Buenos Aires 01 de Febrero del 2021

Distribution and Prognostic Value of High-Sensitive Troponin T Measurement in Patients With Type 2 Diabetes Without Cardiovascular History



Distribution and Prognostic Value of High-Sensitive Troponin T Measurement in Patients With Type 2 Diabetes Without Cardiovascular History


                  Ignacio Bluro*, María Nardi**, Raúl de Miguel**, Martín Fernandez*,
                  Jesica Rolando**,Lisandro Denaday**, Rodolfo Pizarro*, Aída Furci**. 

*   Servicio de Cardiología Médica, Dpto Clínica Medica, Hospital Italiano de Buenos Aires, Argentina
** Servicio de Laboratorio Central, Dpto Diagnóstico y Tratamiento, Hospital Italiano de Buenos Aires, Argentina


                                    Endocrinol. Diabetes Nutr. 2020 (Received 29 January 2020, accepted 25 Mayo 2020)



Aim: The objective of this work is to define a value for the 99th percentile of high-sensitive troponin T and to evaluate the prognostic value of this biomarker in a population of patients with type 2 diabetes without cardiovascular history.

Methods: In this prospective. observational and analytic study. 482 patients with type 2 diabetes were enrolled. Patients were asymptomatic. with no history of cardiovascular events. renal insufficiency. or inflammatory or systemic disease. As events we considered a combined end point of major adverse cardiac events.

Results: 94.9% of patients had detectable troponin values. 20.7% of patients had troponin values above the healthy population reference upper threshold (14 pg/mL). The 99th percentile value for this patient population was 48 pg/mL. Age. sex. glomerular filtration rate and hypertension were associated with positive troponin values. 20 major cardiovascular events were observed after a median follow-up of 2.5 years. The incidence of events was 3.2% in patients with negative troponin values and 8% among patients with positive troponin values (p=0.02).

Conclusions: The 99th percentile value of troponin T in a population of patients with type 2 diabetes is 3-fold higher than the value proposed by the manufacturer for a healthy population. We also observed a significant difference in the distribution of troponin T values among men and women. This biomarker may be a valuable prognostic factor. since troponin T values above the reference upper threshold were associated with an increase in risk of cardiovascular events in these patients.


Keywords: Troponin. diabetes Mellitus Type 2. cardiovascular disease. 

Acknowledgements: Alejandro P. Adam PhD

Funding sources: This research did not receive any specific grant from funding agencies in the public. commercial. or not-for-profit sector.



Cardiac troponins (cTn I and T) are used as both diagnostic and prognostic biomarkers of cardiac injury. This is because the damage to cardiomyocytes leads too cTn release and thus a rise in blood concentration [1]. The use of cardiac troponins as biomarkers for ACS began in the early ‘90s. The original diagnostic assay was further developed to gain higher sensitivity with each iteration [2]. Since the first radioimmunoassay for cTn I in 1987. new methodologies were developed in order to improve diagnostic sensitivity. allowing for the detection of very low levels of circulating cTn. The current assays are termed high-sensitive troponin (hs-cTn) [3]. According to the recommendations of the International Federation of Clinical Chemistry and Laboratory Medicine task force on clinical applications of cardiac biomarkers. to qualify as high-sensitive. an essay must detect cTn in over 50% of healthy subjects [4]. However. this recommendation led to a reduction of assay specificity. because cTn were detectable in patients without heart damage due to acute myocardial infarction (AMI) [5].

Because the diagnosis of acute coronary syndrome (ACS) relies heavily on the troponin values. it is crucial to establish precise reference limits to evaluate hs-cTn results [1.6]. It has been proposed as reference range for the analysis of hs-cTn the 99th percentile of the healthy population [6]. However. there are still aspects of this proposal that require further debate: First, the “healthy population” used in the studies to determine the reference values is poorly defined. Second, a segregation based on sex [7.8] and age [9] is needed; third. differences in circulating cTn levels may exist among different groups of patients. Specifically. it has been shown that the increased glucose levels in patients with type 2 diabetes lead to coronary microvascular dysfunction and myocardial injury [10.11]. due to cellular oxidative stress [12]. This progressive myocardial damage at a sub-clinical stage of cardiovascular disease leads drives higher values of circulating cTn than the healthy population. even in the absence of ACS [13].

