High Mannose Correlates With Surrogate Indexes of Insulin Resistance and Is Associated With an Increased Risk of Cardiovascular Events Independently of Glycemic Status and Traditional Risk Factors

Among people with coronary artery disease (CAD), those with dysglycemia have a worse prognosis (1–3). Insulin resistance (IR) represents a central shared factor to the pathophysiology of CAD and dysglycemia, promoting endothelial damage and proatherosclerotic processes (4,5). However, evaluation of IR in the clinical setting is challenging because of the impracticalities of performing dynamic assessments (6). High circulating mannose is emerging as a suitable candidate marker of IR and CAD, even in individuals without overt dysglycemia, but needs further validation in longitudinal settings (7–10). The objective of our study is to test the hypothesis that mannose selectively tracks IR and is associated with higher cardiovascular (CV) risk independently of the glycemic state and traditional CV risk factors.

The Swedish multicenter, case-control study Periodontitis and Its Relation to Coronary Artery Disease (PAROKRANK) included 805 patients <75 years old hospitalized for a first myocardial infarction (index MI) between May 2010 and February 2014 and 805 sex-, age-, and postal code area-matched control patients chosen from the Swedish National Population Registry (11). Individuals without known diabetes underwent a 2-h standard oral glucose tolerance test (OGTT) to be further characterized as having normal glucose tolerance (NGT), impaired glucose tolerance, or type 2 diabetes (12); the latter two categories are jointly referred to as dysglycemia.

Study visits and laboratory investigations were performed 6–10 weeks after the hospitalization for MI case participants, and within 10 days thereafter for control participants. Fasting plasma mannose concentrations were measured by means of high-performance liquid chromatography coupled to tandem mass spectrometry at the Pisa Metabolism Laboratory (Italy) (13). Indexes of IR and insulin secretion were calculated according to the equations shown in Supplementary Table 1.

Univariate and multivariate associations between mannose and each of the calculated indexes were explored using linear regression models. The primary outcome was time to the first of a composite of major adverse cardiac events (MACE), including CV death, nonfatal MI, and nonfatal stroke. Event curves for MACE-free survival across mannose quartiles were estimated using Kaplan-Meier functions and compared using log-rank test. The independent associations between both mannose (categorized as high [top quartile] vs. low [lowest three quartiles]) and/or Cederholm index and outcomes were investigated by fitting Cox proportional hazards models. All analyses were performed using Stata/IC 16.1 statistical software.

PAROKRANK was approved by the Stockholm Regional Ethics Committee and was conducted according to the recommendations outlined in the Declaration of Helsinki, with all enrolled participants providing written informed consent. More detailed information on laboratory measurements and statistical analyses are reported in the Supplementary Methods.

The data underlying this study are available from authors G.F. and L.R. upon reasonable request.

This study comprised 1,403 participants (696 case and 707 control) whose glycemic status was categorized by an OGTT as outlined in Supplementary Fig. 1. Baseline clinical characteristics of the study population across mannose quartiles are shown in Table 1. The prevalence of traditional CV risk factors progressively increased from the lowest to the highest quartile of mannose concentrations. Baseline characteristics of case versus control individuals in the complete study population and in glycemic subgroups are presented in Supplementary Table 2. As shown in Supplementary Fig. 2 and Supplementary Table 3, in participants with NGT, but not those with dysglycemia, glucose, insulin, and C-peptide levels were consistently higher in case individuals than in control individuals, both fasting and after glucose load.

Indexes of IR and insulin secretion were higher in case individuals than in control individuals (Table 2 and Supplementary Fig. 3). In NGT, case individuals had higher HOMA-IR, but lower Cederholm index, Matsuda index, and Stumvoll metabolic clearance rate at 120 min (MCR₁₂₀) values and a higher basal insulin secretion rate than control individuals.

In the total study population, mannose levels were statistically significantly correlated with indexes of IR and some indexes of insulin secretion (Spearman ρ 0.25–0.34) (Supplementary Table 4), and the association with IR indexes remained statistically significant after adjustments for CV risk factors (regression model 2, all P ≤ 0.001) (Supplementary Table 5). Most associations with indexes of insulin secretion were no longer statistically significant after controlling for HOMA-IR (model 3). Similar results were obtained in participants with NGT and dysglycemia, separately.

There was no statistically significant difference in the 10-year crude MACE-free survival across mannose quartiles (Fig. 1). On univariate analysis, there were no statistically significant associations between MACE and mannose or IR indexes or several CV risk factors (Supplementary Table 6). On multivariate analysis, high versus low mannose emerged as independently associated with an increased risk of MACE in the overall population (hazard ratio 1.54; 95% CI 1.07–2.20; P = 0.019) (Fig. 2 and Supplementary Table 7) and remained statistically significant even after adjusting for Cederholm index and Stumvoll MCR₁₂₀ index (Supplementary Fig. 4).

The main findings of this study are that 1) mannose was a consistent marker of IR in participants with and without dysglycemia and 2) baseline mannose levels were positively associated with an increased 10-year risk of MACE, independently of established CV risk factors and more strongly than other indexes of IR. We put all effort into appropriately assessing the associations between mannose and dynamic insulin metabolism. First, to overcome the difficulties of estimating the β-cell secretory function using OGTT-derived indexes, all models including insulin secretion indexes were further adjusted for HOMA-IR. Moreover, we used insulinogenic indexes based on C-peptide, as C-peptide is not extracted by the liver and reflects prehepatic insulin secretion more accurately (14). Despite correlation coefficients between mannose and IR indexes being quite low, the pattern was consistent across indexes based on both fasting and postload values, with the latter providing information on insulin action in a more dynamic state of stimulation. Besides, correlations between mannose and typical CV risk factors (e.g., BMI, hs-CRP, and HDL cholesterol) reported in our previous work (9) were very close in magnitude to those between IR indexes and such factors. Altogether, these findings support the notion that mannose selectively tracks IR, even when expressed by surrogate indexes in individuals with supposedly normal glucose metabolism, corroborating results of previous clamp-based studies (10).

