Dairy consumption and its association with metabolic disturbance in a prospective study of urban adults

Abstract

The role of dairy foods and related nutrients in cardio-metabolic health etiology is poorly understood. We investigated longitudinal associations between the metabolic syndrome (MetS) and its components with key dairy exposures. We used prospective data from a bi-racial cohort of urban adults [30–64y at baseline (N=1,371)], the Healthy Aging in Neighborhoods of Diversity across the Life Span (HANDLS), in Baltimore City, MD (2004–2013). The average of 2 24-hour dietary recalls measured 4–10d apart was computed at baseline (V1) and follow-up (V2) waves. Annual rates of change (Δ) in dairy foods and key nutrients were estimated. Incident obesity, central obesity and metabolic syndrome (MetS) were determined. Among key findings, in the overall urban adult population, both cheese and yogurt (V1 and Δ) were associated with an increased risk of central obesity [HR=1.13, 95% CI:1.05,1.23 per oz equivalent of cheese (V1); HR=1.21, 95%CI: 1.01, 1.44, per fl oz equivalent of yogurt (V1)]. Baseline fluid milk intake (V1 in cup equivalents) was inversely related to MetS [HR=0.86, 95%CI: 0.78,0.94], specifically to dyslipidemia-triacylglycerol (TA) [HR=0.89, 95% CI:0.81,0.99], though it was directly associated with dyslipidemia-High Density Lipoprotein-Cholesterol (HDL-C), [HR=1.10, 95% CI:1.01,1.21]. Furthermore, Δcalcium and Δphosphorus were inversely related to dyslipidemia-HDL and MetS incidence, respectively, while Δdairy fat was positively associated with incident TA- and HDL-C-dyslipidemias and MetS. A few of those associations were sex- and race-specific. In sum, various dairy exposures had differential associations with metabolic disturbances. Future intervention studies should uncover how over-time changes in dairy components may affect metabolic disorders.

INTRODUCTION

The metabolic syndrome (MetS) is a clustering of cardio-metabolic risk factors, namely central obesity, hyperglycemia, hypertension and dyslipidemia [hypertriglyceridemia and reduced high-density lipoprotein cholesterol (HDL-C)].⁽¹⁾ Increasing CVD and type 2 diabetes risk by 1.7- and 5-folds^{(2; 3)} respectively, MetS is a threat to public health, with rising all-cause mortality rates, disability and health care costs.^{(4; 5; 6; 7; 8; 9; 10; 11; 12)} Dairy consumption’s effect on MetS remains controversial.⁽¹³⁾ Among dairy constituents, saturated fat shows deleterious effect on weight and cardiovascular disease,^{(14; 15; 16; 17; 18)} while calcium and magnesium may carry beneficial effects.^{(13; 19; 20; 21; 22; 23; 24; 25; 26; 27)} Notably, dietary calcium, a key weight regulator, affects adipocyte intracellular calcium concentration, thus decreasing fatty acid synthesis, while upregulating lipolysis and reducing net triglyceride stores.^{(24; 28)}

Most guidelines recommend 2–3 dairy servings/day, a goal unreachable by many US adults.⁽²³⁾ Optimal dairy intake may prevent adverse health outcomes and related risk factors, including obesity, central obesity and MetS.⁽²³⁾ Recent observational and experimental studies suggest that dairy and calcium consumption may reduce obesity risk,^{(29; 30)} excess central⁽³¹⁾ fat distribution, type-2 diabetes ^{(32; 33)}, hypertension⁽³⁴⁾, and the MetS ^{(31; 34; 35; 36; 37; 38; 39; 40; 41; 42; 43; 44; 45; 46; 47; 48; 49; 50; 51; 52; 53; 54; 55; 56; 57; 58; 59; 60; 61; 62; 63; 64; 65; 66; 67; 68)}, while mixed or negative finding were reported by others.^{(69; 70; 71; 72; 73; 74; 75; 76; 77; 78; 79; 80; 81; 82; 83)}

To our knowledge, our study is the first to assess, in an urban population, the association between consumption of dairy and related nutrients and obesity, central obesity and MetS, with repeated measures on dietary and metabolic parameters. We further examined socio-demographic correlates of dairy foods, dairy-related nutrient intakes and metabolic disturbances. Finally, we tested gender and race-specific associations between dairy intake and metabolic disturbances.

METHODS

Database

Initiated in 2004, the Healthy Aging in Neighborhoods of Diversity across the Life Span (HANDLS) study is a prospective, population based, longitudinal study. The sample is a fixed cohort of participants based on household screenings from an area probability sample of 13 neighborhoods (areas of contiguous census tracts) in Baltimore City, Maryland. Neighborhoods were selected as contiguous groups of census tracts that were likely to yield sufficient numbers of participants to fill a 4-way design of race, sex, age, and socioeconomic status assessed by 125% of the Federal poverty level. Recruitment and sampling contractors produced household listings to identify residential dwellings in each neighborhood. The contractors performed doorstep interviews, identified eligible persons in each household, selected 1 of 2 eligible persons per household and invited the eligible candidates to participate in HANDLS. Participants had to be aged 30–64 years, to self-identify as White or African-American, have the ability to give informed consent, perform at least 5 study measures, and present valid picture identification. Individuals were excluded from the study if they were pregnant, within 6 months of active cancer treatment, or multiethnic individuals who did not identify strongly with either the Black or White race. ⁽⁸⁴⁾ The present study uses baseline visit 1 (V1: 2004–2009) and the first follow-up visit 2 (V2: 2009–2013). All participants provided written informed consent, after accessing a protocol booklet in layman’s terms and a video detailing all procedures and future re-contacts. HANDLS study was ethically approved by the National Institute on Environmental Health Sciences, National Institutes of Health Institutional Review Board.

Study participants

Of the original HANDLS sample (N=3,720), 24-hour dietary recall data were collected for each of the two visits (i.e. V1 and V2) for 1,513 participants (Sample 2c, Figure 1). Among those, data was complete on metabolic outcomes at each of the two visits as outlined in Figure 1 (Samples 3a–3h). The final analytic samples consisted of individuals with complete data on dietary intakes and metabolic outcomes at both visits (Sample 5h, Figure 1: N=1,371), and metabolic disturbance-free participants for each of the metabolic outcomes [Samples 6a–6b (N=588–859) and 7a–7d (N=915–1,171), Figure 1]. Mean follow-up time±SE was estimated at 4.62y±0.95 (range: 0.42–8.20).

FIGURE 1.

Participant flow chart

Dietary assessment

At each visit, the average nutrient and food group intakes from 2 24-hour dietary recalls were estimated. Each 24-hour dietary recall was obtained using the US Department of Agriculture (USDA) Automated Multiple Pass Method, a computerized structured interview⁽⁸⁵⁾ utilizing measurement aids (e.g. cups, spoons, ruler, illustrated Food Model Booklet). At first-visit, both recalls were administered in-person by trained interviewers, 4–10 days apart, while at follow-up (visit 2), the second recall was administered using telephone interviews. Using Survey Net, trained nutrition professionals matched foods consumed with 8-digit codes from the Food and Nutrient Database for Dietary Studies version 3.0,⁽⁸⁶⁾ and My pyramid equivalents database was used to create food groups (MPED 2: http://www.ars.usda.gov/SP2UserFiles/Place/80400530/pdf/mped/mped2_doc.pdf).

Dietary exposures

Dietary exposures of interest included: (1) Dairy foods, namely, total dairy intake (servings/d), total fluid milk intake (servings/d), total cheese intake (servings/d) and total yogurt intake (servings/d). One serving of dairy is calculated in terms of cup equivalents. For milk, a serving is 1 cup, while for cheese it ranges between 1.5 oz for hard cheese to 2 cups for ricotta cheese. For yogurt, on the other hand, a serving is 1 cup or 8 fl oz (2) Dairy-related nutrients, namely calcium (mg/d), magnesium (mg/d), phosphorus (mg/d) and dairy fat % of total fat [myristic acid (14:0)×100/total fat]. Fluid milk was also categorized into whole vs. reduced fat milk (grams/d), while DRIs of calcium and total dairy were estimated. Key exposures were measured as baseline (V1) values and annualized rates of change [i.e. Δdairy=(dairy_follow-dairy_base)/(Age_follow-Age_base)].

