Homocysteine (Hcy), a sulphur-containing amino acid, is a metabolite of methionine(Reference Finkelstein1). Over the past 10 years, Hcy has been described as a well-established risk factor for the development of arteriosclerotic vascular diseases(Reference Boushey, Beresford and Omenn2). At present, the generally accepted reference range for Hcy level is 5–15 µmol/l. Hcy concentrations greater than 15 µmol/l are considered indicative of hyperhomocysteinaemia (HHcy), which occurs in 5–7 % of the general population(Reference Ueland and Refsum3,Reference Mccully4) . The association between total plasma homocysteine (tHcy) and the risk of some disease outcomes has been studied. A growing body of evidence indicates that elevated tHcy is associated with an increased risk of CVD(Reference Boushey, Beresford and Omenn2,Reference Martí-Carvajal, Solà and Lathyris5,Reference Perry, Refsum and Morris6,Reference Graham, Daly and Refsum7,Reference Hankey and Eikelboom8) . After adjustments were made for known cardiovascular risk factors, one meta-analysis of twelve prospective studies suggested that a 25 % reduction in Hcy levels can reduce the risk of ischaemic heart disease by 11 % and the risk of stroke by 19 %(9). Another recent meta-analysis showed that participants with higher Hcy levels had a 58 % increased risk of stroke and 55 % increased risk of ischaemic stroke, compared with those with lower Hcy levels(Reference Wu, Zhou and Chen10). In addition, elevated tHcy levels are related to several age-related diseases, such as essential hypertension(Reference Zhong, Zhuang and Wang11), Parkinson’s disease(Reference Christine, Auinger and Joslin12), Alzheimer’s disease(Reference Morris13), diabetes(Reference Huang, Ren and Huang14) and osteoporosis(Reference Verhoef and de Groot15).
The relationship between tHcy levels and the risk of all-cause mortality has also been analysed. Multiple studies have indicated that HHcy is associated with a higher risk of mortality in populations with specific chronic diseases, such as coronary artery disease(Reference Acevedo, Pearce and Jacobsen16), type 2 diabetes(Reference Hoogeveen, Kostense and Jakobs17) and renal failure(Reference Buccianti, Baragetti and Bamonti18), as well as in renal transplant recipients(Reference Connolly, Cunningham and McNamee19) and frail individuals(Reference Wong, Almeida and McCaul20). However, it should be noted that these populations already have a higher overall risk of mortality due to their underlying disease conditions. Furthermore, large-scale epidemiological data on the relationship between tHcy levels and the risk of all-cause mortality are still lacking, especially in the general population.
Considering the results of previous research, some pertinent questions remain unanswered. First, does a linear or non-linear relationship exist between tHcy and the risk of mortality? Second, is the relationship between tHcy and mortality based on the independent role of tHcy or on a product of important confounding factors such as vitamin B12, folate, renal function or disease susceptibilities? Thus, we conducted this study to investigate the association between HHcy and the risk of all-cause and cause-specific mortality in a large, population-based sample of adults in the USA.
Methods
Study design
In this study, we analysed data extracted from the National Health and Nutrition Examination Survey (NHANES, 1999–2002), a complex, stratified, multistage probability survey conducted by the Centers for Disease Control and Prevention (CDC). The NHANES programme began in the early 1960s and collected data on the demographics, lifestyle and health status of the US population using questionnaires. Biomarker data of the participants were also collected. This annual cross-sectional survey examines a nationally representative sample of about 5000 persons from counties across the country. Since NHANES 1999–2000, data for public use have been released almost every 2 years. We pooled data from two 2-year survey cycles of NHANES (1999–2000 and 2001–2002) for this study. We conducted a prospective secondary analysis by linking NHANES data with mortality data from the National Death Index. The National Center for Health Statistics (NCHS) Ethics Review Board approved the NHANES programme and released its documents for public use. Written informed consent was obtained from each participant. More information regarding the methodological details of the NHANES is available on the NHANES official website (www.cdc.gov/nchs/nhanes/). We obtained the NHANES datasets from DataDryad (https://doi.org/10.5061/dryad.d5h62).
