2025 年 32 巻 6 号 p. 734-752
Aims: Plasma S-adenosylhomocysteine (SAH) level is positively associated with cardiovascular risk. However, the relationship between plasma SAH levels and the risk of all-cause and cardiovascular mortality remains unknown. This study aimed to explore the relationship between plasma SAH levels and the risk of all-cause and cardiovascular mortality in patients with coronary artery disease (CAD).
Methods: Plasma SAH levels were measured in 1553 patients with CAD. The association between plasma SAH level and the risk of all-cause and cardiovascular mortality was estimated using Cox Proportional hazards regression models.
Results: Relative to participants in the lowest quartile of plasma SAH levels, those in the highest quartile of plasma SAH levels had a higher risk of all-cause death (adjusted Hazard Ratio [HR], 2.15; 95% CI, 1.54-3.01; P＜0.001) and cardiovascular death (adjusted HR, 2.20; 95% CI, 1.49-3.25; P=0.001) in the age- and sex-adjusted model. The results of the multivariable adjusted analysis were similar (all-cause death [adjusted HR, 1.81; 95% CI, 1.27-2.58; P=0.002] and cardiovascular death [adjusted HR, 1.84; 95% CI, 1.21-2.79; P=0.031]). The age- and sex-adjusted HRs for each 1 SD increase in plasma SAH level were 1.30 (95% CI, 1.22-1.38) for all-cause mortality, and 1.34 (95% CI, 1.25-1.43) for cardiovascular mortality, respectively. A 1 SD increase in the SAH level was associated with a 25% higher risk of total death (adjusted HR, 1.25; 95% CI, 1.17-1.34) and a 29% greater risk of cardiovascular death (adjusted HR, 1.29; 95% CI, 1.20-1.39) in multivariable adjusted analysis.
Conclusions: We found that the plasma SAH level is positively correlated with the risk of all-cause and cardiovascular mortality in patients with CAD in both age- and sex-adjusted and multivariable-adjusted models.
Si Liu, Yongyi Wang, and Mengfeng Yang contributed equally to this work.
Yunjun Xiao, Changhua Zhang and Haiyan Huang are joint senior authors.
Abbreviations: SAH, S-adenosylhomocysteine; CAD, coronary artery disease; CVD, cardiovascular disease; SAM, S-adenosylmethionine; Hcy, homocysteine; GPCDC, Guangdong Provincial Centers for Disease Control and Prevention; GS, Gensini Score; FFQ, food frequency questionnaire; tCys, total cysteine; tHcy, total homocysteine; CV, coefficient of variation; BMI, Body mass index; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; hsCRP, high-sensitivity C-reactive protein; LDL, low-density lipoprotein; HR, hazard ratio.
According to the most recent data from the WHO, it is estimated that approximately 17.9 million people die from cardiovascular diseases (CVD) each year1). Coronary artery disease (CAD), a major CVD, is the leading cause of death and disability worldwide2, 3). It manifests as partial or total obstruction of the coronary arteries supplying the myocardium and is characterized by angina, myocardial infarction, or sudden cardiac death4). Early intervention by addressing the main modifiable risk factors for CAD is critical to reduce the incidence of coronary events and mortality5). Therefore, improving risk prediction for the identification of individuals who are at highest risk for the development of CAD and providing them appropriate treatment will prevent premature deaths.
Methionine is a sulfur-containing amino acid essential to the human body and must be obtained from methionine- or methionine-containing proteins6). In addition to serving as a substrate for protein synthesis, methionine plays a major role in methylation reactions7, 8). Methionine is the direct metabolic precursor of S-adenosylmethionine (SAM) and this conversion is mediated by methionine adenosyltransferase6). SAM is a central one-carbon cycle metabolite that also serves as a methyl donor and plays a critical role in the metabolism and methylation of DNA and RNA9). S-adenosylhomocysteine (SAH) is formed when SAM donates a methyl group to a methyl acceptor by methyltransferases10) and is then converted to adenosine and homocysteine (Hcy) by SAH hydrolase through a hydrolysis reaction11). Hcy can be remethylated to methionine via the folic acid and betaine pathways or enter the sulfate assimilation and transsulfuration pathways (Fig.1)9, 11). Elevated plasma Hcy levels have been reported as an independent risk factor for vascular disease by many studies12-14). However, clinical intervention studies have shown that vitamin B6, B12 and folate could lower elevated Hcy levels, and there was no correlation between lowered Hcy levels and cardiovascular events15-17). Since the conversion of SAH to Hcy is reversible, an increase in Hcy concentration is usually accompanied by an increase in the SAH18). In an in-house cohort study, we found that plasma SAH levels were positively associated with the risk of CVD events19), which was consistent with the results of previous cross-sectional and case-control studies from other laboratories18, 20). To date, no studies have investigated the association between plasma SAH levels and the risk of all-cause or cardiovascular mortality. Combined with our previous studies, we hypothesized that plasma SAH levels are positively correlated with the risk of cardiovascular mortality. Therefore, in this prospective cohort study, we evaluated the relationship between plasma SAH levels and the risk of mortality in patients with CAD.
Met, methionine; MAT, methionine adenosyltransferase enzyme; SAM, S-adenosylmethionine; X, indicates DNA, RNA, histones or other proteins; MTFs, various methyltransferases; SAH, S-adenosylhomocysteine; SAHH, SAH hydrolase; Hcy, homocysteine; BHMT, betaine-Hcy S-methyltransferase; DMG, dimethylglycine; THF, tetrahydrofolate; MTHFR, methylene-THF reductase; MetS, Met synthase.
The cohort we used included patients with CAD; the details have been reported elsewhere19). Briefly, patients of 40–85 years of age were recruited from the cardiology departments of three major hospitals (General Hospital of Guangzhou Military Command of People’s Liberation Army, the First Affiliated Hospital, and the Second Affiliated Hospital of Sun Yat-sen University) in Guangzhou, South China, between October 2008 and December 2011. The exclusion criteria were surgery or trauma within the previous month, critical illness or hemodynamic instability other than acute coronary syndromes, known cancer, hepatic failure, hepatitis, or the use of drugs (e.g., anticancer agents), which would affect plasma methionine metabolism. Different from previous short-term follow-up periods, the median follow-up time for this study was 9.2 years (interquartile range: 8.5–10.2 years). Follow-up ended in December 2019 or at the death of the patient, whichever occurred first. In the early stages of follow-up, we sent follow-up letters to the participants and received response letters to evaluate the percentage of patients who were lost to follow-up. However, this resulted in an 80% response rate, with ＞20% showing no response. In the later follow-up, we called the patients by telephone or visited them indoors and tracked them through the Guangdong Provincial Centers for Disease Control and Prevention (GPCDC) system, and reduced the percentage of patients who were lost to follow-up to ＜5%. Among these subjects, approximately 2% emigrated abroad or could not be contacted for other reasons. We could not use the GPCDC system to track deaths among these subjects. A total of 1977 patients who underwent coronary angiography were diagnosed with CAD according to the WHO 1999/2000 guidelines. After excluding 389 participants with insufficient plasma samples and 35 participants with missing baseline data, a total of 1553 patients were included in the analysis of the present study (Supplementary Fig.1). The primary outcomes were the all-cause and cardiovascular mortality rates. The study was approved by the Ethics Committee of Sun Yat-sen University, China. Informed consent was obtained from the volunteers who participated in the study. This research was conducted in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans. Our cohort study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
CAD, coronary artery disease; GPCDC, Guangdong Provincial Centers for Disease Control and Prevention system; SAH, S-adenosylhomocysteine
Angiograms were interpreted by at least 2 independent cardiologists who were blinded to the patients’ risk factor profiles. The degree to which we evaluated CAD was based on the Gensini Score (GS)21, 22). For each coronary artery lesion, the GS was the sum of all lesion scores and involved 3 main parameters: the severity score, the region multiplying factor, and the collateral adjustment factor. According to the score, the GS reflects the relative severity of the lesion, impact on blood flow, usual blood flow to the left ventricle of each vessel or vessel segment, and functional significance of the area. CAD was defined at 3 levels: mild CAD (visible plaque resulting in ＞20% but ＜50% luminal narrowing stenosis), moderate CAD (≥ 50% to ＜70% stenosis), and obstructive CAD (≥ 50% stenosis in the left main coronary artery, ≥ 70% in any other coronary artery, or both). Patients with obstructive CAD were further categorized according to the number of diseased vessels (i.e., single, double, or triple-vessel disease). In brief, the extent of CAD was comprehensively evaluated using the GS in our study.
