The influence of triglycerides on small dense low‐density lipoprotein cholesterol levels is attenuated in low low‐density lipoprotein‐cholesterol range: Implications for the negative results of the PROMINENT trial

Abstract

The Pemafibrate to Reduce Cardiovascular Outcomes by Reducing Triglycerides in Patients with Diabetes (PROMINENT) trial failed to show the preventive effects of pemafibrate, a triglyceride (TG)‐lowering drug, on atherosclerotic cardiovascular disease in patients with type 2 diabetes and dyslipidemia. We recently reported that TG‐lowering with pemafibrate did not decrease the calculated small dense (sd) low‐density lipoprotein cholesterol (LDL‐C), and speculated that the effect of TG on sdLDL‐C is attenuated in low LDL‐C levels. This report examined this possibility in 1,508 patients with type 2 diabetes and 670 healthy controls. LDL‐C ranges were classified as ≤69, 70–99, 100–139 and 140≤ mg/dL. The slope of the regression curve between sdLDL‐C and TG was found to flatten as LDL‐C decreased; 0.18, 0.13, 0.10 and 0.04 for controls, and 0.18, 0.13, 0.09 and 0.07 for diabetes patients. Correspondingly, the lower the LDL‐C range, the lower the sdLDL‐C/TG ratio. These results suggest that when LDL‐C is tightly controlled, TG‐lowering has only a weak inhibitory effect on sdLDL‐C.

Triglycerides and small dense LDL‐C stratified by LDL‐C.

INTRODUCTION

The Pemafibrate to Reduce Cardiovascular Outcomes by Reducing Triglycerides in Patients with Diabetes (PROMINENT) trial was recently published in The New England of Journal of Medicine ¹ , and failed to show the preventive effects of pemafibrate on atherosclerotic cardiovascular disease (ASCVD) in patients with type 2 diabetes and dyslipidemia. The present study was designed to determine whether pemafibrate, a triglyceride (TG)‐lowering drug, can reduce the incidence of ASCVD in hypertriglyceridemic and low high‐density lipoprotein cholesterol (HDL‐C) patients, whose low‐density lipoprotein (LDL)‐C is controlled <70 mg/dL with high‐intensity statins. Small dense (sd) LDL‐C is a powerful risk factor for ASCVD beyond LDL‐C ² , ³ , and its levels are positively correlated with TG levels ⁴ , ⁵ . Therefore, the TG‐lowering effect of pemafibrate would be expected to reduce sdLDL‐C levels. However, we recently reported that sdLDL‐C levels estimated by Sampson's equation ⁶ using TG and LDL‐C shown in the PROMINET trial were not reduced by pemafibrate treatment ⁷ . The fact that pemafibrate did not change sdLDL‐C levels might partly explain the negative results of this trial. We speculated that tight control of LDL‐C with statins might attenuate the effect of TG changes on sdLDL‐C levels. Therefore, we explored this possibility in diabetes patients and healthy controls to understand the limitation of TG‐lowering for ASCVD prevention.

METHODS

Patients with type 2 diabetes (n = 1,508) and healthy control participants (n = 670) were examined. Most patients with type 2 diabetes were treated with oral hypoglycemic agents or in combination with insulin (17%). Dyslipidemic patients were treated with statins (55%), ezetimibe (7%), fibrates (7%), or omega‐3 fatty acids (4%) alone or in combination. The healthy participants (n = 670) were staff members of Denka Co, Niigata, Japan. They were basically healthy and did not have diabetes, but hyperlipidemic individuals were allowed to enroll. Patients with severe hypertriglyceridemia (TG ≥500 mg/dL) were excluded. Plasma samples were taken in the morning after overnight fasting. sdLDL‐C was measured directly in plasma by our established homogeneous method ⁸ . LDL‐C, HDL‐C and apolipoprotein B (only available in diabetes patients) were measured by commercially available test kits. non‐HDL‐C was calculated by subtracting HDL‐C from total C. TG‐rich lipoprotein (TRL)‐C was calculated by subtracting non‐HDL‐C from total cholesterol. Continuous variables were expressed as the mean ± standard deviation. Differences between groups were examined by anova. A simple linear regression analysis was carried out to calculate the coefficient of determination (R ²) between the two variables. P‐value <0.05 was considered statistically significant.

