Abstract
Previous studies in patients with hypertension have demonstrated that there is a U‐shaped association between HDL‐C (high‐density lipoprotein cholesterol) and the risk of cardiovascular events in male patients with hypertension. However, to the best of our knowledge, the relationship between HDL‐C and intensive blood pressure control in specific cardiovascular events has never been investigated. To fill this knowledge gap, the authors analyzed the relationship between HDL‐C levels and cardiovascular events in hypertensive patients within the Systolic Blood Pressure Intervention Trial (SPRINT). The SPRINT evaluated the impact of intensive blood pressure control (systolic blood pressure < 120 mm Hg) versus standard blood pressure control (systolic blood pressure < 140 mm Hg). The Cox proportional risk regression was used to investigate the association between different HDL‐C status and clinical outcomes. Additional stratified analyzes were performed to evaluate the robustness of sex difference. A total of 9323 participants (6016 [64.53%] males and 3307 [35.47%] females) with hypertension from the SPRINT research were included in the analysis. The median follow‐up period was 3.26 years. Our population was divided into five groups based on the HDL‐C plasma levels: HDL‐C < 30 mg/dL, HDL‐C between 30 and 40 mg/dL, HDL‐C between 40 and 60 mg/dL, HDL‐C between 60 and 80 mg/dL and HDL‐C > 80 mg/dL. Sensitivity analyzes showed that in the SPRINT, women in the HDL‐C high population had a higher risk of mortality from all causes than men. In this cohort study, results suggest that patients with HDL‐C levels higher than 80 mg/dL had lower risk of SPRINT primary outcome, cardiovascular death, and stroke, but this study tested association, not causation. HDL‐C levels were associated with composite cardiovascular outcomes in male but not female patients. Our results demonstrated that in patients with hypertension, the association between HDL‐C and risk of cardiovascular events is L‐shaped.
Keywords: composite cardiovascular outcomes, high‐density lipoprotein cholesterol, hypertension, intensive blood pressure control, sex difference
1. INTRODUCTION
Hypertension is one of the main causes and risk factors for various heart, brain, kidney and vascular diseases, affecting the structure and function of these organs and ultimately leading to functional failure. As the global population ages, the incidence of hypertension is increasing, as is the socio‐economic burden. 1 , 2 , 3 Abnormal lipids often co‐exist with hypertension, which is an important risk factor for cardiovascular and cerebrovascular diseases. In fact, lipid metabolism disorder can exacerbate the autonomic nerve disorder in hypertensive patients. When the deposition of lipid in the blood vessel wall leads to damage of the intima of the artery, endothelial dysfunction, and the formation of atherosclerotic plaque, the change of vascular pressure compliance leads to the imbalance of autonomic nervous function and abnormal blood pressure regulation.
High‐density lipoprotein cholesterol (HDL‐C) levels have historically been inversely associated with increased cardiovascular disease (CVD) risk. 4 However, newer epidemiological and genetic investigations have suggested that increasing HDL‐C levels might not benefit all patients' cardiovascular outcomes. 5 , 6 , 7 For example, studies performed in populations free of cardiovascular disease 8 , 9 have proposed that very high HDL‐C levels can be associated with an increased mortality risk.
Previous studies found benefits of intensive blood pressure control, including a reduced risk of primary cardiovascular events and all‐cause mortality in the Systolic Blood Pressure Intervention Trial (SPRINT) and stroke in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) blood pressure trial. 10 , 11 However, it is unclear whether the protective effects of HDL‐C will change in patients receiving different antihypertensive treatment. In addition, a study of patients with hypertension showed a U‐shaped relationship between HDL‐C and cardiovascular risk. 12 Nevertheless, there are several limitations in the study, one is the relatively small number of individuals at the highest end of the HDL‐C concentration spectrum 12 the other is the lack of information on alcohol intake, which was linked to higher HDL‐C levels. 12 Using SPRINT data, we analyzed the relationship between HDL‐C levels and composite cardiovascular events, and explored the relationship between different blood pressure reduction strategies. We used SPRINT data to analyze the association between HDL‐C levels and composite cardiovascular events, and to explore whether there were differences between different antihypertensive treatments.
2. MATERIALS AND METHODS
2.1. Data source and study of population
We performed a secondary analysis of the SPRINT trial. Data were obtained from the National Institutes of Health Biologic Specimen and Data Repository Information Coordinating Center (https://biolincc.nhlbi.nih.gov/studies/sprint/). We obtained baseline demographic data, laboratory data, physical examination data, prior disease and medication history, and endpoints from the dataset provided by the SPRINT research group. The SPRINT research was conducted in 102 clinical sites in the United States and enrolled 9361 participants, 12 all of whom were randomly assigned to either the intensive treatment group (systolic blood pressure < 120 mm Hg) or standard treatment group (systolic blood pressure < 140 mm Hg). All the participants were at least 50 years old and had a systolic blood pressure of 130 mm Hg or higher. The populations included in the study must have had at least one of the following increased cardiovascular risk factors: clinical or subclinical cardiovascular disease (CVD), chronic kidney disease (defined as an estimated glomerular filtration rate of 20 to <60 mL/min/1.73 m2), Framingham 10‐year CVD risk score ≥15% on the basis of laboratory examination in the last 12 months, or age ≥75 years. Patients with type 2 diabetes mellitus, prior stroke, polycystic kidney, and participants with a standing systolic blood pressure of <110 mm Hg at baseline were excluded. The intervention was stopped after a median follow‐up of 3.26 years owing to a significant reduction in primary cardiovascular events and all‐cause mortality in the intensive treatment group compared with that in the standard treatment group. 12
2.2. Clinical outcomes
The primary outcome of our study was composite cardiovascular outcomes. Secondary outcomes included stroke and death attributable to CVD. The composite cardiovascular outcomes were the first occurrence of cardiovascular events after randomization, including myocardial infarction (MI), non‐MI acute coronary syndrome (non‐MI ACS), new‐onset stroke, heart failure, and death attributable to CVD. Clinical outcome definitions were previously published in the SPRINT protocol. 13
2.3. Statistical analysis
Analysis of variance, Mann–Whitney U test, and χ2 tests were used where appropriate to compare baseline characteristics between five categories of HDL‐C: 30 mg/dL or less, 30–40 mg/ dL, 40–60 mg/dL, 60–80 mg/dL, and greater than 80 mg/ dL. Continuous variables were reported as mean (SD) or median (IQR); categorical variables were reported as frequency (proportion).
