SEARCH

SEARCH BY CITATION

Keywords:

  • hormone replacement therapy;
  • lung cancer;
  • survival;
  • smoking;
  • female

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

BACKGROUND:

Hormone replacement therapy (HRT) may play a role in the development of lung cancer and subsequent survival. Results from studies exploring these issues are inconsistent. A retrospective study in a rural population was conducted to determine whether a history of HRT use is associated with survival of postmenopausal women with lung cancer.

METHODS:

A retrospective medical chart review of 648 postmenopausal women, diagnosed with a first primary lung cancer between1995 and 2005, was conducted in a regional hospital in Paducah, Kentucky. History of HRT use was collected. Log-rank test and multivariate Cox regression analysis were performed to examine the effects of HRT on survival.

RESULTS:

The median survival for women with a history of HRT use was 16.4 months, compared with 10.5 months for women without a history of HRT use. However, this difference in survival was not statistically significant (hazard ratio, 1.09; 95% confidence interval, 0.82-1.44). Women with a history of HRT use were younger on average (64.3 years) at diagnosis than women without a history of HRT use (69.5 years, P < .01). Cigarette smoking was adversely associated with survival (P = .03), as were age (P < .01) and TNM stage (P < .01).

CONCLUSIONS:

In contrast to previous studies, within this population, a history of HRT use in postmenopausal lung cancer patients was not associated with decreased survival. Because most of the published studies on this issue are retrospective, the discrepant findings reflect the complexity of the role of HRT use in the survival of lung cancer patients. Cancer 2009. © 2009 American Cancer Society.

In the United States, lung cancer is the most commonly diagnosed malignancy and the leading cause of cancer death. In 2008, the American Cancer Society estimated that 215,020 new cases of lung cancer will have been diagnosed and approximately 161,840 individuals will have died from the disease.1 Women account for approximately 47% of the newly diagnosed cases and 44% of the deaths from lung cancer.1 In Kentucky, although the age-adjusted lung cancer incidence rate for men decreased steadily through 1995 to 2005, the age-adjusted lung cancer incidence rate for women increased 13% in the same time period.2 A similar trend occurred in the age-adjusted mortality rates for lung cancer among women in Kentucky.2

Compared with lung cancer in men, lung cancer in women has different characteristics. Women are more likely than men to develop adenocarcinoma and small cell lung cancer.3, 4 Although cigarette smoking is the leading cause of lung cancer in both men and women, women appear to be more susceptible to the carcinogenic effects of tobacco smoke on the lung. In fact, individuals who have never smoked, and yet are diagnosed with lung cancer, are more likely to be women than men.5-9 Several studies have suggested that hormone replacement therapy (HRT) in women is associated with the increased incidence of adenocarcinoma in women and sex differences in the histological distribution of lung carcinoma.6, 10-13 Conversely, other investigators have reported that HRT use is associated with a decrease in lung cancer risk.14-18

Another unanswered question regarding HRT use and lung cancer is its impact on survival. Female lung cancer patients consistently have better survival than men.9, 19-22 The reasons for this sex-related disparity in lung cancer survival are still not totally understood. A possible factor in improved lung cancer survival among women is estrogen. To date, published research regarding HRT use and lung cancer mortality in women is limited, and the published findings are inconsistent. Some studies have shown that HRT use is associated with a decreased risk for lung cancer mortality,17, 23 whereas several other studies show that HRT use is associated with an increased risk of mortality in women with lung cancer.15, 24

Most of these published studies were designed to assess whether a history of HRT use affected the risk of developing lung cancer. The impact of HRT use on the survival of female lung cancer patients has typically been of secondary interest. However, a recent retrospective study conducted by Ganti looked at survival as the primary endpoint.25 Ganti et al reported that patients using HRT had a significantly higher of risk of dying (hazard ratio, 1.97) compared with patients not using HRT.

To further examine the association between HRT use and survival in female lung cancer patients, this retrospective study of postmenopausal women with lung cancer was conducted.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

The Kentucky Cancer Registry (KCR) is a population-based registry that is part of the National Cancer Institute's Surveillance, Epidemiology, and End Results program. KCR collaborated with Western Baptist Hospital to conduct this retrospective study with approval from the institutional review board at the University of Kentucky. Western Baptist Hospital is the largest hospital in western Kentucky, whose patients typically come from the Paducah area and surrounding rural counties. The 2003 Beale codes for the Western Baptist Hospital service area range from 7 to 9 and reflect the rurality of this area.26 Only women who were postmenopausal at the time of their diagnosis with lung cancer were included in this study. The medical records of all 671 postmenopausal female lung cancer cases diagnosed between 1995 and 2005 from the Western Baptist Hospital were reviewed. The data analysis included 648 first primary lung cancer cases.

