Lung Cancer 124 (2018) 40–44

Contents lists available at ScienceDirect

Lung Cancer
journal homepage: www.elsevier.com/locate/lungcan

Smoking, alcohol, and nutritional status in relation to one-year mortality in T

Danish stage I lung cancer patients
⁎
Niels Lyhne Christensena,b, , Anders Løkkeb, Susanne Oksbjerg Daltonc, Jane Christensena,
Torben Riis Rasmussenb
a
Department of Documentation and Quality, Danish Cancer Society, Strandboulevarden 49, 2100, Copenhagen Ø, Denmark
b
Department of Respiratory Diseases and Allergy, Aarhus University Hospital, Nørrebrogade 44, 8000, Aarhus, Denmark
c
Danish Cancer Society Research Center, Danish Cancer Society, Strandboulevarden 49, 2100, Copenhagen Ø, Denmark

A R T I C LE I N FO A B S T R A C T

Keywords: Introduction: In addition to the highest incidence rate of lung cancer among the Nordic countries, Denmark has
Stage I the highest mortality rate. Moreover, rates of tobacco and alcohol consumption are among the highest in these
Prognosis countries.
Tobacco Method: In a population-based matched case/control study, we aimed to assess the association between one-year
Smoking
all-cause mortality and a number of smoking-related parameters, high-risk alcohol intake, and nutritional status
Alcohol
in clinical stage I lung cancer patients.
Clinical practice
Results: We included 221 patients who died within one year after diagnosis (early death) and 410 matched
controls who survived more than one year (survivor). The odds ratio (OR) for early death among never-smokers
was 0.3 (CI 95%: 0.1–0.9). There was no significant difference between patients who died early and survivors in
proportions of current smokers (49 vs. 45%), number of cumulated pack-years (45 vs. 46), daily tobacco con-
sumption (15 vs. 14 cigarettes/day), patients who quit smoking after diagnosis (25 vs. 40%) and the prevalence
of chronic obstructive pulmonary disease (COPD) (43 vs. 38%). Patients that died early received more medi-
cations for COPD (p = 0.03) and smoked more after diagnosis, 14 vs. 10 cigarettes per day (p = 0.03). The
unadjusted OR for high-risk alcohol intake was 2.2 (CI 95% 1.3–3.7) in the early death group vs. the survivors.
However, in a treatment-stratified analysis this was observed only for surgically treated patients (OR, 3.2; CI
95% 1.7–6.1). Low nutritional status was associated with early death, unadjusted (OR 2.3; CI 95% 1.4–3.7),
while OR was 1.8 (95% CI 1.0–2.3) adjusted for high-risk alcohol intake and COPD. Treatment selection ac-
cording to and interventions against these factors before and after lung cancer diagnosis may improve outcomes.

1. Introduction
Primarily due to tobacco smoking and excessive alcohol consump-
tion, Denmark ranks as the unhealthiest among the Nordic countries [1]
and the average lifespan is also the lowest (2015) [2]. Furthermore,
Denmark has the highest incidence of lung cancer. There is a growing
body of evidence indicating that continued smoking after the diagnosis
of lung cancer is related to an adverse outcome and that Danish lung
cancer patients also have the highest mortality rate (2015) [3–7]. Since
2006, evidence supporting the effectiveness of the current smoking
cessation therapy has been solidified [8–10]. However, in contrast to
other Nordic countries, the Danish lung cancer guidelines place little
emphasis on smoking cessation and have not included treatment
algorithms against tobacco smoking [11]. The Danish follow-up algo-
rithm, with regular short-interval visits, allows for structured long-term
interventions against risk factors, e.g. continued smoking, for recur-
rence and mortality. However, to our knowledge, no study has eval-
uated the extent of smoking cessation initiatives by physicians in re-
lation to the treatment of or follow-up on lung cancer patients in a
Danish setting.
Alcohol is a risk factor for several types of cancer, but appears to be
only vaguely related to the development of lung cancer, and, in fact,
both protective and oncogenic mechanisms from alcohol have been
described [12–14]. How high alcohol intake affects the prognosis after
lung cancer is not well established, even though a register-based study
did find a negative association between alcohol abuse at diagnosis and

⁎
Corresponding author at: Department of Documentation and Quality, Danish Cancer Society, Strandboulevarden 49, 2100, Copenhagen Ø, Denmark.
