Dairy Food Consumption Is Associated with Reduced Risk of Heart Disease Mortality, but Not All-Cause and Cancer Mortality in US Adults

Previous evidence has linked animal protein intake, including dairy foods, with an increased risk in mortality from all-causes and certain chronic diseases, including cancer and heart disease. The objective of the current analysis was to examine associations between total dairy consumption with mortality from all-causes, cancer, and heart disease. Data for adults (≥19 y; n = 54,830) from the Third National Health and Nutrition Examination Survey (NHANES) and NHANES 1999–2014 were linked with mortality data through 2015. Individual usual intake for dairy foods were estimated using the National Cancer Institute method. Hazard ratio (HR) models were fit for mortality types (all cause, cancer, heart disease) and measures of usual intakes of dairy. Multivariable analysis further adjusted for age, gender, ethnicity, waist circumference, smoking status, education level, chronic condition status (i.e., based on cancer, myocardial infarct, and diabetes/diabetes medication reported), weight loss attempts, and % kcal from animal protein. No associations were seen between dairy food intake and mortality risk from all-causes [HR = 0.97; confidence intervals (CI): 0.85–1.11; p = 0.67], and cancer [HR = 0.95; CI: 0.75–1.20; p = 0.65] when comparing the lowest quartile to the highest quartile of consumption. Dairy food consumption was associated with a 26% reduced risk for heart disease mortality when comparing the lowest quartile to the highest quartile [HR = 0.74; CI: 0.54–1.01; p = 0.05]. Further analyses in different age groups showed that dairy food consumption was associated with 39% and 31% reduced risk for heart disease mortality in older adults 51–70 and ≥51 y, respectively [adults 51–70 y: HR = 0.61; CI: 0.41–0.91; p = 0.01; adults ≥51 y: HR = 0.69; CI: 0.54–0.89; p = 0.004]. These results contradict previous findings that have linked dairy foods to increased mortality risk. Further, dairy foods as part of a healthy dietary pattern, may help lower heart disease mortality risk.


Introduction
Higher dietary intakes of animal protein have been associated with an elevated mortality risk from all-cause, cardiovascular disease (CVD), and cancer. Dairy protein remains a predominant animal protein source in current US dietary patterns, with estimates showing all milk types (i.e., whole, non-fat, reduced fat and low-fat milk) combined contribute 28% of total dairy protein intake in a typical US adult dietary pattern [1]. Furthermore, a substantial level of evidence, stemming from both randomized clinical trials and epidemiological studies, supports dairy foods as part of healthy dietary patterns [2]. Indeed, both the previous two iterations of the Dietary Guidelines for Americans have recommended low-fat and fat-free dairy foods as part of healthy dietary patterns [3,4]. The 2020-2025 Dietary Guidelines for Americans (2020-2025 DGA) emphasizes the increased consumption of nutrient-dense foods and beverages, including low-fat and fat-free dairy products, due to their vitamin, mineral and health promoting components. Further, the 2020-2025 DGA policy document is supported by consistent and accumulating evidence showing positive associations between dietary patterns and health outcomes [4]. Specifically, strong or moderate evidence reviewed by the 2020 Dietary Guidelines Advisory Committee (2020 DGAC) concluded healthy dietary patterns, including patterns that include low-fat and fat-free dairy foods, are associated with lower risk of all-cause mortality, lower risk of cardiovascular mortality, reduced risk of type 2 diabetes, and certain cancers [4].
In contrast to the scientific position adopted by the 2020 DGAC, previous research using cross-sectional data reported positive associations between animal protein consumption and health outcomes. Data in adults using NHANES III found that higher protein intake was associated with a 75% increase in overall mortality and a 4-fold increased risk of cancer and diabetes in older adults. These findings became null or attenuated if the protein source consumed was derived from plants [5]. In contrast, higher protein intake in adults over 65 years of age was associated with reduced risk of cancer-related and allcause mortality [5] despite that age is generally considered a primary risk factor for cancer mortality [6]. Nonetheless, the investigators did not consider usual protein intakes (i.e., an established measure of long-term intake), as recommended by the National Cancer Institute methodology when assessing mortality risk [7], suggesting substantial methodological concerns, as usual intake has been established as a validated measure of habitual intake and a critical dietary intake assessment method when examining mortality risk outcomes.
The objective of the current study was to examine associations between total dairy usual intake with mortality risk from all-causes, cancer, and heart disease using data from the Third National Health and Nutrition Examination Survey (NHANES) and NHANES 1999-2014 with links to mortality data through 2015. It was hypothesized that usual protein intakes from dairy foods would be associated with a decreased risk for mortality from all-causes, as well as CVD and cancer, but that this relationship may be dependent on age.

