Homocysteine (Hcy) is a non-protein amino acid that contains sulfur and is derived from the essential amino acid methionine [1, 2]. Previous studies reported that Hcy was associated with cardiovascular diseases such as aortic atherosclerosis, hypertension, coronary artery disease and acute myocardial infarction [3‐9].

The QT interval is the electrocardiographic manifestation of ventricular depolarization and repolarization. A prolonged QT interval is closely related to ventricular arrhythmias, which can cause sudden cardiac death (SCD) [10]. In mice, the QT interval was found to be affected by Hcy levels [11]. Recent studies determined that Hcy was also related to arrhythmias and SCD, which indicated the possibility that Hcy was associated with cardiac electrophysiology [7, 8, 12‐15]. However, few studies focused on the relationship between Hcy and the QTc interval in the general population. Detection of the QTc interval can help predict the risk of cardiovascular disease (CVD) events [16‐19]. Therefore, in this study, we evaluated plasma Hcy levels and the QTc interval in a general population to determine whether there is an association between Hcy levels and the QT interval.

The present study of 7002 participants consisted of 3260 males and 3742 females. The mean participant age was 54 years, and the participant age ranged from 35 to 93 years. The mean Hcy level of the population was 17.32 ± 12.34 umol/L. Among the included population, 4695 and 2307 demonstrated QTc ≤ 440 ms and QTc >440 ms The distribution of Hcy levels was determined for the entire population after the data were grouped into quartiles: Q1 contained Hcy levels ≤11.1 umol/L; Q2 contained Hcy levels >11.1 umol/L and ≤13.8 umol/L; Q3 contained Hcy levels >13.8 umol/L and ≤18.2 umol/L; and Q4 contained Hcy levels >18.2 umol/L. All variables are presented in the Table 1. Except for sleep duration, serum potassium, and medication used, all variables were significantly different across the four groups (all p < 0.05).

Table 2 presents the baseline characteristics of the study population according to the QTc interval. Participants with QTc > 440 ms had a greater history of heart disease, lower physical activity and medication use, but lower rates of current smoking and drinking than those participants with QTc ≤ 440 ms (all P < 0.001). Participants with QTc > 440 ms had higher BMI, SBP, DBP, FPG, TC, TG and LDL-C levels than those participants with QTc ≤ 440 ms (all P < 0.001). No significant differences in serum calcium or magnesium were observed between the two groups. These results were consistent with the results obtained using the other two methods for QT correction.

Figures 1 and 2 illustrates the QTc intervals according to the Hcy quartiles in boxplot. Using the Bazett formula. Regardless of the correction method, higher mean QTc values were observed in Q1 and Q4, compared to those in Q2 (p < 0.01). There were no significant differences in the mean QTc values between Q2 and Q3. And, as shown in Fig. 3, the mean Hcy levels among participants who had QTc > 440 ms were higher than those who had QTc ≤ 440 ms(p = 0.031).

Table 3 presents the relative odds of QT intervals. Using QTc intervals >440 ms as dependent variables and Hcy quartile as independent variables, multivariable logistic regression analyses were applied to assess the association between Hcy levels and the QTc interval. Model 1 was adjusted for age and gender, model 2 was adjusted for factors in model 1 and lifestyle variables including education level, family income, diet score, sleep duration, current smoking status, drinking status, and physical activity, model 3 was adjusted for factors in model 2 and CV risk factors included body mass index, waist circumference, systolic blood pressure, diastolic blood pressure, and total cholesterol, and model 4 was a multivariable model adjusted for factors in model 3 and other factors which was considered could empirically affect the Qtc directly including serum calcium, potassium and magnesium, history of heart disease, and any medication which . As shown in Table 3, in model 3, the risk of QTc intervals >440 ms in the first and fourth quartiles increased compared with those in the second quartile (P < 0.05). The odds ratio (OR) was 1.20 (95% CI 1.03–1.40) and 1.41 (95% CI 1.120–2.01). We also considered serum calcium, potassium and magnesium, history of heart disease, and any medication into the final, fully adjusted model (model 4) to rule out their impacts. The results were consistent. Compared with the second quartile, the risk of QTc intervals >440 ms in the first and fourth quartile increased (P < 0.05), and the OR was 1.23 (95% CI: 1.05–1.43) for Q1 and 1.40 (95% CI: 1.19–1.65) for Q4.

