IJC International Journal of Cancer
Smoking is associated with lower age, higher grade, higher stage, and larger size of malignant bladder tumors at diagnosis Eline H. van Roekel1,2, Kar K. Cheng1, Nicholas D. James3, D. Michael A. Wallace4, Lucinda J. Billingham1,3, Paul G. Murray3, Richard T. Bryan3 and Maurice P. Zeegers1,5 1
Department of Public Health, Epidemiology and Biostatistics, School of Health and Population Sciences, University of Birmingham, Birmingham, United Kingdom 2 Department of Epidemiology, GROW 2 School for Oncology and Developmental Biology, Maastricht University, Maastricht, The Netherlands 3 School of Cancer Sciences, University of Birmingham, Birmingham, United Kingdom 4 School of Surgery, University of Western Australia, Fremantle, Australia 5 Department of Complex Genetics, Cluster of Genetics and Cell Biology, NUTRIM School for Nutrition, Toxicology and Metabolism, Maastricht University Medical Center1, Maastricht, The Netherlands
Epidemiology
Smoking is a strong risk factor of bladder cancer (BC), but it is currently unclear whether smoking is also associated with severity of BC at diagnosis. We performed a large hospital-based case-comparison study, examining the relation between smoking and clinical characteristics of BC at diagnosis. A total of 1,544 cases from participating hospitals in the West Midlands were recruited between 19 December 2005 and 21 April 2011. Eligible cases were adult BC patients without a previous history of this disease. At time of diagnosis, semi-structured face-to-face interviews were conducted by trained research nurses to collect smoking information. Clinical characteristics were obtained from medical records. Linear mixed models were performed to calculate predicted means in clinical outcomes for a variety of smoking behaviors. A p < 0.05 was considered statistically significant. After adjustment for age and gender, current smokers were on average 4.0 years younger at diagnosis (95% CI: 25.9 to 22.0), had larger tumors (mean difference: 0.48 cm, 95% CI: 0.0420.91), a higher T stage (mean difference: 0.25, 95% CI: 0.0820.41), and a borderline significantly higher grade than never smokers (mean difference: 0.15, 95% CI: 0.0020.30). Our results suggest that smoking could be associated with a more malignant phenotype of BC at diagnosis. More research is needed on the relation between smoking and prognosis, but our results could strengthen the message about the potential risks of smoking to these patients.
Bladder cancer (BC) is the ninth most common cancer worldwide for both sexes combined. In 2008, the global incidence of BC was 386,300, and 150,200 people died of the disease.1 Given the fact that BC mainly occurs in elderly people2 and the increasingly ageing population, it can be expected to remain an important health problem in the future.3 In Europe and North America, more than 90% of all BCs are urothelial carcinomas (UCs, or transitional cell carcinoKey words: smoking, bladder cancer, epidemiology Grant sponsor: Cancer Research UK; Grant sponsor: the Comprehensive Local Research Networks of the West Midlands region; Grant sponsor: the Department of Public Health, Epidemiology and Biostatistics and the School of Cancer Sciences, University of Birmingham; Grant sponsor: the University Hospital Birmingham Charities; Grant sponsor: the Dutch Cancer Society; Grant sponsor: the European Union Erasmus programme; DOI: 10.1002/ijc.28017 History: Received 29 Mar 2012; Accepted 16 Nov 2012; Online 7 Jan 2013 Correspondence to: Kar K. Cheng, Department of Public Health, Epidemiology and Biostatistics, School of Health and Population Sciences, College of Medical and Dental Sciences, University of Birmingham, Birmingham B15 2TT, United Kingdom, Tel.: 144-1214146757, Fax: 144-121-4146216, E-mail:
[email protected]
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
mas, TCCs).4 Most cases (70280%) present with non-muscle-invasive disease (NMIBC, stage Ta, T1, and carcinoma in situ [Tis]), which is rarely lethal, but shows a high recurrence rate of 50270%.5 In about 10220% of patients with Ta/T1 UCs, the disease progresses to muscle-invasion (T2 lesions), which can lead to metastasis and death.5 For patients with NMIBC, it has been observed that tumor grade and stage, and also tumor number, size, presence of carcinoma in situ, and age at diagnosis are risk factors of progression.6,7 The most important environmental risk factor for BC is smoking,8 explaining approximately half of the cases in men and one-third of the cases in women in Europe.9 It has been demonstrated that an increased smoking frequency and duration, and a lower age at initiation are associated with an increased risk of BC, while cessation seems to reduce the risk.9 The relationship between smoking and other prognostic factors is interesting, as it could give insight into biological mechanisms of disease and, perhaps more importantly, have clinical implications by increasing our ability to identify patients at risk of more malignant disease. To our knowledge, the literature on the relationship between lifetime smoking characteristics and the clinical presentation of BC is sparse. We found four studies in which smoking was found to be significantly related to a higher grade at diagnosis.10–13 In contrast, Ouerhani et al.14 and
447
van Roekel et al.