Experimental Section
The US NHANES is a database conducted by the National Center for Health Statistics (NCHS) of the Center for Disease Control and Prevention (CDC), and represents a nationally representative, cross-sectional survey of U.S. free-living, civilian residents. Thorough data collection methods and analytics of NHANES have been previously documented in peer-reviewed publications [1,8,9] and are available online [10]. Comprehensive ethical considerations, including, but not limited to, informed consent and privacy protections for all study subjects have been undertaken, and the NHANES protocol has been previously approved by the Research Ethics Review Board at NCHS. For the current analysis, data were obtained from NHANES III, which comprised data collected from 1988 to 1994 and NHANES 1999-2014, and included energy, protein, and total dairy intake data on all adults ≥19 years old. The analysis contained 15,937 participants following exclusions for unreliable intake data, mortality follow-up ineligible data, and pregnant or lactating females (refer to Figure 1 14]. Animal protein was determined by a scientific approach previously published [15], which defined animal protein foods using MPED/FPED variables for type of protein food.

Methods, Subjects and Statistical Analysis
SAS 9.4 (SAS Institute, Cary, NC, USA) was used for all analyses and NHANES survey parameters including examination weights, strata and primary sampling units were implemented. Data in men (≥19 years old, n = 27,200) and women (n = 27,630) were linked with mortality data through 2015 from the National Death Index-a centralized registry of all US deaths [16]. Individual usual intakes were estimated for energy, animal protein (as % kcal) and dairy intake variables (using the National Cancer Institute method (Version 2.1), which has been previously described [7]. Covariates for usual intake estimation were dietary recall sequence, Dietary Reference Intake age group, weekend (Friday, Saturday or Sunday) intakes, and race/ethnicity. Hazard ratios (HR) models were fit for mortality types (all-cause, cancer, CVD) with usual dairy food intake. Quartiles of total dairy food intake were established.  Methods, Subjects and Statistical Analysis SAS 9.4 (SAS Institute, Cary, NC, USA) was used for all analyses and NHANES survey parameters including examination weights, strata and primary sampling units were implemented. Data in men (≥19 years old, n = 27,200) and women (n = 27,630) were linked with mortality data through 2015 from the National Death Index-a centralized registry of all US deaths [16]. Individual usual intakes were estimated for energy, animal protein (as % kcal) and dairy intake variables (using the National Cancer Institute method (Version 2.1), which has been previously described [7]. Covariates for usual intake estimation were dietary recall sequence, Dietary Reference Intake age group, weekend (Friday, Saturday or Sunday) intakes, and race/ethnicity. Hazard ratios (HR) models were fit for mortality types (all-cause, cancer, CVD) with usual dairy food intake. Quartiles of total dairy food intake were established.
Survival models were created using follow-up months from examination date reported in NHANES. While usual dairy intake was the primary variable of interest (with the lowest quartile of usual dairy intake being the reference group (HR + 1.0), other variables in the models collected at baseline included, age, gender, race/ethnicity, waist circumference, smoking status, education level, chronic condition status (i.e., based on cancer, myocardial infarct, and diabetes/diabetes medication reported), weight loss attempts, energy intake and % kcal from animal protein. The final analytical sample was 46,588 after eliminating subjects with missing elements of the model. For participants with mortality recorded, follow-up time was defined as months until death date, while for other participants, followup was defined as months until 31 December 2015. The p for trend for usual dairy intake was considered significant at p < 0.05. Table 1 provides an overview of subject characteristics at baseline and describes follow-up time before death.  Table 2 provides an overview of gender combined mean usual intakes of dairy foods by age and gender. In general, as expected, males had higher usual intakes of dairy foods compared to females.  Table 3 shows the HR for usual dairy food intakes and all-cause mortality risk. Dairy food was not associated with increased risk; however, age, smoking, having a chronic condition, and male sex were associated with increased all-cause mortality risk, while higher education and trying to lose weight in the previous year were associated with decreased all-cause mortality risk. Further, being Mexican American and non-Hispanic Black (NHB) race/ethnicity was associated with increased all-cause mortality risk.

Usual Dairy Food Intake and Cancer Mortality Risk, ≥19 Years Old
HR for usual dairy food intake and cancer-related mortality risk can be viewed in Table 4. No significant associations were observed for quartiles of usual dairy food intake and cancer-related mortality risk. Increased age, NHB race/ethnicity, smoking, having a chronic condition and male sex showed increased risk for cancer mortality, while trying to lose weight and having an education greater than high school were associated with reduced cancer mortality risk.

