Tobacco smoke overload and ethnic, state, gender, and temporal cancer mortality disparities in Asian-Americans and Pacific Islander-Americans

doi:10.1016/j.ypmed.2005.12.015

Preventive Medicine

Volume 42, Issue 6, June 2006, Pages 430-434

https://doi.org/10.1016/j.ypmed.2005.12.015 Get rights and content

Abstract

Background.

Asians and Pacific Islanders (APIs) are important populations nationally and globally. So we assessed cumulative tobacco smoke overexposure (smoke overload)/cancer mortality associations across states, ethnicities, years, and genders among API-Americans.

Methods.

Death rates were adjusted to the 2000 United States age standard, lung cancer death rates used as a smoke overload bio-index, and lung/non-lung cancer death rate linear regressions run. Cancer death rate smoking-attributable fractions (SAFs) are equal to 1 − estimated unexposed rate/observed rate.

Results.

The two lowest smoke overload and non-lung cancer death rates were in South Asian (Indo)-Californian females and males. The highest were in Korean-Californian males. Non-lung cancer death rates were tightly and steeply associated with smoke overload across ethnicity, state, year, or gender. Cancer death rate smoking-attributable fractions ranged from 0 in female and 6% in male Indo-Californians, to 39% in female and 57% in male API-Americans in 2002, to 71% in Korean-Californian and 69% in API Hawaiian males.

Discussion.

Many API American cancer death rate disparities across genders, ethnicities, states, or years can be explained by smoke overload disparities. Tobacco control may greatly reduce cancer death rates and disparities among API-Americans and, likely, others.

Introduction

Asians and Pacific Islanders are a large, heterogeneous population whose exposure/cancer associations merit more study. Asians and Pacific Islanders represent over 60% of the global and 4% of the United States (US) populations (Ezzati et al., 2005, SEER⁎Stat Database, 2005). Cancer is the leading cause of death among API-Americans, despite their wide, unexplained all-sites cancer death rate disparities across gender, ethnicity, place, and time (SEER⁎Stat Database, 2005). For example, in 1997–2001, API-Californian all-sites cancer death age-adjusted rates (rates)/100,000 ranged nearly four-fold from 57.8 in South Asian (Indo)-Californian females to 203.4 in Korean-Californian males (Kwong, 2004).

Cumulative tobacco smoke overexposure (smoke overload) disparities can account for most temporal and regional cancer mortality variation in African-American males (Leistikow and Tsodikov, 2005), but have not been studied in Asian populations. So we assessed smoke overload/cancer death rate associations and cancer death smoking-attributable fractions (SAFs) in API-Americans stratified by time, place, ethnicity, and gender.

Section snippets

Methods

We used published 1980, 1990–2002 (Health, United States, 2004, 2005, SEER⁎Stat Database, 2005), and 1997–2001 lung and all-sites cancer death rates adjusted to the 2000 US age standard (Kwong, 2004, Kwong et al., 2005). Subtracting lung cancer rates from all-sites rates gave us “non-lung” rates.

Lung cancer death rates were used as a smoke overload bio-index (Leistikow and Tsodikov, 2005). Given many states' high API death rate standard errors, we limited state analyses to the 5 lowest lung

Results

The first, second, and third highest all-sites cancer death rates and SAFs studied were in 1997–2001 Korean-Californian, 1990–2002 API-Hawaiian, and 1997–2001 Japanese-Californian males, respectively (Table 1). The first, second, and third lowest all-sites cancer death rates and SAFs were in Indo-Californian females and males in 1997–2001, and female API-Americans in New Jersey in 1990–2002, respectively.

Lung and non-lung cancer death rates were tightly, steeply, and significantly associated

Discussion

The observed strong API lung/non-lung cancer death rate associations across diverse stratifications and death rates suggest that the same thing, most likely tobacco smoke overload, may cause most prematurely fatal cancers in API-American males, and many such female deaths. These API population-level associations, and the increasing frequency of lung and non-lung cancers in the same API individual (Liu et al., 2002), resemble the Kaposi's sarcoma/pneumocystosis associations seen in the early

Acknowledgment

Financial Support: University of California, Davis.