The objective of this study was to define the value of the 99th percentile for high-sensitive troponin T (hs-TnT) and to determine its prognostic value as biomarker in a population of patients with type 2 diabetes without cardiovascular history.


Materials and Methods

Research methodology and study population

This is an observational. analytical study with a prospective cohort. 482 patients were enrolled with a diagnosis of type 2 diabetes. asymptomatic. with no history of cardiovascular events. renal insufficiency. inflammatory disease or systemic disease.

Serum samples for measuring hs-TnT levels were obtained during routine patient controls. within the period between the years 2013 and 2017.

This study was approved by the Ethics Committee for Research Protocols of the Hospital Italiano de Buenos Aires.

Recruitment criteria were the following:

Inclusion criteria: patients over 18 years of age with a diagnosis of type 2 diabetes (according the American Diabetes Association) [14].

Exclusion criteria: creatinine clearance below 60 mL/min. according the CKD-EPI formula; history of acute coronary events and/or revascularization procedures; severe valvulopathy; clinically decompensated cardiac insufficiency; left ventricle systolic dysfunction with an ejection fraction below 35%; clinically manifest inflammatory disease; acute illness; anemia; neoplasms.

Laboratory methodology

Serum samples were obtained through conventional extraction and immediate centrifugation. An aliquot was stored at -70 °C until further processing. hs-TnT measurements were performed using the electrochemiluminescence assay Elecsys Troponin T hs STAT (Roche) and a COBAS e61 analyzer. The analytic sensibility of this method is 3 pg/mL (lower measurable concentration). As a reference value. the manufacturer proposes the use of the 99th percentile of the healthy population (14 pg/mL). which was verified in our laboratory for this system. The coefficient of variation (CV%) of this methodology is 1.8% for the level 1 (mean of control 1 = 29 pg/mL) and 1.3% for the level 2 (mean of control 2 = 2078 pg/mL). Glucose and creatinine measurements were performed using AU-5800 instruments (Beckman Coulter) via the UV-hexokinase method and the kinetic reaction (Jaffé). respectively. HbA1 measurement was performed via HPLC using a Variant II Turbo system.

Follow-up and event recording

We considered as final combined event an event comprised by death. ACS. requirement of revascularization. or stroke. The events were determined by manual. systematic evaluation of the clinical electronic history of each of the patients. Patients were followed until the occurrence of a first event or until the last medical control recorded in the clinical history. or until September 2017.

Statistical analysis

The statistical analysis was performed using the Stata V13.1 software (StataCorp. TX). Continuous variables are described as mean +- standard deviation when presenting a normal distribution or as median and interquartile range if the distribution was non-parametric. Analyses performed comprised Student’s T tests or Kruskal-Wallis for normal or non-parametric distributions. respectively. Qualitative variables are shown as number and frequency and were analyzed by Mann-Whitney χ2. A multivariate analysis was performed using logistic regression including all variables that resulted in statistically significant changes in univariate analyses. The β coefficients. 95% confidence intervals and statistical significance (p) were determined.

For the survival studies. two groups were separated based on the basal troponin levels being below or above the reference threshold established by the manufacturer. A Cox proportional risk model was used. and a Kaplan-Meier survival study was performed to establish the relationship between the basal troponin levels and major cardiovascular event incidence.



482 patients with type 2 diabetes were enrolled. with no history of cardiovascular disease and with conserved renal function. The basal characteristics of the population are described in Table 1. The median age of the population was 65 (iq25-75: 55-72) years. with 44% female and 56% male. Median serum troponin level was 9 (iq25-75: 6-13) pg/mL. 94.9% of patients showed measurable serum troponin levels. and 20.7% of patients showed a serum hs-TnT level above the reference upper threshold established by the manufacturer. The 99th percentile was 48 pg/mL. Women showed serum troponin levels significantly lower than men (Table 2).

Age. sex. glomerular filtration rate. and a diagnosis of hypertension (HT) were associated with positive troponin levels (> 14 pg/mL). A progressive increase in troponin levels was observed with age and with reduced glomerular filtration rate. Neither fasting glucose or glycosylated hemoglobin. or body mass index. were correlated with serum troponin levels. The patients with positive troponin levels showed lower levels of total cholesterol than those with negative troponin levels (172±44 vs 185±39 mg/dL; p=0.0071).

