Leading dietary determinants identified using machine learning techniques and a healthy diet score for changes in cardiometabolic risk factors in children: a longitudinal analysis

Shang, Xianwen; Li, Yanping; Xu, Haiquan; Zhang, Qian; Liu, Ailing; Du, Songming; Guo, Hongwei; Ma, Guansheng

doi:10.1186/s12937-020-00611-2

Table 3 Changes in cardiometabolic risk factors during follow-up associated with Healthy Diet Score at baseline

From: Leading dietary determinants identified using machine learning techniques and a healthy diet score for changes in cardiometabolic risk factors in children: a longitudinal analysis

	Healthy Diet Score						P-trend*
	≤3	4	5	6	7	≥8
Change in BMI^†
Participants	842	1011	1307	1381	766	298
β (95% CI), Model 1^‡	0	−0.03 (− 0.08, 0.03)^§	− 0.06 (− 0.11, − 0.00)	− 0.07 (− 0.12, − 0.01)	−0.08 (− 0.14, − 0.02)	−0.09 (− 0.17, − 0.02)	0.0004
β (95% CI), Model 2	0	− 0.02 (− 0.08, 0.04)	−0.06 (− 0.11, − 0.00)	−0.07 (− 0.13, − 0.01)	−0.07 (− 0.13, − 0.01)	−0.09 (− 0.17, − 0.02)	0.0007
β (95% CI), Model 3	0	− 0.02 (− 0.08, 0.04)	− 0.06 (− 0.11, − 0.00)	−0.06 (− 0.12, − 0.01)	−0.07 (− 0.13, − 0.01)	− 0.08 (− 0.16, − 0.00)	0.0041
Change in WC
Participants	844	1004	1304	1374	764	299
β (95% CI), Model 1	0	−0.06 (− 0.11, − 0.01)	−0.01 (− 0.06, 0.03)	− 0.04 (− 0.08, 0.01)	−0.04 (− 0.09, 0.01)	−0.05 (− 0.12, 0.01)	0.18
β (95% CI), Model 2	0	−0.05 (− 0.10, − 0.01)	−0.01 (− 0.06, 0.03)	−0.04 (− 0.09, 0.00)	− 0.03 (− 0.08, 0.01)	−0.04 (− 0.10, 0.02)	0.23
β (95% CI), Model 3	0	−0.05 (− 0.10, − 0.01)	−0.02 (− 0.06, 0.03)	−0.04 (− 0.09, 0.00)	− 0.04 (− 0.08, 0.01)	− 0.05 (− 0.11, 0.01)	0.15
Change in PBF
Participants	813	977	1275	1349	743	294
β (95% CI), Model 1	0	0.01 (− 0.06, 0.08)	0.00 (− 0.07, 0.07)	− 0.08 (− 0.15, − 0.01)	−0.12 (− 0.19, − 0.04)^bc	−0.08 (− 0.18, 0.01)	< 0.0001
β (95% CI), Model 2	0	−0.00 (− 0.07, 0.07)	−0.01 (− 0.08, 0.06)	− 0.10 (− 0.17, − 0.03)	−0.14 (− 0.21, − 0.06)^abc	− 0.10 (− 0.19, − 0.00)	< 0.0001
β (95% CI), Model 3	0	0.00 (− 0.07, 0.07)	−0.01 (− 0.08, 0.06)	−0.10 (− 0.17, − 0.03)	− 0.13 (− 0.21, − 0.06)^bc	− 0.09 (− 0.18, 0.01)	< 0.0001
Change in SBP
Participants	847	1010	1307	1369	760	298
β (95% CI), Model 1	0	−0.08 (− 0.18, 0.01)	− 0.18 (− 0.27, − 0.09)^a	−0.33 (− 0.42, − 0.24)^abc	−0.42 (− 0.52, − 0.32)^abc	−0.53 (− 0.66, − 0.41)^abcd	< 0.0001
β (95% CI), Model 2	0	− 0.07 (− 0.17, 0.02)	−0.17 (− 0.25, − 0.08)^a	−0.29 (− 0.38, − 0.21)^ab	−0.39 (− 0.48, − 0.29)^abc	−0.48 (− 0.60, − 0.36)^abc	< 0.0001
