Table 4

Categorized duration of breastfeeding and timing of introduction of complementary foods and risk of IA (any IA or GADA)

IADietary exposure*Timing in months*Number of children who developed any IA or GADA, N (%)Number of children who did not develop any IA or GADA, N (%)Unadjusted HR (95% CI)P valueAdjusted HR (95% CI)P value
Any IAGluten-containing cereals<428 (6)445 (94)0.68 (0.47, 0.99)0.0470.67 (0.54, 0.98)0.037
4–9637 (10)6,048 (90)11
>931 (14)185 (86)1.57 (1.07, 2.31)0.0221.44 (0.97, 2.16)0.074
Any breastfeeding≤7334 (9)3,575 (91)0.93 (0.80, 1.08)0.3260.94 (0.81, 1.09)0.426
>7360 (11)3,042 (89)11
Cow’s milk≤5584 (10)5,922(91)0.83 (0.68, 1.00)0.0550.85 (0.69, 1.04)0.115
>5117 (11)902 (89)11
Cereals≤4483 (9)4,765 (91)0.99 (0.85, 1.17)0.9451.09 (0.92, 1.30)0.309
>4216 (10)2,045 (90)11
Rice≤7618 (9)6,196 (91)0.79 (0.63, 0.99)0.0460.87 (0.69, 1.10)0.233
>775 (13)502 (87)11
Fruit and berries≤4460 (9)4,552 (91)0.99 (0.85, 1.15)0.8681.00 (0.85, 1.18)0.999
>4240 (10)2,254 (90)11
Potato≤7578 (10)5,345 (90)1.19 (0.98, 1.45)0.0770.98 (0.77, 1.25)0.898
>7121 (8)1,353 (92)11
Meat≤8615 (10)5,762 (90)1.27 (1.01, 1.59)0.0401.13 (0.88, 1.45)0.344
>877 (8)914 (92)11
Egg≤9282 (9)3,002 (91)0.86 (0.74, 0.99)0.0450.84 (0.72, 0.99)0.035
>9402 (10)3,456 (90)11
GADAAny breastfeeding≤6120 (3)3,476 (97)1.20 (0.95, 1.52)0.1261.17 (0.92, 1.49)0.196
>6174 (5)3,541 (95)11
Rice≤6264 (4)6,435 (96)1.25 (0.87, 1.78)0.2241.10 (0.76, 1.58)0.623
>628 (4)664 (96)11
Fish and seafood≤9188 (4)4,538 (96)0.88 (0.70, 1.12)0.2901.11 (0.81, 1.53)0.507
>992 (4)2,054 (96)11
  • *Dietary exposure: timing of breastfeeding or food; categorization of timing of gluten-containing cereals’ introductions was based on martingale residuals whereas dichotomizations of timing of other dietary exposures in relation to the risk of any IA or GADA were based on change-point methods using log-rank test (28); only statistically significant (P < 0.05) cut points are shown. No statistically significant cut point of timing of a dietary exposure for IAA was detected.

  • †The Cox regression model was adjusted for country, HLA genotype, FDR status, sex of the child, and probiotic use <28 days.