Predictive Value of Small Dense Low-density Lipoprotein Cholesterol and Remnant-like Particle Cholesterol for Cardiovascular Events in Chinese Elder Diabetes Mellitus Patients

are presented as mean mean SD. cholesterol; sdLDL-C, small dense low-density


Abstract
Background As a subcomponent of lipoprotein cholesterol (LDL-C), small dense LDL-C (sdLDL-C) have been suggested to be a better predictor of cardiovascular diseases(CVD). This research was to evaluate the predictive of the sdLDL-C in cardiovascular events (CVs) in Chinese elder type 2 diabetes mellitus(DM) patients.

Methods
Serum sdLDL-C measured by homogeneous method was compared in 386 consecutive type 2 DM patients between December 2014 and December 2016. Finally, 92 type 2 DM patients had CVs during the 48-month follow-up period. Receiver operating characteristic (ROC) curves were used for assess the predictive value of baseline parameters to major CVs.

Results
Ninety-two CVs occurred during the study period.The ROC curve manifested that sdLDL-C in the study population had a matchable discriminatory power (AUC for sdLDL-C was 0.7366, P = 0.003). In addition, kaplan-Meier event-free survival curves displayed a obvious increase of CVs risk for sdLDL-C ≧ 26 mg/dL (log-rank = 9.10,P = 0.003). This phenomenon had analogous results in patients who received statins at baseline (log rank = 7.336 P = 0.007).The study discovered that the increase in HbA1c, glucose, LDL-C, sdLDL-C, non-HDL-C and ApoB and the decrease in ApoA1 were obviously interrelated with heightened CVs risk through Cox regression analysis. Multivariate analysis demonstrated that the increase of sdLDL-C and HbA1c was obviously correlated with CVs. The results of the study indicated that sdLDL-C (per 10 mg/dL) was a increased risk for CVs in the multivariate model (HR 1.281, 95% CI 1.225-16.032; P < 0.01).

Conclusion
The consequences demonstrated that sdLDL-C was more effective than RLP-C in predicting the future CVs of Chinese elder type 2 DM patients Background Type 2 diabetes mellitus (DM) patients are more likely to have cardiovascular disease (CVD) in part owing to dyslipidemia characterized by elevated low-density lipoprotein cholesterol (LDL-C), small dense LDL-C(sdLDL-C) and remnant lipoprotein cholesterol (RLP-C) in serum [1,2]. Moreover, prior research had discovered that higher sdLDL-C and RLP-C levels are interrelated with elevated risk of CVD events [3][4][5].
Cross-sectional studies have shown that sdLDL-C concentration is closely related to the severity of cardiovascular disease and is independent of classic coronary risk factors [6].Moreover, another study had revealed that a higher RLP-C concertration as a risk factor for cardiovascular events (CVs) independent of other risk factors in diabetic patients [7].However, very little studies in the past have assessed the relationship between sdLDL-C and RLP-C and CVs in Chinese elder type 2 diabetes mellitus(DM) patients.
SdLDL-C levels could be easily detected through automated analysis [8]. But these methods have not yet been impressed on large number of type 2 DM patients, especially among the Chinese elder population.
The study attempted to investigate whether sdLDL-C can predict CVs in Chinese elder type 2 DM patients.  [9]. The body mass index (BMI), estimate of glomerular ltration rate (eGFR) and smoking( current smokers and at least one cigarette per day )were recorded. The diagnosis of hypertension was based on a hypertension or systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg, and/or use of antihypertensive medications within 2 weeks of enrollment [10]. The diagnostic criterion for dyslipidemia was de ned as the fasting srum lipid levels as follows LDL-C ≥ 140 (mg/dl), high-density lipoprotein cholesterol (HDL-C) < 40 (mg/dl) or triglyceride (TG) ≥ 150 (mg/dl) and/or the current use of lipid-lowering medication [11].

Subjects and study design
Exclusion criteria were : age ≥ 90 years (n = 2), presence of malignancy (n = 3), known thyroid disorders (n = 3), lost of blood examination data (n = 6), infectious disease (n = 7), lost during follow-up (n = 7) and severe hepatic and nephrotic disease (n = 4). Finally, 386 patients ( mean age of 72.7 ± 5.4 year, range from 65 to 86 years) were effectively included in the study. All parameters were measured for subjects yearly during the follow-up, Clinical and laboratory data was collated between September and November 2019. The endpoints were: (1)CVD death, (2) the date of the rst occurrence of CVs, and (3) the patient visit to the Suzhou Ninth People′s Hospital for the last time, Table 1 shows the CVs. Laboratory measurements The blood samples were collected after 12 hours of fasting in the morning. After collection, the samples were centrifuged immediately and stored at -80℃ until assay within the same day. High-sensitivity Creactive protein(hsCRP),HDL-C, Hemoglobin A1c (HbA1c), LDL-C, TG,, fasting blood glucose, apolipoprotein A-I ApoA-I , apolipoprotein B ApoB and lipoprotein (a) were detected through standard biochemical tests as earlier report [12]. The sdLDL-C [5] and RLP-C[13]were discovered by a detergentbased fully-automatic homogeneous method(Denka Seiken kit).

Statistical analyses
All statistical analyses were performed using the SAS9.1 solftware package. The Chi-square test was used to analyze classi cation variables. Wilcoxon test for abnormal distribution parameters and independent sample t-test was used for normal distribution parameters to compare the baseline characteristics of CVs group and non-CVs group. Correlation coe cients between sdLDL-C and various parameters were ditermind by Spearman's rank analysis. Kaplan-meier method was used to compare the occurrence of CVs above or below the median sdLDL-C level, differents were assessed with log-rank test.
The receiver operating characteristic (ROC) curves and area under the curve (AUC) were used for determining the ability of sdLDL-C, HbA1c, and RLP-C to predict CVs. Cox regression and multivariate Cox regression analysis were employed to regulate these independent predictors. All statistical analyses were double-tailed, statistically signi cant was considered at the level P < 0.05.

