Skip to main content
letter
. 2012 Nov 20;108(1):188–192. doi: 10.1038/bjc.2012.521

Table 2. HRs of overall mortality, breast cancer-specific mortality, and other mortality according to quartiles of dietary patterns in the MARIE study, Germany, 2001–2009.

    Overall mortality
Breast cancer mortality
Other mortality
Quartiles of dietary pattern No. of subjects No. of deaths HR 95% CI No. of deaths HR 95% CI No. of deaths HR 95% CI
‘Healthy’ pattern
Model 1a
  Q1 631 101 1.00   75 1.00   26 1.00  
  Q2 630 68 0.71 0.52, 0.97 50 0.67 0.47, 0.96 18 0.84 0.45, 1.54
  Q3 631 74 0.77 0.56, 1.04 57 0.76 0.54, 1.09 17 0.77 0.41, 1.45
  Q4 630 73 0.77 0.56, 1.05 53 0.73 0.51, 1.05 20 0.89 0.49, 1.63
  P-trend     0.06   0.04   0.80
                     
Model 2b
  Q1 614 97 1.00   72 1.00   25 1.00  
  Q2 616 62 0.77 0.55, 1.08 46 0.76 0.51, 1.14 16 0.81 0.41, 1.57
  Q3 617 72 0.81 0.58, 1.13 55 0.83 0.56, 1.23 17 0.82 0.42, 1.58
  Q4 609 68 0.87 0.61, 1.23 50 0.89 0.59, 1.35 18 0.81 0.40, 1.61
  P-trend     0.20   0.25   0.66
                     
‘Unhealthy’ pattern
Model 1a
  Q1 631 72 1.00   61 1.00   11 1.00  
  Q2 630 69 0.95 0.68, 1.32 48 0.77 0.52, 1.13 21 1.96 0.93, 4.12
  Q3 631 76 1.06 0.76, 1.47 60 1.00 0.69, 1.43 16 1.42 0.65, 3.08
  Q4 630 99 1.39 1.02. 1.89 66 1.06 0.74, 1.50 33 3.41 1.69, 6.85
  P-trend     0.01   0.44   <0.001
                     
Model 2b
  Q1 619 69 1.00   59 1.00   10 1.00  
  Q2 615 67 1.03 0.72, 1.47 46 0.88 0.58, 1.33 21 2.06 0.94, 4.54
  Q3 611 72 1.11 0.77, 1.59 58 1.07 0.71, 1.61 14 1.46 0.63, 3.39
  Q4 611 91 1.34 0.93, 1.94 60 0.99 0.64, 1.52 31 3.69 1.66, 8.17
  P-trend     0.03   0.59   <0.001

Abbreviations: CI=confidence interval; ERPR=oestrogen receptor/progesterone receptor; HR=hazard ratio; HRT=hormone replacement therapy; Q=quartile.

a

The model was stratified by age at diagnosis and study centre.

b

The model was stratified by age at diagnosis and study centre, and adjusted for tumour size, nodal status, metastases, tumour grade, ERPR status, radiotherapy, HRT use at diagnosis, mode of detection, and total energy intake; the model for other mortality was additionally adjusted for cardiovascular disease; because of missing covariate values, 66 observations were not included in model 2. Other potentially confounding variables, as specified in Supplementary Table 1, were not statistically significant and did not change the risk estimates by ⩾10% when tested in the model and were therefore not included in the final model.