Abstract
Background
The determinants of prostate cancer – aside from established but non-modifiable risk factors of increased age, black ethnicity and a positive family history – are poorly understood.
Methods
We examined the association of a series of baseline socioeconomic, behavioural and metabolic characteristics with the risk of prostate cancer mortality in a 40 year follow-up of the original Whitehall cohort study which gave rise to 578 prostate cancer deaths in 17,934 men.
Results
After adjustment for a series of baseline covariates, results from proportional hazards regression analyses indicated that marital status (hazard ratio; 95% confidence interval: widowed/divorced vs. married: 1.44; 0.95, 2.18), raised blood cholesterol (tertile 3 vs. 1: 1.35; 1.11, 1.65), and increased physical stature (tertile 3 vs. 1: 1.37; 1.09, 1.74) were associated with death from prostate cancer, although statistical significance at conventional levels was not apparent in all analyses. There was no evidence that physical activity, smoking habit, socio-economic status, either component of blood pressure, nor diabetes predicted the risk of death from this malignancy herein.
Conclusions
In the present study there was a suggestion that marital status, blood cholesterol and height were risk indices for death from prostate cancer.
Keywords: epidemiology, risk factors, prostate cancer
Introduction
Although carcinoma of the prostate is the most common malignancy in men in western societies,1 its aetiology – beyond the established but non-modifiable risk indices of increased age, black ethnicity, and a positive family history2 – is very poorly understood. The observation that men migrating from areas of low to high prostate cancer incidence adopt the disease rates of their new country within only one generation3 strongly implicates a key role for environmental factors in the occurrence of this neoplasm.
To date, and despite having a priori biological plausibility, results from population-based studies exploring behavioural risk factors for both fatal and non-fatal prostate cancer, such as physical inactivity4;5 and smoking,6 are highly inconsistent. Recently, investigators have also examined the capacity of the metabolic correlates of these behaviours such as blood pressure,7 serum cholesterol,8–10 obesity2 and diabetes11 to predict future prostate cancer. Again, findings from a modest series of cohort studies, many of which are underpowered, are discordant or suggest null associations.
Four decades of mortality surveillance of men in the original Whitehall study12 have given rise to a large number of deaths ascribed to prostate cancer. This allows us address these issues of data paucity, suboptimal study size, and inconsistent results by examining the relation, if any, of a comprehensive range of socioeconomic, physiological, and behavioural risk factors with prostate cancer mortality.
Methods
In the Whitehall study, data were collected on 19,019 male, non-industrial, government employees aged from 40 to 69 years when examined between 1967 and 1970 in London (UK), representing a 77% response.12 This involved the completion of a study questionnaire and participation in a medical examination, both of which have been described in detail elsewhere.12 In brief, the questionnaire included enquiries regarding civil service employment grade (an indicator of socio-economic status13), smoking habits,14 marital status,15 and physical activity. During the course of baseline data collection, the physical activity questions were modified such that 12889 men responded to enquiries about travel activity16 and 6954 were asked about their leisure activity.17 For travel activity, cohort members indicated how many minutes they walked or bicycled to work each day (single journey). For leisure activity, study members were asked to specify their hobbies or sports; they were then classified as inactive, moderately active, or active during leisure.24
Blood pressure,18 height,19 weight,20 pulmonary function (indexed by forced expiratory volume in one second, FEV1),21 and, following an overnight fast, plasma cholesterol22 and post-challenge blood glucose concentration,23 were all determined using standard protocols.12 Body mass index (BMI) was computed using the standard formulae (weight[kg]/height2[m2]).
In this study, a 10% representative subgroup of 1618 participants completed a three day (Saturday/Sunday and two week days) dietary record during the period of baseline examination.24 The standard dietary records, listing unweighed but semi-quantitative descriptions of all food and drink consumed, were coded with use of a comprehensive food table. Data were extracted on dietary characteristics which, a priori, we had reason to anticipate an association with carcinoma of the prostate:2 intake of alcohol, fat (including animal, dairy, saturated, monounsaturated, polyunsaturated), calcium, linoleic acid, dietary cholesterol, retinol, and beta-carotene.
A total of 18,880 men (99.3% of participants in baseline survey) were traced using the UK National Health Service Central Registry and prostate cancer deaths were ascertained from death certificates (coded as ICD 8/9: 185, ICD 10: C61). The present analyses are based on 17,934 men with complete data; the outcome of interest throughout our analyses was prostate cancer mortality.
The person years of follow-up for each man was partitioned by age at risk using 5-years age groups. To summarise the relationship between each risk factor and prostate cancer mortality we used Cox's proportional hazards regression model to produce hazard ratios (HRs) with accompanying 95% confidence intervals. Censoring was made for age at death, age at loss to follow-up, or age at end of follow-up – whichever came first. These analyses were first adjusted for age at risk and then fully-adjusted for all potential confounding and mediating factors. Tests for trend across physical activity levels were computed by fitting a linear trend term.