E-mail addresses: niechris@rm.dk (N.L. Christensen), andloe@rm.dk (A. Løkke), sanne@cancer.dk (S.O. Dalton), jane@cancer.dk (J. Christensen),
torbrasm@rm.dk (T.R. Rasmussen).

https://doi.org/10.1016/j.lungcan.2018.07.025
Received 14 May 2018; Received in revised form 19 July 2018; Accepted 19 July 2018
0169-5002/ © 2018 Elsevier B.V. All rights reserved.
N.L. Christensen et al.
one-year mortality in a surgical cohort [15], These findings need to be
confirmed in a more recent data material.
Nutritional status is most likely not a risk factor per se for lung
cancer [16]. As it is an indicator of a systemic inflammatory response to
the tumor and increases the patient’s susceptibility to treatment com-
plications, low nutritional status is, however, considered an established
risk factor for adverse outcome in lung cancer patients [17–19]. Since
low nutritional status could also be related to severe comorbidity or
alcohol abuse it may confound associations between these factors and
lung cancer mortality.
In a Danish population-based setting, we sought to establish how
these lifestyle factors and nutritional status were associated with short-
term prognosis, as measured by all-cause one-year mortality in stage I
lung cancer patients (TNM 7th edition) [20]. Moreover, we aimed to
establish the extent of smoking cessation guidance and medical treat-
ment against nicotine dependency provided during the follow-up pro-
gram.
2. Materials and method
The study was designed as a population-based matched case/control
study.
We identified patients registered in the Danish lung Cancer Registry
(DLCR) with clinical stage I lung cancer, diagnosed between January 1,
2011 and December 31, 2014, who died from any cause within the first
year after diagnosis (early death group) and for whom we had a
treatment registration, These patients were then matched with similar
patients according to stage (IA/IB), age, gender, same or previous year
of diagnosis who survived more than one year (one-year survivors).
3. Study variables
3.1. Smoking status
According to the Danish guidelines, patients suspected of having
lung cancer should have their full smoking history included in their
medical record during the diagnostic work-up for lung cancer. We es-
timated the number of pack-years in cigarette equivalents. One pack-
year was defined as 20 cigarette-equivalents per day for 365 days. For
current smokers, the daily self-reported tobacco consumption was es-
tablished (one cigarette = 1 g of tobacco, one cigar = 4 g, one pipe
tobacco and cheroot = 3 g). Intervals were rounded to the lower whole
number. Thus, a patient who had formerly smoked but had not smoked
for at least six months was considered a former smoker. Those who had
quit smoking within six months of the lung cancer diagnosis were re-
gistered as current smokers. For former smokers, the interval in years
between smoking cessation and lung cancer diagnosis was established.
We also collected data on smoking habits after diagnosis. Data were
collected in connection with the follow-up visits or from other hospital
contacts until one year after treatment. Patients were divided into three
categories. Nonsmoker, persistent smoker and quit after diagnosis. For
persistent smokers, we assessed the difference in reported daily tobacco
consumption before and after treatment (cigarette equivalents) by
subtracting the post-treatment consumption from the pretreatment
consumption.
3.2. Alcohol
In accordance with recommendations from the Danish Health
Authority, we defined high-risk alcohol intake as above two and three
units of alcohol/day on average for women and men respectively.
Periodic alcohol abuse and a history of alcohol abuse were also con-
sidered as high-risk alcohol intake.
41
Lung Cancer 124 (2018) 40–44
3.3. Nutritional status
As a part of the general assessment of the patient, nutritional status
is typically registered in the medical record as either low (or poor),
normal or above. If nutritional status was not mentioned and only the
height and weight of the patient were available, the body mass index
(BMI) was calculated and the nutritional status was registered as either
low if BMI < 18.5, normal if BMI was 18.5–24.9 or as above if
BMI > = 25. We did not register data on BMI.
