The plasma levels of free melatonin is not associated with cardiovascular events after acute myocardial infarction

Background Lower circulating levels of total melatonin is associated with adverse cardiovascular (CV) events in acute myocardial infarction (AMI) patients. Free melatonin is easier to measure in clinical practice compared with total melatonin. Whether free melatonin is associated with follow-up CV events in AMI patients has not been determined yet. Methods A total of 732 consecutive AMI patients treated with percutaneous coronary intervention between January 2013 and January 2015 participated in the study. Blood samples were collected as fast samples on the �rst morning after admission. The plasma levels of free melatonin were determined using non-extraction radioimmunoassays. The cox regression was used to explore the association between circulating melatonin and endpoints. The median follow-up was 31.6 months. Results Patients with high melatonin levels were more likely to be younger and to have poorer blood lipid control. Multivariate cox-regression analyses (adjusted for confounding variables) showed that one unit increase in log-transformed melatonin was not associated with increased risks of major adverse CV events (MACE, composite of cardiovascular death, myocardial infarction, stroke and heart failure, hazard ratio [HR], 1.74; 95% con�dence interval [CI] 0.94 to 3.21; p =0.078). Conclusions Higher free melatonin levels on the onset of AMI is not associated with MACE in AMI patients, independent of established conventional risk factors.


Background
Melatonin is the metabolite of tryptophan, which is an important regulator of the body's internal timekeeping system.It participates in major physiological processes including the sleep wake cycle, pubertal development and seasonal adaptation [ 1].Melatonin is mainly synthesized in the pineal gland, while is also produced by retina, gastrointestinal tract, skin, lymphocytes, platelets and bone marrow [ 2].Plasma melatonin is mainly produced during the night while keeps at low levels during the day, following circadian rhythm [ 3].Melatonin receptors have been identi ed in the human coronary arteries, aorta and left ventricles [ 4], and melatonin participates in a variety of cardiovascular (CV) pathophysiological processes including anti-in ammatory, antioxidant, anti-hypertensive and possibly as an antilipidemic function [ 5].
In the past decade, accumulating evidence has demonstrated that melatonin might serve as a potential prognostic factor in myocardial infarction (MI).Nocturnal total melatonin levels decreased after acute MI (AMI) [ 6], and lower nocturnal total melatonin concentrations after AMI are associated more left ventricular remodeling [ 7] and heart failure or cardiac deaths in 6-month follow-up [ 8].Moreover in women with increased BMI, lower nocturnal total melatonin secretion is associated with higher risks of MI [ 9].Two major method exists to detect melatonin in clinical practice.The nonextraction melatonin method mainly detect free melatonin while the methanol-based extraction method detect both bound and free melatonin [ 10].Compared with total melatonin measurement, free melatonin is easier to measure in clinical practice both as it does not need extraction which is hard to chieve in clinical practice and also it require much less sample volumes (100uL vs. 500uL).Whether free melatonin can provide prognostic information in patients with AMI has not been determined yet.The circulating levels of free melatonin and its relations to clinical characteristics, cardiac injury biomarkers, and adverse cardiovascular events in AMI patients were studied in this research.

Study population
A total of 732 consecutive patients with diagnosed AMI, consisting of ST-segment elevation (STEMI) and non-ST-segment elevation myocardial infarction (NSTEMI), admitted to Department of Cardiology, People's Liberation Army General Hospital (PLAGH) between January 2013 and January 2015 were included.The study protocol was approved by the local research ethics committee (PLAGH).The de nition of AMI is the presence of a positive cardiac troponin T or creatine kinase MB test, typical chest pain lasting at least 20 min, and electrocardiograph changes indicative of ischemia (ST segment elevation or depression).The PCI operation and postoperative medication (including aspirin, clopidogrel, β-blockers, statins, angiotensin-converting-enzyme inhibitors or angiotensin II receptor blockers and so on) were conducted according to the current standard guidelines [ 11,12].Patients with autoimmune diseases, collagen tissue diseases, drug addiction, patients receiving immunosuppressive treatment, taking sedatives, antiepileptic drugs, tricyclic antidepressants or any medication known to in uence melatonin metabolism, psychiatric sleeping disorders, shift workers, and subjects with jet-lag syndrome were excluded from this study.Written consents were gotten from all patients.The median follow-up time is 31.6 months.This study was registered on clinicaltrial (clinicaltrials.gov,NCT03230630).

Outcome events and follow-up
Demographic, clinical, and biochemical data were obtained from electronic medical record system.The primary endpoint was major adverse cardiovascular events (MACE) including cardiovascular death, myocardial infarction, stroke and heart failure.Cardiovascular death is de ned as death in the presence of ACS, signi cant cardiac arrhythmia, or refractory congestive heart failure.Hospitalization for heart failure was de ned as a hospital readmission for which heart failure was the primary reason.Stroke is de ned as persistent central nervous system de cit, usually with con rmatory CT imaging.The endpoints were obtained from outpatient follow-up, in-hospital clinical records of the re-hospitalized patients or by contacting each patient or their relatives individually.