Anthropometric measures and metabolic outcome variables

Body mass index and waist circumference

Body Mass Index (BMI=weight/height², kg/m²) was calculated for each participant using measured weight and height. Waist circumference [WC (in cm.)] was measured using a tape measure starting from the hip bone and wrapping around the waist at the level of the navel.

Systolic and diastolic blood pressure

Systolic and diastolic blood pressure levels (SBP and DBP) were measured by averaging right and left sitting non-invasive assessments using brachial artery auscultation with an aneroid manometer, a stethoscope, and an inflatable cuff.

Other metabolic risk factors

Following an overnight fast (8–12 hours), blood was drawn and collected from an antecubital vein. Total cholesterol, High density lipoprotein-cholesterol (HDL-C), Triacylglycerols (TA), and fasting glucose (FG) were assessed using a spectrophotometer (Olympus 5400).

Classification of key health outcomes

Obesity was defined as BMI≥30 kg/m²; and central obesity, waist circumference (WC) ≥ 102 cm or 40 inches (men), ≥ 88 cm or 35 inches (women) ⁽⁸⁷⁾.

Participants were classified as MetS-positive if they screened positive on at least 3 of 5 conditions⁽¹⁾: (1) central obesity (see above); (2) dyslipidemia: TA≥1.695 mmol/L (150 mg/dl); (3) dyslipidemia: HDL-C<40 mg/dL (male), <50 mg/dL (female); (4) blood pressure≥130/85 mmHg; (5) fasting plasma glucose≥6.1 mmol/L (110 mg/dl).⁽⁸⁸⁾ Similarly, continuous annual rates of change (Δ) in metabolic outcomes were considered, namely number of metabolic disturbances (MetD), BMI, WC, SBP, DBP, TA, HDL-C, and Glucose. Binary incident outcomes, were obesity, central obesity, MetS and other metabolic disturbance (i.e. hypertension, dyslipidemia-TA, dyslipidemia-HDL and hyperglycemia).

Covariates

Covariates included in our analyses were baseline age, sex, race, poverty status, education, self-rated health, smoking and drug use among fixed or baseline covariates. Annual rates of change (Δ) in covariates were considered, except when baseline dairy exposures were examined. Those were total energy intake (kcal/d), caffeine intake (mg/d), and MyPyramid equivalents of total fruit, dark green and orange vegetables, whole and non-whole grains, legumes, nuts/seeds, soy, total meat/poultry/fish, eggs, discretionary solid fat and oils (grams), added sugars (teaspoons) and alcoholic beverages (servings).

Statistical analyses

Using Stata release 14.0,⁽⁸⁹⁾ we first described the sex and race differences in dairy consumption and metabolic outcomes, comparing means using independent samples t-tests, and testing associations with χ² tests. Second, Cox proportional hazards (PH) regression models were fit to test independent associations of socio-demographic factors with dairy consumption and incident metabolic outcomes.

Importantly, two sets of models included as exposures dairy foods (model 1) and dairy-related nutrients (model 2), respectively. Cox PH models tested associations of baseline dairy (V1) and Δdairy exposures with incident binary metabolic outcomes. To account for potential selection bias in our multivariate models due to the non-random selection of participants with complete data from the target study population, a 2-stage Heckman selection process was used. ⁽⁹⁰⁾ A probit model was constructed to obtain an inverse mills ratio at the first stage (derived from the predicted probability of being selected, conditional on the covariates in the probit model, mainly baseline age, sex, race, poverty status and education), as was done in earlier studies.^{(91; 92; 93)} This inverse mills ratio was then entered as a covariate in the main models to adjust for sample selectivity. Type I error was set at 0.05.

RESULTS

Baseline study characteristics

Key study characteristics and socio-demographic correlates of dairy consumption and metabolic outcomes are presented in Tables 1. Our final sample of 1,371 urban adults had a mean age of 48.4 with a SE of 0.24, with 40.6% being men and 48.5% being African-American. Only 36.6% had >High School educational attainment and the proportion above poverty was 60.1%. Around 23% reported their health as being fair or poor. Socio-economic, lifestyle and health-related factors differed markedly by sex and by race, reflecting lower SES among African-Americans and women, and higher prevalence of risky healthy behaviors among men as wel as among African-Americans. In terms of dietary intakes, overall, mean baseline dairy servings/day was 1.02 [fluid milk (0.51), cheese (0.48) and yogurt (0.03)]. Moreover, the 3 servings/day goal for dairy was reached by 6.6% of men and 3.8% of women (p<0.05, χ² test), with lower proportions among African-Americans vs. Whites, who consistently consumed less calcium, magnesium, phosphorus and dairy fat, as is the case for women vs. men. With men having higher caloric intake than women, baseline intakes of orange vegetables, whole grains, nuts/seeds, soy, and caffeine were lower among African-Americans with a reverse trend observed for meat/poultry/fish and eggs.

TABLE 1.

Gender and racial differences in intakes of dairy foods and related nutrients, obesity, and metabolic outcomes: HANDLS 2004–2009 and 2009–2013¹