3.4. Comorbidity and cause of death
We registered if a patient either had an existing diagnosis or had
been diagnosed with chronic obstructive pulmonary disease (COPD) or
cardiovascular disease (including cerebrovascular events and excluding
arterial hypertension) in connection with the lung cancer diagnosis. We
registered the types of daily medications for COPD (not including cor-
ticosteroids and/or antibiotics prescribed for an acute exacerbation or a
complication in the diagnostic work-up (DWU)). In addition, we re-
trieved data on forced expiratory volume (liters) in one second (FEV1)
from the DLCR.
Outcome in the present study was all-cause mortality. However, on
the basis of the medical records, we also registered the primary cause of
death. This was categorized as lung cancer, comorbidity, treatment
complications, other (suicide, accidents, etc.), or unknown.
3.5. Statistical analysis
Categorical variables were compared by proportional and Pearson
chi-squared distribution. The Mann-Whitney U test was used to assess
differences in continuous variables between the two groups. Tests for
associations of selected study variables between the two groups were
performed using a conditional logistic regression model. Since age,
gender, stage (Ia/Ib), and year were conditioned upon in the regression
analyses, the analyses were unadjusted and unstratified unless other-
wise stated. It is not mandatory to register normal findings in the
medical records in Denmark. Thus, in the regression analysis, if there
was no mention of either alcohol intake or nutritional status, these
factors were categorized as low-risk alcohol intake and normal status,
respectively.
Calculations were performed with SAS software (SAS system, SAS
Institute, Cary, NC) and Stata software (StataCorp 4905 Lakeway Drive
College Station, Texas 77,845 USA).
4. Results
We finally included 221 early death patients and 410 survivors.
Baseline characteristics are given in Table 1.
In the early death group, 31% died of lung cancer, 22% of co-
morbidity, 18% from treatment complications, and 5% from other
causes while in 24% of the cases the cause of death was registered as
unknown.
4.1. Lifestyle factors at diagnosis
There was a strong adverse association between never-smoked and
death within one year (OR 0.3; CI 95%: 0.1-0.9), as compared to the
survivor group. Aside from never-smoked, none of the smoking-related
factors were associated with early death (Table 2).
There was a significant difference between early death and the
survivor groups in the proportion of patients who had high-risk alcohol
use prior to diagnosis (20 vs. 11%; p = 0.002) corresponding to an OR
of 2.2 (CI 95% 1.4–3.5). When adjusted for nutritional status, smoking
status, daily tobacco consumption, and cumulated pack-years the as-
sociation was unchanged (adjusted OR of 2.2; CI 95% 1.3–3.7). When
analyses were stratified by treatment type, the OR was 3.1 (95% CI,
N.L. Christensen et al. Lung Cancer 124 (2018) 40–44

Table 1 Table 3
Baseline clinic-pathological variables of a population-based case/control study Smoking habits after diagnosis in a population-based case control study of stage
of stage I lung cancer patients diagnosed from 2011 to 2014 in Denmark, ac- I lung cancer patients diagnosed from 2011 to 2014 in Denmark by survival
cording to survival status one year after diagnosis. status one year after diagnosis.
Early death One-year Smoking after Early death (n = 221) Survivor (n = 410) p-value
survivors diagnosis
N = 221 N = 410 n % n %
% % p-value
No 79 36 187 46 0.02
Age mean (range) in years (n) 73.1 (45-89) 72.8 (49-87) 0.59 Yes 33 15 75 18 0.29
Gender female/male 43/57 44/56 0.96 Quit after diagnosis 22 10 68 17 0.02
TNM Unknown/irrelevant 87 39 80 20 < 0.001
T1aN0M0 – stage IA 34 36 0.67 Cigarettes/day 14 * N/A 10** N/A 0.03
T1bN0M0 – stage IA 25 25 0.79 Pre vs. post −1 N/A −3 N/A 0.25
T2aN0M0 – stage IB 40 40 0.85
Histology Pre vs. post, difference in cigarette equivalents smoked before and after treat-
Adenocarcinoma 43 49 0.16 ment. *21 observations. ** 50 observations.
Squamous-cell carcinoma 30 25 0.16
Non small-cell carcinoma 16 16 0.82
Other 10 10 0.99 1.7–5.9) among surgically treated patients (n = 389), while it was 0.8
Treatment (CI 95% 0.3–1.8) among oncologically treated patients (n = 242).
