Lipoprotein ratios are better than conventional lipid parameters in predicting arterial stiffness in young men

Abstract

Although dyslipidemia is associated with cardiovascular disease, there are conflicting data about the role of serum lipids and their ratios in promoting arterial stiffness. The authors aimed to compare serum lipid profiles to predict arterial stiffness, which was assessed by brachial‐ankle pulse wave velocity in young Chinese men. A total of 1015 participants aged 18 to 44 years without serious comorbidities were recruited for conventional detection. Anthropometrics, brachial‐ankle pulse wave velocity, serum lipids, and other laboratory data were measured. Univariate analysis and multivariate logistic regression were performed to examine the relationship between serum lipid profiles and brachial‐ankle pulse wave velocity. Participants with high brachial‐ankle pulse wave velocity exhibited higher levels of total cholesterol, triglyceride (TG), low‐density lipoprotein cholesterol, non–high‐density lipoprotein cholesterol (HDL‐C), total cholesterol/HDL‐C, TG/HDL‐C, low‐density lipoprotein cholesterol/HDL‐C, and non–HDL‐C/HDL‐C. The subsequent multivariable logistic regression showed that TG/HDL‐C, total cholesterol/HDL‐C, non–HDL‐C/HDL‐C, and TG significantly increased the risk for arterial stiffness after adjustment for confounding factors. Results indicate that lipid ratios are superior to conventional lipid parameters for predicting arterial stiffness in young men and that the TG/HDL‐C ratio has the strongest association with arterial stiffness.

1. INTRODUCTION

Atherosclerosis, a progressive process responsible for most cardiovascular events, is a chronic disease that remains asymptomatic for decades. 1 Although the pathogenesis of atherosclerosis remains incompletely understood, dyslipidemia is thought to play an important role. Increased serum concentrations of total cholesterol (TC), triglyceride (TG), and low‐density lipoprotein cholesterol (LDL‐C) are recognized as risk factors for cardiovascular disease (CVD), whereas increased concentration of high‐density lipoprotein cholesterol (HDL‐C) is considered cardioprotective.2 However, several epidemiologic studies have reported the possibility that lipid‐related ratios such as TC/HDL‐C, TG/HDL‐C, and LDL‐C/HDL‐C may be better predictors of CVD risk than any other single lipid marker.3, 4, 5 Furthermore, the Adult Treatment Panel of the National Cholesterol Education Program has recommended using non–high‐density lipoprotein cholesterol (HDL‐C) as a secondary target of therapy and a study has demonstrated that non–HDL‐C is more closely associated with CVD compared with other lipid parameters.6, 7

Arterial stiffness occurs as a result of structural changes in the medial layer of the elastic arteries, including fragmentation and degeneration of elastin, increase in collagen, and thickening of the arterial wall, which are potentially related to the risk of development and progression of atherosclerosis.8 This process is known to begin in childhood.9 Thus, early detection of arterial stiffening before the development of clinical atherosclerosis may help to reduce CVD progression. Among the several noninvasive methods available to assess arterial stiffness, brachial‐ankle pulse wave velocity (baPWV), defined as the speed with which a pulse wave travels along the arterial length, has been broadly used as a marker of arterial stiffness in healthy volunteers as well as in patients with diabetes mellitus or hypertension and is considered a powerful predictor of CVD mortality and all‐cause mortality.10

Many factors such as age, hypertension, glycosylated hemoglobin, γ‐glutamyl transpeptidase (γ‐GTP), diabetes mellitus, and metabolic syndrome have been confirmed to be associated with arterial stiffness.11, 12, 13, 14, 15 However, findings about the relationship between serum lipid profiles and arterial stiffness have been inconsistent, and the superiority of lipid profiles was unconfirmed and questioned by different studies. One Chinese study reported that an increased level of LDL‐C was associated with arterial stiffness,16 but another study supported the use of non–HDL‐C as a superior predictor to LDL‐C in identifying arterial stiffness.17 In addition, studies on the use of lipid‐related ratios for prediction of arterial stiffness in young adults are still relatively lacking. Young men were chosen for this study because the relationship between arterial stiffness and lipid profiles may be sex specific, as suggested by a recent study.18 This population is relatively more vigorously engaged in social activities and less concerned about CVD. Thus, an evaluation of which factors are related to arterial stiffness in this population can be meaningful. Therefore, we performed a cross‐sectional study in apparently healthy young men to directly compare the association of serum lipids and their ratios with arterial stiffness as assessed by baPWV, with the goal of optimizing risk factors.

