PROSTATE CANCER AND SMOKING
Kym Hickey MBBS, MPH Repatriation Medical Authority, Australia
Prostate Cancer and Smoking • Master’s thesis on this topic • Review article on prostate cancer and smoking (Hickey et al 2001, Epidemiologic Reviews 23(1): 115-125)
Learning Objectives • Be aware of data concerning prostate cancer and smoking • Understand epidemiological criteria used to assess causality
Performance Objectives • How to assess new evidence on prostate cancer & smoking • How to apply epidemiological criteria to other questions of causality
Why Prostate Cancer & Smoking • Role of smoking in prostate cancer unclear • Important public health issue as both smoking and prostate cancer are prevalent in community
Identification of Studies • Medline search • Review of bibliographies of identified studies • Conference papers
Assessment of Study Quality • Study design / selection issues • Measurement of prostate cancer & smoking • Control of confounding • Statistical issues
Results • 65 controlled studies reported results for pc & smoking • Results for both current & past smoking varied from protective for pc to increasing the risk of pc
Study 1 2 3 4 5 6 7
No. cases 1369 69 71 180 238 198 707
RR, CI 1.04, 0.85-1.27 0.56, 0.36-0.83 2.2, 1.2-4.4 0.49, 0.16-1.57 1.46, 1.07-1.94 1.0, 0.6-1.6 1.1, 0.9-1.3
Study 8 9 10 11 12 13 14
No. Cases 174 220 2,368 209 109 138 406
RR, CI 0.87, 0.61-1.23 1.08, 0.90-1.30 1.11, 1.01-1.23 1.11, 0.90-1.36 0.82, 0.57-1.14 0.9, 0.40-1.73 1.00, 0.71-1.39
Study No. Cases RR, CI 15 16 17 54 166 243 1.3, 0.61-2.79 0.76, 0.51-1.14 1.1, 1.0-1.3
Study 1 2 3 4 5 6 7
No. Cases 103 709 1748 134 147 4607 569
RR, CI 1.58, 0.81-3.10 1.26, 1.06-1.50 1.34, 1.16-1.56 1.75, 1.37-2.19 1.1, 0.7-1.5 1.18, 1.09-1.28 0.99, 0.87-1.34
Study 8 9 10 11 12 13
No. Cases 149 32 826 319 193 30
RR, CI 2.0, 1.1-3.7 1.83, 1.01-3.05 1.31, 1.13-1.52 1.02, 0.81-1.28 0.93, 0.72-1.18 1.38, 0.67-2.85
Population-based Case-Control • 10 out of 15 p-b case-control studies found no association between current or ever smoking & pc as did 4 out of 5 case-control studies that used hospital cases and population controls
Hospital-based Case-Control • 12 out of 16 hospital-based case-control studies found no association between current or ever smoking and prostate cancer
Methodological Criticisms I • ? Differential measurement of prostate cancer in mortality cohorts • ? Differential screening of prostate cancer in mortality cohorts
Methodological Criticisms II • ? Confounding by dietary fat • ? Differential treatment of prostate cancer by smoking status
Causal Criteria - Consistency • There was inconsistency between the results of the mortality cohorts (positive association between pc and current smoking) and incidence cohorts (no association)
Causal Criteria – Strength of Association
• Association observed between current smoking and fatal prostate cancer was weak – about a 30% increase in risk
Causal Criteria – Dose Response • 2 cohorts found a D-R relation between current smoking and risk of incident pc • 2 cohorts found a D-R relation between current or recent smoking and risk of fatal pc
Prostate Cancer More Aggressive in Smokers ? • Does the increased risk of advanced incident or fatal pc in recent or current smokers mean that pc follows a more aggressive course in smokers compared to nonsmokers ?
Causal Criteria – Biological Plausibility • Few studies on the effects of smoking on prostate cancer in animals • Nitrosamines and pc in rats, tobacco smoke and pc tumor cell fraction in rats
Causal Criteria – Biological Plausibility • Various mechanisms proposed – cadmium, male hormones, genetic mutations, immune dysfunction – but none clearly established
Male Hormones • Probably most widely discussed of the proposed mechanisms • Rests on two assumptions: smoking is anti-estrogenic and male sex hormones increase risk of prostate cancer but ??
Male Hormones • No diff. in mean serum level of T, free T, dihydroT, estrone, estradiol, androstenedione, DHEAS, between pc cases and controls in meta-analysis of 10 prospective studies (Eaton et al 1999 Br J Cancer 80: 930-4)
Conclusion • Positive association observed between smoking and fatal prostate cancer could be due to a methodological bias or prostate cancer may be more aggressive in smokers