Surgery 57 64 0.07 Compared to a normal nutritional status, low nutritional status was
Stereotactic body radiotherapy 35 31 0.30 associated with early death (OR 2.3; CI 95% 1.4–3.7). Adjustment for
Other 8 5 0.13
high-risk alcohol intake and COPD resulted in an adjusted OR of 1.8,
Patient-related factors
Never-smoker 2 6 0.02 failing to reach statistical significance (CI 95% 1.0–3.2). Due to a high
Current smokers at diagnosis 49 46 0.37 number of missing values, we did not include FEV-1 in the regression
Former smokers analyses.
Pack-years (n) 45 46 0.75
Mean cigarette equivalents/day (n) 15 14 0.51
Years since quitting (Former 13 13 0.9 4.2. Smoking habits and interventions against smoking after diagnosis
smokers) (mean)
COPD 43 38 0.19
COPD medications (n) 2.7 2.3 0.03 Given the difference in follow-up time and potential for immortal
FEV-1 (mean liters) 1.72* 1.86** 0.02 time bias, we did not assess risk estimates for post-diagnosis smoking.
Cardiovascular disease 48 43 0.25 However, among persistent smokers (Table 3), early death patients
High risk alcohol intake 20 11 0.002 appeared to smoke more than the survivors with 14 vs. 10 cigarettes per
No/low risk alcohol intake 59 72 0.001
Alcohol information missing 21 17 0.22
day (p = 0.03). In order to assess interventions against smoking in the
Low nutritional status 24 12 < 0.01 follow-up program, we conducted a subgroup analysis of patients who
Normal nutritional status 37 47 0.02 were smokers at diagnosis, completed curative treatment, and entered
High nutritional status 19 21 0.47 the follow-up program. This subgroup consisted of 52 early death pa-
Nutritional status missing 21 20 0.70
tients and 154 survivors. The medical records from the follow-up visits
P-value for difference of proportions (%) and Mann-Whitney U test for numeric
contained no information on smoking status for 33% of patients in the
variables (n). TNM, tumor, node and metastasis 7th edition Pack-year, 20 ci- early death group and for 21% of the survivors. The proportions of
garette-equivalents/day for 365 days. COPD, chronic obstructive pulmonary registered persistent smokers was 42% (22/52) in the early death group
disease. COPD medications, inhaled beta2agonists, anticholinergics and corti- and 40% (62/154) in the survivor group (p = 0.8). Twenty-five percent
costeroids, per oral montelukast, theophylline and stable low dose prednisolone in the early death group and 40% in the survivor group quit smoking
treatment. FEV-1 forced expiratory volume in one second. Cardiovascular dis- after diagnosis (p = 0.07).
ease, ischemic heart disease, arrhythmias, heart failure, thromboembolism
(incl. cerebrovascular events), aortic aneurisms and syndromes (arterial hy-
pertension not included). High alcohol intake > 14/21 units weekly for women 5. Discussion
and men respectively. Nutritional status, based on assessment of the physician
or the body mass index and considered low if = BMI < 18.5, normal if = BMI We have evaluated the association between three lifestyle-related
ranged from 18.5 to 24.9 and as above if = BMI > =25. * 199 observations. ** risk factors and early death among stage I lung cancer patients.
389 observations. Common to these risk factors is the fact that they can be intervened
against as a supplement to the primary cancer treatment [10,21–23].
Regarding smoking, we found significantly fewer never-smokers in the
Table 2 early death group. This finding supports that the course of disease in
Unadjusted odds ratios with corresponding 95% confidence intervals for early never-smokers is different and associated with a more favorable prog-
death in a population-based case/control study of stage I lung cancer patients nosis [24]. However, the proportion of patients who were never-smo-
diagnosed from 2011 to 2014 in Denmark. kers (2% and 6%, respectively) in our study was much lower than in
Lifestyle factor OR 95%CI other studies. For instance, in a Chinese study of 2289 non-small-cell
lung cancer patients, the proportion of never-smokers was 51% [25],
Never smoker 0.3 0.1-0.9
but given the sociodemographic differences between the study popu-
Smoking at diagnosis 1.1 0.7–1.5
Cigarette equivalents/day 1.0 0.99–1.01
lations, direct comparisons may not be tenable. However, it is likely
Former smoker 0.9 0.7–1.3 that smoking interventions might have a greater impact in our popu-
High-risk alcohol intake 2.2 1.4–3.5 lation with its high prevalence of current and former smokers.
