 (68)
14.8 (94)
94.1 (80)
3.5 (3)
1.18 (1)
1.18 (1)
0.0 (0)
49.4 (42)
89.0 (565)
3.6 (230)
0.5 (140)
2.2 (3)
4.7 (30)
50.6 (321)
ns
ns
DISCUSSION
nsb
ns
ns: not significant.
a
All P values are 2-sided. For age and tumor size, P values were derived from Student’s t tests. All
other P values were derived from chi-square tests.
b
More than 20% of the cells have expected frequencies of less than 5. Chi-square may not be a valid test.
c
Stage was assigned according to the TNM system for breast carcinoma.27
free interval, prior stage, prior treatment, or crude
mortality. However, African Americans with recurrent
cancers were significantly younger than whites. The
mean age for African Americans was 45.9 years; for
whites, 54.3 years (P Å 0.0001).
Multiple Models
Three prognostic variables were determined to be crucial after several Cox proportional hazards survival
models were examined. Those three variables — ethnicity, age, and stage — produced a hazard ratio for
African American women of 1.41 (95% CI, 1.16 – 1.70).
After controlling for each major category of covariates — waiting time, tumor pathology, alcohol and tobacco usage, family history, marital status, and dependent status — the hazard ratio for death from all causes
remained unchanged (1.41; 95% CI, 1.16 – 1.71). Adding
treatment to the model changed the risk of death to
1.48 (95% CI, 1.22 – 1.80).
Diagnostic Trends
By 1994, the percentage of white women in ACTUR
who were diagnosed with Stage 0 (in situ) breast carci-
/ 7bbe$$0446
03-10-98 12:41:29
noma had grown to 12.02% — more than double the
percentage of in situ diagnoses in 1981 (Table 5). During the same period, the percentage of African Americans diagnosed in situ went from 0.00% to 11.45%,
approaching the percentage of whites diagnosed at
that stage. As the percentage of early-stage diagnoses
increased, there was a corresponding decrease in the
percentage of diagnoses in Stages II, III, and IV for
both African Americans and whites (Table 5). Nonetheless, the percentage of African American women
diagnosed at the early Stages 0 and I was lower than
for whites, and the percentage of African Americans
diagnosed at the advanced Stages II, III, and IV remained higher.
cana
African American women treated in military hospitals
fared better than African American women nationally,
whose 5-year odds of survival are reported in the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute.28 In the population examined in this study, the 5-year risk of death,
from all causes, was 1.37 for African American women
with breast carcinoma compared with that for white
women; in other words, the mortality rate among African Americans was 24.77% compared with 18.08% for
whites. This value is much better than the 5-year relative risk of death of 1.86 for African American women
compared with white women in the latest SEER, but
the survival rate for white DoD beneficiaries is comparable to that of white women in SEER. The mortality
rate in SEER is 34.2% for African Americans and 18.4%
for whites.28 These calculations suggest that ready access to medical facilities and the full complement of
treatment options improves survival for African American women. However, a significant difference in survival rates still exists between African American and
white military beneficiaries.
ACTUR contains a variable for any family member
with any cancer: no specificity regarding type of cancer
or type of familial relationship is available. This variable shows no relationship between breast carcinoma
and a generalized family history of cancer; however,
previous studies29 – 32 have found a link between breast
carcinoma and a family history of breast carcinoma.
Anderson and Badzioch31 have found the relative risk
associated with a family history of breast carcinoma to
be inversely related to age at initial breast carcinoma
diagnosis. Given that women in our study who were
diagnosed before age 40 years fared worse than any
other group except those older than 65 years, and
given that African American women were diagnosed
at a significantly younger age, a more careful analysis
of family relationships and breast carcinoma is war-
W: Cancer
Breast Carcinoma Survival Analysis/Wojcik et al.
1317
TABLE 5
Percentage of Breast Carcinoma Diagnoses by Stage and Ethnicity, ACTUR, 1975–1994
Stage
Ethnicity
Yrs
0
I
II
III
IV
African American
õ1981
1981–85
1986–90
ú1990
õ1981
1981–85
1986–90
ú1990
6.3
0.0
6.5
11.5
5.0
5.0
8.4
12.0
43.8
29.7
39.9
37.4
41.2
39.0
47.8
48.7
31.3
41.9
42.0
36.1
41.6
40.2
31.9
29.7
6.3
18.9
8.7
9.0
5.7
6.9
6.6
5.6
12.5
9.5
2.9
6.0
6.5
8.8
5.2
4.4
White
P valuea
0.004
0.001
ACTUR: Department of Defense Central Tumor Registry.
a
Two-sided P values were derived from chi-square tests.
ranted. Some studies have concluded that the relative
risk associated with having an affected mother31,32 is
lower than the risk associated with having an affected
sister. The authors have proceeded with a follow-up
study to examine possible links between breast carcinoma and a family history of breast carcinoma.