Results
During a maximum of 40 years of follow-up there were 13,948 deaths from all-causes in the present analytical sample, 578 (4.2 %) of which were ascribed to prostate cancer. As expected, age was strongly related to death from carcinoma of the prostate: HR per 10 year increase; 95% confidence interval: 3.52; 3.21, 3.85). In table 1 we show the associations between prostate cancer and a series of baseline physiological characteristics. Although categories of body mass index were not associated with prostate mortality risk, there was a higher rate of mortality from this carcinoma apparent in men with elevated plasma cholesterol levels, an effect that was robust to statistical adjustment for a range of covariates. Neither component of blood pressure nor diabetes offered any predictive capacity for prostate cancer. Men with increased pulmonary function, as indexed by FEV1, and those who were taller, experienced a raised risk of prostate cancer. After mutual control, the impact of FEV1 was greatly reduced, while the effect for taller men held.
Table 1.
Physiological risk factors for prostate cancer mortality in the original Whitehall Study (578 prostate cancer deaths among 17934 men)
| Risk factor, units (mean [SD]) |
No. deaths / Person years (1000s) |
Hazard ratio (95% confidence intervals) |
||
|---|---|---|---|---|
| Age-adjusted | Multiple- adjustmenta |
|||
| Body mass index, kg/m2 | Underweight (<18.5) | 5 / 5.0 | 1.00 (0.41, 2.42) | 1.00 (0.41, 2.44) |
| (24.75 [ 2.96]) | Normal weight (18.5–24.9) | 323 / 260.9 | 1.0 (ref) | 1.0 (ref) |
| Overweight (25.0–29.9) | 233 / 189.1 | 1.03 (0.93, 1.15) | 1.04 (0.93, 1.16) | |
| Obese (30+) | 17 / 17.1 | 0.93 (0.57, 1.53) | 1.00 (0.61, 1.65) | |
| Per 1 SD increase | 1.02 (0.94, 1.12) | 1.04 (0.94, 1.14) | ||
| P-value for trend | 0.74 | 0.91 | ||
| Plasma cholesterol (mmol/l) | Tertile 1 (<4.54) | 176 / 163.0 | 1.0 | 1.0 |
| ( 5.10 [1.21]) | Tertile 2 (4.54–5.54) | 177 / 151.5 | 1.08 (0.88, 1.33) | 1.07 (0.87, 1.32) |
| Tertile 3 (>5.54) | 225 / 157.8 | 1.36 (1.12, 1.66) | 1.35 (1.11, 1.65) | |
| Per 1 SD increase | 1.12 (1.04, 1.22) | 1.12 (1.03, 1.22) | ||
| P-value for trend | 0.002 | 0.003 | ||
| Diastolic blood pressure (mmHg) | Tertile 1 (<78) | 206 / 159.2 | 1.0 | 1.0 |
| (84.5 [13.8]) | Tertile 2 (78–88) | 200 / 168.8 | 0.94 (0.77, 1.14) | 0.91 (0.75, 1.11) |
| Tertile 3 (>88) | 172 / 144.2 | 1.00 (0.82, 1.22) | 0.98 (0.79, 1.21) | |
| Per 1 SD increase | 1.00 (0.91, 1.09) | 0.99 (0.90, 1.08) | ||
| P-value for trend | 0.96 | 0.82 | ||
| Systolic blood pressure (mmHg) | Tertile 1 (<126) | 217 / 169.3 | 1.0 | 1.0 |
| (136.4 [21.2]) | Tertile 2 (126–142) | 200 / 165.3 | 0.94 (0.78, 1.14) | 0.94 (0.77, 1.13) |
| Tertile 3 (>142) | 161 / 137.6 | 0.92 (0.75, 1.13) | 0.93 (0.75, 1.15) | |
| Per 1 SD increase | 0.96 (0.88, 1.05) | 0.96 (0.88, 1.06) | ||
| P-value for trend | 0.40 | 0.47 | ||
| Diabetesb | Normoglycaemia | 548 / 445.1 | 1.0 | 1.0 |
| (4.17 [0.71]) | IGT | 29 / 22.7 | 1.11 (0.76, 1.61) | 1.16 (0.80, 1.70) |
| Diabetes | 1 / 4.4 | 0.22 (0.03, 1.56) | 0.24 (0.03, 1.73) | |
| P-value for trend | 0.50 | 0.72 | ||
| Lung function (FEV1, L) | Tertile 1 (<2.9) | 131 / 117.8 | 1.0 | 1.0 |
| (3.14 [0.72]) | Tertile 2 (2.9–3.4) | 232 / 161.4 | 1.36 (1.10, 1.68) | 1.25 (1.00, 1.56) |
| Tertile 3 (>3.4) | 215 / 193.0 | 1.18 (0.95, 1.47) | 1.02 (0.80, 1.31) | |
| Per 1 SD increase | 1.08 (0.99, 1.18) | 1.01 (0.91, 1.12) | ||
| P-value for trend | 0.23 | 0.89 | ||
| Height (m) | Tertile 1 (<1.73) | 133 / 123.0 | 1.0 | 1.0 |
| (1.758 [0.068]) | Tertile 2 (1.73–1.78) | 215 / 179.1 | 1.19 (0.96, 1.48) | 1.19 (0.95, 1.48) |
| Tertile 3 (>1.78) | 230 / 170.1 | 1.39 (1.12, 1.72) | 1.37 (1.09, 1.74) | |
| Per 1 SD increase | 1.12 (1.03, 1.22) | 1.11 (1.01, 1.22) | ||
| P-value for trend | 0.002 | 0.008 | ||
multiple adjustment is adjustment for BMI, plasma cholesterol, physical activity, socio-economic status, diabetes/blood glucose, marital status, FEV1, height, age at risk, smoking, and diastolic and systolic blood pressure (estimates for components of blood pressure are not mutually adjusted).
