Abstract
Background
Although the clinical significance of lipidemic profiles has been reported in ischemic stroke, fewer studies reported the clinical significance of these lipidemic profiles in intracerebral hemorrhage. This study aimed to investigate the relationship between high‐density lipoprotein to low‐density lipoprotein ratios and long‐term clinical outcomes including functional dependence and mortality.
Methods
Based on a multicenter, real‐world registry, our study included patients with intracerebral hemorrhage hospitalized within 72 hours from symptoms onset. Divided into quartiles (Q1–Q4), the association between high‐density lipoprotein to low‐density lipoprotein ratios and poor functional outcomes (modified Rankin Scale scores 3–6) was investigated using multivariable logistic regression models.
Results
The final study included 869 intracerebral hemorrhage patients, with a median National Institutes of Health Stroke Scale score of 9 (3–16) and median hematoma volume of 13.4 mL (5.7–31). In the fully adjusted logistic model, where the high‐density lipoprotein to low‐density lipoprotein ratio was analyzed as a continuous variable, a higher ratio was significantly associated with an elevated odds of poor functional outcome at 90 days (adjusted odds ratio (OR), 4.57 [95% CI 1.70–12.29], P=0.003) and at 1 year (adjusted OR, 2.93 [95% CI 1.08–7.99], P=0.036). The sensitivity analyses confirmed that, compared with high‐density lipoprotein or low‐density lipoprotein separately, the high‐density lipoprotein to low‐density lipoprotein ratios emerged as a more robust and consistent predictor in patients with intracerebral hemorrhage without receiving surgical intervention.
Conclusions
Our study delineated the association between high‐density lipoprotein to low‐density lipoprotein ratios and poor functional outcomes in patients with intracerebral hemorrhage.
Keywords: cholesterol ratio, intracerebral hemorrhage, lipid profile, prognosis
Subject Categories: Intracranial Hemorrhage
Nonstandard Abbreviations and Acronyms
- ICH
intracerebral hemorrhage
- mRS
modified Rankin Scale
Clinical Perspective.
What Is New?
This multicenter study identifies the high‐density lipoprotein to low‐density lipoprotein ratio ratio as a novel, independent predictor of long‐term functional dependence after intracerebral hemorrhage.
What Are the Clinical Implications?
This ratio can serve as a practical biomarker for stratifying patients by their odds of poor outcome.
It may aid clinicians in identifying high‐risk individuals to guide personalized management strategies.
Compared with ischemic stroke, intracerebral hemorrhage (ICH) causes higher death and disability rates worldwide, 1 , 2 especially in Asian countries. 3 Evidence on medical or surgical treatment is not as sufficient as these in ischemic stroke. 4 , 5 , 6 Among patients with ischemic stroke, lipid‐lowering therapy is a cornerstone of secondary prevention alongside antiplatelet agents. 7 However, the role of lipidemic‐lowering drugs remained uncertain in ICH. 8 , 9 , 10 In a prior analysis using a national cohort, we evaluated the effects of lipid‐lowering therapies on ICH outcomes but found no clear association with functional recovery. 11
When we divided patients by prior history of dyslipidemia, we saw a different pattern. 11 Subgroup analyses, nevertheless, revealed that among patients without a preexisting diagnosis of dyslipidemia, statin treatment was associated with a significant reduction in in‐hospital mortality. This association was not evident in patients with established dyslipidemia. Moreover, statin therapy corresponded with a lower rate of hematoma evacuation in the group without dyslipidemia, whereas evacuation rates did not differ by statin use among those with dyslipidemia. These divergent findings may reflect differences in baseline lipid profiles. In ischemic stroke, lower low‐density lipoprotein cholesterol (LDL‐C) and higher high‐density lipoprotein cholesterol (HDL‐C) have been correlated with reduced risk of recurrent events by preventing arterial stenosis and stabilizing atherosclerotic plaques. 7 By contrast, the impact of lipid parameters on ICH prognosis has not been fully characterized. LDL‐C has been reported to be associated with hematoma expansion, neurological deterioration, 12 and 90‐day functional outcomes 13 (as well as non‐LDL parameter 14 ). Considering the potential neuroprotective effect of HDL, the combined index of LDL and HDL might be a superior marker. 15 , 16 HDL‐C/LDL‐C was reported to be correlated with lower myocardial infarction risk, all‐cause mortality, hemorrhagic stroke, and ischemic stroke. 17 A registry study reported the independent relationship between LDL‐C/HDL‐C and 90‐day functional outcomes 18 (including mortality 19 ), whereas its relationship with the 1‐year functional outcomes were unknown.