	All (n=1,371)		Men (n=557)		Women (n=814)		Whites (n=568)		African-Americans (n=803)
	Mean, %	SE	Mean, %	SE	Mean, %	SE	Mean, %	SE	Mean, %	SE
Socio-demographic and health characteristics, V1
Age (y)	48.4	0.24	48.6	0.4	48.2	0.3	48.9	0.4	48.0	0.3
Men (%)	40.6		—		—		39.8		41.2
African-American (%)	58.6		59.4		58.0		—		—
Above poverty (%)	60.1		63.9		57.52		70.1		53.12
Education
<High School	6.8		8.4		5.6		9.5		4.92
High School	56.6		55.8		57.0		50.0		61.3
>High School	36.6		35.7		37.1		40.5		33.9
Self-rated health (%)
Poor/fair	22.8		23.2		22.6		24.1		21.92
Good	41.2		39.5		42.4		37.2		44.0
Very good/excellent	36.0		37.3		35.1		38.7		34.1
Current smoker, yes (%)	40.5		43.8		38.2		36.7		43.22
Current smoker, missing (%)	8.1		8.1		8.1		8.4		7.9
Current illicit drug user, yes (%)	15.6		20.5		12.32		11.4		18.62
Current illicit drug user, missing (%)	8.1		8.1		8.1		8.4		7.9
Dairy and related nutrients, V1
Fluid milk (g)
All milk (g)	64.3	2.9	66.1	4.4	63.0	3.9	93.2	5.5	43.82	3.0
Whole milk (g)	32.4	2.2	34.0	3.3	31.4	2.9	41.5	4.1	25.92	2.3
Low fat/fat free milk (g)	31.8	2.2	32.1	3.2	31.6	3.0	51.7	4.4	17.92	2.0
All dairy (servings)	1.02	0.03	1.15	0.04	0.932	0.03	1.33	0.05	0.802	0.03
All dairy, ≥ 3 servings/d (%)	4.96		6.64		3.812		8.98		2.122
Fluid milk (servings)	0.51	0.02	0.56	0.03	0.472	0.02	0.67	0.03	0.392	0.02
Yogurt (servings)	0.027	0.003	0.022	0.005	0.030	0.005	0.042	0.006	0.0162	0.003
Cheese (servings)	0.48	0.02	0.56	0.03	0.422	0.02	0.61	0.03	0.392	0.02
Calcium (mg/d)	725.6	11.7	821.9	20.2	659.72	13.6	829.1	20	652.42	14
Calcium, > recommended mg/d (%)	18.3		26.8		12.52		26.2		12.72
Magnesium(mg/d)	241.1	3.3	271.8	5.7	220.42	3.8	266.5	5.6	223.52	3.9
Phosphorus(mg/d)	1,141	15	1343	27	1,0042	16	1,231	25	1,0792	19
Dairy fatty acids (g/100 g fat)	2.44	0.04	2.34	0.05	2.512	0.05	2.89	0.06	2.122	0.04
Dairy and related nutrients, Δ
Fluid milk
All milk (g)	+2.69	0.87	+2.45	1.26	+2.85	1.18	+4.27	1.67	+1.57	0.89
Whole milk (g)	+0.73	0.67	+0.91	1.03	+0.60	0.88	+1.80	1.28	−0.03	0.70
Low fat/fat free milk (g)	+1.96	0.65	+1.54	0.95	+2.25	0.89	+2.47	1.33	1.60	0.60
All dairy (servings)	+0.05	0.01	+0.07	0.01	+0.04	0.01	+0.06	0.01	+0.05	0.01
Fluid milk (servings)	+0.02	0.01	+0.02	0.01	+0.02	0.01	+0.03	0.01	+0.02	0.01
Yogurt (servings)	+0.005	0.001	+0.004	0.002	+0.006	0.002	+0.008	0.003	+0.003	0.001
Cheese (servings)	+0.025	0.006	+0.034	0.010	+0.019	0.007	+0.022	0.012	+0.028	0.006
Calcium (mg/d)	+34.6	3.6	+41.0	6.2	+30.3	4.3	+30.4	6.4	+37.7	4.1
Magnesium(mg/d)	+3.47	0.81	+3.62	1.44	+3.36	0.94	+1.92	1.43	+4.60	0.93
Phosphorus(mg/d)	+25.7	4.0	+26.5	7.3	+25.2	4.5	+22.8	6.7	+27.8	4.8
Dairy fatty acids (g/100 g fat)	+0.03	0.01	+0.03	0.02	+0.02	0.02	+0.04	0.02	+0.02	0.01
Other dietary factors, V1 and Δ
Energy (kcal/d)
V1	2,003	26	2,382	46	17432	26	2,043	39	1,974	34
Δ	+8.09	6.18	+3.59	11.54	+11.17	6.79	+0.19	10.48	+13.7	7.51
Total fruits (servings)
V1	0.73	0.03	0.81	0.05	0.67	0.03	0.76	0.05	0.70	0.03
Δ	+0.026	0.008	+0.019	0.014	+0.030	0.009	+0.01	0.01	+0.03	0.01
Dark green vegetables (servings)
V1	0.12	0.01	0.11	0.01	0.12	0.01	0.11	0.01	0.12	0.01
Δ	+0.009	0.002	+0.005	0.003	+0.012	0.003	+0.012	0.004	+0.007	0.003
Orange vegetables (servings)
V1	0.068	0.004	0.066	0.006	0.069	0.006	0.087	0.007	0.0542	0.005
Δ	+0.042	0.002	+0.047	0.003	+0.039	0.003	+0.052	0.004	+0.0362	0.002
Whole grains (servings)
V1	0.65	0.03	0.68	0.05	0.63	0.04	0.78	0.05	0.562	0.03
Δ	+0.014	0.008	+0.015	0.014	+0.013	0.010	−0.006	0.014	+0.0282	0.009
Non-whole grains (servings)
V1	5.30	0.09	6.39	0.16	4.552	0.09	5.70	0.14	5.022	0.11
Δ	−0.046	0.024	−0.087	0.042	−0.018	0.028	−0.089	0.040	−0.016	0.029
Legumes (servings)
V1	0.05	0.01	0.06	0.01	0.04	0.01	0.06	0.01	0.05	0.01
Δ	+0.032	0.004	+0.036	0.006	+0.028	0.005	+0.039	0.006	+0.026	0.005
Nuts/seeds (servings)
V1	0.47	0.05	0.52	0.08	0.44	0.07	0.60	0.09	0.382	0.07
Δ	+0.003	0.015	+0.019	0.023	−0.008	0.019	+0.007	0.026	+0.000	0.017
Soy (servings)
V1	0.06	0.01	0.04	0.01	0.07	0.01	0.09	0.02	0.042	0.01
Δ	−0.002	0.002	−0.001	0.001	−0.003	0.003	+0.007	0.026	+0.000	0.017
Meat, poultry, fish (servings)
V1	5.49	0.11	6.82	0.21	4.592	0.11	4.77	0.16	6.002	0.15
Δ	−0.031	0.029	−0.063	0.054	−0.010	0.032	−0.056	0.045	−0.014	0.038
Eggs (servings)
V1	0.58	0.02	0.74	0.04	0.472	0.02	0.47	0.03	0.662	0.03
Δ	+0.009	0.006	+0.006	0.011	+0.011	0.007	+0.016	0.011	+0.004	0.007
Discretionary fat (servings)
V1	43.5	0.77	52.5	1.39	37.32	0.82	45.2	1.3	42.3	1.0
Δ	−0.79	0.20	−1.04	0.38	−0.62	0.21	−1.16	0.36	−0.53	0.23
Discretionary oil (servings)
V1	17.4	0.44	19.6	0.8	15.92	0.5	17.6	0.6	17.3	0.6
Δ	+0.91	0.13	+1.15	0.23	+0.74	0.15	+1.01	0.21	+0.83	0.16
Added sugars (servings)
V1	20.1	0.4	23.5	0.8	17.82	0.5	20.0	0.7	20.2	0.5
Δ	+0.06	0.12	+0.02	0.20	+0.07	0.14	−0.10	0.21	+0.15	0.13
Alcoholic beverages (servings)
V1	0.55	0.04	0.77	0.08	0.392	0.04	0.54	0.07	0.55	0.06
Δ	−0.01	0.01	−0.01	0.02	−0.01	0.01	−0.00	0.02	−0.01	0.01
Caffeine, mg/d
V1	137.9	8.3	159.0	8.3	123.52	5.5	227.5	9.3	74.62	3.2
Δ	+0.60	1.01	−0.78	1.74	+1.55	1.22	−0.36	2.21	+1.28	0.73
Metabolic outcomes, V1, V2, Δ
BMI (kg/m2)
V1	29.8	0.2	28.2	0.3	30.92	0.3	29.8	0.3	29.9	0.3
V2	30.5	0.2	28.6	0.3	31.82	0.3	30.4	0.3	30.5	0.3
Δ	+0.14	0.02	+0.09	0.03	+0.182	0.03	+0.16	0.03	+0.13	0.03
Waist circumference (cm)
V1	100.0	0.8	100.4	1.8	99.7	0.6	102.4	1.7	98.32	0.6
V2	102.9	0.5	102.3	0.7	103.4	0.6	104.1	0.7	102.12	0.6
Δ	+0.62	0.14	+0.41	0.34	+0.76	0.09	+0.40	0.34	+0.77	0.08
SBP (mm Hg)
V1	119.7	0.5	120.5	0.7	119.2	0.6	117.6	0.7	121.22	0.6
V2	122.2	0.5	122.0	0.7	122.3	0.6	119.0	0.7	124.52	0.6
Δ	+0.49	0.12	+0.27	0.17	+0.64	0.17	+0.23	0.21	+0.69	0.15
DBP (mm Hg)
V1	72.6	0.3	74.3	0.4	71.42	0.4	71.8	0.4	73.12	0.4
V2	70.8	0.3	72.3	0.4	69.72	0.3	68.7	0.4	72.22	0.4
Δ	−0.40	0.08	−0.44	0.12	−0.37	0.12	−0.73	0.15	−0.162	0.10
HDL-C (mg/dL)
V1	53.1	0.5	48.8	0.7	56.02	0.6	49.7	0.6	55.52	0.6
V2	56.7	0.5	52.2	0.7	59.72	0.6	52.7	0.7	59.52	0.7
Δ	+0.79	0.08	+0.71	0.13	+0.85	0.10	+0.71	0.11	+0.85	0.11
TA (mg/dL)
V1	124.0	2.6	138.0	5.1	114.42	2.6	146.6	4.7	108.02	2.8
V2	123.6	2.1	130.3	3.7	119.02	2.4	143.5	4.0	109.42	2.1
Δ	−0.06	0.49	−1.60	0.96	+1.002	0.51	−0.66	0.95	+0.37	0.51
Fasting blood glucose (mg/dL)
V1	104.4	1.1	106.9	1.9	102.7	1.4	106.4	1.8	103.0	1.4
V2	104.5	1.1	108.5	2.0	101.72	1.2	106.0	1.8	103.4	1.4
Δ	+0.06	0.24	+0.30	0.41	−0.10	0.29	+0.07	0.42	+0.06	0.27
Obesity (%, BMI≥30)
V1	42.1		33.0		48.32		41.7		42.3
V2	47.3		37.0		54.32		47.5		47.1
Incident	14.0		11.6		16.12		14.3		13.8
Central obesity (%)
V1	59.8		40.9		72.72		63.7		57.02
V2	68.3		48.5		81.92		71.5		66.02
Incident	29.6		20.7		43.02		30.5		29.1
MetS (%)
V1	25.8		25.1		26.3		31.0		22.22
V2	25.8		22.3		28.32		31.5		21.82
Incident	12.6		9.6		14.72		14.5		11.4
Number of metabolic disturbances
V1	1.66	0.03	1.49	0.06	1.782	0.04	1.82	0.05	1.552	0.04
V2	1.70	0.03	1.49	0.06	1.852	0.04	1.81	0.06	1.622	0.04
Δ	+0.006	0.007	−0.007	0.011	+0.014	0.009	−0.01	0.01	+0.01	0.01