The review consisted of case information retrieved from medical chart review, microfilmed charts, and off-site review of physician office charts. All forms of HRT used before the diagnosis of lung cancer were considered as positive for a history of HRT use, including either a combination of estrogen and progestin or estrogen alone. The menopausal status of women in the study was determined from the medical record review. The postmenopausal group included women who were documented as postmenopausal, were older than 60 years, or who had undergone a hysterectomy before age 60 years. The data collected from Western Baptist Hospital were merged with KCR data, and an updated survival status was obtained through the routine linkage of KCR data with National Death Index data. Women who had a history of smoking >5 pack years were considered “smokers,” whereas women who smoked <5 pack years or never smoked were considered “nonsmokers.” TNM stage and type of lung cancer were determined according to the guidelines of the American Joint Committee on Cancer.27

All of the statistical analyses were performed using the SAS 9.1 statistical software package (SAS Institute, Cary, NC). The Pearson chi-square test was used to test the difference between the groups with and without a history of HRT use in terms of age at diagnosis, TNM stage, health insurance, smoking history, race, cancer cell type, family history, and first course of treatment. A Student t test was also used to test the mean difference in age at diagnosis between the groups with and without a history of HRT use. Multivariate Cox proportion hazard regression models were used to assess the association between a history of HRT use and survival. Cases with unknown TNM stage and smoking status were excluded in the multivariate Cox regression analysis. The final mode was selected based on the backward stepwise selection method. Log-rank test was used to test for the equality of survival functions, and the Kaplan-Meier survival curves were plotted by history of HRT use and smoking status. The criterion for statistical significance was P value <.05.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

A large majority (n = 620, 96%) of the 648 postmenopausal women with lung cancer were white. Six in 10 of the women (n = 410, 61.9%) had a history of at least 5 pack years of smoking, and 12.6% had smoked, but <5 pack years. Unfortunately, a quarter of the sample (n = 165, 25.5%) had either incomplete smoking information or were missing pack years of smoking. The mean age at diagnosis was 68.4 years (range, 37-90 years); 31.8% (n = 206) of the women were <65 years of age. Because of the older age of the study population, the majority of women (n = 466, 71.9%) were Medicare beneficiaries. Only 26 women (4%) had no insurance coverage. Over half of the women (52.9%) had no family history of lung cancer (Table 1).

Table 1. Characteristics of Women With First Primary Lung Cancer
CharacteristicsNo. of Patients, n=648Percentage
  1. HRT indicates hormone replacement therapy.

Race  
 White62096.0
 Black264.0
 Other20.0
Smoking history  
 No8212.7
 Yes40161.9
 Missing16525.5
Medical insurance  
 No insurance264.0
 Insured11617.9
 Medicaid314.8
 Medicare46671.9
 Other40.6
 Unknown50.8
Age at cancer diagnosis, y  
 <6520631.8
 65-7426741.2
 75+17527.0
 Mean68.4 
 Range37-90 
Family history  
 Yes7711.9
 No34352.9
 Unknown22835.2
History of HRT use  
 No47873.8
 Yes11417.6
 Unknown568.6
Type of cancer  
 Small cell12819.8
 Nonsmall cell52080.2
  Adenocarcinoma23445.0
  Squamous cell carcinoma10319.8
  Adenosquamous carcinoma71.4
  Large cell carcinoma10319.8
  Other7314.0
TNM classification  
 Stage I13520.8
 Stage II314.8
 Stage III19530.1
 Stage IV24237.4
 Unknown stage456.9
First course treatment  
 No therapy7311.3
 Surgery only10115.6
 Chemotherapy only6910.7
 Radiation only12619.4
 Surgery+chemotherapy152.3
 Surgery+radiation162.5
 Chemotherapy+radiation21332.9
 Surgery+chemotherapy+radiation355.4