Study population
We conducted a secondary analysis of data from two 2-year NHANES survey cycles: 1999–2000 and 2001–2002. Given the possibility of survival bias among old adults, the participants were limited to adults aged 18–70 years (n 9446). Individuals without data on tHcy levels (n 996) and all-cause mortality (n 8) were excluded. Finally, 8442 eligible participants were enrolled in the final analysis (Fig. 1).
Fig. 1. Flow chart of participants.
Exposure variable and endpoint
The exposure variable was tHcy level (µmol/l) and HHcy. HHcy was defined as a tHcy level greater than or equal to 15 µmol/l. Hcy was measured using the Abbott IMX analyser in 1999–2000, the Abbott Homocysteine IMX in 2001 and the Abbott Axsym in 2002. The details of the tHcy measurement process are available at http://cdc.gov/nchs/nhanes.
The primary endpoint was all-cause mortality, and the secondary endpoint was cause-specific mortality including CVD, cancer and respiratory disease mortality. The mortality information including cause and time of death was obtained from the 2015 NCHS Public-Use Linked Mortality Files. Follow-up data were taken for the period from the date of participation in the NHANES survey until the date of death or 31 December 2015. Cause of mortality was ascertained by the NCHS based on the International Classification of Diseases, 10th revision. CVD mortality was defined as death due to diseases of the heart (Codes: I00–I09, I11, I13, I20–I51) and cerebrovascular disease (Codes: I60–I69). Cancer mortality was defined as death due to malignant neoplasms (Codes: C00–C97), and respiratory disease mortality was defined as death due to chronic lower respiratory diseases (Codes: J40–J47), influenza and pneumonia (Codes: J09–J18).
Covariates
Statistical analyses were adjusted for a priori covariates based on well-known risk factors for mortality(21). The following covariates were included as continuous variables: age, BMI (kg/m2), mean systolic blood pressure (SBP, mmHg), mean diastolic blood pressure (mmHg), C-reactive protein (mg/dl) level, glycohaemoglobin (%) level, total cholesterol (mg/dl) level, albumin (g/dl) level, alanine aminotransferase (U/l), aspartate aminotransferase (U/l), γ-glutamyl transferase (U/l), alkaline phosphatase (U/l), uric acid (mg/dl) level, blood urea nitrogen (mg/dl), estimated glomerular filtration rate (ml/min per 1·73 m2), serum vitamin B12 (pg/ml) level, serum folate (ng/ml) level, dietary factors (total monounsaturated fatty acids (g), total polyunsaturated fatty acids (g), total saturated fatty acids (g), total fat intake (g), protein intake (g), dietary fibre (g), energy intake (kcal)) and supplement use (vitamin B12 (mg), folic acid (mg)). The categorical variables included sex, race (grouped as non-Hispanic White, Black, Mexican American, other Hispanic or other), education status (dichotomised as below high school diploma, high school diploma or any training above high school diploma), smoking status (grouped as never smoker, current smoker, former smoker), alcohol consumption (classified as less than 5 g or more than 5 g drinks/d), physical activity (classified as sedentary, low, moderate and high level based on the distribution of metabolic equivalent of task (MET)-minute levels in the present NHANES sample), history of diseases (coronary atherosclerotic heart disease (CAD), hypertension, diabetes and cancer) and medication use (lipid-lowering drugs, antihypertensive drugs and glucose-lowering drugs). CAD was defined as a self-reported physician diagnosis of CAD. Hypertension was defined as meeting any of the following criteria: SBP ≥ 135 mmHg, diastolic blood pressure ≥ 85 mmHg or self-reported physician diagnosis of hypertension. Diabetes was defined as a self-reported physician diagnosis of diabetes or a fasting glucose concentration >126 mg/dl. Cancer was defined as a self-reported cancer or malignancy (any type).
Statistical analyses
Statistical analyses were performed following the guidelines of the CDC (https://wwwn.cdc.gov/nchs/nhanes/tutorials/default.aspx). Each participant in the NHANES survey was assigned a sample weight(Reference Johnson, Paulose-Ram and Ogden22). The proposed weighting methodology in the analytical guidelines of the NCHS was adopted. Continuous variables were presented as mean values with their standard error using weighted linear regression models, and categorical variables were presented as proportions ± se using weighted chi-square tests.