Biochemical measurements
The blood were collected as fast samples in an EDTA-tube between 5:30 a.m. and 06:00 a.m. on the rst day after admission.We collect blood samples accompanied with the patient's routine fast morning blood collection.The patients were gently awakened from sleep just before the blood collecting with only bed lamp on.Lights were turned off at 10:00 p.m. and on at 06:00 a.m. in all the wards.The plasma was frozen at -80°C until further analysis.The plasma melatonin levels were determined using a nonextraction radioimmunoassay (RIA) kit (Cat: BA R-3300, LDN, Germany) on XH6080 Radioimmuno-detector (Xi'an Nuclear Instrument Factory, China) according to the manufacturer's instructions.The detection limit for melatonin was 2.3 pg/mL.Total cholesterol, triglycerides, low-density lipoprotein (LDL) cholesterol, highdensity lipoprotein (HDL) cholesterol, fast plasma glucose (FPG), cardiac troponin T (cTNT), myoglobin, MB isoenzyme of creatine kinase (CK-MB), NT-pro-brain natriuretic peptide (NT-proBNP), aspartate aminotransferase (AST), gamma-glutamyltransferase (GGT), alanine aminotransferase (ALT) and creatinine were measured on autoanalyzer (Cobas C and E system, Roche, Swiss).Haemostatic pro le, including thrombin time (TT), prothrombin time (PT), activated partial thromboplastin time (APTT) and brinogen were measured on the STA-R Evolution® device (Diagnostica Stago®, Asnières sur Seine, France).We collected the rst biochemical test results after admission.All biochemical measurements were performed by investigators who were blinded to patients' characteristics and outcomes.

Echocardiographic measurements
Echocardiography was performed during hospital stay.Transthoracic two-dimensional and M-mode echocardiographic data were obtained.Doppler recordings of mitral in ow were also performed by placing a 2.5 mm sample volume at the tip of the mitral valve lea ets and recording the pulsed wave Doppler signal using VIVID 7 system (GE, USA).Peak velocity of early (E) and atrial (A) diastolic lling were measured and the E/A ratio calculated.Left ventricular ejection fraction (LVEF), left ventricular systolic volumes (LVESV) and left ventricular end-diastolic volumes (LVEDV) were calculated as previously reported [ 13].All analyses were performed by an experienced operator blinded to melatonin values and clinical parameters of patients.

Statistical analysis
Analyses were conducted using SPSS 13.0 software (IBM, USA).Patients were divided into tertiles based on the values of plasma melatonin levels.Normality of continuous variables was determined by the Kolmogorov-Smirnov test.Normally distributed continuous variables were presented as mean with standard deviation (SD), variables with skewed distribution as median with interquartile range, and categorical variables as number with percentages.Baseline characteristics were compared across plasma melatonin tertiles using ANOVA test for normal distribution and Kruskal-Wallis for asymmetric distribution and chi-square test for discrete variables.Trend analysis were conducted using trend test from generalized linear model for continuous variables and Cochran-Armitage trend test for discrete variables.The plasma melatonin levels were left-skewed, and therefore, log-transformed (natural logarithms) prior to inference testing.
Hazard ratios (HR) and 95% con dence intervals (CI) for endpoints were calculated using Cox proportional hazard analyses by melatonin tertiles (using the rst tertile as reference).For logtransformed melatonin as a continuous variable, the HR (with 95% CI) for endpoints were calculated per unit increment of log melatonin.In addition to the variables that resulted signi cant (p < 0.05) from univariable Cox proportional hazards model using the forward variable selection method, some other factors biologically shown to be associated with death in AMI patients were also included into the multivariable model [ 14].

Baseline data
The mean age of 732 study participants was 58.7 ± 10.6 years; 20.2% were female.The distribution of plasma melatonin is left-skewed (Supp.Figure 1).The median plasma melatonin levels were 27.47 (interquartile range, 13.82-53.62)pg/ml.There was no signi cant difference between male and female patients (28.29 (14.18-53.68)pg/ml for the male versus 24.56 (12.53-52.29)pg/ml for the female; P=0.37), between diabetic and non-diabetic patients (24.76 (11.66-49.00)pg/ml for the diabetic versus 28.58 ( Kaplan-Meier survival analysis was carried out to compare the difference in survival rate in AMI patients according to tertiles of plasma melatonin.The results showed that higher melatonin was not associated with composite CV outcomes (Figure 1).Treating log melatonin as a continuous variable that 1 unit increase of log melatonin was not associated CV outcomes (HR: 1.74; 95% CI: 0.94-3.21;p =0.078), and patients in the highest tertile of plasma melatonin levels (≥41.50 pg/mL) were not associated with composite CV events (HR: 1.79; 95% CI: 0.99-3.25;p =0.054) compared with patients in the lowest tertile (< 19.03 pg/mL).Subgroup indicate there is a possible interaction between baseline pro-BNP and melatonin as to composite CV outcomes.