The present study complied with the principal of the Declaration of Helsinki. The study was explained in detail to all individuals who agreed to participate, and written informed consent were obtained from all diabetes patients and healthy controls. This study was approved by the Ebina General Hospital Ethics Committee and by the internal regulations of Denka Co.

RESULTS

Table 1 lists the clinical characteristics and measurements of diabetes patients and healthy controls. Diabetes patients were older, and had higher body mass index, hyperglycemia and high prevalence of ASCVD, but lower estimated glomerular filtration rate than healthy controls. Diabetes patients showed higher TG and TRL‐C, and lower total‐C, LDL‐C, HDL‐C and non‐HDL‐C compared with healthy controls. A possible reason for the low cholesterol levels in diabetes patients is that more than half of these individuals were taking statins. Nevertheless, sdLDL‐C was significantly higher in diabetes patients than in healthy controls.

Table 1.

Clinical characteristics and measurements in patients with type 2 diabetes and healthy control participants

	Diabetes patients		Healthy controls
n (male : female)	1,508 (972:536)		670 (414:256)
	Mean	SD	Mean	SD	P
Age (years)	66	11	39	11	<0.0001
Body mass index (kg/m2)	25.3	4.2	22.4	3.4	<0.0001
Coronary artery disease, n (%)	162 (10)		0		<0.0001
Stroke, n (%)	119 (7.7)		0		<0.0001
Glucose (mg/dL)	149	40	82	8	<0.0001
HbA1c (%)	7.3	0.8	5.3	0.2	<0.0001
eGFR (mL/min/1.73 m2)	71	20	82	14	<0.0001
Lipids (mg/dL)
Total C	180	31	202	35	<0.0001
TG	126	73	92	65	<0.0001
Ln TG*	4.6	0.5	4.3	0.5	<0.0001
HDL‐C	54	14	61	16	<0.0001
TRL‐C	25	11	18	12	<0.0001
Non‐HDL‐C	126	28	140	35	<0.0001
LDL‐C	101	24	123	34	<0.0001
Small dense LDL‐C	31	14	28	16	0.0001
Apolipoprotein B	87	18	NA

^*Logarhymically transformed triglycerides. C, cholesterol; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; NA, not available; SD, standard deviation; TG, triglyceride; TRL, triglyceride‐rich lipoprotein.

Figure 1 shows the linear regression analysis between sdLDL‐C and TG, LDL‐C, non‐HDL‐C or apolipoprotein B in controls and diabetes patients (apolipoprotein B was only available in the diabetes group). The coefficient of determination (R ²) for TG, LDL‐C, non‐HDL‐C and apolipoprotein B was 0.499, 0.514 and 0.645, respectively, for controls, and 0.347, 0.293, 0.471,and 0.572 for diabetes patients, suggesting that the association between sdLDL‐C and conventional lipid parameters is rather weak in diabetes patients.

Figure 1.

Relationship between small dense low‐density lipoprotein cholesterol (sdLDL‐C) and conventional lipid parameters in healthy controls and diabetes patients.

Figure 2 shows the relationship between sdLDL‐C and TG stratified by LDL‐C range. LDL‐C range was designated as L1 ≤69, L2 = 70–99, L3 = 100–139 and L4 ≥140 mg/dL. The intercepts and slope steepness values dropped as the LDL‐C range decreased (L4, 3, 2 and 1): sdLDL‐C = 23.3 + 0.18 × TG, 13.9 + 0.13 × TG, 10.7 + 0.10 × TG and 8.8 + 0.04 × TG for controls, and sdLDL‐C = 23.3 + 0.18 × TG, 19.0 + 0.13 × TG, 15.0 + 0.09 × TG and 10.5 + 0.07 × TG for diabetes patients, respectively. This trend was observed when diabetes patients were divided into statin users and non‐users (Figure S1). To clarify the involvement of TG in sdLDL‐C levels in each LDL‐C range, the sdLDL‐C/TG ratio was calculated. The sdLDL‐C/TG ratio decreased significantly with lower LDL‐C range in both controls and diabetes patients (Figure 3).

Figure 2.

Relationship between small dense low‐density lipoprotein cholesterol (LDL‐C) and triglyceride (TG) stratified by LDL‐C range in healthy controls and diabetes patients.

Figure 3.

The small dense low‐density lipoprotein cholesterol (sdLDL‐C)‐C/triglyceride (TG) ratio stratified by LDL‐C range in healthy controls and diabetes patients.