A generalised additive model and smooth curve fitting (restricted cubic spline) based on Cox proportional hazards models were applied to estimate the relationship between tHcy level and mortality. We utilised four models simultaneously according to the STROBE guidelines: model 1 (not adjusted for any covariates), model 2 (adjusted for age (smooth), sex, race), model 3 (adjusted for all covariates shown in Table 1 except for dietary factors and supplement use) and model 4 (adjusted for all covariates shown in Table 1). Age covariates were entered into the equation using smooth curve fitting to account for the potential non-linear relationship between age and mortality. Hazard ratios (HR) and 95 % CI were estimated in the four models. tHcy level was included as a continuous variable and categorical variable (with and without HHcy). Cumulative survival rate analysis was performed using Kaplan–Meier curves with log-rank statistics according to different groups (with and without HHcy).
Table 1. Characteristics of study participants
(Mean values and standard errors)
HHcy, hyperhomocysteinaemia; CAD, coronary atherosclerotic heart disease; ACEi, angiotensin-converting enzyme inhibitors; CRP, C-reactive protein; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyl transferase; ALP, alkaline phosphatase; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; TMFA, total monounsaturated fatty acids; TPFA, total polyunsaturated fatty acids; TSFA, total saturated fatty acids.
* Unweighted number of observations in dataset.
Continuous variables were calculated by weighted linear regression model. Categorical variables were calculated by weighted chi-square test.
Subgroup analyses were performed according to age group (above and below 50 years), sex, race/ethnicity, survey cycles, lifestyle (smoking status, alcohol consumption and physical activity), history of chronic disease, medication use, BMI, estimated glomerular filtration rate, vitamin B12 level (tertile grouping) and folate level (tertile grouping) using stratified Cox proportional hazards models.
To confirm the robustness of our results, we quantified unmeasured confounders between HHcy and all-cause mortality by calculating E-values(Reference Haneuse, VanderWeele and Arterburn23), as unmeasured confounding factors may overturn the observed association between HHcy and all-cause mortality. E-values can estimate the validity required for a confounding factor.
All tests were two-sided and statistical significance was set at P < 0·05. All analyses were performed using the R statistical software package (http://www.R-project.org, The R Foundation for Statistical Computing), EmpowerStats (http://www.empowerstats.com, X&Y Solution, Inc.) and Free Statistics software versions 1.5(Reference Yang, Zheng and Chen24).
Results
Baseline characteristics of the participants
The weighted distribution of the baseline characteristics of the participants according to the presence or absence of HHcy is shown in Table 1. There were 235 (2·78 %) participants with HHcy. Compared with the participants without HHcy, participants with HHcy were slightly older, to be males and more likely to be drinker and former smokers, have a less physically activity, have a lower education level (below high school diploma), have a diagnosis of CAD, hypertension and/or diabetes, have a significantly higher SBP, C-reactive protein level, aspartate aminotransferase level, γ-glutamyl transferase level, alkaline phosphatase level, uric acid level and blood urea nitrogen level, have a lower estimated glomerular filtration rate, serum vitamin B12 and serum folate level and take less total monounsaturated fatty acids, vitamin B12 and folic acid.
Endpoints of mortality
A total of 8442 individuals (1999–2000 survey cycle: 3978 subjects; 2001–2002 survey cycle: 4464 subjects) aged 18–70 years with 119 364·8 person-years of follow-up (median follow-up duration, 14·7 years; interquartile range, 13·7–15·8 years) were included in the final data analysis. We noted that a total of 1007 (11·9 %) deaths occurred including 197 CVD-related deaths, 255 cancer-related deaths and fifty-eight respiratory disease deaths during follow-up. The participants with HHcy had a higher all-cause and cause-specific mortality rate (per 1000 person-years) than those without HHcy. P-values were all less than 0·001 (Table 2).
Table 2. The endpoints in participants without and with HHcy
HHcy, hyperhomocysteinaemia.