What’s new? Smoking is associated with an increased risk of bladder cancer, but the relationship between lifetime smoking characteristics and clinical presentation of the disease at diagnosis remains limited. In this study, one of the first to investigate this relationship, current smokers were found to be younger and to have larger tumors, higher T stage, and marginally higher tumor grade at diagnosis compared with never smokers. The findings suggest that smoking could be associated with a more malignant type of bladder cancer at diagnosis and could help strengthen the message about the potential risks of smoking.
Material and Methods Setting and participants
This study is part of the West Midlands Bladder Cancer Prognosis Programme (BCPP), which is an ongoing multicenter cohort study in the West Midlands, United Kingdom. Details of the study have been published previously.33 Briefly, adult patients (age 18 years) presenting with symptoms suspicious of BC (predominantly haematuria) and referred to one of the participating urology centers within the region were enrolled into the study on the basis of initial findings suggestive of BC (predominantly abnormal cystoscopy). Those who had a previous diagnosis of cancer of the urethra, bladder, ureter, or renal pelvis within the last decade, HIV infection, or any other condition that might interfere with the ability of the participant to participate fully were excluded. During the enrollment period (19 December 2005 to 21 April 2011), 2,603 new BC cases were eligible within C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
the catchment areas of whom 1,544 (59%) were recruited. The study protocol was approved by the Nottingham Multicenter Research Ethics Committee (06/MRE04/65), and written informed consent was obtained from all participants. Exposure assessment
At the time of diagnosis, semi-structured face-to-face interviews were conducted by trained research nurses to collect information on socio-demographics, health-related lifestyle, medical and drug history, dietary intake, social support, and quality of life. The questions concerning lifetime smoking behavior assessed current and former smoking status, age at initiation, average weekly consumption, and, if applicable, age at cessation. These were measured for filter cigarette, non-filter cigarette, hand-rolled cigarette, pipe, and cigar smoking separately. The classification as former smoker was based on data provided by the patient, and no cut-off point for time since cessation was used. We used the following assumptions for the weighting of different smoking products to calculate a weightand time-adjusted average of the different product frequencies in grams per day: one cigarette equals 1 g, one pipe equals 3 g, and one cigar equals 4 g of tobacco. Furthermore, duration of smoking was calculated by subtracting the age at initiation from the current age or age at cessation, depending on smoking status. We assumed a smoking duration of 6 months if the age at initiation was equal to the age at cessation or the current age. Cumulative smoking amount in kilograms of tobacco was calculated by multiplying smoking frequency by smoking duration. All continuous smoking variables were categorized by creating logical categories with groups of comparable sizes with at least 50 patients per group. For age at initiation we used the categories 20 and >20 years. Outcome assessment
Medical records of all patients were inspected by trained research nurses to gather information on clinico-pathological characteristics of BC at diagnosis. This comprised T stage (according to the TNM 2002 classification system34), grade (according to the World Health Organization 1973 system35), size of the largest tumor, and the number of visible tumors. If at initial resection the tumor was reported as stage Ta/T1, but a repeat resection within 3 months indicated muscle-invasive disease, then a muscle-invasive tumor at diagnosis was assumed. We calculated age at interview by subtracting date of birth from date of interview, and used this as a proxy measure for age at
Epidemiology
Sturgeon et al.15 found smoking to be associated with the development of low-grade tumors. The evidence for smoking being associated with a higher T stage is stronger, as the number of studies reporting a statistically significant association is somewhat higher.10,13,16–20 A few studies17,21–23 have observed a lower age at diagnosis in current smokers than in ex- and/or non-smokers: Michalek et al.22 demonstrated a negative correlation between lifetime exposure to smoking and age at diagnosis, while a Danish study23 showed a higher percentage of heavy smokers amongst younger patients. In addition, four studies11,12,17,21 have indicated that smoking is associated with larger tumor size, which was statistically significant in only two studies.11,12 In a study of Mohseni et al.,12 a higher percentage of patients with multiple lesions was observed amongst smokers than non-smokers. However, several other studies investigating smoking and at least one of these variables have not found any association.4,24–32 Overall, the number of participants in studies so far has been low: we identified only five studies15–17,20,28 reporting results from over 1,000 patients, but these were registry-based, possibly indicating less detail and poorer quality of smoking and clinical outcome assessment. In addition, some studies included only Ta/T1 cases.11,21,24,30 We performed a large hospital-based case-comparison study of BC cases, examining the relationship between smoking (with emphasis on status, frequency, cumulative smoking amount, age at initiation, and time since cessation) and clinical characteristics of BC at presentation (including age and tumor grade, stage, size, and number).