Most earlier studies of smoking and breast carcinoma have found either a very small positive association or no association.18 – 21 In our study, the significance of smoking to breast carcinoma was much
stronger.
Previous studies of marital status by Kelsey and
Horn-Ross33 set the relative risk for never-married
women between 1.1 and 1.8. Our analysis produced
an estimate of 1.37 (95% CI, 0.84 – 1.96).
We can speculate on several socioeconomic factors not in our data bases that should be considered
in future analyses of military populations. Due to recruiting standards and self-screening, military personnel share similarities in education, life-style, and socioeconomic status that may affect differences in survival rates among different racial groups in the
military. Besides health care entitlements, military
members and their families benefit from equal wage
and housing policies and common access to community education and support. The educational levels,
resources, and life-styles of African American and
white military personnel and their families are more
comparable than in the general population. The military emphasis on fitness and health for service members may carry over to health-conscious behaviors
among their spouses, while creating a more homogeneous study population.
A positive diagnostic trend was observed, with an
increase in early stage diagnoses and a corresponding
decrease in later stage diagnoses for both African
Americans and whites. It is unclear whether the grow-
/ 7bbe$$0446
03-10-98 12:41:29
cana
ing number of early diagnoses are due to improved
methods of diagnosis available to practitioners (such
as mammograms), an increase in breast self-examinations and health awareness, or an increased frequency
of women seeking medical care; but the diagnostic
pattern indicates that one or more of these positive
trends is in effect. The fact that the increase in early
stage diagnoses for African Americans grew from 0%
to over 11% (Table 5), and the increase was double
the increase for whites in the same period of time,
is an encouraging sign that more African American
women are seeking medical care sooner. Even so, African American military beneficiaries are at a significantly higher risk of dying from breast carcinoma than
their white cohorts. Researchers have yet to account
for the reasons why African American women, on average, develop breast carcinoma at a significantly
younger age than whites, and why all women diagnosed with cancer before age 40 years have poorer
survival rates than women ages 40 – 65 years. Ready
access to definitive medical care, then, is a necessary
but not sufficient condition for equalizing survival
rates between African American and white women.
Policymakers have solid evidence in these findings
of the positive impact of equal access to medical care.
However, researchers need to consider other explanations, such as differences in tumor biology, socioeconomic status, or sociocultural factors, that may contribute to diagnoses of African American women at
younger ages and at later, higher-risk stages of the
disease.
REFERENCES
1.
Vernon SW, Tilley BC, Neale AV, Steinfeldt L. Ethnicity, survival, and delay in seeking treatment for symptoms of breast
cancer. Cancer 1985;55:1563–71.
W: Cancer
1318
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
CANCER April 1, 1998 / Volume 82 / Number 7
Muss HB, Hunter CP, Wesley M, Correa P, Chen VW,
Greenberg RS, et al. Treatment plans for black and white
women with Stage II node-positive breast cancer: the National Cancer Institute Black/White Cancer Survival Study
experience. Cancer 1992;70:2460–7.
Eley JW, Hill HA, Chen VW, Austin DF, Wesley MN, Muss
HB, et al. Racial differences in survival from breast cancer:
results of the National Cancer Institute Black/White Cancer
Survival Study. JAMA 1994;272:947–54.
Kimmick G, Muss HB, Case LD, Stanley V. A comparison of
treatment outcomes for black patients and white patients
with metastatic breast cancer: the Piedmont Oncology Association experience. Cancer 1991;67:2850–4.
Bain RP, Greenberg RS, Whitaker JP. Racial differences in
survival of women with breast cancer. J Chronic Dis 1986;
39:631–42.
Dayal HH, Power RN, Chiu C. Race and socio-economic status
in survival from breast cancer. J Chronic Dis 1982;35:675–83.
Elledge RM, Clark GM, Chamness GC, Osborne CK. Tumor
biologic factors and breast cancer prognosis among white,
Hispanic, and black women in the United States. J Natl Cancer Inst 1994;86:705–12.
Garfinkel L, Poindexter CE, Silverberg E. Cancer in black
Americans. CA Cancer J Clin 1980;30:39–44.
Crowe JP Jr., Gordon NH, Hubay CA, Pearson OH, Marshall
JS, McGuire WL, et al. The interaction of estrogen receptor
and race in predicting prognosis for stage I breast cancer
patients. Surgery 1986;100:599–605.
Stanford JL, Szklo M, Boring CC, Brinton LA, Diamond EA,
Greenberg RS, et al. A case–control study of breast cancer
stratified by estrogen receptor status. Am J Epidemiol 1987;
125:184–94.