Association between total plasma homocysteine and hyperhomocysteinaemia an all-cause and cause-specific mortality
Four models were constructed using a generalised additive model to analyse the effect of tHcy levels and HHcy on all-cause and cause-specific mortality. The HR and 95 % CI for these equations are shown in Table 3. In model 4 (the fully adjusted model), a 1 µmol/l higher tHcy level was associated with 4 % increased risk of all-cause mortality (HR 1·04; 95 % CI (1·03, 1·05)), 6 % increased risk of CVD mortality (HR 1·06; 95 % CI (1·03, 1·09)), 1 % increased risk of cancer mortality (HR 1·01; 95 % CI (0·98, 1·05)) and an 8 % increased risk of respiratory disease mortality (HR 1·08; 95 % CI (1·03, 1·14)). The results of the association between tHcy level and cancer mortality did not reach statistical significance.
Table 3. Association of HHcy with the risk of all-cause and cause-specific mortality
(Hazards ratios and 95 % confidence intervals)
HR, hazards ratio; HHcy, hyperhomocysteinaemia.
*Cox proportional hazards models were used to estimate HR and 95 % CI.
† Model 1: no covariates were adjusted.
‡ Model 2: adjusted for age (smooth), sex, race/ethnicity.
§ Model 3: adjusted for age (smooth), sex, race/ethnicity, education status, smoking status, alcohol consumption, physical activity, coronary atherosclerotic heart disease, hypertension, diabetes, cancer, glucose-lowering drugs, statin use, ACEi use, BMI, SBP, DBP, CRP, glycohaemoglobin, total cholesterol, albumin, ALT, AST, GGT, ALP, uric acid, BUN, eGFR, serum vitamin B12, serum folate.
‖ Model 4: adjusted for age (smooth), sex, race/ethnicity, education status, smoking status, alcohol consumption, physical activity, coronary atherosclerotic heart disease, hypertension, diabetes, cancer, glucose-lowering drugs, statin use, ACEi use, BMI, SBP, DBP, CRP, glycohaemoglobin, total cholesterol, albumin, ALT, AST, GGT, ALP, uric acid, BUN, eGFR, serum vitamin B12, serum folate, total monounsaturated fatty acids, total polyunsaturated fatty acids, total saturated fatty acids, total fat intake, protein intake, dietary fibre, energy intake, and supplement use (vitamin B12, folic acid).
Participants with HHcy had a 93 % increased risk of all-cause mortality (HR 1·93; 95 % CI (1·48, 2·51)), a 160 % increased risk of CVD mortality (HR 2·60; 95 % CI (1·52, 4·45)), an 82 % increased risk of cancer mortality (HR 1·82; 95 % CI (1·03, 3·21)) and a 146 % increased risk of respiratory disease mortality (HR 2·46; 95 % CI (0·89, 6·81)) compared with those without HHcy. However, the P value was greater than 0·05 (not significant) for the risk of respiratory disease mortality.
Analyses of the dose–response relationship between total plasma homocysteine and all-cause and cause-specific mortality
The association between tHcy level and all-cause and cause-specific mortality was evaluated on a continuous scale using a generalised additive model and smooth curve fitting (restricted cubic spline method) based on Cox proportional hazards models. The fully adjusted smooth curve fitting showed a linear association between tHcy level and all-cause and cause-specific mortality (Fig. 2).
Fig. 2. Dose–response associations of homocysteine level with risk of all-cause (a), CVD (b), cancer (c) and respiratory disease mortality (d). The red solid line represents the estimated risk of all-cause and cause-specific mortality, with cyan dashed lines showing 95 % CI. Analyses were adjusted for age (smooth), sex, race/ethnicity, education status, smoking status, alcohol consumption, physical activity, coronary atherosclerotic heart disease, hypertension, diabetes, cancer, glucose-lowering drugs, statin use, ACEi use, BMI, SBP, DBP, CRP, glycohaemoglobin, total cholesterol, albumin, ALT, AST, GGT, ALP, uric acid, BUN, eGFR, serum vitamin B12, serum folate, total monounsaturated fatty acids, total polyunsaturated fatty acids, total saturated fatty acids, total fat intake, protein intake, dietary fibre, energy intake, and supplement use (vitamin B12, folic acid). ACEi, angiotensin-converting enzyme inhibitor; SBP, systolic blood pressure; DBP, diastolic blood pressure; CRP, C-reactive protein; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, γ-glutamyl transferase; ALP, alkaline phosphatase; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate.
