Abstract
Background: Smoking and drinking alcohol are major risk factors for cancer development, and we investigated their effects on gastric cancer prognosis following initial resection.
Methods: Data from male patients with stage III–IV gastric adenocarcinoma who underwent surgery between 2001 and 2006 were retrospectively reviewed. Patients were followed up until 2011. Kaplan–Meier plots and Cox proportional hazards regressions were applied for survival rates.
Results: Among 238 patients, 151 (63.4%) smoked and 146 (61.3%) drank alcohol. Current smokers had an increased risk of cancer recurrence or death from any cause [adjusted HR (aHR), 1.94; 95% confidence interval (CI), 1.18–3.21], cancer recurrence (aHR, 1.89; 95% CI, 1.12–3.21), and overall mortality (aHR, 2.14; 95% CI, 1.23–3.73) compared with never-smokers. Patients with a lifetime cigarette smoking of <40 and ≥40 pack-years had increased cancer recurrence or death from any cause (aHR, 1.72 and 2.43, respectively; 95% CI, 1.03–2.86 and 1.38–4.30, respectively), cancer recurrence (aHR, 1.63 and 2.61, respectively; 95% CI, 0.95–2.79 and 1.43–4.77, respectively), and overall mortality (aHR, 1.92 and 2.75, respectively; 95% CI, 1.09–3.38 and 1.47–5.12, respectively) compared with never-smokers. However, drinking alcohol was not associated with postsurgery survival.
Conclusions: Cigarette-smoking history at the time of diagnosis, but not drinking history, is associated with cancer recurrence and poor survival after surgery in male patients with stage III–IV gastric cancer.
Impact: These findings encourage physicians to advise patients with gastric cancer to stop smoking to obtain a general health benefit and likely improvement in the gastric cancer course. Cancer Epidemiol Biomarkers Prev; 22(10); 1805–12. ©2013 AACR.
Introduction
In Republic of Korea, the overall age-standardized gastric cancer incidence rate in 2009 was 43.0 per 100,000 persons (64.5 and 25.5 in males and females, respectively), and gastric cancer was the second most common cancer in the total population and the most common cancer in males (1). Gastric cancer survival rates have been increasing steadily due to better detection methods and more effective adjuvant therapies (1–3). Recently, the effects of modifiable risk factors such as smoking and alcohol drinking on cancer patient survival were studied as part of an effort to improve survival. Smoking was associated with poor prognosis in patients with lung (4), liver (5), colon (6, 7), and prostate cancer (8), and in non-Hodgkin lymphoma (NHL; ref. 9). Heavy drinkers had worse outcomes for head, neck, liver (5), and breast cancer (10), and NHL (9), in a dose-dependent manner.
Although smoking and alcohol drinking are well-known preventable risk factors for developing gastric cancer (11–15), the precise impact of smoking and alcohol drinking history on gastric cancer survival rates remained controversial. In a Japanese study with 877 patients with gastric cancer, the risk ratio for death was 2.53 [95% confidence interval (CI), 1.22–5.29] for habitual smokers, and an inverse dose–response relationship was found between ever smoking and gastric cancer patient survival (16). On the other hand, in a study with 14,578 Korean men, Park and colleagues identified that patients with gastric cancer with a prediagnosis smoking history had better survival rates than nonsmokers in multivariate-adjusted analyses (HR for mortality, 0.83; 95% CI, 0.72–0.95; ref. 5). Alcohol intake was also associated with an excess risk of dying, and a dose–response relationship with increasing levels of consumption was shown (Ptrend = 0.02). Patients in the highest tertile of alcohol intake had a 40% increased risk of all-cause death (HR, 1.41; 95% CI, 1.04–1.91; ref. 17). However, in a Japanese study, current or ever-habitual drinking did not significantly increase mortality HRs (16).
Because the effect of smoking and alcohol drinking on the survival of patients with gastric cancer remained unclear, we carried out a study to evaluate the effect of these factors on the prognoses of patients with gastric cancer in a Korean population.
Materials and Methods
Study population
In this retrospective study, patients presenting to the Center for Gastric Cancer at the National Cancer Center (NCC), a tertiary referral cancer center hospital, in Goyang, Republic of Korea between August 2001 and June 2006 were potentially eligible for inclusion. Male patients were included if they: (i) were of ages 20 years or older, (ii) had a histologically confirmed stage III/IV gastric adenocarcinoma, (iii) had newly diagnosed cancer without previous treatment, (iv) underwent R0 total or subtotal gastrectomy, and (v) completed an interview about health behavior (smoking and alcohol drinking history and current status). Finally, our study included 238 male patients. Only patients with stage III–IV gastric cancer were analyzed because the survival of patients with cancer differs substantially between stages (18). This study was approved by the Institutional Review Board of the NCC in Republic of Korea (NCCNCS-13-721).