β (95% CI), Model 3	0	− 0.07 (− 0.16, 0.02)	−0.16 (− 0.25, − 0.07)	−0.28 (− 0.37, − 0.19)^ab	−0.37 (− 0.46, − 0.27)^abc	−0.46 (− 0.58, − 0.34)^abc	< 0.0001
Change in DBP
Participants	848	1010	1309	1373	760	298
β (95% CI), Model 1	0	−0.03 (−0.12, 0.07)	−0.17 (− 0.26, − 0.08)^a	−0.36 (− 0.44, − 0.27)^abc	−0.45 (− 0.54, − 0.35)^abc	−0.49 (− 0.61, − 0.37)^abc	< 0.0001
β (95% CI), Model 2	0	− 0.03 (− 0.12, 0.07)	−0.16 (− 0.25, − 0.07)	−0.32 (− 0.41, − 0.23)^abc	−0.43 (− 0.53, − 0.33)^abc	−0.46 (− 0.59, − 0.34)^abc	< 0.0001
β (95% CI), Model 3	0	−0.03 (− 0.12, 0.07)	−0.16 (− 0.25, − 0.07)	−0.32 (− 0.41, − 0.23)^abc	−0.43 (− 0.52, − 0.33)^abc	−0.46 (− 0.58, − 0.34)^abc	< 0.0001
Change in MAP
Participants	847	1010	1307	1371	759	298
β (95% CI), Model 1	0	−0.05 (−0.14, 0.05)	−0.19 (− 0.28, − 0.10)^a	−0.37 (− 0.46, − 0.28)^abc	−0.47 (− 0.56, − 0.37)^abc	−0.55 (− 0.67, − 0.42)^abc	< 0.0001
β (95% CI), Model 2	0	− 0.04 (− 0.14, 0.05)	−0.18 (− 0.26, − 0.09)^a	−0.34 (− 0.42, − 0.25)^abc	−0.44 (− 0.54, − 0.35)^abc	−0.51 (− 0.63, − 0.38)^abc	< 0.0001
β (95% CI), Model 3	0	−0.04 (− 0.14, 0.05)	−0.17 (− 0.26, − 0.09)^a	−0.33 (− 0.42, − 0.24)^abc	−0.43 (− 0.53, − 0.34)^abc	−0.50 (− 0.62, − 0.38)^abc	< 0.0001
Change in TC
Participants	798	945	1227	1306	719	278
β (95% CI), Model 1	0	−0.01 (−0.08, 0.07)	0.05 (−0.01, 0.12)	0.19 (0.13, 0.26)^abc	0.20 (0.13, 0.28)^abc	0.21 (0.12, 0.31)^abc	< 0.0001
β (95% CI), Model 2	0	−0.01 (− 0.09, 0.06)	0.03 (− 0.04, 0.10)	0.17 (0.10, 0.24)^abc	0.18 (0.11, 0.26)^abc	0.20 (0.11, 0.30)^ab	< 0.0001
β (95% CI), Model 3	0	−0.01 (− 0.08, 0.06)	0.03 (− 0.04, 0.10)	0.17 (0.10, 0.23)^abc	0.17 (0.09, 0.24)^abc	0.18 (0.08, 0.27)^b	< 0.0001
Change in HDL-C
Participants	798	943	1223	1309	720	278
β (95% CI), Model 1	0	0.02 (−0.08, 0.12)	0.05 (− 0.05, 0.15)	0.28 (0.18, 0.37)^abc	0.42 (0.31, 0.53)^abc	0.37 (0.23, 0.50)^abc	< 0.0001
β (95% CI), Model 2	0	0.01 (−0.10, 0.11)	0.03 (−0.07, 0.12)	0.21 (0.11, 0.31)^abc	0.36 (0.26, 0.47)^abc	0.30 (0.17, 0.44)^abc	< 0.0001
β (95% CI), Model 3	0	0.01 (−0.09, 0.11)	0.02 (−0.07, 0.12)	0.20 (0.11, 0.30)^bc	0.36 (0.25, 0.46)^abc	0.28 (0.15, 0.42)^abc	< 0.0001
Change in LDL-C
Participants	797	945	1229	1307	718	279
β (95% CI), Model 1	0	0.02 (−0.05, 0.10)	0.12 (0.05, 0.20)	0.30 (0.22, 0.37)^abc	0.42 (0.34, 0.50)^abc	0.47 (0.37, 0.58)^abcd	< 0.0001
β (95% CI), Model 2	0	0.03 (−0.05, 0.11)	0.12 (0.05, 0.20)	0.32 (0.24, 0.39)^abc	0.42 (0.34, 0.50)^abc	0.49 (0.38, 0.59)^abc	< 0.0001
β (95% CI), Model 3	0	0.03 (−0.05, 0.11)	0.11 (0.04, 0.19)	0.30 (0.22, 0.37)^abc	0.39 (0.31, 0.47)^abc	0.44 (0.34, 0.55)^abc	< 0.0001