¹HANDLS, Healthy Aging in Neighborhoods of Diversity Across the Life Span; SBP, systolic blood pressure; DBP, diastolic blood pressure; TA, triacylglycerols; MetS, metabolic syndrome.

²P < 0.05 for testing the null hypothesis that means or proportions are the same between groups.

³Defined as waist circumference > 102 cm for men and > 88 cm for women.

⁴Defined based on NCEP ATP III described in Methods.

⁵Three or more metabolic disturbances as listed above represent MetS. Metabolic disturbances may range between 0 and 5.

Furthermore, hypertension was more prevalent among African-Americans vs. Whites, while lipid profiles reflected poorer cardio-metabolic health among Whites. Central obesity and MetS were also more prevalent among Whites, though incidence proportions in metabolic outcomes did not differ by race.

Socio-demographic correlates of dairy consumption and metabolic outcomes

Moreover, dairy intake was higher among Whites and those with >High School education, independently of age, sex and poverty status (Table 2). Nevertheless, above poverty status was directly linked to obesity and central obesity incidence, particularly among women. Both central obesity and MetS incidence rates increased with age, consistently among women, who simultaneously had lower incidence rates of both outcomes compared to men. Most notably, MetS incidence was lower among African-Americans vs. Whites.

TABLE 2.

Associations of sociodemographic characteristics with baseline dairy consumption, incident obesity, central obesity, and metabolic syndrome: HANDLS, 2004–2009 and 2009–2013¹

	Dairy consumption (servings) (n=1,371)			Obesity (n=859)			Central obesity (n=588)			Metabolic syndrome (n=1,017)
	β	(SE)	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
All subjects
Men (vs. Women)	+0.28	(0.22)	0.22	0.75	(0.52,1.08)	0.13	0.51	(0.37,0.70)	<0.001	0.61	(0.42,0.89)	0.011
Age(y)	−0.006	(0.003)	0.10	1.00	(0.98,1.02)	0.78	1.02	(1.00,1.04)	0.014	1.03	(1.01,1.05)	0.018
African-American vs. White	−0.54	(0.06)	<0.001	0.97	(0.66,1.45)	0.90	1.10	(0.78,1.55)	0.60	0.67	(0.46,0.98)	0.038
Above vs. below poverty	−0.05	(0.10)	0.61	2.02	(1.34,3.02)	0.001	1.72	(1.23,2.43)	0.002	1.26	(0.85,1.86)	0.26
Education
<High School	—			—			—			—
High School	+0.14	(0.10)	0.22	1.01	(0.44,2.34)	0.98	0.78	(0.40,1.53)	0.47	0.79	(0.40,1.54)	0.48
>High School	+0.35	(0.11)	0.001	0.89	(0.37,2.11)	0.79	0.93	(0.47,1.84)	0.83	0.84	(0.42,1.69)	0.63
Men
Age(y)	−0.008	(0.010)	0.41	1.02	(0.99,1.06)	0.27	1.02	(0.99,1.05)	0.21	1.04	(1.00,1.08)	0.045
African-American vs. White	−0.70	(0.12)	<0.001	0.66	(0.35,1.23)	0.19	0.83	(0.48,1.43)	0.50	0.64	(0.33,1.25)	0.19
Above vs. below poverty	+0.06	(0.33)	0.86	1.17	(0.62,2.20)	0.62	1.56	(0.93,2.63)	0.09	1.18	(0.57,2.46)	0.65
Education
<High School	—			—			–			—
High School	+0.11	(0.16)	0.49	0.75	(0.26,2.20)	0.60	0.60	(0.25,1.44)	0.26	0.67	(0.20,2.27)	0.52
>High School	+0.33	(0.16)	0.050	0.86	(0.28,2.63)	0.78	0.79	(0.32,1.97)	0.61	1.02	(0.29,3.56)	0.98
Women
Age(y)	+0.004	(0.005)	0.38	0.98	(0.96,1.10)	0.22	1.03	(1.00,1.05)	0.018	1.03	(1.0,1.05)	0.021
African-American vs. White	−0.53	(0.07)	<0.001	1.29	(0.77,2.16)	0.34	1.24	(0.77,2.00)	0.39	0.69	(0.41,1.09)	0.11
Above vs. below poverty	+0.22	(0.15)	0.15	2.93	(1.71,5.01)	<0.001	2.03	(1.27,3.24)	0.003	1.33	(0.83,2.12)	0.24
Education
<High School	—			—			—			—
High School	+0.14	(0.14)	0.30	1.52	(0.36,6.39)	0.57	0.99	(0.33,2.99)	0.98	0.86	(0.39,1.92)	0.72
>High School	+0.38	(0.14)	0.007	1.11	(0.26,4.76)	0.89	1.00	(0.36,2.97)	1.00	0.76	(0.33,1.76)	0.52

¹See Table 1 for definitions of obesity, central obesity, and metabolic syndrome. These were based on multivariate regression analyses. Linear regression was conducted for dairy consumption, and Cox PH regression models was conducted for obesity, central obesity, and metabolic syndrome. HANDLS, Healthy Aging in Neighborhoods of Diversity Across the Life Span; HR, hazard ratio; Ref, reference; y, years.

²P< 0.05 for the null hypothesis that the regression coefficient β = 0 or that regression coefficient corresponding to HR is equal to zero.

Dairy consumption and incident metabolic outcomes

Furthermore, in the overall population, cheese and yogurt (both V1 and Δ) were directly related to central obesity incidence (Table 3), while Δdairy fat was positively associated with dyslipidemia disturbances (TA and HDL) and with MetS incidence. Moreover, higher milk consumption (both V1 and Δ) was inversely related to dyslipidemia-TA, whereas only its baseline value (i.e. All milk (V1)) was inversely related to MetS, while being directly related to dyslipidemia-HDL.

TABLE 3.

Five-Year Multivariable-Adjusted HRs (95% CIs) for cases of Incident metabolic disturbances by baseline and annual rates of change in dairy food and dairy-related nutrient Intake among disturbance-free (at baseline) HANDLS Participants: HANDLS 2004–2009 and 2009–2013¹,⁴

