Introduction

Hypertensive disorders of pregnancy (HDP), a common and serious complication of pregnancy, occur in 6~10% of pregnancies [1]. The etiology of HDP remains to be clarified, although there are various hypotheses [2, 3]. Because essential treatment for HDP has not yet been established, its early detection is important for HDP prevention and the achievement of better prognosis. Using an effective peptidomic analysis [4], we identified seven circulating HDP-associated peptides (P-2081, P-2091, P-2127, P-2209, P-2378, P-2858 and P-3156) (Table 1) and proposed them as biomarkers for the diagnosis of HDP [3, 5, 6]. These peptides are fragments of some proteins: P-2081, P-2127 and P-2209 originated from kininogen-1; P-2091 originated from fibrinogen-α; P-2378 originated from complement component 4 (C4); P-2858 originated from α2-HS-glycoprotein; and P-3156 originated from inter-α-trypsin inhibitor heavy chain H4 (ITIH4) [3, 5, 6]. It was reported that the levels of these seven peptides in patients with HDP were significantly different from the levels in pregnant women without HDP [5]. However, it remains unknown whether and how these HDP-associated peptides are related to blood pressure in a general population.

Table 1 Characteristics of the peptides associated with hypertensive disorders of pregnancy
Full size table

Women with HDP have been reported to show accelerated cardiovascular aging and a high incidence of cardiovascular events later in life [7, 8]. Hypertension is a major risk factor for atherosclerotic diseases, including coronary heart disease, heart failure, stroke and peripheral artery disease [9]. Individuals with hypertension are prone to other cardiovascular risk factors, including visceral obesity, dyslipidemia and hyperglycemia [9]. The accumulation of these risk factors is defined as metabolic syndrome, which is particularly dangerous due to the associated risk of developing ischemic heart disease, stroke and diabetes [10]. Interestingly, the parent proteins of the above seven peptides (fibrinogen, C4, α2-HS-glycoprotein, ITIH4) are acute-phase proteins, which have been proposed to be biomarkers for atherosclerotic disease [11]. Thus, it would be of interest to know whether HDP-associated peptides are related to hypertension and other cardiovascular risk factors in a general population.

The purpose of this study was therefore to clarify the relationships of HDP-associated peptides with hypertension in the general population. We also investigated the relationships between the peptides and other cardiovascular risk factors, including adiposity-related indices, blood lipids, glycemic status, blood lipid-related indices and metabolic syndrome.

Methods

Participants

A total of 69 Japanese male workers who had received periodic health checkup examinations at their workplace (mean age, 46.4 years) were included in this study. The protocol of this study was approved by the Hyogo College of Medicine Ethics Committee (No. 3449). All participants provided written informed consent to participate in the study. All methods were performed in accordance with the relevant guidelines and regulations. There were no participants with acute infection or serious diseases, including cancer and cardiovascular disease. Histories of alcohol consumption, cigarette smoking, illness, and therapy for illness were surveyed by self-report questionnaires. The weekly average amount of alcohol consumption was reported in the questionnaires. The frequency of alcohol consumption was asked as “How frequently do you drink alcohol?” and was categorized as “every day” (regular drinkers), “sometimes” (occasional drinkers) and “never” (nondrinkers). Habitual smokers were categorized according to daily average cigarette consumption as light smokers (20 cigarettes or less), heavy smokers (between 21 and 40 cigarettes) and very heavy smokers (>40 cigarettes) by the questionnaires. Smokers were defined as persons who had smoked cigarettes in the past month and for 6 months or longer.

Blood sample collection

Venous blood was collected from each participant in the morning after overnight fasting, and serum was separated from the blood by centrifugation. Concentrations of blood lipids, creatinine and hemoglobin A1c in the serum were measured by the methods described below. To avoid proteolysis of the peptides, we tried as much as possible to keep blood samples at a low temperature (below 4 °C) after collecting blood. A portion of the serum was stored at −80 °C until determination of peptide concentrations by mass spectrometry.

Measurements of the seven HDP-associated peptides in serum

The concentration of each peptide in serum was measured according to the method described previously [12,13,–14] with some modifications. Briefly, 50 μl of each serum sample was spiked with stable isotope-labeled (SI) synthetic standard peptides and heated at 100 °C for 10 min. After centrifugation, the supernatant was applied to a graphite carbon GL-tip GC device (GL Sciences Inc., Tokyo, Japan) equilibrated with a solution containing 3% acetonitrile and 0.1% trifluoroacetic acid (TFA), and bound materials were eluted with 60 µL of a solution containing 30% acetonitrile and 0.1% TFA. The eluate was evaporated and reconstituted with a solution containing 2% acetonitrile and 0.1% TFA.