Discussion
This study demonstrates that increased plasma free melatonin levels are not associated with cardiovascular outcomes.There exists a possible interaction between pro-BNP and melatonin.
In cell and small animal MI model studies with no comorbidities and comedications, melatonin could attenuate post-MI injury by breaking the cycle of mitochondrial impairment and ROS generation [ 15].
Melatonin could up-regulate autophagy, decrease apoptosis and modulate mitochondrial integrity and biogenesis thus alleviating post-infarction cardiac remodeling and dysfunction [ 16].However it's also reported physiological melatonin concentrations are important in reducing the I/R-induced myocyte damage, while additional pharmacological concentrations did not add to the bene cial effect [ 17].In a large animal closed-chest porcine model, melatonin usage immediately prior to reperfusion failed to reduce MI size [ 18], suggesting complex cardioprotection with melatonin in different MI models [ 19].In addition to experimental studies in animal models, a recent clinical randomized trial has found intravenous and intracoronary usage of melatonin is not associated with a reduction in infarct size in STEMI patients and has an unfavourable effect on the ventricular volumes and LVEF evolution [ 20,21], however in subgroups of timely treated patients (shorter symptoms onset to balloon time) the administration of melatonin is associated with a signi cant reduction in the infarct size [ 22].In another study, oral melatonin started on the night following primary PCI and continued daily during the hospitalization had con icting results on enzymatic markers of MI injury following STEMI [ 23].These results suggest that melatonin may be effective only in certain kinds of MI or within a speci c period of time.
Nocturnal plasma total melatonin levels decreased in AMI patients compared with controls without cardiovascular diseases [ 6].The AMI patients who experienced adverse events during 6 month follow-up had signi cantly lower nocturnal total melatonin levels than patients without events in this study [ 6].And a recent published study has reported that lower uric melatonin secretion was signi cantly associated with a higher risk of MI [ 9].In our study both STEMI and non-STEMI patients were included, and diabetic and cancer patients were not excluded, while in the previous study only STEMI patients without diabetes and cancers were included.Moreover the plasma samples were obtained between 5:30 and 6:00 a.m. on the rst day after admission in our study while serum samples were collected at 2:00 a.m in theirs.Until now no study has reported the changes of melatonin levels over time in AMI patients.
It should be pointed out that the detection methods is different between our study and previous ones.We used a nonextraction melatonin detection method, while previous studies used methanol-based extraction method.The nonextraction melatonin method mainly detect free melatonin while the methanol-based extraction method detect both bound and free melatonin [ 10].It's reported in amniotic uid there was no signi cant difference between melatonin concentration measured by non-extraction versus extraction method in early pregnancy women while in late pregnancy the melatonin concentration was signi cantly lower when measured by non-extraction method than when by extraction one [ 24].Future studies are needed to clarify this by using both two methods at the same time in the AMI population.

Limitations
We only used the non-extraction but not the extraction method to detect plasma melatonin, as mentioned above the results may vary greatly between the two methods.We did not measure the nocturnal levels of melatonin at different times at night, it is possible that at 5:30 a.m., the levels of melatonin do not coincide with its peak.Moreover it's a single center study with a Chinese population, future studies is need to clarify the results in different populations.The study is observational and the authors cannot exclude the possibility of residual confounding.

Conclusion
The current study shows that free circulating melatonin levels is associated with risks of cardiovascular death, and is associated with decreasing risks of non-cardiovascular death in AMI patients.In summary, the current study shows that free circulating melatonin levels is not associated with risks of cardiovascular outcomes in AMI patients.Tables Table 1 Patients' characteristics according to plasma free melatonin tertiles Data are presented as mean ± SD, median (interquartile range) or numbers (percentages).ACEI, angiotensin converting enzyme inhibitors; ARB, angiotensin receptor blocker; BMI, body mass index; DBP, diastolic blood pressure; MI, myocardial infarction; SBP, systolic blood pressure; STEMI, ST-elevation myocardial infarction.

Supplemental Figure Legend
Supp Figure 1.Plasma levels of free melatonin was left-skewed in AMI patients.

Figures
Figure 1 Results of Kaplan-Meier analysis of cumulative event-free rates in AMI patients.
Besides presenting a crude model, variables considered for multivariate model 1 included age, sex, for model 2 age, sex, smoking, body mass index, creatinine, triglyceride, peak CK-MB, peak NT-proBNP, GGT, type 2 diabetes and previous myocardial infarction and types of myocardial infarction (STEMI, NSTEMI).All reported p values are based on two-tailed tests with <0.05 considered as statistical signi cance.

Table 3 .
Multivariable Cox Proportional Hazards Models for predicting composite CV outcomes by tertiles of free melatonin Model 1 adjusted for age and sex.Model 2 adjusted for age, sex, smoking, body mass index, creatinine, triglyceride, peak CK-MB, peak NT-proBNP, type 2 diabetes and previous myocardial infarction.