diabetes was defined as blood glucose of ≥11.1 mmol/l (≥200 mg/100ml) and/or a positive response to the questionnaire enquiry “are you, or have you been, diabetic?”; and impaired glucose tolerance (IGT) as 5.4 to 11.0 mmol/l (96 to 199 mg/100ml). All other men were denoted normoglycaemic. The Mean (SD) blood glucose reported here is for the 17699 men who were classified as either normoglycaemic or as having IGT.
In table 2 we depict the association of behavioural and psychosocial risk factors with death from carcinoma of the prostate. There was no suggestion that smoking or either index of physical activity was associated with this malignancy. There was also no evidence of a relation between socioeconomic status and prostate cancer, however, there was an association with marital status such that being single at baseline conferred protection, while being widowed or divorced was associated with elevated risk compared to married men.
Table 2.
Behavioural and psychosocial risk factors for prostate cancer mortality in the original Whitehall Study (578 prostate cancer deaths among 17934 men)
| Risk factor, units | Percent | No. deaths / Person years (1000s) |
Hazard ratio (95% confidence intervals) |
||
|---|---|---|---|---|---|
| Age-adjusted | Multiple- adjustment* |
||||
| Leisure activity | None / inactive | 35.6 | 55 / 57.0 | 1.0 (ref) | 1.0 (ref) |
| Moderately active | 41.1 | 96 / 71.5 | 1.30 (0.93, 1.81) | 1.24 (0.88, 1.73) | |
| Active | 23.3 | 53 / 43.3 | 1.19 (0.81, 1.73) | 1.12 (0.76, 1.64) | |
| P-value for trend | 0.35 | 0.56 | |||
| Travel activity (mins./day) | 0 – 9 | 19.6 | 63 / 61.5 | 1.0 | 1.0 |
| 10 – 19 | 44.8 | 182 / 144.1 | 1.23 (0.92, 1.64) | 1.24 (0.93, 1.66) | |
| 20 – 29 | 25.3 | 108 / 83.7 | 1.23 (0.90, 1.67) | 1.26 (0.92, 1.72) | |
| 30 – 39 | 8.3 | 36 / 27.0 | 1.28 (0.85, 1.92) | 1.30 (0.86, 1.97) | |
| 40+ | 2.0 | 11 / 6.2 | 1.60 (0.84, 3.03) | 1.65 (0.87, 3.15) | |
| P-value for trend | 0.14 | 0.10 | |||
| Socio-economic position | Administrative (highest) | 5.1 | 35 / 27.3 | 0.98 (0.70, 1.39) | 0.93 (0.65, 1.31) |
| Professional/executive | 65.1 | 401 / 325.8 | 1.0 | 1.0 | |
| Clerical | 15.8 | 73 / 62.9 | 0.90 (0.70, 1.16) | 0.96 (0.74, 1.25) | |
| Other | 9.4 | 35 / 32.7 | 0.78 (0.55, 1.10) | 0.83 (0.58, 1.19) | |
| BC & DS‡ | 4.7 | 34 / 23.5 | 1.20 (0.84, 1.69) | 1.14 (0.80, 1.62) | |
| P-value for trend+ | 0.15 | 0.50 | |||
| Marital status | Married | 88.1 | 527 / 420.0 | 1.0 | 1.0 |
| Single | 8.5 | 27 / 38.9 | 0.67 (0.45, 0.98) | 0.70 (0.47, 1.03) | |
| Widowed / divorced | 3.4 | 24 / 13.3 | 1.37 (0.91, 2.06) | 1.44 (0.95, 2.18) | |
| P-value for heterogeneity | 0.03 | 0.03 | |||
| Smoking habit | Never smoked | 18.3 | 123 / 99.5 | 1.0 | 1.0 |
| Ex-smoker | 36.6 | 225 / 179.8 | 0.95 (0.76, 1.18) | 0.94 (0.76, 1.18) | |
| Pipe / cigar smoker | 3.5 | 27 / 17.0 | 1.23 (0.81, 1.87) | 1.24 (0.82, 1.88) | |
| Current cigarette smoker | 41.6 | 203 / 175.8 | 1.10 (0.88, 1.37) | 1.14 (0.91, 1.44) | |
| P-value for trend | 0.20 | 0.09 | |||
British Council & Diplomatic Service are included as a separate category as they are not directly comparable with other civil service departments.