The prepared sample corresponding to 7.5 µL of each serum sample was applied to the nanoflow reverse-phase C18 LC–MS/MS system (QTRAP 6500, SCIEX, Framingham, MA, USA) and quantified by the multiple reaction monitoring (MRM) mode. The MRM transitions were set into two periods according to the retention time of each peptide on LC: four peptides (P-2081, P-2091, P-2127 and P-2209) had an early retention time, and three peptides (P-2378, P-2858 and P-3156) had a late retention time [5]. The mass chromatographic peak area of each peptide was calculated using the MQ4 algorithm of MultiQuant 3.0 software (SCIEX), and the peptide concentration was quantified on the basis of the ratio of the peak areas of the natural and SI peptides.

Measurements of cardiovascular risk factors

At the health checkup examination, height and body weight were measured, and body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Waist circumference was measured at the navel level according to the definition of the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome [15], and visceral obesity was evaluated by using the waist-to-height ratio defined as waist circumference (cm) divided by height (cm). Resting blood pressure was measured with each participant in a sitting position with an oscillometric device (TM-2580, A&D Co., Ltd, Tokyo, Japan). Concentrations of triglycerides, HDL cholesterol, LDL cholesterol and creatinine in the serum were measured by enzymatic methods using the following commercial kits: pureauto S TG-N, cholestest N-HDL, cholestest LDL and pureauto S CRE-L (Sekisui Medical Co., Ltd, Tokyo, Japan), respectively. The hemoglobin A1c concentration was measured by the latex cohesion method using a commercial kit (Determiner HbA1c, Kyowa Medex, Tokyo, Japan) and calibrated by using a formula proposed by the Japan Diabetes Society [16]. The cardiometabolic index (CMI) was defined as the product of the ratio of waist circumference to height (cm/cm) and the ratio of triglycerides to HDL cholesterol (mg/dl/mg/dl) [17]. To evaluate renal function, the estimated glomerular filtration rate (eGFR) was calculated by using the following equation developed by the Japanese Society of Nephrology: eGFR (mL/min/1.73 m2) = 194 × serum creatinine−1.094 × age−0.287 [18].

Criteria of metabolic syndrome

Metabolic syndrome was defined according to the criteria proposed by the IDF (International Diabetes Federation) [19] with a slight modification as the presence of two or more risk factors in addition to visceral obesity diagnosed according to a large waist circumference. Risk factors included in the criteria are visceral obesity (high waist-to-height ratio), high blood pressure, dyslipidemia (low HDL cholesterol and/or high triglycerides) and hyperglycemia (high hemoglobin A1c). The criteria for each risk factor are as follows: visceral obesity, waist-to-height ratio ≥ 0.5 [20]; high blood pressure, systolic blood pressure ≥ 130 mmHg and/or diastolic blood pressure ≥ 85 mmHg [21]; low HDL cholesterol, HDL cholesterol < 40 mg/dl; high triglycerides, triglycerides ≥ 150 mg/dl [19, 22]; and diabetes, hemoglobin A1c ≥ 6.5% [23]. Drug therapy for hypertension, dyslipidemia or diabetes was also included in the above definitions of high blood pressure, dyslipidemia and diabetes, respectively.

Statistical analysis

Pearson’s correlation coefficients or Spearman’s rank correlation coefficients were calculated in univariable correlation analysis. Standardized partial regression coefficients (β) were calculated in multivariable linear correlation analysis. Since the levels of each peptide, triglycerides, triglycerides-to-HDL-cholesterol ratio and CMI did not display a normal distribution, their values are presented as medians with interquartile (25%, 75%) ranges and were analyzed nonparametrically or parametrically after base-10 logarithmic transformation. To compare the mean levels of each variable between two groups, Student’s unpaired t test was used in univariable analysis, and analysis of covariance (ANCOVA) with Student’s t test was used in multivariable analysis. In ANCOVA, the dependent variable was each peptide level after base-10 logarithmic transformation, the independent variable was each cardiovascular risk factor (diastolic hypertension, low HDL cholesterol or hypertriglyceridemia), and the covariates were possible confounding factors including age, BMI, history of smoking, history of alcohol consumption and history of medication for hypertension or dyslipidemia. Age, BMI, habits of smoking and alcohol consumption, and history of medication for hypertension or dyslipidemia were also used as explanatory variables in multivariable linear regression analyses. After arranging the participants with HDL cholesterol levels in ascending order, they were divided into three almost equal-sized tertile groups. Because the percentage of subjects showing abnormally low HDL cholesterol levels (<40 mg/dl) was low (5.8%), means of peptide levels were compared between the lowest tertile of HDL cholesterol (34~51 mg/dl) and the two higher tertiles (HDL cholesterol: 53~137 mg/dl). Statistical significance was defined when probability (p) values were <0.05. A computer software program (IBM SPSS Statistics for Windows, Version 25.0., IBM Corp., Armonk, NY, USA) was used for the statistical analyses.