Usual Dairy Food Intake and CVD Mortality Risk, ≥19 Years Old
HR for usual dairy food intake and CVD-related mortality risk can be viewed in Table 5. Usual dairy food consumption was associated with a 26% reduced risk for heart disease mortality when comparing the lowest quartile to the highest quartile [HR = 0.74; CI: 0.54-1.01; p = 0.05]. A 28% reduced risk in CVD mortality was seen when comparing the third quartile of usual intake with the lowest quartile of usual intake [HR = 0.72; CI: 0.58-0.90; p = 0.003]. Further, attainment of an education greater than high school was associated with a reduced risk, while increased age, smoking and being a male were associated with increased risk for CVD mortality risk as was waist circumference and presence of chronic conditions.

Usual Dairy Food Intake and CVD Mortality Risk in Older Adults (51-70 Years Old)
HR for usual dairy food intake and CVD-related mortality risk in adults 51-70 years of age can be viewed in Table 6. Usual dairy food consumption was associated with a 31% reduced risk for CVD mortality when comparing the lowest quartile to the highest quartile [HR = 0.61; CI: 0.41-0.91; p = 0.015]. Results also showed a 7% risk reduction associated with 1% higher percentage of calories coming from animal protein (with a 2.5% unit change in % animal protein (% kcal), risk is estimated to be reduced by 19%). Further, attainment of an education greater than high school was associated with a reduced risk, while increased age, smoking and male sex were associated with increased risk for CVD mortality risk.

Usual Dairy Food Intake and CVD Mortality Risk in Older Adults (≥51 Years Old)
HR for usual dairy food intake and CVD-related mortality risk in older adults ≥51 years of age can be viewed in Table 7. Usual dairy food intake was associated with a 31% reduced risk for heart disease mortality when comparing the lowest quartile to the highest quartile [HR = 0.69; CI: 0.54-0.89; p = 0.004]. A 21% reduced risk in CVD mortality was seen when comparing the third quartile of usual intake with the lowest quartile of usual intake [HR = 0.79; CI: 0.65-0.98; p = 0.031]. As seen with the other age groups, attainment of an education greater than high school was associated with a reduced risk, while increased age, smoking, being a male, and having a chronic condition were associated with increased risk for CVD mortality risk in adults ≥51 years old.