Even though no differences were observed with the use of statins. it is not known the strength and quantity levels of each treatment. We did not find significant differences in the rate of use of insulin. oral hypoglycemic medications. aspirin or beta blockers. We observed a significant increase in the rate of use of angiotensin-converting enzyme inhibitors (ACEi) or angiotensine II receptor blockers (ARB) in patients with positive troponin values. This finding is consistent with an increased prevalence of hypertension among patients with positive troponin. We performed a multivariate logistic regression analysis with elevated troponin (>14 pg/mL) as dependent variable for all other variables that were significantly associated (p<0.05) with troponin levels in all univariate tests. Only increased age. male gender. and lower glomerular filtration rate were significantly associated with elevated troponin levels (Table 3).

On follow-up with a median of 2.5 years. we observed 18 major cardiovascular events (3.84%). These include 6 deaths (1.28%). 9 ACS events (1.93%) and 3 stroke events (0.6%). While the mortality rate was higher among patients with positive troponin levels. this difference did not reach statistical significance (1.48 vs. 0.32%; HR=4.65 IC95 0.94-23%; p=0.06) (Table 4).

The yearly rate of major cardiovascular events on the total population was 1.58% per year. mortality 0.5% per year. and the rate of coronary events 0.79% per year. The patients with troponin levels above the reference threshold showed a rate of major cardiovascular events of 3.96%. while all others showed a rate of 1.07% (HR=3.83. IC95 1.49-9.58%. p=0.005). Similarly. the rate of ACS events in patients with troponin levels above the threshold was 1.98%. (Table 5).

Overall. we found that in patients with type 2 diabetes without any other associated pathology. a measurement of elevated hs-cTn levels is a marker with high specificity (80%) but poor selectivity (44%) and low predictive value for cardiovascular events (8%). On the contrary. a normal troponin level in these patients is a strong negative predictor for cardiovascular events (97%) (Figure 1).



The data presents in this study shows that detectable troponin levels can be measured in 95% of patients with type 2 diabetes. asymptomatic. with no history of cardiovascular events. renal insufficiency. inflammatory disease or systemic disease. Of these. 21% presented hs-TnT levels above the reference threshold. In this population. the 99th percentile was three times as high as the reference threshold proposed by the manufacturer. Similar results were observed in diabetic patients with an established cardiovascular disease [15]. Nevertheless. it is noteworthy that we observed these high values in patients with type 2 diabetes with no history of cardiovascular disease. normal renal function and no clinical evidence of systemic or inflammatory disease.

It is well known that conventional assays allow the measurement of serum troponin levels in less than 1% of the population. while the use of high-sensitive assays allows for the detection of much lower troponin levels [16]. thus increasing the rate of detection among patients. Because of this. the distribution of troponin values depends on the biochemical characteristics of each assay and of the population under study [4.17]. The value of the 99th percentile proposed by the manufacturer as a threshold for medical decision of AMI depends on the corresponding population [17]. because these threshold values are established on populations that are presumably healthy or “normal”. However. because troponin levels are affected by the gender. age. ethnicity. and other factors that may influence this value on a given population. it has been proposed to determine different threshold values according to the corresponding population [16].

The studies to determine the range of reference values for troponin assays demonstrated that the 99th percentile for men is substantially higher than for women. which could be explained by the higher heart mass in men [18]. This difference can be observed not only on the value for the 99th percentile. but also on the percentage of patients with undetectable hs-TnT levels (10.9% in women vs 0.74% in men). Due to this. it may be necessary to use gender-specific values for the 99th percentile for patient evaluation [4]. For example. Roche included gender-specific threshold values on a assay recently approved by the FDA [17].

Our data also shows a progressive increase in hs-TnT levels with age. similar to other reports [19]. Even in healthy individuals (according to clinical. biochemical and imaging criteria). increasing hs-cTn levels with age were observed. suggesting that this increase may be due to subtle changes that may have been undetected in the clinical characterization of these individuals [19]. In another study performed on aging patients. the use of an age-adjusted reference threshold reduced the rate of hs-cTn false positives by approximately 90%. greatly improving its diagnostic precision [19].

