Table 4

Association between total alcohol and wine intake and ovarian borderline tumors, OCAC studies
Intake/d Controls N=10,358 Serous borderline N=818 Mucinous borderline N=561
Total alcohol * Ca OR (95% CI) Ca OR (95% CI) P value†
 None 4,296 297 1.0 (Ref) 181 1.0 (Ref)
 Up to 1 drink 3,928 349 0.87 (0.72-1.05) 245 0.95 (0.76-1.20)
 1-2 drinks 1,112 86 0.91 (0.69-1.20) 59 0.86 (0.62-1.20)
 2-3 drinks 400 32 1.04 (0.69-1.56) 23 0.94 (0.58-1.51)
 >3 drinks 622 54 1.19 (0.86-1.66) 53 1.40 (0.99-1.20)
P trend 0.45 0.22 0.39
 >2 drinks § 1,022 86 1.13 (0.86-1.49) 76 1.22 (0.90-1.66)
P trend § 0.63 0.42 0.67
Wine
 None 5,307 405 1.0 (Ref) 263 1.0 (Ref)
 Up to 4 oz 3,984 336 0.89 (0.74-1.06) 236 0.89 (0.72-1.10)
 4-8 oz 522 38 0.99 (0.68-1.43) 26 0.88 (0.57-1.36)
 >8 oz 545 39 0.99 (0.69-1.44) 36 1.03 (0.69-1.52)
P trend 0.66 0.72 0.87
 >4 oz § 1,067 77 0.99 (0.75-1.31) 62 0.96 (0.70-1.31)
P trend § 0.52 0.52 0.69

Adjusted for age (<40; 40-49; 50-59; 60-69; 70+ years), smoking status (never, former, current), site (AUS, CON, DOV, GER, HAW, HOP, MAL, NCO, NEC, NJO, POL, SON), race/ethnicity (white nonHispanic; white Hispanic; black non Hispanic; Asian; other or unknown); menopausal status (pre/peri-menopausal; postmenopausal, unknown or missing), oral contraceptive use (<6mo, 6-22 mo, 23+ mo, unknown or missing), tubal ligation (yes; no; unknown or missing), endometriosis (yes; no; unknown or missing), hysterectomy (yes; no; unknown or missing), family history of breast or ovarian cancer in first-degree relatives (no; yes; unknown; no daughters or sisters), parity/age at last birth (nulliparous; 1-2 births/age ≤25 yrs at last pregnancy; 3+ births/age ≤25 yrs at last pregnancy; 1-2 births/age >25 years at last pregnancy; 3+ births/age >25 years at last pregnancy; yes if ever pregnant but unknown or missing age at last pregnancy age; no or unknown if ever pregnant and missing age at last pregnancy, interview year (1990-1994; 1995-1999; 2000-2004; 2005-2009; missing), age at menarche (8-10 yrs; 11 yrs; 12 yrs; 13 yrs; 14-21 yrs; <8 or ≥ 22 yrs), body mass index (continuous) and education (less than high school, high school, some college, completed college or university, completed graduate or professional degree, missing). Wine consumption was additionally adjusted for other alcoholic beverage types. Models include interaction terms between site and each covariate except alcohol.

* 1 drink = 10 grams ethanol.

P for tumor heterogeneity derived from testing the trend variable for alcohol or wine intake in polytomous regression models with 2 df (see Statistical analysis).

§ Risk estimates and P trend values are from models that collapse the two highest intake categories.

Models are also simultaneously adjusted for consumption of beer and liquor intake.

Kelemen et al.

Kelemen et al. BMC Cancer 2013 13:28   doi:10.1186/1471-2407-13-28

Open Data