British Council & Diplomatic Service are excluded from test for trend.
multiple adjustment is adjustment for BMI, plasma cholesterol, physical activity, socio-economic status, diabetes/blood glucose, marital status, FEV1, height, age at risk, smoking, and diastolic and systolic blood pressure (estimates for components of blood pressure are not mutually adjusted).
Finally, in the 1618 men with data on dietary characteristics there were 60 deaths from prostate cancer during follow-up (Table 3). There was no strong suggestion of an association between any of the selected dietary characteristics – alcohol, fat (including animal, dairy, saturated, monounsaturated, polyunsaturated), calcium, linoleic acid, dietary cholesterol and beta-carotene – and death from this malignancy.
Table 3.
Dietary risk factors for prostate cancer mortality in the original Whitehall Study (60 prostate cancer deaths among 1618 men)
| Risk factor, units (mean [SD]) |
No. deaths / Person years (1000s) |
Hazard ratio (95% confidence intervals) | ||
|---|---|---|---|---|
| Age-adjusted | Multiple-adjustmenta | |||
| Alcohol, g/day | None | 17 / 13.7 | 1.0 (ref) | 1.0 (ref) |
| (14.3 [22.2]) | 1 – 9 | 22 / 12.3 | 1.31 (0.70, 2.47) | 1.55 (0.81, 3.00) |
| 10 – 34 | 14 / 11.6 | 0.80 (0.39, 1.59) | 0.96 (0.46, 1.98) | |
| 35+ | 7 / 5.4 | 1.11 (0.46, 2.67) | 1.52 (0.60, 3.85) | |
| P-value for trend | - | 0.70 | 0.74 | |
| Fat, g | Tertile 1 (<98.3) | 25 / 13.6 | 1.0 | 1.0 |
| (110.5 [ 29.2]) | Tertile 2 (98.3–120.8) | 19 / 14.5 | 0.73 (0.40, 1.33) | 0.77 (0.42, 1.42) |
| Tertile 3 (>120.8) | 16 / 14.9 | 0.58 (0.31, 1.08) | 0.57 (0.30, 1.09) | |
| Per 1 SD increase | - | 0.87 (0.66, 1.15) | 0.85 (0.64, 1.13) | |
| P-value for trend | - | 0.08 | 0.09 | |
| Animal fat, g | Tertile 1 (<30.7) | 27 / 13.8 | 1.0 | 1.0 |
| (38.4 [15.7]) | Tertile 2 (30.7–43.6) | 16 / 14.5 | 0.51 (0.28, 0.95) | 0.49 (0.26, 0.94) |
| Tertile 3 (>43.6) | 17 / 14.7 | 0.61 (0.33, 1.12) | 0.53 (0.28, 1.01) | |
| Per 1 SD increase | - | 0.85 (0.65, 1.11) | 0.79 (0.60, 1.05) | |
| P-value for trend | - | 0.09 | 0.05 | |
| Dairy fat, g | Tertile 1 (<42.3) | 19 / 14.2 | 1.0 | 1.0 |
| (50.9 [18.4]) | Tertile 2 (42.3–56.7) | 24 / 14.2 | 1.34 (0.73, 2.45) | 1.40 (0.76, 2.57) |
| Tertile 3 (>56.7) | 17 / 14.6 | 0.88 (0.46, 1.70) | 0.94 (0.48, 1.85) | |
| Per 1 SD increase | - | 0.92 (0.71, 1.19) | 0.94 (0.72, 1.22) | |
| P-value for trend | - | 0.72 | 0.91 | |
| Saturated fat, g | Tertile 1 (<50.5) | 20 / 13.7 | 1.0 | 1.0 |
| (56.9 [15.0]) | Tertile 2 (50.5–62.1) | 25 / 14.4 | 1.20 (0.67, 2.16) | 1.30 (0.72, 2.35) |
| Tertile 3 (>62.1) | 15 / 14.9 | 0.68 (0.35, 1.33) | 0.68 (0.34, 1.35) | |
| Per 1 SD increase | - | 0.86 (0.66, 1.13) | 0.85 (0.65, 1.12) | |