In this study, we investigate the association between HDL‐to‐LDL (HDL/LDL) ratio and functional outcomes in patients with ICH using data from a multicenter, prospective registry cohort. Besides, we also aim to determine the significance of HDL/LDL ratio in long‐term mortality in patients with ICH.
METHODS
The data that support the findings of this study are available from the corresponding author upon reasonable request. Before initiating patient enrollment, the study protocol received approval from the Institutional Review Board of Beijing Tiantan Hospital (IRB No. KY2014‐023‐02), and written informed consent was obtained from each patient or their legal representative. Between 2014 and 2016, we established a prospective registry of patients with ICH at 13 stroke centers in China, to describe the clinical practice in patients with ICH.
We enrolled consecutive adults (age ≥18 years) who presented with ICH within 72 hours of symptom onset. Exclusion criteria were (1) primary intraventricular hemorrhage, (2) incomplete modified Rankin Scale (mRS) data, or (3) incomplete lipidemic parameters data.
All clinical data were recorded prospectively by trained site investigators using deidentified case report forms. Before patient recruitment at each center, study physicians completed a standardized training program to ensure uniform data collection. We documented baseline demographics (age, sex), vascular risk factors (hypertension, diabetes, dyslipidemia, atrial fibrillation, prior stroke, and alcohol use), and pre‐ICH medications (antihypertensives, including α/β‐blockers, angiotensin‐converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, diuretics; antiplatelets, including aspirin, clopidogrel, dipyridamole, cilostazol; anticoagulants, including unfractionated heparin, low‐molecular‐weight heparin, warfarin; and statin agents). Neurological deficits at admission were quantified using the National Institutes of Health Stroke Scale, which ranges from 0 (no deficit) to 42 (severe impairment). 20 At discharge, we recorded length of stay and total and medication‐specific hospital expenditures. Fasting venous blood samples were obtained from the antecubital vein on the second morning following admission, after at least 8 hours of fasting. Levels of HDL‐C and LDL‐C were determined using conventional enzymatic assays.
Radiological assessments were performed by local neuroradiologists blinded to clinical outcomes. Hematoma volume was calculated using the ABC/2 method, 21 and pathogenetic subtypes of primary ICH were assigned based on combined clinical and imaging findings. Functional status was evaluated by mRS at 90 days and 1 year post ICH (blinded to baseline characteristics and prognostic factors), with higher scores indicating greater disability. 20 Primary study outcome was defined as the poor functional outcome (mRS score 3–6). Functional outcomes were primarily evaluated through in‐person interviews; when direct assessment was not feasible, structured telephone follow‐up was conducted. This longitudinal design allowed us to examine outcomes in relation to baseline clinical and radiological parameters.
Statistical Analysis
Continuous variables are presented as mean±SD or median (interquartile range) depending on their distribution. Categorical variables are reported as count (percentage). In univariate analyses, normally distributed data among multiple groups were compared using the 1‐way ANOVA test, and nonnormally distributed data were compared using the Kruskal–Wallis test. Categorical variables among multiple groups were compared using the χ2 test or Fisher’s exact test, as appropriate. We also calculated P values for trend to examine potential linear trend of baseline characters across HDL/LDL ratio categories.