Change in TG
Participants	798	945	1228	1308	720	279
β (95% CI), Model 1	0	−0.02 (−0.12, 0.07)	0.00 (−0.08, 0.09)	− 0.06 (− 0.15, 0.03)	−0.12 (− 0.22, − 0.03)	−0.04 (− 0.17, 0.08)	0.50
β (95% CI), Model 2	0	−0.02 (− 0.11, 0.07)	0.01 (− 0.08, 0.09)	−0.03 (− 0.11, 0.06)	−0.10 (− 0.19, − 0.00)	0.00 (− 0.12, 0.12)	0.96
β (95% CI), Model 3	0	−0.02 (− 0.11, 0.07)	0.01 (− 0.08, 0.09)	−0.03 (− 0.12, 0.06)	−0.11 (− 0.20, − 0.01)	−0.01 (− 0.13, 0.11)	0.49
Change in fasting glucose
Participants	796	943	1229	1311	720	278
β (95% CI), Model 1	0	−0.10 (−0.18, − 0.02)	−0.18 (− 0.26, − 0.11)^a	−0.25 (− 0.32, − 0.17)^ab	−0.23 (− 0.31, − 0.15)^a	−0.22 (− 0.33, − 0.12)^a	< 0.0001
β (95% CI), Model 2	0	−0.09 (− 0.17, − 0.01)	−0.16 (− 0.23, − 0.08)^a	−0.21 (− 0.29, − 0.14)^ab	−0.21 (− 0.29, − 0.13)^ab	−0.21 (− 0.32, − 0.11)^a	< 0.0001
β (95% CI), Model 3	0	−0.09 (− 0.16, − 0.01)	−0.16 (− 0.23, − 0.08)^a	−0.21 (− 0.29, − 0.14)^ab	−0.21 (− 0.29, − 0.13)^ab	−0.22 (− 0.32, − 0.11)^a	< 0.0001
Change in insulin
Participants	700	810	1062	1148	667	257
β (95% CI), Model 1	0	−0.08 (−0.24, 0.07)	−0.13 (− 0.28, 0.01)	−0.35 (− 0.49, − 0.21)^abc	−0.56 (− 0.72, − 0.40)^abc	−0.64 (− 0.84, − 0.44)^abc	< 0.0001
β (95% CI), Model 2	0	− 0.07 (− 0.22, 0.08)	−0.09 (− 0.23, 0.05)	−0.28 (− 0.42, − 0.13)	−0.48 (− 0.63, − 0.33)^abc	−0.55 (− 0.74, − 0.36)^abc	< 0.0001
β (95% CI), Model 3	0	− 0.06 (− 0.20, 0.09)	−0.07 (− 0.21, 0.07)	−0.25 (− 0.39, − 0.11)	−0.45 (− 0.60, − 0.30)^abc	−0.52 (− 0.71, − 0.32)^abc	< 0.0001
Change in HOMA-IR
Participants	699	810	1062	1148	667	256
β (95% CI), Model 1	0	−0.10 (−0.25, 0.04)	−0.17 (− 0.31, − 0.03)	−0.39 (− 0.53, − 0.25)^abc	−0.59 (− 0.74, − 0.44)^abc	−0.67 (− 0.86, − 0.48)^abc	< 0.0001
β (95% CI), Model 2	0	− 0.09 (− 0.24, 0.05)	−0.13 (− 0.27, 0.01)	−0.32 (− 0.46, − 0.18)^a	−0.52 (− 0.67, − 0.37)^abc	−0.58 (− 0.77, − 0.39)^abc	< 0.0001
β (95% CI), Model 3	0	− 0.08 (− 0.22, 0.07)	−0.11 (− 0.25, 0.02)	−0.30 (− 0.43, − 0.16)^a	−0.49 (− 0.63, − 0.34)^abc	−0.55 (− 0.73, − 0.36)^abc	< 0.0001
Change in CMRS
Participants	713	852	1118	1198	668	265
β (95% CI), Model 1	0	−0.25 (−0.48, − 0.01)	−0.38 (− 0.61, − 0.16)	−0.94 (− 1.16, − 0.72)^abc	−1.22 (− 1.46, − 0.97)^abc	−1.21 (− 1.51, − 0.90)^abc	< 0.0001
β (95% CI), Model 2	0	−0.21 (− 0.44, 0.01)	−0.32 (− 0.53, − 0.11)	−0.79 (− 1.00, − 0.58)^abc	− 1.10 (− 1.33, − 0.87)^abc	−1.05 (− 1.34, − 0.76)^abc	< 0.0001
β (95% CI), Model 3	0	−0.22 (− 0.44, 0.01)	−0.31 (− 0.52, − 0.10)	−0.77 (− 0.98, − 0.55)^abc	− 1.08 (− 1.31, − 0.85)^abc	−1.02 (− 1.31, − 0.73)^abc	< 0.0001