| P-value for trend | - | 0.27 | 0.30 | |
| Monounsaturated fat, g | Tertile 1 (<40.0) | 25 / 13.6 | 1.0 | 1.0 |
| (45.4 [12.3]) | Tertile 2 (40.0–49.7) | 17 / 14.6 | 0.63 (0.34, 1.16) | 0.65 (0.35, 1.22) |
| Tertile 3 (>49.7) | 18 / 14.8 | 0.66 (0.36, 1.21) | 0.64 (0.34, 1.21) | |
| Per 1 SD increase | - | 0.87 (0.66, 1.15) | 0.85 (0.64, 1.13) | |
| P-value for trend | - | 0.16 | 0.15 | |
| Polyunsaturated fat, g | Tertile 1 (<6.8) | 24 / 13.7 | 1.0 | 1.0 |
| (7.9 [2.8]) | Tertile 2 (6.8–8.7) | 18 / 14.9 | 0.72 (0.39, 1.33) | 0.74 (0.39, 1.37) |
| Tertile 3 (>8.7) | 18 / 14.4 | 0.81 (0.44, 1.49) | 0.81 (0.43, 1.52) | |
| Per 1 SD increase | - | 0.90 (0.68, 1.20) | 0.89 (0.67, 1.20) | |
| P-value for trend | - | 0.46 | 0.48 | |
| Calcium, mg | Tertile 1 (<983) | 17 / 13.7 | 1.0 | 1.0 |
| (1140 [321]) | Tertile 2 (983–1255) | 25 / 14.9 | 1.19 (0.64, 2.20) | 1.24 (0.66, 2.31) |
| Tertile 3 (>1255) | 18 / 14.4 | 0.96 (0.50, 1.87) | 0.90 (0.46, 1.77) | |
| Per 1 SD increase | - | 1.07 (0.83, 1.38) | 1.01 (0.78, 1.31) | |
| P-value for trend | - | 0.90 | 0.75 | |
| Linoleic acid, g | Tertile 1 (<4.86) | 22 / 13.7 | 1.0 | 1.0 |
| (5.75 [ 2.19]) | Tertile 2 (4.86–6.27) | 18 / 14.6 | 0.78 (0.42, 1.45) | 0.80 (0.43, 1.50) |
| Tertile 3 (>6.27) | 20 / 14.6 | 0.87 (0.48, 1.60) | 0.84 (0.45, 1.56) | |
| Per 1 SD increase | - | 0.93 (0.70, 1.24) | 0.91 (0.68, 1.22) | |
| P-value for trend | - | 0.66 | 0.57 | |
| Cholesterol, g | Tertile 1 (<449) | 25 / 14.4 | 1.0 | 1.0 |
| (558 [212]) | Tertile 2 (449–631.8) | 17 / 14.1 | 0.72 (0.39, 1.34) | 0.75 (0.40, 1.40) |
| Tertile 3 (>631.8) | 18 / 14.5 | 0.74 (0.40, 1.36) | 0.75 (0.40, 1.38) | |
| Per 1 SD increase | - | 0.81 (0.62, 1.07) | 0.82 (0.62, 1.09) | |
| P-value for trend | - | 0.32 | 0.34 | |
| Retinol, µg | Tertile 1 (<565) | 25 / 13.6 | 1.0 | 1.0 |
| (1432 [2041]) | Tertile 2 (565–867) | 16 / 14.4 | 0.55 (0.29, 1.02) | 0.57 (0.30, 1.08) |
| Tertile 3 (>867) | 19 / 15.0 | 0.61 (0.34, 1.11) | 0.57 (0.31, 1.04) | |
| Per 1 SD increase | - | 0.74 (0.55, 0.99) | 0.72 (0.54, 0.97) | |
| P-value for trend | - | 0.10 | 0.07 | |
| Beta-carotene, µg | Tertile 1 (<1082) | 13 / 13.5 | 1.0 | 1.0 |
| (2634 [2665]) | Tertile 2 (1082–2403) | 23 / 14.6 | 1.50 (0.76, 2.96) | 1.61 (0.81, 3.19) |
| Tertile 3 (>2403) | 24 / 14.9 | 1.42 (0.72, 2.78) | 1.33 (0.67, 2.64) | |
| Per 1 SD increase | - | 1.08 (0.84, 1.39) | 1.01 (0.79, 1.30) | |
| P-value for trend | - | 0.36 | 0.50 | |
multiple adjustment is adjustment for BMI, plasma cholesterol, physical activity, socio-economic status, diabetes/blood glucose, marital status, FEV1, height, age at risk, smoking, and diastolic and systolic blood pressure (estimates for components of blood pressure are not mutually adjusted).
