| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* From the Department of Epidemiology, School of Public Health, Clinical Research, Division of Cardiothoracic Surgery, University of Alabama Birmingham, Birmingham, AL.
Correspondence to: Robert J. Cerfolio, MD, FCCP, Professor of Surgery, Chief of Thoracic Surgery, Division of Cardiothoracic Surgery, University of Alabama at Birmingham, 1900 University Blvd, THT 712, Birmingham, AL 35294; e-mail: Robert.cerfolio{at}ccc.uab.edu
Abstract
Background: The impact that smoking cigarettes has on the characteristics and survival of patients with non-small cell lung cancer (NSCLC) is disputed.
Methods: A retrospective cohort study using a prospective database of patients with NSCLC over a 6-year period. Clinical and histologic characteristics and survival rates were compared between smokers and never-smokers.
Results: There were 730 patients; 562 patients (77%) were smokers and 168 patients (23%) were never-smokers. The overall 5-year survival rate was greater in never-smokers (64%) compared to smokers (56%; p = 0.031). Never-smokers were more likely to be younger (p = 0.04), female (p = 0.01), symptomatic at the time of presentation (p < 0.001), have poorly differentiated tumors (p = 0.04), and have a higher maximum standardized uptake value (maxSUV) on positron emission tomography (PET) (p = 0.026) than smokers. The stage-specific 5-year survival rate was greater for never-smokers compared to smokers for stage I disease (62% vs 75%, respectively; p = 0.02), stage II disease (46% vs 53%, respectively; p = 0.09), and stage III disease (36% vs 41%, respectively; p = 0.13). The 5-year survival rate was significantly lower in patients who had a smoking history of > 20 pack-years.
Conclusions: Never-smokers in whom NSCLC develops are more likely to be young, female, and have poorly differentiated tumors with higher maxSUV values on PET scans. Never-smokers with early-stage cancer have a significantly better survival rate than smokers. Patients with a smoking history of 
20 pack-years have worse survival. Thus, smoking not only causes lung cancer, but once NSCLC is diagnosed, the prognosis becomes worse. A biological, hormonal, and genetic explanation is currently lacking to explain these findings, and these data may help to improve treatment and surveillance.
Key Words: lung • lung cancer • smoking • surgery
Lung cancer is responsible for more deaths than the next three most common solid organ cancers combined in both men and women. In 2006, there were estimated to be 174,470 new patients in whom lung cancer was diagnosed in the United States and 162,460 deaths attributed to it.1 In fact, even after 3 decades of improvements in staging and therapy, the 5-year mortality rate for all patients with lung cancer remains at an extraordinarily high rate of 85%.2 Despite evidence that smoking cigarettes causes lung cancer, the effect that smoking status has on survival after diagnosis is controversial. For example, several studies345678 have reported that smoking cigarettes is a negative prognostic factor, whereas other reports9101112 have found no association between lung cancer and smoking cigarettes. This controversy is enhanced by the limited availability of data. For example, even large national registries, such as the Survival, Epidemiology and End Results database do not contain information on smoking status for patients with non-small cell lung cancer (NSCLC).
It has been estimated that 10% of men and 20% of women in whom lung cancer develops are never-smokers. The fact that lung cancer afflicts lifelong never-smokers has recently received national attention with the death of actress Dana Reeve. The purpose of this study was to examine the differences in the epidemiology and histologic distribution of NSCLC in smokers with NSCLC compared to never-smokers with NSCLC. We also wanted to assess the survival difference between smokers and nonsmokers who were carefully staged and then underwent similar treatment strategies based on the stage of their disease.
Methods and Definitions
This is a retrospective cohort study using an electronic prospective database that was conducted between January 1999 and December 2005. Patients with pathologically proven NSCLC who presented to our lung cancer clinic were eligible for this study. Patients with carcinoid tumors and bronchoalveolar cancers were included in this study. Patients were excluded if their smoking history was unknown or if they had not undergone pathologic staging. The demographics, histology, and survival of never-smokers and current smokers in whom NSCLC had been diagnosed were obtained and compared. The Institutional Review Board at the University of Alabama at Birmingham approved both the electronic prospective database used for this study and this prospective trial. Patient consent was obtained for entry into the prospective database.