2.3 Baseline MeasurementsGeneral information and family medical history, such as family history of CAD, hypertension, and diabetes mellitus, were surveyed in detail before enrollment. A family history of CAD was defined as a diagnosis of CAD among immediate family members before 60 years of age. Diabetes was considered to be present if there was a fasting blood glucose concentration of ＞126 mg/dL, if there was a history of diabetes, or if the patient was taking antidiabetic medication. Hypertension was defined as the use of antihypertensive medication, diastolic blood pressure ＞90 mm Hg, and/or systolic blood pressure ＞140 mm Hg. The dietary intake of participants was collected using a paper-based semi-quantitative food frequency questionnaire (FFQ), which was used to assess the usual consumption of major nutrients and food groups. The FFQ, which includes 8 food groups with a total of 81 food items, was verified to have satisfactory reproducibility and reasonable validity in a previous study23). The frequency and portion-size data of the participants were calculated using the China Food Composition Table 2004 24). Medical records and complete data were reviewed by medical staff and trained interviewers. Anthropometric measurements were performed according to standard clinical test procedures. All baseline measurements and disease definitions have been detailed in previous studies19).
2.4 Biochemical AnalysesThe collection and preparation of fasting blood samples have been described elsewhere19). Routine blood chemistry, plasma lipid levels, and hematological investigations were performed using standard methods. Serum folate concentration25), plasma total cysteine (tCys), and total homocysteine (tHcy)26) were measured according to a previously reported method. All inter-assay and intra-assay coefficients of variation (CVs) were ≤ 8.6%.
2.5 Plasma SAM and SAH MeasurementsPlasma SAH and SAM are likely to degrade. However, this can be avoided by the immediate addition of appropriate amounts of acid after sample collection27). The samples were stored at −80℃ until use. Plasma SAH and SAM were estimated using high-performance liquid chromatography-mass spectrometry (Agilent Technologies Inc.)19, 28). The stable-isotope method was used for the quantitative analysis of SAH and SAM. Accuracy and precision were evaluated by injecting standards between the sample runs. The inter- and intra-assay CVs were 4.5% and 3.9%, respectively, for SAM and 4.2% and 3.6% for SAH. The limits of quantitation (LOQ) for SAM and SAH were at the nmol level. The concentrations of plasma SAH and SAM were not significantly different after storage at −80℃ for 1 day, 1 week, and 1 month.
2.6 Outcomes of Follow-upThe primary outcomes were the all-cause and cardiovascular mortality rates. Annual follow-up information, all-cause mortality, and cardiovascular mortality were collected and confirmed by a combination of medical records, telephone interviews with patients or family members, and deaths registered in the GPCDC. Collection continued until the end of December 2019 or until the death of the patient, whichever occurred first. The median follow-up time was 9.2 years (interquartile range: 8.5–10.2 years). Death certificates were coded by nosologists, according to the International Classification of Diseases. Cardiovascular mortality was defined as death attributable to an ischemic cardiovascular cause (including fatal MI, stroke, and peripheral arterial disease) or sudden death due to an unknown but presumed cardiovascular cause in high-risk patients, according to the International Classification of Diseases, Tenth Revision codes I00-I99.
2.7 Statistical AnalysesParticipants were divided into quartiles based on their plasma SAH levels. Continuous and normally distributed variables are presented as mean±standard deviation or expressed as median with interquartile range. Categorical variables are presented as frequencies and percentages (%). Differences in plasma SAH levels were assessed using 1-way ANOVA, Kruskal-Wallis H, or the χ2 test. Competing risk analyses were performed using the “cmprsk” package in R to generate cumulative incidence curves of cardiovascular and non-cardiovascular mortality according to SAH quartiles. Three models were created: a model was adjusted for age and sex; multivariable-adjusted model 1 was adjusted for age, sex, body mass index (BMI), physical activity, smoking status, alcohol consumption, family history of CAD, hypertension, diabetes mellitus, systolic blood pressure, GS, total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, fasting glucose, use or nonuse of statins, aspirin, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, β-blockers, SAM, tCys, tHcy, estimated glomerular filtration rate (eGFR), and high-sensitivity C-reactive protein (hsCRP); multivariable-adjusted model 2 was adjusted for folate, vitamin B12, and other variables in model 1. To display continuous dose-response relationships of SAH with mortality, we examined the Cox regression model using restricted cubic splines with 4 degrees of freedom. The Cox regression analysis with restricted cubic splines was performed using R.
All statistical analyses were performed using SPSS (version 25.0; IBM SPSS Inc., Chicago, IL, USA) or R (Version R-4.0.3). P values of ＜0.05 were considered to indicate statistical significance unless otherwise specified.
Among the 1977 patients, 79 patients who could not be contacted or tracked by the GPCDC system were excluded because there was insufficient plasma to measure their concentrations of SAH and SAM, resulting in incomplete baseline data. There were no significant differences in baseline characteristics between these patients and the remaining cohort (Supplementary Table 1). The baseline characteristics and risk factors of the 1553 participants are presented in Table 1. The average age of the overall population was 63.9±11.9 years, and 67.5% of the participants were male. Participants were categorized into 4 grades based on the quartiles of plasma SAH levels. The median (interquartile range) plasma SAH was 17.78 (10.74, 27.49) nmol/L. According to the quartiles of the plasma SAH level, participants in the third and fourth quartiles of SAH were more likely to have a higher presence of alcohol consumption, higher prevalence of hypertension, higher systolic blood pressure, and greater CAD burden (higher GS) than those in the first and second quartiles (P＜0.05). Relative to patients in the first and second quartiles of SAH, the plasma SAM/SAH ratios and eGFR levels were much lower among patients in the third and fourth quartiles of SAH (P＜0.001). As for the tHcy levels, those in the last quartile were higher than those in the other 3 quartiles of SAH (P＜0.001).
