Dietary Glutamine and Glutamate in Relation to Cardiovascular Disease Incidence and Mortality in the United States Men and Women with Diabetes Mellitus

Abstract

Background

Evidence regarding the potential health effects of dietary amino acids glutamine and glutamate among individuals with type 2 diabetes (T2D) is limited.

Objectives

The aim was to examine dietary glutamine and glutamate in relation to subsequent risk of cardiovascular disease (CVD) and mortality among individuals with T2D.

Methods

We prospectively followed 15,040 men and women with T2D at baseline or diagnosed during follow-up (Nurses' Health Study: 1980–2014 and Health Professionals Follow-Up Study: 1986–2018). Diet was repeatedly assessed using validated food frequency questionnaires every 2–4 y. Associations of energy-adjusted glutamine and glutamate intake, as well as their ratio, with CVD risk and mortality, were assessed using Cox proportional-hazards models with adjustments for demographics, dietary and lifestyle factors, and medical history.

Results

During 196,955 and 225,371 person-years of follow-up in participants with T2D, there were 2927 incident CVD cases and 4898 deaths, respectively. Higher intake of glutamine was associated with lower risk of CVD incidence, CVD mortality, and total mortality: comparing extreme quintiles, the hazard ratios (HRs) (95% confidence intervals [CIs]) were 0.88 (0.77, 0.99), 0.78 (0.65, 0.92), and 0.84 (0.76, 0.92), respectively (all P-trend < 0.05). In contrast, higher intake of glutamate was associated with a higher risk of CVD incidence, CVD mortality, and total mortality; the HRs were 1.30 (1.15, 1.46), 1.46 (1.24, 1.72), and 1.20 (1.09, 1.32), respectively (all P-trend < 0.05). Furthermore, comparing extreme quintiles, a higher dietary glutamine-to-glutamate ratio was associated with a lower risk of CVD incidence (0.84 [0.75, 0.95]), CVD mortality (0.66 [0.57, 0.77]), and total mortality (0.82 [0.75, 0.90]). In addition, compared with participants with stable or decreased consumption of glutamine-to-glutamate ratio from prediabetes to postdiabetes diagnosis, those who increased the ratio had a 17% (5%, 27%) lower CVD mortality.

Conclusions

In adults with T2D, dietary glutamine was associated with a lower risk of CVD incidence and mortality, whereas the opposite was observed for glutamate intake.

Introduction

Type 2 diabetes (T2D) has become a global public health challenge [1]. Currently, >450 million people live with T2D globally, and the number is projected to reach 700 million by 2045 [1]. Cardiovascular disease (CVD) is the primary complication and the leading cause of death in individuals with diabetes [2]. It is of particular importance to develop effective strategies, such as dietary interventions, to reduce CVD risk among individuals with diabetes.

Glutamine and glutamate are highly abundant amino acids in plant and animal protein sources. They are also the most abundant amino acids in the human body, and their metabolism plays an important role in various physiologic functions, such as synthesis of muscle proteins, activation of immune system, and protection of gastrointestinal mucosa [3]. Furthermore, glutamine is closely related to the Krebs cycle [4], which is involved in Nicotinamide Adenine Dinucleotide Hydrogen (NADH)-dependent synthase mechanisms [5]. As such, glutamine may also influence CVD health through this pathway [6].

Recent evidence indicated that both dietary and circulating levels of glutamine were associated with lower risk of chronic diseases (e.g., T2D, CVD, and cancers) [[7], [8], [9], [10], [11]] and mortality (e.g., CVD mortality) [12], whereas dietary and circulating glutamate was linked to a higher risk of these chronic diseases [7] and mortality [12]. However, these studies were mainly conducted in generally healthy populations. Because of the altered metabolism of nutrients (e.g., dietary protein), dyslipidemia, and the prothrombotic profile in patients with diabetes who have an elevated risk of developing CVD and mortality [13], whether these associations observed in healthy individuals could be extrapolated to individuals with diabetes has yet to be elucidated. Our previous study indicated genetic variation in endogenous glutamine synthase was associated with a higher risk of developing coronary artery disease in individuals with diabetes [14]. Furthermore, one small clinical trial has reported potential cardioprotective effects of short-time (e.g., 6 wk) glutamine supplementation among 66 participants with T2D [15]. However, to date, whether long-term habitual dietary intakes of glutamine and glutamate are linked to subsequent CVD events among individuals with diabetes remains largely unknown.

To fill this knowledge gap, we aimed to prospectively investigate the associations of dietary glutamine and glutamate with CVD incidence and total and cause-specific mortality among adults with T2D in 2 large prospective cohort studies with dietary data repeatedly measured every 2–4 y during >30 y of follow-up. We hypothesized that higher intakes of glutamine and glutamine-to-glutamate ratio are associated with lower risk of CVD and total and cause-specific mortality among individuals with T2D, whereas higher intake of glutamate is associated with higher risk of CVD and mortality among these participants.

Methods

Study population

The Nurses’ Health Study (NHS) was established in 1976 among 121,701 US registered female nurses aged 30–55 y [16]. The Health Professionals Follow-Up Study (HPFS) is a parallel cohort that started in 1986 and enrolled 51,529 US male health professionals aged 40–75 y [17]. Details of the 2 cohorts have been described elsewhere [16,17]. Information on lifestyle factors, medical history, and incident diseases was collected using questionnaires every 2 y. The cumulative response rate during follow-up was >90% in both cohorts.

For the current study, we included participants with prevalent diabetes at baseline (1980 for the NHS and 1986 for the HPFS), as well as incident diabetes cases diagnosed during follow-up through 2018. Participants were excluded if they had CVD or cancer at baseline, reported CVD or cancer before diabetes diagnosis during follow-up, reported implausible daily caloric intake (<500 or >3500 kcal/d for women, and <800 or >4200 kcal/d for men [18,19]), or had missing information on dietary glutamine and glutamate consumption at baseline. After the exclusions, 15,040 participants with T2D (10,679 in the NHS, 4361 in the HPFS) remained in the final analyses, of whom 1298 were prevalent T2D at baseline, and 13,742 were incident T2D developed during follow-up (Supplementary Figure 1).

The study protocol was approved by the institutional review boards of the Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health, and the return of completed questionnaires was considered to imply informed consent.

Assessment of dietary glutamine and glutamate

Diet was assessed using validated semiquantitative food frequency questionnaire (FFQ) with ∼131 food items administered in 1980, 1984, and 1986, and every 4 y thereafter in the NHS and every 4 y since 1986 in the HPFS, as described previously [20,21]. In all FFQs, participants were asked about the mean consumption of foods, beverages, and supplements (with a prespecified portion size) during the previous year based on 9 categories of intake frequency, ranging from less than once per month to 6 or more times per day. The overall validity and reliability of the semiquantitative FFQs have been reported previously [[20], [21], [22]]. Nutrient composition came from the Harvard University Food Composition Database, which is continuously updated to account for changes in nutrient contents and food processing.

