Lung Cancer Incidence and Mortality Review Shows No Ethnic Differences in Women

Special Reports, NSCLC (Issue 10), Volume 10, Issue 10

There are no differences in lung cancer incidence or mortality among women of different ethnic groups when adjusted for age, smoking, and other factors, according to an article recently published in the Journal of Clinical Oncology.

Manali Patel, MD, MPH

There are no differences in lung cancer incidence or mortality among women of different ethnic groups when adjusted for age, smoking, and other factors, according to an article recently published in theJournal of Clinical Oncology.The findings, based on a large-scale review of data from the Women’s Health Initiative (WHI), differ from some other studies that have found increased lung cancer risk among non-Hispanic white women.

Unadjusted data from the cohort of 129,951 women found that women from other ethnic groups (except for American Indian or native Alaskan) had decreased incidence of lung cancer compared with non-Hispanic white women. The finding was statistically significant (P<.001), said the article.

After adjustment for age, smoking, education, calcium, vitamin D, body mass index, alcohol, family history, oral contraceptives, hormones, physical activity, and diet, the disparities between ethnic groups were no longer statistically significant.

“Interventions focused on these factors may reduce racial/ethnic differences in lung cancer incidence and mortality,” the article observed. Among both men and women, lung cancer is the primary cause of cancer death, and the second most common cancer in the United States, according to the Centers for Disease Control and Prevention.

Researchers found that smoking was highest in white women, who also had the highest rate of smoking cessation. Black women had the lowest smoking cessation rate, the researchers found.

No Significant Differences in a Fully-Adjusted Model

“There was a decreased risk of mortality among Hispanic women, but it attenuated toward the null after adjustment,” said Manali Patel, MD, MPH, instructor in the Division of Oncology, Stanford University School of Medicine and lead author on the paper, in an interview withTargeted Oncology.

“The hazard ratio moved from really statistically significant to nonstatistically significant. After adjustment, we did not find any decreased or increased incidence with patients from different racial and ethnic groups, compared with non-Hispanic whites,” she explained.

In unadjusted models, Hispanic and Asian/Pacific Islander women had a considerably reduced risk of lung cancer death (odds ratio [OR] 0.30) compared with white women (OR 0.34), the article said.

Like most of the other groups, non-Hispanic black women also had a reduced incidence of lung cancer compared with non-Hispanic white women (OR 0.75). When adjusted for age, however, there were no differences in incidence for black women compared with white women.

But even in a partially-adjusted model, Hispanic women continued to have reduced incidence of, and risk of death from, lung cancer, the study found. After adjustment for smoking, Hispanic women had reduced incidence (OR 0.47) and mortality (OR 0.43) compared with white women (P<.001), the study found.

In the fully-adjusted model, Hispanic women continued to have reduced incidence (OR 0.59) and mortality (OR 0.52) compared with white women, but it was not deemed statistically significant (P=.26 andP=.22, respectively).

The &lsquo;Hispanic Paradox&rsquo;

“We previously discussed this Hispanic paradox among women in California, which has been well-described for almost a decade,” said Patel.

“Despite having lower socioeconomic status, they tend to have better outcomes for non-small cell lung cancer (NSCLC). We also found that relevant factors included recent immigration status, nativity, and living in neighborhoods (which could be considered a proxy for social support),” she explained.

While the improved outcomes for Hispanic women were attenuated after full adjustment, the decreased lung cancer incidence nonetheless warrants additional investigation, according to the article.

The original WHI study enrolled 161,809 women from 40 clinical centers between October 1993 and December 1998, and included both observational study and clinical trial arms. Reviewing the data, Patel et al excluded those participants with missing outcomes or covariates from their analysis, leaving them with a total cohort of 129,951 women. Of the cohort, 83% were white, 8% black, 4% Hispanic, 3% Asian/Pacific Islander, and 0.4% were American Indian/Alaskan Native, according to the article.

Postmenopausal women between 50 and 79 years of age were eligible. Participants self-reported characteristics such as smoking, use of hormones or birth control agents, and other factors; cancer incidence was confirmed by medical review of health care records. Follow-up continued through 2009.

In the cohort used by Patel et al, 1044 women developed lung cancer, and 613 women died.

As a secondary endpoint, the researchers looked for a relationship between ethnicity and lung cancer subtype. Black women were more likely to have NSCLC compared with white women (OR 1.24), and less likely to have small cell lung cancer (OR 0.65), while Hispanic women were less likely to have small cell lung cancer (OR 0.57). Additionally, among other findings, American Indian/Alaska Native women were considerably more likely to have small cell lung cancer (OR 2.52). However, none of the findings reached statistical significance (P= .44), the article said.

Researchers also conducted a statistical analysis to test the hypothesis that race-specific cancer risk may be modified by smoking intensity, but found no statistically significant cross-associations.

Weaknesses of the Study

Researchers had to exclude a large number of participants in the study from their analysis because of missing data. “A good rule of epidemiologic studies is that you can’t include or evaluate those with missing covariates or factors,” explained Patel.

Another weakness in the WHI study was that participants were volunteers who self-selected. Thus, it is possible that the participants enjoyed increased health status and education, the article said. The study also noted that the WHI data showed no differences in tumor characteristics, while previous studies showed cancer stage and histology at the time of diagnosis to be differential among black women, the article said.

Additionally, the design of WHI was different from the use to which the data were put. “Lung cancer was not the primary outcome for the WHI, which was looking prospectively at hormones, and ways to prevent heart disease, breast and colorectal cancer, and osteoporosis,” noted Patel. “But our results are similar to those found in the SEER database.”

Researchers were unable to analyze the relevance of molecular and genetic factors for incidence and mortality of lung cancer in various ethnic groups. However, such factors may be discussed by recent studies, such as the WHI Life and Longevity After Cancer study, or the European Prospective Investigations Into Cancer studies, the article said. The United Kingdom Biobank may also contain data relevant to such analysis, it added.


  1. Patel M, Wang A, Kapphahn K, et al. Racial and ethnic variations in lung cancer incidence and mortality: results from the Women&rsquo;s Health Initiative [published online ahead of print December 23, 2015]. J Clin Oncol. pii:JCO635789. Accessed January 3, 2016.
  2. Women&rsquo;s Health Initiative.
  3. Patel M, Schupp C, Gomez S, et al. How do social factors explain outcomes in non—small-cell lung cancer among Hispanics in California? Explaining the Hispanic paradox. J Clin Oncol. 2013;31:3572-3578.
  4. Women&rsquo;s Health Initiative.
  5. World Health Organization.
  6. United Kingdom Biobank.