Introduction
Prior to an individual developing overt type 2 diabetes, there appears to be a period of subclinical metabolic abnormality, manifesting in the altered circulating levels of many metabolites [
1,
2]. Specifically, several studies have now reported that circulating concentrations of amino acids predict the development of type 2 diabetes. A nested case–control study from the Framingham Offspring study showed that branched-chain amino acids (BCAAs) isoleucine, leucine and valine and aromatic amino acids (AAAs) tyrosine and phenylalanine showed positive associations with insulin resistance and risk of type 2 diabetes [
3]. The European Investigation into Cancer and Nutrition (EPIC) Potsdam study, the Metabolic Syndrome in Men (METSIM) study, the Cardiovascular Risk in Young Finns (CRY) study and the Southall and Brent Revisited (SABRE) study reported similar findings [
4‐
7]. Glycine and glutamine have also been reported to be consistently inversely associated with risk of type 2 diabetes in a meta-analysis [
8].
In general population studies, elevated levels of BCAAs and AAAs also appear to be associated with increased risk of cardiovascular disease [
9‐
12], although these associations have not been entirely consistent [
13]. In the large Estonian Biobank study, inverse associations between the concentration of several amino acids (including BCAAs) and all-cause mortality were observed [
14]. An inverse association between BCAAs and clinical dementia or Alzheimer’s disease has also been observed [
15]. Finally, we have recently reported in a randomised placebo-controlled trial that metformin treatment (for 18 months in men with CHD but without type 2 diabetes) led to improved insulin sensitivity and was associated with increases in alanine and histidine and reductions in phenylalanine and tyrosine concentrations, with no effect on BCAAs [
16].
Therefore, the existing literature highlights inconsistent associations of amino acids with different outcomes in different studies, apparently contrary to the observations made in general population studies wherein elevated levels of BCAAs and AAAs are an adverse signal. This raises the possibility that the mechanisms that influence circulating amino acids might be more subtle than previously thought and as such it is worth investigating and contrasting the associations of measurable circulating amino acids with different adverse outcomes in people with type 2 diabetes.
A very small (
N = 80) nested case–control study did not find that amino acids were associated with diabetic retinopathy [
17]. However, we are aware of no large studies investigating the association of circulating amino acids with outcomes in individuals with type 2 diabetes.
Developing an understanding of any relationship between amino acids and a range of adverse outcomes in diabetes is important from an aetiological perspective, to develop hypotheses for intervention studies and potentially to develop clinical risk scores. We thus aimed to simultaneously investigate the association of circulating amino acids with the following outcomes in people with type 2 diabetes: (1) macrovascular disease; (2) microvascular disease and (3) all-cause mortality.
Methods
Statistical analysis
Continuous data with approximately normal distributions (including all amino acids) are presented as mean ± SD; those with skewed distributions are presented as median (with interquartile range). Categorical data are presented as n (%). Pearson correlations were used to explore associations of the amino acids with each other. Associations of amino acids with classical risk factors were investigated across quarters of the distribution of each amino acid.
Cox regression models were fitted using the STSELPRE procedure for case-cohort analyses (StataCorp, College Station, TX, USA). Models estimated HRs for a 1 SD increase in each amino acid with each of the endpoints. Two models, with different potential confounding variables, were fitted for each amino acid/outcome combination: model 1 with age, sex, region and randomised treatment; model 2 with, additionally, a prior macrovascular complication of diabetes (myocardial infarction, stroke, hospital admission for a transient ischaemic attack or for unstable angina, coronary or peripheral revascularisation, or amputation secondary to peripheral vascular disease), duration of diabetes, current smoking, systolic blood pressure, BMI, urinary ACR, eGFR, HbA1c, plasma glucose, total cholesterol, HDL-cholesterol, triacylglycerols, aspirin or other antiplatelet agent, statin or other lipid-lowering agent, β-blocker, ACE inhibitor or angiotensin receptor blocker, metformin use, history of heart failure, participation in moderate and/or vigorous exercise for >15 min at least once weekly, and high-sensitivity C-reactive protein (CRP). A third adjustment model, attempting to include all amino acids in the same model, resulted in collinearity and estimates were thus not available. Non-linearity was tested by comparing the deviances of linear and categorical models and by the inclusion of polynomial components (quadratic and cubic terms). Other analyses were performed using SAS v9.2 (SAS Institute, Cary, NC, USA). All p values reported are two-sided, with the 5% threshold used to determine significance.
For the random subcohort, the ability of amino acids to discriminate between those who will and those who will not go on to suffer each of the three adverse outcomes were estimated, in the context of model 2, using
c statistics for 5 year risk, accounting for censoring. In addition, the ability of amino acids to reclassify participants according to 5 year risk, using the continuous net reclassification index (NRI), was assessed by methods suitable for survival data, using bootstrapping [
27].
Primary results came from use of all available data; sensitivity analyses using only participants with complete data were also performed.
