Introduction

Homocysteine is a sulfur-containing amino acid formed during metabolism of the essential amino acid methionine [1, 2]. Serum total homocysteine concentration (tHcy) has been recognized as an atherogenic factor, promoting oxidative stress, inflammation, thrombosis, endothelial dysfunction, and cell proliferation [2–4]. In addition, many previous epidemiologic studies suggest that elevated tHcy is a new, strong, and independent risk factor of coronary heart disease (CHD) and stroke [5–7]. Also in Japan, several previous studies reported that tHcy was an independent risk factor of ischemic stroke [8, 9].

It has also been reported that several factors, including age, gender, smoking, alcohol consumption, malignancies, thyroid disease, renal failure, nutrition (deficiency of folic acid, Vitamin B6 and B12), medication, and gene type of methylene tetrahydrofolate reductase (MTHFR) can affect tHcy [10–12]; Jacques [13–15]. Thus, the changes in tHcy might derive from a combination of these factors [10]. As regards the effects of age and gender, it was clearly shown that tHcy increased with age, and was higher in males than in females [16].

Diabetic patients have increased risk of vascular complications or CHD, since hyperglycemia leads to vascular dysfunction [17, 18]. In Japan, diabetics have a two to four-fold greater risk of vascular disease occurrence as compared with non-diabetics [19]. In recent years, it has been suggested that atherosclerosis develops markedly even in the phase of mild impairment of glucose tolerance [20]. In addition, several studies have shown the positive and negative association between tHcy and pathophysiology of diabetes mellitus (DM) [14, 21–24]. In Japan, some studies have reported the factors associated with tHcy in DM [25]. However, the clinical guideline for lowering tHcy has not yet been fully established in Japanese community health-care activities. Accordingly, it will be necessary that the determinants of tHcy in DM patients will be concurrently analyzed by multi-factors.

The aim of this study was, first, to investigate the cross-sectional association of tHcy with DM in middle-aged and old-aged Japanese according to sex. The second objective was to clarify the associations of tHcy with other various biomarkers or health indices. Moreover, the third objective was to clarify which type of DM treatment and complication could further specifically affect tHcy in DM patients.

Subjects and methods

Subjects

A total of 1,276 individuals, who had visited a medical clinic in Tokyo, from 1st August 2003 to 31st July 2005, were enrolled in the study. They consisted of 587 males and 689 females, who had had a complete medical check-up or diabetic outpatient care at the clinic. Two hundred and eighty of the 1,276 subjects were diagnosed with type 2 DM. No one suffered from type 1 or other DM types. The diabetic patients in this study were treated in one of three ways: with special dietetic therapy and exercise by dieticians and exercise therapists; with oral agents; or with insulin therapy. All patients received the complete course of their therapy. Patients with malignancy or thyroid disease were excluded from the study. One-hundred and eighty-six of the 1,276 subjects had experienced myocardial infarction (MI) or cerebral infarction (CI). All the subjects gave written informed consent before participating in the study. This study was approved by the Human Ethics Committee of Kyorin University School of Medicine.

Clinical measurements and questionnaire

All subjects underwent the following physical examinations: tHcy, fasting plasma glucose level (FPG), hemoglobin A1c level (HbA1c), total cholesterol level in serum (TC), serum triglyceride level (TG), serum creatinine level (Cre), blood pressure (BP), and body mass index (BMI). Furthermore, estimated glomerular filtration rate (eGFR) was calculated by the modification of diet in renal disease (MDRD) method [26].

Serum total homocysteine concentration was determined by HPLC (LC-9A, Shimadzu, Kyoto, Japan) [11]. Cre was assayed by an enzyme method. The glucose-hydrogenase method was used for plasma glucose assay [27]. The latex agglutination turbidimetric immunoassay method was used for HbA1c assay [28]. TC was determined by the cholesterol dehydrogenase (UV-End) method [29]. TG was determined by an enzyme method [30] and transformed to a logarithm because of non-normal distribution. Brachial blood pressure of the subjects was measured by use of a mercury sphygmomanometer in the sitting position after sufficient rest. BP was classified into the following three categories: hypertension, systolic blood pressure (SBP) ≥ 140 mmHg, or diastolic blood pressure (DBP) ≥ 90 mmHg; high level of normal range, 130 mmHg ≤ SBP < 140 mmHg, and/or 80 mmHg ≤ DBP < 90 mmHg except for hypertension; and normal, SBP < 130 mmHg and DBP < 80 mmHg [31].

Personal information regarding the subjects was collected from their medical records and from a questionnaire.