Low nutritional status 2.3 1.4–3.7 Between 45 and 49% of this population were smokers at diagnosis,
High nutritional status 1.0 0.7–1.5
which is in line with the findings of Hildebrand et al. [6], who in an
OR, odds ratio; CI, confidence intervals. American single-center study assessed the extent of smoking cessation
guidance after diagnosis in 948 lung cancer patients diagnosed between
2008 and 2010. On the other hand, Garces et al., who studied the

42
N.L. Christensen et al.
relationship between smoking and quality of life after lung cancer di-
agnosis among 1028 patients diagnosed between 1997 and 2002 [26],
reported a lower proportion (24%) of current smokers at diagnosis and
a higher percentage of never-smokers (18%). This might be due to
differences in smoking definition [27], data collection or study popu-
lations.
We found no difference between the two groups in the proportion of
patients who continued to smoke after diagnosis. However, the persis-
tent smokers among patients in the early death group smoked sig-
nificantly more than the survivors did. Among patients who partici-
pated in follow-up, some 25% of smokers in the early death group quit
after diagnosis while among the survivors 40% reported that they had
quit. Although the low numbers preclude statistical significance for this
difference, our findings suggest that interventions against smoking after
diagnosis may be beneficial in relation to short-term outcome. In a
pivotal systematic review, Parsons et al. assessed the benefits of
smoking cessation after diagnosis. They included both European and
Asian studies of primarily surgically treated patients. All studies had a
minimum of four years of follow-up. It was found that continued
smoking was associated with a hazard ratio for overall mortality of 2.94
(CI95%: 1.15–7.54) [7]. Furthermore, a more recent paper by Andreas
et al, has reviewed the effects of smoking cessation in more advanced
stages of lung cancer. The authors found that smoking cessation was
associated with improved treatment outcomes in patients treated with
radio-chemotherapy and improved quality of life in palliatively treated
patients [28].
There was no mention of smoking habits in between a fifth and a
third of the medical records for patients who participated in the follow-
up program. Furthermore, drawing on hospital medical records, we
registered no prescription of smoking cessation medications for any of
the patients in connection with follow-up visits. Thus, our findings in-
dicate little emphasis from the physician on smoking cessation in the
Danish lung cancer follow-up program. This may reflect the sparse and
unstructured smoking-cessation guidance in the national lung cancer
guidelines that were used nationwide during the study period.
However, since the follow-up of the early death patients, where the
primary cause of death was lung cancer, was more complex than that of
the survivors due to a higher rate of disease recurrence and concurrent
treatments, it seems likely that the focus of the clinical interventions
had different focuses in the two groups. Differences in follow-up time
between the two groups may have introduced information bias and our
findings on post-treatment risk factors should be regarded only as de-
scriptive (Table 3).
The number of COPD patients was similar in both groups, which is
contrary to findings in a meta-analysis of 3761 surgically treated pa-
tients, where COPD was associated with a lower overall survival rate
[29]. The short-term outcome used in our study may explain this dis-
cordance. We did find, however, that the early death patients received
more COPD-medications and had a lower mean FEV-1 than the survi-
vors. Furthermore, COPD seemed to confound the observed association
between poor nutritional status and early death, which may indicate
that the degree of COPD tended to be more severe among early death
patients. Similarly, in a retrospective study of 250 surgically treated
NSCLC patients, Saji et al. found that having Global Initiative for
Chronic Obstructive Lung Disease (GOLD) spirometrically defined stage
2–3 was independently associated with a poor prognosis [30]. Having
applied a broad categorization of cardiovascular disease, there was no
significant overall difference between the two groups. A disease-specific
subcategorization may have revealed differences between the two
groups. However, this was beyond the scope of the present paper.