448
Smoking and bladder cancer clinical characteristics at diagnosis
diagnosis. Stage was recoded into a numerical outcome variable suitable for analysis, by coding Ta tumors with a value “0” and all other stages according to their T-number. Patients with Tis tumors (n510) were excluded from the analysis of stage, as these tumors generally have a relatively more aggressive clinical behavior.36 In addition, patients with a tumor number higher than ten (n57) were regarded as outliers and therefore recoded as missing. Further, the outcome variables were dichotomized by making logical groups (T stage: muscle-invasive vs. non-muscle-invasive; grade: high (III) vs. low (I/II); tumor number: multiple vs. one tumor) or by using the median (tumor size) or mean (age at diagnosis) value as cut-off point.
unavailable in 101 patients (9%), resulting in a number of 1,067 patients suitable for analysis. Most patients were former smokers (n 5 616), with the remaining being either current (n 5 227) or never smokers (n 5 224). The median time since cessation for former smokers was 23.9 years (range: 0–64.7 years). A total of 39 patients (6%) had stopped within the previous year. Table 1 shows the demographics and clinical characteristics of included patients with BC by smoking status. Overall, nearly four times more males than females were included, and the percentage of males was higher amongst current (82%) and former (84%) than amongst never smokers (60%). The mean age at diagnosis of current (65.7 years), former (72.7 years), and never smokers (69.9 years) was significantly different (p < 0.001). Most patients (80%) were diagnosed with UC, 43 patients (4%) had non-UC (12 squamous carcinomas, 12 neuroendocrine tumors, 1 adenocarcinoma, 2 sarcomas, and 16 other unspecified types), and in 16% of patients this was unknown. The percentage of non-UCs in current (6%) and former smokers (4%) was higher than in never smokers (1%), which was borderline significant (p 5 0.049). Muscle-invasive BC (T2) was less frequent in never (12%) than in current (23%) and former smokers (24%), which was reflected by a significant difference in T stage between the smoking groups (p 5 0.002). Crude estimates by ANOVA tests showed no significant differences in grade, tumor size, and tumor number. The patients with Tis tumors (n 5 10) are included in Table 1, but were excluded from the statistical analyses of T stage.
Epidemiology
Data analysis
For current, former, and never smokers frequencies per category of gender, histological type of BC, as well as T stage, grade, tumor number, age, and tumor size at diagnosis were presented and compared. Afterwards, multivariable analyses were performed using generalized linear mixed models to analyze the relation between smoking and clinical outcomes, in which random intercepts for hospital site were incorporated to correct for potential clustering of patients in hospital reference areas. Separate models were made for each smoking type and for the sum of all smoking types combined (total tobacco smoking). Smoking status was fitted as a categorical variable with never smokers as the reference group. If the number of current smokers of any type of tobacco was less than 50, current and former smokers were combined into one category “ever smokers.” Never smokers were excluded from the analyses of smoking characteristics other than status. Two types of models were fitted with either the smoking variable as a categorical or an ordinal variable (dose–response analysis). In models with categorical smoking variables, the lowest exposure was chosen as the reference category. The results of the dose–response analysis were shown if a statistically significant dose–response relationship according to the Wald-test of the fitted ordinal smoking variable was found. All analyses were adjusted for age and gender, except for models with age at diagnosis as the outcome in which only gender was adjusted for. Sensitivity analyses were performed by fitting mixed effects logistic regression models with dichotomized clinical outcomes. In addition, we investigated a potential interaction effect by gender. From the fitted linear mixed models the predicted means in clinical outcomes in smoking groups, as well as increments for ordinal smoking variables, with 95% confidence intervals (CIs) were calculated. For calculation of predicted means in adjusted models the mean values of age and gender were used. Statistical analyses were performed using STATA version 11.0,37 and a pvalue of 20
