Obesity, Mortality, and Life Years Lost Associated With Breast Cancer in Nonsmoking US Women, National Health Interview Survey, 1997–2000

Introduction The relationship between obesity and breast cancer has been extensively investigated. However, how obesity and breast cancer interplay to affect mortality and life expectancy of women in the United States has not been well studied. Methods We used data from the National Health Interview Survey, 1997–2000. Our sample included nonsmoking, nonpregnant women who reported a body mass index of at least 18.5 kg/m2 and no cancer other than breast cancer at the time of the survey. A survival model with Gamma frailty and Gompertz baseline was used to estimate relative risks of total mortality and project life years lost associated with breast cancer by obesity status and age. Results Breast cancer increased risk of mortality depending on degree of obesity and decreased life years by 1 to 12 years depending on race, age, and obesity status. Relative risks for death increased with degree of obesity. Obese women under age 50 across all racial groups were predicted to lose the most life years; racial groups other than whites and blacks lost the most life years (11.9 y), followed by whites (9.8 y) and blacks (9.2 y). Conclusion The number of life years lost associated with breast cancer was more marked for more obese than for less obese women and for women under age 50 and women aged 70 or older than for women aged 50 through 69. Public health initiatives should put more emphasis on the prevention and control of obesity for these target populations.

Medscape, LLC designates this Journalbased CME activity for a maximum of 1 AMA PRA Category 1 Credit(s)™. Physicians should claim only the credit commensurate with the extent of their participation in the activity.
All other clinicians completing this activity will be issued a certificate of participation. To participate in this journal CME activity: (1) review the learning objectives and author disclosures; (2) study the education content; (3) take the post-test with a 70% minimum passing score and complete the evaluation at www.medscape.org/journal/pcd ; (4) view/print certificate.
Release date: November 13, 2013; Expiration date: November 13, 2014 Learning Objectives Upon completion of this activity, participants will be able to: • Analyze the effect of overweight and obesity on the risk of mortality among women with breast cancer • Compare the effects of obesity on the risk of mortality among women with and without breast cancer • Assess the effect of age on the relationship between obesity and the risk of mortality among women with breast cancer • Assess the effect of race/ethnicity on the relationship between obesity and the risk of mortality among women with breast cancer

Introduction
Breast cancer is the most commonly diagnosed cancer in women and the second leading cause of death among women (1). Obesity is a known risk factor for postmenopausal breast cancer (2,3) but may be inversely related to premenopausal breast cancer incidence (4). Obesity is also associated with increased all-cause mortality (5)(6)(7). However, it is unclear how obesity and breast cancer interplay to affect mortality and life expectancy of adult women. Some studies have focused on obesity in relation to breast cancer incidence (8)(9)(10); others have explored the relationship between obesity and mortality from breast cancer (11)(12)(13). Few studies have examined life expectancy associated with breast cancer (14,15). To our knowledge, no studies have considered this association in relation to body mass index (BMI). Our study is the first to consider BMI in the relationship between breast cancer and mortality rate as well as life years lost associated with breast cancer. Our study used both life years lost and mortality rate, which are accepted outcome measures of population health (16,17).
The goal of this study is to investigate the mortality risks associated with breast cancer according to BMI level and race and to predict life years lost among women in the United States. We used data from a national probability sample of the US population and a class of survival model. The projection of life years is computed on the basis of mortality estimates and the observed characteristics of the sample as a snapshot of their life span. This approach is not restricted by the length of mortality follow-up and the observed mortality rates, and thus provides an alternative to the life table approach (14). We focused on female breast cancer; therefore, our sample included only women. The sample exclusion criteria were the following: 1) women with any missing data; 2) women who had smoked more than 100 cigarettes in their entire life, because analyses can be confounded by illnesses associated with smoking (6,7); 3) women pregnant at the time of survey, because BMI levels are unstable during pregnancy; 4) women who reported having had any type of cancer other than breast cancer, to exclude the possibility of breast cancer as a secondary cancer; and 5) underweight women, because disease-driven weight loss could perplex the analysis.

Data and the sample
Outcome variable and covariates 2 The outcome variable was age at death or censor (December 31, 2006 The additional covariates were dichotomized as follows: educational attainment (whether a person was a high school graduate); alcohol consumption (whether a person had had no more than 12 drinks of any type of alcoholic beverage in the person's entire life at time of survey); and physical activity (whether a person engaged in light/modest or vigorous physical activity for at least 10 minutes more than once per week).