Subgroup analyses
The results of the subgroup analyses of the association between HHcy and all-cause mortality are presented in Fig. 3. The association between HHcy and all-cause mortality in the stratified analysis was consistent with that in the multivariable Cox regression analysis, except for Mexican American.
Fig. 3. Association between hyperhomocysteinaemia and all-cause mortality according to subgroup. Analyses were adjusted for age (smooth), sex, race/ethnicity, education status, smoking status, alcohol consumption, physical activity, CAD, hypertension, diabetes, cancer, glucose-lowering drugs, statin use, ACEi use, BMI, SBP, DBP, CRP, glycohaemoglobin, total cholesterol, albumin, ALT, AST, GGT, ALP, uric acid, BUN, eGFR, serum vitamin B12, serum folate, total monounsaturated fatty acids, total polyunsaturated fatty acids, total saturated fatty acids, total fat intake, protein intake, dietary fibre, energy intake, and supplement use (vitamin B12, folic acid), except for the stratification variable. ACEi, angiotensin-converting enzyme inhibitor; SBP, systolic blood pressure; DBP, diastolic blood pressure; CRP, C-reactive protein; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, γ-glutamyl transferase; ALP, alkaline phosphatase; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; CAD, coronary atherosclerotic heart disease.
Survival analyses
Kaplan–Meier analysis showed that the survival probability among participants with HHcy was significantly lower compared with those without HHcy (both P < 0·0001) (Fig. 4).
Fig. 4. Kaplan–Meier curves for all-cause (a), CVD (b), cancer (c) and respiratory disease mortality (d). Unadjusted Kaplan–Meier estimates for all-cause and cause-specific mortality for HHcy. HHcy, hyperhomocysteinaemia.
Sensitivity analyses
To test the robustness of the primary results, we calculated an E-value to assess the effect of unmeasured confounding factors. The association between HHcy and the risk of all-cause mortality was found to be robust, unless the HR of all-cause mortality risk of an unmeasured confounder was >3·06.
Discussion
Our findings
In this study, we used the generalised additive model to illustrate the relationship between HHcy and the risk of all-cause and cause-specific mortality among adults under the age of 70 years from the general population of the USA. After adjusting for indicators such as demographics, traditional cardiovascular risk factors and laboratory test results, we observed that HHcy was associated with an increased risk of all-cause and cause-specific mortality after a median follow-up of 14·7 years. However, the risk of respiratory disease mortality did not reach statistical significance.
Previous studies
Our results are consistent with the follow-up report of older Framingham subjects(Reference Bostom, Silbershatz and Rosenberg25) and residents of Jerusalem(Reference Kark, Selhub and Adler26), which states that tHcy was an effective predictor of all-cause mortality. There is increasing evidence that elevated Hcy levels are associated with an increased risk of all-cause mortality(Reference Toyomasu, Adachi and Enomoto27,Reference Mendonça, Jagger and Granic28) . A meta-analysis of nineteen studies with 4340 subjects showed that elevated Hcy levels were associated with a 3·19-fold increased risk of all-cause mortality(Reference Zhang, Xiao and Yang29). Another meta-analysis of ten prospective studies with 11 061 participants found that stroke and ischaemic stroke risk increases in a dose-dependent manner with increases in tHcy level(Reference Wu, Zhou and Chen10). This is consistent with the results of the present study. However, one previous study reported that a high tHcy level is not associated with the risk of cardiovascular mortality after 10·3 years of follow-up in healthy individuals aged 20–59 years(Reference de Bree, Verschuren and Blom30). Nonetheless, that study may have been limited by the selection of relatively young healthy subjects. In addition, some important confounding variables, such as folic acid and vitamin B levels, were not adjusted in that study.