2. METHODS

2.1. Study participants

The survey population included 1137 men aged 18 to 44 years who visited the Health Examination Centre of Jiangmen Central Hospital (GuangDong, China) for a health checkup from July 2012 to August 2016. Information including age, height, weight, and history of previous diseases was collected. The following exclusion criteria were used: (1) history of diabetes mellitus or fasting glucose concentration ≥7.0 mmol/L, (2) history of hypertension or blood pressure ≥140/90 mm Hg, (3) active infection, (4) history of renal disease or creatinine ≥150 μmol/L, (5) CVD (defined as a clinical history or evidence on examination), and (6) active use of lipid‐lowering drugs. After exclusion of ineligible participants, a total of 1015 apparently healthy men were finally enrolled. This study was approved by the Hospital Ethics Committee. All participants provided written informed consent.

2.2. Data collection

All participants were assessed after overnight fasting for at least 10 hours. Blood pressure was measured using an automated sphygmomanometer (HBP‐9021, Omron, Tokyo, Japan) after the participants had rested for at least 5 minutes. Body mass index was calculated as body weight in kilograms divided by height in square meters. Fasting serum TC, HDL‐C, LDL‐C, TG, fasting glucose, aminotransferase, alanine aminotransferase, γ‐GTP, creatinine, and uric acid were measured by standard laboratory methods. Non–HDL‐C was calculated by subtracting HDL‐C from TC. Estimated glomerular filtration rate was calculated using the abbreviated equation from the Modification of Diet in Renal Disease study: 186.3 × (serum creatinine^−1.154) × (age^−0.203) × 0.742 (if female).19

2.3. Measurement of baPWV

baPWV was used to estimate arterial stiffness using a noninvasive vascular screening device (VP‐2000; Colin Co Ltd, Komaki, Japan). The participants were examined while in the supine position. Electrocardiographic electrodes were placed on both wrists, a microphone was placed on the left edge of the sternum to detect heart sounds, and pneumatic cuffs were placed on both arms and ankles. The participant's bilateral arm and ankle blood pressure and pulse volumes in the brachial and posterior tibial arteries were recorded. Although baPWV values were measured bilaterally, we used the higher value of baPWV for statistical analyses. A high baPWV was defined as the highest quartile of the values among the study participants; in the present study, the cutoff value was 1359 cm/s.

2.4. Statistical analysis

Continuous variables with normal distribution were expressed as mean±standard deviation, whereas continuous variables with skewed distribution were expressed as median (and interquartile range). The characteristics of the study sample with and without high baPWV were compared using the independent two‐sample t test for continuous variables with a normal distribution and the Mann‐Whitney U test for continuous variables with a skewed distribution. The comparison of the prevalence of high baPWV according to tertiles of lipid profile was done using chi‐square test. A receiver operating characteristic curve analysis was performed for each of the lipid profiles to compare the abilities of these measures to correctly discriminate high baPWV. The overall diagnostic accuracy was quantified using the area under the receiver operating characteristic curve and comparing the area under the receiver operating characteristic curves with the z statistic.20 Odds ratios (ORs) and 95% confidence intervals (CIs) for high baPWV were calculated using logistic regression analysis after adjustment for other covariates across tertiles of lipid profiles. P value for trend was determined by linear regression analysis. A two‐sided P<.05 was considered statistically significant. All statistical analyses were conducted using SPSS statistical software, version 16.0 (SPSS Inc, Chicago, IL).

