Table 3 Summary of observational studies on the association between ARA and risk of breast cancer

Study design: cohort study

Exposure assessment: dietary intake

Holmes et al. 1999 [51]

NHS, USA, 1976- 1994, prospective cohort design (14 year biennial follow-up, follow-up rate = 95%)

88,795 female nurses aged 30-55, no prior history of cancer other than nonmelanoma skin cancer

Semiquantitative FFQ, 131 items, validated against 2 x 7-day WR

Self-reported physician diagnosis, deaths identified by family member of participants, postal services and National Death Index, supplemented by medical record

Total energy intake, age, energy-adjusted vitamin A intake, alcohol intake, time period, height, parity, age at first birth, weight change, BMI, age at menopause, menopausal status, use of HRT, family history, benign breast disease, age at menarche

19

%energy increment of dietary ARA intake per day 0.03

P

Not shown

Study design: nested case-control study

Exposure assessment: dietary intake

Voorrips et al. 2002 [52]

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

941 breast cancer patients from entire cohort, 1,598 subcohort members (selection criteria not shown), aged 55-69 at baseline, no prior history of cancer other than nonmelanoma 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, history of benign breast disease, maternal breast cancer, breast cancer in one or more sisters, age at menarche, age at menopause, oral contraceptive use, parity, age at first birth, Quetelet index, educational level, alcohol intake, smoking status, total energy intake, total energy-adjusted fat intake

19

Dietary ARA intake, g/day, quintile, median

Ptrend

Q1: 0.05

0.93

Q2: 0.07

Q3: 0.09

Q4: 0.11

Q5: 0.15

Exposure assessment: blood ARA level

Saadatian-Elahi et al. 2002 [29]

NYUWHS, USA, 1985-1995 (average 4.3 years follow-up), nested case-control design

197 breast cancer patients, 197 controls (free of cancer), aged 34-65, matched by age, menopausal status, date of blood sampling, number of blood samplings, day of menstrual cycle

Serum phospholipids, GC analysis, precision indicated

Self-reported physician diagnosis, combined with tumor registries, mortality databases and review of clinical and pathological documents

Family history, age at first full-term birth, total cholesterol, history of treatment for benign breast conditions

19

ARA composition%, quartile

P for the overall categorial variable:

Q1

0.80

Q2

Q3

Ptrend with the score variable

Q4

0.66

Pala et al. 2001 [53]

ORDET study, Italy, 1987-1995 (average 5.5 years follow-up)

71 breast cancer patients, 141 controls (free of cancer), 1 case matched with 2 controls by age, menopausal status at recruitment, daylight-saving period at blood sampling, recruitment center and date

Erythrocyte phospholipids (fasting blood), GC analysis blinded to case-control status, precision indicated

Lombardy Cancer Registry

None (BMI, WHR, age at menarche, age at first birth, age at menopause, months of lactation, parity and educational level were investigated)

23

ARA composition%, tertile

Ptrend

T1: <16.67

0.42

T2: ≥16.67-

<17.94

T3: ≥17.94

Chajès et al. 1999 [54]

Three ongoing cohort studies in Sweden, VIP(1986- 1997), northern Sweden component of the WHO MONICA(1986, 1990 and 1994), MSP(1995-1997), nested case-control design

196 breast cancer patients (VIP 103, MONICA 9, MSP 84), 388 controls (VIP 214, MONICA 6, MSP 168), 1 case matched with 2 controls by age, age of blood sample, sampling center

Serum phospholipids (for VIP and MONICA fasting blood, for MSP very little fasting blood), GC analysis, precision indicated

Regional cancer registry, National Cancer Registry, follow-up for vital status (death) or losses to follow-up determined through local and national population registries

Age at menarche, parity, age at first full-term pregnancy, use of hormones, menopausal status

19

ARA composition%, quartile

Ptrend

Q1

0.091

Q2

Q3

Q4

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

Exposure assessment: dietary intake

Nkondjock et al. 2003 [55]