	Obesity		Central obesity		Hypertension		Hyperglycemia		Dyslipidemia-TA		Dyslipidemia-HDL		Metabolic syndrome2
	HR	95% CI	HR	95% CI	HR	95% CI	HR	95% CI	HR	95% CI	HR	95% CI	HR	95% CI
All subjects
Model 1
All milk (V1)	0.98	(0.90,1.06)	1.00	(0.75,1.36)	0.94	(0.88,1.00)	0.93	(0.83,1.04)	0.893	(0.81,0.99)	1.103	(1.01,1.21)	0.863	(0.78,0.94)
Cheese (V1)	0.99	(0.89,1.10)	1.133	(1.05,1.23)	1.03	(0.96,1.11)	0.99	(0.88,1.11)	1.02	(0.91,1.14)	1.01	(0.90,1.13)	1.02	(0.92,1.13)
Yogurt (V1)	0.97	(0.87;1.08)	1.213	(1.01,1.44)	1.01	(0.93,1.10)	0.97	(0.85,1.12)	1.05	(0.95,1.15)	1.06	(0.94,1.20)	0.94	(0.85,1.05)
ΔAll milk	0.76	(0.51,1.15)	1.01	(0.75,1.36)	0.95	(0.72,1.27)	0.93	(0.60,1.45)	0.583	(0.36,0.93)	1.32	(0.94,1.87)	0.67	(0.45,1.02)
ΔCheese	1.15	(0.81,1.64)	1.903	(1.47,2.47)	1.08	(0.81,1.45)	0.78	(0.51,1.18)	1.06	(0.73,1.53)	0.95	(0.63,1.43)	0.93	(0.62,1.41)
ΔYogurt	0.87	(0.63,1.20)	1.213	(1.01,1.44)	0.99	(0.79,1.25)	0.88	(0.62,1.25)	1.17	(0.93,1.48)	1.12	(0.83,1.51)	1.04	(0.81,1.33)
Model 2
Calcium (V1)	1.00	(0.90,1.08)	1.02	(0.95,1.11)	1.02	(0.95,1.11)	1.01	(0.91,1.13)	0.92	(0.79,1.07)	0.98	(0.89,1.08)	1.06	(0.95,1.18)
Phosphorus (V1)	0.98	(0.82,1.17)	1.06	(0.93,1.21)	1.04	(0.92,1.17)	0.96	(0.79,1.15)	0.98	(0.78,1.22)	1.00	(0.83,1.21)	0.823	(0.68,0.98)
Magnesium (V1)	1.01	(0.82,1.17)	0.96	(0.91,1.01)	0.97	(0.93,1.02)	0.95	(0.88,1.03)	1.00	(0.92,1.09)	1.00	(0.93,1.08)	0.97	(0.91,1.04)
Dairy fat (V1)	0.98	(0.88,1.08)	1.01	(0.91,1.11)	0.97	(0.89,1.04)	1.01	(0.91,1.13)	1.14	(0.99,1.31)	1.10	(0.98,1.24)	1.05	(0.95,1.05)
ΔCalcium	1.03	(0.67,1.58)	1.14	(0.79,1.63)	1.14	(0.83,1.56)	0.93	(0.62,1.40)	0.71	(0.38,1.32)	0.713	(0.51,1.00)	1.25	(0.77,2.03)
ΔPhosphorus	1.12	(0.55,2.25)	1.14	(0.64,2.04)	1.54	(0.93,2.53)	1.31	(0.65,2.64)	0.83	(0.31,2.18)	1.43	(0.71,2.86)	0.50	(0.23,1.10)
ΔMagnesium	0.88	(0.68,1.13)	0.88	(0.72,1.08)	0.86	(0.73,1.03)	0.77	(0.57,1.04)	0.92	(0.65,1.30)	0.98	(0.75,1.29)	0.86	(0.65,1.14)
ΔDairy fat	0.93	(0.63,1.36)	1.31	(0.93,1.86)	0.84	(0.60,1.17)	1.00	(0.65,1.55)	1.863	(1.12,3.08)	1.873	(1.28,2.73)	1.553	(1.09,2.21)
Men
Model 1
All milk (V1)	0.94	(0.81,1.10)	0.893	(0.79,0.99)	0.92	(0.84,1.01)	0.92	(0.74,1.14)	0.733	(0.58,0.91)	1.11	(0.91,1.35)	0.763	(0.63,0.91)
Cheese (V1)	1.17	(0.68,2.00)	1.05	(0.93,1.18)	1.00	(0.90,1.12)	1.08	(0.90,1.29)	1.14	(0.97,1.34)	0.90	(0.74,1.10)	1.04	(0.86,1.27)
Yogurt (V1)	1.455	(0.97,2.16)	0.95	(0.70,1.29)	0.98	(0.83,1.16)	0.89	(0.66,1.20)	1.413,	(1.11,1.80)	1.16	(0.95,1.41)	0.84	(0.68,1.04)
ΔAll milk	0.61	(0.31,1.21)	0.82	(0.50,1.32)	0.90	(0.61,1.32)	0.72	(0.33,1.56)	0.223	(0.08,0.66)	0.915	(0.44,1.90)	0.383	(0.18,0.77)
ΔCheese	1.17	(0.68,2.00)	1.843	(1.27,2.66)	0.83	(0.52,1.31)	1.16	(0.62,2.14)	1.40	(0.78,2.51)	0.94	(0.46,1.90)	0.98	(0.50,1.91)
ΔYogurt	1.45	(0.97,2.16)	1.27	(0.90,1.79)	0.66	(0.37,1.18)	0.81	(0.34,1.92)	1.51	(0.92,2.46)	1.41	(0.74,2.68)	1.07	(0.55,2.08)
Model 2
Calcium (V1)	1.055	(0.81,1.35)	0.87	(0.73,1.04)	0.95	(0.82,1.11)	0.79	(0.37,1.67)	0.613	(0.44,0.84)	1.11	(0.85,1.44)	0.94	(0.71,1.23)
Phosphorus (V1)	1.145	(0.79,1.65)	1.17	(0.92,1.11)	1.12	(0.90,1.39)	3.02	(0.95,9.60)	1.306	(0.85,1.98)	0.71	(0.46,1.10)	1.08	(0.71,1.64)
Magnesium (V1)	0.875	(0.75,1.02)	1.01	(0.92,1.11)	0.913	(0.83,0.99)	0.68	(0.40,1.16)	1.056	(0.90,1.22)	1.05	(0.89,1.23)	0.925	(0.78,1.09)
Dairy fat (V1)	0.753,5	(0.59,0.96)	0.985	(0.84,1.15)	0.97	(0.84,1.14)	1.16	(0.60,2.26)	1.483	(1.14,1.91)	1.14	(0.91,1.43)	0.94	(0.76,1.18)
ΔCalcium	1.39	(0.69,2.80)	0.94	(0.53,1.66)	0.70	(0.40,1.21)	0.79	(0.37,1.67)	0.45	(0.16,1.24)	1.29	(0.53,3.13)	0.68	(0.28,1.66)
ΔPhosphorus	1.14	(0.40,3.29)	1.48	(0.62,3.53)	1.84	(0.86,3.94)	3.02	(0.95,9.60)	1.03	(0.26,4.00)	0.31	(0.07,1.35)	1.115	(0.24,5.11)
ΔMagnesium	0.583	(0.34,0.98)	1.05	(0.75,1.47)	0.613	(0.46,0.82)	0.68	(0.40,1.16)	0.71	(0.38,1.32)	1.32	(0.78,2.25)	0.865	(0.49,1.53)
ΔDairy fat	0.193,5	(0.08,0.46)	0.95	(0.57,1.59)	0.763	(0.46,0.82)	1.16	(0.60,2.26)	3.663	(1.45,9.22)	1.345	(0.60,2.95)	1.67	(0.89,3.14)
Women
Model 1
All milk (V1)	1.05	(0.91,1.22)	1.07	(0.94,1.21)	0.90	(0.81,1.01)	0.97	(0.81,1.15)	0.87	(0.73,1.05)	1.07	(0.91,1.27)	0.803	(0.71,0.91)
Cheese (V1)	0.97	(0.82,1.15)	1.183	(1.02,1.37)	1.04	(0.92,1.18)	0.94	(0.79,1.13)	0.90	(0.70,1.17)	1.12	(0.93,1.36)	1.00	(0.86,1.15)
Yogurt (V1)	0.833	(0.70,0.99)	0.96	(0.86,1.07)	1.01	(0.91,1.13)	1.07	(0.91,1.27)	0.96	(0.78,1.17)	1.06	(0.85,1.32)	0.99	(0.87,1.12)
ΔAll milk	1.00	(0.53,1.88)	0.97	(0.60,1.57)	0.88	(0.53,1.46)	1.21	(0.62,2.38)	0.50	(0.21,1.19)	1.77	(0.94,3.34)	0.62	(0.34,1.12)
ΔCheese	1.42	(0.82,2.44)	2.393	(1.47,3.99)	1.29	(0.82,2.02)	0.70	(0.35,1.38)	0.47	(0.19,1.18)	1.44	(0.65,3.20)	0.87	(0.49,1.56)
ΔYogurt	0.53	(0.28,1.01)	1.363	(1.03,1.80)	1.07	(0.82,1.39)	1.04	(0.67,1.61)	1.01	(0.64,1.60)	1.22	(0.82,1.81)	0.99	(0.72,1.37)
Model 2
