Longitudinal trajectories of diet quality and subsequent mortality among Chinese adults: results from the China health and nutrition survey 1997–2015

Table 3 Associations of trajectories of quintile of Chinese Healthy Eating Index during the 10-year period (1997–2006) with the subsequent risk of death in all respondents and subgroups (n = 6398)

	Trajectories of quintile of Chinese Healthy Eating Index
	Group 1 (Worsening)	Group 2 (Low-moderate-low)	Group 3 (Improving)	Group 4 (High-moderate-high)
All respondents
N, death/total	110/1469	178/1649	80/1434	114/1846
Basic model, HR (95% CI)^a	0.99 (0.78, 1.26)	Ref.	0.80 (0.61, 1.04)	0.79 (0.62, 1.002)
Multivariable model, HR (95% CI)^b	0.91 (0.71, 1.18)	Ref.	0.73 (0.55, 0.97)^*	0.76 (0.59, 0.98)^*
Baseline age
20 to < 65 years
N, death/total	75/1387	109/1474	54/1361	88/1762
Multivariable model, HR (95% CI)²	0.80 (0.59, 1.09)	Ref.	0.62 (0.44, 0.87)^**	0.74 (0.55, 0.99)^*
≥ 65 years
N, death/total	35/82	69/175	26/73	26/84
Multivariable model, HR (95% CI)^b	1.14 (0.70, 1.84)	Ref.	1.02 (0.61, 1.69)	0.85 (0.49, 1.48)
Sex
Male
N, death/total	60/713	104/815	44/688	73/913
Multivariable model, HR (95% CI)^b	0.75 (0.53, 1.04)	Ref.	0.58 (0.40, 0.85)^**	0.61 (0.44, 0.85)^**
Female
N, death/total	50/756	74/834	36/746	41/933
Multivariable model, HR (95% CI)^b	1.21 (0.80, 1.81)	Ref.	1.00 (0.64, 1.56)	1.07 (0.70, 1.64)
Baseline residence
Rural
N, death/total	21/389	50/465	18/398	25/615
Multivariable model, HR (95% CI)^b	1.25 (0.70, 2.23)	Ref.	0.95 (0.50, 1.80)	0.64 (0.36, 1.13)
Urban
N, death/total	89/1080	128/1184	62/1036	89/1231
Multivariable model, HR (95% CI)^b	0.87 (0.65, 1.15)	Ref.	0.71 (0.51, 0.98)^*	0.80 (0.60, 1.08)

^a Basic model adjusted for age and sex
^b Multivariable model adjusted for age, sex, residence (rural or urban), initial household income and changes in household income (all in quantile, inflated to 2015), changes in marital status, changes in physical activity, changes in smoke status, sleep duration (short, meet the US National Sleep Foundation’s recommendation, long), initial BMI (< 18.5, 18.5–23.9, 24–27.9, or ≥ 28), weight change (in quintiles), history of hypertension, and history of diabetes. The changes referred to the ten-year period (1997–2006)
^*P < 0.05, ^**P < 0.01, ^***P < 0.001

ISSN: 1479-5868