The majority of patients with stage III–IV cancer with lymph node metastasis received adjuvant chemotherapy. Several combination chemotherapies were administered every 3 weeks for six to eight cycles. The detailed information about the chemotherapy regimen used is described elsewhere (19).
Clinical data
Clinical data were obtained from a review of medical records at the NCC. Clinicopathologic characteristics included tumor location, tumor size, Lauren classification, depth of invasion, lymph node metastasis, adjuvant chemotherapy, operation method, lymph node dissection, tumor stage, and first-degree family history of gastric cancer. The stage at diagnosis was determined with adherence to the sixth edition of the Union for International Cancer Control/American Joint Committee on Cancer classification system (20). Family history of cancer was ascertained by self-reporting and positive first-degree relatives including siblings, parents, or offsprings with gastric cancer.
Smoking and alcohol consumption assessment
Smoking status, smoking duration, number of cigarettes smoked per day, and cumulative pack-years information was collected by self-reporting through an interview with the physicians at the time of cancer diagnosis. Smoking status was categorized as never-smoker, ex-smoker, or current smoker based on the smoking history that the patients reported at the time of cancer diagnosis. Smokers were defined as patients who had smoked at least 1 year in their lifetime. Ex-smokers were defined as patients who had stopped smoking at least 12 months before cancer diagnosis. Current smokers were patients who were smoking within 1 year before the diagnosis, regardless of their smoking status thereafter. So, those who had given up smoking more recently (less than 12 months prior) were considered as current smokers. From former smokers, data about age at first smoking, the time since smoking cessation, and the number of cigarettes previously smoked per day were collected. From current smokers, data about the age at first smoking and the number of cigarettes smoked per day were collected. Smoking duration was categorized as never-smoker, less than 30 years, and 30 or more years. The number of cigarettes smoked per day was categorized as never-smoker, 20 or less per day, and more than 20 per day. Lifetime pack-years were defined as the estimated number of packs of cigarettes smoked per year over the subject's lifetime. Pack-years were divided into three groups: never-smoker, <40 pack-years, and ≥40 pack-years.
Alcohol drinking status was categorized as never, ex-drinker, or current drinker. In ex-drinkers or current drinkers, alcohol drinking frequency (never, ≤3 times/week, 4–5 times/week, or every day), duration of alcohol drinking (never, <20, 20–39, and ≥40 years), and alcohol-consumption amount (never, <30 g/day, and ≥30 g/day) data were additionally collected. Alcohol-consumption amount was calculated as grams of pure alcohol, based on the beverage type (soju, beer, makgeolli, and spirits) and the amount of each type consumed on a typical drinking day, assuming the following alcohol content by volume: soju 25%, beer 4%, makgeolli 5%, and spirits 40%. Soju (distilled spirits) and makgeolli (unrefined rice wine) are traditional alcoholic beverages in Republic of Korea.
Statistical analysis
The main outcome was disease-free survival (DFS), which was defined as the time from surgery to tumor recurrence, occurrence of a new primary gastric cancer, or death from any cause. Recurrence-free survival (RFS) was defined as the time from surgery to tumor recurrence, death with evidence of recurrence, or occurrence of a new metachronous gastric tumor. Patients who died without confirmed tumor recurrence were censored at the last documented evaluation. Overall survival (OS) was defined as the time from surgery to death from any cause.
Follow-up data through June 2011 were obtained for recurrence or death. Survival curves were constructed using the Kaplan–Meier method, and a log-rank test was used for comparison. Multivariate analyses were conducted using the Cox proportional hazards model adjusting for age, smoking, drinking, tumor location, tumor size, Lauren classification, adjuvant chemotherapy, tumor stage, and first-degree family history of gastric cancer for DFS, RFS, and OS. All analyses were conducted using SAS software (version 9.1; SAS Institute).
Results
Among the 478 patients with stage III–IV gastric cancer who underwent total or subtotal gastrectomy from 2001 to 2006, all 150 female patients were excluded. Furthermore, 90 male patients were excluded because of missing values in smoking or alcohol drinking status. Finally, 238 men were included in this study and their mean age was 56.0 ± 11.8 years. More than 77% of the tumors were located in the lower two third of the stomach. The mean tumor diameter was 6.7 ± 2.8 cm and 18.9% patients had a first-degree family history of gastric cancer (Table 1).