Calcium (V1)	0.99	(0.86,1.14)	1.06	(0.94,1.20)	1.133	(1.01,1.26)	1.03	(0.87,1.22)	1.19	(0.96,1.49)	0.693	(0.55,0.86)	1.193	(1.01,1.39)
Phosphorus (V1)	0.93	(0.70,1.23)	1.00	(0.80,1.26)	0.98	(0.44,2.19)	0.80	(0.59,1.10)	0.633	(0.43,0.93)	1.903	(1.30,2.78)	0.613	(0.47,0.80)
Magnesium (V1)	1.02	(0.93,1.11)	0.92	(0.84,1.00)	1.01	(0.79,1.30)	0.98	(0.87,1.11)	0.99	(0.87,1.13)	0.833	(0.71,0.96)	0.98	(0.91,1.07)
Dairy fat (V1)	1.02	(0.88,1.18)	1.00	(0.86,1.17)	0.76	(0.49,1.18)	1.06	(0.89,1.26)	1.10	(0.88,1.36)	1.08	(0.88,1.33)	1.08	(0.94,1.25)
ΔCalcium	0.86	(0.43,1.64)	1.05	(0.54,2.05)	1.883	(1.15,3.08)	1.11	(0.53,2.32)	1.90	(0.65,5.59)	0.103	(0.04,0.24)	1.82	(0.90,3.66)
ΔPhosphorus	2.09	(0.61,7.17)	1.30	(0.42,3.99)	0.98	(0.44,2.19)	0.80	(0.25,2.57)	0.123	(0.02,0.73)	36.23	(8.6,151.8)	0.243	(0.08,0.76)
ΔMagnesium	0.74	(0.52,1.05)	0.693	(0.49,0.98)	1.01	(0.79,1.30)	0.83	(0.51,1.37)	1.13	(0.69,1.87)	0.453	(0.25,0.81)	0.91	(0.65,1.27)
ΔDairy fat	1.55	(0.94,2.55)	1.46	(0.88,2.43)	0.76	(0.49,1.18)	1.02	(0.49,2.12)	1.51	(0.67,3.41)	3.96	(1.94,8.04)	1.59	(0.94,2.67)
Whites
Model 1
All milk (V1)	0.92	(0.79,1.06)	1.143,6	(1.00,1.29)	0.98	(0.89,1.18)	0.803	(0.65,0.98)	0.90	(0.78,1.03)	1.07	(0.91,1.27)	0.813	(0.70,0.92)
Cheese (V1)	1.06	(0.90,1.24)	1.15	(0.99,1.34)	1.03	(0.90,1.18)	0.793	(0.63,1.00)	0.88	(0.71,1.07)	1.12	(0.93,1.36)	0.91	(0.78,1.06)
Yogurt (V1)	0.90	(0.77,1.06)	0.98	(0.88,1.09)	1.01	(0.89,1.14)	0.826	(0.45,1.47)	1.03	(0.91,1.18)	1.06	(0.85,1.32)	0.94	(0.79,1.12)
ΔAll milk	0.59	(0.32,1.08)	1.10	(0.69,1.76)	0.92	(0.60,1.43)	0.61	(0.29,1.27)	0.62	(0.31,1.22)	1.77	(0.94,3.34)	0.473	(0.26,0.85)
ΔCheese	1.40	(0.87,2.27)	2.113	(1.33,3.34)	1.046	(0.64,1.70)	0.403	(0.17,0.94)	0.86	(0.46,1.64)	1.44	(0.65,3.20)	0.79	(0.46,1.38)
ΔYogurt	0.85	(0.57,1.26)	1.29	(0.98,1.70)	0.97	(0.71,1.33)	0.82	(0.45,1.47)	1.27	(0.90,1.79)	1.22	(0.82,1.81)	0.83	(0.59,1.18)
Model 2
Calcium (V1)	1.00	(0.80,1.26)	1.10	(0.94,1.28)	1.07	(0.92,1.24)	0.88	(0.60,1.29)	0.273	(0.08,0.94)	0.87	(0.57,1.34)	0.97	(0.75,1.27)
Phosphorus (V1)	1.09	(0.77,1.55)	1.18	(0.89,1.56)	1.02	(0.79,1.31)	1.16	(0.68,1.96)	0.78	(0.17,3.59)	0.95	(0.54,1.68)	0.80	(0.55,1.16)
Magnesium (V1)	0.843	(0.73,0.96)	0.873	(0.79,0.97)	0.95	(0.88,1.03)	0.813	(0.69,0.95)	0.89	(0.45,1.75)	0.99	(0.85,1.15)	0.89	(0.78,1.01)
Dairy fat (V1)	0.98	(0.82,1.18)	0.99	(0.82,1.19)	0.91	(0.88,1.03)	1.02	(0.79,1.31)	1.24	(0.58,2.66)	1.25	(0.95,1.64)	1.10	(0.92,1.30)
ΔCalcium	1.11	(0.51,2.41)	1.64	(0.76,3.54)	1.076	(0.61,1.89)	0.56	(0.19,1.65)	0.87	(0.65,1.16)	0.283	(0.09,0.88)	0.52	(0.19,1.46)
ΔPhosphorus	1.34	(0.41,4.36)	0.92	(0.28,3.01)	1.19	(0.50,2.80)	1.49	(0.38,5.79)	1.00	(0.64,1.54)	1.50	(0.29,7.66)	0.656	(0.15,2.82)
ΔMagnesium	0.413	(0.24,0.70)	0.653	(0.43,0.98)	0.96	(0.73,1.27)	0.493	(0.26,0.94)	0.92	(0.78,1.09)	1.12	(0.66,1.90)	0.686	(0.39,1.19)
ΔDairy fat	1.25	(0.66,2.36)	1.64	(0.81,3.31)	0.576	(0.32,1.03)	1.90	(0.74,4.84)	0.98	(0.80,1.20)	4.103,6	(2.07,8.12)	2.533,6	(1.41,4.57)
AA
Model 1
All milk (V1)	1.14	(0.98,1.34)	0.87	(0.75,1.00)	0.873	(0.79,0.97)	0.87	(0.71,1.06)	0.693	(0.51,0.93)	1.18	(0.99,1.40)	0.87	(0.72,1.05)
Cheese (V1)	0.99	(0.54,1.80)	1.09	(0.96,1.25)	0.99	(0.88,1.11)	1.06	(0.89,1.25)	1.15	(0.89,1.48)	1.07	(0.86,1.34)	1.04	(0.86,1.26)
Yogurt (V1)	0.88	(0.43,1.78)	0.99	(0.77,1.28)	0.97	(0.83,1.12)	1.09	(0.93,1.27)	1.15	(0.82,1.60)	0.98	(0.77,1.26)	1.00	(0.85,1.16)
ΔAll milk	1.24	(0.64,2.45)	0.76	(0.44,1.34)	0.78	(0.49,1.24)	0.71	(0.32,1.58)	0.153	(0.04,0.63)	1.21	(0.65,2.26)	0.73	(0.35,1.50)
ΔCheese	0.99	(0.54,1.80)	1.873	(1.22,2.88)	1.24	(0.81,1.90)	1.04	(0.52,2.06)	1.19	(0.53,2.67)	1.05	(0.51,2.19)	0.71	(0.35,1.44)
ΔYogurt	0.88	(0.43,.1.78)	1.28	(0.84,1.97)	0.77	(0.47,1.25)	1.05	(0.61,1.80)	0.94	(0.27,3.28)	1.63	(0.90,2.97)	1.55	(0.96,2.49)
Model 2
Calcium (V1)	0.99	(0.83,1.18)	0.97	(0.82,1.14)	1.04	(0.93,1.16)	0.97	(0.80,1.17)	0.80	(0.57,1.11)	0.98	(0.82,1.18)	1.52	(0.79,2.94)
Phosphorus (V1)	1.01	(0.77,1.32)	1.10	(0.88,1.37)	0.98	(0.83,1.15)	1.05	(0.94,1.17)	1.05	(0.69,1.61)	1.21	(0.90,1.63)	1.05	(0.33,3.35)
Magnesium (V1)	1.133	(1.02,1.24)	1.04	(0.95,1.13)	0.95	(0.89,1.02)	1.02	(0.86,1.21)	1.07	(0.92,1.25)	0.94	(0.81,1.08)	0.85	(0.58,1.24)
Dairy fat (V1)	1.01	(0.85,1.18)	1.02	(0.88,1.18)	1.04	(0.93,1.15)	1.05	(1.01,1.09)	1.373	(1.06,1.77)	1.09	(0.91,1.32)	0.68	(0.33,1.42)
ΔCalcium	0.92	(0.46,1.84)	0.96	(0.56,1.65)	1.31	(0.86,2.00)	1.30	(0.69,2.46)	0.49	(0.16,1.46)	0.96	(0.47,1.95)	1.06	(0.88,1.26)
ΔPhosphorus	1.45	(0.53,4.02)	1.94	(0.83,4.57)	1.38	(0.72,2.63)	1.24	(0.45,3.41)	1.66	(0.31,8.84)	2.07	(0.72,5.91)	0.92	(0.68,1.23)
ΔMagnesium	1.34	(0.95,1.91)	1.02	(0.76,1.39)	0.703	(0.53,0.93)	1.03	(0.69,1.52)	0.99	(0.31,1.66)	0.72	(0.43,1.19)	1.02	(0.92,1.14)
ΔDairy fat	0.94	(0.51,1.75)	1.15	(0.71,1.87)	1.20	(0.79,1.83)	0.72	(0.35,1.48)	2.38	(0.95,5.94)	1.22	(0.66,2.27)	0.95	(0.79,1.14)