Definitions
A never-smoker was defined as a person who had never smoked or who had smoked < 20 cigarettes in his or her lifetime. The pack-year history was defined as the number of total years smoked multiplied by the number of packs of cigarettes smoked on average per day. Since tumor size was most commonly reported by the pathologist and the radiologist in only two dimensions (ie, the width and height), it is reported in square centimeters.
Procedures/Staging/Surgery
Patients who were referred to our surgical clinic from general physicians, internists, pulmonologists, and oncologists were included in this series. Disease in all patients was initially clinically staged using the TNM classification system13 and was eventually staged pathologically. A clinical stage was assigned for the patient based on the fluorodeoxyglucose (FDG)-positron emission tomography (PET) and CT scan results by one physician. All suspicious N2, N3, or M1 areas (maximum standardized uptake value [maxSUV], > 2.5) were sampled prior to pulmonary resection. Mediastinoscopy was used to sample suspicious lymph nodes in the paratracheal area (ie, stations 2R, 4R, 2L, 4L, and the superior part of 7), and/or transesophageal endoscopic ultrasound [EUS] fine-needle aspiration [FNA]14) was used to sample suspicious posterior aortopulmonary window lymph nodes (5, 6), subcarinal lymph nodes (7), periesophageal lymph nodes (8), and inferior pulmonary ligament lymph nodes (9).
Patients with suspected M1 disease in the liver, adrenal gland, or contralateral lung underwent definitive biopsy to prove or disprove the presence of M1 cancer. If the bone or brain was suspected to harbor metastases, an MRI was considered the standard reference test.
Treatment Algorithms
If patients had biopsy-proven N3 or M1 disease, the stage was recorded, but they did not undergo resection. Patients with proven N2 disease (stage IIIa) determined by mediastinoscopy and/or EUS-FNA underwent preoperative chemoradiotherapy using a carboplatinum-based regimen. In general, these patients underwent surgical resection only if the mediastinal lymph node that was initially involved with cancer was downstaged by the neoadjuvant therapy. If there was no evidence of N2 or higher disease after mediastinoscopy or EUS-FNA, patients underwent thoracotomy, pulmonary resection, and complete thoracic lymphadenectomy. In general, after 2005 patients < 75 years of age with completely resected pathologic stage Ib to IIIa disease with normal renal function received adjuvant chemotherapy. Pathologic review was performed via standard techniques, and immunohistochemical staining was employed when appropriate.
Survival Data
Survival data were obtained using the following multiple registries: Social Security death index; Survival, Epidemiology and End Results database linked to the Medicare registry; our own hospital databases such as the tumor registry; our prospective database that had been approved by our institutional review board; the oncology clinic database; and follow-up letters from patient’s primary care or oncology physicians.
Statistical Analysis
The association between smoking status and various outcomes for the entire cohort was evaluated by comparing demographic and histologic characteristics and rates of treatment associated with morbidity in resected patients. The data were entered into a spreadsheet (Excel; Microsoft Corp; Redmond, WA) and subsequently were imported into a statistical software package (SAS, version 9.0; SAS Institute; Cary, NC) for analysis.
We also compared the survival rates between never-smokers and smokers within relevant subgroups, such as those defined by age, race, gender, stage of disease, histology, comorbidity, and treatment-related morbidity. Survival curves were constructed with the Kaplan-Meier method and compared with the use of the log-rank statistic.15 For analyses involving adjustments for potential confounding factors, we used the Cox proportional hazards method. Operative mortalities were excluded from the survival analysis. Patients who were alive at the end of the follow-up period were censored. All p values are two-sided, with a p value of < 0.05 considered to be a statistically significant outcome that was unlikely due to chance.
Results
There were 730 patients who met the entry criteria for this study (512 smokers and 168 never-smokers). The patient characteristics and treatments are shown in Table 1 . Never-smokers were younger than smokers (median ages, 53.6 and 66.2 years, respectively; p = 0.042). Women were disproportionately represented in the group of lung cancer patients who were never-smokers (59%) compared to those who were smokers (21%). Histologic characteristics for the two groups are shown in Table 2 .
|
|
|
|
|
20 pack-years (Fig 3
). The median follow-up time was 3.6 years. The survival results were not different if the patients with carcinoid tumors were eliminated from this analysis.
|
Although the exact mechanism of how bronchogenic malignancy develops is not fully understood, smoking cigarettes remains the primary etiologic culprit. Perhaps, the most important message that health professionals can deliver to the public is to stop smoking and, even better, to never start. Smoking causes about 90% of lung cancer deaths in men and almost 80% in women in the United States.2 Compared to never-smokers, men who smoked cigarettes are about 23 times more likely to develop lung cancer, and women who smoked are about 13 times more likely. As shown in this study, smokers who have early stage NSCLC (ie, stage I and II) have a worse prognosis compared to never-smokers. Thus, smoking not only causes lung cancer but it also worsens the prognosis once the patient received a diagnosis.