70.1–72.3
69.92
16–20
71.2
130
139
11–15
70.4–73.4
64.2–67.2
71.8–73.8
68.1–71.2
95% CI
69.7–71.4
107
6–10
2
71.92
65.71
72.8
1
69.7
Pred mean
70.62
128
5
Smoking frequency (g/day)
616
Former smoker
224
Current smoker
Never smoker
Smoking status
n
Age at diagnosis (years)
Table 2. Adjusted predicted means of clinical variables at diagnosis in relation to different smoking characteristics
Epidemiology
123
112
117
200
641
114
185
90
92
103
167
189
120
127
95
115
209
560
197
n
1.83
1.92
1.81
2.01
2.00
1.71
1.83
1.91
2.18
2.15
1.77
1.88
1.83
2.10
2.14
1.72
2.08
1.94
2.10
Pred mean
1.50–2.17
1.57–2.27
1.47–2.15
1.72–2.30
1.81–2.19
1.36–2.06
1.53–2.13
1.51–2.31
1.78–2.57
1.77–2.52
1.46–2.09
1.58–2.18
1.47–2.19
1.75–2.44
1.75–2.54
1.36–2.09
1.81–2.35
1.75–2.12
1.82–2.37
95% CI
Tumor number
128
112
117
199
646
117
186
91
96
102
170
189
121
131
99
114
213
565
205
n
3.352
3.14
2
2.932
2.732
3.12
2.90
3.31
2.92
2.97
3.22
2.87
3.27
2.94
3.04
3.21
2.80
3.311
3.00
2.84
Pred mean
2.93–3.76
2.81–3.47
2.62–3.25
2.34–3.12
2.82–3.41
2.42–3.38
2.90–3.73
2.38–3.45
2.44–3.50
2.71–3.73
2.45–3.30
2.86–3.68
2.46–3.42
2.57–3.50
2.69–3.73
2.32–3.29
2.95–3.68
2.72–3.28
2.46–3.21
95% CI
Tumor size (cm)
450 Smoking and bladder cancer clinical characteristics at diagnosis
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
501
164
Current smoker
439
Ever smoker
117
54
Former smoker
Current smoker
219
Ever smoker
168
Ever smoker
69.1–70.8 72.2–75.6
73.91
70.3–73.4
70.0
71.9
69.4–71.1
62.2–68.0
65.11
70.2
68.9–73.0
71.0
70.0–71.8
73.7–75.9
74.81
70.9
66.6–68.5
67.5
65.6 63.9–67.3
69.4–71.6
70.51 1
71.6–74.0
72.8
95% CI
165
855
216
807
52
111
859
426
596
158
490
384
n
Grade
2.361
2.22
2.30
2.23
2.481
2.23
2.23
2.30
2.20
2.28
2.26
2.20
Pred mean
2.23–2.49
2.15–2.29
2.19–2.42
2.16–2.30
2.26–2.70
2.07–2.39
2.16–2.30
2.21–2.39
2.12–2.28
2.15–2.42
2.17–2.34
2.10–2.29
95% CI
164
857
215
809
52
113
858
427
596
158
487
388
n
T stage
1.81
1.72
1.80
1.72
2.031
1.81
1.71
1.841
1.66
1.77
1.76
1.69
Pred mean
1.66–1.96
1.64–1.81
1.67–1.94
1.63–1.81
1.78–2.28
1.63–1.98
1.62–1.80
1.73–1.94
1.57–1.76
1.61–1.92
1.65–1.86
1.58–1.81
95% CI
151
801
201
754
50
110
794
408
546
153
452
358
n
2.10
1.99
2.16
1.96
1.91
2.22
1.97
2.04
1.97
2.10
1.95
2.01
Pred mean
1.79–2.41
1.83–2.14
1.89–2.43
1.81–2.12
1.39–2.43
1.86–2.58
1.81–2.14
1.84–2.25
1.79–2.15
1.80–2.41
1.75–2.14
1.80–2.23
95% CI
Tumor number
152
816
204
767
50
109
811
407
563
155
460
364
n
2.96
3.06
2.79
3.11
3.44
3.14
3.00
3.11
2.99
3.25
3.01
2.98
Pred mean
Epidemiology
2.53–3.38
2.79–3.33
2.41–3.18
2.83–3.39
2.77–4.11
2.65–3.62
2.72–3.28
2.79–3.44
2.69–3.30
2.83–3.67
2.71–3.31
2.67–3.30
95% CI
Tumor size (cm)
Pred: Predicted; CI: confidence interval. 1 Indicator variable is statistically significantly different from the baseline category “never smoker” (p20; mean difference: 3.7 years, 95% CI: 0.327.1) and a higher grade (mean difference: 0.29, 95% CI: 0.0420.54), while in pipe smokers this was associated with a higher tumor size (mean difference: 1.08 cm, 95% CI: 0.3421.81). In addition, we found a dose2response relationship in cigar smokers between a higher grade and less time since cessation (20 vs. >20 years since cessation; mean difference: 0.26, 95% CI: 0.0420.48) and we observed cigar smokers with a higher cumulative smoking amount to have a lower age at diagnosis (>15 kg vs. 15 kg; mean difference: 22.9 years, 95% CI: 25.3 to 20.5). Furthermore, in non-filter smokers a positive relation was observed between cumulative smoking amount and age at diagnosis (increment: 1.7 years per 50 kg, 95% CI: 0.922.6). Finally, a lower pipe smoking frequency was associated with a higher T stage at diagnosis (2 vs. >2 pipes per day; mean difference: 0.32, 95% CI: 0.0320.61).