Discussion
Although previous observational studies have reported associations of circulating BCAAs and AAAs with adverse outcomes in healthy people, the present report contrasts for the first time the associations of multiple circulating amino acids with the major vascular complications of diabetes. Rather than one (or more) amino acids being a consistent signal for adverse outcomes of any kind, we report that their associations with risk of macrovascular events, microvascular events and all-cause mortality are strikingly different from each other. A key finding is the inverse association of tyrosine with risk of microvascular events, independent of eGFR and urinary ACR. Although the evidence from the present study suggests that these might only be very moderately useful biomarkers in incremental prediction of adverse events in individuals with type 2 diabetes, the pathophysiology underlying these associations and the possibility of intervention studies are intriguing and worthy of further investigation.
The association of high circulating concentrations of BCAAs and AAAs with obesity has been known since the 1960s [
29] and has been proposed to be at least partially mediated by insulin resistance. Insulin is thought to be a regulator of branched-chain α-keto acid dehydrogenase complex [
30]. Insulin resistance may hence suppress BCAA catabolism, as suggested by associations noted in observational epidemiology studies [
31‐
33]. The causal pathway may not be unidirectional; a recent Mendelian randomisation study suggests that genetically elevated BCAAs (via impaired catabolism) are associated with increased risk of type 2 diabetes [
34], although a better understanding of the underlying pathway is required to increase confidence in this observation [
35]. There is also the possibility that amino acids themselves (particularly BCAAs) may affect metabolism by suppressing postprandial glucose levels [
36]. Increased protein turnover in people with central obesity may result in higher circulating levels of amino acids [
37] and might therefore cause elevations in amino acids in people who are overweight and have type 2 diabetes. There are hence a variety of potential mechanisms related to type 2 diabetes pathologies that might influence circulating amino acids in individuals with type 2 diabetes.
Given this background, and prior findings in general populations of associations of specific amino acids with CVD [
9‐
12], we wished to examine whether amino acid levels associated with adverse outcomes in individuals with type 2 diabetes. In the ADVANCE study, the BCAAs leucine, valine and isoleucine showed no association with macrovascular events, but low levels of leucine and valine were associated with increased all-cause mortality. However, the positive, albeit not independently predictive, association of phenylalanine with CVD and all-cause mortality we observed is broadly in line with other published data. There are limited intervention studies investigating the effect of amino acid supplements on health outcomes, with most research coming from short-term trials examining surrogate health markers in the sports science area [
38]. Our data strongly support the need for further studies to determine why higher phenylalanine appears to be a consistently adverse signal for CVD outcomes. Our study provides observations that are the basis for testable hypotheses investigating the effect of genetic variants, which are instrumental variables for circulating amino acids, on health outcomes [
39,
40].
The inverse association of tyrosine with risk of microvascular events is perhaps the most intriguing individual finding from this study. Tyrosine itself was positively associated with baseline eGFR and inversely associated with baseline HbA
1c and urinary ACR. Impaired conversion of phenylalanine to tyrosine has been reported in renal disease [
41,
42]. Low tyrosine levels might therefore simply reflect impaired kidney function, which itself predicts future microvascular events. Tyrosine is also linked to catecholamine synthesis, which, also speculatively, might be relevant to our findings [
43]. That noted, counter-intuitively, we have reported that metformin in fact lowers, not raises, tyrosine levels in individuals with CVD and at high risk of diabetes [
16]. The effect of other glucose-lowering drugs on amino acid profiles in individuals with diabetes would now be of interest. Further studies are now needed to validate our novel observations and to examine whether our findings may represent causal pathways.
Strengths of the study include the use of a well-characterised clinical trial cohort, an efficiently designed case-cohort study to yield a powerful study for a range of endpoints, which were independently adjudicated according to pre-defined criteria. Like other RCT populations, ADVANCE study participants represent a selected cohort. For instance, ADVANCE study participants were required to have a history of CVD or CVD risk factors. Therefore, our results may not be generalisable to all individuals with diabetes, although other risk factors we measured are generally associated with risk of major endpoints in the expected directions. Amino acids were measured in pragmatically collected plasma samples in the context of a multinational RCT and we cannot rule out the potential for differential pre-analytical sample handling or sample degradation during storage, which may have biased our results [
44], although these samples were analysed at first thaw. We also present data suggesting broadly comparable concentrations of amino acids relative to other cohorts. Another potential limitation is the analysis of samples from non-fasted participants, although in clinical practice, fasting is rarely required among individuals with type 2 diabetes. NMR spectroscopy has been used to investigate changes in amino acids 30 min after a standardised liquid meal [
45] and effects sizes were generally relatively small, although the immediate postprandial state is likely to give larger effect estimates than are at play in this study.
In conclusion, we report distinct associations of different amino acids with risk of major adverse endpoints in individuals with type 2 diabetes. Most notably, the identification of tyrosine as a potential marker of microvascular risk requires further study.
Acknowledgements
We thank E. Butler, University of Glasgow, UK for technical assistance in conducting the study.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.