Results

Characteristics of participants

Table 2 shows the characteristics of the participants. The percentages of smokers and alcohol consumers were 43.5% and 75.4%, respectively. The mean concentrations of the seven peptides are shown as ng/ml: among the seven peptides, the most abundant peptide in serum was P-2858, and P-2091 and P-2378 were relatively abundant. The concentrations of P-2081, P-2127 and P-2209 were lower than 1 ng/ml. Approximately one-fourth and one-third of the participants had high diastolic blood pressure (≥85 mmHg) and high triglycerides (≥150 mg/dl), respectively.

Table 2 Characteristics of the participants
Full size table

Correlations between the levels of each peptide and blood pressure (systolic, diastolic and mean blood pressure) in univariable analysis

Among the seven HDP-associated peptides tested, only P-2091 showed significant correlations with diastolic blood pressure and mean blood pressure (Table 3). However, there was no significant correlation between P-2091 and systolic blood pressure. Levels of the other six peptides were not significantly correlated with systolic, diastolic or mean blood pressure levels.

Table 3 Correlations between each peptide level and blood pressure in univariable analysis
Full size table

Correlations between the levels of each peptide and other cardiovascular risk factors in univariable analysis

Table 4 shows correlation coefficients between the levels of each cardiovascular risk factor (other than blood pressure) and each peptide in univariable analysis. The P-2378 level showed a significant inverse correlation with HDL cholesterol and a significant positive correlation with the ratio of LDL cholesterol to HDL cholesterol. The P-3156 level showed significant inverse correlations with triglycerides, the ratio of triglycerides to HDL cholesterol and CMI. The P-2378 level showed a significant positive correlation with BMI, while the P-3156 level showed significant inverse correlations with BMI and waist-to-height ratio. eGFR did not show a significant correlation with any of the seven HDP-associated peptides (Spearman’s rank correlation coefficient: P-2081, 0.076 [p = 0.536]; P-2091, 0.135 [p = 0.270]; P-2209, −0.028 [p = 0.819]; P-2127, 0.117 [p = 0.340]; P-2378, 0.141 [p = 0.247]; P-2858, 0.073 [p = 0.550]; P-3156, −0.119 [p = 0.328]).

Table 4 Correlations between each peptide level and variables related to cardiovascular risk
Full size table

Distributions of serum levels of P-2091, P-2378 and P-3156

As shown in the histograms in Fig. 1, the levels of P-2091, P-2378 and P-3156 did not show a normal distribution and were normalized by base-10 logarithmic transformation. Therefore, they were used after logarithmic transformation in the following analyses: linear regression analysis, univariable comparison of means, and ANCOVA.

Fig. 1
figure 1

Histograms of concentrations of P-2091 (A), P-2378 (C) and P-3156 (E) in serum and their logarithmically transformed values (B, D and F, respectively)

Full size image

Correlations between the level of P-2091 and blood pressure (systolic, diastolic and mean blood pressure) in multivariable analysis

Table 5 shows the results of multivariable linear regression analysis of the relationships of the level of P-2091 with blood pressure after adjustment for age, BMI, a history of therapy for hypertension, and histories of alcohol consumption and smoking. The level of log-transformed P-2091 concentration showed significant inverse correlations with diastolic blood pressure and mean blood pressure but not systolic blood pressure in multivariable analysis. Age was significantly correlated with systolic, diastolic and mean blood pressure levels, and BMI was significantly correlated with systolic and mean blood pressure levels (Table 5).