We derived dietary glutamine and glutamate values using gene sequencing methods (Swiss Institute of Bioinformatics), which have been validated against the US Department of Agriculture and modified biochemical (Khun) methods [23]. Briefly, the proteinogenic amino acid content of food proteins consumed in the nutrient database was calculated based on [12]: 1) the identification of protein fractions of food proteins reported in the literature; 2) the identification of the amino acid composition of protein fractions using the Expert Protein Analysis System (ExPASy) sequence database server; 3) the weighted sum of amino acids content of protein fractions in food proteins; and 4) the incorporation of amino acid content of food proteins in recipes used to create the nutrient database. The entire sequence of the protein fraction examined was applied to calculate the amino acid content of food protein fractions. The terminal residues were excluded from the amino acid calculations as terminal amino acids have a short lifetime, even if terminal amino acids have little impact on the final composition of proteins, due to their small contribution to the total sequence. The formula to calculate the content of amino acids in food protein derived from sequencing data was described in earlier studies [24]. For example, the glutamine content of β-casein in milk was calculated as [% β-casein × # glutamine in sequence molecular weight of glutamine (g/mole)]/molecular weight of β-casein(g/mole) = (23.4% × 20 × 146.15)/(23583.2) = 2.9%. Data on dietary glutamine and glutamate were further adjusted for total energy intake using the residual method [25].

The age-adjusted Spearman correlations between dietary glutamine and glutamate in the current study are presented in Supplementary Table 1. In our cohorts, the major food contributors to glutamine intake were whole grains, cold cereals, and refined grains, and the major food contributors to glutamate were red and processed meats, poultry, fish, and dairy products [12]. Our primary exposures of interest were dietary glutamine and glutamate, assessed after a diabetes diagnosis. Due to the potential contrasting impact of dietary glutamine and glutamate on health, we also calculated dietary intake of glutamine-to-glutamate ratio [12]. Particularly, our additional exposures of interest were changes in dietary intake of glutamine-to-glutamate ratio before and after diabetes diagnosis. The prediabetes dietary intake of glutamine-to-glutamate ratio was assessed from the most proximal questionnaires before diabetes was ascertained, with a mean of 1.8 y for the time gap between the most proximal questionnaires and the ascertainment of diabetes.

Ascertainment of T2D

T2D cases were first identified by self-reports in the biennial questionnaires and subsequently confirmed by a validated supplementary questionnaire on the symptoms, diagnostic tests, and treatment of diabetes. Cases before 1998 were confirmed in accordance with National Diabetes Data Group criteria [26] by at least one of the following: 1) one or more classic symptoms (excessive thirst, polyuria, weight loss, and hunger) and fasting glucose concentrations ≥7.8 mmol/L or random glucose concentrations ≥11.1 mmol/L, 2) 2 or more elevated glucose concentrations on different occasions (fasting concentrations ≥7.8 mmol/L, random glucose concentrations ≥11.1mmol/L, and/or concentrations of ≥11.1 mmol/L after ≥2 h shown by oral glucose tolerance testing) in the absence of symptoms, or 3) treatment with hypoglycemic medication (insulin or oral hypoglycemic agent). After 1998, the diagnosis criterion of fasting glucose was lowered to 7.0 mmol/L according to the American Diabetes Association diagnostic criteria [27]. The validity of the supplementary questionnaire for diagnosis of T2D was previously demonstrated [28,29]. In the validation studies, 98% (61/62 cases) of diabetes cases confirmed by the supplementary questionnaire were reconfirmed by medical record review in the NHS, and 97% (57/59 cases) were reconfirmed in the HPFS [28,29].

Ascertainment of CVD and mortality

CVD incidence and total and cause-specific mortality were the primary outcomes of the current study. Incident CVD was defined as fatal and nonfatal CAD (including nonfatal myocardial infarction [MI], coronary artery bypass graft surgery, and coronary angioplasty and stent) and fatal and nonfatal stroke. When nonfatal CVD events were reported on any biennial questionnaires, we requested permission to access the medical records. Physicians blinded to the participant questionnaire data reviewed all medical records. Nonfatal MI was ascertained according to the WHO criteria, including typical symptoms, elevated cardiac enzyme levels, and electrocardiographic findings [30]. Nonfatal stroke was defined according to the National Survey of Stroke criteria, requiring evidence of neurologic deficits with sudden or rapid onset, which persisted for ≥24 h or until death [31]. The diagnoses of coronary artery bypass graft, coronary angioplasty, and stent were based on self-reports, for which the validity had been demonstrated [32]. Furthermore, fatal CAD was defined if CAD was listed as the cause of death on the death certificate and a history of CAD was evident from previous reports, medical records, or autopsy reports. Similarly, fatal stroke was identified and confirmed by reviewing death certificates, hospital records, or autopsy records.

Deaths were identified by reports by next of kin, the United States postal authorities, or searching the National Death Index. Using these methods, ≥98% of deaths were ascertained [33]. The underlying cause of death was assigned by study physicians through the review of death certificates, autopsy reports, and hospital records. For the current analysis, the cause of death was classified according to the International Classification of Diseases Eighth Revision (ICD-8) in the NHS and ICD-9 in the HPFS. Specifically, CVD deaths were determined by ICD-8 codes 390–458 in NHS and ICD-9 codes 390–495 in HPFS, and cancer deaths were determined by ICD-8 codes 140–209 in NHS and ICD-9 codes 140–208 in HPFS.

Assessment of covariates

Information on anthropometric and lifestyle factors and medication or supplement use was assessed at baseline and updated during follow-up through biennial questionnaires, including age, ethnicity, body weight, physical activity, cigarette smoking, alcohol consumption, menopausal status, and use of postmenopausal hormones (women only), medical history, family history of MI, diabetes or cancer, history of hypertension, hypercholesterolemia, CVD, cancer, or other diseases. As plant-based foods are among the top contributors to dietary glutamine intake, and dietary glutamate is often from animal products, we used a healthy plant-based diet index (hPDI) to assess overall diet quality for better control for confounding. The hPDI positively rated healthy plant foods and inversely rated less healthy plant-based foods and animal foods [34].