Participants were categorized into quartiles (Q1–Q4) based on HDL/LDL ratio distribution. One multivariable logistic regression, where the HDL/LDL ratio was analyzed as a continuous variable, was used to evaluate the association between HDL/LDL ratio and the odds of poor functional outcome (mRS score 3–6) at 90 days and 1 year. The covariates in this multivariable logistic regression model included variates with P<0.1 in the univariable analyses and those known as key predictors of functional outcomes in ICH (age, National Institutes of Health Stroke Scale score, hematoma size, intraventricular hemorrhage, and hematoma location). Statin use at admission was also included as a covariate considering its impact on the lipid profiles. Before modeling, we assessed collinearity and tested for interactions. As a sensitivity analysis, this model was repeated after excluding patients who underwent surgical hematoma evacuation, or, considering the potential collinearity, performed to test the independent association of LDL or HDL without the inclusion of HDL/LDL ratio in the model among the full included cohort, separately.
We also generated Kaplan–Meier curves using Cox regression to compare 1‐year cumulative mortality and plotted receiver operating characteristic curves to evaluate the predictive performance of HDL/LDL ratio for poor functional outcomes at 90 days and 1 year. Statistical analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC), with 2‐sided P<0.05 indicating significance. Due to the potential type 1 error derived from multiplicity tests, the Benjamini–Hochberg method was applied to minimize the false positives (P‐false discovery rate [FDR]).
RESULTS
Baseline characteristics of the 869 enrolled patients with ICH are summarized by HDL/LDL ratio quartiles (Figure 1). The median age was 57 years (49–66), and 69.9% were male. On admission, the median National Institutes of Health Stroke Scale score was 9 (3–16), and the median hematoma volume was 13.4 mL (5.7–31). Hypertensive hemorrhage accounted for 91.2% of primary ICH cases. At 1 year, overall mortality was 13.5%, whereas 46.8% and 39.8% achieved mRS scores 3 to 6 at 90 days and 1 year, respectively (Table 1).
Figure 1. Flow‐chart.
HDL indicates high‐density lipoprotein; ICH, intracerebral hemorrhage; LDL, low‐density lipoprotein; and mRS, modified Rankin Scale.
Table 1.
Baseline Characteristics Across Different HDL/LDL Ratio Levels
| Overall | HDL/LDL ratio | HDL/LDL ratio | HDL/LDL ratio | HDL/LDL ratio | P value | P for trend | |
|---|---|---|---|---|---|---|---|
| Q1 (0.07–0.33) | Q2 (0.33–0.43) | Q3 (0.43–0.57) | Q4 (0.57–1.02) | ||||
| No.=869 | No.=217 | No.=217 | No.=217 | No.=218 | |||
