Table 2.

Crude and adjusted HRs of epithelial ovarian cancer death according to anthropometric variables (n = 207)

Anthropometric variablesNo. cases% DeadCrude HR (95% CI)Adjusted HR (95% CI)
BMI at diagnosis (kg/m2)
    <20.040451.01.0
    20.0-22.475360.79 (0.44-1.44)0.60 (0.32-1.14)
    22.5-24.953531.23 (0.68-2.23)0.65 (0.34-1.24)
    ≥25.039541.15 (0.61-2.17)0.76 (0.38-1.52)
    χ12 for trend*0.84 (P = 0.36)0.52 (P = 0.47)
BMI at 5 y before diagnosis (kg/m2)
    <20.046281.01.0
    20.0-22.464451.99 (1.04-3.85)1.79 (0.90-3.55)
    22.5-24.954441.78 (0.91-3.50)1.71 (0.84-3.46)
    ≥25.043653.32 (1.72-6.42)2.33 (1.12-4.87)
    χ12for trend*11.39 (P < 0.01)6.13 (P = 0.01)
BMI at age 21 years (kg/m2)
    <20.068381.01.0
    20.0-22.487441.28 (0.78-2.11)0.99 (0.59-1.69)
    22.5-24.942571.83 (1.05-3.19)1.03 (0.56-1.89)
    ≥25.010602.60 (1.07-6.33)3.31 (1.26-8.73)
    χ12for trend*5.37 (P = 0.02)1.54 (P = 0.22)
  • NOTE: Estimates from Cox regression models included terms for age at diagnosis (in y, continuous), total energy intake (kcal, continuous), menopausal status (premenopause and postmenopause), Fesddration Internationale des Gynaecologistes et Obstetristes stage (I, II, III, and IV), histopathologic grade of differentiation (well, moderately, poorly differentiated, and not available), ascites (no/yes), residual lesions (<2 and ≥2 cm), and chemotherapy (no/yes).

  • * Quantity of anthropometric variables was tested for linear trend.