Discussion
Dairy foods remain an integral part of recommended US dietary patterns and are robust contributors of nutrient density to several healthy eating patterns, including the Dietary Approaches to Stopping Hypertension (DASH), Mediterranean and vegetarian diets. The present analyses demonstrate no associations between usual dairy food intake and mortality risk from all-causes, and cancer when comparing the lowest quartile to the highest quartile of consumption. Usual dairy food consumption was associated with a 26% reduced risk for heart disease mortality when comparing the lowest quartile to the highest quartile. In older adults, 51-70 and ≥51 y, usual dairy food consumption was associated with 39% and 31% reduced risk for heart disease mortality, respectively. These results contradict previous findings that have linked dairy foods to increased mortality risk. Further, our analyses using the National Cancer Institute's usual intake method show that dairy foods, as part of a healthy dietary pattern, may help lower heart disease mortality risk.
Current and previous US dietary guidelines have recommended limiting consumption of higher fat dairy foods, particularly as such dairy foods are a source of saturated fats which are presumed to be a contributing factor to adverse cardiovascular outcomes and cardiovascular-related mortality risk. A recent review by Mozaffarian [17] established dietary priorities for cardiometabolic health included whole-fat or low-fat dairy foods and reiteration of the notion that "specific foods and overall dietary patterns, rather than single isolated nutrients, are most relevant for cardiometabolic health". The review [17] further summarizes that longitudinal studies do not support harmful health outcomes associated with whole-fat milk consumption and chronic diseases, including cardiovascular disease, diabetes and obesity in adult populations [18][19][20][21][22]. Further, whole-fat dairy foods may offer potential positive health benefits, including reducing risks for cardiovascular and diabetes-related outcomes [23][24][25].
The current data indicate contradictory results for mortality risk when using the usual dairy food intake validated and preferred methodology in regression analyses in comparison to previous work [5] that have not used the method recommended by the National Cancer Institute [7], suggesting substantial methodological concerns. However, the current findings are supportive and aligned with recent data from a large 21-country cohort [26]. Similar to the current methodology, a validated food frequency questionnaire assessed dairy food intake in nearly 137,000 adults and reported that higher intakes of total dairy foods (i.e., >2 servings per day vs. no dairy food consumption) was associated with a 16% reduced risk of the composite outcome (i.e., the primary outcomes assessed the composite of mortality or major cardiovascular failure), a 17% risk reduction in total mortality, a 14% lowered risk of non-cardiovascular mortality, a 23% lowered risk of cardiovascular mortality and a 22% decreased risk of cardiovascular diseases [26]. Previous research has also examined the clinical outcomes when replacing low-fat dairy with higherfat dairy foods. Indeed, while the original DASH diet trials [27,28] included mostly low-fat dairy foods, a subsequent clinical trial showed that the DASH diet is equally effective in contributing to cardiovascular benefits when higher-fat dairy foods are substituted for lowfat dairy foods [29]. Researchers further reduced sugars from carbohydrates in the DASH diet to compensate for additional energy provided by the higher-fat dairy foods. Results showed that the higher-fat dairy DASH diet with lowered carbohydrates had the same beneficial effect of lowering blood pressure, reducing triglycerides and very-low-density lipoprotein cholesterol levels, without elevating low-density lipoprotein cholesterol levels, which was similar to the original DASH diet with low-fat dairy foods [27,28].
Our study has limitations which are inherent in epidemiological research using NHANES data which have been previously reported [1,8,9]. The results reported here are observational in nature, and thus, only indicate an association between usual dairy food intakes and mortality outcomes. We acknowledge that there is controversy over the accuracy of the methods used to assess food intake using the methods employed in NHANES [30]. While an in-depth discussion is not warranted here as these arguments have previously been documented, we acknowledge that it has long been recognized that self-reported dietary intakes show consistent underreporting of energy intake and, to some extent, macronutrients [30]. The utilization of biomarkers may improve estimates of intake and urinary urea nitrogen could be used to correct for the underreporting of protein intake [31]. We are unaware, however, of data to suggest that reporting of usual dairy food intake would be systematically biased in those consuming higher or lower dairy food intakes, and thus would propose that our risk estimates are robust. Furthermore, mortality dates were only publicly available through 2015 [16] at the completion of the current analyses, and as such, additional and further research examining usual dairy food consumption and risk of mortality is warranted and recommended.
In conclusion, the present analyses did not demonstrate a relationship between usual dairy food intake and mortality risk from all-causes and cancer, when comparing the lowest quartile to the highest quartile of dairy food consumption. In contrast, usual dairy food intake was significantly associated with cardiovascular disease-related mortality risk. Specifically, a 26% reduced risk for heart disease mortality was observed when comparing the lowest quartile to the highest quartile of usual dairy food intake. In older adults, 51-70 and ≥51 y, usual dairy food intake was associated with 39% and 31% reduced risk for heart disease mortality, respectively. These results contradict previous findings that have linked dairy foods to increased mortality risk. The present analyses, which incorporated the National Cancer Institute's usual intake method, provides evidence that dairy foods within a dietary pattern, may help lower heart disease mortality risk in adults.
Author Contributions: Conceptualization, Y.P. and V.L.F.III; methodology, Y.P. and V.L.F.III; software, V.L.F.III; validation, Y.P. and V.L.F.III; formal analysis, V.L.F.III; investigation, Y.P. and V.L.F.III; resources, Y.P. and V.L.F.III; data curation, Y.P. and V.L.F.III; writing-original draft preparation, Y.P.; writing-review and editing, Y.P. and V.L.F.III; visualization, Y.P. and V.L.F.III; supervision, Y.P. and V.L.F.III; project administration, Y.P.; funding acquisition, Y.P. All authors have read and agreed to the published version of the manuscript.

Funding:
The present research was funded and supported by the National Dairy Council.

Institutional Review Board Statement:
The studies involving human participants were reviewed and approved by the Research Ethics Review Board at the National Center for Health Statistics.

Informed Consent Statement:
Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin as per guidelines established by the Research Ethics Review Board at the National Center for Health Statistics.

Data Availability Statement:
Publicly available datasets were analyzed in this study. This data can be found here: https://wwwn.cdc.gov/nchs/nhanes (accessed on 29 December 2022).

Conflicts of Interest:
The sponsor (National Dairy Council) had no role in the design of the study, in the collection, analyses, or interpretation of the data. Y.P., as President of Nutritional Strategies provides food, nutrition and regulatory affairs consulting services for food and beverage companies and food-related associations. V.L.F.III, as Senior Vice-President of Nutrition Impact, provides food and nutrition consulting services for food and beverage companies. V.L.F.III also conducts analyses of NHANES data for members of the food industry.