Possible explanations for our findings
HHcy can cause atherosclerosis and promote thrombus formation. This mechanism may cause endothelial dysfunction through increased oxidative stress(Reference Welch and Loscalzo31,Reference Kanani, Sinkey and Browning32,Reference Tawakol, Omland and Gerhard33,Reference Al-Obaidi, Philippou and Stubbs34) . Hcy can also affect the properties of the extracellular matrix, increase smooth muscle cell proliferation and induce platelet enrichment(Reference Bellamy, McDowell and Ramsey35,Reference Chambers, McGregor and Jean-Marie36,Reference Harker, Slichter and Scott37) . Studies have shown that Hcy level is positively correlated with age(Reference Selhub, Jacques and Wilson38). In the present study, participants with HHcy were older and more likely to have complications. Hcy level may be an indirect marker of serious diseases. The association between HHcy and the risk of all-cause and CVD mortality decreased significantly after adjusting for age and sex. Moreover, the subgroup analysis showed that there is a strong correlation between HHcy and the risk of all-cause mortality among participants with risk factors for CVD, such as current smokers, drinkers and those with a history of hypertension and diabetes. And, the participants with HHcy were more likely to be drinker, former smokers, have a less physically activity, have a diagnosis of CAD, hypertension and diabetes. People who smoke heavily for a long time may be prone to vitamin deficiency due to less intake of vegetables and fruits in their diet, which in turn leads to increased tHcy levels. Chronic alcohol consumption affects post-translational modification of hepatic methionine synthase(Reference Kharbanda39). Inhibition of methionine synthase by alcohol reduces re-methylation of Hcy, resulting in HHcy(Reference Kenyon, Nicolaou and Gibbons40). It may be a combination of multiple factors such as age, sex, unhealthy lifestyle (smoking and drinking habits, less physically activity) and chronic diseases that may interact with each other affecting the methionine-homocysteine cycle, thereby influencing the final adverse outcomes. Azarpazhooh et al.(Reference Azarpazhooh, Andalibi and Hackam41) found the metabolic syndrome, smoking and HHcy interact with each other and ultimately contribute to increased cardiovascular risk. Another study also showed that a healthy lifestyle such as physical activity, intaking of more fruit and quitting smoking can help prevent HHcy(Reference Han, Liu and Wang42).
Hypertension and hyperhomocysteinaemia
Subgroup analysis in our study showed that HHcy was associated with a higher risk of all-cause mortality in participants with hypertension than those without hypertension, which was in line with our previous study(Reference Zhao, Gao and Lv43). The Hordaland Homocysteine Study(Reference Nygård, Vollset and Refsum44) that included 16 176 individuals indicated that tHcy level was positively correlated with SBP and diastolic blood pressure. Results from another NHANES study(Reference Lim and Cassano45) also suggested that a 5-μmol/l increase in Hcy levels was associated with 0·7 and 0·5 mmHg increases in SBP and diastolic blood pressure, respectively. Zhong et al. (Reference Zhong, Zhuang and Wang11) found that elevated Hcy levels were associated with risk of essential hypertension. This may be due to the destruction of elastic fibres and increased arterial stiffness as a result of Hcy. Symons et al.(Reference Symons, Mullick and Ensunsa46) observed that the carotid elasticity was less in hyperhomocysteinemic rats compared with control rats.