Table 5 Correlations between P-2091 and blood pressure in multivariable analysis
Full size table

Correlations between P-2378 and HDL cholesterol and between P-3156 and triglycerides in multivariable analysis

In multivariable analysis adjusted for age, BMI, a history of therapy for dyslipidemia, and histories of alcohol consumption and smoking, the log-transformed level of P-3156 was significantly correlated with the log-transformed triglyceride level, while there was no significant correlation between the log-transformed level of P-2378 and HDL cholesterol (Table 6). Alcohol consumption was positively correlated with HDL cholesterol, while smoking showed an inverse correlation with HDL cholesterol. Histories of medication therapy for dyslipidemia and smoking showed significant positive correlations with the log-transformed triglyceride level (Table 6).

Table 6 Correlations of P-2378 with HDL cholesterol and of P-3156 with triglycerides in multivariable analysis
Full size table

Correlations of the levels of P-2378 and P-3156 with lipid-related indices in multivariable analysis

Correlations of the levels of P-2378 and P-3156 with lipid-related indices were investigated in multivariable analysis adjusted for age, a history of therapy for dyslipidemia, and histories of alcohol consumption and smoking. P-2378 was significantly correlated with the LDL-C/HDL-C ratio (standardized partial regression coefficient [β], 0.289 [p = 0.019]) but not with the TG/HDL-C ratio (β, −0.046 [p = 0.703]) or CMI (β, −0.015 [p = 0.899]), while P-3156 was significantly correlated with the TG/HDL-C ratio (β, −0.384 [p = 0.001]) and CMI (β, −0.396 [p < 0.001]) but not with the LDL-C/HDL-C ratio (β, −0.012 [p = 0.924]).

Comparison of mean levels of P-2091 in the groups of subjects with and without high diastolic blood pressure

In both univariable analysis and multivariable analysis, the mean log-transformed P-2091 levels were significantly lower in subjects with high diastolic blood pressure than in subjects with normal diastolic blood pressure (Fig. 2A).

Fig. 2
figure 2

Comparisons of peptide levels between normal and high or low groups of diastolic blood pressure (DBP) (A), HDL cholesterol (HDL-C) (B) and triglycerides (TG) (C). Peptide levels were used for analysis after base-10 logarithmic transformation in order to normalize them. The values at the y-axes show levels of each peptide after base-10 logarithmic transformation. Asterisks denote significant differences from the normal groups (A, C) and the higher tertile group (B) (*p < 0.05; **p < 0.01)

Full size image

Comparison of mean levels of P-2378 in the groups of subjects with and without relatively low HDL cholesterol

In both univariable analysis and multivariable analysis, the mean log-transformed P-2378 levels were significantly higher in the lowest tertile group of HDL cholesterol than in the group of the higher (2nd and 3rd) tertiles (Fig. 2B).

Comparison of mean levels of P-3156 in the groups of subjects with and without high triglycerides

Both in univariable analysis and multivariable analysis, the mean log-transformed P-3156 levels were significantly lower in the subjects with high triglycerides than in the subjects with normal triglycerides (Fig. 2C).

Correlations between levels of each peptide pair in univariable analysis

Table 7 shows Spearman’s rank correlation coefficient between the levels of two peptides. The level of P-2081 showed significant positive correlations with the levels of P-2127, P-2209 and P-3156. The level of P-3156 also showed significant positive correlations with the levels of P-2091 and P-2209.

Table 7 Correlations between each pair from the seven peptides associated with hypertensive disorders of pregnancy
Full size table

Comparison of peptide levels between subjects with and without metabolic syndrome

Approximately one-fourth of the participants (n = 19) had metabolic syndrome. Mean levels of the seven peptides related to HDP were compared in subjects with and without metabolic syndrome in univariable analysis and multivariable analysis after adjustment for age and histories of smoking and alcohol consumption. As shown in Table 8, there were no significant differences in the levels of each peptide between the groups with and without metabolic syndrome in both univariable analysis and multivariable analysis.

Table 8 Comparison of each peptide level between subjects with and those without metabolic syndrome
Full size table

Discussion

This is the first study demonstrating the significance of several potential biomarkers for HDP in relation to cardiovascular risk in the general population. The peptides P-2091, P-2378 and P-3156 were inversely associated with diastolic blood pressure, HDL cholesterol and triglycerides, respectively. High blood pressure, low HDL cholesterol and high triglycerides are major cardiovascular risk factors. Thus, the above three HDP-associated peptides were also associated with cardiovascular risk factors in men, implying that these peptides are indirectly related to cardiovascular disease and possibly discriminate cardiovascular risk. Further studies including prospective studies are needed to clarify the direct relations of the peptides with cardiovascular disease and to determine the significance of the peptides in the pathogenesis of cardiovascular disease.

