Table 2.

Time-dependent HR (95% CI) of all-cause death for smoking after cancer diagnosis for all patients with cancer and for current smokers at cancer diagnosis, by cancer sites, The Shanghai Cohort Study 1986–2010

All patientsLung cancerStomach cancerColorectal cancerBladder cancerProstate cancerOther cancers
All patients with cancer
 No. of cases1,632288362248107132495
 No. of deaths (%)931 (57.0%)229 (79.5%)159 (43.9%)152 (61.3%)40 (37.4%)47 (35.6%)304 (61.4%)
 Person-years at risk6,9718381,8051,0776694452,137
 Average person-years4.32.95.04.36.33.44.3
 Average age at diagnosis68.868.169.667.969.273.367.9
 Median survival time (y)5.42.110.94.015.08.45.2
 HR (95%CI) time-dependent smoking statusa1.59 (1.36–1.86)1.92 (1.41–2.60)1.76 (1.11–2.79)1.65 (1.14–2.38)3.66 (1.56–8.60)1.94 (0.79–4.73)1.17 (0.87–1.58)
Patients with cancer who were current smokers at cancer diagnosis and survived until the next annual follow-up interview
 No. of cases7471841351144932233
 No. of deaths (%)491 (65.7)151 (82.1)72 (53.3%)73 (64.0%)23 (46.9%)15 (46.9%)157 (67.4%)
 Person-years at risk3,3315476475813111181,126
 Average person-years4.53.04.85.16.43.74.8
 Average age at diagnosis67.066.969.465.467.770.165.9
 Median survival time (y) for all patients4.12.17.34.112.46.34.3
  Median survival time (y) for quitters (r = 0)4.42.13.87.9N/Ac9.83.3
  Median survival time (y) for intermittent smokers (0<r<1.0)7.12.810.99.112.46.07.7
  Median survival time (y) for continued smokers (r = 1)2.11.73.82.12.911.12.1
 HR (95% CI) intermittent smokers vs. quittersb0.89 (0.70–1.12)0.91 (0.59–1.39)0.69 (0.38–1.27)1.35 (0.68–2.67)1.75 (0.28–11.01)2.88 (0.22–37.20)0.76 (0.49–1.17)
 HR (95% CI) continued smokers vs. quittersb1.76 (1.37–2.27)1.89 (1.18–3.02)1.36 (0.70–2.67)3.46 (1.69–7.10)17.29 (2.25–132.64)0.93 (0.08–11.24)1.49 (0.92–2.40)
 HR (95% CI) time-dependent smoking statusa1.79 (1.49–2.16)2.36 (1.63–3.42)1.63 (0.98–2.73)2.31 (1.40–3.81)2.95 (1.09–7.95)2.27 (0.40–12.80)1.34 (0.96–1.88)
  • aDerived from a yearly based time-dependent variable of smoking status (yes vs. no) after cancer diagnosis with the adjustment for age at diagnosis (continuous), level of education (primary school, middle school, or college or above), cumulative number of pack-year of smoking before diagnosis (continuous), and treatment for cancer: surgery (yes/no), chemotherapy (yes/no), and radiation therapy (yes/no). Cox models for all patients were additionally adjusted for cancer site.

  • bAll Cox models were adjusted for age at diagnosis (continuous), level of education (primary school, middle school, or college or above), cumulative number of pack-year of smoking before diagnosis (continuous), and treatment modalities: surgery (yes/no), chemotherapy (yes/no), and radiation therapy (yes/no). Cox models for all patients were additionally adjusted for cancer site.

  • cMedian survival (50% of the patients died) was not reached in this group of patients with bladder cancer.