We chose to assess several OGTT-based surrogate indexes because they have been suggested to be more strongly associated with future MACE than HOMA-IR, probably because the latter as a fasting index mainly represents hepatic IR (15,16). Among such indexes, the Cederholm index has been reported as having the strongest association with both future CV disease and type 2 diabetes (15). In the current study, the Cederholm index was not statistically significantly associated with future MACE and only marginally weakened the independent association between mannose and MACE when added to the multivariate model, further opening the way for mannose as a potential, accessible marker of IR without needing to perform an OGTT.

Few studies have explored the potential role of mannose as a marker of CV disease (7–9). A recent validation study reported a statistically significant association between mannose and CAD as well as a worse prognosis in patients with mannose levels ≥84.6 μmol/L compared with those with lower concentrations (8). Despite the independent association between mannose as a continuous variable and MACE falling short of statistical significance in our study, an increased risk of MACE was found for the higher quartile versus the lower three quartiles of mannose concentrations, corresponding to a value of 82.2 μmol/L, which is very close to the one in the aforementioned study. This association could be even stronger considering that we excluded patients with events in the first 6–10 weeks after the index MI, i.e., when the risk was highest.

To our knowledge, this investigation is the first to explore the association between plasma mannose and different OGTT-derived indexes of IR and insulin secretion with a CV perspective in a large and well-characterized population with a long follow-up (11). All surrogate indexes used in this analysis have been validated against clamp techniques in patients with varying degrees of glucose tolerance (6,17–19). Some limitations have to be acknowledged. The present cohort included Caucasian patients living in Sweden who had a relatively benign risk factor profile, possibly as a result of selecting patients with a first MI only and high-quality medical care. This, together with the fact that people willing to participate in clinical studies are usually healthier and have a higher compliance with medical treatment, may limit the generalizability of our results. However, it can be hypothesized that in a sicker and/or more unselected patient population, mannose might be more strongly associated with poor prognosis. It is also important to note that even though adjustments for several covariates were performed, residual unknown confounding cannot be ruled out. Finally, and related to the observational design of the current study, the present findings cannot establish whether the nature of the described associations is causal, making it important to confirm these results prospectively.

In conclusion, this study reinforces the notion that mannose is an informative correlate of IR and high concentrations are associated with a higher risk of long-term MACE. Mannose could represent a new biomarker able to track early, potentially detrimental glucometabolic alterations, even in patients who are classified as having NGT. Further studies are needed to investigate mannose’s predictive utility and possible targeted interventions for CV prevention.

Funding. The present study was supported by grants from the Erling Persson Foundation. The PAROKRANK study was academy driven and supported by grants from the AFA Insurance Foundation, Swedish Heart-Lung Foundation, Swedish Research Council, Eklund Foundation, Swedish Society of Medicine, and Stockholm County Council (ALF project and Steering Committee KI/SLL for Odontological Research).

Duality of Interest. L.G.M. reports personal fees from Novo Nordisk, Sanofi Aventis, AstraZeneca, MSD, Boehringer Ingelheim, Novartis, and Amgen outside the submitted work. A.N. has received research grants from the Swedish Heart and Lung Foundation and Stockholm County Council and honoraria for advisory board meetings from AstraZeneca, Novo Nordisk, MSD Sweden, and Boehringer Ingelheim outside the submitted work. L.R. has received research grants from the Swedish Heart and Lung Foundation, Stockholm County, Erling-Persson Family Foundation, and private foundations, outside the submitted work. G.F. has received grant support from the Erling-Persson Family Foundation and reports personal fees from the Swedish Heart and Lung Foundation, European Society of Cardiology, Boehringer Ingelheim, and AstraZeneca outside the submitted work. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. E.Fo. contributed to the study concept, literature search, statistical analysis, data interpretation, manuscript drafting, critical revision of the manuscript for important intellectual content, and administrative handling. B.C. contributed to the methodology, quality assessment, and critical revision of the manuscript for important intellectual content. E.Fe. contributed to the study concept, data interpretation, and critical revision of the manuscript for important intellectual content. A.M. contributed to thee data handling and data interpretation and critical revision of the manuscript for important intellectual content. L.G.M. contributed to the data interpretation, supervision, and critical revision of the manuscript for important intellectual content. A.N. contributed to the original design of the PAROKRANK study and to the data interpretation, supervision, and critical revision of the manuscript for important intellectual content. P.N. contributed to the data handling, statistical analysis and data interpretation, and critical revision of the manuscript for important intellectual content. L.R. contributed to the original design of the PAROKRANK study and to study concept, supervision, data interpretation, and critical revision of the manuscript for important intellectual content. A.S. contributed to the methodology, data interpretation, and critical revision of the manuscript for important intellectual content. G.F. was responsible for the study concept, statistical analysis, data interpretation, manuscript drafting, and critical revision of the manuscript for important intellectual content. All authors approved the final version of the manuscript. L.R. and G.F. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.