| Characteristics | Follow-up | Lost of follow-up | P value |
|---|---|---|---|
| n | 1553 | 79 | |
| Clinical characteristics | |||
| Age, y | 63.9±11.5 | 62.4±11.4 | 0.277 |
| Male, n (%) | 1049 (67.5) | 50 (63.3) | 0.431 |
| BMI, kg/m2 | 24.0±3.3 | 24.3±3.5 | 0.426 |
| Current smoking, n (%) | 569 (36.6) | 27 (34.2) | 0.658 |
| Alcohol drinker, n (%) | 403 (25.9) | 19 (24.1) | 0.707 |
| Hypertension, n (%) | 858 (55.2) | 39 (49.4) | 0.305 |
| Diabetes mellitus, n (%) | 493 (31.7) | 29 (36.7) | 0.356 |
| Family history of CAD, n (%) | 113 (7.3) | 6 (7.6) | 0.915 |
| Systolic blood pressure, mm Hg | 134.3±22.1 | 129.6±20.3 | 0.061 |
| Diastolic blood pressure, mm Hg | 76.5±12.8 | 77.2±11.9 | 0.637 |
| Extent of CAD | |||
| Gensini score | 24 (10, 48) | 25 (12, 41.5) | 0.645 |
| Statin, n (%) | 515 (33.2) | 26 (32.9) | 0.963 |
| Aspirin, n (%) | 531 (34.2) | 23 (29.1) | 0.352 |
| ACE inhibitor or ARB, n (%) | 495 (31.9) | 24 (30.4) | 0.781 |
| β blocker, n (%) | 663 (42.7) | 35 (44.3) | 0.778 |
| Laboratory measurements | |||
| Total cholesterol, mmol/L | 4.64±1.10 | 4.68±1.28 | 0.780 |
| Triglycerides, mmol/L | 1.54 (1.11, 2.21) | 1.55 (1.24, 2.12) | 0.332 |
| HDL cholesterol, mmol/L | 1.06±0.27 | 1.08±0.26 | 0.609 |
| LDL cholesterol, mmol/L | 2.97±0.99 | 2.83±1.02 | 0.223 |
| Fasting glucose, mmol/L | 6.60±2.79 | 6.70±2.44 | 0.755 |
| SAH, nmol/L | 17.8 (10.7, 27.6) | / | / |
| SAM, nmol/L | 93.0±13.2 | / | / |
| SAM/SAH ratio | 5.21 (3.28, 8.64) | / | / |
| tHcy, μmol/L | 14.5±5.13 | 14.2±4.32 | 0.716 |
| tCys, μmol/L | 254.0±63.9 | 259.5±35.2 | 0.442 |
| Folate, nmol/L | 18.4±3.8 | 18.7±3.6 | 0.478 |
| Vitamin B12, pmol/L | 300 (214, 354) | 292 (197, 347) | 0.240 |
| eGFR, mL min－1 1.73 m－2 | 78.4 (64.2, 95.1) | 78.1 (64.2, 96.6) | 0.357 |
| hsCRP, mg/L | 4.55 (1.8, 12.2) | 4.27 (2.15, 13.8) | 0.617 |
1Values expressed as mean±SD values, median (interquartile range) values, and percentage (%) unless otherwise stated. Significance tests for comparisons by quartiles of plasma SAH based on analysis of variance for continuous variables and Pearson χ2 test for categorical variables.
ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; hsCRP, high-sensitivity C-reactive protein; LDL, low-density lipoprotein; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; tCys, total cysteine; and tHcy, total homocysteine.
| Characteristics | Baseline Plasma SAH Quartiles | P value | |||
|---|---|---|---|---|---|
| Q1 (＜10.74 nmol/L) | Q2 (10.74-17.77 nmol/L) | Q3 (17.78-27.49 nmol/L) | Q4 (＞27.49 nmol/L) | ||
| n | 389 | 388 | 388 | 388 | |
| Clinical characteristics | |||||
| Age, y | 63.5±11.4 | 64.9±11.6 | 63.4±11.8 | 64.7±11.4 | 0.399 |
| Male, n (%) | 262 (67.4) | 275 (70.9) | 247 (63.7) | 265 (68.3) | 0.192 |
| BMI, kg/m2 | 24.4±3.3 | 23.8±3.3 | 23.9±3.2 | 24.1±3.5 | 0.047 |
| Current smoking, n (%) | 136 (35.0) | 155 (39.9) | 136 (35.1) | 142 (36.6) | 0.436 |
| Alcohol drinker, n (%) | 85 (21.9) | 94 (24.2) | 103 (26.5) | 121 (31.2) | 0.022 |
| Hypertension, n (%) | 161 (41.4) | 213 (54.9) | 231 (59.5) | 253 (65.2) | ＜0.001 |
| Diabetes mellitus, n (%) | 121 (31.1) | 125 (32.2) | 127 (32.7) | 120 (30.9) | 0.939 |
| Family history of CAD, n (%) | 29 (7.5) | 29 (7.5) | 30 (7.7) | 25 (6.4) | 0.906 |
| Physical activity ＜4.0 h/wk, n (%) | 237 (60.9) | 223 (57.5) | 234 (60.3) | 231 (59.5) | 0.779 |
| Systolic blood pressure, mm Hg | 130.9±20.1 | 132.4±20.6 | 135.4±21.9 | 138.6±24.7 | ＜0.001 |
| Diastolic blood pressure, mm Hg | 76.5±12.2 | 76.6±12.6 | 75.8±13.1 | 77.3±13.2 | 0.390 |
| Extent of CAD | |||||
| Gensini score | 18.5 (7.8, 42) | 24 (10, 42) | 25 (8.2, 48) | 28 (12.6, 52) | ＜0.001 |
| Medication usage | |||||
| Statin, n (%) | 132 (33.9) | 133 (34.3) | 125(32.2) | 125 (32.2) | 0.888 |
| Aspirin, n (%) | 140 (36.0) | 126 (32.5) | 128 (33.0) | 137 (35.3) | 0.675 |
| ACE inhibitor or ARB, n (%) | 122 (31.4) | 119 (30.7) | 135 (34.8) | 119 (30.7) | 0.554 |
| β blocker, n (%) | 151 (38.8) | 169 (43.6) | 184 (47.4) | 159 (41.0) | 0.089 |
| Laboratory measurements | |||||
| Total cholesterol, mmol/L | 4.65±1.16 | 4.61±1.06 | 4.65±1.07 | 4.66±1.11 | 0.945 |
| Triglycerides, mmol/L | 1.60 (1.15, 2.34) | 1.50 (1.14, 2.16) | 1.61 (1.13, 2.19) | 1.51 (1.05, 2.17) | 0.217 |
| HDL cholesterol, mmol/L | 1.05±0.26 | 1.06±0.25 | 1.08±0.30 | 1.07±0.28 | 0.404 |
| LDL cholesterol, mmol/L | 2.94±1.01 | 2.96±0.94 | 2.97±0.96 | 3.02±1.03 | 0.689 |
| Fasting glucose, mmol/L | 6.55±2.41 | 6.62±3.01 | 6.70±3.00 | 6.55±2.69 | 0.858 |
| SAM, nmol/L | 92.8±14.5 | 93.5±11.7 | 93.5±13.4 | 92.1±13.1 | 0.384 |
| SAM/SAH ratio | 12.7 (10.0, 20.2) | 6.49 (5.70, 7.58) | 4.16 (3.62, 4.76) | 2.16 (1.43, 2.76) | ＜0.001 |
| tHcy, μmol/L | 13.7±4.54 | 14.8±5.33 | 13.0±4.39 | 16.4±5.53 | ＜0.001 |
| tCys, μmol/L | 258.4±60.1 | 252.5±64.3 | 251.1±66.7 | 253.7±64.2 | 0.415 |
| Folate, nmol/L | 18.4±3.8 | 18.4±3.7 | 18.5±4.1 | 18.3±3.9 | 0.957 |
| Vitamin B12, pmol/L | 303.6 (226.7, 354.0) | 299.0 (214.8, 348.5) | 300.3 (214.7, 355.0) | 297.7 (210.4, 351.7) | 0.843 |
| eGFR, mL min－1 1.73 m－2 | 85.2 (72.7, 101.9) | 77.6 (64.3, 95.5) | 74.9 (61.6, 92.1) | 74.1 (58.7, 90.9) | ＜0.001 |
| hsCRP, mg/L | 4.45 (1.87, 12.51) | 4.83 (1.88, 12.60) | 4.50 (1.73, 11.53) | 4.41 (1.82, 12.00) | 0.763 |
1 Values are expressed as mean±SD deviation, median (interquartile range) values, and percentage (%) unless otherwise stated. Significance tests for comparisons by quartiles of plasma SAH were based on analysis of variance for continuous variables and Pearson χ2 test for categorical variables.
ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; hsCRP, high-sensitivity C-reactive protein; LDL, low-density lipoprotein; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; tCys, total cysteine; tHcy, total homocysteine.
The bivariate correlation between plasma SAH and clinical characteristics is shown in Supplementary Table 2. Higher plasma SAH levels were associated with higher alcohol consumption, prevalence of hypertension, systolic blood pressure, greater CAD burden (GS), and higher levels of tHcy in univariate analyses, as well as in the multivariate model (P＜0.05). However, the plasma SAH level was inversely associated with SAM and eGFR levels in the univariate analyses, whereas eGFR levels remained significant determinants of plasma SAH levels in the multivariate model (P＜0.001).
| Variables | Plasma SAH | |||
|---|---|---|---|---|
| Univariate | Multivariate | |||
| β | P value | β | P value | |
| Age | 0.035 | 0.172 | -0.008 | 0.745 |
| Male | 0.004 | 0.872 | 0.011 | 0.666 |
| BMI | -0.011 | 0.678 | -0.017 | 0.516 |
| Current smoking | -0.009 | 0.730 | -0.027 | 0.294 |
| Alcohol drinker | 0.052 | 0.042 | 0.058 | 0.023 |
| Hypertension | 0.084 | 0.001 | 0.082 | 0.001 |
| Diabetes mellitus | -0.033 | 0.194 | -0.038 | 0.133 |
| Systolic blood pressure | 0.104 | ＜0.001 | 0.090 | ＜0.001 |
| Diastolic blood pressure | 0.031 | 0.216 | -0.014 | 0.579 |
| Gensini score | 0.065 | 0.011 | 0.043 | 0.094 |
| Total cholesterol | -0.017 | 0.516 | -0.015 | 0.564 |
| Triglycerides | -0.036 | 0.157 | -0.029 | 0.249 |
| HDL cholesterol | 0.002 | 0.942 | -0.010 | 0.709 |
| LDL cholesterol | 0.003 | 0.913 | 0.016 | 0.520 |
| Fasting glucose | 0.031 | 0.225 | 0.031 | 0.232 |
| SAM | -0.062 | 0.014 | -0.032 | 0.218 |
| tHcy | 0.129 | ＜0.001 | 0.116 | ＜0.001 |
| tCys | -0.019 | 0.464 | -0.037 | 0.152 |
| Folate | -0.026 | 0.303 | -0.026 | 0.311 |
| Vitamin B12 | 0.012 | 0.643 | 0.026 | 0.301 |
| eGFR | -0.076 | 0.003 | -0.056 | 0.028 |
| hsCRP | 0.007 | 0.772 | -0.012 | 0.632 |
1Conventional risk factors and the independent determinants of SAH in the bivariate correlation analysis were included in the multivariate model. BMI, body mass index; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; hsCRP, high- sensitivity C-reactive protein; LDL, low-density lipoprotein; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; tCys, total cysteine; and tHcy, total homocysteine.
The median follow-up period was 9.2 (8.5-10.2) years, and there were 321 deaths (20.6%), including 227 deaths (14.6%) from CVD. Competing risk analyses showed significant positive associations between SAH levels and the risk of cardiovascular (P＜0.001) and non-cardiovascular mortality (P＜0.05) (Fig.2). In the unadjusted models, the hazard ratios (HRs) among the 4 quartiles were 1.00, 1.35, 1.96, and 2.22 for all-cause mortality (P for trend ＜0.001) and 1.00, 1.43, 1.68, and 2.25 for cardiovascular mortality (P for trend ＜0.001) (Table 2). For all-cause mortality and cardiovascular mortality, in age- and sex-adjusted and multivariable-adjusted models, the trends of HRs were consistent with those in the unadjusted model. Furthermore, relative to participants in the lowest quartile of plasma SAH levels, participants in the highest quartile of plasma SAH levels had a higher risk of all-cause death (adjusted HR, 2.15; 95% CI, 1.54-3.01; P＜0.001) and cardiovascular death (adjusted HR, 2.20; 95% CI, 1.49-3.25; P=0.001) in the age- and sex-adjusted model, which is similar to the multivariable adjusted model for all-cause death (adjusted HR, 1.81; 95% CI, 1.27-2.58; P=0.002) and cardiovascular death (adjusted HR, 1.84; 95% CI, 1.21-2.79; P=0.031) (Table 2). Continuous SAH values were evaluated using Cox proportional hazard regression models. The age- and sex-adjusted HRs for each 1 SD increase in plasma SAH level were 1.30 (95% CI, 1.22-1.38) for all-cause mortality, and 1.34 (95% CI, 1.25-1.43) for cardiovascular mortality, respectively (Table 2). A 1 SD increase in the SAH level was associated with a 25% higher risk of total death (adjusted HR, 1.25; 95% CI, 1.17-1.34) and a 29% greater risk of cardiovascular death (adjusted HR, 1.29; 95% CI, 1.20-1.39) in the multivariable adjusted analysis (Table 2). The dose-response association between plasma SAH and all-cause mortality (Fig.3A) or cardiovascular mortality (Fig.3B) was further demonstrated using restricted cubic splines.