| Age, y, median [interquartile range] | 57.0 [49.0–66.0] | 55.0 [45.0–63.0] | 57.0 [50.0–65.0] | 56.0 [48.0–65.0] | 62.0 [53.0–70.0] | <0.001 | <0.001 |
| Male sex, n (%) | 607 (69.9%) | 160 (73.7%) | 144 (66.4%) | 152 (70.0%) | 151 (69.3%) | 0.416 | 0.588 |
| National Institutes of Health Stroke Scale score at admission, median [interquartile range] | 9.0 [3.0–16.0] | 8.0 [3.0–16.0] | 9.0 [3.0–15.0] | 11.0 [4.0–20.0] | 9.5 [3.0–16.0] | 0.059 | 0.047 |
| Onset‐to‐admission intervals, h, median [interquartile range] | 3.8 [1.8–10.9] | 3.8 [1.8–14] | 4.0 [1.8–10] | 3.8 [1.7–11.5] | 3.6 [1.7–9.9] | 0.823 | 0.424 |
| Medical history | |||||||
| Hypertension, n (%) | 601 (69.3%) | 153 (70.5%) | 155 (71.8%) | 155 (71.4%) | 138 (63.6%) | 0.208 | 0.076 |
| Diabetes, n (%) | 133 (15.3%) | 40 (18.4%) | 38 (17.5%) | 26 (12.0%) | 29 (13.3%) | 0.174 | 0.085 |
| Dyslipidemia, n (%) | 77 (8.9%) | 21 (9.7%) | 25 (11.5%) | 16 (7.4%) | 15 (6.9%) | 0.292 | 0.141 |
| Atrial fibrillation, n (%) | 6 (0.7%) | 2 (0.9%) | 2 (0.9%) | 1 (0.5%) | 1 (0.5%) | 0.867 | 0.489 |
| Ischemic stroke, n (%) | 124 (14.3%) | 34 (15.7%) | 28 (12.9%) | 33 (15.2%) | 29 (13.3%) | 0.801 | 0.647 |
| ICH, n (%) | 24 (2.8%) | 2 (0.9%) | 4 (1.8%) | 10 (4.6%) | 8 (3.7%) | 0.077 | 0.057 |
| Alcohol abuse, n (%) | 344 (39.6%) | 91 (41.9%) | 86 (39.6%) | 80 (36.9%) | 87 (39.9%) | 0.926 | 0.688 |
| Pre‐mRS score, median [interquartile range] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.874 | 0.663 |
| Medication history | |||||||
| Antiplatelet agents, n (%) | 138 (15.9%) | 31 (14.3%) | 34 (15.7%) | 33 (15.2%) | 40 (18.3%) | 0.623 | 0.257 |
| Anticoagulant agents, n (%) | 7 (0.8%) | 1 (0.5%) | 2 (0.9%) | 3 (1.4%) | 1 (0.5%) | 0.821 | 0.929 |
| Antihypertensive agents, n (%) | 291 (33.5%) | 82 (37.8%) | 75 (34.6%) | 69 (31.8%) | 65 (29.8%) | 0.197 | 0.075 |
| Statin agents, n (%) | 51 (5.9%) | 11 (5.1%) | 9 (4.1%) | 12 (5.5%) | 19 (8.7%) | 0.199 | 0.054 |
| HDL/LDL ratio, median [interquartile range] | 0.4 [0.3–0.6] | 0.3 [0.3–0.3] | 0.4 [0.3–0.4] | 0.5 [0.5–0.5] | 0.7 [0.6–0.8] | <0.001 | <0.001 |
| LDL‐C levels, mmol/L, median [interquartile range] | 2.8 [2.2–3.4] | 3.6 [3.2–4.2] | 3.0 [2.6–3.6] | 2.6 [2.1–3.0] | 2.1 [1.8–2.4] | <0.001 | <0.001 |
| HDL‐C levels, mmol/L, median [interquartile range] | 1.2 [1.0–1.5] | 1.0 [0.9–1.1] | 1.1 [1.0–1.3] | 1.3 [1.1–1.5] | 1.5 [1.3–1.7] | <0.001 | <0.001 |
| Total cholesteral levels, mmol/L, median [interquartile range] | 4.6 [3.9–5.3] | 5.2 [4.7–6.0] | 4.7 [4.1–5.4] | 4.3 [3.7–5.0] | 4.1 [3.5–4.6] | <0.001 | <0.001 |
| Triglyceride levels, mmol/L, median [interquartile range] | 1.3 [0.9–1.8] | 1.7 [1.3–2.2] | 1.4 [1.1–1.8] | 1.1 [0.8–1.5] | 0.9 [0.7–1.2] | <0.001 | <0.001 |