Additionally, surprising to some, people who have never smoked cigarettes do develop NSCLC. There are several characteristics that are more commonly seen in never-smokers with NSCLC compared to smokers. In this study, we found that never-smokers were predominantly female and were more likely to be symptomatic (specifically, to have a cough, dyspnea, or weight loss) at the time of presentation. Nonsmokers had a greater incidence of carcinoid and bronchoalveolar cancers. In 2000, Liu et al16 also found that never-smokers with NSCLC were more likely to be women, younger, and to have adenocarcinoma. Interestingly, others have also reported17 that adenocarcinoma is more common in women and that there may be hormonal or genetic factors that explain this finding. For example, women are known to have a different nicotine metabolism than men,18 and to have higher incidences of p53 and K-ras mutations1920 and gender-specific viral infections such as human papilloma virus. These have all been linked to NSCLC and perhaps to adenocarcinoma.2122
In previous studies, we and others have shown1723242526 that women with NSCLC have a better survival rate than men, and this holds true for careful, pathologically staged patients who have undergone complete resection and who had similar adjuvant therapy. Thus, we performed a Cox hazards regression analysis and adjusted for factors found to be significant on univariate analysis (ie, stage of disease at diagnosis, completion of resection, and gender) and found that only stage of disease was an independent predictor of survival. However, smoking status, favoring the never-smokers, approached statistical significance (p = 0.067). A stage-specific multivariate survival analysis, however, did discover a difference for smoking status.
The overall 5-year survival rate in this series is high. This is most likely secondary to several components such as the following: all patients underwent a dedicated FDG-PET scan so the presence of unsuspected M1 disease was minimized; most patients underwent resection, and all patients who underwent resection had a complete thoracic lymphadenectomy so nodal disease was not understaged; many patients with N2 disease had microscopic and not bulky N2 disease; and, finally, this was a cohort composed of predominantly surgical patients. These factors led to careful pathologic staging instead of clinical staging and thus to a higher overall 5-year survival rate. It also may be secondary to the fact that 66 patients had carcinoid tumors. We included these patients since carcinoid tumor is a type of NSCLC,27 and it is commonly seen in nonsmokers and thus is pertinent to any study that compares smokers to nonsmokers in whom NSCLC develops. Finally, even if the patients with carcinoid tumors were eliminated from the survival analysis, the results are unchanged.
Stage-Specific Survival
In this study, never-smokers had a greater 5-year survival rate for stages I, II, and III disease. However, statistical significance was achieved only for patients with stage I disease. Similarly, in 2006 Yoshino et al28 showed that smoking habit and stage, but not gender or histologic subtype, were independent prognostic factors in patients with stage I adenocarcinoma. In 2005, Kawai et al,29 who reported on disease-specific survival, found that men with stage Ia NSCLC who fared better were younger, had better performance status, and had a smoking history of fewer pack-years.
We did not find a survival advantage for nonsmokers who had stage IIIa disease from N2 disease. Similarly, in 2006 Tomita et al30 found that smoking status was not related to patient survival in those patients with pathologic proven N2 or stage IIIa NSCLC. In 2006, Tsao et al31 observed that there was no difference between in survival rates and treatment responses between smokers and nonsmokers in those patients who underwent chemoradiation therapy for NSCLC. Given these data as well as our own experience, perhaps the advantage of being a never-smoker in whom NSCLC develops is restricted to those patients with early-stage disease. More likely is the explanation that once the cancer has metastasized to mediastinal N2 lymph nodes (stage IIIa disease) the bad effects of smoking are blurred by the already poor prognosis.