Survey, Canada, 1989-1993, case-control design

414 primary breast cancer patients aged 35-79, 688 controls (eligibility criteria not shown), population-based, matched by age, language, place of residence

French version FFQ, >200 items, validated against 7-day FD

Histological diagnosis

Age at first full-term pregnancy, smoking status, family history of breast cancer, history of benign breast disease, marital status, number of full-term pregnancies, total energy intake

20

Dietary ARA intake, g/day, quartile

Ptrend

Q1

0.723

Q2

Q3

Q4

Exposure assessment: blood ARA level

Vatten et al. 1993 [56]

Janus Serum Bank, Norway, 1973-1991, case-control design

87 breast cancer patients, 235 controls with no prior history of cancer, matched by age, date of blood sampling

Serum phospholipid, GC analysis blinded to case-control state, precision indicated

National cancer registry linked to Janus Serum Bank donor information

None

20

ARA concentration, mg/l, mean(SD) 78(30)

P

Not significant

Exposure assessment: tissue ARA level

London et al. 1993 [57]

Survey, USA, 1986-1988, case-control design

Postmenopausal women, 380 breast cancer patients, 573 controls with breast abnormality (free of breast cancer), matching not indicated

Buttock adipose tissue fatty acids, GC analysis, precision indicated

Physician diagnosis (detail not shown)

Age, alcohol intake, age at first birth, parity, family history of breast cancer, age at menopause, age at menarche, history of benign breast disease, weight

19

ARA composition%, quintile

Ptrend

Q1

0.60

Q2

Q3

Q4

Q5

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

Exposure assessment: dietary intake

Zhu et al. 1995 [58]

Survey, Finland, 1990-1992

17 premenopausal, 32 postmenopausal primary breast cancer patients, 34 premenopausal, 16 postmenopausal controls with benigh breast disease (eligibility criteria not shown), matching not indicated

Semiquantitative FFQ, 110 items, validated against 14-day DR

Histological diagnosis

Age, total energy intake

13

Dietary ARA intake, mg/day, mean(SD)

P

Premenopausal case:

Premenopausal:

58(27)

Not significant

Postmenopausal case:

Postmenopausal:

90(191)

Not significant

Exposure assessment: blood ARA level

Aro et al. 2000 [59]

Kuopio Breast Cancer Study, Finland, 1992-1995, case-control design

195 primary breast cancer patients aged 25-75, 208 controls drawn randomly from the National Population Register, matched by age, long-term area of residence

Serum fatty acids (fasting blood), GC analysis, precision indicated

Histological diagnosis

Age, area, age at menarche, age at first full-term pregnancy, use of oral contraceptives, use of HRT, family history of breast cancer, history of benign breast disease, educational level, alcohol intake, smoking status, physical activity, WHR, BMI

15

ARA composition%, quintile, median

Ptrend

Postmenopausal:

Postmenopausal:

Q1: 3.84

Signifncant

Q2: 4.89

Q3: 5.46

Q4: 6.04

Q5: 7.15

ARA composition%, mean(SD)

P

Premenopausal case:

Premenopausal:

5.68(1.01)

Not significant

Zaridze et al. 1990 [60]

Survey, now-defunct Union of Soviet Socialist Republics, case-control design

25 premenopausal, 21 postmenopausal primary breast cancer patients, 20 premenopausal, 33 postmenopausal neighborhood controls (eligibility criteria not shown), matching not indicated

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

Not shown

None

11

ARA concentration, μg/mg phospholipids, bisectional, (Summer-Autumn/Winter-Spring)

P

Premenopausal:

Premenopausal:

≤11.70/9.89 vs

0.122

>11.70/9.89

Postmenopausal:

Postmenopausal:

≤11.70/9.89 vs

0.018

>11.70/9.89

Exposure assessment: tissue ARA level

Bagga et al. 2002 [61]

Survey, USA, 1995-1996, case-control design

73 breast cancer patients, 73 controls undergoing reduction mammoplasty for mastomegaly, matching not indicated