Kaplan–Meier curves were generated for SPRINT primary outcome across HDL‐C categories. Cox proportional hazards models were used to compute hazard ratios (HRs) and 95% CIs for adverse outcomes (SPRINT primary outcome, cardiovascular death, and stroke). Individuals with HDL‐C levels ≤30 mg/dL formed the reference. Both unadjusted and adjusted HRs were computed. According to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) statement, Model 1 was adjusted for none; Model 2 was adjusted for age (<75 and ≥75 years of age), sex (female, male), ethnicity (HISPANIC, BLACK, WHITE, OTHER), and education; and Model 3 was further adjusted for age, sex, ethnicity, education, body mass index, smoking status (never smoked, former smoker, current smoker), alcoholic drinking status, baseline systolic blood pressure, baseline heart rate, glucose, total triglycerides, LDL cholesterol, aspirin use, statin use, number of distinct anti‐hypertensive agents. The severity of multicollinearity in Cox model was measured by the variance inflation factor (VIF). If the VIF is ≥5, then multicollinearity existed among variables. Subsequently, a two‐stage linear regression model was used to calculate the threshold effect of the relationship between HDL‐C levels and SPRINT primary outcome. We used a trial method to determine the threshold value by moving the trial turning point along a predefined interval and selecting the value that provided the maximum model likelihood. Next, we performed log‐likelihood ratio analysis comparing the one‐line linear regression model with the two‐piecewise linear model. We performed interaction and stratified analyses according to age (48–62 years, 63−72 years, and ≥73 years), CVD subgroup.
All analyzes were performed using the statistical software packages R (The R Foundation; http://www.R‐project.org) and EmpowerStats (X&Y Solutions, Inc., Boston, Massachusetts, USA; http://www.empowerstats.com). Statistical significance was set at p < .05.
3. RESULTS
3.1. Baseline characteristics of included hypertension patients
A total of 9323 participants with hypertension from the SPRINT research were included in the analysis. The median follow‐up period was 3.26 years. The mean age for all the participants was 67.92 ± 9.42 years; 3332 (35.59%) participants were females, 2802 (29.93%) were black, 1877 (20.05%) had a history of CVD, and 60 (3.34%) had a history of stroke. Individuals with an HDL‐C level greater than 80 mg/dL constituted 4.66% (434 of 9323) of this cohort and were likely to be older (mean [SD] age, 70.62 [9.53]), female (308 [70.97%]), with a higher prevalence of cardiovascular risk factors (mean [SD] systolic blood pressure (142.8 [16.31] mm Hg); mean [SD] heart rate (69.05 [12.70] bpm)) and a lower mean [SD] fast glucose levels (94.82 [13.26]mg/dL) compared with those with lower HDL‐C levels. Individuals with an HDL‐C level greater than 80 mg/dL also had lower median (IQR) triglyceride levels (69.00 [54.00–89.0] mg/dL; to convert to millimoles per liter, multiply by 0.0113) and higher mean (SD) LDL‐C levels (223.26 [36.77] mg/dL; to convert to millimoles per liter, multiply by 0.0259) (Table 1) than those with lower HDL‐C levels.
TABLE 1.
Characteristics of our population.