Identification of DM cases was fulfilled according to diagnostic criteria set by Japan Diabetes Society recommendations [32]. Presence of MI or CI, contents of treatment regimen for DM and complication status of DM characterized by neuropathy, retinopathy, or nephropathy, were also confirmed from subjects’ medical records at the point of the examinations. In this study, patients with nephropathy had 300 mg/day or more urinary protein.

An interview-based questionnaire was conducted to investigate each subject’s background and lifestyle. The questionnaire included questions concerning the issues:

  1. 1

    smoking habits—non-smokers or current smokers; and

  2. 2

    drinking habits—non/occasional drinkers or habitual drinkers.

Statistical analysis

In this study, the data of age, gender, tHcy, FPG, HbA1c, TC, logarithmic TG (logTG), Cre, eGFR, BMI, BP (normal = 0, high level of normal range = 1, hypertension = 2), current smoking habit (non-smokers = 0, current smokers = 1), routine drinking habit (non-or occasional drinkers = 0, habitual drinkers = 1), DM (non-diabetics = 0, diabetics = 1), diabetic complication (neuropathy, retinopathy, nephropathy), types of treatment regimen for DM (only diet and exercise = 0, oral agent = 1, insulin therapy = 2), previous history of MI or CI (not yet = 0, presence of MI or CI = 1), and information about taking antihypertensive agent (no = 0, routine intake = 1) or lipid-lowering agent (no = 0, routine intake = 1) were subjected to analyses as a cross-sectional study. In addition to the analyses for all the subjects (n = 1,276), the DM group (n = 280) and the non-DM group (n = 996) were also conducted as the subgroups. Multiple regression analysis was performed to analyze the association of tHcy with selected factors. The 0.05 level was regarded as the cut-off point of statistical significance.

All the statistical analyses were performed using SAS statistical software for Windows, version 8.2.

Results

Detailed characteristics of subjects are shown in Table 1. The mean levels of tHcy were 11.26 μmol/l for males and 8.60 μmol/l for females. tHcy was higher in males than in females, and was higher in DM than in non-DM.

Table 1 Basic characteristic data for all subjects
Full size table

As shown in Table 2, in both sexes, DM was significantly associated with increased tHcy after adjustment for age and Cre; partial regression coefficient (β) ± standard error in tHcy between DM and non-DM was 1.409 ± 0.458 (p = 0.002) in males and 1.004 ± 0.275 (p < 0.001) in females. The association of HbA1c with tHcy was significant in males only after adjustment for multi-variables. Substituted DM or FPG as variables for HbA1c in Model B produced the same results: standard partial regression coefficient of DM for tHcy was 0.107 (p < 0.05) in males and 0.060 (p > 0.05) in females; that of FPG for tHcy was 0.098 (p < 0.05) in males and 0.047 (p > 0.05) in females, (data not shown in Table 2). eGFR had the strongest association with tHcy in both sexes.

Table 2 Correlations with serum total homocysteine concentration level in all subjects, calculated by multiple regression analysis
Full size table

As shown in Table 3, eGFR and Cre had a strong association with tHcy in both sexes according to multiple regression analysis, irrespective of presence or absence of DM. Furthermore, eGFR of the diabetics except patients with nephropathy had also strong association with tHcy in both sexes (p = 0.023 and p = 0.017 in males and females, respectively, data not shown in Table 3). eGFR and Cre had also strong association with tHcy after adjustment for Model B. Table 3 shows that the factors except eGFR and Cre associated with tHcy were different between diabetics and non-diabetics in males. The FPG and the logTG after adjustment for Model A were significantly positively associated with tHcy in diabetic males only. In contrast, taking anti-hypertension agents, BP, and drinking habits were associated with tHcy in non-diabetic males. However, there was no difference in the factors associated with tHcy between DM and non-DM in females. Age, BP, Cre and eGFR were significantly associated with tHcy in females. With regard to association with lifestyle, current smoking habits had no significant association, but they had a close positive relation with tHcy in diabetic males, whereas drinking habits were negatively associated with tHcy in non-diabetic males.

Table 3 Correlations with serum total homocysteine concentration level in diabetics and non-diabetic subjects according to multiple regression analysis
Full size table

In Table 4, tHcy is compared among different subgroups by type of DM treatment and diabetic complications according to analysis of covariance. As regards the association with type of treatment, tHcy was highest in subjects receiving insulin therapy and lowest in those using only diet and exercise, after adjustment for age, HbA1c, and Cre. It was found that nephropathy was significantly associated with tHcy after adjustment for age and HbA1c.