Statistical analysis

Person-time was calculated from the date of a diabetes diagnosis to the occurrence of study outcomes, the last return of a valid follow-up questionnaire, or the end of follow-up (June 30, 2014, for the NHS and January 31, 2018, for the HPFS), whichever came first. We stopped updating dietary variables on a report of CVD or cancer because changes in diet after diagnosis of these diseases may distort the true associations of interest (only for mortality analysis). Cox proportional hazard models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations of dietary glutamine, glutamate, and glutamine-to-glutamate ratio with total CVD, CAD, and stroke incidence, and all-cause and cause-specific mortality. We modeled dietary glutamine, glutamate, and glutamine-to-glutamate ratio as time-varying variables. Time-varying covariates were also considered in the multivariable models, which included age, sex, ethnicity, diabetes duration, BMI at diabetes diagnosis, smoking status, alcohol consumption, physical activity, healthy plant-based diet score, intake of total energy, menopausal status and use of postmenopausal hormones (women only), family history of MI, diabetes, and cancer, current aspirin use, current multivitamin use, history of hypertension, use of lipid-lowering drugs, and diabetes medication use. These covariables were selected because they were known or potential predictors of exposures and/or outcomes. We also mutually adjusted for dietary glutamine and glutamate. In the current study, the proportional-hazards assumption was tested by using a likelihood ratio test comparing models with and without multiplicative interaction terms between exposures and calendar year, and we did not find evidence of violation of the proportional-hazards assumption. We further tested the significance of linear trends by modeling the median value within each category as a continuous variable and then examining the significance of this variable. To explore the dose-response relationship of dietary glutamine-to-glutamate ratio with CVD incidence and mortality, we fitted cubic spline regressions, where the same covariates in the primary analyses were adjusted.

In a secondary analysis, we explored the associations of changes in dietary glutamine-to-glutamate ratio from prediabetes to postdiabetes diagnosis with CVD incidence and mortality. Changes in intakes of glutamine-to-glutamate ratio from prediabetes to postdiabetes diagnosis were defined as the absolute difference in consumption of dietary glutamine-to-glutamate ratio, i.e., time-varying postdiabetes glutamine-to-glutamate intake minus the prediabetes intake. We further adjusted for changes in lifestyle and dietary factors and intake of glutamine-to-glutamate ratio before diabetes diagnosis in the multivariate model.

We stratified analyses by age at diabetes diagnosis (<65 y or ≥65 y), sex (women or men), BMI (<25kg/m² or ≥25kg/m²), diabetes duration (<10 y or ≥10 y), smoking status after diabetes diagnosis (never smoker, past, or current smoker), alcohol consumption (<5g/d or ≥5g/d, approximately the population mean), physical activity (<median or ≥median), hypertension or hypercholesterolemia at diabetes diagnosis (yes or no), glutamine-to-glutamate ratio consumption before diabetes diagnosis (<median or ≥median), and hPDI score (<median or ≥median). We used a Bonferroni-corrected P value threshold (0.05/10 = 0.005) to account for multiple comparisons in the interaction tests.

Sensitivity and exploratory analyses

We conducted several sensitivity analyses to test the robustness of our findings and a set of exploratory analyses to derive additional insights based on fully adjusted models (model 2). First, to minimize within-person variation, the average of the last 2 FFQs was used to estimate the consumption of dietary glutamine, glutamate, and glutamine-to-glutamate ratio. Second, we excluded deaths that occurred within 4 y after diagnosis of diabetes and excluded prevalent T2D to examine whether the results were impacted by reverse causation bias. Third, we adjusted for major food items (e.g., fruits, vegetables, whole grains, refined grains, legumes, and red meat) and alternative healthy eating index instead of hPDI, respectively. Fourth, to examine whether these associations of dietary glutamine and glutamate were independent of other major dietary amino acids, we additionally adjusted for dietary branched-chain amino acids. Last, in an exploratory analysis, we examined the correlations between dietary and plasma glutamine and glutamate in a subgroup (n = 637). Furthermore, as our previous analyses indicated that a variant on chromosome 1q25 (rs10911021_T) in endogenous glutamine metabolism synthase was significantly associated with CAD risk among individuals with T2D in the same cohorts, we explored potential interactions between rs10911021_T and glutamine/glutamate intake on CVD incidence and mortality in a subgroup of participants (n = 2052) for whom existing genetic data were available.

To ensure adequate statistical power for analyses, we pooled participants from the 2 cohorts (no significant heterogeneity was found). All statistical analyses were performed with SAS software, version 9.4 (SAS Institute Inc). A P value of < 0.05 (2-sided) was considered statistically significant.

Results

Population characteristics

Table 1 shows the age-standardized baseline characteristics of the participants by quintiles of dietary glutamine and glutamate, respectively. Participants with a higher intake of glutamine had a higher consumption of whole and refined grains and plant protein and a lower intake of red meat, whereas those with a higher intake of glutamate had a higher consumption of animal protein and red meat and a lower intake of plant protein and refined grains. Supplementary Table 2 presents the characteristics of the participants by quintiles of dietary glutamine-to-glutamate ratio.

TABLE 1.

Baseline characteristics of participants according to dietary glutamine and glutamate in the Nurses’ Health Study and Health Professionals Follow-up Study

	Dietary glutamine intake			Dietary glutamate intake
	Q1 (Low)	Q3	Q5 (High)	Q1 (Low)	Q3	Q5 (High)
Intake of glutamine or glutamate (g/d)	5.6 (0.8)	7.2 (0.6)	8.9 (0.9)	5.5 (1)	7.6 (0.9)	10.3 (1.8)
Person (n)	3600	2985	2350	3633	2979	2262
Age (y)	62.1 (10.3)	61.5 (10)	61.4 (9.8)	62.6 (10.7)	61.4 (9.8)	61 (9.6)
Sex (men, %)	26.5	29.2	31.6	25.8	29.9	31.8
BMI (kg/m2)	29.9 (5.9)	30.2 (5.8)	30.1 (6)	29.8 (5.9)	30.1 (5.9)	30.8 (5.8)
Physical activity (MET-h/wk)	14.4 (21.8)	17.6 (23.3)	18.8 (28.4)	14.8 (21.6)	17.1 (22.3)	19.8 (29.9)
Current smoker (%)	15.9	11.5	10.4	13.9	11.8	11.5
Hypertension (%)	63.8	62.3	63.6	61.9	62.9	65.8
Hypercholesterolemia (%)	28.1	32.4	33.4	29.5	31.1	32.2
Family history of MI (%)	24.0	25.9	26.7	24.0	25.8	26.8
Aspirin use (%)	54.3	56.5	54.8	53.7	55.4	53.7
Use of antihypertensive drugs (%)	38.3	38.5	38.6	37.3	39.8	39.8
Use of lipid-lowering medication (%)	19.1	22.7	22.8	20.5	21.4	21.8
Total energy intake, kcal/d	1796 (629)	1817 (583)	1737 (573)	1792 (613)	1823 (581)	1712 (592)
Animal protein (g/d)	50.0 (20.7)	61.1 (22.4)	67.3 (27.9)	42.4 (16.4)	61.4 (19.6)	79.6 (28.6)
Plant protein (g/d)	20.4 (9.1)	23.4 (9.5)	25.5 (11.4)	23.3 (9.8)	23.4 (10.1)	20.7 (9.7)
Alcohol consumption (g/d)	7.9 (14.6)	4.8 (9.3)	3 (6.9)	6.8 (13.6)	5.4 (10.6)	3.7 (7.3)
Fruits (servings/d)	1.4 (1.4)	1.5 (1.3)	1.5 (1.2)	1.4 (1.3)	1.5 (1.2)	1.5 (1.5)
Vegetables (servings/d)	2.6 (2.0)	3.0 (2.4)	3.1 (2.4)	2.4 (1.8)	3.0 (2.3)	3.3 (2.9)
Nuts (servings/d)	0.4 (0.7)	0.3 (0.5)	0.3 (0.5)	0.3 (0.5)	0.4 (0.6)	0.3 (0.5)
Legumes (servings/d)	0.3 (0.3)	0.3 (0.4)	0.4 (0.5)	0.3 (0.4)	0.3 (0.4)	0.4 (0.5)
Whole grains (servings/d)	0.7 (0.7)	1.2 (1.1)	1.8 (1.5)	1.2 (1.2)	1.2 (1.2)	1.1 (1.1)
Refined grains (servings/d)	1.3 (1.1)	1.7 (1.3)	1.8 (1.6)	1.9 (1.5)	1.6 (1.2)	1.1 (1)
Red meat (Servings/d)	1.1 (0.8)	1.1 (0.8)	0.8 (0.7)	0.9 (0.7)	1.1 (0.8)	1.1 (0.9)
Healthy plant-based diet score	52.5 (7.6)	53.4 (7.2)	56.3 (7.1)	52.7 (7.6)	53.5 (7.3)	55.7 (7.1)