| HDL‐C level | ||||||
|---|---|---|---|---|---|---|
| Variable | ≤30 mg/dL (152) | 30–40 mg/dL (1517) | 40–60 mg/dL (5342) | 60–80 mg/dL (1878) | >80 mg/dL (434) | p‐value |
| AGE, mean (SD), years | 63.74 ± 9.05 | 66.17 ± 9.27 | 67.50 ± 9.29 | 70.18 ± 9.32 | 70.62 ± 9.53 | <.001 |
| Female, No. (%) | 13 (8.55%) | 221 (14.57%) | 1721 (32.22%) | 1044 (55.59%) | 308 (70.97%) | |
| Body mass index, mean (SD) | 31.68 ± 5.14 | 31.47 ± 5.54 | 30.30 ± 5.69 | 27.94 ± 5.44 | 26.29 ± 5.50 | <.001 |
| Race and ethnicity, No. (%) | <.001 | |||||
| HISPANIC | 97 (63.82%) | 979 (64.54%) | 3053 (57.15%) | 1017 (54.15%) | 240 (55.30%) | |
| BLACK | 29 (19.08%) | 321 (21.16%) | 1625 (30.42%) | 654 (34.82%) | 156 (35.94%) | |
| WHITE | 20 (13.16%) | 192 (12.66%) | 570 (10.67%) | 167 (8.89%) | 29 (6.68%) | |
| OTHER | 6 (3.95%) | 25 (1.65%) | 94 (1.76%) | 40 (2.13%) | 9 (2.07%) | |
| Fasting total cholesterol, mean (SD), mg/dL | 173.87 ± 56.02 | 175.30 ± 40.16 | 187.68 ± 40.09 | 202.67 ± 36.75 | 223.26 ± 36.77 | <.001 |
| Fasting glucose, mean (SD), mg/dL | 103.00 ± 16.08 | 102.42 ± 15.53 | 99.03 ± 13.20 | 95.86 ± 11.60 | 94.82 ± 13.26 | <.001 |
| Fasting total triglycerides, median (IQR), mg/dL | 226.50 (162.25–334.50) | 151 (111.00–209.00) | 109.00 (81.00–146.00) | 80.00 (63.00–108.00) | 69.00 (54.00–89.0) | <.001 |
| Aspirin use, N (%) | 70 (46.36%) | 806 (53.20%) | 2738 (51.38%) | 924 (49.28%) | 209 (48.16%) | .086 |
| Statin use, N (%) | 58 (38.93%) | 665 (44.04%) | 2420 (45.58%) | 757 (40.61%) | 146 (33.95%) | <.001 |
| Number of distinct anti‐hypertensive agents, N (%) | <.001 | |||||
| No anti‐hypertensives at baseline visit | 11 (17.74%) | 157 (28.65%) | 502 (24.50%) | 177 (22.84%) | 33 (17.84%) | |
| Anti‐hypertensive agents prescribed at baseline visit | 51 (82.26%) | 391 (71.35%) | 1547 (75.50%) | 598 (77.16%) | 152 (82.16%) | |
| Smoking status, N (%) | <.001 | |||||
| Never smoked | 55 (36.18%) | 612 (40.34%) | 2350 (43.99%) | 898 (47.82%) | 196 (45.16%) | |
| Former smoker | 55 (36.18%) | 678 (44.69%) | 2304 (43.13%) | 748 (39.83%) | 178 (41.01%) | |
| Current smoker | 41 (26.97%) | 225 (14.83%) | 682 (12.77%) | 230 (12.25%) | 60 (13.82%) | |
| Missing data | 1 (0.66%) | 2 (0.13%) | 6 (0.11%) | 2 (0.11%) | 0 (0.00%) | |
| Framingham 10‐year CVD risk, N (%) | 133 (87.50%) | 1172 (77.26%) | 3359 (62.88%) | 914 (48.67%) | 159 (36.64%) | <.001 |
| Previous CVD, N (%) | 35 (23.03%) | 380 (25.05%) | 1114 (20.85%) | 275 (14.64%) | 65 (14.98%) | <.001 |
| STROKE.TIA, N (%) | 5 (3.29%) | 42 (2.77%) | 184 (3.45%) | 63 (3.35%) | 17 (3.92%) | .705 |
| Baseline systolic blood pressure, mean (SD), mm Hg | 138.81 ± 16.26 | 137.56 ± 14.66 | 139.34 ± 15.43 | 141.65 ± 16.17 | 142.80 ± 16.31 | <.001 |
| Baseline heart rate, mean (SD), bpm | 67.02 ± 12.26 | 66.42 ± 11.73 | 65.88 ± 11.55 | 66.34 ± 11.26 | 69.05 ± 12.70 | <.001 |
Note: Values are mean ± SD or number and percentage. The p value indicates p for trend.
SI conversion factors: To convert total cholesterol to millimoles per liter, multiply by 0.0259; HDL‐C to millimoles per liter, multiply by 0.0259; LDL‐C to millimoles per liter, multiply by 0.0259; triglycerides to millimoles per liter, multiply by 0.0113.
Abbreviations: CVD, cardiovascular diseases; HDL‐C, high‐density lipoprotein cholesterol; IQR: interquartile range; LDL‐C, low‐density lipoprotein cholesterol.
3.2. Clinical outcomes
Key outcome events were identified in 16 (10.53%) ≤ 30 mg/dL, 110 (7.25%) ≤ 40 mg/dL, 318(5.95%) 40−60 mg/dL, 96(5.11%) ≤ 60 mg/dL, and 21 (4.84%) > 80 mg/dL. The between‐group differences were consistent across the components of the primary outcome and other prespecified secondary outcomes (Table 2). A total of 561 primary outcome occurred — 16 in the ≤30 mg/dL group and 21 in the > 80 mg/dL group. Decrease in primary outcome observed as HDL increases (p = .010) (Table 2).
TABLE 2.
Primary and secondary outcomes.