SAH, S-adenosylhomocysteine.
| Plasma SAH | P for Trend |
SAH as a Continuous Variable (1 SD Increase) |
||||
|---|---|---|---|---|---|---|
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | |||
| All-cause mortality | ||||||
| No. of deaths/person-years | 51/3434 | 68/3368 | 95/3243 | 107/3242 | ||
| Unadjusted HR (95% CI) | 1 | 1.35 (0.94, 1.95) | 1.96 (1.40, 2.76) | 2.22 (1.59, 3.09) | ＜0.001 | 1.31 (1.23, 1.39) |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.34 (0.93, 1.93) | 1.99 (1.41, 2.80) | 2.15 (1.54, 3.01) | ＜0.001 | 1.30 (1.22, 1.38) |
| Multivariable-adjusted HR of model 1 (95% CI)2 | 1 | 1.26 (0.87, 1.84) | 1.91 (1.34, 2.71) | 1.85 (1.31, 2.63) | ＜0.001 | 1.26 (1.18, 1.35) |
| Multivariable-adjusted HR of model 2 (95% CI)3 | 1 | 1.25 (0.86, 1.81) | 1.77 (1.24, 2.53) | 1.81 (1.27, 2.58) | 0.002 | 1.25 (1.17, 1.34) |
| Cardiovascular mortality | ||||||
| No. of deaths/person-years | 37/3434 | 52/3368 | 59/3243 | 79/3242 | ||
| Unadjusted HR (95% CI) | 1 | 1.43 (0.93, 2.18) | 1.68 (1.11, 2.54) | 2.25 (1.52, 3.33) | ＜0.001 | 1.35 (1.25, 1.44) |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.41 (0.92, 2.15) | 1.71 (1.13, 2.58) | 2.20 (1.49, 3.25) | 0.001 | 1.34 (1.25, 1.43) |
| Multivariable-adjusted HR of model 1 (95% CI)2 | 1 | 1.38 (0.89, 2.13) | 1.67 (1.09, 2.55) | 1.92 (1.27, 2.91) | 0.013 | 1.30 (1.21, 1.40) |
| Multivariable-adjusted HR of model 2 (95% CI)3 | 1 | 1.35 (0.88, 2.09) | 1.57 (1.02, 2.41) | 1.84 (1.21, 2.79) | 0.031 | 1.29 (1.20, 1.39) |
1 HRs and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression models. CAD, coronary artery disease; SAH, S-adenosylhomocysteine.
2 Adjusted for age, sex, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, systolic blood pressure, Gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose,triglycerides, use or nonuse of statins, aspirin, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, β-blockers, S-adenosylmethionine, total cysteine, total homocysteine, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
3 Adjusted for folate, vitamin B12, and other variables in model 2.
HRs and 95% CIs were estimated using Cox proportional hazards regression models (n=1553). SAH, S-adenosylhomocysteine.
As shown in Supplementary Table 3, there was no significant heterogeneity in the association between SAH and all-cause mortality when patients were stratified by age, sex, BMI, alcohol consumption, hypertension, tHcy, folate, and eGFR in the multivariable analyses (P for interaction ＞0.05). The same was true for the association between SAH and cardiovascular mortality.
| Subpopulation |
No. of Death/ No. of Participants |
SAas a Continuous Variable (1 SD Increase) |
P value | P for Interaction |
|---|---|---|---|---|
| All-cause mortality | ||||
| Age groups, y | 0.594 | |||
| ≤ 60 | 84/539 | 1.38 (1.17, 1.63) | 0.001 | |
| ＞60 | 237/1014 | 1.23 (1.14, 1.33) | ＜0.001 | |
| Sex groups | 0.327 | |||
| Men | 221/1049 | 1.21 (1.11, 1.32) | ＜0.001 | |
| Women | 100/504 | 1.28 (1.13, 1.45) | ＜0.001 | |
| BMI, kg/m2 | 0.520 | |||
| ＜24 | 178/794 | 1.26 (1.14, 1.39) | ＜0.001 | |
| ≥ 24 | 143/759 | 1.23 (1.12, 1.36) | ＜0.001 | |
| Alcohol drinker | 0.734 | |||
| Yes | 79/398 | 1.33 (1.14, 1.56) | ＜0.001 | |
| No | 242/1155 | 1.24 (1.15, 1.35) | ＜0.001 | |
| Hypertension | 0.723 | |||
| Yes | 196/921 | 1.24 (1.13, 1.36) | ＜0.001 | |
| No | 125/632 | 1.26 (1.13, 1.41) | ＜0.001 | |
| tHcy, μmol/L | 0.415 | |||
| ＜13.64 | 155/776 | 1.31 (1.16, 1.47) | ＜0.001 | |
| ≥ 13.64 | 166/777 | 1.25 (1.14, 1.36) | ＜0.001 | |
| Folate, nmol/L | 0.623 | |||
| ＜17.71 | 167/776 | 1.28 (1.16, 1.42) | ＜0.001 | |
| ≥ 17.71 | 154/777 | 1.25 (1.13, 1.38) | ＜0.001 | |
| eGFR, mL min－1 1.73 m－2 | 0.768 | |||
| ＜60 | 85/283 | 1.19 (1.03, 1.37) | 0.012 | |
| ≥ 60 | 236/1270 | 1.26 (1.16, 1.38) | ＜0.001 | |
| Cardiovascular mortality | ||||
| Age groups, y | 0.816 | |||
| ≤ 60 | 65/539 | 1.38 (1.14, 1.67) | 0.001 | |
| ＞60 | 162/1014 | 1.28 (1.18, 1.40) | ＜0.001 | |
| Sex groups | 0.752 | |||
| Men | 161/1049 | 1.28 (1.17, 1.40) | ＜0.001 | |
| Women | 66/504 | 1.21 (1.03, 1.42) | 0.018 | |
| BMI, kg/m2 | 0.554 | |||
| ＜24 | 123/794 | 1.31 (1.17, 1.45) | ＜0.001 | |
| ≥ 24 | 104/759 | 1.25 (1.12, 1.39) | ＜0.001 | |
| Alcohol drinker | 0.666 | |||
| Yes | 63/398 | 1.39 (1.16, 1.66) | ＜0.001 | |
| No | 164/1155 | 1.29 (1.18, 1.41) | ＜0.001 | |
| Hypertension | 0.922 | |||
| Yes | 135/921 | 1.28 (1.16, 1.42) | ＜0.001 | |
| No | 92/632 | 1.27 (1.13, 1.45) | ＜0.001 | |
| tHcy, μmol/L | 0.284 | |||
| ＜13.64 | 110/776 | 1.35 (1.19, 1.54) | ＜0.001 | |
| ≥ 13.64 | 117/777 | 1.28 (1.15, 1.42) | ＜0.001 | |
| Folate, nmol/L | 0.580 | |||
| ＜17.71 | 118/776 | 1.35 (1.20, 1.51) | ＜0.001 | |
| ≥ 17.71 | 109/777 | 1.28 (1.14, 1.43) | ＜0.001 | |
| eGFR, mL min－1 1.73 m－2 | 0.778 | |||
| ＜60 | 67/283 | 1.27 (1.08, 1.49) | 0.003 | |
| ≥ 60 | 160/1270 | 1.31 (1.18, 1.43) | ＜0.001 | |
1HRs and 95% CIs were estimated by Cox proportional hazards regression models. BMI, body mass index; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; SAH, S-adenosylhomocysteine; tHcy, total homocysteine.