| Systolic pressure at admission, median [interquartile range], mm Hg | 161.0 [147.0–184.0] | 160.0 [140.0–178.0] | 167.0 [150.0–187.8] | 161.0 [143.0–186.0] | 164.0 [149.5–186.5] | 0.006 | 0.128 |
| International normalized ratio, median [interquartile range] | 1.0 [0.9–1.0] | 0.9 [0.9–1.0] | 1.0 [0.9–1.0] | 1.0 [0.9–1.0] | 1.0 [0.9–1.0] | 0.462 | 0.163 |
| Serum creatinine, μmol/L, median [interquartile range] | 63.9 [52.8–76.5] | 64.7 [52.0–76.1] | 64.4 [52.9–80.0] | 61.8 [50.9–75.9] | 63.2 [54.0–72.5] | 0.372 | 0.322 |
| Statin at discharge, n (%) | 82 (96.5%) | 45 (97.8%) | 19 (100.0%) | 9 (81.8%) | 9 (100.0%) | 0.135 | 0.554 |
| Admission hematoma volume, [interquartile range], mL | 13.4 [5.7–31.0] | 11.8 [4.8–24.8] | 13.1 [5.8–28.1] | 17.7 [6.9–37.9] | 14.6 [6.6–30.0] | 0.030 | 0.01 |
| Surgical intervention, n (%) | 94 (10.8%) | 18 (8.3%) | 22 (10.1%) | 30 (13.8%) | 24 (11.0%) | 0.311 | 0.5 |
| Hematoma location | |||||||
| Lobar hematoma, n (%) | 195 (22.4%) | 47 (21.7%) | 49 (22.6%) | 49 (22.6%) | 50 (22.9%) | 0.990 | 0.777 |
| Basal ganglia hematoma, n (%) | 450 (51.8%) | 115 (53.0%) | 108 (49.8%) | 118 (54.4%) | 109 (50.0%) | 0.719 | 0.703 |
| Thalamus hematoma, n (%) | 155 (17.8%) | 47 (21.7%) | 35 (16.1%) | 33 (15.2%) | 40 (18.3%) | 0.301 | 0.554 |
| Brainstem hematoma, n (%) | 45 (5.2%) | 10 (4.6%) | 16 (7.4%) | 14 (6.5%) | 5 (2.3%) | 0.083 | 0.114 |
| Cerebellum hematoma, n (%) | 52 (6.0%) | 10 (4.6%) | 15 (6.9%) | 10 (4.6%) | 17 (7.8%) | 0.381 | 0.262 |
| Intraventricular hemorrhage, n (%) | 74 (8.5%) | 16 (7.4%) | 13 (6.0%) | 29 (13.4%) | 16 (7.3%) | 0.028 | 0.654 |
| Subtype of primary intracerebral hemorrhage, n (%)* | 0.071 | 0.066 | |||||
| Hypertensive | 717 (91.2%) | 179 (92.3%) | 186 (94.4%) | 175 (91.1%) | 177 (87.2%) | ||
| Cerebral amyloid angiopathy | 18 (2.3%) | 2 (1.0%) | 1 (0.5%) | 7 (3.6%) | 8 (3.9%) | ||
| Other/unknown | 51 (6.5%) | 13 (6.7%) | 10 (5.1%) | 10 (5.2%) | 18 (8.9%) | ||
| 90‐d mRS score, [interquartile range] | 2.0 [1.0–4.0] | 2.0 [1.0–4.0] | 2.0 [1.0–4.0] | 3.0 [1.0–4.0] | 3.0 [1.0–4.0] | 0.001 | 0.704 |
| 90‐d mRS score 3–6, (%) | 407 (46.8%) | 79 (36.4%) | 89 (41.0%) | 115 (53.0%) | 124 (56.9%) | <0.001 | <0.001 |
| Death at 1 y, n (%) | 117 (13.5%) | 20 (9.2%) | 24 (11.1%) | 35 (16.1%) | 38 (17.4%) | 0.033 | 0.006 |
| mRS score at 1 y, [interquartile range] | 2.0 [1.0–3.0] | 2.0 [1.0–3.0] | 2.0 [1.0–3.0] | 2.0 [1.0–4.0] | 2.0 [1.0–4.0] | <0.001 | 0.301 |
| mRS score 3–6 at 1 y, n (%) | 346 (39.8%) | 66 (30.4%) | 74 (34.1%) | 101 (46.5%) | 105 (48.2%) | <0.001 | <0.001 |
HDL‐C indicates high‐density lipoprotein cholesterol; ICH, intracranial hemorrhage; LDL‐C, low‐density lipoprotein cholesterol; and mRS, modified Rankin Scale.
Secondary patients with ICH were not accounted for.