| HDL‐C level | ||||||
|---|---|---|---|---|---|---|
| ≤30 mg/dL (152) | 30–40 mg/dL (1517) | 40–60 mg/dL (5342) | 60–80 mg/dL (1878) | >80 mg/dL (434) | p‐value | |
| Primary outcome | 16 (10.53%) | 110 (7.25%) | 318 (5.95%) | 96 (5.11%) | 21 (4.84%) | .010 |
| Secondary outcomes | ||||||
| Myocardial infarction | 8 (5.26%) | 43 (2.83%) | 122 (2.28%) | 32 (1.70%) | 8 (1.84%) | .023 |
| Acute coronary syndrome | 3 (1.97%) | 18 (1.19%) | 43 (0.80%) | 16 (0.85%) | 0 (0.00%) | .089 |
| Stroke | 6 (3.95%) | 27 (1.78%) | 70 (1.31%) | 22 (1.17%) | 7 (1.61%) | .045 |
| Heart failure | 3 (1.97%) | 32 (2.11%) | 93 (1.74%) | 29 (1.54%) | 5 (1.15%) | .634 |
| Death from cardiovascular causes | 5 (3.29%) | 22 (1.45%) | 58 (1.09%) | 13 (0.69%) | 3 (0.69%) | .017 |
| Death from any cause | 10 (6.58%) | 68 (4.48%) | 203 (3.80%) | 60 (3.19%) | 22 (5.07%) | .074 |
| Primary outcome or death | 21 (13.82%) | 144 (9.49%) | 425 (7.96%) | 129 (6.87%) | 34 (7.83%) | .005 |
3.3. Association between HDL‐C levels and adverse outcomes
We constructed the Cox proportional hazard regression models to estimate the association between HDL‐C levels and adverse outcomes, respectively. The results of the three models were presented in Table 3. In unadjusted models, compared with the reference category with low HDL‐C levels ≤30 mg/dL, those with an HDL‐C level greater than 80 mg/dL had an expected lower risk of SPRINT primary outcome, cardiovascular deaths and stroke, even after adjustment for the aforementioned covariates (Model 1, primary outcome: HR, 0.47; 95% CI, 0.25–0.90; p = .0231; cardiovascular death: HR, 0.22; 95% CI, 0.05–0.94 ; p = .0409; stroke: HR, 0.42 ; 95% CI, 0.14–1.24; p = .1165; Model 2, primary outcome: HR, 0.39; 95% CI, 0.20–0.76; p = .0059; cardiovascular death: HR, 0.19; 95% CI, 0.04–0.84; p = .0279; stroke: HR, 0.26 ; 95% CI, 0.08–0.81; p = .0198; Model 3, primary outcome: HR, 0.45; 95% CI, 0.21–0.93; p = .0301; cardiovascular death: HR, 0.20; 95% CI, 0.04–0.97; p = .0457; stroke: HR, 0.23; 95% CI, 0.06–0.80; p = .0207). Intriguingly, this phenomenon cannot be found in individuals with HDL‐C levels (30–40 mg/dL). However, after adjustment, the group also had a decreased risk of stroke (Model 2, HR, 0.38; 95% CI, 0.16–0.92; p < .0326; Model 3, HR, 0.35; 95% CI, 0.14–0.89; p < .0276). Further, individuals with HDL‐C levels in the range of 40–60 mg/dL and 60–80 mg/dL also had decreased risk for adverse outcomes (primary outcome, cardiovascular death and stroke).
TABLE 3.
Association between HDL‐C levels and adverse outcomes.
| HDL‐C level | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 30–40 mg/dL | 40–60 mg/dL | 60–80 mg/dL | >80 mg/dL | ||||||
| ≤30 mg/dL | HR (95% CI) | p value | HR (95%CI) | p value | HR (95% CI) | p value | HR (95% CI) | p value | |
| EVENT‐PRIMARY | |||||||||
| Model 1 | 1.0 | 0.70 (0.41,1.17) | .1744 | 0.57 (0.34,0.94) | .0278 | 0.50 (0.29, 0.84) | .0092 | 0.47 (0.25, 0.90) | .0231 |
| Model 2 | 1.0 | 0.61 (0.36,1.03) | .0645 | 0.49 (0.30, 0.81) | .0058 | 0.41 (0.24, 0.70) | .0011 | 0.39 (0.20, 0.76) | .0059 |
| Model 3 | 1.0 | 0.65 (0.37, 1.13) | .1260 | 0.54 (0.31, 0.96) | .0320 | 0.47 (0.26, 0.86) | .0147 | 0.45 (0.21, 0.93) | .0301 |
| EVENT‐CVDDEATH | |||||||||
| Model 1 | 1.0 | 0.45 (0.17, 1.20) | .1115 | 0.34 (0.14, 0.85) | .0214 | 0.22 (0.08, 0.62) | .0043 | 0.22 (0.05, 0.94) | .0409 |
| Model 2 | 1.0 | 0.37 (0.14, 0.98) | .0465 | 0.28 (0.11, 0.70) | .0065 | 0.18 (0.06, 0.53) | .0017 | 0.19 (0.04, 0.84) | .0279 |
| Model 3 | 1.0 | 0.39 (0.14, 1.11) | .0775 | 0.30 (0.11, 0.83) | .0211 | 0.20 (0.06, 0.65) | .0077 | 0.20 (0.04, 0.97) | .0457 |
| EVENT‐STROKE | |||||||||
| Model 1 | 1.0 | 0.45 (0.19, 1.09) | .0783 | 0.33 (0.14, 0.76) | .0095 | 0.30 (0.12, 0.74) | .0091 | 0.42 (0.14, 1.24) | .1165 |
| Model 2 | 1.0 | 0.38 (0.16, 0.92) | .0326 | 0.26 (0.11, 0.60) | .0016 | 0.20 (0.08, 0.51) | .0007 | 0.26 (0.08, 0.81) | .0198 |
| Model 3 | 1.0 | 0.35 (0.14, 0.89) | .0276 | 0.23 (0.09, 0.60) | .0025 | 0.18 (0.06, 0.52) | .0015 | 0.23 (0.06, 0.80) | .0207 |
Note: Model 1: adjusted for none. Model 2: adjusted for age, female, race, education. Model 3: adjusted for age, female, race, education, BMI; baseline systolic blood pressure; heart rate; smoking status; alcoholic drinking status; glucose; triglycerides; LDL; statin use; aspirin use; number of distinct anti‐hypertensive agents.