Adjusted for age, sex, body mass index, smoking status, alcohol drinker, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, and β-blockers, S-adenosylmethionine, total cysteine; total homocysteine, folate, vitamin B-12, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
We further analyzed the association between SAH and the risk of all-cause and cardiovascular mortality using a propensity score-matched analysis. The number of patients in the low SAH (＜17.78 nmol) group was equal to that in the high SAH (＞17.78 nmol) group, and the number of patients at both levels was 647. High SAH was associated with a higher risk of all-cause mortality in both age- and sex-adjusted models (HR, 1.95; 95% CI, 1.51-2.53; P＜0.001) and the multivariable adjusted analysis (HR, 1.85; 95% CI, 1.42-2.41; P＜0.001). Similarly, high SAH was associated with an increased risk of cardiovascular mortality in both age- and sex-adjusted models (HR, 1.87; 95% CI, 1.37-2.55; P＜0.001) and the multivariable adjusted analysis (HR, 1.72; 95% CI, 1.26-2.36; P=0.001) (Supplementary Table 4). In addition, a stratified analysis based on the baseline severity of CAD showed that plasma SAH was significantly associated with all-cause death and cardiovascular death in mild, moderate, and severe patients in both age- and sex-adjusted models and the multivariable adjusted analysis (Supplementary Table 5). Additionally, after excluding 283 patients with eGFR ＜60 mL min-1 1.73 m-2, the associations between SAH and all-cause and cardiovascular mortality remained significant in the multivariable-adjusted model (Supplementary Table 6).
| Low SAH (＜17.78 nmol/L) | High SAH (≥ 17.78 nmol/L) | P value | |
|---|---|---|---|
| All-cause mortality | |||
| No. of deaths/ Participants | 89/647 | 165/647 | |
| Unadjusted HR (95% CI) | 1 | 1.96 (1.51, 2.53) | ＜0.001 |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.95 (1.51, 2.53) | ＜0.001 |
| Multivariable-adjusted HR (95% CI)2 | 1 | 1.85 (1.42, 2.41) | ＜0.001 |
| Cardiovascular mortality | |||
| No. of deaths/Participants | 63/647 | 112/647 | |
| Unadjusted HR (95% CI) | 1 | 1.88 (1.38, 2.55) | ＜0.001 |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.87 (1.37, 2.55) | ＜0.001 |
| Multivariable-adjusted HR (95% CI)2 | 1 | 1.72 (1.26, 2.36) | 0.001 |
1HRs and 95% CIs were estimated by Cox proportional hazards regression models. SAH, S-adenosylhomocysteine.
2Adjusted for age, sex, body mass index, smoking status, alcohol drinker, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, and β-blockers, S-adenosylmethionine, total cysteine; total homocysteine, folate, vitamin B12, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
| Severity of CAD (Gensini score) | |||
|---|---|---|---|
| Mild (＜13) | Moderate (13-39) | Severe (≥ 40) | |
| All-cause mortality | |||
| No. of deaths/ Participants | 75/504 | 125/543 | 121/506 |
| Unadjusted HR (95% CI) | 1.32 (1.17, 1.48), P＜0.001 | 1.28 (1.16, 1.41), P＜0.001 | 1.36 (1.19, 1.54), P＜0.001 |
| Age- and sex- adjusted HR (95% CI) | 1.32 (1.17, 1.48), P＜0.001 | 1.25 (1.14, 1.38), P＜0.001 | 1.36 (1.20, 1.55), P＜0.001 |
| Multivariable-adjusted HR (95% CI)2 | 1.24 (1.07, 1.43), P = 0.003 | 1.24 (1.11, 1.38), P＜0.001 | 1.31 (1.14, 1.50), P＜0.001 |
| Cardiovascular mortality | |||
| No. of deaths/Participants | 44/504 | 93/543 | 90/506 |
| Unadjusted HR (95% CI) | 1.38 (1.21, 1.57), P＜0.001 | 1.28 (1.15, 1.43), P＜0.001 | 1.42 (1.24, 1.63), P＜0.001 |
| Age- and sex- adjusted HR (95% CI) | 1.37 (1.20, 1.55), P＜0.001 | 1.26 (1.13, 1.41), P＜0.001 | 1.43 (1.25, 1.63), P＜0.001 |
| Multivariable-adjusted HR (95% CI)2 | 1.21 (1.02, 1.42), P = 0.023 | 1.30 (1.14, 1.47), P＜0.001 | 1.35 (1.17, 1.56), P＜0.001 |
1HRs and 95% CIs were estimated by Cox proportional hazards regression models. CAD, coronary artery disease; SAH, S-adenosylhomocysteine. 2Adjusted for age, sex, body mass index, smoking status, alcohol drinker, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, and β-blockers, S-adenosylmethionine, total cysteine; total homocysteine, folate, vitamin B12, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
| Plasma SAH | P for Trend | ||||
|---|---|---|---|---|---|
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | ||
| All-cause mortality | |||||
| No. of deaths/ Participants | 47/371 | 51/311 | 67/307 | 65/281 | |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.29 (0.87, 1.93) | 1.81 (1.25, 2.63) | 1.82 (1.25, 2.65) | 0.003 |
| Multivariable-adjusted HR (95% CI)2 | 1 | 1.28 (0.86, 1.92) | 1.65 (1.11, 2.44) | 1.76 (1.19, 2.59) | 0.019 |
| Cardiovascular mortality | |||||
| No. of deaths/ Participants | 33/371 | 39/311 | 44/307 | 44/281 | |
| Age- and sex- adjusted HR (95% CI) | 1 | 1.41 (0.88, 2.23) | 1.70 (1.09, 2.68) | 1.75 (1.12, 2.75) | 0.015 |
| Multivariable-adjusted HR (95% CI) 2 | 1 | 1.42 (0.88, 2.28) | 1.61 (1.00, 2.58) | 1.69 (1.06, 2.71) | 0.045 |
1HRs and 95% CIs were estimated by Cox proportional hazards regression models. CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; SAH, S-adenosylhomocysteine.
2Adjusted for age, sex, body mass index, smoking status, alcohol drinker, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, and β-blockers, S-adenosylmethionine, total cysteine; total homocysteine, folate, vitamin B12, and high-sensitivity C-reactive protein.
We analyzed the combined effects of plasma SAH and tHcy or tCys on the risk of mortality. First, we analyzed the combination of SAH and tHcy levels. Taking patients with low SAH and low tHcy as a reference, it was observed that when SAH increased, patients with either low tHcy or high tHcy had a higher risk of all-cause and cardiovascular death in both unadjusted and adjusted models (P＜0.001). However, this association was not observed in patients with low SAH and high tHcy levels in unadjusted or adjusted models. Second, we analyzed the mortality risk for different SAH and tCys levels. Relative to participants with low SAH and low tCys, those with high SAH or high tCys had a higher risk of total mortality and cardiovascular death (P＜0.05). Participants with high SAH and high tCys had a 123% higher risk of total death (fully adjusted HR, 2.23; 95% CI, 1.56-3.18; P＜0.001) and a 119% greater risk of cardiovascular death (multivariate adjusted HR, 2.19; 95% CI, 1.42-3.38; P＜0.05), which was higher than the other combination of SAH and tCys (Table 3). Additionally, we analyzed whether the association between SAH and SAM and mortality was mediated by tHcy and other one-carbon metabolism markers. Adjustment of tHcy, tCys, folate, and vitamin B12 levels had no significant effect on the association between SAH and SAM and mortality (Supplementary Table 7).