3. RESULTS

The anthropometric and biochemical characteristics of the high baPWV and control groups are shown in Table  1. Age, systolic blood pressure, diastolic blood pressure, alanine aminotransferase, aminotransferase, γ‐GTP, and fasting glucose were greater in the high baPWV group compared with the control group (P<.05 for all). There were no significant differences between the groups in body mass index, uric acid, and estimated glomerular filtration rate. Participants with high baPWV had increased plasma levels of TC (P<.001), LDL‐C (P=.012), TG (P<.001), and non–HDL‐C (P<.001). There was no significant difference between the two groups in HDL‐C (P=.087). Compared with the control group, the high baPWV group had significant increases in all of the lipid ratio indices, including TC/HDL‐C (P<.001), TG/HDL‐C (P<.001), LDL‐C/HDL‐C (P=.004), and non–HDL‐C/HDL‐C (P<.001).

Table 1.

Comparison of clinical characteristic between the high baPWV and control groups

	High baPWV group (n=252)	Control group (n=763)	P value
Age, y	39 (35–42)	37 (33–41)	<.001
BMI, kg/m2	24.51±3.19	24.07±3.29	.064
Systolic BP, mm Hg	122.64±8.21	114.62±8.21	<.001
Diastolic BP, mm Hg	75.79±7.8	69.44±7.28	<.001
baPWVmax, mm/s	1437.5 (1398.25–1508)	1222 (1154–1280)	<.001
ALT, U/L	28.5 (20–42.75)	24 (17–35)	<.001
AST, U/L	29 (24–35)	27 (23–32)	.009
γ‐GTP, U/L	34.5 (24.25–53)	29 (21–44)	<.001
Uric acid, mg/dL	412.9 (356.83–473.43)	406.6 (355.3–467.4)	.555
eGFR	94.06 (85.13–106.31)	95.05 (84.27–107.42)	.765
Fasting glucose, mmol/L	5.17±0.5	5.07±0.46	.004
TC, mmol/L	5.52 (4.81–6.25)	5.22 (4.66–5.92)	<.001
LDL‐C, mmol/L	3.22±0.84	3.07±0.82	.012
TG, mmol/L	1.81 (1.22–2.62)	1.40 (0.99–2.08)	<.001
HDL‐C, mmol/L	1.22 (1.07–1.43)	1.27 (1.09–1.48)	.087
Non–HDL‐C, mmol/L	4.31±1.13	4.01±0.97	<.001
TC/HDL‐C	4.53±1.13	4.22±1.02	<.001
TG/HDL‐C	1.50 (0.92–2.16)	1.08 (0.72–1.75)	<.001
LDL‐C/HDL‐C	2.61±0.75	2.45±0.78	.004
Non–HDL‐C/HDL‐C	3.53±1.13	3.22±1.02	<.001

Abbreviations: γ‐GTP, γ‐glutamyl transferase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; baPWV, brachial‐ankle pulse wave velocity; BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride.

The prevalence of high baPWV according to tertiles of lipid profile in Table  2 and receiver operating characteristic analysis was performed to test the accuracy of each of the nine lipid profiles in predicting the presence of high baPWV. The area under the receiver operating characteristic curve values for TC/HDL‐C (OR, 0.581; 95% CI, 0.541–0.621 [P<.001]), TG/HDL‐C (OR, 0.614; 95% CI, 0.574–0.654 [P<.001]), LDL‐C/HDL‐C (OR, 0.564; 95% CI, 0.524–0.605 [P=.002]), and non–HDL‐C/HDL‐C ratios (OR, 0.581; 95% CI, 0.541–0.621 [P<.001]) were relatively higher compared with those for TC (OR, 0.571; 95% CI, 0.530–0.612 [P=.001]), TG (OR, 0.610; 95% CI, 0.570–0.0649 [P<.001]), LDL‐C (OR, 0.556; 95% CI, 0.514–0.597 [P=.008]), and non–HDL‐C levels (OR, 0.580, 95% CI, 0.540–0.621 [P<.001]), but the differences were not significant (all P>.05). For all variables of interest, the TG/HDL‐C ratio presented the largest area under the curve. However, HDL‐C (OR, 0.464; 95% CI, 0.423–0.505 [P=.087]) was also not significantly associated with high baPWV by receiver operating characteristic analysis.