Abbreviations: CI, confidence interval; HR, hazard ratio
3.4. HDL‐C levels as a continuous variable and composite cardiovascular outcomes
We then used a GAM nonlinear fit to estimate relative risk in the population, demonstrating a nonlinear L‐type relationship between HDL‐C levels and composite cardiovascular outcomes (Figure 1). As shown in Table 4, when we used HDL‐C levels as a continuous covariate, each 1 unit increase in HDL‐C levels decreased the risk of composite cardiovascular outcomes in total participants (HR, 0.99; 95% CI, 0.98−1.00; p < .0076) and male (HR, 0.98; 95% CI, 0.97−0.99; p < .0001). However, HDL‐C levels were not associated with primary outcome among female hypertensive patients (HR, 1.01; 95% CI, 0.99−1.02; p = .2817). Next, we used the two‐stage linear regression model to calculate the threshold effect. Table 4 shows the results of the two‐stage linear regression model. The inflection point was 50 mg/dL in all participants; on the left inflection point, the effect size, 95% CI, and p value were .97, .96−.99, and .0028, respectively. In female participants, HDL‐C levels were not associated with primary outcome in the two‐stage linear regression model or on‐line linear regression model. In male participants, when the HDL‐C levels were lower than 47 mg/dL, the risk of composite cardiovascular outcomes significantly decreased with HDL‐C levels increasing (HR, 0.97; 95% CI, 0.95−0.99; p = .0042). When the HDL‐C levels were greater than 47 mg/dL, the risk of composite cardiovascular outcomes also decreased, as increased HDL‐C levels (HR, 0.98; 95% CI, 0.97−1.00; p = .00319), although the downward trend was slower. The log likelihood ratio test was 0.046, and there was an interaction between the sex and HDL‐C levels (p for interaction = .0045).
FIGURE 1.
A L‐shaped association links HDL‐C (high‐density lipoprotein cholesterol) and the risk of cardiovascular events in patients with hypertension. The red line is the trend line and the blue line is the 95% confidence interval.
TABLE 4.
Results of two‐piecewise linear‐regression model.
| Male | Female | Total | |
|---|---|---|---|
| One linear‐regression model | 0.98 (0.97, 0.99) < 0.0001 | 1.01 (0.99, 1.02) 0.2817 | 0.99 (0.98, 1.00) 0.0076 |
| Inflection point (K) | 47 | 47 | 50 |
| <K Effect size β (95%CI) | 0.97 (0.95, 0.99) 0.0042 | 1.02 (0.96, 1.09) 0.4673 | 0.97 (0.96, 0.99) 0.0028 |
| >K Effect size β (95%CI) | 0.98 (0.97, 1.00) 0.0319 | 1.00 (0.99, 1.02) 0.4366 | 1.00 (0.99, 1.01) 0.5013 |
| Log likelihood ratio test | 0.292 | 0.580 | 0.046 |
Note: Two‐piecewise linear‐regression model was used to calculate the threshold effect of the HDL levels. If the log likelihood ratio test >0.05, it means the two‐piecewise linear regression model is not superior to the single‐line linear regression model.
According to Kaplan–Meier survival curves (Figure 2), the median survival between the two groups was different. Male patients lived longer than female patients (log‐rank test; p = .0014).
FIGURE 2.
Shown are the survival probability for the primary outcome (a composite of myocardial infraction, acute coronary syndrome, stroke, heart failure, or death from cardiovascular cause). p = .0014.
3.5. LDL‐C levels and composite cardiovascular outcomes
The association between LDL‐C levels and composite cardiovascular outcomes in hypertensive patients is shown in Table S1. In the unadjusted model, composite cardiovascular outcomes in the third quartile had the lowest risk (HR, 0.62; 95% CI, 0.49−0.79; p < .0001) as compared with those in the first quartile. Quartiles of LDL‐C levels as a continuous variable did not change this trend.
As shown in Figure S1, when we used LDL‐C levels as a continuous covariate and plotted the spline curves of the Cox regression models to estimate the relative hazard ratio in our population, evidencing a nonlinear U‐shaped association between LDL‐C levels and composite cardiovascular outcomes.
Next, we used the two‐stage linear regression model to calculate the threshold effect. Table S2 shows the results of the two‐stage linear regression model. In female participants, LDL‐C levels were associated with composite cardiovascular outcomes in the two‐stage linear regression model. When the LDL‐C levels were lower than 112 mg/dL, the risk of composite cardiovascular outcomes significantly decreased with LDL‐C levels increasing (HR, 0.98; 95% CI, 0.97−0.99; p = .0012). When the LDL‐C levels were greater than 112 mg/dL, the risk of composite cardiovascular outcomes also increased, (HR, 1.01; 95% CI, 1.00−1.01; p = .0291), although the downward trend was slower. The log likelihood ratio test was 0.001.
3.6. Subgroup analyses of outcomes by potential effect modifiers
Additional interaction and stratified analyses were performed to evaluate the robustness of the association between the different HDL‐C levels and risk of SPRINT primary outcome. Each subgroup analysis was adjusted for all factors in Model 3, except for the stratification factor itself. The results are shown in Figure 3. Generally, age and sex played an interactive role in the association between HDL‐C levels and composite cardiovascular outcomes among hypertensive patients (p = .0081 and p = .0045, respectively). HDL‐C levels were not associated with the risk of composite cardiovascular outcomes in female patients (HR: 1.01; 95%CI, 0.99−1.02; p = .2809).