| No. of Death/No. of Participants | Unadjusted HR (95% CI) | Age- and sex- adjusted HR (95% CI) | Multivariable-adjusted HR (95% CI) 2 | |
|---|---|---|---|---|
| All-cause mortality | ||||
| Low SAH and low tHcy | 54/389 | 1 | 1 | 1 |
| Low SAH and high tHcy | 65/387 | 1.30 (0.91, 1.87) | 1.32 (0.93, 1.91) | 1.19 (0.82, 1.75) |
| High SAH and low tHcy | 101/390 | 1.99 (1.43, 2.78) | 2.02 (1.45, 2.81) | 1.66 (1.17, 2.35) |
| High SAH and high tHcy | 101/387 | 2.07 (1.48, 2.88) | 2.07 (1.48, 2.88) | 1.82 (1.29, 2.57) |
| P for Trend | ＜0.001 | ＜0.001 | 0.001 | |
| Low SAH and low tCys | 45/379 | 1 | 1 | 1 |
| Low SAH and high tCys | 74/397 | 1.67 (1.15, 2.42) | 1.69 (1.17, 2.45) | 1.66 (1.14, 2.42) |
| High SAH and low tCys | 98/380 | 2.35 (1.65, 3.35) | 2.36 (1.66, 3.36) | 1.97 (1.36, 2.84) |
| High SAH and high tCys | 104/397 | 2.37 (1.67, 3.36) | 2.38 (1.67, 3.38) | 2.23 (1.56, 3.18) |
| P for Trend | ＜0.001 | ＜0.001 | ＜0.001 | |
| Cardiovascular mortality | ||||
| Low SAH and low tHcy | 40/389 | 1 | 1 | 1 |
| Low SAH and high tHcy | 49/387 | 1.32 (0.87, 2.01) | 1.34 (0.88, 2.03) | 1.12 (0.72, 1.75) |
| High SAH and low tHcy | 68/390 | 1.86 (1.26, 2.75) | 1.88 (1.27, 2.78) | 1.65 (1.10, 2.47) |
| High SAH and high tHcy | 70/387 | 1.87 (1.27, 2.77) | 1.89 (1.28, 2.79) | 1.44 (0.95, 2.18) |
| P for Trend | 0.003 | 0.003 | 0.056 | |
| Low SAH and low tCys | 31/379 | 1 | 1 | 1 |
| Low SAH and high tCys | 58/397 | 1.90 (1.23, 2.94) | 1.91 (1.23, 2.96) | 1.88 (1.21, 2.93) |
| High SAH and low tCys | 69/397 | 2.28 (1.49, 3.48) | 2.28 (1.49, 3.49) | 1.97 (1.27, 3.06) |
| High SAH and high tCys | 69/380 | 2.41 (1.57, 3.68) | 2.43 (1.59, 3.71) | 2.19 (1.42, 3.38) |
| P for Trend | ＜0.001 | ＜0.001 | 0.004 | |
1HRs and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression models. CAD, coronary artery disease; SAH, S-adenosylhomocysteine; tHcy, total homocysteine; tCys, total cysteine.
2Adjusted for age, sex, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, sex, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, β-blockers, folate, vitamin B12, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
| SAH | SAM | SAM/SAH ratios | SAH | SAM | SAM/SAH ratios | |
|---|---|---|---|---|---|---|
| Q4 vs Q1 (P value) | Q1 vs Q4 (P value) | Q1 vs Q4 (P value) | 1-SD Increase (P value) | 1-SD Decrease (P value) | 1-SD Decrease (P value) | |
| All-cause mortality | ||||||
| Multivariable Model2 | 0.552±0.179 (0.002) | 0.487±0.168 (0.004) | 0.711±0.178 (＜0.001) | 0.242±0.034 (＜0.001) | 0.358±0.073 (＜0.001) | 0.463±0.117 (＜0.001) |
| Multivariable Model + tHcy | 0.529±0.181 (0.003) | 0.478±0.169 (0.005) | 0.702±0.180 (＜0.001) | 0.241±0.035 (＜0.001) | 0.354±0.073 (＜0.001) | 0.456±0.117 (＜0.001) |
| Multivariable Model + tCys | 0.551±0.179 (0.002) | 0.494±0.168 (0.003) | 0.711±0.178 (＜0.001) | 0.245±0.034 (＜0.001) | 0.363±0.073 (＜0.001) | 0.462±0.117 (＜0.001) |
| Multivariable Model + folate | 0.551±0.179 (0.002) | 0.490±0.168 (0.004) | 0.713±0.178 (＜0.001) | 0.240±0.034 (＜0.001) | 0.360±0.073 (＜0.001) | 0.463±0.116 (＜0.001) |
| Multivariable Model + VB12 | 0.553±0.179 (0.002) | 0.488±0.168 (0.004) | 0.713±0.178 (＜0.001) | 0.242±0.034 (＜0.001) | 0.358±0.073 (＜0.001) | 0.463±0.117 (＜0.001) |
| Cardiovascular mortality | ||||||
| Multivariable Model2 | 0.574±0.210 (0.006) | 0.819±0.213 (＜0.001) | 0.721±0.208 (0.001) | 0.276±0.038 (＜0.001) | 0.515±0.093 (＜0.001) | 0.394±0.130 (0.002) |
| Multivariable Model + tHcy | 0.553±0.213 (0.009) | 0.808±0.214 (＜0.001) | 0.711±0.210 (0.001) | 0.274±0.038 (＜0.001) | 0.512±0.093 (＜0.001) | 0.384±0.129 (0.003) |
| Multivariable Model + tCys | 0.572±0.210 (0.007) | 0.829±0.213 (＜0.001) | 0.718±0.208 (0.001) | 0.279±0.038 (＜0.001) | 0.525±0.093 (＜0.001) | 0.392±0.130 (0.003) |
| Multivariable Model + folate | 0.575±0.210 (0.006) | 0.823±0.213 (＜0.001) | 0.726±0.208 (＜0.001) | 0.274±0.038 (＜0.001) | 0.520±0.093 (＜0.001) | 0.394±0.130 (0.002) |
| Multivariable Model + VB12 | 0.576±0.210 (0.006) | 0.819±0.213 (＜0.001) | 0.722±0.208 (0.001) | 0.276±0.038 (＜0.001) | 0.516±0.093 (＜0.001) | 0.393±0.130 (0.002) |
1Results are presented as β±SE estimated by Cox proportional hazards regression models. CAD, coronary artery disease; SAH, S-adenosylhomocysteine, SAM, S-adenosylmethionine.
2Adjusted for age, sex, body mass index, smoking status, alcohol drinker, hypertension, diabetes mellitus, physical activity, family history of coronary artery disease, gensini score, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, triglycerides, use or nonuse of statins, aspirin, and β-blockers, estimated glomerular filtration rate, and high-sensitivity C-reactive protein.