Vitamin B12, folate and hyperhomocysteinaemia
Vitamin B12 and folate are important factors in Hcy metabolism and important determinants of tHcy concentration. Elevated tHcy levels may reflect a lack of folate and vitamin B12 (Reference Vermeulen, Stehouwer and Twisk47,Reference Voutilainen, Rissanen and Virtanen48) . B-complex vitamins can reduce tHcy levels by promoting Hcy metabolism(Reference McNulty, Pentieva and Hoey49). Bertoia et al. revealed a negative correlation between high folate intake and tHcy levels(Reference Bertoia, Pai and Cooke50). Studies have shown that two-thirds of patients with HHcy have lower plasma folate and vitamin B12 concentrations than normal(Reference Ubbink, Vermaak and van der Merwe51). A study of 1041 relatively older adults also showed that vitamin B12 plays an important role in the pathogenesis of HHcy(Reference Selhub, Jacques and Wilson38). Similarly, in the present study, we found that participants with HHcy had lower vitamin B12 and folate levels. In addition, our subgroup analysis showed that tHcy levels showed a stronger relationship with the risk of all-cause mortality in the lower tertile of vitamin B12 than in the higher tertile, a finding which is consistent with that of previous studies(Reference Jacques, Selhub and Bostom52). However, a meta-analysis of a large, randomised trial did not show that vitamin B therapy has a beneficial effect on the mortality of individuals at risk of CVD or those suffering from CVD(Reference Clarke, Halsey and Lewington53). Therefore, further observation and clinical trials are necessary to develop appropriate primary prevention strategies.
Effects of drugs and hyperhomocysteinaemia
Numerous epidemiological studies have shown that statin use is significantly associated with a reduction in all-cause and cardiovascular mortality(Reference Orkaby, Driver and Ho54,Reference Baigent, Keech and Kearney55,Reference Rea, Biffi and Ronco56) and angiotensin-converting enzyme inhibitors can significantly reduce the risk of all-cause mortality in patients with co-morbidities such as hypertension, diabetes, CHD and chronic kidney disease(Reference Savarese, Costanzo and Cleland57,Reference Cheng, Zhang and Zhang58) . However, in our subgroup analysis, participants with HHcy who took statins or angiotensin-converting enzyme inhibitors were associated with a higher risk of all-cause mortality than those who did not. This may be because of a higher prevalence of co-existing cardiovascular risk factors such as hyperlipidaemia and hypertension in participants with HHcy as they had a higher percentage of statin and angiotensin-converting enzyme inhibitor drug usage. In addition, statins may not reduce tHcy concentrations. A meta-analysis of seven studies showed no significant alteration in Hcy levels following treatment with statins(Reference Sahebkar, Pirro and Reiner59).
Limitations
The limitations of our study should be noted as well. First, the prevalence of HHcy in kidney transplant recipients is higher than that in the general population(Reference Kang, Nigwekar and Perkovic60). Our study data did not include information on history of kidney transplantation; therefore, we could not assess the role of kidney transplants in the relationship between HHcy and the risk of mortality. Second, the possibility of the residual confounding effect of incomplete adjustment of some cardiovascular risk factors cannot be excluded. However, an E-value analysis was conducted to quantify the potential impact of unmeasured confounders. The results showed that an unknown confounder was unlikely to explain the effect of the risk of all-cause mortality. Third, all blood tests including tHcy levels were based on a single measure, and regression dilution bias may underestimate the strength of the association. Therefore, larger-scale studies among the general population are needed to estimate the strength of the correlation between tHcy level and mortality more accurately.
Conclusion
The main finding of this study was that HHcy was associated with high risk of all-cause and cause-specific (CVD, cancer) mortality among adults aged below 70 years in the USA, which suggests that maintaining tHcy at normal levels may be beneficial in reducing the risk of mortality. Future prospective studies are needed to evaluate the clinical benefit of a Hcy-lowering intervention.
Acknowledgements
The authors thank all the staff members of the NHANES study for their valuable contributions. The authors also thank Jie Liu of Department of Vascular and Endovascular Surgery, Chinese PLA General Hospital, Haibo Li of Fujian Maternity and Child Health Hospital, Changzhong Chen, Xinglin Chen of Yi-er College for their contribution to the statistical support. The authors would like to thank Editage (www.editage.cn) for English language editing.
The authors appreciate for partial financial supports done by Zhejiang Province Public Welfare Technology Application Research Project (No. LGF20G030011) and Zhejiang Province Health Science and Technology Plan (No. 2021KY467).
W. Y. Z.: writing original draft. Y. L.: analysed the data. H. B. H. and H. L. M.: conceptualisation. W. Y.: data curation. Q. H.: design of methodology. X. C. and F. L. G.: writing – review and editing.
The authors declare that no conflicts of interest exist.