Values are means (SD) or percentages (%).Except for hypertension and antihypertensive drugs, all variables in Table 1 were statistically different across the quintiles of dietary glutamine and glutamate consumption.

Abbreviations: MET, metabolic equivalent. MI, myocardial infarction.

Main results

During 196,955 and 225,371 person-y of follow-up, there were 2927 incident CVD cases (including 2275 CHD cases and 652 stroke cases) and 4898 deaths (including 1632 CVD deaths, 1114 cancer deaths, and 2152 other deaths), respectively.

After multivariable adjustments, including diabetes duration, BMI at diagnosis, other lifestyle and dietary factors, and medication use, higher intake of glutamine was associated with lower risk of CVD and CHD incidence, CVD mortality, total mortality, and other mortality: comparing extreme quintiles, the HRs (95% CIs) were 0.88 (0.77, 0.99, P-trend = 0.04) for CVD incidence, 0.87 (0.75, 1.00, P-trend = 0.04) for CHD incidence, 0.78 (0.65, 0.92, P-trend = 0.003) for CVD mortality, 0.84 (0.76, 0.92, P-trend = 0.0008) for total mortality, and 0.85 (0.73, 0.98, P-trend = 0.07) for other mortality, whereas intake of glutamine was not associated with cancer mortality (TABLE 2, TABLE 3). In contrast, a higher intake of dietary glutamate was associated with a higher risk of CVD and CAD incidence and mortality because of CVD and all causes, although not with cancer and other mortality. The HRs comparing the extreme quintiles were 1.30 (1.15, 1.46) for CVD incidence, 1.32 (1.15, 1.51) for CHD incidence, 1.46 (1.24, 1.72) for CVD mortality, and 1.20 (1.09, 1.32) for total mortality (all P-trend < 0.05). Furthermore, comparing extreme quintiles, higher glutamine-to-glutamate ratio was associated with lower risk of CVD incidence (0.84 [0.75, 0.95]), CHD incidence (0.87 [0.76, 1.00]), stroke incidence (0.75 [0.58, 0.96]); CVD mortality (0.66 [0.57, 0.77]), and total mortality (0.82 [0.75, 0.90]) (all P-trend < 0.0001), but not cancer and other mortality (TABLE 2, TABLE 3).

TABLE 2.

Hazard ratios (95% CIs) of CVD incidence according to dietary glutamine and glutamate consumption after diabetes diagnosis

Dietary glutamine intake
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
CVD
No. of case/person-y	654/39,102	531/39,455	590/39,375	551/39,513	601/39,413
Model 1	1	0.83 (0.74, 0.94)	0.91 (0.81, 1.02)	0.82 (0.73, 0.93)	0.84 (0.75, 0.95)	0.003
Model 2	1	0.87 (0.77, 0.98)	0.94 (0.84, 1.06)	0.87 (0.77, 0.99)	0.88 (0.77, 0.99)	0.04
CHD
No. of case/person-y	511/39,102	418/39,455	449/39,375	423/39,513	474/39,413
Model 1	1	0.83 (0.73, 0.95)	0.87 (0.76, 0.99)	0.79 (0.69, 0.91)	0.83 (0.73, 0.95)	0.003
Model 2	1	0.87 (0.76, 1.00)	0.90 (0.79, 1.03)	0.84 (0.73, 0.97)	0.87 (0.75, 1.00)	0.04
Stroke
No. of case/person-y	143/39,102	113/39,455	141/39,375	128/39,513	127/39,413
Model 1	1	0.84 (0.65, 1.09)	1.07 (0.83, 1.36)	0.95 (0.74, 1.22)	0.89 (0.69, 1.15)	0.46
Model 2	1	0.87 (0.67, 1.13)	1.11 (0.86, 1.42)	0.98 (0.76, 1.27)	0.93 (0.71, 1.22)	0.71
Dietary glutamate intake
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
CVD
No. of case/person-y	557/39,286	550/39,516	535/39,552	562/39,504	723/39,000
Model 1	1	1.06 (0.94, 1.20)	1.07 (0.94, 1.21)	1.15 (1.01, 1.30)	1.44 (1.28, 1.61)	<0.0001
Model 2	1	1.07 (0.95, 1.21)	1.05 (0.92, 1.18)	1.11 (0.98, 1.26)	1.30 (1.15, 1.46)	<0.0001
CHD
No. of case/person-years	424/39,286	421/39,516	407/39,552	447/39,504	576/39,000
Model 1	1	1.07 (0.93, 1.22)	1.06 (0.92, 1.23)	1.21 (1.05, 1.39)	1.50 (1.31, 1.71)	<0.0001
Model 2	1	1.07 (0.93, 1.23)	1.03 (0.90, 1.19)	1.16 (1.00, 1.33)	1.32 (1.15, 1.51)	<0.0001
Stroke
No. of case/person-years	133/39,286	129/39,516	128/39,552	115/39,504	147/39,000
Model 1	1	1.06 (0.82, 1.35)	1.07 (0.83, 1.37)	0.96 (0.74, 1.25)	1.22 (0.96, 1.56)	0.12
Model 2	1	1.07 (0.84, 1.38)	1.07 (0.83, 1.38)	0.96 (0.74, 1.25)	1.20 (0.93, 1.55)	0.22
Dietary glutamine-to-glutamate ratio
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
CVD
No. of case/person-years	659/39,065	579/39,420	528/39,582	604/39,418	557/39,374
Model 1	1	0.87 (0.77, 0.97)	0.78 (0.69, 0.87)	0.87 (0.78, 0.97)	0.77 (0.69, 0.86)	<0.0001
Model 2	1	0.94 (0.84, 1.05)	0.86 (0.77, 0.97)	0.96 (0.86, 1.07)	0.84 (0.75, 0.95)	0.01
CHD
No. of case/person-years	510/39,065	444/39,420	421/39,582	462/39,418	438/39,374
Model 1	1	0.85 (0.75, 0.97)	0.79 (0.70, 0.90)	0.86 (0.75, 0.97)	0.78 (0.69, 0.89)	0.0008
Model 2	1	0.94 (0.83, 1.07)	0.91 (0.79, 1.03)	0.96 (0.84, 1.09)	0.87 (0.76, 1.00)	0.07
Stroke
No. of case/person-years	149/39,065	135/39,420	107/39,582	142/39,418	119/39,374
Model 1	19	0.92 (0.73, 1.16)	0.71 (0.55, 0.92)	0.93 (0.74, 1.18)	0.72 (0.57, 0.92)	0.03
Model 2	1	0.95 (0.75, 1.20)	0.74 (0.58, 0.96)	0.97 (0.76, 1.22)	0.75 (0.58, 0.96)	0.05