In the Additional file 2, we presented the result from linear regression analyses for association between homocysteine and QTc interval.

In the present study, we evaluated the relationship of different Hcy levels to the QTc interval in the general population. We found that the odds for QTc > 440 ms in the first and fourth quartile of the Hcy level increased compared with those in the second quartile.

Prior studies explored the relationship between Hcy levels and arrhythmias. Duyuler et al. found an independent relationship among the iron status, Tp-e interval and Tp-e/QT ratios of elite sport players that were also not correlated with serum homocysteine levels [33]. Previous researchers also found that the homocysteine levels determined prior to electrical cardioversion could predict recurrence of AF after successful restoration of sinus rhythms [34]. However, few studies focused on the QTc interval. Rosenberger et al. found a highly significant prolongation of the QTc interval in male C57/BL6J mice that received standard rodent chow and drinking water supplemented with 400 mg DL-homocysteine compared with the controls, which received standard rodent chow and water ad libitum [11]. Acampa et al. measured Hcy plasma levels, P wave dispersions (PWD), and the QTc interval in 32 patients who had orthotropic heart transplantation (OHT) and in 20 control subjects. Acampa found that, in OHT patients, plasma Hcy levels significantly correlated with PWD, whereas no correlation was found with the QTc interval [13]. Lee Yin Leng et al. concluded that in community populations, homocysteine was moderately elevated when QRS durations were >120 ms, but QTc intervals were not significantly associated with plasma homocysteine [26]. In our study, Linear regression and multiple logistic regression analyses results showed an association between QTc interval and Hcy levels. Similar to other studies in China [35], we obtained the elevated average level of Hcy in this general population. Some studies reported a high prevalence of HHCY in Chinese populations, and smoking status, alcohol consumption, education level, physical activity and MTHFR genotypes were significant determinants. These factors also may explain the high levels of Hcy in this study.

To the best of our knowledge, this study is the first to demonstrate a correlation between Hcy levels and the QTc interval in a general population. Our results confirmed that a link exists between Hcy levels and cardiac electrophysiology. However, the mechanism of Hcy-mediated cardiac arrhythmias is still unclear. According to laboratory studies, elevations in Hcy levels can inhibit cardiac K1 channels, which may delay ventricular repolarization [8, 36]. Homocysteine may decrease the Na + channel activity in in vitro cardiac preparations [9, 37‐39]. Moreover, biochemical damage from high homocysteine levels on the atrial extracellular matrix involves the activation of matrix metalloproteinases-9 and extracellular signal regulated kinase and causes subsequent atrial fibrosis with slow and a heterogeneous atrial conduction, favoring the appearance of a vulnerable reentrant substrate [40‐42]. However, these results do not explain the link between Hcy levels and the QTc interval. Therefore, future research is needed regarding this relationship. In addition, research is necessary to identify a mechanism that explains how both higher and lower levels of Hcy affect the QTc interval.

Our study had some limitations. First, we sampled adults only in the Liaoning Province, and could expect varying results if we had sampled adults in other Chinese provinces. Second, electrocardiograms for the participants were performed at one time, which may have affected our results. In addition, our results were based on a cross-sectional design, and thus no cause-and-effect relationships could be established. Longitudinal studies are required for further investigation of these findings. And, in this study we used 440mms as the cut off, we agree with that opinion that dichotomization is debatable and variability of inter observer cannot be completely avoided. However, the linear regression gave the similar result. In this paper we can only present an association between homocysteine levels and QTc but can’t conclude the correlation between homocysteine and prolonged QTc. Finally, beats with arrhythmias including AV-blocks were not deleted from QT measurements. Even though we used the heart rate with the correct QT interval, the effects of irregular heart rates could not be avoided.

In this general population, Hcy levels are associated with the QTc interval. The result provides novel, population-based evidence of the relationship between Hcy levels and the QTc interval.