EUF Vol. 1 No. 1

20 96

Studies reporting associations between smoking cessa-
tion and disease outcomes were also subjected to the
aforementioned criteria. In addition, we required that the
time period over which smoking cessation took place was
clearly delineated and that the comparator group contained
patients who were smoking at the time of diagnosis.
2.3. Data extraction
J.J.C. tabulated the data from all the studies included and
M.R. reviewed the tables for accuracy. For studies reporting
associations between smoking and UCB risk, we required
that odds ratios (ORs) were available with 95% confidence
intervals (CIs). When not available, ORs were calculated if
the data permitted this. For associations between smoking
and clinical outcomes, we recorded p values (or 95% CIs
if p values were not provided), along with hazard ratios
(HRs) or ORs if they were available. If both multivariate
and univariate analyses were performed for the same
smoking-endpoint association, we recorded the result of
the multivariate analysis. Statistical significance was
defined as p < 0.05.
3. Evidence synthesis
3.1. Bladder cancer risk
We estimate that several hundred case-control and cohort
studies have been conducted to estimate the risk of UCB
attributable to smoking. An analysis performed in 2011 of
467 528 men and women found that former and current
smokers had two- and fourfold increases, respectively, in
UCB risk relative to never smokers . The population risk
of bladder cancer attributable to smoking was approxi-
mately 50% for both men and women.
This reviewsummarizes the impact of smokingonbladder
cancer risk in studies published since 2011. Among the eight
case-control studies selected ) , UCB risk
was two- to fourfold higher for current smokers, and up
to threefold higher for former smokers. Risk estimates
were similar for men and women across studies. One study
found that smoking was more strongly associated with the
development of MIBC than with NMIBC . While smoking
cessation attenuates UCB risk, former smokers are still
approximately twice as likely to develop the disease 20 yr
after quitting
Overall, the most current risk estimates are consistent
with those of Freedman et al from 2011. Interestingly,
changes in the composition of cigarette smoke that might be
more likely to induce bladder carcinogenesis could explain
the nearly unchanged incidenceinUCB despite the decreased
prevalence of smoking. Although the molecular mechanisms
behind bladder carcinogenesis remain incompletely eluci-
dated, several have been proposed . In addition,
investigating specific genetic susceptibilities may help in
understanding smoking-related risk profiles for individual
patients
It is important to note that case-control studies such
as those reported here are limited by their retrospective
nature and demographic and clinical differences often
unaccounted for between case and control populations.
However, they are often adequate for obtaining risk
estimates, especially in light of the high cost of performing
a prospective cohort study.
Table 1 – Selected studies reporting associations of smoking and risk of urothelial carcinoma of the bladder
Study Years Cases (n) Controls (n) Mean or median
age (years)
Male (%) Smoking
categor
Estimate
Alguacil et al. (2011) 1998-2001 712 611 NR 89 Ever OR=2.4y
Former OR=1.8y
Current OR=3.7y
Jiang et al. (2012)z 1987-1996 1586 725 56 78 Ever OR=2.2 (1.8-2.8)
Former OR=1.7 (1.3-2.1)
Current OR=3.2 (2.5-4.1)
Zheng et al. (2012)z 2006-2010 765 1651 58 100 Ever OR=2.0y
Former OR=1.2 (0.9-1.7)
Current OR=2.1 (1.7-2.6)
Ferreccio et al. (2013) 2007-2010 232 640 NR 69 Ever OR=1.6 (1.1-2.2)
Erdurak et al. (2014)z 2011 173 282 6 6 NR Ever OR=3.2y
Former OR=3.3 (1.8-5.8)
Current OR=3.0y
Moura et al. (2014) 1998-2011 4312 26971 NR 5 Ever (male) OR=3.0 (2.8-3.2)
Ever (female) OR=3.0 (2.6-3.4)
Welty et al. (2014)z 2000-2008 378 76055 NR 48 Ever OR=2.8y
Former HR=2.0 (1.6-2.6)
Current HR=3.8 (2.7-5.4)
Wu et al. (2014) 2002-2009 261 672 61 67 Ever OR=2.3 (1.6-3.3)
NR=not reported; OR=odds ratio; HR=hazard ratio.
* All risk estimates are relative to never smokers.
y Calculated based on available raw data.
z Refer to full text for smoking quantity and duration data.
§ Applies to cases only.
k Applies to controls only.
E U R O P E A N U R O L O G Y F O C U S 1 ( 2 0 1 5) 1 7 – 2 7 20