A history of HRT use was found in 478 (73.8%) women; 114 (17.6%) women had no history of HRT use. HRT information could not be retrieved for 56 (8.6%) women. No statistical differences were found between groups with HRT information and with missing HRT information in terms of age at diagnosis, TNM stage, smoking status, and cancer survival. The most common diagnosis was nonsmall cell lung cancer (n = 520, 80.2%). Within this specific lung cancer cell type, 234 (45.0%) had adenocarcinoma, followed by squamous cell carcinoma (n = 103, 19.8%), large cell carcinoma (n = 103, 19.8%), and adenosquamous carcinoma (n = 7, 1.4%). Only 166 (25.2%) women had lung cancer diagnosed at an early stage (stage I and II); 437 women (67.5%) had cancer diagnosed at a late stage (stage III and IV). Chemotherapy plus radiation was the most commonly reported treatment (n = 213, 32.9%) for this study population, followed by radiation only (n = 126, 19.4%) and surgery only (n = 101, 15.6%). About 11.3% of patients did not receive any treatment.

Comparison between the groups of women with a history of HRT use and no history of HRT use showed no statistical difference in terms of race (P = .44), smoking status (P = .33), family history (P = .07), cancer type (P = .32), cancer stage (P = .61), or history of cancer (P = .57). Women with a history of HRT use were younger (mean age, 64.3 years) than women without a history of HRT use (mean age, 69.5 years, P < .01). This explains why a higher percentage of Medicare beneficiaries were found in women without a history of HRT use (Table 2).

Table 2. Characteristics of Women With First Primary Lung Cancer by History of HRT Use
CharacteristicsNo History of HRT Use (n=478)Had History of HRT Use (n=114)P
No. of PatientsPercentageNo. of PatientsPercentage
  1. HRT indicates hormone replacement therapy.

Race    .44
 White/other45695.411197.4 
 Black204.232.6 
 Other20.400.0 
Smoking history    .33
 No6513.6119.7 
 Yes29661.96960.5 
 Missing11724.53429.8 
Medical insurance    .01
 No insurance194.065.3 
 Insured7114.93228.1 
 Medicaid224.676.1 
 Medicare35975.16859.7 
 Other30.610.9 
 Unknown40.800.0 
Age at cancer diagnosis, y    <.01
 <6513127.45649.1 
 65-7419540.84842.1 
 75+15231.8108.8 
 Mean69.5 64.3 <.01
 Range41-90 46-87  
Family history    .07
 Yes6012.687.0 
 No24551.37162.3 
 Unknown17336.23530.7 
Type of cancer     
 Small cell9519.91815.8.32
 Nonsmall cell38380.19684.2 
  Adenocarcinoma17044.44546.9.91
  Squamous cell carcinoma7419.31818.8 
  Adenosquamous carcinoma51.322.1 
  Large cell carcinoma7820.41616.7 
  Other5614.61515.6 
TNM stage    .61
 Stage I9519.93026.3 
 Stage II214.465.3 
 Stage III14229.73228.1 
 Stage IV18538.73934.2 
 Unknown stage357.376.1 
First course treatment    .06
 No therapy6012.697.9 
 Surgery only7215.12219.3 
 Chemotherapy only4810.01513.2 
 Radiation only9820.51210.5 
 Surgery+chemotherapy112.343.5 
 Surgery+radiation112.343.5 
 Chemotherapy+radiation15632.63833.3 
 Surgery+chemotherapy+radiation224.6108.8 

Women with a history of HRT use had an increased overall survival (median survival, 16.4 months; 95% confidence interval [CI], 12.5-20.4 months) compared with women without a history of HRT use (median survival, 10.5 months; 95% CI, 9.2-12.1 months; Fig. 1). However, the difference was not statistically significant (P = .12). The increased survival in women with a history of HRT use is likely because of a younger average age. Smokers had significantly worse survival than nonsmokers (median survival, 11.3 vs 16.9 months; P = .03; Fig. 2). Among nonsmokers, overall survival by HRT status was not different. Among smokers, women with a history of HRT use had significantly better survival (median survival, 16.2 months; 95% CI, 9.8-21.4 months) than women without a history of HRT use (median survival, 10.4 months; 95% CI, 9.1-12.1 months; P = .04; Fig. 3).

thumbnail image

Figure 1. Kaplan-Meier survival estimate of overall survival by history of hormone replacement therapy (HRT) use is shown. The women with a history of HRT use had a median survival of 16.4 months, a 1-year survival rate of 46.3%, and a 5-year survival rate of 15.0% (n = 114). The women without a history of HRT use had a median survival of 10.5 months, a 1-year survival rate of 60.5%, and a 5-year survival rate of 18.3% (n = 478). Log-rank test resulted in a P value of .12.