Table 4 Comparisons of serum total homocysteine level among different subgroups in diabetic subjects according to analysis of covariance
Full size table

Even after further adjustment for past history of cardiovascular events and intake of metformin, one of the important agents affecting tHcy [33], all the above results were basically unchanged (data were not shown).

Discussion

First, the association between tHcy and glycemic control, DM treatment type, and diabetic complication were considered. Glycemic status had an independent close association with tHcy in diabetic males only. In this study, the association of tHcy and glycemic control in females was not significant, suggesting that this sex difference was possibly caused by sex hormones. As estrogen inhibits the progression of atherosclerosis [34], shortage of estrogen in males may lead to elevation of tHcy, which promotes thrombogenesis.

Hyperglycemia increases oxidative stress and asymmetric dimethylarginine (ADMA) in the vascular endothelium [35]. Meanwhile, elevated tHcy induces endothelial dysfunction via increased oxidative stress and ADMA [4, 36, 37]. This oxidative stress affecting tHcy may arise as a result of oxidative damage of vascular endothelial cells and increased proliferation of vascular smooth muscle cells after oxidative metabolism of homocysteine to homocysteine and homocysteine thiolactone. However, it remains unclear whether hyperglycemia leads to elevation of tHcy or, conversely whether tHcy leads to hyperglycemia [2]. For males in this study, the glycemic status had close association with tHcy in DM but not in non-DM. Therefore, tHcy may increase after the onset of DM.

Although we did not measure plasma insulin levels in this study, some previous studies have shown the positive association between insulin-resistant status and tHcy, and indicated that insulin levels may affect the activity of enzymes involved in homocysteine metabolism [38]. Fonseca et al. [38] have reported that insulin treatment in rats caused a reduction of the enzyme cystathionine β-synthase and increasing MTHFR activity, leading to elevated tHcy. This may suggest that not only chronic hyperinsulynemia but also acute hyperinsulynemia by insulin injection increase tHcy. In this study, DM treatment levels after adjustment for HbA1c, age, and Cre were not significantly associated with tHcy. However, tHcy of the subjects receiving insulin therapy was relatively high compared with the other two categories of treatment. These findings were consistent with previous studies.

With respect to the effects of diabetic complications, a strong association between nephropathy and tHcy was found in this study. This finding was consistent with the results from previous studies [39–41]). It has been revealed that the deterioration of renal function is associated with elevated tHcy. The transformation of methionine to homocysteine occurs via a demethylation pathway, providing a methyl group with glycocyamine in the creation of creatinine [42]. Thus, in renal metabolism, homocysteine is closely linked with formation of creatinine [43], through decreased renal homocysteine excretion, impaired renal metabolism, or generally reduced B vitamins status in renal failure [44]. Therefore, Elias and Eng [24] suggested that adequate management of tHcy could enable prevention of renal dysfunction.

Second, it is noteworthy that eGFR and Cre within their normal ranges were strongly associated with tHcy in every group, DM and non-DM, in both sexes of this study. These results were consistent with a report by Abdella, et al. [45]. Subtle changes in eGFR and Cre after adjustment reflect minimal abnormality of renal function, even though within their normal ranges they may affect tHcy. It has been established that the incidence and mortality of CKD increase with the deterioration of eGFR [46]. This study suggests that the strong association between tHcy and eGFR may represent a link between heart and kidney. There is a possibility that a slight change of tHcy in DM patients or prediabetics can indicate early microangiopathy of the kidney. However a previous study gave conflicting results, showing that tHcy in early diabetic nephropathy was not increased [47]. As another possibility, elevated tHcy might cause microvascular endothelial damage in a renal circulation. Clarification of these mechanisms may be an important issue in future research.

One of the limitations of this study is that we did not have information on nutrition status of vitamin B6, B12 and folic acid, or gene type such as MTHFR. Furthermore, since this study was based on a cross-sectional design, many statistical associations with tHcy found in this study could not be correctly interpreted.

In the future, further follow-up or interventional studies will be necessary in order to increasingly clarify the significance of tHcy in diabetics.

In conclusion, these findings suggest there are some differences in the factors associated with tHcy between diabetics and non-diabetics, and between males and females. Glycemic status in DM was more strongly associated with tHcy in males than in females. There is, therefore, circumstantial evidence a prevention and treatment regimen may be required for elevated tHcy. Because tHcy was strongly associated with eGFR and Cre even within the normal ranges, tHcy may have important implications regarding the microangiopathy of the kidney and atherosclerosis.