Model 1: Age, sex (men/women), and race (Caucasian/others).

Model 2: Model 1+ diabetes duration (years), BMI at diabetes diagnosis (<23.0, 23.0–24.9, 25.0–29.9, 30.0–34.9, or ≥35.0 kg/m²), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, or ≥27.0 MET-h/wk), smoking status (never, past, current 1–14 cigarettes/d, or current ≥15 cigarettes/d), alcohol consumption (0, 0.1–4.9, 5.0–14.9, or ≥15.0 g/d), family history of MI (yes/no), family history of cancer (yes/no), family history of diabetes (yes/no), menopausal status and use of postmenopausal hormones (premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users, only for women), current aspirin use (yes/no), multivitamin use (yes/no), presence of hypertension (yes/no), use of lipid-lowering medication (yes/no), diabetes medication use (insulin, oral medication, or others), intake of total energy (quintiles), and healthy plant-based diet index (quintiles). Mutual adjustment was conducted for dietary glutamine and glutamate. CVD, cardiovascular disease

Abbreviations: MET, metabolic equivalent; MI, myocardial infarction.

TABLE 3.

Hazard ratios (95% CIs) of total and cause-specific mortality according to dietary glutamine and glutamate consumption after diabetes diagnosis

Dietary glutamine intake
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
Total mortality
No. of case/person-years	1090/44,840	927/45,168	982/45,126	926/45,229	973/45,008
Model 1	1	0.80 (0.73, 0.88)	0.84 (0.77, 0.92)	0.78 (0.71, 0.86)	0.76 (0.69, 0.84)	<0.0001
Model 2	1	0.90 (0.82, 0.99)	0.92 (0.84, 1.01)	0.89 (0.81, 0.98)	0.84 (0.76, 0.92)	0.0008
CVD mortality
No. of case/person-years	360/44,840	315/45,168	328/45,126	305/45,229	324/45,008
Model 1	1	0.81 (0.69, 0.95)	0.83 (0.71, 0.98)	0.73 (0.62, 0.86)	0.71 (0.60, 0.84)	<0.0001
Model 2	1	0.91 (0.78, 1.07)	0.91 (0.78, 1.07)	0.85 (0.72, 1.00)	0.78 (0.65, 0.92)	0.003
Cancer mortality
No. of case/person-years	232/44,840	221/45,168	229/45,126	211/45,229	221/45,008
Model 1	1	0.88 (0.73, 1.07)	0.90 (0.74, 1.09)	0.84 (0.69, 1.03)	0.85 (0.69, 1.03)	0.12
Model 2	1	0.97 (0.80, 1.17)	0.98 (0.81, 1.19)	0.93 (0.76, 1.14)	0.93 (0.76, 1.15)	0.48
Other mortality
No. of case/person-years	498/44,840	391/45,168	425/45,126	410/45,229	428/45,008
Model 1	1	0.76 (0.66, 0.87)	0.82 (0.71, 0.94)	0.79 (0.68, 0.91)	0.78 (0.68, 0.90)	0.004
Model 2	1	0.86 (0.75, 0.98)	0.90 (0.78, 1.03)	0.90 (0.78, 1.03)	0.85 (0.73, 0.98)	0.07
Dietary glutamate intake
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
Total mortality
No. of case/person-years	1001/45,009	924/45,291	895/45,216	960/45,071	1,118/44,785
Model 1	1	1.07 (0.98, 1.18)	1.13 (1.02, 1.24)	1.26 (1.15, 1.39)	1.47 (1.34, 1.62)	<0.0001
Model 2	1	1.07 (0.98, 1.18)	1.05 (0.96, 1.16)	1.18 (1.07, 1.30)	1.20 (1.09, 1.32)	<0.0001
CVD mortality
No. of case/person-years	330/45,009	286/45,291	272/45,216	305/45,071	439/44,785
Model 1	1	1.03 (0.88, 1.21)	1.08 (0.91, 1.28)	1.27 (1.07, 1.49)	1.84 (1.56, 2.16)	<0.0001
Model 2	1	1.03 (0.87, 1.21)	1.00 (0.84, 1.18)	1.17 (0.99, 1.38)	1.46 (1.24, 1.72)	<0.0001
Cancer mortality
No. of case/person-years	215/45,009	211/45,291	215/45,216	249/45,071	224/44,785
Model 1	1	1.06 (0.88, 1.29)	1.15 (0.95, 1.40)	1.37 (1.13, 1.67)	1.24 (1.00, 1.52)	0.008
Model 2	1	1.05 (0.86, 1.27)	1.11 (0.91, 1.35)	1.33 (1.09, 1.61)	1.11 (0.90, 1.36)	0.11
Other mortality
No. of case/person-years	456/45,009	427/45,291	408/45,216	406/45,071	455/44,785
Model 1	1	1.12 (0.97, 1.28)	1.15 (1.00, 1.32)	1.18 (1.02, 1.37)	1.31 (1.14, 1.52)	0.0004
Model 2	1	1.12 (0.98, 1.29)	1.07 (0.93, 1.23)	1.10 (0.95, 1.27)	1.04 (0.89, 1.20)	0.81
Dietary glutamine-to-glutamate ratio
	Q1 (Low)	Q2	Q3	Q4	Q5 (High)	P-trend
Total mortality
No. of case/person-years	1193/44,657	995/45,057	846/45,306	909/45,166	955/45,186
Model 1	1	0.83 (0.76, 0.90)	0.68 (0.62, 0.74)	0.70 (0.64, 0.77)	0.67 (0.61, 0.73)	<0.0001
Model 2	1	0.97 (0.89, 1.06)	0.81 (0.74, 0.89)	0.85 (0.78, 0.93)	0.82 (0.75, 0.90)	<0.0001
CVD mortality
No. of case/person-years	447/44,657	330/45,057	270/45,306	302/45,166	283/45,186
Model 1	1	0.75 (0.65, 0.86)	0.60 (0.51, 0.69)	0.63 (0.54, 0.73)	0.53 (0.45, 0.61)	<0.0001
Model 2	1	0.88 (0.76, 1.02)	0.73 (0.62, 0.85)	0.77 (0.66, 0.90)	0.66 (0.57, 0.77)	<0.0001
Cancer mortality
No. of case/person-years	257/44,657	243/45,057	198/45,306	192/45,166	224/45,186
Model 1	1	0.95 (0.79, 1.13)	0.75 (0.62, 0.90)	0.71 (0.59, 0.86)	0.78 (0.65, 0.94)	0.0009
Model 2	1	1.06 (0.88, 1.26)	0.85 (0.71, 1.03)	0.80 (0.67, 0.97)	0.89 (0.74, 1.07)	0.04
Other mortality
No. of case/person-years	489/44,657	422/45,057	378/45,306	415/45,166	448/45,186
Model 1	1	0.85 (0.75, 0.97)	0.74 (0.64, 0.84)	0.78 (0.69, 0.89)	0.75 (0.66, 0.86)	<0.0001
Model 2	1	1.01 (0.88, 1.15)	0.89 (0.78, 1.03)	0.97 (0.85, 1.11)	0.95 (0.83, 1.09)	0.30