In our study population, there was a significant association between
early death and high-risk alcohol use. The available data in the patients’
medical records show that 20% of the patients who died early had high-
risk intake and/or abuse. This association remained after adjustment for
potential confounders, but subgroup analysis revealed that the asso-
ciation was restricted to the surgically treated patients. Thus, our
43
Lung Cancer 124 (2018) 40–44
findings based on more recent data and a more inclusive definition of
high-risk alcohol use confirm the findings of a Danish register-based
study comprising surgically treated lung cancer patients from 2007 to
2011 (n = 3363) where alcohol abuse was related to increased mor-
tality [31].
5.1. Strengths and limitations
We have retrieved detailed data on smoking, alcohol intake, and
nutritional status in a population-based setting, which is the major
strength of the present study. We were able to take recent advances in
treatment and changes in e.g. smoking and alcohol patterns into ac-
count by including patients diagnosed in recent years. However, by
selecting the study population on the basis of short-term outcome, we
were unable to assess the long-term impact of these adverse lifestyle
factors. We included only stage I patients, who constitute the patients
with the best prognosis, and the impact of the observed associations
might be different in more advanced stages of lung cancer. However,
concerning smoking, there is no indication that the benefits of cessation
interventions would be limited to a given stage or histologic subtype
[7,28]. The recently revised TNM classification may have produced
slightly different results, but we do not believe that using TNM, 7th
revision has altered the overall conclusions of the present study [32].
Drinking habits and nutritional status were less specified than
smoking history in the medical records. This allowed only for a crude
categorization of high-risk alcohol use. We based the categorization of
the nutritional status on either the stated assessment in the medical
record or on BMI and cannot assess the validity of this variable.
Moreover, the assumption that no mention of alcohol and nutritional
status was equal to no high-risk alcohol intake and normal nutritional
status respectively may have resulted in misclassification. As these
factors were recorded at time of diagnosis and thus independently of
the outcome, it is likely that this misclassification was non-differential
and thus may have led to an underestimation of the observed associa-
tions. We only had access to the physicians’ notes in the hospital’s
medical record and could not identify non-pharmacological smoking
cessation interventions initiated by e.g. nurses and/or general practi-
tioners.
Finally, all data on smoking and drinking habits were self-reported
with a potential for underreporting. Moreover, daily tobacco con-
sumption (cigarettes/day) and duration of smoking in years were often
reported in intervals of five in the medical records, which affected the
precision of the estimates. However, there is no reason to believe that
this potential information imprecision was more predominant in one
particular group, and that it would therefore have altered the direction
of the observed associations.
6. Conclusions
COPD, high-risk alcohol intake, low nutritional status, and the
number of cigarettes smoked after diagnosis were associated with death
within one year among recently diagnosed Danish stage I lung cancer
patients, whereas being a never-smoker was associated with a better
prognosis. Overall, our findings suggest that these adverse lifestyle
factors, which are frequent in Danish society, may in part explain dif-
ferences in lung cancer mortality between Denmark and other coun-
tries. Moreover, we found that interventions against tobacco smoking in
the Danish follow-up program were sparse. Incorporating structured
smoking cessation interventions into the follow-up program may
therefore improve the outcome for Danish lung cancer patients. High-
risk alcohol intake was related to early death in surgically, but not in
oncologically treated patients, and this finding should be taken into
account when deciding upon treatment modality among stage I lung
cancer patients.
N.L. Christensen et al. Lung Cancer 124 (2018) 40–44

Funding [15] A.J. Mehta, Alcoholism and critical illness: a review, World J. Crit. Care Med. 5 (1)
(2016) 27–35, https://doi.org/10.5492/wjccm.v5.i1.27.
[16] A.V. Patel, B.D. Carter, V.L. Stevens, M.M. Gaudet, P.T. Campbell, S.M. Gapstur, The
This work was funded by The Danish Cancer Society. relationship between physical activity, obesity, and lung cancer risk by smoking
status in a large prospective cohort of US adults, Cancer Causes Control 28 (12)
Conflicts of interest (2017) 1357–1368, https://doi.org/10.1007/s10552-017-0949-0.
[17] T. Tsukioka, N. Nishiyama, N. Izumi, et al., Sarcopenia is a novel poor prognostic
factor in male patients with pathological stage I non-small cell lung cancer, Jpn. J.