Table 2.

Prevalence of high baPWV according to tertiles of lipid profile

	Tertiles of lipid profiles, %
	Tertile 1	Tertile 2	Tertile 3	P value for trend
Single lipid measures
TC	20.6	22.5	31.4	.001
TG	18.0	23.0	33.4	<.001
LDL‐C	18.5	28.3	27.7	.006
Non–HDL‐C	19.2	25.4	29.9	.001
HDL‐C	26.1	27.5	20.8	.115
Lipid ratios
TC/HDL‐C	17.5	26.9	30.2	<.001
TG/HDL‐C	16.3	24.6	33.6	<.001
LDL‐C/HDL‐C	20.1	26.0	28.3	.014
Non–HDL‐C/HDL‐C	17.5	26.6	30.4	<.001

Abbreviations: baPWV, brachial‐ankle pulse wave velocity; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride.

Finally, multiple logistic regression analysis was performed using the lipid profiles and seven other variables (age, alanine aminotransferase, aminotransferase, γ‐GTP, systolic and diastolic blood pressure, and fasting plasma glucose) with a significance level <0.05 in previous univariate analysis as independent variables. As shown in Table  3, the presence of high baPWV was significantly and independently predicted by TG/HDL‐C (OR, 1.872; 95% CI, 1.224–2.863), TC/HDL‐C (OR, 1.623; 95% CI, 1.075–2.449), non–HDL‐C/HDL‐C (OR, 1.633; 95% CI, 1.080–2.469), and TG (OR, 1.560; 95% CI, 1.020–2.386). However, we did not observe a significant association between high baPWV and TC (OR, 1.347; 95% CI, 0.899–2.019), LDL‐C (OR, 1.245; 95% CI, 0.828–1.871), HDL‐C (OR, 0.848; 95% CI, 0.570–1.262), non–HDL‐C (OR, 1.289; 95% CI, 0.853–1.946), and LDL‐C/HDL‐C (OR, 1.374; 95% CI, 0.919–2.054) after multivariate adjustment. When comparing all lipid parameters in terms of their strength of association with high baPWV, lipid ratios performed better overall than any of the individual variables used alone. The lipid parameter with the strongest association was TG/HDL‐C.

Table 3.

Odds ratios and 95% confidence intervals for high baPWV by tertiles of lipid profiles

	Tertiles of lipid profiles
	Tertile 1	Tertile 2	Tertile 3
Single lipid measures
TC	1	0.951 (0.630–1.436)	1.347 (0.899–2.019)
TG	1	1.031 (0.677–1.571)	1.560 (1.020–2.386)a
LDL‐C	1	1.409 (0.942–2.107)	1.245 (0.828–1.871)
Non–HDL‐C	1	1.103 (0.730–1.665)	1.289 (0.853–1.946)
HDL‐C	1	1.235 (0.843–1.811)	0.848 (0.570–1.262)
Lipid ratios
TC/HDL‐C	1	1.290 (0.855–1.947)	1.623 (1.075–2.449)a
TG/HDL‐C	1	1.316 (0.864–2.004)	1.872 (1.224–2.863)a
LDL‐C/HDL‐C	1	1.259 (0.840–1.886)	1.374 (0.919–2.054)
Non–HDL‐C/HDL‐C	1	1.281 (0.848–1.936)	1.633 (1.080–2.469)a

Abbreviations: baPWV, brachial‐ankle pulse wave velocity; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride. ^a P<.05 vs tertile 1.

