Table 3

Logistic regression analyses of foods with large coefficients in the fruit and vegetable pattern score for hip

Food item

Model 1

Model 2

OR (95% CI)

OR (95% CI)


Allium vegetables1

0.748 (0.577,0.971)

0.702 (0.506,0.973)

Cruciferous vegetables2

0.819 (0.640,1.049)

0.719 (0.466,1.110)

Green leafy vegetables3

1.024 (0.810,1.296)

1.051 (0.725,1.525)

Yellow vegetables4

0.918 (0.732,1.150)

1.117 (0.751,1.661)

Other vegetables5

0.962 (0.736,1.257)

0.907 (0.614,1.339)

Citrus fruit6

0.858 (0.570,1.293)

0.784 (0.565,1.089)

Non-citrus fruit7

0.676 (0.493,0.927)

0.560 (0.396,0.790)

Chips and roast potatoes

1.173 (0.959,1.434)

1.314 (0.940,1.836)


1 garlic, onions and leeks 2 cabbage, broccoli, Brussels sprouts, cauliflower and coleslaw 3 green salad and spinach 4 carrots and tomatoes 5 parsnips, marrow, watercress, sweetcorn, avocado, beetroot, mushrooms and sweet peppers 6 oranges and grapefruit 7 apples, bananas, pears, peaches, melon, grapes, tinned fruit and dried fruit

Model 1: individual food intakes only; model 2: OR shown for individual food intakes, but obtained with twin pair mean intakes in the same model

All models also adjusted for age, BMI and physical activity; Kellgren-Lawrence scores coded as 0 and 1 versus 2+

All foods listed had high positive loadings in the fruit and vegetable pattern, except chips and roast potatoes, which had a high negative loading

Williams et al. BMC Musculoskeletal Disorders 2010 11:280   doi:10.1186/1471-2474-11-280

Open Data