DISCUSSION

Although many cohort studies and Mendelian randomized analyses have shown that TG is causally related to the development of ASCVD, the TG‐lowering drug pemafibrate failed to prevent ASCVD in patients with type 2 diabetes ¹ . The atherogenicity of TG is mainly explained by TRL remnants and sdLDL particles ⁹ . The former is exclusively controlled by TG levels, whereas the latter is also controlled by LDL levels ⁴ , ⁵ . The PROMENENT study showed that pemafibrate significantly reduced remnant cholesterol (calculated and measured) with a decrease in TG, but rather increased apolipoprotein B (a surrogate marker for LDL particle count) and LDL‐C levels ¹ . As shown in Figure 1, apolipoprotein B levels are strongly correlated with sdLDL‐C levels, so it is not surprising that the estimated sdLDL‐C levels were not reduced in the PROMINET trial. However, it remained unclear why TG‐lowering with pemafibrate did not lead to a decrease in sdLDL‐C.

In the present study, we found that the involvement of TG in sdLDL‐C levels is significantly reduced as LDL‐C levels decrease. Statin use/non‐use and diabetes/non‐diabetes did not significantly affect the relationship between sdLDL‐C and TG stratified by LDL‐C range, suggesting that LDL‐C level itself is an essential factor in determining sdLDL‐C. The PROMINENT trial is designed for diabetes patients with LDL‐C levels <70 mg/dL on high‐intensity statins. This rigorous control of LDL‐C might minimize the inhibitory effect of pemafibrate on sdLDL‐C production by lowering TG.

We reported that sdLDL‐C levels calculated by the Sampson's formula were more modestly associated with direct sdLDL‐C measurement in diabetes patients compared with non‐diabetes controls ¹⁰ .

The present study also confirmed a modest association between sdLDL‐C and lipid parameters in diabetes patients (Figure 1). Increased very LDL1 production and hepatic TG lipase activity are both involved in the generation of sdLDL particles ¹¹ , which are not necessarily associated with plasma TG or LDL‐C levels ⁹ . The production of sdLDL‐C in diabetes is complex, suggesting that TG‐lowering therapy alone cannot efficiently suppress sdLDL‐C levels. Yamashita et al. ¹² reported in a pooled analysis of phase II and III trials on the efficacy and safety of pemafibrate that pemafibrate treatment reduced the sdLDL fraction. The reason pemafibrate did not reduce calculated sdLDL‐C in the PROMINENT trial and reduced the sdLDL fraction in Yamashita's study might be that baseline LDL‐C concentrations in the phase II and III trials were much higher, >110 mg/dL, than the 70 mg/dL in the PROMINENT trial. Also, the PROMINENT trial used mainly strong statins, whereas Yamashita's study used mainly mild statins. A possible reason why lowering TG does not reduce sdLDL‐C when LDL‐C levels are very low is that TG is not sufficiently transported from TRL to LDL due to the limited number of LDL that are acceptors for TG.

There are other reasons that explain the negative results of the PROMINENT study. For example, the unexpectedly small decrease in TG with pemafibrate administration and the lack of a decrease in apolipoprotein B levels ¹ . Further subanalyses, including our speculations, are required to clarify these possibilities.

Lower LDL‐C levels attenuate the effect of TG on sdLDL‐C levels, which might explain in part why TG‐lowering with pemafibrate did not suppress ASCVD in the PROMINENT trial, in which LDL‐C levels were tightly controlled.

DISCLOSURE

Tsutomu Hirano receives advisor fees from Denka Co, and lecture fees from Kowa Co. Yasuki Ito is an employee of Denka Co. and declares no conflict of interest.

Approval of the research protocol: The study complied with the principal of the Declaration of Helsinki. This study was approved by the Ebina General Hospital Ethics Committee and by the internal regulations of Denka Co. This study was approved by the Ethics Committee of Ebina General Hospital (EGHCS no 20190911‐1). Approval date: 11 September 2019.

Informed consent: The study was explained in detail to all individuals who agreed to participate, and written informed consent was obtained from all diabetes patients and healthy controls.

Animal studies: N/A.

Registry: ViNA Cohort. no 20190911‐1.

Supporting information

Figure S1 | Relationship between small dense low‐density lipoprotein cholesterol and triglyceride stratified by low‐density lipoprotein cholesterol range in diabetes patients with/without statin use.