might explain why in some cases no effects were found. The fact that we also found dose–response relationships and effects in separate types of smoking strengthens our belief of a genuine effect. It is already known that smoking increases the risk of BC,9 and an interaction effect with NAT2 acetylation status has been reported.38,39 Our results suggest that smoking might also promote tumor growth and invasiveness. Jiang et al.16 hypothesized that this could be related to the fact that different molecular pathways are involved in the pathogenesis of muscle-invasive and non-muscle-invasive tumors. Highgrade muscle-invasive tumors are characterized by mutations in the TP53 and RB1 tumor-suppressor genes, while lowgrade non-muscle-invasive tumors are characterized by activating FGFR3 gene mutations.40,41 Interestingly, smoking seems to be associated with TP53, but not with FGFR3 mutations.13 In addition, patients with BC with overexpression of p53, which is associated with the presence of a TP53 mutation, have higher progression and recurrence risks, and a lower survival.42,43 A detailed review of smoking and molecular pathways of BC is beyond the scope of this article. However, from what we already know we can speculate that smoking may further promote invasion and growth of BC by changing the micro-environment of tumors via overexpression of cyclooxygenase 2 (COX-2), which is a downstream target of p53 that regulates cell growth, apoptosis, and angiogenesis.44 Quite unexpectedly and unlike two previous studies,17,22 we found former smokers to have a significantly higher age at diagnosis than never smokers. In addition, we found inconsistencies in the associations between age at diagnosis and smoking status in different smoking types. We cannot think of an explanation of these effects and therefore believe that these results should be interpreted with caution. In females, dose–response relationships were observed between less time since cessation and a higher T stage and tumour size at diagnosis, which were not found in men. This implies that the effect of smoking could be stronger in females than in males. In accordance with our results, Jiang et al.16 observed that women have a higher risk of invasive BC than men who smoke comparable amounts. A possible explanation for these effects could be that estrogen promotes tumour progression through COX-2, as it has been observed that estrogen activates COX-2 in female mice.45 To our knowledge, this is the first study looking at the relation between smoking and clinical characteristics of BC at diagnosis in which several smoking features and different types of smoking were taken into account. Face-to-face interviews of patients by trained research nurses were performed, which is associated with a higher accuracy of self-reported smoking in comparison to self-administered questionnaires.46 Nevertheless, this study also has some limitations. First of all, a reasonably large percentage of patients was either missed, refused to participate, or did not provide smoking data which could have introduced selection bias. When we compared the age and gender distribution of all patients in the West
Interaction by gender
We found a number of significant interaction effects by gender (results not shown). First, the effect of smoking status on age at diagnosis was different, which was related to a lower average age at diagnosis of never smokers in males than in females (predicted means: 68.1 years, 95% CI: 66.3270.0 and 72.6 years, 95% CI: 70.2274.9, respectively), while the age of current and formers smokers was similar in stratified analyses. In addition, in women we found dose2response relationships between a lower time since cessation and both a higher T stage and tumor size at diagnosis (increments: 0.30, 95% CI: 0.1220.47 and 1.00 cm, 95% CI: 0.3821.62 per 10 years less time since cessation, respectively), which were not observed in men.