The authors have no conflicts of interest in relation to the present Clin. Oncol. 47 (4) (2017) 363–368, https://doi.org/10.1093/jjco/hyx009.
study. [18] S. Mori, N. Usami, K. Fukumoto, et al., The significance of the prognostic nutritional
index in patients with completely resected non-small cell lung cancer, PLoS One 10
(9) (2015) 1–10, https://doi.org/10.1371/journal.pone.0136897.
References [19] H. Kawai, Y. Saito, Y. Suzuki, Gender differences in the correlation between prog-
nosis and postoperative weight loss in patients with non-small cell lung cancer,
Interact. Cardiovasc. Thorac. Surg. 25 (2) (2017) 272–277, https://doi.org/10.
[1] S.S. Lim, K. Allen, L. Dandona, et al., Measuring the health-related sustainable
1093/icvts/ivx092.
development goals in 188 countries: a baseline analysis from the global burden of
[20] S. Mirsadraee, D. Oswal, Y. Alizadeh, et al., The 7th lung cancer TNM classification
disease study 2015, Lancet 388 (10053) (2016) 1813–1850, https://doi.org/10.
and staging system: review of the changes and implications, World J. Radiol. 4 (4)
1016/S0140-6736(16)31467-2.
(2012) 0, https://doi.org/10.4329/wjr.v4.i4.128.
[2] World Health Organization. http://www.who.int/gho/mortality_burden_disease/
[21] M. Akbar, M. Egli, Y.E. Cho, B.J. Song, A. Noronha, Medications for alcohol use
life_tables/situation_trends/en/.
disorders: an overview, Pharmacol. Ther. (2017), https://doi.org/10.1016/j.
[3] Nordcan. http://www-dep.iarc.fr/NORDCAN.
pharmthera.2017.11.007.
[4] A.K. Dobson, A. Hyland, R. Reed, et al., Tobacco cessation May improve lung cancer
[22] E. Leedo, J. Gade, S. Granov, A. Mellemgaard, The effect of a home delivery meal
patient survival, J. Thorac. Oncol. 10 (5) (2015) 1014–1019, https://doi.org/10.
service of energy- and protein-rich meals on quality of life in malnourished out-
1097/JTO.0000000000000578.
patients suffering from lung cancer: a randomized controlled trial, Nutr. Cancer 69
[5] M.C. Roach, S. Rehman, T.A. Dewees, C.D. Abraham, J.D. Bradley, C.G. Robinson,
(2017) 444–453, https://doi.org/10.1080/01635581.2017.1283421.
It’s never too late: smoking cessation after stereotactic body radiotherapy for non-
[23] E.R. Park, J.S. Ostroff, G.K. Perez, et al., Integrating tobacco treatment into cancer
small cell lung carcinoma improves overall survival, Pract. Radiat. Oncol. 6 (1)
care: study protocol for a randomized controlled comparative effectiveness trial,
(2016) 12–18, https://doi.org/10.1016/j.prro.2015.09.005.It.
Contemp. Clin. Trials 50 (2016) 54–65, https://doi.org/10.1016/j.cct.2016.07.016.
[6] J.R. Hildebrand, S. Sastry, Stop smoking!” Do We say It enough? J. Oncol. Pract. 9
[24] M. Dias, R. Linhas, S. Campainha, S. Conde, A. Barroso, Lung cancer in never-
(5) (2013) 230–232, https://doi.org/10.1200/JOP.2013.000890.
smokers–what are the differences? Acta Oncol. (Madr) 56 (7) (2017) 931–935,
[7] A. Parsons, A. Daley, R. Begh, P. Aveyard, Influence of smoking cessation after
https://doi.org/10.1080/0284186X.2017.1287944.
diagnosis of early stage lung cancer on prognosis: systematic review of observa-
[25] S. Zheng, R. Wang, Y. Zhang, et al., Former smokers with non-small-cell lung
tional studies with meta-analysis, BMJ 340 (2010) 1–7, https://doi.org/10.1136/
cancers: a comprehensive investigation of clinicopathologic characteristics, onco-
bmj.b5569.
genic drivers, and prognosis, Cancer Med. 5 (8) (2016) 2117–2125, https://doi.org/
[8] D. Gonzales, S.I. Rennard, M. Nides, et al., Varenicline, an alpha4beta2 nicotinic
10.1002/cam4.764.
acetylcholine receptor partial agonist, vs sustained-release bupropion and placebo
[26] Y.I. Garces, P. Yang, J. Parkinson, J.A. Sloan, Smoking and quality of life after lung
for smoking cessation: a randomized controlled trial, JAMA 296 (1) (2006) 47–55,
cancer diagnosis, Chest 126 (6) (2004) 1733–1741, https://doi.org/10.1378/chest.
https://doi.org/10.1001/jama.296.1.47.