Discussion
In this large prospective cohort study with almost complete follow-up of its members, of the eleven risk factors examined in relation to prostate cancer mortality, there was a suggestion that being widowed or divorced, higher levels of plasma cholesterol, and increased physical stature were associated with elevated risk. The magnitude of these relationships was generally modest and, although the associations of marital status, plasma cholesterol and physical stature were not attenuated with adjustment for the measured covariates, they could plausibly be explained by residual confounding by other unmeasured factors or, given the large number of exposure variables, the problem of multiple comparisons. There was essentially no association between the dietary variables and later prostate cancer mortality risk, although these analyses were hampered by a modest number of cases.
In keeping with our own results, studies examining the relationship between plasma or blood cholesterol and prostate cancer risk typically reveal positive associations,25;26 although this is not a universal observation.27–29 A series of reports suggest taller men are more likely to experience prostate cancer than their shorter counterparts.30–32 Our observation in some analyses that elevated lung function confers some increased risk of prostate carcinoma has, to our knowledge, not been previously examined. Both lung function and stature may be conceptualised as markers of pre-adult exposures,21 and, while under a degree of genetic control, probably capture nutrition during this period.33 It is plausible that over-feeding during selected periods across the life course stimulate carcinogenesis.34 Underlying this relation may be levels of insulin-like growth factors (IGF) which correlate directly with caloric intake (in animals),35 height in children,36 and risk of prostate cancer in adult humans.37 A separate explanation for the positive height–cancer effects evident in the present analysis simply posits that taller people have larger bodies and therefore more cells that may potentially undergo malignant transformation to cancer.38
This study has several strengths, including: its large sample size which leads to a greater number of fatal prostate cancer events than most other studies, prospective design, statistical control of a range of potential mediating and confounding variables, and almost complete follow-up for mortality. Weaknesses include: the absence of measurement of other potential risk factors such as hormone levels and exposure to environmental pesticides; some of the risk factor-prostate cancer mortality relationships we explored were not hypothesis-driven; examining the influence of multiple exposures on prostate cancer death may have led to positive results surfacing by chance alone; and, owing to a lack of data, we were unable to examine if the risk factors for prostate cancer mortality – our only endpoint of interest – differed to those for prostate cancer incidence. While misclassification of the exposures variables may have occurred, this is likely to be random with respect to the outcome of interest, so resulting in an underestimation of the magnitude of the effect estimates.
In conclusion, in the present study there was a suggestion that being widowed or divorced, and higher levels of plasma cholesterol, and physical stature were associated with higher rates of prostate cancer mortality.
Acknowledgements
The original screening of the Whitehall study was funded by the Department of Health and Social Security and the Tobacco Research Council. David Batty is a Wellcome Trust Fellow; Michael Marmot is a MRC Research Professor. Martin Shipley is supported by the British Heart Foundation and Mika Kivimaki by the Academy of Finland, Finland, the National Heart, Lung, and Blood Institute (R01HL036310-20A2) and the National Institute on Aging (R01AG034454-01), NIH, US. The Medical Research Council (MRC) Social and Public Health Sciences Unit receives funding from the MRC and the Chief Scientist Office at the Scottish Government Health Directorates.
References
- 1.Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55:74–108. doi: 10.3322/canjclin.55.2.74. [DOI] [PubMed] [Google Scholar]
- 2.Hsing AW, Chokkalingam AP. Prostate cancer epidemiology. Front Biosci. 2006;11:1388–1413. doi: 10.2741/1891. [DOI] [PubMed] [Google Scholar]