In this large-scale prospective cohort study, we analyzed the association between plasma SAH and the risk of all-cause and cardiovascular mortality, which is the first large-scale cohort study of SAH and the risk of mortality in patients with CAD to date. In our analysis, we found that the increased risks of all-cause and cardiovascular mortality were associated with higher plasma SAH levels. After adjusting for age and sex, there was still a positive correlation between plasma SAH levels and the risks of all-cause and cardiovascular mortality. The plasma SAH level was associated with baseline characteristics such as alcohol consumption, higher prevalence of hypertension, higher systolic blood pressure, greater CAD burden (higher GS), plasma SAM/SAH ratio, eGFR, and tHcy. After adjusting for the above factors, conventional CVD risk factors, and other potential confounders29), there was still a significant association between plasma SAH levels and the risk of all-cause and cardiovascular mortality.
Methionine is essential to the human body and plays a major role in the methylation reaction and intervenes in lipid metabolism8). When methionine enters the ‘methionine cycle,’ it can be metabolized into SAM and then converted to SAH, Hcy, cysteine, and so on by a series of enzymes30). SAM is the principal methyl donor that generates SAH while donating its methyl group to the acceptor molecules31). Methylation plays an important role in regulating gene expression, protein function, and RNA processing. Several studies have shown that methylation status is associated with CVD. The methylation status of blood DNA at CpG sites is related to acute coronary syndrome32). Previous studies indicate that changes in DNA methylation states are conducive to the regulation of the biological processes of atherosclerosis33), hypertension34), and inflammation35). Methyltransferases catalyze a variety of methylation reactions, and different methyltransferases show different sensitivities to SAH36). Therefore, cells must maintain a low SAH concentration to ensure normal methylation. Previous studies have demonstrated that high levels of SAH affect the cellular methylation status37-40). SAH is a strong inhibitor of DNA methyltransferase 1 (DNMT1), which is essential for DNA methylation. High levels of SAH inhibit the expression of DNMT1, inducing hypomethylation in the p66shc gene promoter and transcriptional activation of the p66shc gene, and then p66shc promotes oxidative stress and vascular endothelial dysfunction41). In addition, elevated SAH activates oxidative stress, which activates the ERK signaling pathway, induces vascular smooth muscle cell proliferation and migration, and promotes the early formation of atherosclerosis42). SAH induces macrophage apoptosis by inhibiting histone methyltransferase expression, inhibiting histone methylation, and activating endoplasmic reticulum stress to accelerate atherosclerosis43). SAH is positively correlated with coronary artery calcification and promotes Runx2-dependent atherosclerotic calcification44).
The concentration of Hcy in plasma is at the micromolar level, while the concentration of SAH and SAM is at the nanomolar level, which makes the detection of SAH and SAM a difficult problem45). In addition, SAH and SAM exhibit poor stability under physiological conditions and are easily degraded during processing46). These factors limit research on SAH and SAM in CVD. With the development of detection technology and the optimization of sample processing, it is possible to accurately quantify SAH and SAM21, 47). In recent comprehensive studies, SAH was a more sensitive indicator of CVD20). In a case-control study, Huang et al. found that coronary artery lesions were associated with SAH in 160 patients who underwent coronary angiography, and the concentration of plasma SAH was independently related to coronary lesions48). In another case-control study, elevated plasma SAH was shown to be a more sensitive risk predictor of cerebral venous thrombosis (CVT)49). Stojan et al. showed that plasma SAH was positively associated with calcified plaque, non-calcified plaque, and total coronary plaque in patients with SLE50). In our median follow-up period of a 3.0-year cohort study, we found a positive correlation between plasma SAH and the risk of CVD events19). Although the relationship between SAH and cardiovascular risk has been verified by many studies, there have been no studies on the relationship between SAH and cardiovascular mortality risk. In this study, we followed 1553 participants with CAD for a median of 9.2 years. This was a larger sample size and longer follow-up time in comparison to previous studies. Among the 1553 patients, 321 died, including 227 who died of CVD. We observed a significant association between the risk of plasma SAH and all-cause and cardiovascular mortality using Cox proportional hazards regression models. The positive relationship was identical in the propensity score-matched and stratified analyses.
We further evaluated the relationship between plasma SAH and all-cause and cardiovascular mortality among different subgroups according to baseline characteristics. Although SAH has more advantages than Hcy in predicting CVD, there is still a relationship between tHcy and CVD. Previous studies have shown that plasma tHcy and tCys levels are related to mortality51-53). Additionally, cysteine is a controversial factor in risk prediction for CVD. Several studies have indicated that cysteine has a U-shaped relationship with peripheral vascular and cerebrovascular diseases51) and is a better predictor of CAD than conventional Hcy54). However, other studies found that cysteine was not an independent risk factor for CAD55, 56). Considering the possibility of using tHcy and tCys in the prediction of CVD, we evaluated the relationship between mortality and different combinations of SAH and tHcy or tCys. The combination of plasma SAH and tCys could predict CAD better than the combination of plasma SAH and tHcy.
The present study was associated with some limitations. First, only patients of Chinese origin were included in this study. Thus, it is not clear whether the results can be applied to other ethnicities. Second, we did not consider the influence of other factors, such as dietary patterns, age, and time point of detection, on the plasma concentration of SAH. However, we found that dietary betaine intake was inversely associated with plasma SAH levels, suggesting that betaine supplementation may be a potential strategy for lowering plasma SAH, which is consistent with a previous study reporting that betaine could reduce intracellular SAH levels57). Given that folate, vitamin B6 and B12 were not associated with plasma SAH levels58), vitamin B6 and B12 supplementation could not lower plasma SAH levels59) or reduce the cardiovascular risk, despite the plasma tHcy level being lowered15-17). This evidence suggests that lowering plasma SAH by other strategies, such as betaine supplementation, might be used to guide the treatment of patients with CAD. Finally, all-cause and cardiovascular mortality may be affected by other confounding factors. For example, renal dysfunction is an independent risk factor for CVD, patients with renal insufficiency or end-stage renal disease also have dramatically increased plasma SAH levels60, 61). Hence, the positive association between SAH and the risk of mortality may be overestimated due to potential residual confounding effects of renal dysfunction. Large-scale prospective cohort studies are needed to demonstrate the relationship between SAH and mortality in patients with renal insufficiency.
In summary, we found that plasma SAH levels were positively correlated with the risk of all-cause and cardiovascular mortality in patients with CAD. In future studies, a more comprehensive consideration should be given to the influence of other complex factors on mortality. The results should be validated in various ethnic groups to determine whether SAH can be regarded as a risk factor for mortality in CAD.
YX, SL, YW, MY, XD, and HH: designed essential reagents and materials; YX, HS, PL, and XD: collected the data; YX, CZ, YW, MY, XD, and RL provided essential reagents and materials; YX, HS, PL, and YC analyzed the data; YX, CZ, and HH wrote the manuscript; and all authors read and approved the final manuscript. None of the authors have any conflicts of interest to declare.
This work was supported by grants from the National Natural Science Foundation of China (82373553, 82073530, and 82404263), Guangdong Basic and Applied Basic Research Foundation (No.2022B1515020108), Guangdong Provincial Key Laboratory of Digestive Cancer Research (No.2021B1212040006), Shenzhen Medical Research Fund (B2403003), and Shenzhen Science and Technology Plan Project (JCYJ20220530145015035).
None of the authors have any conflicts of interest to declare.
We are particularly grateful to all patients and volunteers for participating in the present study and for their kind assistance in collecting data and samples.
The data described in the manuscript, code book, and analytic code will be made available upon request, pending application, approval, and payment.