Stanford JL, Greenberg RS. Breast cancer incidence in young
women by estrogen receptor status and race. Am J Public
Health 1989;79:71–3.
Mohla S, Sampson CC, Khan T, Enterline JP, Leffall L Jr.,
White JE. Estrogen and progesterone receptors in breast
cancer in black Americans: correlation of receptor data with
tumor differentiation. Cancer 1982;50:552–9.
Chung M, Chang HR, Bland KI, Wanebo HJ. Younger women
with breast carcinoma have a poorer prognosis than older
women. Cancer 1996;77:97–103.
Longnecker MP, Berlin JA, Orza MJ, Chalmers TC. A metaanalysis of alcohol consumption in relation to risk of breast
cancer. JAMA 1988;260:652–6.
Longnecker MP, Newcomb PA, Mittendorf R, Greenberg ER,
Clapp RW, Bogdan G, et al. Risk of breast cancer in relation
to past and recent alcohol consumption [SER abstract]. Am
J Epidemiol 1992;136:1001.
Harvey EB, Schairer C, Brinton LA, Hoover RN, Fraumeni JF
Jr. Alcohol consumption and breast cancer. J Natl Cancer
Inst 1987;78:657–61.
/ 7bbe$$0446
03-10-98 12:41:29
cana
17. Hiatt RA, Klatsky AL, Armstrong MA. Alcohol consumption
and the risk of breast cancer in a prepaid health plan. Cancer
Res 1988;48:2284–7.
18. Palmer JR, Rosenberg L. Cigarette smoking and the risk of
breast cancer. Epidemiol Rev 1993;15:145–56.
19. O’Connell DL, Hulka BS, Chambless LE, Wilkinson WE, Deubner DC. Cigarette smoking, alcohol consumption, and
breast cancer risk. J Natl Cancer Inst 1987;78:229–34.
20. Chu SY, Stroup NE, Wingo PA, Lee NC, Peterson HB, Gwinn
ML. Cigarette smoking and the risk of breast cancer. Am J
Epidemiol 1990;131:244–53.
21. Palmer JR, Rosenberg L, Clarke EA, Stolley PD, Warshauer
ME, Zauber AG, et al. Breast cancer and cigarette smoking:
a hypothesis. Am J Epidemiol 1991;134:1–13.
22. Hunter CP, Redmond CK, Chen VW, Austin DF, Greenberg
RS, Correa P, et al. Breast cancer: factors associated with
stage at diagnosis in black and white women. J Natl Cancer
Inst 1993;85:1129–37.
23. Optenberg SA, Thompson IM, Friedrichs P, Wojcik B, Stein
CR, Kramer. Race, treatment, and long-term survival from
prostate cancer in an equal-access medical care delivery
system. JAMA 1995;274:1599–605.
24. Defense Enrollment Eligibility Reporting System (DEERS)
automated central tumor registry data dictionary. Washington, DC: U.S. Department of Defense, 1993.
25. Chevarley F, White E. Recent trends in breast cancer mortality among white and black U.S. women. Am J Public Health
1997;87:775–81.
26. Kerlikowske K, Grady D, Barclay J, Sickles EA, Ernster V.
Effect of age, breast density, and family history on the sensitivity of first screening mammography. JAMA 1996;276:
33–8.
27. Beahrs OH, Myers MH. Manual for staging of cancer. 2nd
edition. American Joint Committee on Cancer. Philadelphia:
J. B. Lippincott, 1983.
28. Ries LA, Miller BA, Hankey BF, Kosary CL, Harrad A, Edwards BK. SEER cancer statistics review, 1973 – 1991: tables
and graphs. Bethesda, MD: National Cancer Institute,
1994.
29. Gammon MD, John EM. Recent etiologic hypotheses concerning breast cancer. Epidemiol Rev 1993;15:163–8.
30. Horn PL, Thompson WD. Risk of contralateral breast cancer:
associations with factors related to initial breast cancer. Am
J Epidemiol 1988;128:309–23.
31. Anderson DE, Badzioch MD. Bilaterality in familial breast
cancer patients. Cancer 1985;56:2092–8.
32. Bernstein JL, Thompson WD, Risch N, Holford TR. The genetic epidemiology of second primary breast cancer. Am J
Epidemiol 1992;136:937–48.
33. Kelsey JL, Horn-Ross PL. Breast cancer: magnitude of the
problem and descriptive epidemiology. Epidemiol Rev 1993;
15:7–16.
W: Cancer
Документ
Категория
Без категории
Просмотров
4
Размер файла
100 Кб
Теги
389
1/--страниц
Пожаловаться на содержимое документа