Results
This study included 269 males (69.7%) and 117 (30.3%) females. The age of these persons was 65 years or older. They received follow-ups ranging from 20 to 48 months, with an average of 28 months. At baseline the patients with CVs had obviously higher BMI and were more likely to be under calcium channel blockers and insulin therapy compared with non-CVs patients ( Table 2). A comparison of laboratory ndings (Table 3) manifested the levels of glucose, HbA1c,LDL-C, non-HDL-C, sdLDL-C, RLP-C, ApoB and ApoA-1 in the CVs group were signi cantly different from those of the non-CVs group. There was no signi cantly difference in TG, hsCRP, HDL-C, eGFR, sdLDL-C /LDL-C ratioand lipoprotein (a) between the two groups.  In this study, rst-time CVs were observed in 92 patients. (Fig. 1a) displayed a obvious increase of CVs risk for the median levels of sdLDL-C. This phenomenon had analogous results in patients who received statins at baseline (Fig. 1b). Cox regression analysis showed that increase in sdLDL-C and HbA1c revealed a higher risk for CVs. However, the elevated RLP-C level was not (Table 4).To determine whether sdLDL-C was an independent risk factor, we performed cox multivariate regression analysis.The models were built after adjustment age, gender and CVs risk factors. Model 1 including Glucose, HbA1c, LDL-C, Non-HDL-C,sdLDL-C, ApoA-I and ApoB and Model 2 only including sdLDL-C and HbA1c showed that just sdLDL-C and HbA1c remained signi cantly associated with the risk of CVs These results suggest that elevated Glucose and dyslipidemia might contribute to CVs.  Spearman's correlation analysis between sdLDL-C and various parameters are shown in the Table 5. These result suggest that the level of sdLDL-C exhibited signi cant positive correlations with triglyceride and RLP-C than LDL-C. These result suggest that sdLDL-C might the major factor among LDL-C contribute to CVs. We performed the ROC analysis in order to test the discriminatory power of sdLDL-C for the cardiovascular events (Fig. 2). The result indicate that the AUC of the sdLDL-C has a strong discriminating power against CVs, and its optimal cut-off value is 36.2 mg/dL( AUC = 0.736, P = 0.003) than HbA1C and RLP-C.

Discussion
Diabetes mellitusis patients are often accompanied by dyslipidemia, which is the major contollable risk factor associated with cardiovascular disease(CVD) events. Dyslipidemia has been con rmed as one of the principal processes underlying CVD, while sdLDL-C is considered as an emerging risk factor for CVD. Indeed, about 70% of elder type 2 DM patients die from CVD [14][15][16].
To the best of our knowledge, few study have investigated on the in uence of sdLDL-C levels on the onset of CVs in Chinese elderly type2 DM patients.
Furthermore, previous studies concerning the elevated levels of RLP-C and sdLDL-C have been associated with CVs [17][18][19]. In the present study, Spearman's correlation analysis between sdLDL-C and various parameters are shown in the Table 5. These result suggest that the level of sdLDL-C exhibited signi cant positive correlations with triglyceride and RLP-C than LDL-C. These result suggest that sdLDL-C might the major factor among LDL-C contribute to CVs. We performed the ROC analysis in order to test the discriminatory power of sdLDL-C for the cardiovascular events (Fig. 2). The result indicate that the AUC of the sdLDL-C has a strong discriminating power against CVs, and its optimal cut-off value is 36.2 mg/dL( AUC = 0.736, P = 0.003) than HbA1C and RLP-C.
Moreover,kaplan-Meier event-free survival curve displayed a obvious increase of CVs risk for the median levels of sdLDL-C (Fig. 1a). This phenomenon had analogous results in patients who received statins at baseline (Fig. 1b). Cox regression analysis showed that increase in sdLDL-C and HbA1c revealed a higher risk for CVs. However, the elevated RLP-C level was not (Table 4).To determine whether sdLDL-C was an independent risk factor, we performed cox multivariate regression analysis.The models were built after adjustment age, gender and CVs risk factors. Model 1 including Glucose, HbA1c, LDL-C, Non-HDL-C,sdLDL-C, ApoA-I and ApoB and Model 2 only including sdLDL-C and HbA1c showed that just sdLDL-C and HbA1c remained signi cantly associated with the risk of CVs These results suggest that elevated Glucose and dyslipidemia might contribute to CVs.
Indeed, sdLDL-C levels has the ability to predict CVD better than total LDL-C [18].In addition, the Québec Cardiovascular Study has shown that sdLDL-C is interrelated with an raised risk of CAD in men. [6,7] On the other hand, remnant lipoproteins are rich in TG and the main components include VLDL in the fasting state [20].Obviously, the current study not only con rmed the sdLDL-C concentrations was an independent risk predictor for CVs, but also provided novel information concerning the role of RLP-C in predicting CVs in diabetic patients [21][22][23][24].
The limitations of this study mainly include: First, this cohort study is so small that it inevitably leads to a bias to fully observe the results and/or severity of CAD. Second, the effects of statin therapy needs further investigation. Finally, the comparison of the predictive ability of sdLDL-C for CVs to the patient subgroup needs further research. In future research, we will strive to solve these issues.

Conclusion
The consequences demonstrated that sdLDL-C was more effective than RLP-C in predicting the future CVs of elderly diabetic patients.