FIGURE 3.
Subgroup analysis of the association between the different HDL‐C levels and risk of SPRINT primary outcome. Each subgroup analysis was adjusted for all factors in Model 3, except for the stratification factor itself.
4. DISCUSSION
In this post‐hoc analysis involving 9323 adults with hypertension but without diabetes from the SPRINT, our results indicated that those with very high HDL‐C levels (>80 mg/dL) have a lower risk of composite cardiovascular outcomes, independent of both traditional cardiovascular risk factors and alcohol consumption. Which uses Generalized Additive Model (GAM) to estimate the relative risk in the population, in this study, we demonstrated a non‐linear L‐type relationship between HDL‐C levels and composite cardiovascular outcomes.
The relationship between high HDL‐C and cardiovascular events has been explored in many observational studies in recent years and reported the increased risk of all‐cause death in patients with high HDL‐C levels. According to Hirata and coworkers 14 reported a significantly higher risk of atherosclerotic cardiovascular mortality in Japanese individuals with extremely high HDL‐C levels (≥90 mg/dL), while Hamer and coworkers 15 the increased hazard ratio for all‐cause mortality seen in patients with HDL‐C > 90 mg/dL does not appear to be explained by an increase in cardiovascular mortality. Another study, conducted in older people, found that a hazard ratio for all‐cause mortality for HDL‐C > 90 mg/dL was driven by both cardiovascular mortality and non‐cardiovascular mortality, and the results were similar. 16 The discrepancy between these data and ours may be explained by the different global cardiovascular risk of the study populations: general population versus patients with arterial hypertension, or by the inclusion in the cardiovascular events also of heart failure requiring hospitalization, incident coronary revascularization, angina, and atrial fibrillation at the time of their first presentation, which represent earlier evidence of atherosclerotic disease. A possible explanation for the association between high HDL cholesterol levels and cardiovascular risk is due to genetic variants. 17 Certain genetic variants may have some adverse effects causing elevated risk of disease and mortality, this is the case for some mutations in CETP, ABCA1, LIPC, and SCARB1, which are both associated with an elevated risk of coronary artery disease. 18 , 19 , 20 , 21 The observed associations in this study could also be an epiphenomenon where there is a pathophysiologic abnormality, perhaps genetic, which increases risk in ways we do not understand also influences HDL, suggesting that the physiology of HDL is complex and perhaps not well understood. Conformation and functional properties of the HDL particle may also be altered in individuals with extreme high HDL cholesterol. Another possible explanation includes differences in risk factors associated with both high HDL cholesterol and mortality. Although our multifactorially adjusted analysis included the most important risk factors for all‐cause mortality, 22 residual confounding cannot be discarded completely. Whether the association between extreme high HDL cholesterol and the risk of composite cardiovascular outcomes is causal is an important unresolved question in relation to the findings in this study.
Our results demonstrated that in patients with hypertension the association between HDL‐C and risk of cardiovascular events is L‐shaped. More recently, Valentino and coworkers 12 analyzed the relationship between HDL‐C levels and cardiovascular events in hypertensive patients as part of the Campania Salute Network in Southern Italy by showing that a non‐linear U‐shaped association between HDL‐C levels and cardiovascular health outcomes. Interestingly, the increased cardiovascular risk associated with high HDL‐C was not confirmed in female patients. However, this finding conflicts with the results of our study. This might be due to differences in the population of the participants. Valentino and coworkers 12 studied a group from a general population. Conversely, SPRINT enrolled an older cohort (mean age, 68 years), with 28% of participants 75 years of age or older, and also included participants with chronic kidney disease. This study only less than 3% of the participants had HDL‐C levels greater than 80 mg/dL, a much lower prevalence than the 7% participants reported in the study of general population in Canada, as well as approximately 8%–9% reported in the general population of northern China.
This highlights the complexity of the relationship between HDL‐C levels and CVD. HDL‐C levels may not be able to fully reflect the real cardioprotective effect of HDL. The association between very high HDL‐C levels and risk of mortality may be driven by heterogenous groups of HDL‐C particles, which may play critical roles beyond the measured levels of HDL‐C. Further, the function of HDL‐C particles could be converted from anti‐inflammatory to proinflammatory under certain circumstances, such as increased oxidative stress. 8 Future studies are warranted to identify specific mechanisms by investigating the functionality of various major HDL‐C apolipoprotein components.