BMI, body mass index; CMRS, cardiometabolic risk score; DBP, diastolic blood pressure; HOMA-IR, homeostatic model assessment of insulin resistance; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MAP, mean arterial pressure; SBP, systolic blood pressure; SE, standard error; TC, total cholesterol; TG, triglyceride
*GLM was used to estimate beta coefficients (β) and 95% CIs of cardiometabolic risk factors between quintiles. Benjamin-Hochberg’s procedure was used to control the false discovery rate at level 5% for multiple comparisons with the P-value cut-off point of significance was 0.0433 for HDS and changes in CMR factors (Model 3)
^†Changes in CMR factors were calculated by subtracting the results at baseline from those at follow-up
^‡Model 1 was adjusted for classes in school as clustering effects and characteristics of individuals including age, sex, and corresponding CMR factor at baseline as fixed effects; Model 2 was adjusted for Model 1 plus puberty, grade, intervention, BMI, physical activity, and intake of energy, fiber, vegetable, fruit, pork, legumes, and nuts at baseline as fixed effects; Model 3 was adjusted for Model 2 plus birthweight, household income, mother’s education, father’s education, mother’s BMI, and father’s BMI as fixed effects
^§All these data are β (95% CI) of changes in CMR factors
^abcdBonferroni Post-hoc test was used to examine the difference between every two groups of the healthy diet score with ^a indicating significance compared with HDS ≤ 3, ^b indicating significance compared with HDS = 4, ^c indicating significance compared with HDS = 5, and ^d indicating significance compared with HDS = 6. The comparisons with HDS = 7 were also conducted, but no significant associations were found

Back to article page

ISSN: 1475-2891

Contact us

General enquiries: journalsubmissions@springernature.com

Nutrition Journal

Contact us