Epidemiology
Discussion In this study, we observed current smokers to be diagnosed with BC at an earlier age, having a higher T stage, larger tumor size, and a higher grade at diagnosis than never smokers; there was no difference in the number of tumors diagnosed. In addition, we found significant associations between on the one hand increased frequency, lower age at initiation, and less time since cessation of smoking, and on the other hand a lower age at diagnosis. Tumor size was found to increase with less time since cessation. The findings of this study are consistent with those of a recent Japanese study,17 in which current smokers were also found to be diagnosed at an earlier age, and having a higher T stage and grade, and larger tumor size at diagnosis compared to current non-smokers, and in this study no relation with tumor number was observed as well. Although there is limited literature on this topic, the existing literature is rather inconsistent. Some studies find smoking to be associated with at least one of the clinical factors we studied,10–23 but other studies found no association at all.4,24–32 Most studies which did not find any effect had a rather small study size, which
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
453
van Roekel et al.
Midlands during the period 2006–2010 with the patients included in analyses and recruited during the same period (n 5 1,058), we found that patients in this study were on average only 3 years younger, and that the percentage males was only 6% higher (data provided by the West Midlands Cancer Intelligence Unit: Incidence data of bladder cancer patients (ICD-10 C67) diagnosed 2005–2010 and resident in the West Midlands, 2011). However, we do not think this could have influenced our results as we have adjusted for both these factors. In addition, the results we obtained could be caused by smokers possibly being less likely to seek medical care and therefore presenting themselves later than non-smokers (lead time bias). Nevertheless, in a previous study in UC patients in the West Midlands, no relationship was found between smoking and the median delay from onset of symptoms to GP referral and we therefore think that this is not plausible.47 Finally, we performed a linear regression analysis with semicontinuous outcome variables (tumor size, number, grade, and stage). We addressed this by dichotomizing clinical outcome variables and performing mixed effects logistic regression analyses; the results of these analyses showed no substantial differences with the original analyses. Our findings may have some important implications. We think clinicians should be made aware of the possible relationship between smoking and severity of BC, as our results suggest smoking patients could have a higher chance of presenting with more malignant disease. It could mean that a
thorough resection, which includes detrusor muscle in the sample to determine muscle-invasiveness of the tumor, is even more important in smokers. As we and others found the effects of smoking on clinical presentation of BC to be less pronounced in former smokers, it is likely that cessation of smoking after BC diagnosis might lead to a better prognosis in terms of mortality and/or recurrences; although there is some evidence for this effect,24 large longitudinal studies are needed to establish this relationship more clearly.48 Currently, knowledge about the relation between smoking and BC is not well-known among patients.49 It is important that patients are well-informed about the potential risks of smoking and the possible relation with prognosis could strengthen this message.
Acknowledgements With thanks to all participating clinicians and their staff from within the West Midlands region for their continued support: P. Cooke, K. Jefferson, H. Krasnowski, J. Parkin, B.D. Sarmah, P. Ryan, R. Bhatt, M. Foster, K. Desai, L.A. Emtage, K.W. Kadow, C. Luscombe, S. Khwaja, and A. Makar. We would also like to recognize the invaluable contribution made by the BCPP nursing, research and administrative staff to the program: A. McGuire, C. Langford, C. Letchford, C. Slater, C. Taylor, D. Bird, G. Heritage, H. Shackleford, J. Sears, J. Maiden, J. Goodall, J. Allison, J. Hutton, J.Y. Robinson, K. Castro, L.R. Moore, L. Benson, M. Grant, R. Abid, S. Collins, T. Martin, and T. Coles. In addition, we thank R.C. Reulen and D. Nekeman for their substantial help with data management and the West Midlands Cancer Intelligence Unit for providing data on patients with bladder cancer in the West Midlands.
1.
Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. 2. Kirkali Z, Chan T, Manoharan M, et al. Bladder cancer: epidemiology, staging and grading, and diagnosis. Urology 2005;66:4–34. 3. Ploeg M, Aben KK, Kiemeney LA. The present and future burden of urinary bladder cancer in the world. World J Urol 2009;27:289–93. 4. Morrison AS, Proppe KH, Verhoek WG, et al. Histologic features of bladder cancer in Boston, USA, Manchester, UK, and Nagoya, Japan. Int J Cancer 1982;30:701–05. 5. Kaufman DS, Shipley WU, Feldman AS. Bladder cancer. Lancet 2009;374:239–49. 6. Kiemeney LA, Witjes JA, Verbeek AL, et al. The clinical epidemiology of superficial bladder cancer. Dutch South-East Cooperative Urological Group. Br J Cancer 1993;67:806–12. 7. Millan-Rodriguez F, Chechile-Toniolo G, Salvador-Bayarri J, et al. Multivariate analysis of the prognostic factors of primary superficial bladder cancer. J Urol 2000;163:73–8. 8. Zeegers MP, Kellen E, Buntinx F, et al. The association between smoking, beverage consumption, diet and bladder cancer: a systematic literature review. World J Urol 2004;21:392–401. 9. Zeegers MP, Tan FE, Dorant E, et al. The impact of characteristics of cigarette smoking on urinary tract cancer risk: a meta-analysis of epidemiologic studies. Cancer 2000;89:630–39. 10. Thompson IM, Peek M, Rodriguez FR. The impact of cigarette smoking on stage, grade and
11.
12.
13.
14.
15.
16.
17.
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
number of recurrences of transitional cell carcinoma of the bladder. J Urol 1987;137:401– 03. Garcia Mediero JM, Angulo CJ, Lujan GM, et al. [Smoking habit and superficial bladder tumours’ natural history]. Actas Urol Esp 2007;31:349–54. Mohseni MG, Zand S, Aghamir SM. Effect of smoking on prognostic factors of transitional cell carcinoma of the bladder. Urol J 2004;1:250–52. Wallerand H, Bakkar AA, de Medina SG, et al. Mutations in TP53, but not FGFR3, in urothelial cell carcinoma of the bladder are influenced by smoking: contribution of exogenous versus endogenous carcinogens. Carcinogenesis 2005;26:177–84. Ouerhani S, Rouissi K, Marrakchi R, et al. Do smoking and polymorphisms in xenobiotic metabolizing enzymes affect the histological stage and grade of bladder tumors? Bull Cancer 2009;96:E23–E29. Sturgeon SR, Hartge P, Silverman DT, et al. Associations between bladder cancer risk factors and tumor stage and grade at diagnosis. Epidemiology 1994;5:218–25. Jiang X, Castelao JE, Yuan JM, et al. Cigarette smoking and subtypes of bladder cancer. Int J Cancer 2012;130:896–901. Hinotsu S, Akaza H, Miki T, et al. Bladder cancer develops 6 years earlier in current smokers: analysis of bladder cancer registry data collected by the cancer registration committee of the Japanese Urological Association. Int J Urol 2009;16:64–9.
18. Khedhiri S, Stambouli N, Ouerhani S, et al. The impact of smoking and polymorphic enzymes of xenobiotic metabolism on the stage of bladder tumors: a generalized ordered logistic regression analysis. J Cancer Res Clin Oncol 2010;136:1111– 16. 19. Prout GR Jr, Wesley MN, Greenberg RS, et al. Bladder cancer: race differences in extent of disease at diagnosis. Cancer 2000;89:1349–58. 20. Brooks DR, Geller AC, Chang J, et al. Occupation, smoking, and the risk of high-grade invasive bladder cancer in Missouri. Am J Ind Med 1992;21:699–713. 21. Fleshner N, Garland J, Moadel A, et al. Influence of smoking status on the disease-related outcomes of patients with tobacco-associated superficial transitional cell carcinoma of the bladder. Cancer 1999;86:2337–45. 22. Michalek AM, Cummings KM, Pontes JE. Cigarette smoking, tumor recurrence, and survival from bladder cancer. Prev Med 1985;14:92–8. 23. Lockwood K. On the etiology of bladder tumors in Kobenhavn-Frederiksberg. An inquiry of 369 patients and 369 controls. Acta Pathol Microbiol Scand Suppl 1961;51(Suppl 145):1–166. 24. Chen CH, Shun CT, Huang KH, et al. Stopping smoking might reduce tumour recurrence in nonmuscle-invasive bladder cancer. BJU Int 2007;100:281–86. 25. Persad R, Fleming C, Chern HD, et al. Environmental procarcinogen hypothesis of bladder cancer in humans: dapsone hydroxylation
Epidemiology
References
454
as a susceptibility risk factor for aggressive bladder cancer. Urol Oncol 1997;3:18–26. 26. Jensen OM, Wahrendorf J, Blettner M, et al. The Copenhagen case-control study of bladder cancer: role of smoking in invasive and non-invasive bladder tumours. J Epidemiol Community Health 1987;41:30–6. 27. Lopez-Beltran A, Croghan GA, Croghan I, et al. Prognostic factors in survival of bladder cancer. Cancer 1992;70:799–807. 28. Nieder AM, MacKinnon JA, Fleming LE, et al. Bladder cancer clusters in Florida: identifying populations at risk. J Urol 2009;182:46–50. 29. Hayes RB, Friedell GH, Zahm SH, et al. Are the known bladder cancer risk-factors associated with more advanced bladder cancer? Cancer Causes Control 1993;4:157–62. 30. Serretta V, Altieri V, Morgia G, et al. Cigarette smoking and drinking water source: correlation with clinical features and pathology of superficial bladder carcinoma. Urol Int 2009;82:318–23. 31. Carpenter AA. Clinical experience with transitional cell carcinoma of the bladder with special reference to smoking. J Urol 1989;141:527–28.