. | N = 238 . |
---|---|
Age, y | |
≤49 | 76 (31.9) |
50–59 | 65 (27.3) |
60–69 | 60 (25.2) |
≥70 | 37 (15.6) |
Overall mean patient age ± SD, y | 56.0 ± 11.82 |
Location of tumor | |
Upper one third | 46 (19.3) |
Lower two third | 185 (77.7) |
Whole stomach | 7 (2.9) |
Tumor diameter, cm | |
<4.0 | 29 (12.2) |
4.0–5.9 | 71 (29.8) |
≥6.0 | 138 (58.0) |
Overall mean tumor diameter ± SD | 6.7 ± 2.80 |
Lauren classification | |
Intestinal | 121 (50.8) |
Diffuse or mixed type | 117 (49.2) |
Depth of invasiona | |
≤pT2 | 68 (28.6) |
pT3 | 112 (47.1) |
pT4 | 58 (24.4) |
LN metastasisa | |
≤N1 | 63 (26.5) |
N2 | 154 (64.7) |
N3 | 21 (8.8) |
Adjuvant chemotherapy | |
No | 31 (13.0) |
Yes | 207 (87.0) |
Operation method | |
Subtotal gastrectomy | 131 (55.0) |
Total gastrectomy | 107 (45.0) |
LN dissection | |
D1 | 2 (0.8) |
D2 | 236 (99.2) |
Tumor stagea | |
III | 170 (71.4) |
IV | 68 (28.6) |
First-degree family history of gastric cancer | |
No | 193 (81.1) |
Yes | 45 (18.9) |
. | N = 238 . |
---|---|
Age, y | |
≤49 | 76 (31.9) |
50–59 | 65 (27.3) |
60–69 | 60 (25.2) |
≥70 | 37 (15.6) |
Overall mean patient age ± SD, y | 56.0 ± 11.82 |
Location of tumor | |
Upper one third | 46 (19.3) |
Lower two third | 185 (77.7) |
Whole stomach | 7 (2.9) |
Tumor diameter, cm | |
<4.0 | 29 (12.2) |
4.0–5.9 | 71 (29.8) |
≥6.0 | 138 (58.0) |
Overall mean tumor diameter ± SD | 6.7 ± 2.80 |
Lauren classification | |
Intestinal | 121 (50.8) |
Diffuse or mixed type | 117 (49.2) |
Depth of invasiona | |
≤pT2 | 68 (28.6) |
pT3 | 112 (47.1) |
pT4 | 58 (24.4) |
LN metastasisa | |
≤N1 | 63 (26.5) |
N2 | 154 (64.7) |
N3 | 21 (8.8) |
Adjuvant chemotherapy | |
No | 31 (13.0) |
Yes | 207 (87.0) |
Operation method | |
Subtotal gastrectomy | 131 (55.0) |
Total gastrectomy | 107 (45.0) |
LN dissection | |
D1 | 2 (0.8) |
D2 | 236 (99.2) |
Tumor stagea | |
III | 170 (71.4) |
IV | 68 (28.6) |
First-degree family history of gastric cancer | |
No | 193 (81.1) |
Yes | 45 (18.9) |
NOTE: Data are expressed as number (%).
aUnion for International Cancer Control/American Joint Committee on Cancer (UICC/AJCC) staging system, sixth edition.
Abbreviation: LN, lymph node.
Smoking and alcohol consumption behavior
Among 238 male patients with gastric cancer, 18.5% and 63.4% were ex- and current smokers, respectively. Approximately, 39% of patients smoked for 30 years or more and 15.5% of patients smoked more than 20 cigarettes (one pack) per day. Furthermore, 58.8% and 23.1% of the patients had less than and more than 40 pack-years, respectively (Table 2).
