Table 4 Summary of observational studies on the association between ARA and risk of prostate cancer

Study design: cohort study

Exposure assessment: dietary intake

Leitzmann et al. 2004 [67]

HPFS, USA, 1986-2000, prospective cohort design (14 years follow-up)

47,866 health professionals aged 40-65, no prior history of cancer

Semiquantitative FFQ, 131 items, validated against 2 x 1-week DR

Self-reported physician diagnosis supplemented by medical record and pathology report

Age, time period, race, family history of prostate cancer, history of type 2 DM and vasectomy, BMI, height, smoking status, physical activity, total energy intake, % of energy from protein intake, monounsaturated fat intake, saturated fat intake and trans unsaturated fat intake, calcium intake, supplemental vitamin E and lycopene

21

Dietary ARA intake, %energy, quintile

Ptrend

Q1: <0.028

0.44

Q2: 0.028-0.035

Q3: 0.036-0.041

Q4: 0.042-0.049

Q5: >0.049

Study design: nested case-control study

Exposure assessment: dietary intake

Männistö et al. 2003 [68]

ATBC study, Finland, 1985-1993, nested case-control design (5-8 years follow-up)

198 prostate cancer patients, 198 controls (free of prostate cancer) matched by age, trial supplementation group

Self-administered dietary questionnaire, 276 items, validated against 12 x 2-day DR

Finnish Cancer Registry and Register of Causes of Death

Resident area, educational level, BMI, alcohol intake, smoking period

23

Dietary ARA intake, g/day, median

Ptrend

Q1: 0.04

0.23

Q2: 0.06

Q3: 0.07

Q4: 0.10

Schuurman et al. 1999 [69]

NLCS, Netherlands, 1986-1992 (6.3 years follow-up), case-cohort design

642 primary prostate cancer patients from entire cohort, 1,525 subcohort members (selection criteria not shown) aged 55-69 at baseline, without prevalent cancer other than skin cancer, matching not indicated

Semiquantitative FFQ, 150 items, validated against 3 x 3-day DR

All regional cancer registries and Dutch national database of pathology reports

Age, family history of prostate carcinoma, socioeconomic status, total energy intake, total energy-adjusted fat intake

23

Dietary ARA intake, g/day, quintile, median

Ptrend

Q1: 0.06

0.30

Q2: 0.09

Q3: 0.11

Q4: 0.13

Q5: 0.17

Exposure assessment: blood ARA level

Crowe et al. 2008 [70]

EPIC study, Denmark, Germany, Greece, Italy, Netherlands, Spain, Sweden, UK, 1992-2000, nested case-cohort design

962 prostate cancer patients, 1,061 controls without prevalent cancer other than NMSC, 1 case matched with 1-2 control(s) by study center, age, time of blood sampling, time between blood sampling and last consumption of food or drink

Plasma phospholipids, GC analysis, precision indicated

Regional or national cancer registries or combination of health insurance records, cancer and pathology registries and self-report

BMI, smoking status, alcohol intake, educational level, marital status, physical activity

26

ARA composition mol%, quintile

Ptrend

Q1: 4.40–7.93

0.419

Q2: 7.93–8.89

Q3: 8.90–9.86

Q4: 9.86–10.98

Q5: 10.99–19.14

Chavarro et al. 2007 [71]

PHS, USA, 1982-1995, nested case-control design within a randomized, double-blind, placebo-controlled factorial aspirin and beta-carotene trial (13 years follow-up)

476 prostate cancer patients, 476 controls, male physicians without history of cancer except NMSC, 1 case matched with 1 control by age, smoking status, with consideration for trial intervention

Whole blood fatty acids, GC analysis blinded to case-control status, precision indicated

Self-report, combined with review of hospital records and pathology reports

Age, smoking status, length of follow-up

22

ARA concentration (%,), quintile, median

Ptrend

Q1: 7.9

0.98

Q2: 9.3

Q3: 10.1

Q4: 10.9

Q5: 12.3

Männistö et al. 2003 [68]

ATBC study, Finland, 1985-1993, nested case-control design (5-8 years follow-up)

198 prostate cancer patients, 198 controls (free of prostate cancer) matched by age, trial supplementation group

Serum cholesterol ester fatty acids, GC analysis, precision indicated

Finnish Cancer Registry and Register of Causes of Death

Resident area, educational level, BMI, alcohol intake, smoking period

23

ARA composition %, quartile, median

Ptrend

Q1: 3.96

0.34

Q2: 4.55

Q3: 5.09

Q4: 5.89

Harvei et al. 1997 [72]

Janus serum bank, Norway, 1973-1994, nested case-control design

141 prostate cancer patients, 282 controls (eligibility criteria not shown), 1 case matched with 2 controls by age, date of blood sampling, resident area

Serum phospholipids, GC analysis, blinded to case-control status, precision not indicated

Cancer Registry and Statistics Norway

None

14

ARA concentration mg/l, quartile, upper limit

Ptrend

Q1: 4.86

0.6

Q2: 5.68

Q3: 6.68

Q4: >6.68

Gann et al. 1994 [73]

PHS, USA, 1982-1988, nested case-control design within a randomized, double-blind, placebo-controlled factorial aspirin and beta-carotene trial (6 years follow-up)

120 prostate cancer patients, 120 controls, male physicians without history of cancer except NMSC, 1 case matched with 1 control by age, smoking status without regard to trial intervention