4. DISCUSSION

Previous studies have indicated that lipids and the related ratios are strong predictors of arterial stiffness in selected populations, such as middle‐aged and elderly individuals and patients with essential hypertension and diabetes mellitus.21, 22, 23 However, most studies focused on only one ratio or traditional lipids. As an extension of previous studies, we directly compared serum lipid and lipid ratios as predictors of arterial stiffness measured by baPWV in young men. After adjusting for age, systolic and diastolic blood pressure, alanine aminotransferase, aminotransferase, γ‐GTP, and fasting glucose, the TG/HDL‐C ratio appears to be most closely associated with arterial stiffness, followed by non–HDL‐C/HDL‐C and then TC/HDL‐C and TG.

There have been several reports suggesting that traditional lipid ratios have greater predictive power for CVD risk than each of the single standard lipid measures.24, 25, 26 To date, TG/HDL‐C has become increasingly recognized as an important index of atherogenic particles. Our findings are consistent with some of the previous studies. The TG/HDL‐C ratio was originally proposed by Gaziano and colleagues27 as an atherogenic index and was proven to be a highly significant independent predictor of coronary heart disease, even stronger than TC/HDL‐C and LDL‐C/HDL‐C. Subsequent studies reported that TG/HDL‐C was strongly associated with predicting coronary artery disease (defined as stenosis >50%), coronary heart disease incidence, cardiovascular death, and total mortality.28, 29 Recently, Urbina and colleagues30 reported a relationship between TG/HDL‐C and arterial stiffness, measured by carotid‐femoral pulse wave velocity, in adolescents and young adults aged 10 to 26 years. Our findings further suggest that TG/HDL‐C was a better predictor of arterial stiffness than isolated lipid parameters and other traditional lipid ratios in apparently healthy young Chinese men. In addition, our findings were also consistent with previous reports which showed that in the general population, TG was independently associated with arterial stiffness but not non–HDL‐C, LDL‐C, or TC.

In theory, elevated serum TG may generate reactive oxygen species and induce insulin resistance, and thus influence several proatherogenic signaling pathways. Importantly, TG level is associated with smaller LDL particles, which are more atherogenic.31 Regarding vascular function, increased TG levels are accompanied by abnormal arterial smooth muscle and endothelial vasodilatory response.32 Conversely, HDL‐C is the major antiatherogenic lipoprotein. HDL stimulates the efflux of excess cellular cholesterol and facilitates reverse cholesterol transport to the liver.33 In addition, HDL also protects against atherosclerosis by inhibiting lipoprotein oxidation. The antioxidant properties of HDL are due, in part, to serum paraoxonase, an esterase carried on HDL that can degrade certain biologically active oxidized phospholipids.34 Based on these known mechanisms, it is no wonder that TG/HDL‐C, which indicates the proportion between the atherogenic and protective lipoproteins, has greater predictive power for assessing the extent of early‐stage atherosclerosis. Furthermore, TG/HDL‐C has been shown to have a stronger correlation with insulin resistance when compared with other traditional lipid or lipid ratios. In a cross‐sectional study of 6546 Korean adults who underwent routine health examinations, the TG/HDL‐C ratio was found to be significantly associated with insulin resistance in patients without metabolic syndrome.24 Insulin resistance has been found to play a significant role in the pathogenesis of CVD and was independently associated with baPWV even in healthy individuals.35 Insulin resistance causes increased assembly and secretion of TG‐rich very LDL and decreased HDL‐C.36 McLaughlin and colleagues37 suggested that TG/HDL‐C ratios were able to identify insulin‐resistant overweight individuals with normal glucose tolerance and are markers of insulin resistance with specificities and sensitivities similar to those for fasting plasma insulin concentration. Since the TG/HDL‐C ratio has been available as a simple clinical indicator of insulin resistance, and this ratio probably provides a prediction of arterial stiffening related to insulin resistance.