In the case of patients with type 2 diabetes. the interpretation of hs-cTn values must be made with caution. taking into consideration the observed difference between the threshold values proposed by the manufacturer and the threshold value determined by this study. In these patients. the use of a lower threshold value determined for a healthy population may lead to an overestimation of ACS [16].

Rubin et al [20] demonstrated that. in 9661 patients from the ARIC study. the troponin levels above the manufacturer’s reference threshold increased progressively with increasing levels of glycosylated hemoglobin. We did not. however. observe a relationship between troponin and glycosylated hemoglobin levels. This difference may be due to the fact that in our study all patients were diabetics and that the vast majority were under pharmacological treatment. which can clearly affect the levels of glycosylated hemoglobin [20].

Our study opens the question as to why these diabetic patients with no history of significant coronary disease had elevated levels of troponin and what is its clinical meaning. The mechanisms by which hyperglycemia leads to myocardial damage. as evidenced by elevated troponin values. are not well understood. It has been proposed that the deleterious effect of chronic hyperglycemia in diabetic patients is due to an increased oxidative stress that leads to microcirculatory dysfunction and thus myocardial damage [21]. Previously published mechanisms may explain this process. such an increase in polyols flux. an increase in the intracellular formation of final products of exacerbated glycosylation. the activation of protein kinase C. and an increase in the hexosamine pathway [22]. We cannot also exclude a role for possible atherosclerosis leading to a degree of silent ischemia. Diabetes also produces an increase in subclinical atheromatosis as evidenced by imaging studies. which then leads to increased silent myocardial damage as evidenced by higher biomarker levels [23].

This would not only allow the use of troponin levels as a diagnostic tool for ACS. but also would provide prognostic information in diverse clinical situations: lung thromboembolism. patients with stable coronary disease. and even in the general population [24]. We can conclude from the data presented in this study that the detection of hs-TnT levels above the reference threshold is associated with an increased risk of cardiovascular events. However. this association must be adjusted by age and creatinine clearance. If we consider the troponin values as a continuous variable. its levels are associated with the incidence of cardiovascular events independently of age and glomerular filtration rate. This finding is in agreement with a previous study performed in diabetic women with no history of cardiovascular disease that showed that detectable hs-TnT levels. even below the reference threshold. were independently associated with the incidence of cardiovascular events [25].

There is evidence of an increased risk of cardiovascular events in individuals under primary prevention with elevated hs-cTn levels [21.26]. It is also known that changes in hs-cTn levels over time are associated with increased cardiovascular risk [27]. However. there is no evidence of a treatment that can modify the troponin levels and alter the patient’s clinical evolution [28].

Lastly. the finding that the 99th percentile hs-TnT levels in a given population is higher than the reference threshold raises the question of whether the reference value for ACS diagnosis on a patient that shows at the emergency department should be adjusted. Twerenbold [29] proposes that the threshold values for diagnosis of ACS should be adjusted in a population of patients with diminished renal function. This value would be between 2 and 3 times the value for the reference population. Similar to others [30]. we do not share this view. because in the presence of either diabetes or diminished renal function. even if the specificity and predictive value of this test are reduced. neither the sensitivity nor the negative predictive value are affected [30]. In the case of patients that belong to a population of chronically elevated troponin levels. it is important to evaluate the changes above or below the basal troponin levels. considering significant any change beyond 20% to discriminate between chronic damage and acute injury [6].

The lack of imaging data to confirm the absence of cardiovascular disease and the exclusion of inflammatory disease via biomarkers are limitations of the current study. Even if it is possible that a bias may be generated by the lack of evaluation of these entities. we consider that our study reflects the standard clinical practice. where a physician at an emergency department must evaluate the patient and determine a course of action in the absence of such information. The data obtained did not include the period of evolution of diabetes and we do not know the intensity of pharmacological treatment that these patients received. This may be a cause for the lack of correlation between glycosylated hemoglobin and troponin levels. Finally. it is important to note that the rate of events may be biased by the characteristics of the population under study.


The 99th percentile of hs-TnT levels in a population of patient with type 2 diabetes with no history of cardiovascular disease. normal renal function and no systemic or inflammatory disease was 3 times higher than the reference threshold proposed by the manufacturer for a healthy population. Also. we observed a difference in the distribution of hs-TnT levels between men and women. The measurement of elevated hs-TnT levels was associated with an increased risk of cardiovascular events.