In addition, we found a significant sex difference in the relationship between HDL‐C levels and adverse cardiovascular events. In male patients, high HDL‐C levels predicted a lower incidence of cardiovascular outcomes, whereas in female patients, HDL‐C levels were not associated with the incidence of adverse cardiovascular events. It is well known that HDL loses its cardiovascular protective function under certain conditions and may promote the development of atherosclerosis, particularly in postmenopausal women. 8 , 21 , 23 , 24 This may be related to the changes in hormone levels that occur following menopause. Several detrimental physiological changes affect the transition to menopause in women, including changes in sex hormones, deposition of body fat, and distribution of lipids. Over time, the accumulation of these changes may trigger a chronic inflammatory state that may hinder the cardioprotective ability of HDL. 16 , 25 Menopausal hormonal changes could lead to the accumulation of CVD risk factors. During the menopausal transition, the reduction of estradiol, an effective antioxidant, may increase lipid peroxidation and the formation of reactive oxygen species in women. Which in turn may affect the protein composition of HDL, thus depleting anti‐inflammatory and/or antioxidant proteins and enriching for pro‐inflammatory proteins. In vitro biochemical analysis of lipoprotein in a small subset of premenopausal and postmenopausal women (30 per group) demonstrated that postmenopausal women displayed an impairment in their ability to limit LDL oxidation, independent of the level of HDL. These results suggest that HDL particles may lose some of their anti‐atherosclerotic properties during the menopausal transition. 26 , 27 , 28 El Khoudary and coworkers reported that menopause may have differing effects on different HDL particles. They found that menopause does not affect the cardioprotective effect of small HDL particles, whereas large HDL particles are affected by menopause. They suggest that high HDL‐C levels in older women may be a marker of HDL dysfunction. It is possible that this is one of the reasons why HDL‐C loses its cardio protective effect in elderly women. 11
As for LDL‐C, in a study involving 27 533 women with an average follow‐up period of 17.2 years, Mora and coworkers revealed that relying on LDL‐C alone might overestimate or underestimate the risk in a subgroup in which LDL‐C was inconsistent with another LDL‐related measurement method (up to a quarter of the total population). Therefore, in our study, no statistically significant correlation between LDL‐C levels and the risk of composite cardiovascular outcomes was identified among patients with hypertension but without diabetes from the SPRINT. This may be attributable to the short follow‐up period or to the fact that the LDL‐C level itself is not a good predictor of composite cardiovascular outcomes.
4.1. Strengths and limitations
There are several key strengths in our study. First, the present study is the first to demonstrate the relationship between HDL‐C and intensive blood pressure control in populations who had long‐term follow‐up for both incident all‐cause and cardiovascular death events. Second, this study used longitudinally and repeatedly collected blood pressure data at the individual level before the incidence of CVD, which allowed us to directly evaluate the long‐term effect of blood pressure and HDL‐C on CVD risk. Third, we adjusted all available confounding factors and conducted several sensitivity analyses, especially the information on alcohol intake, which has been associated with higher HDL‐C levels.
This post‐hoc analysis also has some limitations. First, we did not determine in all patients cholesterol efflux capacity, 29 nor the HDL apolipoproteome, which has been recently associated to cardiovascular mortality in patients with coronary artery disease. 30 Second, this was a retrospective study, and the original study was not designed to examine the relationship between HDL‐C levels and adverse cardiovascular events. Finally, we analyzed the patients' baseline HDL‐C levels; although adjusting the LDL‐C levels in the analysis, we were still unable to control for all variables that might have influenced the results.
5. CONCLUSIONS
In this cohort study, results suggest that patients with HDL‐C levels higher than 80 mg/dL had lower risk of SPRINT primary outcome, cardiovascular death and stroke, but this study tested association, not causation. HDL‐C levels were associated with composite cardiovascular outcomes in male but not female patients. Our results demonstrated that in patients with hypertension the association between HDL‐C and risk of cardiovascular events is L‐shaped.
AUTHOR CONTRIBUTIONS
Rufei Liu wrote the paper. Rufei Liu applied for the database and made statistical analysis. Wenli Cheng was responsible for the revision of the paper. Both authors contributed to the article and approved the submitted version.
CONFLICT OF INTEREST STATEMENT
We have no completing interests to declare.
DATA AVAILABILITY OF STATEMENT
Publicly available datasets were analyzed in this study. This data can be found here: https://biolincc.nhlbi.nih.gov/studies/sprint/.
Supporting information
Supporting Information
ACKNOWLEDGMENTS
We thank Keyang Zheng from the Capital Medical University for assistance in developing the content of this article.
Liu R, Cheng W. Association between HDL‐C and intensive blood pressure control in patients with hypertension: A post‐hoc analysis of SPRINT. J Clin Hypertens. 2024;26:225–234. 10.1111/jch.14754
REFERENCES
- 1. Anstey KJ, Peters R, Clare L, et al. Joining forces to prevent dementia: the International Research Network On Dementia Prevention (IRNDP). Int Psychogeriatr. 2017;29:1757‐1760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Frankish H, Horton R. Prevention and management of dementia: a priority for public health. Lancet. 2017;390:2614‐2615. [DOI] [PubMed] [Google Scholar]