Model 1: Age, sex (men/women), and race (Caucasian/others).

Model 2: Model 1+ diabetes duration (years), BMI at diabetes diagnosis (<23.0, 23.0–24.9, 25.0–29.9, 30.0–34.9, or ≥35.0 kg/m²), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, or ≥27.0 MET-h/wk), smoking status (never, past, current 1–14 cigarettes/d, or current≥15 cigarettes/d), alcohol consumption (0, 0.1–4.9, 5.0–14.9, or ≥15.0 g/d), family history of MI (yes/no), family history of cancer (yes/no), family history of diabetes (yes/no), menopausal status and use of postmenopausal hormones (premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users, only for women), current aspirin use (yes/no), multivitamin use (yes/no), presence of hypertension (yes/no), use of lipid-lowering medication (yes/no), diabetes medication use (insulin, oral medication, or others), intake of total energy (quintiles), and healthy plant-based diet index (quintiles). Mutual adjustment was conducted for dietary glutamine and glutamate.

Abbreviations: MET, metabolic equivalent; MI, myocardial infarction.

In the cubic spline regression, we found linear inverse dose-response relationships of dietary glutamine-to-glutamate ratio with CVD incidence, CVD mortality, and total mortality (P-non-linearity > 0.05) (Figure 1). Each 10% increment in the dietary glutamine-to-glutamate ratio was associated with a 3% (1%, 4%) lower risk of CVD incidence, a 6% (4%, 8%) lower CVD mortality, and a 3% (3%, 4%) lower total mortality (Figure 1).

FIGURE 1.

Dose-response analyses between dietary glutamine-to-glutamate ratio and CVD incidence and mortality. Dose-response relationships of dietary glutamine-to-glutamate ratio with CVD, total, and CVD mortality were estimated by restricted cubic spline Cox proportional-hazards model. Data were truncated at 2.5 and 97.5 percentiles of the dietary glutamine-to-glutamate ratio to limit the impact of extreme values. A multivariable model was adjusted for age, sex, diabetes duration (years), BMI at diabetes diagnosis (<23.0, 23.0–24.9, 25.0–29.9, 30.0–34.9, or ≥35.0 kg/m²), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, or ≥27.0 MET-h/wk), smoking status (never, past, current 1–14 cigarettes/d, or current ≥15 cigarettes/d), alcohol consumption (0, 0.1–4.9, 5.0–14.9, or ≥15.0 g/d), family history of MI (yes/no), family history of cancer (yes/no), family history of diabetes (yes/no), menopausal status and use of postmenopausal hormones (premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users, only for women), current aspirin use (yes/no), multivitamin use (yes/no), presence of hypertension (yes/no), use of lipid-lowering medication (yes/no), diabetes medication use (insulin, oral medication, or others), intake of total energy (quintiles), and healthy plant-based diet index (quintiles). The solid line is a point estimate, and the dashed lines are 95% CIs. CVD, cardiovascular disease; MET, metabolic equivalent; MI, myocardial infarction.

Increment in dietary glutamine-to-glutamate ratio from pre- to postdiabetes diagnosis was significantly associated with a lower risk of subsequent CVD mortality (Figure 2). Compared with participants without increment in dietary glutamine-to-glutamate ratio after diabetes diagnosis, those who increased the ratio after diabetes diagnosis had a 17% (5%, 27%) lower CVD mortality.

FIGURE 2.

Hazard ratios (95% CIs) of CVD incidence and mortality according to changes in consumption of glutamine-to-glutamate ratio before and after diabetes diagnosis. Multivariable analyses were adjusted for age (continuous), diabetes duration (years), sex (men or women), Caucasian (yes/no), family history of MI (yes/no), family history of cancer (yes/no), family history of diabetes (yes/no), hypertension status (no hypertension, new hypertension, or always hypertension), lipid-lowering medication use (never user, new user, or always user), menopausal status and use of postmenopausal hormones (premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users, only for women), aspirin use (never user, new user, or always user), multivitamin use (never user, new user, or always user), changes in smoking status (always never smoker, always past smoker, always current smoker, quit smoking after diabetes diagnosis, or others), changes in physical activity (continuous), changes in alcohol consumption (continuous), changes in BMI (continuous), changes in total energy (continuous), changes in healthy plant-based diet index (continuous), and intake of dietary glutamine or glutamate before diabetes diagnosis (continuous). ^∗No increment included no change and decrease in glutamine-to-glutamate ratio before and after diabetes diagnosis. CVD, cardiovascular disease; MET, metabolic equivalent; MI, myocardial infarction.

Consistent results were observed when analyses were stratified by age (<65 y or ≥65 y), sex (women or men), BMI (<25kg/m² or ≥25kg/m²), diabetes duration (<10 y or ≥10 y), smoking status after diabetes diagnosis (never smoker, past, or current smoker), alcohol consumption (<5g/d or ≥5g/d, approximately the population mean), physical activity (<median or ≥median), hypertension or hypercholesterolemia at diabetes diagnosis (yes or no), glutamine-to-glutamate ratio consumption before diabetes diagnosis (<median or ≥median), and hPDI score (<median or ≥median). None of the interaction tests was statistically significant (all P-interaction > 0.005 [adjusted for multiple comparisons]) (Supplementary Table 3).