References (30)

P. Froom et al.
Healthy volunteer effect in industrial workers
J. Clin. Epidemiol.
(1999)
B.N. Leistikow et al.
Cancer death epidemics in United States black males: evaluating courses, causation, and cures
Prev. Med.
(2005)
Y.Y. Liu et al.
Multiple primary malignancies involving lung cancer-clinical characteristics and prognosis
Lung Cancer
(2002)
M.J. Thun et al.
Changes in mortality from smoking in two American Cancer Society prospective studies since 1959
Prev. Med.
(1997)
Adult Smoking
(2005)
J.D. Bailar et al.
Progress against cancer?
N. Engl. J. Med.
(1986)
Centers for Disease Control and Prevention, Office on Smoking and Health, 2002. Methodology: Estimating the health...
M. Cockburn et al.
R. Doll et al.
The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today
J. Natl. Cancer Inst.
(1981)
M. Ezzati et al.
Measuring the accumulated hazards of smoking: global and regional estimates for 2000
Tobacco Control
(2003)

M. Ezzati et al.

Role of smoking in global and regional cancer epidemiology: current patterns and data needs

Int. J. Cancer

(2005)

S. Goodman

Setting the Stage for Discussion: Attributable Risk in Epidemiology

(2004)

M.S. Gottlieb et al.

The acquired immunodeficiency syndrome

Ann. Intern. Med.

(1983)

E.R. Gritz et al.

Smoking, the missing drug interaction in clinical trials: ignoring the obvious

Cancer Epidemiol., Biomarkers Prev.

(2005)

Health, United States, 2004, 2005. Hyattsville, MD: National Center for Health Statistics, CDC,...

Cited by (24)