Figure - Tables



Table 1. Population characteristics according to the presence of hs-TnT ≤ or> 14 pg/mL


Total (N=482)

hs-TnT ≤14 pg/mL (N=382)

hs-TnT >14 pg/mL (N=100)


Age (median - IQ 25-75) (years)

65 (55-72)

63 (53-69)

72 (65-79)


Female sex - N (%) 

211 (43.7)

184 (48.2)

27 (27)


Fasting glucose (median- IQ 25-75


169 (151 - 205)

169 (151 - 205)

167 (150 - 206)


HbA1c % (median - IQ 25-75)

7.8 (7.1 - 9.1)

7.8 (7.1-9.0)

7.9 (7-9.3)


BMI (median - IQ 25-75)

30.5 (27 - 35)

31 (27-35)

30 (26-34)


Hypertension - N (%)

315 (66.8)

241 (64.3)

74 (76.3)


Total Cholesterol 

(median - IQ 25-75) (mg/dL)

179 (154 - 210)

181 (156-211)

167 (140-204)


CrCl (median - IQ 25-75) (mL/min)

84 (70 - 96)

87 (73-100)

76 (60-84)


Insulin - N (%)

137 (29.2)

106 (28.3)

31 (32.3)


Hypoglycemic drugs - N (%)

400 (85.4)

324 (86.9)

76 (79.2)


Aspirin - N (%)

156 (33.5)

126 (33.9)

30 (31.6)


ACEi/ARB - N (%)

255 (54.6)

191 (51.2)

64 (67.4)


Statins - N (%)

241 (51.6)

189 (50.7)

52 (54.7)


Beta Blockers - N (%)

126 (26.9)

95 (25.5)

31 (32.6)


CrCl: creatinine clearance; HbA1c: glycated hemoglobin A1c; BMI: body mass index; ACEi: angiotensin-converting enzyme inhibitors; ARB: angiotensine II receptor blockers



Table 2. Distribution of hs-TnT values ​​by sex


Total = 482

Women = 211

Men =271


Median hs-TnT (IQ 25-75) (pg/mL)

9 (6-13)

7 (4-11)

10 (7-15)


99th percentile hs-TnT (pg/mL)





Below lower limit of detection - N (%)

25 (5.2)

23 (10.9)

2 (0.74)


hs-TnT >14 pg/mL - N (%)

100 (20.7)

27 (12.8)

73 (26.9)



Tabla 3: Characteristics associated with the elevation of hs-TnT


Univariate analysis

Multivariate analysis










0.05- 0.09

< 0.0001


0.02 – 0.08

< 0.0001



0.43 – 1.40

< 0.0001


0.61 – 1.76

< 0.0001



-0.06 – -0.03

< 0.0001


-0.04 – -0.01


Fasting Glucose


-0.00 – 0.00





-0.10 – 0.15





0.06 – 1.09



-1.25 – 0.33


Total Cholesterol


-0.01 – -0.00



-0.008 – 0.004




-0.06 – 0.01





-0.29 – 0.67



Hypoglycemic drugs


-1.13 – 0.02





-0.58 – 0.37





0.20 – 1.15



-0.27 – 1.14




-0.28 – 0.61



Beta Blockers


-0.14 – 0.83



CrCl: creatinine clearance; HbA1c: glycated hemoglobin A1c; BMI: body mass index; ACEi: angiotensin-converting enzyme inhibitors; ARB: angiotensine II receptor blockers.


Tabla 4: Risk of vascular events in patients with hs-TnT above the reference limit ( >14 pg/mL).





















                            A.C.S: Acute coronary síndrome; MACE: major adverse cardiac events


Tabla 5: MACE predictors

Univariate analysis

Multivariate analysis










1.00 - 1.08



0.95 - 1.05




0.98 - 8.99





0.00 - 56



Total Cholesterol


0.98 - 1.00



Fasting Glucose


0.99 - 1.01





0.84 - 1.37





0.93 - 0.99



0.93 - 1.01














hs-TnT > 14 pg/mL







HbA1c: glycated hemoglobin A1c; CrCl: creatinine clearance.