- 3.Shimizu H, Ross RK, Bernstein L, Yatani R, Henderson BE, Mack TM. Cancers of the prostate and breast among Japanese and white immigrants in Los Angeles County. Br.J Cancer. 1991;63:963–966. doi: 10.1038/bjc.1991.210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zeegers MP, Dirx MJ, van den Brandt PA. Physical activity and the risk of prostate cancer in the Netherlands cohort study, results after 9.3 years of follow-up. Cancer Epidemiol.Biomarkers Prev. 2005;14:1490–1495. doi: 10.1158/1055-9965.EPI-04-0771. [DOI] [PubMed] [Google Scholar]
- 5.Wannamethee SG, Shaper AG, Walker M. Physical activity and risk of cancer in middle-aged men. Br.J Cancer. 2001;85:1311–1316. doi: 10.1054/bjoc.2001.2096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hickey K, Do KA, Green A. Smoking and prostate cancer. Epidemiol.Rev. 2001;23:115–125. doi: 10.1093/oxfordjournals.epirev.a000776. [DOI] [PubMed] [Google Scholar]
- 7.Friedman GD. Blood pressure and heart rate: no evidence for a positive association with prostate cancer. Ann.Epidemiol. 1997;7:486–489. doi: 10.1016/s1047-2797(97)00083-5. [DOI] [PubMed] [Google Scholar]
- 8.Huxley R. The impact of modifiable risk factors on mortality from prostate cancer in populations of the Asia-Pacific region. Asian Pac.J Cancer Prev. 2007;8:199–205. [PubMed] [Google Scholar]
- 9.Eichholzer M, Stahelin HB, Gutzwiller F, Ludin E, Bernasconi F. Association of low plasma cholesterol with mortality for cancer at various sites in men: 17-y follow-up of the prospective Basel study. Am.J Clin.Nutr. 2000;71:569–574. doi: 10.1093/ajcn/71.2.569. [DOI] [PubMed] [Google Scholar]
- 10.Mondul AM, Clipp SL, Helzlsouer KJ, Platz EA. Association between plasma total cholesterol concentration and incident prostate cancer in the CLUE II cohort. Cancer Causes Control. 2009 doi: 10.1007/s10552-009-9434-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kasper JS, Giovannucci E. A meta-analysis of diabetes mellitus and the risk of prostate cancer. Cancer Epidemiol.Biomarkers Prev. 2006;15:2056–2062. doi: 10.1158/1055-9965.EPI-06-0410. [DOI] [PubMed] [Google Scholar]
- 12.Reid DD, Hamilton PJS, McCartney P, Rose G, Jarrett RJ, Keen H, et al. Cardiorespiratory disease and diabetes among middle-aged male civil servants. Lancet. 1974;i:469–473. doi: 10.1016/s0140-6736(74)92783-4. [DOI] [PubMed] [Google Scholar]
- 13.Davey Smith G, Leon D, Shipley MJ, Rose G. Socioeconomic differentials in cancer among men. Int J Epidemiol. 1991;20:339–345. doi: 10.1093/ije/20.2.339. [DOI] [PubMed] [Google Scholar]
- 14.Batty GD, Kivimaki M, Gray L, Smith GD, Marmot MG, Shipley MJ. Cigarette smoking and site-specific cancer mortality: testing uncertain associations using extended follow-up of the original Whitehall study. Ann Oncol. 2008;19:1002. doi: 10.1093/annonc/mdm578. [DOI] [PubMed] [Google Scholar]
- 15.Ben Shlomo Y, Davey Smith G, Shipley M, Marmot MG. Magnitude and causes of mortality differences between married and unmarried men. J Epidemiol Community Health. 1993;47:200–205. doi: 10.1136/jech.47.3.200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Batty GD, Shipley M, Marmot M, Davey Smith G. Physical activity and cause-specific mortality in men: further evidence from the Whitehall study. Eur J Epidemiol. 2002;17:863–869. doi: 10.1023/a:1015609909969. [DOI] [PubMed] [Google Scholar]
- 17.Davey Smith G, Shipley MJ, Batty GD, Morris JN, Marmot M. Physical activity and cause-specific mortality in the Whitehall study. Public Health. 2000;114:308–315. doi: 10.1038/sj.ph.1900675. [DOI] [PubMed] [Google Scholar]
- 18.Batty GD, Shipley MJ, Marmot MG, Davey Smith G. Blood pressure and site-specific cancer mortality: evidence from the original Whitehall study. Br.J.Cancer. 2003;89:1243–1247. doi: 10.1038/sj.bjc.6601255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Batty GD, Shipley MJ, Langenberg C, Marmot MG, Davey Smith G. Adult height in relation to mortality from 14 cancer sites in men in London (UK): evidence from the original Whitehall study. Ann.Oncol. 2006;17:157–166. doi: 10.1093/annonc/mdj018. [DOI] [PubMed] [Google Scholar]
- 20.Batty GD, Shipley MJ, Jarrett RJ, Breeze E, Marmot MG, Smith GD. Obesity and overweight in relation to organ-specific cancer mortality in London (UK): findings from the original Whitehall study. Int J Obes (Lond) 2005;29:1267–1274. doi: 10.1038/sj.ijo.0803020. [DOI] [PubMed] [Google Scholar]
- 21.Batty GD, Gunnell D, Langenberg C, Smith GD, Marmot MG, Shipley MJ. Adult height and lung function as markers of life course exposures: associations with risk factors and cause-specific mortality. Eur.J.Epidemiol. 2006;21:795–801. doi: 10.1007/s10654-006-9057-2. [DOI] [PubMed] [Google Scholar]