Download figure to PowerPoint

thumbnail image

Figure 2. Kaplan-Meier survival estimate of overall survival by smoking history is shown. The smokers had a median survival of 11.3 months, a 1-year survival rate of 47.8%, and a 5-year survival rate of 13.6% (n = 401). The nonsmokers had a median survival of 16.9 months, a 1-year survival rate of 57.1%, and a 5-year survival rate of 25.5% (n = 82). Log-rank test resulted in a P value of .03.

Download figure to PowerPoint

thumbnail image

Figure 3. Kaplan-Meier survival estimate of overall survival by hormone replacement therapy (HRT) use among smokers is shown. The smokers with a history of HRT use had a median survival of 16.2 months, a 1-year survival rate of 59.4%, and a 5-year survival rate of 18.7% (n = 69). The smokers without a of history HRT use had a median survival of 10.4 months, a 1-year survival rate of 44.8%, and a 5-year survival rate of 11.9% (n = 296). Log-rank test resulted in a P value of .03.

Download figure to PowerPoint

The multivariate Cox regression analysis found that older age at diagnosis (hazard ratio, 0.03 per year; P < .01), late stage at diagnosis (hazard ratio, 1.56, 3.47, and 5.91; P = .13, <.01, and <.01 for TNM stage II, III, and IV compared with TNM stage I, respectively), and smoking history (hazard ratio, 1.41; P = .03) were significantly associated with a higher risk of dying (Table 3). A history of HRT use was not found to be significantly associated with survival (hazard ratio, 1.09; P = .55). No significant interaction effects were found, including the interaction effect between smoking and HRT use, which was found in pair-wise comparisons.

Table 3. Multivariate Cox Regression Analysis of Overall Survival for History of HRT Use
VariableHazard Ratio95% CIP
  1. HRT indicates hormone replacement therapy; CI, confidence interval.

History of HRT use (yes vs no)1.090.82-1.44.55
Age at cancer diagnosis0.031.02-1.05<.01
TNM stage (baseline=stage I)  <.01
 Stage II1.560.88-2.75.13
 Stage III3.472.51-4.79<.01
 Stage IV5.914.31-8.11<.01
Smoking history (yes vs no)1.411.03-1.91.03

Multivariate Cox regression analyses were also performed for several subsets of data: white women only, nonsmokers only, smokers only, small cell lung cancer only, and nonsmall cell lung cancer only. No significant association was found between a history of HRT use and survival of lung cancer patients in any of the subset data analyses. Because older women are likely to have used HRT for a longer period of time, data analyses were performed for 4 age groups separately (<55, 55-59, 60-64, and 65+ years) to examine if duration of HRT use affects the association of a history of HRT use and patient survival. No significant associations between a history of HRT use and survival of lung cancer patients were found in any of the 4 subage group analyses (results not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

To our knowledge, this is largest retrospective case study assessing the relationship between a history of HRT use and survival in female lung cancer patients. Our results showed no significant correlation between history of HRT use and lung cancer survival in postmenopausal women. Older age, late TNM stage, and smoking were associated with higher risk of mortality, consistent with known negative prognostic findings.

Although the exact role of estrogen in lung cancer tumorigenesis is unclear, it is known that estrogen has the potential to interact with estrogen receptors (ERs) in the lung. Studies have found nuclear ER-β is highly expressed in lung tumor tissue and mediates transcriptional response types of lung cancer cells.11, 28-30 Stabile et al found that estradiol promotes an association between ER-β and GRIP1/TIF2 coactivators that modify gene expression and stimulate cell growth. Epidermal growth factor receptor protein is down-regulated in response to estrogen and up-regulated by antiestrogens in lung cancer cells.31 Progesterones have been found to be present in about half of lung tumors and meditate pathways that induce apoptosis in lung cancer cells and reduce lung tumor growth.32