In the present study, among the seven HDP-associated peptides, only P-2091 showed an inverse association with blood pressure, and blood levels of the other six peptides were not associated with blood pressure. Moreover, P-2091 was inversely associated with high diastolic blood pressure in this study, whereas P-2091 was positively associated with pregnancy-induced hypertension in our previous study [5]. Thus, the significance of the HDP-associated peptides in adult men was completely different from that in pregnant women with HDP. In addition, the relationship between the seven HDP-associated peptides and pregnancy-induced hypertension is thought not to be confounded by general hypertension. P-2091 showed significant relations with diastolic blood pressure and mean blood pressure but not with systolic blood pressure. Possible reasons for this dissociation of the results are the small number of subjects in this study and the greater variability in systolic blood pressure than that in diastolic blood pressure [24]. Therefore, further studies are also needed to test the relation of P-2091 peptide with systolic blood pressure in a larger cohort.

Age, history of smoking and alcohol consumption, and history of medication therapy for hypertension and dyslipidemia were possible confounders for the relationships of the peptides with each cardiovascular risk factor. However, the associations of the above three peptides with blood pressure and blood lipids were not altered in multivariable analysis after adjustment for these possible confounders, indicating that the associations were independent of age, smoking, alcohol consumption and therapy for hypertension and dyslipidemia. When using eGFR as an additional explanatory variable or covariate in multivariable analyses, the associations between P-2091 and diastolic blood pressure, between P-2378 and HDL cholesterol, and between P-3156 and triglycerides were not altered (data not shown). Thus, the relationships between HDP-associated peptides and cardiovascular risk factors are thought not to be confounded by renal function.

We also investigated the relationships of the seven peptide levels with lipid-related indices and metabolic syndrome. The ratio of LDL cholesterol to HDL cholesterol (LDL-C/HDL-C ratio) is a classic atherogenic index and is useful for the prediction of cardiovascular risk [25]. The ratio of triglycerides to HDL cholesterol (TG/HDL-C ratio) has also been shown to be associated with cardiovascular disease and metabolic syndrome [26, 27] and to be more useful than LDL cholesterol alone for the prediction of cardiovascular disease [28]. CMI, defined as the product of the TG/HDL-C ratio and the waist-to-height ratio [17], has been shown to be associated with arterial stiffness in a general population [29] and the progression of atherosclerosis in patients with peripheral arterial disease [30]. In the present study, the level of P-2378 was significantly associated with the LDL-C/HDL-C ratio, and the level of P-3156 was significantly associated with the TG/HDL-C ratio and CMI. These results suggest that P-2378 and P-3156 are useful discriminators for cardiovascular risk, and the results are reasonable since P-2378 was associated with HDL cholesterol and P-3156 was associated with the waist-to-height ratio and triglycerides. Metabolic syndrome is a cluster of conditions that occur together and increase the risk of cardiovascular disease and diabetes; these conditions include high blood pressure, high blood glucose, visceral obesity and dyslipidemia [19]. However, none of the peptides tested showed associations with metabolic syndrome in the present study. Although the reason for this negative finding is unknown, one possible explanation is the small size of the subject population in this study. In addition, there was no association between the levels of each peptide and hemoglobin A1c, which reflects glycemic status, a component of metabolic syndrome.

Although the reasons for the associations of HDP-related peptides with cardiovascular risk factors are unknown, there are possible explanations for them: P-2378 is derived from C4, which is an acute-phase protein and has been shown to be associated with metabolic syndrome [31]. Since HDL cholesterol is included in the criteria of metabolic syndrome, the above finding is in agreement with the results of the present study showing that the level of P-2378 was inversely associated with HDL cholesterol, although P-2378 was not significantly related to metabolic syndrome. Moreover, blood C4 levels have been shown to be inversely correlated with HDL cholesterol in adolescents [32]. P-3156, which was inversely correlated with triglycerides in the present study, is derived from ITIH4. Polymorphism of the ITIH4 gene has been reported to be associated with hypercholesterolemia [33]. In a proteomics study, ITIH4 was shown to be associated with atherosclerotic plaque in the human carotid artery [34]. However, to the best of our knowledge, the relationship of ITIH4 with triglycerides is unknown. P-2091 is a fragment of fibrinogen-α, and interestingly, fibrinopeptide B has been reported to show a vasocontractile effect [35] and a facilitating effect on the proliferation of vascular smooth muscle cells [36]. P-2091 is a fragment of fibrinogen-α, while fibrinopeptide B, a fragment of fibrinogen-β, has been reported to have vasocontractile action [35]. Therefore, it is unlikely that P-2091 is associated with the regulation of arterial tonus, although further studies are needed to determine whether fibrinogen-α-related peptides, including P-2091, are involved in the modulation of blood pressure.