¹Δ, annual rate of change; AA=African-Americans; HANDLS, Healthy Aging in Neighborhoods of Diversity Across the Life Span;

²Based on NCEP ATP III criteria described in Methods.

³P<0.05 for null hypothesis that Log_eHR = 0.

⁴See Table 3 for scaling of exposure variables in each model. Each model controls for age, sex, race, socioeconomic status (education and poverty status), energy intake at baseline, current smoking, current drug use, and self-rated health. Additional control was also made on the following major food group servings and nutrients (baseline (V1) and annual rates of change (Δ)): energy intake, total fruit, dark green vegetables, deep yellow vegetables, whole grains, non-whole grains, legumes, nuts/seeds, soy, total meat/poultry/fish, eggs, grams of discretionary solid fat, grams of discretionary oils, added sugars (teaspoons), alcoholic beverages (servings), and mg of caffeine.

⁵P<0.05 for testing effect modification by sex, in separate models using interaction terms between exposure and each of the two effect modifiers.

⁶P<0.05 for testing effect modification by race, in separate models using interaction terms between exposure and each of the two effect modifiers.

Sex-specific findings indicated some significant differentials in the relationship between dairy intakes (including dairy-related nutrients) and metabolic disturbances. Most notably, Δdairy fat was inversely related with obesity among men while being positively related to dyslipemia-HDL among women (P<0.05 for exposure×sex interaction in a separate model with main effects).

Other relationships were race-specific, specifically a 14% increased risk of central obesity with each 0.20 serving increase in baseline milk intake, observed in Whites only. Moreover, the positive association between Δdairy fat and MetS as well as with dyslipidemia-HDL was restricted to Whites.

DISCUSSION

Main findings

Our study uncovered some important findings regarding the relationship between dairy consumption and various metabolic disturbances, including MetS. Specifically, in the overall urban adult population, both cheese and yogurt (V1 and Δ) were associated with an increased risk of central obesity. Baseline fluid milk intake (V1 in cup equivalents) was inversely related to MetS [HR=0.86, 95%CI: 0.78,0.94], specifically to dyslipidemia-triacylglycerol (TA) [HR=0.89, 95% CI:0.81,0.99], though it was directly associated with dyslipidemia-High Density Lipoprotein-Cholesterol (HDL-C), [HR=1.10, 95% CI:1.01,1.21]. Furthermore, Δcalcium and Δphosphorus were inversely related to dyslipidemia-HDL and MetS incidence, respectively, while Δdairy fat was positively associated with incident TA- and HDL-C-dyslipidemias and MetS. A few of those associations were sex- and race-specific.

Previous studies

Among recent cross-sectional studies, sixteen found an inverse relationship between dairy consumption and adverse metabolic outcomes (positive findings),^{(34; 47; 48; 49; 50; 51; 52; 53; 54; 55; 56; 57;} ₅^{8; 59; 60; 61; 62)} while five had mixed findings^{(71; 74; 75; 76; 77; 78)} and the remaining studies failed to detect an association in the expected direction. ^{(79; 80; 81)} Among positive findings, a study of 827 Iranian adults (18–74y) concluded that the uppermost quartile of dairy consumption (vs. lowest) had reduced odds of central obesity, hypertension and MetS, an association primarily mediated by calcium intake,⁽³⁴⁾ as was replicated in a separate study.⁽⁶⁰⁾ In a US study of adult women (n=10,006, ≥45y), both calcium and dairy products’ intakes were inversely related to MetS, in multivariate-adjusted models. ⁽⁴⁷⁾ Similarly, a large Korean study (N=4,862, ≥19y) replicated those findings for milk and yogurt intake.⁽⁵⁴⁾ Moreover, a large study of middle-aged adults (ELSA-Brazil, n=9,835), concluded that a higher intake of total and full-fat dairy was inversely related to MetS, and that dairy saturated fat may be mediating this effect. Specifically, dairy was inversely related to blood pressure, glucose and TA, and total dairy was positively associated with HDL-C among women.⁽⁶²⁾ Mixed findings were echoed in a recent national study of US adults [the National Health and Nutrition Examination Surveys (NHANES 1999–2004)] whereby metabolic disorders were inversely related to whole milk, yogurt, calcium and magnesium but positively associated with low-fat milk, cheese and phosphorus intakes.⁽⁷¹⁾ Using NHANES 2001–2010, another study concluded that women meeting the recommended dietary allowance (RDA) for magnesium and calcium have lower odds of MetS, unlike men who required above-RDA calcium intakes to be protected.⁽⁷⁷⁾ The inverse relationship between full-fat dairy and insulin resistance was also observed in a study of 496 Japanese adults.⁽⁵³⁾ Studies examining whole dietary patterns also suggested an inverse relationship of the dairy-rich pattern with MetS. ^{(51; 52; 55; 58; 59)}

Most selected cohort studies ^{(35; 37; 38; 39; 40; 41; 42; 43; 44; 45; 46; 72; 73)} concluded that dairy consumption, dairy-related nutrients or dietary patterns that include dairy are inversely related to the risk of MetS and various metabolic disturbances. For instance, after an average 3.2y of follow-up, incident MetS among 1,868 older adults was inversely related to low-fat dairy and yogurt but positively related to cheese intake.⁽³⁷⁾ Similarly, using data from the Framingham Offspring study (n=3,440, baseline mean age: 54.5y), Wang and colleagues found that total dairy and yogurt intake were both related to over-time weight loss, as well as reduction in WC.⁽³⁸⁾ In another follow-up study of Korean middle-aged adults, (n=7,240, average follow-up time: 45 months), higher baseline dairy intake was associated with lower MetS risk and an over-time reduction in WC.⁽³⁹⁾ Regular fat dairy was associated with lower MetS incidence, as was found in our secondary analysis, in another recent study of Australian middle-aged adults.⁽⁴⁰⁾ In a large study combining data from ARIC and MESA studies (n=13,444), incident hypertension was inversely related to phosphorus, particularly when derived from dairy products.⁽⁴⁶⁾ Two selected cohort studies found little evidence of an association between dairy consumption and metabolic disturbance.^{(72; 73)}