Mechanism
There are several theories that may explain these data. Smoking is known to cause a decrease in systemic and local immunity in the airway.32 Moreover, there are distinct molecular changes in the airway seen in smokers (which linger for years in former smokers) who do not have cancer.33 These changes are quite similar to the molecular changes seen in lung cancer cells; thus, they may predispose the epithelium to the development of cancer. Smoking is known to cause an up-regulation of oncogenes and a down-regulation of tumor suppressor genes, and this may play a part in the pathophysiologic mechanisms. In 1999, Lackman et al34 found that neonates born to mothers who smoked cigarettes had high urine levels of N-nitrosamine, 4-(methyl-nitrosamino)-1-(3-pyridyl)-1-butanone. This carcinogen has been linked to adenocarcinoma. This may lead to an increase in their chance of the development of smoking-related cancers later in life despite the fact that they themselves never smoked. Unfortunately, we do not have the maternal smoking history of all of the patients in this study to fully analyze this effect on our data.
Interestingly, we found that cancers in never-smokers had a higher median maxSUV on FDG-PET scans than cancers in smokers. This finding seems to conflict with the findings of our prior study35 and another study36 showing that the higher the maxSUV of the tumor on FDG-PET scan, the worse the survival. However, we37 and others3839 have also shown that smokers have a lower maxSUV than nonsmokers for NSCLC cancer, and that this may be due to a higher background standardized uptake value that artificially reduces the maxSUV of the tumor in smokers. This phenomenon may explain why the maxSUV of the cancer in never-smokers was higher, yet their survival was better.
Dose Response
We found that survival significantly worsened in patients with a cigarette smoking history of > 20 pack-years. Previous studies have also identified threshold values similar to ours. In 1999, Fujisawa et al40 reported a threshold dose response of 30 pack-years. They found that patients with a smoking history of > 30 pack-years showed a significant increase in the prevalence of recurrent disease. In 2004, Haque et al41 reported poor survival in those with a smoking history of > 40 pack-years. They suggested this may be due to the inactivation of the protective p53 pathway. In 1997, Ammenheuser et al42 observed an increased frequency of Hprt mutant lymphocytes with an increased number of cigarettes smoked per day; patients who smoked 10 cigarettes per day had the greatest number of mutant lymphocytes. Finally, Dressler et al43 showed that patients who had a shorter period of smoking cessation also had a higher recurrence rate.
Strengths and Limitations
The strengths of this study include the following: the accuracy of the prospectively entered data that was collected and stored by one physician; the use of pathologic and not clinical staging; staging that employed FDG-PET scanning and more recently integrated FDG-PET/CT scans; and the routine use of mediastinoscopy, EUS-FNA, and complete thoracic lymphadenectomy at the time of thoracotomy allowed accurate staging. The use of one physician to stage and operate on all patients also limits the number of confounders. The limitations to this study include the omission of data about second-hand smoke exposure, the fact that smoking history was obtained from the patient, the use of arbitrary definitions of a smoker and a never-smoker, and the inclusion and labeling of patients as smokers who had smoked > 20 cigarettes but had quit smoking many years prior to presentation. The use of overall survival instead of cancer-specific survival can be seen as a limitation and a strength. In our experience, cancer-specific death data are often flawed. The true cause of death of a patient is often unknown even by the caring physician. Family members commonly are unaware of how a loved one died. It is not uncommon that the suggestion of clinical recurrence is used to mean "the cancer is back" but that the patient did not have biopsy-proven recurrence. For example, did a patient who died in a motor vehicle crash die because he had a stroke and lost consciousness or did he die from an intracranial bleed from a brain metastasis? Few patients undergo autopsy to answers these questions; thus, we prefer overall survival as opposed to cancer-specific survival since it may be more accurate and is easy to obtain.
Conclusions
The occurrence of NSCLC is not uncommon among never-smokers. These patients are more likely to be young, female, symptomatic at the time of presentation, have poorly differentiated tumors, and have bronchoalveolar and carcinoid type cancers. Smokers with completely resected early-stage (stage I and II) disease have a worse prognosis than those who never smoked; thus, perhaps the former patients should be more aggressively treated and/or more frequently followed up after undergoing resection. There is a dose response for smoking. Patients who have a 20-pack-year history of smoking have worse survival rates than those who consumed fewer cigarettes. These data need to be corroborated and then considered to help guide future therapies and surveillance.
Footnotes
Abbreviations: EUS = endoscopic ultrasound; FDG = fluorodeoxyglucose; FNA = fine-needle aspiration; maxSUV = maximum standardized uptake value; NSCLC = non-small cell lung cancer; PET = positron emission tomography
The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Received for publication February 21, 2007. Accepted for publication March 21, 2007.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