. | N = 238 . |
---|---|
Cigarette-smoking status | |
Never-smoker | 43 (18.1) |
Ex-smoker | 44 (18.5) |
Current smoker | 151 (63.4) |
Smoking duration, y | |
Never-smoker | 43 (18.1) |
<30 | 103 (43.3) |
≥30 | 92 (38.7) |
Number of cigarettes smoked per day | |
Never-smoker | 43 (18.1) |
≤20 | 158 (66.4) |
>20 | 37 (15.5) |
Lifetime cigarette smoking (pack-years) | |
Never-smoker | 43 (18.1) |
<40 | 140 (58.8) |
≥40 | 55 (23.1) |
Alcohol-drinking status | |
Never drank | 52 (21.8) |
Ex-drinker | 40 (16.8) |
Current drinker | 146 (61.3) |
Alcohol-drinking frequency, per weeka | |
Never drank | 52 (22.0) |
≤3 | 112 (47.5) |
4–5 | 32 (13.6) |
Every day | 40 (16.9) |
Duration of alcohol drinking, yb | |
Never drank | 52 (22.3) |
<20 | 29 (12.5) |
20–39 | 90 (38.6) |
≥40 | 62 (26.6) |
Alcohol-consumption amount, g/da | |
Never drank | 52 (22.0) |
<30 | 88 (37.3) |
≥30 | 96 (40.7) |
. | N = 238 . |
---|---|
Cigarette-smoking status | |
Never-smoker | 43 (18.1) |
Ex-smoker | 44 (18.5) |
Current smoker | 151 (63.4) |
Smoking duration, y | |
Never-smoker | 43 (18.1) |
<30 | 103 (43.3) |
≥30 | 92 (38.7) |
Number of cigarettes smoked per day | |
Never-smoker | 43 (18.1) |
≤20 | 158 (66.4) |
>20 | 37 (15.5) |
Lifetime cigarette smoking (pack-years) | |
Never-smoker | 43 (18.1) |
<40 | 140 (58.8) |
≥40 | 55 (23.1) |
Alcohol-drinking status | |
Never drank | 52 (21.8) |
Ex-drinker | 40 (16.8) |
Current drinker | 146 (61.3) |
Alcohol-drinking frequency, per weeka | |
Never drank | 52 (22.0) |
≤3 | 112 (47.5) |
4–5 | 32 (13.6) |
Every day | 40 (16.9) |
Duration of alcohol drinking, yb | |
Never drank | 52 (22.3) |
<20 | 29 (12.5) |
20–39 | 90 (38.6) |
≥40 | 62 (26.6) |
Alcohol-consumption amount, g/da | |
Never drank | 52 (22.0) |
<30 | 88 (37.3) |
≥30 | 96 (40.7) |
NOTE: Data are expressed as number (%).
aExcluded two cases whose alcohol-drinking frequency and amount were not obtained.
bExcluded five cases whose alcohol-drinking duration was not obtained.
With regard to alcohol-drinking behavior, 16.8% and 61.3% of patients were ex- and current drinkers, respectively. Approximately, 16.9% of patients drank every day and 26.6% drank for more than 40 years. Furthermore, 40.7% drank more than 30 g alcohol per day (Table 2).
Effect of smoking and alcohol drinking on the survival of stage III–IV gastric cancer patients
The overall DFS, RFS, and OS at 5 years were 43.2%, 47.6%, and 51.3%, respectively. In the Kaplan–Meier survival curves, DFS at 5 years of never-, ex-, and current smokers were 58.1%, 37.2%, and 40.7%, respectively (P = 0.080; Fig. 1). According to the lifetime amount of cigarette smoking, DFS was significantly better in patients who had never smoked compared with smokers with 40 pack-years or more (P = 0.008; Fig. 2). DFS at 5 years was 58.1% for never-smokers compared with 44.6% and 27.7% for smokers with less than 40 pack-years and smokers with 40 pack-years or more, respectively (overall P = 0.008, Ptrend = 0.002; Fig. 2).
The Cox proportional hazard model showed significant associations between cigarette smoking and survival in male patients with stage III–IV gastric cancer. Current smokers had an increased risk of cancer recurrence or death from any cause [adjusted HR (aHR), 1.94; 95% CI, 1.18–3.21], cancer recurrence (aHR, 1.89; 95% CI, 1.12–3.21), and overall mortality (aHR, 2.14; 95% CI, 1.23–3.73) compared with never-smokers after controlling for patient age, alcohol drinking status, tumor location, tumor size, Lauren classification, adjuvant chemotherapy, tumor stage, and first-degree family history of gastric cancer. Also, lifetime cigarette smoking (pack-years) of less than 40 and 40 years or more had increased cancer recurrence or death from any cause (aHR, 1.72 and 2.43, respectively; 95% CI, 1.03–2.86 and 1.38–4.30, respectively), cancer recurrence (aHR, 1.63 and 2.61, respectively; 95% CI, 0.95–2.79 and 1.43–4.77, respectively), and overall mortality (aHR, 1.92 and 2.75, respectively; 95% CI, 1.09–3.38 and 1.47–5.12, respectively) compared with never-smokers. Smokers who had a longer duration of smoking and greater numbers of cigarette smoking per day showed significantly worse survival (Table 3). However, alcohol-drinking status, alcohol-drinking frequency, duration of alcohol drinking, and alcohol-consumption amount in male patients with gastric cancer were not associated with survival (Table 4).