Higher HDL/LDL ratios were significantly associated with older age and higher HDL‐C levels (P<0.001, P for trend <0.001), whereas LDL‐C, total cholesterol, and triglycerides were lower in groups with higher HDL/LDL ratios (all P<0.001). The P value for systolic blood pressure across quartiles was 0.006, indicating a significant between‐group difference but not a clear monotonic linear trend (P for trend 0.128). Those in lower HDL/LDL quartiles had better functional outcomes, demonstrated by lower median mRS scores (P=0.001 at 90 days; P<0.001 at 1 year) and lower rates of mRS scores 3 to 6 (P<0.001 for both time points) (Table 1).
In the final model selected by stepwise backwards elimination, a higher HDL/LDL ratio was significantly associated with an increased odds of poor functional outcome (mRS score, 3–6) at both 90 days and 1 year after ICH. At 90 days, each unit increase in the HDL/LDL ratio was associated with an adjusted odds ratio (OR) of 4.37 (95% CI, 1.70–12.29, P=0.003; FDR‐adjusted P=0.007) (Table 2). At 1 year, the association remained significant with an adjusted OR of 2.93 (95% CI, 1.08–7.99, P=0.036; FDR‐adjusted P=0.079) (Table 2). Although the association at 1 year remained statistically significant at the conventional 0.05 level (P=0.036), it did not survive strict FDR correction (P‐FDR=0.079), suggesting the relationship may be more robust in the short term. Use of statin agents was not significantly associated with outcome in the final model. These results indicate that a higher HDL/LDL ratio is independently associated with unfavorable functional recovery in patients with ICH, with a stronger effect size observed at 90 days than at 1 year. The distribution on mRS score also showed poor functional outcomes in higher HDL/LDL ratio groups, both at 90 days and 1 year (Figure 2).
Table 2.
Multivariable Logistics Regression Analyses
| 90‐d mRS score 3–6 | 1‐y mRS score 3–6 | |||||||
|---|---|---|---|---|---|---|---|---|
| Adjusted OR | 95% CI | P value | P‐FDR | Adjusted OR | 95% CI | P value | P‐FDR | |
| HDL/LDL ratio | 4.57 | 1.7–12.29 | 0.003 | 0.007 | 2.93 | 1.08–7.99 | 0.036 | 0.079 |
| Age | 1.05 | 1.04–1.07 | <0.001 | 1.07 | 1.05–1.09 | <0.001 | ||
| Admission hematoma volume | 1.02 | 1.01–1.03 | <0.001 | 1.02 | 1.01–1.03 | <0.001 | ||
| Statin agents at admission | 1.06 | 0.49–2.3 | 0.886 | 1.41 | 0.66–3.05 | 0.377 | ||
| ICH history | 2.9 | 0.91–9.29 | 0.073 | 2.48 | 0.78–7.85 | 0.222 | ||
| Hematoma location: supratentorial (basal ganglia, thalamus, cerebral lobes) | 0.96 | 0.39–2.38 | 0.932 | 1.08 | 0.44–2.66 | 0.873 | ||
| Hematoma location: infratentorial (cerebellum, brainstem) | 1.1 | 0.39–3.06 | 0.857 | 1.43 | 0.51–3.99 | 0.492 | ||
| Intraventricular hemorrhage | 1.65 | 0.78–3.48 | 0.192 | 1.26 | 0.61–2.63 | 0.533 | ||
| Admission NIHSS score | 1.19 | 1.15–1.22 | <0.001 | 1.17 | 1.14–1.21 | <0.001 | ||
The table shows the variables retained in the final model. The variables initially entered into the model included the following: systolic pressure at admission, ICH history, HDL/LDL ratio, total cholesterol levels, triglyceride levels, primary ICH cause, admission hematoma volume, statin agents use at admission, hematoma location (including the presence or absence of intraventricular hemorrhage), admission NIHSS score. HDL level and LDL level were not included in the models considering collinearity. FDR indicates false discovery rate; HDL‐C, high‐density lipoprotein cholesterol; ICH, intracranial hemorrhage; LDL‐C, low‐density lipoprotein cholesterol; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; and OR, odds ratio.