Sensitivity and exploratory analysis results

In sensitivity analyses, similar results were observed when we used the averages of the last 2 FFQs to estimate the consumption of dietary glutamine, glutamate, and glutamine-to-glutamate ratio after diabetes diagnosis (Supplementary Table 4). For example, compared with participants in the lowest quintile of dietary glutamine-to-glutamate ratio, those in the highest quintile had a 16% (6%, 25%) lower risk for CVD incidence, a 30% (19%, 40%) lower risk for CVD mortality, and a 17% (10%, 24%) lower risk for total mortality. The results did not substantially change when deaths occurred within 4 y after diabetes diagnosis or prevalent T2D were excluded (Supplementary Table 5 and 6). Furthermore, adjustment for major food groups or alternative healthy eating index instead of hPDI did not materially change the results (Supplementary Table 7). Also, additional adjustments for dietary branched-chain amino acids did not substantially alter the associations of dietary glutamine-to-glutamate ratio with the risk of CVD and mortality (Supplementary Table 8). In the exploratory analysis, we observed no significant correlations between dietary and plasma glutamine (r = −0.01; P = 0.72) and glutamate (r = −0.01; P = 0.89). Rs10911021 did not modify associations between the intake of these amino acids and study outcomes (all P for interaction >0.05) in a subgroup analysis of 2052 participants (Supplementary Table 9).

Discussion

In these 2 large prospective cohort studies among US men and women with diabetes, we found that higher intake of glutamine was associated with a lower risk of CVD and CAD incidence, CVD mortality, and total mortality, whereas higher intake of glutamate was associated with higher risk of CVD and CAD incidence and mortality due to CVD and all causes. Furthermore, a higher dietary glutamine-to-glutamate ratio was significantly associated with lower CVD, CAD, and stroke incidence and mortality due to CVD and all causes. The associations were independent of established risk factors, including diabetes duration, BMI, lifestyle and dietary factors, medication use, glutamine and glutamate consumption before diabetes diagnosis, and overall diet quality. In addition, a greater increment in dietary glutamine-to-glutamate ratio from prediabetes to postdiabetes diagnosis was significantly associated with a lower risk of subsequent CVD death. Various sensitivity analyses and stratified analyses demonstrated the robustness of these associations.

Comparison with previous studies

Only a few previous prospective cohort studies have examined associations of dietary glutamine and glutamate with CVD incidence and mortality among general healthy populations, and results were mixed [12,35,36]. For example, in a previous analysis of the NHS and HPFS cohorts, we observed that general, healthy participants in the highest quintile of dietary glutamine had a 13%–15% lower risk for mortality due to CVD and all causes, whereas the participants in the highest quintile of dietary glutamate had a 9% higher risk of CVD mortality, compared with those in the lowest quintiles [12]. In contrast, a Japanese cohort study reported that a higher intake of glutamate was associated with a lower risk of stroke mortality among generally healthy women [35], whereas the Dutch Rotterdam Study observed null association of dietary glutamate with blood pressure [36]. The reason underlying this discrepancy is unknown, although it is noticeable that the food sources of dietary glutamate can be different across study populations. For example, in our cohorts, the main sources of dietary glutamate included red and processed meats, whereas in the Japanese cohort, the main sources of dietary glutamate were cereals and starches [35]. However, to date, data pertaining to the potential health impact of dietary glutamine and glutamate consumption on CVD health among persons with diabetes are sparse. Only one small clinical trial reported that 6-wk supplementation with 30 g/d glutamine improved some CVD risk factors, including blood pressure and fasting glucose, as well as body composition, in 66 patients with T2D (15). Furthermore, to the best of our knowledge, so far, no randomized control studies (RCTs) have investigated the potential effect of dietary glutamine on CVD health among general healthy populations, although a meta-analysis of 12 RCTs among 378 patients (e.g., critically ill patients, burn patients, or multiple-trauma patients) reported beneficial effect of dietary glutamine supplementation on fast plasma glucose [37]. Leveraging the large sample size, long duration of follow-up, and repeated dietary assessments, we made some novel observations. We observed that higher habitual intakes of glutamine and glutamine-to-glutamate ratio were significantly associated with lower risk of CVD incidence and mortality, particularly CVD mortality among individuals with diabetes, whereas higher intake of glutamate was significantly associated with higher risk of CVD incidence and mortality. Moreover, we also found linear inverse dose-response relationships for glutamine-to-glutamate ratio with incident CVD and mortality due to CVD and all causes in our cubic spline regression. To our knowledge, this is the first study that examined the associations of long-term habitual dietary glutamine and glutamate with CVD events and mortality among individuals with diabetes. Of note, compared with our previous analyses in healthy populations, the effect sizes of these associations among participants with diabetes were somewhat stronger than those in general healthy populations. For example, the participants with diabetes in the highest quintile of the glutamine-to-glutamate ratio had an 18%–34% lower risk for total and CVD mortality, whereas the general healthy participants in the highest quintile had a 13%–19% lower risk of total and CVD mortality [12].

Furthermore, in the current analysis, we observed that dietary intake of these 2 amino acids was not correlated with their own circulating levels, which could be explained by the fact that plasma glutamine and glutamate concentrations are mainly dependent on the enzyme glutamine synthetase, phosphate-dependent glutaminase, and health conditions [38]. Interestingly, despite this lack of correlations between diet and circulating levels, the circulating levels of glutamine, glutamate, and glutamine-to-glutamate ratio were also robustly associated with the risk of diabetes and CVD in the same directions in generally healthy populations. A meta-analysis including 8000 individuals reported a 15% decreased risk of T2D per 1 SD increase of plasma glutamine [10]. Zheng et al. [7] also observed that the plasma level of glutamate was associated with an increased risk of CVD, whereas the glutamine-to-glutamate ratio was associated with a 25% lower risk in a Spanish cohort [7]. Nonetheless, the existing evidence suggests that beyond some shared pathways, dietary and plasma glutamine and glutamate may also influence cardiometabolic health through distinct pathways [4,39]. For example, dietary glutamine may favorably influence cardiometabolic health by augmenting the release of glucagon-like peptide-1 (GLP-1) [40]. In contrast, plasma glutamine may influence cardiometabolic health mainly by regulating the expression of many genes related to metabolism, signal transduction, cell defense, and repair and to activate intracellular signaling pathways [4]. Whether circulating levels of these amino acids are also associated with comorbidity and mortality among individuals with diabetes shall be elucidated in future studies.