. | Cancer recurrence or death from any cause . | Cancer recurrence . | Overall mortality . | |||
---|---|---|---|---|---|---|
. | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . |
Cigarette-smoking status | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
Ex-smoker | 1.73 (0.96–3.12) | 1.74 (0.94–3.23) | 1.70 (0.91–3.16) | 1.78 (0.93–3.42) | 2.11 (1.12–4.00) | 2.13 (1.09–4.17) |
Current smoker | 1.73 (1.06–2.83) | 1.94 (1.18–3.21) | 1.72 (1.02–2.89) | 1.89 (1.12–3.21) | 1.89 (1.09–3.27) | 2.14 (1.23–3.73) |
Smoking duration, y | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<30 | 1.59 (0.95–2.67) | 1.64 (0.96–2.81) | 1.63 (0.95–2.81) | 1.64 (0.93–2.88) | 1.78 (1.01–3.15) | 1.87 (1.03–3.39) |
≥30 | 1.89 (1.13–3.16) | 2.25 (1.31–3.88) | 1.81 (1.05–3.13) | 2.20 (1.23–3.93) | 2.11 (1.19–3.73) | 2.45 (1.36–4.44) |
Ptrend | 0.017 | 0.003 | 0.043 | 0.008 | 0.012 | 0.003 |
Number of cigarettes smoked per day | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
≤20 | 1.63 (0.99–2.67) | 1.79 (1.08–2.95) | 1.62 (0.96–2.73) | 1.73 (1.02–2.95) | 1.77 (1.02–3.07) | 1.96 (1.12–3.42) |
>20 | 2.20 (1.22–3.97) | 2.49 (1.36–4.55) | 2.16 (1.16–4.03) | 2.56 (1.36–4.84) | 2.73 (1.45–5.13) | 3.02 (1.58–5.79) |
Ptrend | 0.008 | 0.003 | 0.014 | 0.003 | 0.001 | 0.001 |
Lifetime cigarette smoking (pack-years) | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<40 | 1.54 (0.94–2.55) | 1.72 (1.03–2.86) | 1.52 (0.90–2.58) | 1.63 (0.95–2.79) | 1.71 (0.98–2.97) | 1.92 (1.09–3.38) |
≥40 | 2.29 (1.33–3.96) | 2.43 (1.38–4.30) | 2.29 (1.29–4.08) | 2.61 (1.43–4.77) | 2.63 (1.44–4.78) | 2.75 (1.47–5.12) |
Ptrend | 0.002 | 0.002 | 0.004 | 0.002 | 0.001 | 0.001 |
. | Cancer recurrence or death from any cause . | Cancer recurrence . | Overall mortality . | |||
---|---|---|---|---|---|---|
. | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . |
Cigarette-smoking status | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
Ex-smoker | 1.73 (0.96–3.12) | 1.74 (0.94–3.23) | 1.70 (0.91–3.16) | 1.78 (0.93–3.42) | 2.11 (1.12–4.00) | 2.13 (1.09–4.17) |
Current smoker | 1.73 (1.06–2.83) | 1.94 (1.18–3.21) | 1.72 (1.02–2.89) | 1.89 (1.12–3.21) | 1.89 (1.09–3.27) | 2.14 (1.23–3.73) |
Smoking duration, y | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<30 | 1.59 (0.95–2.67) | 1.64 (0.96–2.81) | 1.63 (0.95–2.81) | 1.64 (0.93–2.88) | 1.78 (1.01–3.15) | 1.87 (1.03–3.39) |
≥30 | 1.89 (1.13–3.16) | 2.25 (1.31–3.88) | 1.81 (1.05–3.13) | 2.20 (1.23–3.93) | 2.11 (1.19–3.73) | 2.45 (1.36–4.44) |
Ptrend | 0.017 | 0.003 | 0.043 | 0.008 | 0.012 | 0.003 |
Number of cigarettes smoked per day | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
≤20 | 1.63 (0.99–2.67) | 1.79 (1.08–2.95) | 1.62 (0.96–2.73) | 1.73 (1.02–2.95) | 1.77 (1.02–3.07) | 1.96 (1.12–3.42) |
>20 | 2.20 (1.22–3.97) | 2.49 (1.36–4.55) | 2.16 (1.16–4.03) | 2.56 (1.36–4.84) | 2.73 (1.45–5.13) | 3.02 (1.58–5.79) |
Ptrend | 0.008 | 0.003 | 0.014 | 0.003 | 0.001 | 0.001 |
Lifetime cigarette smoking (pack-years) | ||||||
Never-smoker | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<40 | 1.54 (0.94–2.55) | 1.72 (1.03–2.86) | 1.52 (0.90–2.58) | 1.63 (0.95–2.79) | 1.71 (0.98–2.97) | 1.92 (1.09–3.38) |
≥40 | 2.29 (1.33–3.96) | 2.43 (1.38–4.30) | 2.29 (1.29–4.08) | 2.61 (1.43–4.77) | 2.63 (1.44–4.78) | 2.75 (1.47–5.12) |
Ptrend | 0.002 | 0.002 | 0.004 | 0.002 | 0.001 | 0.001 |
aaHRs adjusted for age, alcohol drinking status, tumor location, tumor size, Lauren classification, adjuvant chemotherapy, tumor stage, and first-degree family history of gastric cancer.