Schabath et al found that HRT use was associated with a decrease in lung cancer risk (odds ratio, 0.66; 95% CI, 0.51-0.89) and had the strongest protective effect among current smokers (odds ratio, 0.59; 95% CI, 0.38-0.92).17 Their results also suggested that the benefits of HRT include better survival and reduced risk of dying. Schwartz et al also found that having ever used HRT, increased HRT duration of use, and increased estrogen use were significant predictors in reduced risk of nonsmall cell lung cancer.29 In a recent retrospective study, Ganti et al found that overall survival was significantly higher in patients without HRT use, compared with patients who received HRT (hazard ratio, 1.97; 95% CI, 1.14-3.39).25 The effect was more pronounced in smokers than nonsmokers. Among smokers, women who used HRT had a worse survival compared with women who did not use HRT (median survival, 39 vs 73 months, respectively; P = .03).

Although a history of HRT use was not significantly associated with survival in this retrospective review, some of the study findings are interesting. This study is similar to Ganti's study in terms of methodology. Both studies used data from a retrospective medical chart review. However, the results are quite different. In this study, a pair-wise analysis found a history of HRT use to be protective among smokers compared with those with no history of HRT use. The protective effect was not present in the multivariate analysis. This is likely because of younger age at diagnosis among the group with a history of HRT use (mean age, 64.3 years) compared with those without a history of HRT use (mean age, 69.5 years, P < .01). Smoking and late TNM stage were significantly associated with higher risk of mortality in this study. No such significant differences were found in Ganti's study. The median ages of the patients in the 2 studies are similar (68.4 vs 67 years). However, compared with Ganti's study, the sample in this study was larger (648 vs 498), postmenopausal, and mostly rural. In Ganti's study, pre- and postmenopausal women were included. This study had a higher number/proportion of late stage cancers (for nonsmall cell lung cancer, 62% of stage III and IV vs 47%, respectively) and a substantially shorter length of survival (median survival for history of HRT use, 16.4 vs 63 months, respectively; for no history of HRT use, 10.5 vs 68 months).

Our results showed that women with a history of HRT use had a much younger average initial age at diagnosis than women without a history of HRT use (64.3 vs 69.5 years), similar to the results found in Ganti's study (63 vs 68 years). HRT use has decreased sharply in recent years, possibly because of the release of findings from the Women's Health Initiative in 2003. However, this should not have influenced the prevalence of HRT use among women in this study (1995-2005).33 The results suggest that HRT may promote the development of lung cancer at a younger age. It is also possible the younger age at diagnosis in the group with history of HRT use was because of the duration of HRT use. Published results have demonstrated that long-term use of HRT is potentially associated with decreased lung cancer risk.15

Although this is a large population-based study, it does have some limitations. The exact doses and duration of HRT use, which may be risk factors for lung cancer and impact survival, were not collected and controlled for in this study.17, 34 Although it is speculative, we considered age as a surrogate variable for duration of HRT use. As previously noted, the analyses for the data of 4 separate age groups (<55, 55-59, 60-64, 65+ years) provided similar results, which suggests that lacking information on the duration of HRT use may not alter the results of the study. Because it is possible that women who had a hysterectomy but not an oophorectomy, or were older than 60 years and still maintained their menstrual cycle, misclassification of postmenopausal could have occurred. We do not believe the number of misclassified cases is large enough to alter the results of the study analyses. There were also a considerable number of cases with missing information on cigarette smoking history (25.5%) and family history of cancer in this current study (35.2%). To examine if differences existed between patients with known smoking information and with missing smoking information, and between patients with known family history information and with missing family history information, a series of pair-wise comparison analyses were conducted. No significant differences were found in terms of HRT use, race, age, cancer type, cancer stage, insurance status, and survival. Thus, even if biases from missing smoking history and missing family history information existed, they are not likely to be large enough to bias the results of the analysis, but would rather reduce the statistical power of the analysis.

In conclusion, the contrast between the findings in this study and those previously reported further underscores the complexity of the role of HRT use in survival of lung cancer patients. More studies, collecting a detailed history of hormone therapy, including timing and duration, history of smoking, and sociodemographic factors, are clearly warranted to have a better understanding of this issue. Extensive research is needed to gain insight into how HRT biologically affects the development of lung cancer and subsequent outcomes.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

We thank Donna Schmidt and cancer registrars at Western Baptist Hospital for their assistance in data collection. We also thank Dr. Richard Clayton, Dr. Thomas Tucker, and Dr. Brent Shelton for their comments and suggestions.

References

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References