126.6.1733.
[9] K. Cahill, L.F. Stead, T. Lancaster, Nicotine receptor partial agonists for smoking
[27] P. Mitchell, T. Mok, H. Barraclough, A. Strizek, R. Lew, M. Van Kooten, Smoking
cessation, Cochrane Database Syst. Rev. (12) (2011), https://doi.org/10.1002/
history as a predictive factor of treatment response in advanced non-small-cell lung
14651858.CD006103.pub5.
cancer: a systematic review, Clin. Lung Cancer 13 (4) (2012) 239–251, https://doi.
[10] K. Cahill, N. Lindson-Hawley, K. Thomas, T. Fanshawe, T. Lancaster, Nicotine re-
org/10.1016/j.cllc.2011.08.003.
ceptor partial agonists for smoking cessation, Cochrane Database Syst. Rev. (5)
[28] S. Andreas, A. Rittmeyer, M. Hinterthaner, R.M. Huber, Smoking cessation in lung
(2016), https://doi.org/10.1002/14651858.CD006103.pub7 www.
cancer-achievable and effective, Dtsch. Arztebl. Int. 110 (43) (2013) 719–724,
cochranelibrary.com.
https://doi.org/10.3238/arztebl.2013.0719.
[11] N.L. Christensen, A. Jekunen, S. Heinonen, S.O. Dalton, T.R. Rasmussen, Lung
[29] L.-E. Tan, R. A M, C.-S. Lim, Association of chronic obstructive pulmonary disease
cancer guidelines in Sweden, Denmark, Norway and Finland: a comparison, Acta
and postresection lung cancer survival: a systematic review and meta-analysis, J.
Oncol. (Madr) 56 (7) (2017) 943–948, https://doi.org/10.1080/0284186X.2017.
Investig. Med. 65 (2) (2017) 342–352, https://doi.org/10.1136/jim-2016-000059.
1315172.
[30] H. Saji, T. Miyazawa, H. Sakai, et al., Survival significance of coexisting chronic
[12] V. Bagnardi, M. Rota, E. Botteri, et al., Alcohol consumption and lung cancer risk in
obstructive pulmonary disease in patients with early lung cancer after curative
never smokers: a meta-analysis, Ann. Oncol. 22 (12) (2011) 2631–2639, https://
surgery, Thorac. Cancer 9 (1) (2018) 19–24, https://doi.org/10.1111/1759-7714.
doi.org/10.1093/annonc/mdr027.
12507.
[13] F. Islami, A. Goding Sauer, K.D. Miller, et al., Proportion and number of cancer
[31] A. Green, J. Hauge, M. Iachina, E. Jakobsen, The mortality after surgery in primary
cases and deaths attributable to potentially modifiable risk factors in the United
lung cancer: results from the Danish lung cancer registry, Eur. J. Cardio-Thoracic
States, CA Cancer J. Clin. 68 (1) (2018) 31–54, https://doi.org/10.3322/caac.
Surg. (2015) 1–6, https://doi.org/10.1093/ejcts/ezv107.
21440.
[32] F.C. Detterbeck, D.J. Boffa, A.W. Kim, L.T. Tanoue, The eighth edition lung cancer
[14] S.M. Álvarez-Avellón, A. Fernández-Somoano, E.M. Navarrete-Muñoz, J. Vioque,
stage classification, Chest 151 (1) (2017) 193–203, https://doi.org/10.1016/j.
A. Tardón, Efecto del alcohol y sus metabolitos en el cáncer de pulmón: estudio
chest.2016.10.010.
CAPUA, Med. Clin. (Barc) 148 (12) (2017) 531–538, https://doi.org/10.1016/j.
medcli.2016.12.033.

44