Abbreviation: cHR, crude HR.
. | Cancer recurrence or death from any cause . | Cancer recurrence . | Overall mortality . | |||
---|---|---|---|---|---|---|
. | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . |
Alcohol-drinking status | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
Ex-drinker | 0.86 (0.49–1.50) | 0.73 (0.41–1.30) | 0.93 (0.52–1.65) | 0.79 (0.43–1.45) | 0.84 (0.47–1.50) | 0.69 (0.38–1.26) |
Current drinker | 1.01 (0.67–1.53) | 0.99 (0.66–1.51) | 1.02 (0.66–1.59) | 1.00 (0.64–1.57) | 0.99 (0.65–1.53) | 1.00 (0.65–1.55) |
Alcohol-drinking frequency (per weekb) | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
≤3 | 0.87 (0.56–1.33) | 0.79 (0.51–1.22) | 0.91 (0.58–1.43) | 0.81 (0.51–1.29) | 0.91 (0.58–1.41) | 0.84 (0.53–1.33) |
4–5 | 1.49 (0.87–2.54) | 1.40 (0.81–2.42) | 1.55 (0.88–2.74) | 1.48 (0.83–2.63) | 1.28 (0.74–2.24) | 1.14 (0.65–2.01) |
Every day | 1.02 (0.60–1.73) | 1.03 (0.59–1.78) | 0.93 (0.52–1.66) | 0.96 (0.53–1.75) | 0.88 (0.50–1.56) | 0.90 (0.49–1.65) |
Ptrend | 0.459 | 0.366 | 0.690 | 0.479 | 0.974 | 0.874 |
Duration of alcohol drinking, yc | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<20 | 0.81 (0.44–1.49) | 0.72 (0.37–1.37) | 0.80 (0.42–1.53) | 0.63 (0.32–1.26) | 0.95 (0.51–1.75) | 0.86 (0.45–1.65) |
20–39 | 0.97 (0.63–1.50) | 1.03 (0.65–1.63) | 1.02 (0.65–1.60) | 1.01 (0.62–1.64) | 0.84 (0.54–1.32) | 0.89 (0.55–1.44) |
≥40 | 1.11 (0.70–1.77) | 0.94 (0.55–1.59) | 1.07 (0.65–1.76) | 1.02 (0.58–1.81) | 1.13 (0.70–1.83) | 0.91 (0.52–1.58) |
Ptrend | 0.600 | 0.964 | 0.649 | 0.707 | 0.760 | 0.755 |
Alcohol consumption amount, g/db | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<30 | 0.91 (0.58–1.42) | 0.83 (0.52–1.31) | 0.96 (0.60–1.54) | 0.86 (0.53–1.39) | 0.93 (0.58–1.47) | 0.86 (0.53–1.38) |
≥30 | 1.06 (0.69–1.64) | 1.02 (0.66–1.58) | 1.05 (0.66–1.66) | 1.01 (0.63–1.62) | 0.99 (0.63–1.56) | 0.95 (0.60–1.51) |
Ptrend | 0.693 | 0.743 | 0.802 | 0.772 | 0.965 | 0.995 |
. | Cancer recurrence or death from any cause . | Cancer recurrence . | Overall mortality . | |||
---|---|---|---|---|---|---|
. | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . | cHR (95% CI) . | aHR (95% CI)a . |
Alcohol-drinking status | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
Ex-drinker | 0.86 (0.49–1.50) | 0.73 (0.41–1.30) | 0.93 (0.52–1.65) | 0.79 (0.43–1.45) | 0.84 (0.47–1.50) | 0.69 (0.38–1.26) |
Current drinker | 1.01 (0.67–1.53) | 0.99 (0.66–1.51) | 1.02 (0.66–1.59) | 1.00 (0.64–1.57) | 0.99 (0.65–1.53) | 1.00 (0.65–1.55) |
Alcohol-drinking frequency (per weekb) | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
≤3 | 0.87 (0.56–1.33) | 0.79 (0.51–1.22) | 0.91 (0.58–1.43) | 0.81 (0.51–1.29) | 0.91 (0.58–1.41) | 0.84 (0.53–1.33) |
4–5 | 1.49 (0.87–2.54) | 1.40 (0.81–2.42) | 1.55 (0.88–2.74) | 1.48 (0.83–2.63) | 1.28 (0.74–2.24) | 1.14 (0.65–2.01) |
Every day | 1.02 (0.60–1.73) | 1.03 (0.59–1.78) | 0.93 (0.52–1.66) | 0.96 (0.53–1.75) | 0.88 (0.50–1.56) | 0.90 (0.49–1.65) |
Ptrend | 0.459 | 0.366 | 0.690 | 0.479 | 0.974 | 0.874 |
Duration of alcohol drinking, yc | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<20 | 0.81 (0.44–1.49) | 0.72 (0.37–1.37) | 0.80 (0.42–1.53) | 0.63 (0.32–1.26) | 0.95 (0.51–1.75) | 0.86 (0.45–1.65) |
20–39 | 0.97 (0.63–1.50) | 1.03 (0.65–1.63) | 1.02 (0.65–1.60) | 1.01 (0.62–1.64) | 0.84 (0.54–1.32) | 0.89 (0.55–1.44) |
