Table 3: Some epidemiological studies on prostate cancer.

ExposureStudy populationSMR, relative risk or odds ratio (95% confidence interval)Main strengthsMain limitationsReference

Ecological studies

ArsenicMembers of the Mormon church from Millard County, UtahSMR compared to the general population of Utah: 1.45 (1.07–1.91)Quite good characterisation of exposure. Cohort population with restrictive life styleNo clear dose-effect relationship; no correction for confounding[105]
ArsenicResidents of 314 of the 361 districts in TaiwanMultivariate-adjusted regression coefficient for age-adjusted mortality per 100,000 person-years for every 0.1 ppm increase in arseniclevel of well water: 0.5 (SE: 0.2) P < 0.05Quite good characterisation of exposureIncomplete correction for confounding[106]

Case-control studies

Polychlorinated Biphenyl congener 153OR for greater than median adipose tissue concentration: 3.15 (1.04–9.54)
Trans-chlordaneOR for greater than median adipose tissue concentration: 3.49 (1.08–11.2)
Polychlorinated Biphenyl congener 153General population of Örebro County, SwedenOR for prostate cancer with PSA > median, for greater than median adipose tissue concentration: 30.3 (3.24–284)Highly relevant parameter of personal internal exposure. Adequate correction for confounding.Small number of subjects [107]
Trans-chlordaneOR for prostate cancer with PSA > median, for greater than median adipose tissue concentration: 11.0 (1.87–64.9) Small number of subjects. The biological relevance of the arbitrary cut-off point of PSA is unclear
HexachlorobenzeneOR for prostate cancer with PSA > median for greater than median adipose tissue concentration: 9.84 (1.99–48.5),

Cohort studies

Polychlorinated BiphenylsElectrical Capacitor Manufacturing Workers in New York and MassachusettsProstate cancer mortality increased with cumulative exposure to RR = 6.05 for the highest quartile (trend 𝑃 value = 0.0001)Quite elaborate characterisation of exposureLimited correction for confounding[108]

Beta-hexachlorocyclohexaneORs for the 2° and 3° tertiles of detectable values were 1.46 (0.52–4.13) and 3.36 (1.24–9.10) (P for trend = 0.02)
Trans-nonachlorParticipants in the U.S. National Health and Nutrition Examination Survey (1999-2000)ORs for the 2° and 3° tertiles of detectable values were 5.84 (1.06–32.2) and 14.1 (2.55–77.9) (P for trend = 0.002)Measurement of personal internal exposure. Adequate correction for confoundingSelf-reported cancer diagnosis. Only 65 cases[109]
DieldrinORs for the 2° and 3° tertiles of detectable values were 1.06 (0.30–3.73) and 2.74 (1.01–7.49) (P for trend = 0.04)

Prospective cohort studies

Application of pesticidesParticipants in the Agricultural Health Study in Iowa and North CarolinaOR: 1.14 (1.05, 1.24) compared to male populations of Iowa and North CarolinaParticipants: 55,332 male pesticide applicators Elaborate characterization of external exposure. Adequate correction for confoundingShort followup period[110]

Meta-analysis

Application of pesticides22 studies (1986–2003)RR: 1.24 (1.06–1.45) RR's higher in North America than in EuropeRR's homogenous in same region No biasLimited info on specific chemicals[111]
Manufacturing of pesticides18 studies (1984–2004)RR: 1.28 (1.05–1.58). Consistent increase for all chemical classes, significant for phenoxy herbicides No publication biasLimited info on specific chemicals[112]