Mechanisms

The biologic pathways through which dietary glutamine and glutamate intake may affect long-term health are not entirely clear, but accumulating evidence from animal and human trials has suggested that dietary glutamine may improve multiple CVD risk factors among individuals with T2D, including obesity, glycemic control, blood pressure, lipid metabolism, inflammation, and endothelial dysfunction [3,40]. Evidence from human and animal studies has indicated that glutamine supplements may reduce glucose levels [41], adiposity, and blood pressure and improve inflammation biomarkers in patients with T2D by enhancing GLP-1 secretory responses and lowering C-reactive protein levels, respectively [15,[42], [43], [44]]. Moreover, glutamine supplementation may benefit the pathologic process of CVD in diabetes, such as atherosclerosis, by preventing an increase in methylglyoxal, a reactive molecule that is known to induce apoptosis of vascular endothelial cells, as methylglyoxal can be detoxified by NADH-dependent synthase mechanisms (6). Glutamine, as a precursor of NADH/Nicotinamide Adenine Dinucleotide Phosphate, can also improve flow of blood and the ability of hemoglobin in red blood cells to carry oxygen to tissue, which may reduce CVD complications of certain diseases, such as sickle cell disease [45]. NADH pathways are regulated by the Krebs cycle (5), which is closely related to glutamine synthesis in cells through glutamine synthetase [4]. However, it remains unclear how dietary glutamine influences endogenous glutamine synthesis. Another less well-understood pathway could be through the gut microbiome. A pilot study reported that oral glutamine supplementation favorably altered the gut microbiota composition in overweight and obese humans by reducing the Firmicutes to Bacteroidetes ratio [46], which could benefit CVD health [47]. In addition, the beneficial association of dietary glutamine with CVD incidence among individuals with T2D may also involve genetic variation in glutamine metabolism. Our previous analysis indicated that genetic variation (rs10911021_T) in glutamine metabolism was associated with higher CHD risk in individuals with diabetes [14]. However, we did not observe evidence of effect modification by rs10911021_T on these associations of interest in an exploratory analysis. The unfavorable associations of dietary glutamate may be partially due to advanced glycation end products produced from dietary glutamate that are common in red meat browned at high temperatures [40]. Advanced glycation end products may increase the risk of CVD and T2D both directly (vasoconstriction and antinatriuresis) and indirectly (oxidative stress) [40]. Dietary glutamate may also increase appetite by interfering with leptin, which has been linked to obesity [48]. Although these mechanisms may potentially explain the slightly stronger associations observed in individuals with T2D in comparison with healthy populations, more mechanistic studies are needed to further illustrate potential specific mechanisms through which dietary glutamine and glutamate influence CVD complications and mortality among individuals with diabetes.

Strengths and limitations

The strengths of the present study included a prospective design, a relatively large sample size comprised of both men and women, long-term follow-up with a high retention rate, repeated assessments of dietary glutamine and glutamate, and other dietary and lifestyle variables before and after diabetes diagnosis, and analyses of several adjudicated disease outcomes including total CVD, CAD, and stroke incidence, and all-cause and cause-specific mortality. In addition, we conducted a series of sensitivity analyses, which showed the robustness of our findings.

Several limitations should be considered as well. First, measurement errors were inevitable in the estimates of food and nutrient intakes, such as dietary protein and fat intake [49]. However, such measurement errors were likely to be nondifferential in this prospective study and, thus, would be more likely to bias the associations toward the null. Second, our study participants were all health professionals, and most were nonhispanic Whites, which could limit the generalizability of our findings to other ethnic groups or socioeconomic groups. However, the relative homogeneity potentially minimizes confounding by socioeconomic status or cultural practices. Third, our study did not have direct measurements of glycemic control and severity of diabetes. However, the results remained unchanged and persisted when we adjusted for the duration of diabetes or the use of insulin and/or other hypoglycemic drugs and when stratifying the analyses by the duration of diabetes. Fourth, we did not have enough power to test whether these associations could be modified by genetic susceptibility (e.g., rs10911021_T), which warrants further investigation. Also, we did not have sufficient data on blood levels of glutamine and glutamate to explore their associations with these outcomes and compare the effect sizes of the associations of dietary and serum glutamine and glutamate. Last, causality may not be established because of the observational nature of this study. Although we adjusted for multiple important risk factors, including dietary and lifestyle factors and lowering-lipid and lowering-glucose medication use, for CVD incidence and mortality, it was difficult to rule out residual confounding in observational studies because of unmeasured or imprecisely measured confounders. For example, due to a lack of data on specific types of treatments and medications for diabetes and CVD, we could not adjust for these more specific types of treatments and medications and were unable to examine the impact of these factors on associations of interest. More clinical trials are warranted to substantiate the causal effects of glutamine intake on disease outcomes among individuals with diabetes.

In conclusion, among US men and women with T2D, higher intakes of dietary glutamine and glutamine-to-glutamate ratio were significantly associated with lower CVD incidence and mortality, particularly CVD mortality, whereas higher intake of dietary glutamate was significantly associated with higher CVD incidence and mortality. In addition, increased glutamine-to-glutamate ratio consumption before and after diabetes diagnosis was also associated with lower CVD mortality. In light of similar associations observed in healthy individuals in our cohorts, our results suggest that long-term habitual dietary glutamine may facilitate the prevention of CVD complications and premature deaths among adults with or without T2D. Further studies are warranted to elucidate the clinical implications of our findings.

Author contributions

The authors’ responsibilities were as follows – ZC, QS, AD: conceived the study. QS, FBH, JEM, AD: were involved in data collection. ZC: analyzed the data. QS, YH: provided statistical expertise. ZC: wrote the first draft of the paper. All authors: contributed to the interpretation of the results and revision of the manuscript for important intellectual content and approved the final version of the manuscript.

The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The authors assume full responsibility for analyses and interpretation of these data. ZC and QS are the guarantors of this work and, as such, had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis.

Conflicts of interest

The authors report no conflicts of interest.

Funding

The NHS and HPFS studies and the current analysis are supported by National Institutes of Health grants UM1 CA186107, P01 CA87969, R01 CA49449, R01 HL034594, U01 HL145386, R01 HL088521, U01 CA176726, R01 CA67262, U01 CA167552, R01 HL035464, R01 HL060712, R01 HL132254, and R01 DK120870. The funding sources did not play a role in the design or conduct of the study; collection, management, analyses or interpretation of the data; or preparation, review, or approval of the manuscript.

Data availability

Data described in the article, code book, and analytic code will not be made publicly available. Further information, including the procedures to obtain and access data from the Nurses’ Health Study and Health Professionals’ Follow-up Study is described at http://www.nurseshealthstudy.org/researchers (e-mail: nhsaccess@channing.harvard.edu) and http://sites.sph.harvard.edu/hpfs/for-collaborators, respectively.

Appendix A. Supplementary data

The following is the Supplementary data to this article: