For facial acne scarring, the odds ratios for overall prostate cancer was 0.66 (95% CI 0.371.17) for the obvious compared with the not obvious. 16 (i.e., low concentration) to 1 1 : 65 536 (i.e., high concentration; median value=1 : 1024). The odds ratio for prostate cancer associated with titres at or above the median value was 0.73 (95% CI 0.580.91,P=0.005). The association appeared to be particularly strong for advanced prostate cancer (AJCC Stage grouping IIIIV) for which the odds ratio was 0.59 (95% CI NQO1 substrate 0.430.81,P=0.001) but there was insufficient evidence that this association differed by tumour stage (p heterogeneity=0.07). == Conclusion: == These results need to be confirmed in prospective studies but they are consistent with the hypothesis thatP. acneshas a role in prostate cancer. Keywords:prostate cancer, risk factor, acne,P. acnes, casecontrol study Studies investigating a possible association between acne and prostate cancer risk have reported inconsistent results. These include a decreased risk associated with facial acne scarring in our Risk Factors for Prostate Cancer Study (RFPCS) (Gileset al, 2003), no significant associations in other two studies (Lightfootet al, 2004;Galobardeset al, 2005), and NQO1 substrate an increased risk associated with long-term use of tetracyclines that was assumed to indicate treatment of severe acne (Sutcliffeet al, 2007). Most of these studies (Gileset al, 2003;Lightfootet al, 2004;Galobardeset al, 2005) were investigating acne as a marker of androgen activity in puberty, under the hypothesis that Mouse monoclonal to LAMB1 higher androgen levels might predispose to both acne and development of prostate cancer. Subsequent research has indicated another possible connection between acne and NQO1 substrate prostate cancer. Recent reports onPropionibacterium acnes(P. acnes), a bacterium associated with acne that normally lives on the skin and thrives in blocked follicles, suggest that this bacterium is usually prevalent in the prostate, it is associated with acute and chronic prostatic inflammation, and it might have a role in prostate carcinogenesis (Cohenet al, 2005;Alexeyevet al, 2006). Our group (RJC and BAS) culturedP. acnesfrom one-third of a series of radical prostatectomy specimens and NQO1 substrate found thatP. acnes-positive specimens were more likely to contain foci of acute or chronic inflammation thanP. acnes-negative specimens (Cohenet al, 2005). Consistent with these findingsAlexeyevet al(2006)investigating bacterial DNA in prostatic tissue from patients with benign prostatic hyperplasia found thatP. acneswas the NQO1 substrate most prevalent bacterium and that its prevalence was higher in samples from patients subsequently diagnosed with prostate cancer. These observations led to hypothesise thatP. acnesinfection in the prostate gland is usually positively associated with prostate cancer risk, with acne considered as a marker of increased immune sensitivity toP. acnesinfection. The aim of our study was to test whether plasma concentration ofP. acnesantibodies, measured by an ELISA assay we recently developed (Shannonet al, 2008), was associated with prostate cancer risk. We tested this hypothesis using the RFPCS, a population-based casecontrol study. Although the majority of studies discussed above have suggested positive associations between prostate cancer and either acne in puberty orP. acnesinfection in the prostate gland, our RFPCS cohort previously showed an inverse association between facial scarring and prostate cancer risk (Gileset al, 2003), therefore a secondary aim of this study was to investigate this apparently conflicting result. == Materials and methods == == Study design and subjects == The RFPCS study was conducted in 19941998 in Melbourne and Perth, Australia. Details of this study were published previously (Gileset al, 2001;Severiet al, 2003). Cases were men aged less than 70 years at diagnosis with histologically confirmed adenocarcinoma of the prostate and notified to the population-based Cancer Registries in the two States during the period between 1994 and 1997. Tumours were staged according to the TNM system and categorised into American Joint Committee on Cancer stage groupings (IIV) (AJCC, 2002). Cases were classified into two groups according to tumour differentiation: moderate-grade cases (i.e., Gleason score 57) and high-grade cases (i.e., Gleason score 810). Men with well-differentiated tumours or Gleason score less than 5 were excluded. Controls were randomly selected from the state electoral rolls and frequency matched to cases by 5-year age group and city of residence at the time of selection in a ratio of one control per case. Face to face interviews were conducted usually at the men’s home using questionnaires to capture various exposures including demographic factors, medical history, anthropometry, radiation exposures, alcohol consumption, smoking, growth and development, medical and surgical procedures related to the reproductive organs, diet captured using a detailed food frequency questionnaire, lifetime history of sexual activity, and family.