Breast adipose tissue fatty acids, GC analysis, precision not indicated

Not shown

None

15

ARA concentration, μmol/g total fatty acid, mean(SEM)

P

Case:

0.27

15.03(1.20)

Maillard et al. 2002 [62]

Survey, France, 1992-1996, case-control design

241 patients with non-metastatic invasive breast carcinoma, 88 controls with benign breast diseases, matching not indicated

Breast adipose tissue triglycerides, GC analysis blinded to case-control status, precision indicated

Not shown

Age at diagnosis, height, BMI, menopausal status, BMI-menopausal status interaction

16

ARA composition%, tertile

Ptrend

T1

0.32

T2

T3

Zhu et al. 1995 [58]

Survey, Finland, 1990-1992

26 premenopausal, 47 postmenopausal primary breast cancer patients, 35 premenopausal, 20 postmenopausal controls with benign breast disease (eligibility criteria not shown), matching not indicated

Breast adipose tissue triglycerides and phospholipids, GC analysis, precision not indicated

Histological diagnosis

Age

13

Triglyceride ARA composition mol%, mean(SD)

P

Premenopausal case:

Triglyceride

0.33(0.18)

Premenopausal:

Postmenopausal case:

Not significant

0.33(0.18)

Postmenopausal:

Phospholipid ARA composition mol%, mean(SD), Premenopausal case:

<0.01

9.67(2.56)

Phospholipid

Postmenopausal case:

Premenopausal:

9.64(2.26)

Not significant

Postmenopausal:

Not significant

Petrek et al. 1994 [63]

Survey, USA, 1987-1989, case-control design

154 invasive breast cancer patients, 125 controls at average risk of breast cancer, matching not indicated

Breast adipose tissue fatty acids, GC analysis, precision not indicated

Histological diagnosis

None

7

ARA composition weight%, mean(SD)

P

Case:

Not significant

0.40(0.15)

Study design: cross-sectional study

Exposure assessment: blood ARA level

Williams et al. 1993 [64]

Survey, UK

12 malignant breast disease patients, 10 benign breast disease patients, 22 normal controls

Erythrocyte PIs and PCs (fasting blood), GC analysis, precision not indicated

Histological diagnosis

None

8

ARA composition%, only shown as figure:

P

Erythrocyte PIs: not significant

PCs:

Erythrocyte PCs: significantly higher in control compared with benign and malignant group

Malignant/Control:

<0.02

Benign/Control:

<0.02

Hietanen et al. 1994 [46]

Survey, UK, cross-sectional design

20 breast cancer patients aged 37-85, controls matched by age, sex, smoking status

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

Not shown

None

10

ARA composition%, mean(SD)

P

Case:

Not significant

17.5(0.8)

Punnonen et al. 1989 [65]

Survey, Finland

6 breast cancer patients, 9 normal controls

Erythrocyte phospholipids, GC analysis, precision not indicated

Histological diagnosis

None

5

ARA composition%, mean(SEM)

P

Case:

Not significant

12.1(1.5)

Exposure assessment: tissue ARA level

Williams et al. 1993 [64]

Survey, UK

12 malignant breast disease patients, 10 benign breast disease patients, 6 normal controls

Breast tissue PIs and PCs, GC analysis, precision not indicated

Histological diagnosis

None

8

ARA composition%, only shown as figure:

P

Breast tissue PIs: not significant

PCs:

Breast tissue PCs: significantly higher in control compared with benign and malignant group

Malignant/Control:

<0.02

Benign/Control:

<0.02

Eid et al. 1988 [66]

Survey, Israel

85 sequential patients (37 carcinoma, 27 fibroadenoma, 21 others)

Breast adipose tissue fatty acids, GC analysis, precision indicated

Not shown

None

8

ARA composition%, mean(SD)

P

Carcinoma:

Not significant

0.62(0.05)

Fibroadenoma:

0.78(0.18)