Earlier studies have indicated that TC/HDL‐C and non–HDL‐C/HDL‐C are risk indicators for CVD with greater predictive value than each parameter used independently in patients with coronary heart disease.3, 38 However, few studies investigated the effects of TC/HDL‐C and non–HDL‐C/HDL‐C ratios on early‐stage atherosclerosis in the general population. In the present study, the ratio of TC/HDL‐C or non–HDL‐C/HDL‐C, but not the isolated TC or non–HDL‐C, was found to be independently and positively associated with baPWV after adjustment for other non‐lipid risk factors. The results from our study show that the non–HDL‐C/HDL‐C ratio appears to be as useful as the TC/HDL‐C ratio (Table  3). Their similarity can be explained by the fact that non–HDL‐C is the difference between TC and HDL‐C, and both reflect the cholesterol in all lipoprotein particles that are considered to be atherogenic, such as very LDL, LDL‐C, and lipoprotein (a). Individuals with a higher ratio of TC/HDL‐C or non–HDL‐C/HDL‐C had greater cardiovascular risk owing to the imbalance between the cholesterol molecules carried by atherogenic lipoproteins and those carried by antiatherogenic lipoproteins. This may be due to an increase in the atherogenic constituents contained in the numerator, a decrease in the antiatherogenic constituents of the denominator, or both.39

LDL‐C transports cholesterol from the liver to peripheral tissues and promotes the foaming of macrophages via uptake within the arterial wall. The level of LDL‐C is used as a standard therapeutic target in the treatment of patients with CVD worldwide. Furthermore, Tamada and colleagues40 revealed that the serum LDL‐C/HDL‐C ratio was independently associated with increased carotid plaque score and may represent a useful marker for evaluating the extent of atherosclerosis in the early stages in the general population compared with LDL‐C alone. However, according to our study, after adjusting for potential confounders, LDL‐C or LDL‐C/HDL‐C failed to be an independent index for predicting arterial stiffness. The negative result may be attributable to the fact that our study population comprised relatively young healthy participants and excluded patients with CVD and diabetes mellitus. Additional studies with a larger sample size are needed to avoid possible bias. Furthermore, Brinkley and colleagues41 suggested that plasma oxidized LDL levels, a key player in the pathogenesis of atherosclerosis, might also play a role in arterial stiffening. Thus, oxidized LDL‐C may be one of the critical points in our future research.

5. STUDY LIMITATIONS

The present study has several limitations. First, it was a cross‐sectional study. We performed only association analysis, and no causal relationship between lipid profiles and the risk of arterial stiffness can be drawn. Second, although the sample size was relatively large and based on community‐dwelling individuals, participants were recruited at the time of their general health examination in only one location, which may not be a true representative of the general population. Third, although γ‐GTP has long been used as a marker of excessive alcohol intake,42 detailed information on alcohol consumption and smoking status of each individual was not available, which may confound our findings somewhat. Finally, the baPWV measurement describes the flexibility of both the aorta and medium arteries. Therefore, we could not separately evaluate pulse wave velocity of the central and peripheral arteries. However, baPWV, which can be measured more conveniently than aortic pulse wave velocity, has been reported to be well correlated with aortic pulse wave velocity assessed using a direct catheter method.10

6. CONCLUSIONS

Our results indicate that lipid ratios are superior to conventional lipid parameters for predicting arterial stiffness in young Chinese men, and the TG/HDL‐C ratio has the strongest association with arterial stiffness. As the TG/HDL‐C ratio is easily obtained and arterial stiffness may be modifiable, it seems reasonable to propose the TG/HDL‐C ratio as a potential target for intervention to reduce arterial stiffness in clinical practice.

DISCLOSURES

The authors declare that there are no conflicts of interest related to the subject matter or materials discussed in this article.

Wen J, Zhong Y, Kuang C, Liao J, Chen Z, Yang Q. Lipoprotein ratios are better than conventional lipid parameters in predicting arterial stiffness in young men. J Clin Hypertens. 2017;19:771–776. 10.1111/jch.13038