≥40 | 1.11 (0.70–1.77) | 0.94 (0.55–1.59) | 1.07 (0.65–1.76) | 1.02 (0.58–1.81) | 1.13 (0.70–1.83) | 0.91 (0.52–1.58) |
Ptrend | 0.600 | 0.964 | 0.649 | 0.707 | 0.760 | 0.755 |
Alcohol consumption amount, g/db | ||||||
Never drank | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
<30 | 0.91 (0.58–1.42) | 0.83 (0.52–1.31) | 0.96 (0.60–1.54) | 0.86 (0.53–1.39) | 0.93 (0.58–1.47) | 0.86 (0.53–1.38) |
≥30 | 1.06 (0.69–1.64) | 1.02 (0.66–1.58) | 1.05 (0.66–1.66) | 1.01 (0.63–1.62) | 0.99 (0.63–1.56) | 0.95 (0.60–1.51) |
Ptrend | 0.693 | 0.743 | 0.802 | 0.772 | 0.965 | 0.995 |
aaHRs adjusted for age, cigarette smoking status, tumor location, tumor size, Lauren classification, adjuvant chemotherapy, tumor stage, and first-degree family history of gastric cancer.
bExcluded two cases whose alcohol-drinking frequency and consumption amount were not obtained.
cExcluded five cases whose alcohol-drinking duration was not obtained.
Abbreviation: cHR, crude HR.
Discussion
By using a cohort of resected patients with gastric cancer, we evaluated the impact of smoking and alcohol drinking on gastric cancer recurrence and survival. Cigarette smoking was significantly associated with poor survival. Especially, the dose–response association was noted for lifetime pack-years consumed. In contrast, alcohol-drinking behavior was not associated with survival.
In our study subjects, ex- and current smoking rates were 18.5% and 63.4%, respectively. The smoking rate in Korean adult males, which has been reported to be 48.3% in a 2010 survey, seems higher compared with that of other countries (21, 22). Our result was comparable with the previous Korean studies, showing a higher smoking rate in male patients with cancer than the general population. Ex- and current smokers constituted 17.2% and 64.8%, respectively, of the 14,578 Korean male patients with cancer in one study (5) and 17.2% and 56.7%, respectively, of the 12,242 male patients with gastric cancer in another study (15). Smoking is considered a main risk factor for gastric cancer in Republic of Korea. Men who smoked for 20 to 39 years had a 2.09-fold increased risk of gastric cancer compared with nonsmokers, and those who smoked for more than 40 years had a 3.13-fold increased risk (3).
In this study, the overall DFS, RFS, and OS at 5 years were 43.2%, 47.6%, and 51.3%, respectively. These figures are compatible but slightly higher than those reported in a recent Korean study (23). In agreement with our results, a Japanese study of gastric cancer (24) showed a better survival rate than that of Western studies (25, 26), despite an advanced stage. One of the main differences among the studies that can affect the outcome is the extent of lymph node dissection. The D2 dissection rate of both the Japanese and our study was more than 99% (24), whereas in the Western studies, the D2 dissection rate was only 10% in one study (25) and approximately 40% in another study (26). Also, surgical mortality might be low because the NCC in Korea is a high-volume center. Surgical patients in high-volume hospitals have shown significantly better survival than those in low- to medium-volume hospitals (27). Here, we included only patients who had R0 resection carried out. So, the survival rate of patients in our study was higher than that of overall stage III–IV patients.