- 22.Davey Smith G, Shipley MJ, Marmot MG, Rose G. Plasma cholesterol concentration and mortality. The Whitehall Study. JAMA. 1992;267:70–76. [PubMed] [Google Scholar]
- 23.Batty GD, Shipley MJ, Marmot M, Davey Smith G. Diabetes status and post-load plasma glucose concentration in relation to site-specific cancer mortality: findings from the original Whitehall study. Cancer Causes Control. 2004;15:873–881. doi: 10.1007/s10552-004-1050-z. [DOI] [PubMed] [Google Scholar]
- 24.Keen H, Thomas BJ, Jarrett RJ, Fuller JH. Nutrient intake, adiposity, and diabetes. Br.Med.J. 1979;1:655–658. doi: 10.1136/bmj.1.6164.655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Platz EA, Till C, Goodman PJ, Parnes HL, Figg WD, Albanes D, et al. Men with low serum cholesterol have a lower risk of high-grade prostate cancer in the placebo arm of the prostate cancer prevention trial. Cancer Epidemiol.Biomarkers Prev. 2009;18:2807–2813. doi: 10.1158/1055-9965.EPI-09-0472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Platz EA, Clinton SK, Giovannucci E. Association between plasma cholesterol and prostate cancer in the PSA era. Int.J Cancer. 2008;123:1693–1698. doi: 10.1002/ijc.23715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Huxley R. The impact of modifiable risk factors on mortality from prostate cancer in populations of the Asia-Pacific region. Asian Pac.J Cancer Prev. 2007;8:199–205. [PubMed] [Google Scholar]
- 28.Eichholzer M, Stahelin HB, Gutzwiller F, Ludin E, Bernasconi F. Association of low plasma cholesterol with mortality for cancer at various sites in men: 17-y follow-up of the prospective Basel study. Am.J Clin.Nutr. 2000;71:569–574. doi: 10.1093/ajcn/71.2.569. [DOI] [PubMed] [Google Scholar]
- 29.Mondul AM, Clipp SL, Helzlsouer KJ, Platz EA. Association between plasma total cholesterol concentration and incident prostate cancer in the CLUE II cohort. Cancer Causes Control. 2009 doi: 10.1007/s10552-009-9434-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.MacInnis RJ, English DR. Body size and composition and prostate cancer risk: systematic review and meta-regression analysis. Cancer Causes Control. 2006;17:989–1003. doi: 10.1007/s10552-006-0049-z. [DOI] [PubMed] [Google Scholar]
- 31.Engeland A, Tretli S, Bjorge T. Height, body mass index, and prostate cancer: a follow-up of 950000 Norwegian men. Br.J Cancer. 2003;89:1237–1242. doi: 10.1038/sj.bjc.6601206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Davey SG, Hart C, Upton M, Hole D, Gillis C, Watt G, et al. Height and risk of death among men and women: aetiological implications of associations with cardiorespiratory disease and cancer mortality. J Epidemiol.Community Health. 2000;54:97–103. doi: 10.1136/jech.54.2.97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Batty GD, Shipley M, Gunnell D, Kivimaki M, Woodward M, Man Ying Lee C, et al. Height, Wealth, and Health: An Overview With New Data From Three Longitudinal Studies. Econ.Hum.Biol.(in press) 2009 doi: 10.1016/j.ehb.2009.06.004. [DOI] [PubMed] [Google Scholar]
- 34.Frankel S, Gunnell DJ, Peters TJ, Maynard M, Davey Smith G. Childhood energy intake and adult mortality from cancer: the Boyd Orr Cohort Study. BMJ. 1998;316:499–504. doi: 10.1136/bmj.316.7130.499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ruggeri BA, Klurfeld DM, Kritchevsky D, Furlanetto RW. Caloric restriction and 7,12-dimethylbenz(a)anthracene-induced mammary tumor growth in rats: alterations in circulating insulin, insulin-like growth factors I and II, and epidermal growth factor. Cancer Res. 1989;49:4130–4134. [PubMed] [Google Scholar]
- 36.Juul A, Bang P, Hertel NT, Main K, Dalgaard P, Jorgensen K, et al. Serum insulin-like growth factor-I in 1030 healthy children, adolescents, and adults: relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin.Endocrinol.Metab. 1994;78:744–752. doi: 10.1210/jcem.78.3.8126152. [DOI] [PubMed] [Google Scholar]
- 37.Renehan AG, Zwahlen M, Minder C, O'Dwyer ST, Shalet SM, Egger M. Insulin-like growth factor (IGF)-I, IGF binding protein-3, and cancer risk: systematic review and meta-regression analysis. Lancet. 2004;363:1346–1353. doi: 10.1016/S0140-6736(04)16044-3. [DOI] [PubMed] [Google Scholar]
- 38.Albanes D, Jones Y, Schatzkin A, Micozzi MS, Taylor PR. Adult stature and risk of cancer. Cancer Research. 1988;48:1658–1662. [PubMed] [Google Scholar]