Statistical analysis
A survival model with Gamma frailty and Gompertz baseline was used (20). Maximum pseudo-likelihood estimation was performed to adjust for complex sampling design in the NHIS data (21). The estimates were used to compute hazard ratios for death (relative risks, RRs) (22) by BMI and racial groups for people with breast cancer and without any cancer. Life years were predicted by a closed-form expression (23). The sample was divided into subsamples with different combinations of race (whites, blacks, and other), age (aged <50 y, 50-69 y, and ≥70 y), and obesity status (obese: class I-III; nonobese: normal weight and overweight). For each subsample, a sample was selected on the basis of sampling weights, and life years were predicted on the basis of characteristics of the subsample. Within the same subsample, the number of life years lost was projected by comparing the predicted life years of women with breast cancer to the predicted life years of the same women had they not had breast cancer. The bootstrap method was performed to resample the subpopulations 1,000 times to compute the means and standard errors (24).
All estimations and bootstraps were adjusted for the complex sampling design in the NHIS (25). Stata version 11 (Stata Corp LP, College Station, Texas) was used to obtain the summary statistics for the sample and the population, and Matlab version 7.13, R2012b (MathWorks Inc, Natick, Massachusetts) was used to perform estimations and predictions.

Results
The initial sample size of women in our data was 70,651. The remaining sample after applying the exclusion criteria consisted of 35,853 women, representing a US population estimated at 51,860,143 (95% confidence interval [CI], 50,944,982-52,775,304) nonsmoking adult women (aged ≥18 y) ( Figure 1). Of the population, 78% were white and 14% were black ( Table 1). The average BMI was 26.3 kg/m ( Table 2); 50% of the population had normal weight; 28% had a BMI within the overweight range; 14% belonged to class I obese; 5% belonged to class II obese; and 3% fell into class III obese (Table 1).  If breast cancer was present in obese women aged 50 or less, life years lost were 9.8 years (95% CI, 7.6-12.0) for whites, 9.2 (95% CI, 6.4-12.0) for blacks, and 11.9 (95% CI, 5.7-18.0) for all other racial groups. Fewer life years were lost for nonobese women aged 50 or less: 7.9 years (95% CI, 6.7-9.0) for whites, 5.3 (95% CI, 3.0-7.7) for blacks, and 6.0 (95% CI, 5.7-6.2) years for all other racial groups ( A similar trend was found among women aged 70 or older. For obese women aged 70 or older, the number of life years lost was 4.9 (95% CI, 3.9−5.9) for whites and 6.9 (95% CI, 6.6−7.2) for blacks. With the exception of racial groups other than whites and blacks, the number of life years lost was less for nonobese women aged 70 years or older: 3.7 (95% CI, 3.2−4.2) for whites, 4.5 (95% CI, 3.0−6.0) for blacks, and 5.8 (95% CI, 5.0−7.0) for all other racial groups.