The association of harsh working environment and poor behavior habits with neck health
2023, International Journal of Industrial Ergonomics
To aid nurses in preventing cervical spondylosis and neck subhealth, this study intends to investigate the relationship between neck health and the two categories of factors (behavior habits and working environment).
A self-reported questionnaire combining the neck disability index and neck pain intensity measurement tool was used to collect data anonymously. Binary logistic regression was applied to explore the association of the neck health with the working environment and behavior habits.
A total of 2880 nurses from more than ten hospitals in Kunshan participated in the survey. The risk factors of neck health status included cumulative arm lift time >1 h per day (OR = 1.29), typical head posture for using a computer [tilted to the left or right (OR = 1.67), head-down (OR = 1.36)], prolonged use of electronic products (OR = 1.34), prolonged head-down work (OR = 1.69), frequent assistance in turning the patient (OR = 1.27), maintaining a fixed position (OR = 1.70), having overload work (OR = 1.73), working hours >40 h per week (OR = 1.22), sedentary (OR = 1.28), napping while taking transportation (OR = 1.29), staying up late (OR = 1.40), poor ventilation (OR = 1.67), high humidity of working environment (OR = 1.83). The protective factors were sleeping enough hours per day (OR = 0.65) and exercise >2 h per week (OR = 0.81).
Although the results suggest that nurses under 45 are particularly benefited, all nurses should pay more attention to their neck health. Policymakers and relevant authorities should pay attention to nurses' neck health, develop advanced tools through multidisciplinary cooperation, provide ergonomic education, remove adverse factors from the environment, and raise nurses' neck health awareness.
How many deaths are attributable to smoking in the United States? Comparison of methods for estimating smoking-attributable mortality when smoking prevalence changes
2011, Preventive Medicine
The number of smoking-attributable deaths is commonly estimated using current and former smoking prevalences or lung cancer mortality as an indirect metric of cumulative population smoking. Neither method accounts for differences in the timing with which relative risks (RRs) for different diseases change following smoking initiation and cessation. We aimed to develop a method to account for time-dependent RRs.
We used birth cohort lung cancer mortality and its change over time to characterize time-varying cumulative smoking exposure. We analyzed data from the American Cancer Society Cancer Prevention Study II to estimate RRs for disease-specific mortality associated with current and former smoking, and change in RRs over time after cessation.
When lung cancer was used to measure cumulative smoking exposure, 254,700 male and 227,000 female deaths were attributed to smoking in the US in 2005. A modified method in which RRs for different diseases decreased at different rates after cessation yielded similar but slightly lower estimates [251,900 (male) and 221,100 (female)]. The lowest estimates resulted from the method based on smoking prevalence [225,800 (male) and 163,700 (female)].
Although all methods estimated a large number of smoking attributable deaths, future efforts should account for temporal changes in smoking prevalence and in accumulation/reversibility of disease-specific risks.
Smoke-Free Policies Among Asian-American Women. Comparisons by Education Status
2009, American Journal of Preventive Medicine
Citation Excerpt :
SHS exposure may be responsible for the increased risk of lung cancer among Chinese-American and Filipino-American women, compared with the expected risk for their non-Latino white counterparts with a similar proportion of cigarette consumption.7 Asian-American subgroups have different cancer death rate smoking-attributable fractions, with Korean-American women significantly higher than their counterparts.8 Few studies have enough large diverse subgroups to assess the extent of SHS exposure for the Asian-American population.6
California has significantly decreased racial/ethnic and educational disparities in smoke-free home and indoor work policies. California's ethnic-specific surveys present an opportunity to disaggregate data and examine the impact of California's smoke-free social norm campaign for Asian-American women.
The California Tobacco Use Surveys for Chinese Americans and Korean Americans were conducted in 2003 and analyzed in 2008 to compare women with lower (≤ high school graduate) or higher education status for smoke-free policy adoption and enforcement.
Lower-educated and higher-educated women had similar proportions of smoke-free policies at home (58%) or indoor work (90%). However, lower-educated women were more likely than higher-educated women to report anyone ever smoking at home (OR=1.62, 95% CI=1.06, 2.48, p=0.03) and exposure during the past 2 weeks at an indoor workplace (OR=2.43, 95% CI= 1.30, 4.55, p=0.005), even after controlling for ethnicity, smoke-free policy, knowledge about the health consequences of secondhand smoke exposure, and acculturation. There was no interaction between education and knowledge about secondhand smoke health harms.
The intended consequences of California's tobacco-control efforts have resulted in similar rates of smoke-free policies at home and in indoor work environments among Asian-American women across educational levels. However, an unintended consequence of this success is a disparity in enforcement by educational status, with lower-educated Asian-American women reporting greater smoke exposure despite similar rates of knowledge about the health consequences of secondhand smoke exposure. Besides establishing policies, lower-educated Asian-American women may need to be empowered to assert and enforce their right to smoke-free environments.
Lung cancer as an index of tobacco exposure: association with non-lung cancer mortality
2009, Gaceta Sanitaria
Valorar el papel del tabaquismo, empleando como proxy la mortalidad por cáncer de pulmón, en la mortalidad por otros cánceres (excluyendo el de estómago).