Our observed lower survival rates in smokers compared with never-smokers might be due to the biologic effect of smoking and health behavior of smokers. Direct genotoxicity and nongenotoxic (epigenetic) effects of cigarette smoke underlie the mechanisms of tumor promotion and progression (28). Nicotine, although not carcinogenic by itself, has been shown to promote the growth of gastric cancer cells by stimulating cell proliferation and angiogenesis (29). Smokers have more rapidly progressing tumors and cancer metastasis than nonsmokers (28). Also, tobacco smoking has been shown to affect the immune system. Specifically, natural killer cell activity against cancer cells was reduced significantly in smokers (30).
The health behavior of smokers, which is not as good as that of never-smokers, may also contribute to this survival difference (31–34). In the 2006 Korean National Cancer Screening Survey, an annual cross-sectional survey that uses nationally representative random sampling to investigate the cancer-screening rates, current smokers showed lower rates of endoscopy use than nonsmokers (31). Hence, smokers may be at an increased risk of dying of gastric cancer or cancer recurrence due to the differences in the screening behavior rather than the smoking per se. In addition, other unhealthy lifestyles, such as low-fruit diet, inadequate physical activity, and unhealthy diet were risk factors for poor prognoses of gastric cancer (33, 34). These behaviors were observed more frequently in smokers versus nonsmokers (35, 36). The effect of screening participation and other health behavior could not be verified directly in our study because these data were not collected. Further studies are needed to clarify this point.
In this study, dose–response relationships were observed between postsurgery survival and lifetime cigarette smoking. Smokers who had longer pack-year accumulations showed significantly higher aHRs than never-smokers. Two earlier Asian studies showed the dose–response relationship between cigarette smoking and gastric cancer death. In a report studying Japanese men, daily cigarette consumption and mortality from gastric cancer showed a significant dose–response relationship (37). In a study from Taiwan, dose–response relationships between death from gastric cancer and the quantity of smoking, duration of smoking, age of smoking initiation, and cumulative cigarette smoking in pack-years were reported (38). However, these early reports did not control for other known clinicopathologic prognostic factors of gastric cancer. In our study, adjustment for known clinical factors associated with more aggressive tumor behavior did not fully attenuate our finding that greater smoking intensity equals lower survival after surgery.
Our observation of lack of a relationship between alcohol consumption and survival among patients with gastric cancer is consistent with several past studies (39, 40). In contrast, alcohol use was significantly associated with worse outcome among patients with stomach cancer in some studies (17, 41). These inconsistencies in the literature might be due to the way that the alcohol-drinking behavior is assessed, for example, in only assessing the alcohol-drinking status or the alcohol-consumption amount (17, 39, 42–44). In our study, alcohol consumption was assessed in terms of both alcohol-drinking status and amount. Therefore, the method of assessing alcohol drinking does not fully explain these inconsistencies.
There are some study limitations that should be considered. First, these analyses are based on the self-reporting of tobacco and alcohol-use history. Hence, the rates of reported smoking and alcohol consumption might have been underestimated as a result of social desirability. Previous study showed that 39.1% of ex-smokers who reported quit smoking within 1 year had an elevated serum cotinine level (45). Thus, the definition of current smokers in our study may reduce the risk of misrepresentation from the patients, but may also increase the risk of the people who were not actually using tobacco at the time of diagnosis. Second, we did not collect the information on changes in smoking and drinking habits subsequent to cancer diagnosis and surgery. Finally, only male patients with gastric cancer were analyzed, because we did not enroll enough women due to low smoking rate of women in Republic of Korea. In the Korea National Health and Nutrition Examination Survey 2010, the current smoking rate of women was 6.3% (21).
The strengths of the current study include the histologically confirmed incident gastric cancer cases and the detailed interview data that were assembled to evaluate the influence of risk factors on long-term survival. Compared with most previous studies, our study included information about cigarette-smoking volume and alcohol-consumption amount as well as cigarette-smoking status and alcohol-drinking status, which might more fully reflect smoking and alcohol abuse.
In summary, our study shows that the prognosis after surgical resection for stage III–IV gastric cancer is worse for smokers than never-smokers, particularly in current smokers. This finding warrants physicians to strongly advise patients with gastric cancer to stop smoking because of detrimental effects of current smoking on the course of their disease.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: M.A. Han, I.J. Choi, M.G. Oh
Development of methodology: M.A. Han, I.J. Choi
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.-W. Kim, I.J. Choi, J.Y. Lee, B.W. Eom, H.M. Yoon, K.W. Ryu
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.A. Han, I.J. Choi, M.G. Oh
Writing, review, and/or revision of the manuscript: M.A. Han, Y.-W. Kim, M.G. Oh, C.G. Kim, S.-J. Cho
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y.-W. Kim
Study supervision: Y.-W. Kim, I.J. Choi
Grant Support
This work was supported by grant 1310280 from the National Cancer Center, Republic of Korea.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.