This study has some limitations. The number of subjects was small, and the method for statistical analysis was therefore limited to correlation analysis and comparisons of means of two groups. However, in multivariable linear regression analysis in the present study, reasonable results related to cardiovascular risk factors, which had already been confirmed by previous studies [37, 38], were obtained: blood pressure was positively correlated with age and BMI (Table 5); HDL cholesterol was positively and inversely correlated with alcohol consumption and smoking, respectively (Table 6); and triglycerides were positively correlated with smoking (Table 6). This study is exploratory, and the significance of the seven HDP-associated peptides in men is unknown. Thus, we did not consider multiplicity regarding the seven HDP-associated peptides in this study. We confirmed that statistical significance was maintained for the correlations between P-2091 and diastolic blood pressure, between P-2378 and HDL cholesterol, and between P-3156 and triglycerides when Bonferroni correction was performed in univariable correlation analysis regarding the seven HDP-associated peptides. P-2081, P-2127 and P-2209 are fragments from kininogen-1 and have been shown to be associated with HDP [5]. However, these peptides were not associated with blood pressure or other cardiovascular risk factors, including blood lipids and adiposity, in the present study. On the other hand, P-2081 and P-2209 showed relatively strong correlations with P-3156 (Table 7), which was associated with triglycerides. Thus, these kininogen-related peptides were indirectly associated with triglycerides. A possible reason for the lack of direct associations of these peptides with triglycerides is the small sample size in this study, and further studies using a larger population are needed to clarify the relationships of P-2081 and P-2209 with blood lipids. The purpose of this study was to explore the significance of HDP-associated peptides in relation to cardiovascular risk factors, including hypertension, in the general population. Female hormones potently affect blood pressure and other cardiovascular risk factors [39] and thus may confound the relationships of the peptides with cardiovascular risk factors in women. Therefore, we chose healthy male workers as the subjects of this study. However, it would also be interesting to investigate peptide levels in nonpregnant women with hypertension in future studies. Since the seven peptides were identified in the serum of women with HDP and the subjects of this study were Japanese men, further studies are also needed to investigate the relationships of the peptides with cardiovascular risk in women and persons of other ethnicities. Age, BMI, histories related to lifestyle factors such as smoking and alcohol consumption and a history of medication therapy for hypertension or dyslipidemia were adjusted in multivariable analysis. However, there are other possible confounding factors, including physical activity, diet, nutrition and socioeconomic factors (e.g., education and occupation), for which information was not available in this study. In our preliminary study, we confirmed comparable profiles of HDP-associated peptides in the serum and plasma of normal pregnant women, nonpregnant women and healthy men [5]. Thus, although we did not measure each peptide level in plasma in the present study, our findings would not be altered when plasma, instead of serum, was used as a sample for peptide measurement. The three peptides that showed associations with blood pressure and blood lipids are fragments derived from fibrinogen, C4 and ITIH4, for which serum concentrations were unfortunately unavailable in this study. The process of proteolysis is complicated, and thus, it is difficult to know how blood levels of fragments from each protein reflect levels of the parent protein. It would be interesting to know the relationships of P-2378 and P-3156 levels with HDL cholesterol and triglyceride levels in pregnant women. However, information on these lipid levels was not available in our previous study. Instead, we analyzed relationships of the peptide levels with serum total cholesterol levels, for which information was available in both the previous study and the present study: in both pregnant women and adult men, P-2378 and P-3156 levels were not significantly correlated with total cholesterol levels (data not shown). Finally, since this study was cross-sectional, further prospective studies are needed to explore causal relationships between the peptides and cardiovascular risk factors.

In conclusion, among the seven peptides related to HDP, three peptides were associated with cardiovascular risk factors in adult men. The peptides P-2091, P-2378 and P-3156 were inversely associated with diastolic blood pressure, HDL cholesterol and triglycerides, respectively. Therefore, these peptides are proposed as possible biomarkers for discriminating cardiovascular risk, although their clinical implications remain to be clarified.