Análisis de series temporales de mortalidad por cáncer en los hombres españoles (1970-2003) para valorar la posible asociación entre cáncer de pulmón y los cánceres «no pulmón-no estómago» (NPNE). Para evitar el efecto de posibles autocorrelaciones se aplicó la regresión Prais-Winsten.
Las tasas anuales de mortalidad por cánceres NPNE están linealmente relacionadas con la mortalidad por cáncer de pulmón en el período 1970-2003, con una pendiente de la recta de 1,07, intervalo de confianza del 95% de 0,98-1,17 y R² de 0,97.
Las variaciones de las tasas de mortalidad por cánceres NPNE pueden ser modeladas en función de los cambios en las tasas de mortalidad por cáncer de pulmón en el período estudiado. Los resultados presentados parecen mostrar una posible asociación entre el tabaquismo y los cánceres NPNE.
To assess the possible role of tobacco smoke in non-lung cancer (excluding stomach cancer) using changes in lung cancer mortality rates as a proxy for tobacco exposure.
A time series analysis of cancer mortality was performed to evaluate the possible association between changes in mortality rates for lung cancer and for non-lung, non-stomach cancer (NLNS) from 1970 to 2003 in Spanish males. To avoid problems with autocorrelation, Prais-Winsten regression was applied.
Changes in NLNS cancer death rates showed a parallel trend with lung cancer death rates in the study period, with an adjusted slope of 1.07, 95% CI of 0.98-1.17, and R² of 0.97.
Variation in NLNS cancer death rates can be accurately modelled as a function of changes in lung cancer death rates for the study period, suggesting a possible association between tobacco exposure and NLNS cancers.
Oral squamous cell carcinoma incidence by subsite among diverse racial and ethnic populations in California
2008, Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology and Endodontology
Citation Excerpt :
Accordingly, in our data, there were 17 palatal OSCC cases in Filipino females and just 1 case in males. Among Asians, the highest M:F ratio for OSCC was seen in Koreans, and this may well be attributable to the markedly high smoking prevalence among Korean males compared with females.43 In California, Korean men were found to have a smoking prevalence of 27.9%, which is 46% higher than general-population men in California (19.1%) and more than 5 times higher than Korean women (4.3%).44
The aim of this report was to examine the oral cancer incidence by sex, race/ethnicity, and anatomical subsite.
Data from the California Cancer Registry (CCR) were used to calculate the age-adjusted incidence rates of invasive oral squamous cell carcinoma (OSCC) by sex, race/ethnicity, and anatomical subsite among residents in California during 1988 to 2001.
Although non-Hispanic (NH) black men have the highest overall incidence rate for OSCC, NH whites and NH blacks have similar incidence patterns by subsite, but the male-to-female (M:F) rate ratio is higher among NH blacks. The OSCC incidence rates for Hispanics are much lower than those for NH whites and NH blacks and similar to those of Asians. The Asian ethnic groups display dramatic variations in terms of the subsite-specific incidence rates and M:F rate ratios.
The findings illustrate the heterogeneity and complexity of oral cancer by anatomical location and the importance of cultural habits and behavioral factors in the development of oral cancer.
Smoke load/cancer death rate associations in Korea females, 1985-2004
2007, Preventive Medicine
Citation Excerpt :
However, smoking status is a poor proxy for cumulative tobacco smoke damage (smoke load) (2002a,b) and appears to greatly underestimate contributions of smoking to premature cancer death (Leistikow et al., 2006; Leistikow and Tsodikov, 2005; Yoo et al., 2006). Smoke load/cancer mortality associations can explain most temporal cancer mortality variation in Asian-American, African-American, and Korea males (Leistikow et al., 2006; Leistikow and Tsodikov, 2005; Yoo et al., 2006), but not Asian-American females (Leistikow et al., 2006). The lack of temporal association in females may be due to chance given the low numbers of female cancer deaths studied.
Korea female death rates from many cancers have risen rapidly since 1985. The sources of those cancer death epidemics are unclear but may be related to rising cumulative tobacco smoke damage (smoke load). We assessed Korea female smoke load/cancer death rate associations from 1985 to 2004.
Lung cancer rates were used as a smoke load bio-index. Subtracting lung, stomach, and uterine corpus cancer death World age standard rates (rates) from all-sites rates gave us non-lung–stomach–uterine corpus (NLSUc) rates. Lung/NLSUc linear regressions were run, adjusted for autocorrelation. Estimated, lower, and upper bound smoking-attributable fractions (SAFs) were calculated using the formula SAF = 1 − {(unexposeds' cancer death rate) / (observed rate)}, based on the linear regression and respective best, upper, and lower bound estimated lung, stomach, and uterine cancer death rates in the unexposed.
Lung cancer death rates (smoke load) can explain 88% of the variance in NLSUc rates from 1985 to 2004 after adjusting for autocorrelation. The estimated Korea female all-sites cancer death rate SAF in 2004 was 43% (sensitivity range 29–56%).
Smoke load, probably from tobacco given the epidemic time course, may cause a large cancer death burden in Korea females despite their very low self-reported prevalence of smoking.

View all citing articles on Scopus

View full text

Tobacco smoke overload and ethnic, state, gender, and temporal cancer mortality disparities in Asian-Americans and Pacific Islander-Americans

Abstract

Background.

Methods.

Results.

Discussion.

Introduction

Section snippets

Methods

Results

Discussion

Acknowledgment

J. Clin. Epidemiol.

Prev. Med.

Lung Cancer

Prev. Med.

Adult Smoking

Progress against cancer?

N. Engl. J. Med.

The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today

J. Natl. Cancer Inst.

Measuring the accumulated hazards of smoking: global and regional estimates for 2000

Tobacco Control