33. Zeegers MP, Bryan RT, Langford C, et al. The West Midlands Bladder Cancer Prognosis Programme: rationale and design. BJU Int 2010;105:784–8. 34. Sobin LH, Wittekind C, eds. Union Internationale Contre le Cancer. TNM classification of malignant tumors, vol. 6. New York: Wiley-Liss, 2002. 35. Mostofi FK, Sobin LH, Torloni H. Histological typing of urinary bladder tumours. Geneva: World Health Organization, 1973. 36. Chade DC, Shariat SF, Godoy G, et al. Clinical outcomes of primary bladder carcinoma in situ in a contemporary series. J Urol 2010;184:74–80. 37. StataCorp LP. Stata/IC 11.0 statistical package for Windows, 2009. 38. Rothman N, Garcia-Closas M, Chatterjee N, et al. A multi-stage genome-wide association study of bladder cancer identifies multiple susceptibility loci. Nat Genet 2010;42:978–84. 39. Moore LE, Baris DR, Figueroa JD, et al. GSTM1 null and NAT2 slow acetylation genotypes, smoking intensity and bladder cancer risk: results from the New England bladder cancer study and NAT2 meta-analysis. Carcinogenesis 2011;32:182– 9. 40. Wu XR. Urothelial tumorigenesis: a tale of divergent pathways. Nat Rev Cancer 2005;5:713– 25. 41. Goebell PJ, Knowles MA. Bladder cancer or bladder cancers? Genetically distinct malignant
42.
43.
44.
45.
46.
47.
48.
49.
conditions of the urothelium. Urol Oncol 2010;28:409–28. Esrig D, Elmajian D, Groshen S, et al. Accumulation of nuclear p53 and tumor progression in bladder cancer. N Engl J Med 1994;331:1259–64. Masters JR, Vani UD, Grigor KM, et al. Can p53 staining be used to identify patients with aggressive superficial bladder cancer? J Pathol 2003;200:74–81. Huang RY, Chen GG. Cigarette smoking, cyclooxygenase-2 pathway and cancer. Biochim Biophys Acta 2011;1815:158–69. Egan KM, Lawson JA, Fries S, et al. COX-2derived prostacyclin confers atheroprotection on female mice. Science 2004;306:1954–7. Patrick DL, Cheadle A, Thompson DC, et al. The validity of self-reported smoking: a review and meta-analysis. Am J Public Health 1994;84:1086– 93. Wallace DM, Bryan RT, Dunn JA, et al. Delay and survival in bladder cancer. BJU Int 2002;89:868–78. Aveyard P, Adab P, Cheng KK, et al. Does smoking status influence the prognosis of bladder cancer? A systematic review. BJU Int 2002;90:228–39. Strope SA, Montie JE. The causal role of cigarette smoking in bladder cancer initiation and progression, and the role of urologists in smoking cessation. J Urol 2008;180:31–7; discussion 37.
Epidemiology
32. Raitanen MP, Nieminen P, Tammela TL. Impact of tumour grade, stage, number and size, and smoking and sex, on survival in patients with transitional cell carcinoma of the bladder. Br J Urol 1995;76:470–4.
Smoking and bladder cancer clinical characteristics at diagnosis
C 2013 UICC Int. J. Cancer: 133, 446–455 (2013) V