- 3. Mills KT, Stefanescu A, He J. The global epidemiology of hypertension. Nat Rev Nephrol. 2020;16:223‐237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Wilson PW, Abbott RD, Castelli WP. High density lipoprotein cholesterol and mortality. The Framingham Heart Study. Arteriosclerosis. 1988;8:737‐741. [DOI] [PubMed] [Google Scholar]
- 5. Silbernagel G, Schöttker B, Appelbaum S, et al. High‐density lipoprotein cholesterol, coronary artery disease, and cardiovascular mortality. Eur Heart J. 2013;34:3563‐3571. [DOI] [PubMed] [Google Scholar]
- 6. Angeloni E, Paneni F, Landmesser U, et al. Lack of protective role of HDL‐C in patients with coronary artery disease undergoing elective coronary artery bypass grafting. Eur Heart J. 2013;34:3557‐3562. [DOI] [PubMed] [Google Scholar]
- 7. Voight BF, Peloso GM, Orho‐Melander M, et al. Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study. Lancet. 2012;380:572‐580. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Madsen CM, Varbo A, Nordestgaard BG. Extreme high high‐density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. Eur Heart J. 2017;38:2478‐2486. [DOI] [PubMed] [Google Scholar]
- 9. Ko DT, Alter DA, Guo H, et al. High‐density lipoprotein cholesterol and cause‐specific mortality in individuals without previous cardiovascular conditions: the CANHEART study. J Am Coll Cardiol. 2016;68:2073‐2083. [DOI] [PubMed] [Google Scholar]
- 10. Cushman WC, Evans GW, Byington RP, et al. Effects of intensive blood‐pressure control in type 2 diabetes mellitus. N Engl J Med. 2010;362:1575‐1585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Wright JT, Williamson JD, Whelton PK, et al. A randomized trial of intensive versus standard blood‐pressure control. N Engl J Med. 2015;373:2103‐2116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Trimarco V, Izzo R, Morisco C, et al. High HDL (high‐density lipoprotein) cholesterol increases cardiovascular risk in hypertensive patients. Hypertension. 2022;79:2355‐2363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Williamson JD, Supiano MA, Applegate WB, et al. Intensive vs standard blood pressure control and cardiovascular disease outcomes in adults aged ≥75 years: a randomized clinical trial. JAMA. 2016;315:2673‐2682. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hirata A, Sugiyama D, Watanabe M, et al. Association of extremely high levels of high‐density lipoprotein cholesterol with cardiovascular mortality in a pooled analysis of 9 cohort studies including 43,407 individuals: the EPOCH‐JAPAN study. J Clin Lipidol. 2018;12. [DOI] [PubMed] [Google Scholar]
- 15. Li Z‐H, Lv Y‐B, Zhong W‐F, et al. High‐density lipoprotein cholesterol and all‐cause and cause‐specific mortality among the elderly. J Clin Endocrinol Metab. 2019;104:3370‐3378. [DOI] [PubMed] [Google Scholar]
- 16. Hamer M, O'Donovan G, Stamatakis E. High‐density lipoprotein cholesterol and mortality: too much of a good thing? Arterioscler Thromb Vasc Biol. 2018;38:669‐672. [DOI] [PubMed] [Google Scholar]
- 17. Motazacker MM, Peter J, Treskes M, et al. Evidence of a polygenic origin of extreme high‐density lipoprotein cholesterol levels. Arterioscler Thromb Vasc Biol. 2013;33:1521‐1528. [DOI] [PubMed] [Google Scholar]
- 18. Agerholm‐Larsen B, Nordestgaard BG, Steffensen R, Jensen G, Tybjaerg‐Hansen A. Elevated HDL cholesterol is a risk factor for ischemic heart disease in white women when caused by a common mutation in the cholesteryl ester transfer protein gene. Circulation. 2000;101:1907‐1912. [DOI] [PubMed] [Google Scholar]
- 19. Andersen RV, Wittrup HH, Tybjaerg‐Hansen A, et al. Hepatic lipase mutations,elevated high‐density lipoprotein cholesterol, and increased risk of ischemic heart disease: the Copenhagen City Heart Study. J Am Coll Cardiol. 2003;41:1972‐1982. [DOI] [PubMed] [Google Scholar]
- 20. Frikke‐Schmidt R, Nordestgaard BG, Jensen GB, Steffensen R, Tybjaerg‐Hansen A. Genetic variation in ABCA1 predicts ischemic heart disease in the general population. Arterioscler Thromb Vasc Biol. 2008;28:180‐186. [DOI] [PubMed] [Google Scholar]
- 21. Zanoni P, Khetarpal SA, Larach DB, et al. Rare variant in scavenger receptor BI raises HDL cholesterol and increases risk of coronary heart disease. Science. 2016;351:1166‐1171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Rehm J, Mathers C, Popova S, et al. Global burden of disease and injury and economic cost attributable to alcohol use and alcohol‐use disorders. Lancet. 2009;373:2223‐2233. [DOI] [PubMed] [Google Scholar]
- 23. Holmes MV, Asselbergs FW, Palmer TM, et al. Mendelian randomization of blood lipids for coronary heart disease. Eur Heart J. 2015;36:539‐550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Boden WE, Probstfield JL, Anderson T, et al. Niacin in patients with low HDL cholesterol levels receiving intensive statin therapy. N Engl J Med. 2011;365:2255‐2267. [DOI] [PubMed] [Google Scholar]
- 25. Landray MJ, Haynes R, Hopewell JC, et al. Effects of extended‐release niacin with laropiprant in high‐risk patients. N Engl J Med. 2014;371:203‐212. [DOI] [PubMed] [Google Scholar]
- 26. Lincoff AM, Nicholls SJ, Riesmeyer JS, et al. Evacetrapib and cardiovascular outcomes in high‐risk vascular disease. N Engl J Med. 2017;376:1933‐1942. [DOI] [PubMed] [Google Scholar]
- 27. Schwartz GG, Olsson AG, Abt M, et al. Effects of dalcetrapib in patients with a recent acute coronary syndrome. N Engl J Med. 2012;367:2089‐2099. [DOI] [PubMed] [Google Scholar]
- 28. Bowman L, Hopewell JC, Chen F, et al. Effects of anacetrapib in patients with atherosclerotic vascular disease. N Engl J Med. 2017;377:1217‐1227. [DOI] [PubMed] [Google Scholar]
- 29. Catapano AL, Pirillo A, Bonacina F, Norata GD. HDL in innate and adaptive immunity. Cardiovasc Res. 2014;103:372‐383. [DOI] [PubMed] [Google Scholar]
- 30. Lappegård KT, Kjellmo CA, Hovland A. High‐density lipoprotein subfractions: much ado about nothing or clinically important? Biomedicines. 2021;9(7):836. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information
Data Availability Statement
Publicly available datasets were analyzed in this study. This data can be found here: https://biolincc.nhlbi.nih.gov/studies/sprint/.