Discussion
We used the following examples to translate life years lost associated with breast cancer to the proportion of total life years. An obese white woman under age 50 with breast cancer was expected to lose 9.8 years of life, which translates to a 12.1% reduction in the expected total life years (81.4 years) for a white woman aged 40 years (27). Similarly, an obese black woman under age 50 with breast cancer was expected to lose 9.2 years of life, which translates to an 11.9% reduction of the total life years (77.5 years) for a black woman aged 40 years (27).
No obvious racial disparity was found in life years lost associated with breast cancer between blacks and whites. Whites (obese and nonobese) lost more life years than their black counterparts for age groups of under 50 and 50 to 69, whereas black women aged 70 years or older lost more life years than white women of the same age range. A plausible explanation is that our data showed fewer breast cancer cases in blacks than in whites and, on average, black women have a higher prevalence of obesity and a shorter life expectancy than white women (1). Therefore, the extent to which breast cancer shortened life years was not as prominent as the extent to which breast cancer increased mortality risk in blacks. Wong et al (28) used a stochastic model to simulate life expectancy for blacks and whites. They found that, of all cancers, breast cancer had the biggest impact on the racial difference in life years lost, accounting for 0.15 years (life years lost for blacks minus that for whites), compared with 0.03 years for obese women and −0.05 years for nonobese women at all ages in our study. The smaller gap between blacks and whites in our study was because the differences were positive and wider for women aged 70 or older (1.94 years for obese women and 0.80 years for nonobese women), whereas the differences were negative for women aged less than 70.
Previous studies that analyzed breast cancer mortality and BMI found results comparable to those of our study: higher mortality among breast cancer patients with a higher degree of obesity. Reeves et al analyzed 1.2 million UK women aged 50 to 64 from 1996 to 2001 and examined mortality from breast cancer in relation to BMI (10). Like us, they found that increasing BMI was associated with a significant increase in the risk of breast cancer mortality. By using data from 1997, Ni Mhurchu et al examined postmenopausal breast cancer mortality in New Zealand and found that higher than optimal BMI (>21 kg/m ) contributed to postmenopausal breast cancer deaths (29); high BMI contributed to 4% of cancer deaths, though the proportion of these deaths that was specific to breast cancer is unknown (29). Our results with regard to mortality rate are consistent with those of Aragón et al, who found that premature mortality attributable to breast cancer among San Francisco residents was the highest among black women compared with other ethnic groups (Asians/Pacific Islanders, Latinos/Hispanics, and whites) (14). However, the predicted life years lost in their study (13.4 years) was generally higher than that in ours, because we focused on nonsmoking women with no other cancer reported at survey and controlled for several confounding variables.
Our study has several strengths. First, our analysis focused on breast cancer and examined its association with mortality and life years, which has not been well studied and deserves more investigation (14,15). Second, the major outcomes of our study are the basic metrics for population health and provide information about mortality and longevity (16). Third, to our knowledge, ours is the first study to incorporate BMI, age, and race information in the analysis of breast cancer, mortality, and life years lost. Fourth, our study advances methodology by controlling for unobserved heterogeneity, allowing for a more precise estimation and prediction of life years lost. This is important for 2 population studies, which lack information specific to breast cancer. Last, we used a national data set, and our conclusions are generalizable to the population of US women.
This study has several limitations, many of which are due to secondary data constraints. First, breast cancer status was dichotomized, preventing us from differentiating the extent of the disease progression or the effects of different types of breast cancers. Second, the data contained few breast cancer cases for women aged less than 40 and for women other than whites and blacks. The predicted life years lost for adult women under 50 and for racial groups other than whites and blacks should be interpreted with this caveat in mind. Third, BMI information was self-reported and recorded only at the time of survey, the latter of which restricted analyses of BMI and mortality at the diagnosis of breast cancer. Fourth, the data used were cross-sectional, and our model is not time-varying and does not capture the dynamics of disease evolution or weight change over time, limiting its projection capability on life years (30,31). Last, our data does not capture menopausal status -shown to differentiate the associations of BMI with breast cancer mortality risk (10) -but this limitation was compensated for by dividing the sample into 3 age groups (<50 y, 50−69 y, and ≥70 y). Future studies should focus on how the extent of breast cancer progression, different types of breast cancers, treatments, or menopausal status affect the prediction of life years lost associated with breast cancer.
Our study suggests that breast cancer increased the risk of early mortality depending on degree of obesity and decreased life years by 1 to 12 years depending on race, age, and obesity status. The number of life years lost associated with breast cancer was higher for more obese women aged less than 50 than for nonobese women of the same age. Moreover, obese women with breast cancer aged less than 50 and older than 70 bear a higher burden in terms of years of life lost. Public health initiatives should put more emphasis on the prevention and control of obesity for these target populations. Whether an individual engaged in light/modest or vigorous physical activity for at least 10 min more than once per week.

Post-Test Information
To obtain credit, you should first read the journal article. After reading the article, you should be able to answer the following, related, multiple-choice questions. To complete the questions (with a minimum 70% passing score) and earn continuing medical education (CME) credit, please go to http://www.medscape.org/journal/pcd . Credit cannot be obtained for tests completed on paper, although you may use the worksheet below to keep a record of your answers. You must be a registered user on Medscape.org. If you are not registered on Medscape.org, please click on the "Register" link on the right hand side of the website to register. Only one answer is correct for each question.
Once you successfully answer all post-test questions you will be able to view and/or print your certificate. For questions regarding the content of this activity, contact the accredited provider, CME@medscape.net. For technical assistance, contact CME@webmd.net. American Medical Association's Physician's Recognition Award (AMA PRA) credits are accepted in the US as evidence of participation in CME activities. For further information on this award, please refer to http://www.ama-assn.org/ama/pub/about-ama/awards/ama-physicians-recognition-award.page . The AMA has determined that physicians not licensed in the US who participate in this CME activity are eligible for AMA PRA Category 1 Credits™. Through agreements that the AMA has made with agencies in some countries, AMA PRA credit may be acceptable as evidence of participation in CME activities. If you are not licensed in the US, please 2 a a