Plasma cholesterol ester fatty acids, GC analysis blinded to case-control status, precision indicated

Self-report, combined with review of medical records

None

19

ARA composition of plasma cholesterol estel %, quartile

Ptrend

Q1

0.76

Q2

Q3

Q4

Study design: case-control study (temporal relationship among exposure and outcome is unclear)

Exposure assessment: dietary intake

Hodge et al. 2004 [74]

Survey, Australia, 1994-1997, case-control design

858 prostate cancer patients aged <70, 905 controls matched by age

Melbourne FFQ, 121 items, validated against 2 x 4-day WFR

Not shown

Age at selection, study center, calendar year, family history of prostate cancer, country of birth, socioeconomic status

18

Dietary ARA intake, g/day, quintile

Ptrend

Q1: <0.028

0.6

Q2: 0.028-0.036

Q3: 0.037-0.046

Q4: 0.047-0.059

Q5: ≥0.06

Exposure assessment: blood ARA level

Ukori et al. 2010 [75]

Survey, USA and Nigeria, case-control design

48 African American and 66 Nigerian prostate cancer patients, 96 African American and 226 Nigerian controls, aged ≥40, without any cancer history other than skin cancer, matching not indicated

Plasma fatty acids (fasting blood), GC analysis, precision not indicated

Abnormal DRE and/or abnormal PSA (>4ng/ml) with histological diagnosis

Age, educational level, family history of prostate cancer, WHR

14

ARA concentration μg/ml, quartile American African:

Ptrend

Q1 vs Q4

American African:

Nigerian:

Q1 vs Q4

<0.05

Nigerian:

Not significant

Ukori et al. 2009 [76]

Survey, Nigeria, case-control design

66 prostate cancer patients, 226 controls, aged ≥40, matching not indicated (same population as Nigerian participants of Ukori et al. 2010)

Plasma fatty acids (fasting blood), GC analysis, precision not indicated

Abnormal DRE and/or abnormal PSA (>4ng/ml) with histological diagnosis

Age, educational level, family history of prostate cancer, WHR

11

ARA concentration μg/ml, quartile

Ptrend

Q1

0.06

Q2

Q3

Q4

Newcomer et al. 2001 [77]

Survey, USA, case-control design

67 prostate cancer patients, 156 population-based controls, 1 case matched with about 2 controls by age distribution

Erythrocyte fatty acids, GC analysis blinded to case-control status, precision indicated

Not shown

Age

23

ARA composition weight%, quartile

Ptrend

Q1: ≤13.25

0.88

Q2: 13.26-14.12

Q3: 14.13-14.90

Q4: ≥14.91

Yang et al. 1999 [78]

Survey, Korea

19 prostate cancer patients, 24 benign prostatic hyperplasia patients, 21 normal controls, matched by age, demographics

Serum fatty acids, GC-MS analysis, precision not indicated

Not shown

None

4

ARA composition%, mean (SD)

P

Cancer:

Not significant

0.77(0.31)

Benign:

0.95(0.16)

Study design: cross-sectional study

Exposure assessment: blood ARA level

Faas et al. 2003 [79]

Survey, USA, 1995-1998

Prostate cancer patients, benign prostate disease patients

Erythrocyte and plasma phospholipids, GC analysis, precision not indicated

Pathology reports

None

10

Erythrocyte ARA composition%, mean(SEM)

P

Malignant:

Erythrocyte:

16.33(0.28)

Not significant

Plasma ARA composition%, mean(SEM)

Plasma:

Malignant:

Not significant

12.60(0.27)

Hietanen et al. 1994 [46]

Survey, UK, cross-sectional design

10 prostate cancer patients aged 64-85, controls, matched by age, sex, smoking status

Erythrocyte phospholipids (fasting blood), GC analysis, precision not indicated

Not shown

None

8

ARA composition%, mean(SD)

P

Case:

Not significant

17.8(1.3)

Chaudry et al. 1991 [80]

Survey, UK

20 patients admitted for prostatic surgery (10 malignant, 10 benign)

Plasma phospholipids (fasting blood), GC analysis, precision not indicated

Histological diagnosis

None

6

ARA composition%, median(IQR)

P

Malignant:

Not significant

8.93(1.84)

Exposure assessment: tissue ARA level

Faas et al. 2003 [79]

Survey, USA, 1995-1998

Prostate cancer patients, benign prostate disease patients

Prostate tissue phospholipids, GC analysis, precision not indicated

Pathology reports

None

10

ARA composition%, mean(SEM)

P

Malignant:

<0.001

15.20(0.33)

Mamalakis et al. 2002 [81]

Survey, Greece, 1997-1999

36 prostate cancer patients, 35 benign prostate hyperplasia patients

Gluteal adipose tissue and prostate tissue fatty acids, GC analysis, precision not indicated

DRE, serum PSA, transrectal ultrasound, prostate biopsy

None

12

Gluteal adipose tissue ARA composition%, mean(SD)

P

Malignant:

Gluteal adipose tissue:

0.28(0.12)

Not significant

Prostate tissue ARA composition%, mean(SD)

Malignant:

Prostate tissue:

5.99(3.65)

<0.001

Chaudry et al. 1991 [80]

Survey, UK

20 patients admitted for prostatic surgery (10 malignant, 10 benign)

Prostate tissue phospholipids, GC analysis, precision not indicated

Histological diagnosis

None

6

ARA composition%, median(IQR)

P

Malignant:

11.33(4.12)

0.002