Most relevant intervention trials,^{(63; 64; 65; 66; 67; 68; 82; 83)} detected a protective effect of dairy consumption on metabolic endpoints. In 6–12 week follow-up study that randomized adults into 3 groups [n=25 (glucose control), n=20 (casein group), n=25 (whey group)], there were consistently faster reductions in TA, insulin, insulin resistance and LDL-C over-time in the whey group compared to controls.⁽⁶⁶⁾ Those results were replicated in another smaller study (n=20 obese/overweight post-menopausal women) comparing whey and caseinate intervention vs. glucose control on a wider array of metabolic outcomes. The protective effect of caseinate was found by an over-time reduction in post-prandial TA.⁽⁶⁷⁾ Nevertheless, in a randomized cross-over study of 35 subjects (mean age=49.5y), the milk/yogurt arm (vs. fruit juice, fruit biscuit control) had limited effect on metabolic risk factors.⁽⁸²⁾ This null finding was also replicated in an Australian randomized cross-over study (n=71, 18–75y) of high dairy vs. low dairy after 12 months follow-up, measuring glucose, TA, and HDL-C among others.⁽⁸³⁾

Biological plausibility

Some of our key findings have plausible biological underlying mechanisms.^{(21; 24; 26; 27; 28)} Firstly, dairy provides half of dietary calcium,⁽²⁴⁾ and 42 IU/100 ml of vitamin D which promotes calcium gut absorption and helps maintain adequate serum calcium and phosphate concentrations, thus enhancing bone mineralization and preventing hypocalcemia and secondary hyperparathyroidism.⁽²⁴⁾ In fact, serum calcium is tightly regulated whereby minor decreases trigger normalization by parathyroid hormone (PTH) which activates kidney 1α hydroxylase thus converting 25-hydroxyvitamin D to its active form 1,25-dihydroxyvitamin D (1,25-OH₂-D).^{(24; 94)} The latter induces rapid calcium ions (Ca²⁺) increase inhibiting peroxisome proliferator-activated receptor gamma expression, CCAAT/enhancer binding protein alpha, and steroid regulatory binding element protein which are strong inhibitory signals for adipogenesis and inflammation.⁽⁹⁵⁾ A similar mechanism may also be responsible for the calciuretic effect of high salt diets which increase 1,25-dihydroxyvitamin D and vascular smooth muscle intracellular calcium, thereby increasing peripheral vascular resistance and blood pressure.⁽⁹⁶⁾ Secondly, dairy is an important source of beneficial microbiota and two proteins, whey and casein, which along with branched chained amino acids (e.g. leucine), improve complex indigestible polysaccharides utilization and enhance the anti-obesity effects of calcium by suppressing plasma lipids, blood pressure, improving glucose homeostasis and ameliorating pro-inflammatory and oxidative stress.^{(37; 97; 98)} Thirdly, calcium from dairy products can also bind intestinal short chain fatty acids and bile acids causing up-regulation of the LDL receptor and thus reducing serum Low-Density Lipoprotein-Cholesterol (LDL-C) concentration.⁽²⁴⁾ The cholesterol-lowering effects of calcium accompanied by the effects of low-fat milk products enriched with plant stanol esters improve both total and LDL-C concentration in subjects with moderate hypercholesterolemia.^{(24; 99; 100)}

Moreover, magnesium can modulate insulin action and secretion by preserving pancreatic β-cell function through their impact on calcium homeostasis and oxidative stress among others.⁽¹⁰¹⁾ Magnesium can also raise serum HDL-C while reducing LDL-C and TA, through increasing lipoprotein lipase activity among others.⁽¹⁰¹⁾ Magnesium’s potential effect on weight maintenance was also reported, forming an un-absorbable soap with fatty acids and cholesterol, decreasing their absorption, and thus reducing energy intake from the diet.⁽¹⁰¹⁾ Similarly, reduced serum phosphate level, partly ascribed to reduced phosphorus intake, is also a hallmark of MetS, mostly the insulin resistance component as suggested elsewhere.⁽¹⁰²⁾

Strengths, limitations of the study and conclusions

Our study has several strengths, including its prospective design with long follow-up and repeated measures on exposures and outcomes. Further, we studied both major dairy foods and dairy-related nutrients, while distinguishing between low-fat and full milk in part of the analysis. Although fat content was available for all dairy products, we only considered varying fat contents of milk being the important contributor to total dairy intake. Though small randomized trials have already been conducted, larger observational cohort studies remain clinically important to examine this research question over longer follow-up periods. Moreover, our study collected two 24 hr. dietary recalls/wave instead of one, reducing measurement error and enhancing the value of dietary variables in reflecting usual intake. Given the overall lower socioeconomic status of our study sample, dairy consumption was expected to be lower than the national average. ⁽⁷¹⁾In fact, in both nationally representative data and this urban sample of US adults, educational attainment was an important factor determining dairy intake, particularly among women. Our sample had almost half a serving lower mean intake of total dairy compared to the national average, with less than 5% reaching the recommended 3 servings/d in total dairy intake.

Despite its strengths, our study findings should be interpreted in light of some limitations. Some findings may be observed due to selection bias, given that less than half of the original HANDLS sample was included in our present study. This was partly adjusted for using the 2-stage Heckman selection model, as described in previous studies. ^{(91; 92; 93)} Moreover, measurement error in dietary exposures can still be sizeable, even though 2 24-hr recalls per wave are an improvement over many large cross-sectional studies. Those errors are probably random across metabolic disturbance groups, leading to attenuation of true associations. Nevertheless, our findings regarding yogurt intake may not be as reliable as other dairy foods, given the low average consumption (<0.1 serving per day). ⁽¹⁰³⁾ Additionally, data on supplemental intakes of calcium, magnesium and phosphorus were not available for baseline data which precluded the assessment to total intake of those nutrients. Our findings may be generalizable to urban adults in Baltimore city and other cities around the US with similar racial composition. Finally, modest associations observed could be the result of residual confounding by unmeasured lifestyle or health-related factors, while other associations may have been left undetected due to inadequate statistical power. In fact, dairy intake may be a reflection of a healthy lifestyle measured by factors that were not accounted for in our analyses. It is worth noting that in addition to the commonly cited limitations of observational studies, many of the randomized trials to date have failed to use an adequate comparison group that would reflect the dairy-related nutrients’ potential effects on MetS or its components, including calcium and magnesium. It is therefore important to compare dairy consumption to non-dairy products (e.g. soy products) and their potential effect on metabolic parameters over time. Instead, most randomized trials to date have compared individual constituents within dairy (e.g. whey vs. casein) or dairy vs. sugar-sweetened beverages. The latter cannot be considered a good comparison, as sugar-sweetened beverages are well-known to increase blood glucose, insulin and TA over time.⁽¹⁰⁴⁾ While differential composition in calcium, magnesium and phosphorus as well as dairy fat may partially explain differences in the association between various dairy products and metabolic disturbance, further studies are needed to uncover the key mediators.

In sum, various dairy exposures had differential associations with metabolic disturbances. Future intervention studies should uncover how over-time changes in dairy components may affect metabolic disorders, accounting for sex and race differences in those putative effects. Specifically, our study found that some dairy foods (yogurt and cheese) were directly associated, while milk was inversely related to MetS and its components. Furthermore, minerals like calcium and phosphorus are abundantly found in yogurt and cheese as well as in milk. They are also found in other foods like vegetables and whole grains. The latter food groups have been associated with lower incidence of major chronic diseases and thus their consumption should be further encouraged. Replication of our findings by randomized controlled trials with similar exposures would strengthen the case for the public health implications of intakes of various dairy foods and related nutrients on populations and their potential impacts on metabolic disorders, including the metabolic syndrome.

ABBREVIATION

BMI

Body Mass Index

CVD

Cardiovascular Disease

DBP

Diastolic Blood Pressure

HANDLS

Healthy Aging In Neighborhoods of Diversity across the Life Span

HDL-C

High Density Lipoprotein-Cholesterol

LDL-C

Low Density Lipoprotein-Cholesterol

NHANES

National Health and Nutrition Examination Surveys

MetD

Metabolic Disturbances

MetS

Metabolic Syndrome

OLS

Ordinary Least Square

PH

Proportional Hazards

SBP

Systolic Blood Pressure

TA

Triacylglycerol

WC

Waist Circumference