Figure 2. Distribution of mRS score between different HDL/LDL quartiles.
HDL indicates high‐density lipoprotein; LDL, low‐density lipoprotein; and mRS, modified Rankin Scale.
Sensitivity analyses, which excluded patients undergoing hematoma evacuation, were performed to further investigate the association between HDL/LDL ratios and functional outcomes. The results showed that a higher HDL/LDL ratio remained significantly associated with an increased odds of poor outcome at 90 days (adjusted OR, 4.02 [95% CI, 1.45–11.29], P=0.008). At the 1‐year follow‐up, the association remained statistically significant (adjusted OR, 2.87 [95% CI, 1.02–8.08], P=0.046). Additional analyses in the full cohort indicated that a higher HDL level was a significant risk factor at 90 days (OR, 1.75, 95% CI, 1.20–2.69, P=0.008), with a trend persisting at 1 year (OR, 1.36 [95% CI, 1.04–1.95], P=0.052). Conversely, a higher LDL level appeared to be protective at 90 days (OR, 0.83 [95% CI, 0.68–1.00], P=0.045), though this effect was attenuated at 1 year (OR, 0.87 [95% CI, 0.72–1.05], P=0.155). These findings collectively reinforce the value of the HDL/LDL ratio as an integrated marker of lipid metabolism in predicting prognosis, particularly in the short term after ICH (Table 3).
Table 3.
Sensitivity Analyses
| 90‐d mRS score 3–6 | P value | P‐FDR | 1‐y mRS score 3–6 | P value | P‐FDR | ||
|---|---|---|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | ||||||
| Excluding patients receiving surgical intervention | HDL/LDL ratio | 4.02 (1.45–11.29) | 0.008 | 0.016 | 2.87 (1.02–8.08) | 0.046 | 0.046 |
| All included patients | HDL level | 1.75 (1.20–2.69) | 0.008 | 0.016 | 1.36 (1.04–1.95) | 0.052 | 0.052 |
| All included patients | LDL level | 0.83 (0.68–1.00) | 0.045 | 0.090 | 0.87 (0.72–1.05) | 0.155 | 0.155 |
Same adjusted model as Table 2.
FDR indicates false discovery rate; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; mRS, modified Rankin Scale; and OR, odds ratio.
Over the 12‐month follow‐up, a Kaplan–Meier curve (Figure 3) showed that the survival rate was lower in the high HDL/LDL group compared with the low HDL/LDL group (log‐rank test, P=0.0061). However, Cox regression analysis with the same covariates as the previous logistic regression model also showed that HDL/LDL was not significantly associated with 1‐year mortality (hazard ratio [HR], 1.67 [95% CI, 0.64–4.32], P=0.294). To evaluate the predictive efficacy of HDL/LDL ratios for poor functional outcomes after ICH, we constructed receiver operating characteristic curves for both 90‐day and 1‐year mRS scores of 3 to 6 after adjusting for the same covariates as the previous logistic regression model. As shown in Figure 4, the area under the curve was 0.878 (95% CI, 0.85–0.906) for 1‐year mRS 3 to 6 and 0.877 (95% CI, 0.85–0.904) for 90‐day mRS 3 to 6, indicating good discriminatory ability (Figure 4).
Figure 3. Kaplan–Meier estimation of 1‐year survival rate by HDL/LDL quartiles.
HDL indicates high‐density lipoprotein; KM, Kaplan–Meier; and LDL, low‐density lipoprotein.
Figure 4. Receiver operating characteristic curves.
AUC indicates area under the curve; and mRS, modified Rankin Scale.
DISCUSSION
Our analysis of a multicenter, prospective registry demonstrated a robust association between the HDL/LDL ratio and long‐term functional dependence in patients with ICH. Although Kaplan–Meier analysis suggested differences in survival between HDL/LDL ratio groups, Cox regression adjusted for covariates did not confirm a significant association between HDL/LDL ratio and 1‐year mortality. Because HDL and LDL exert contrasting biological effects, a composite measure such as the HDL/LDL ratio may more accurately reflect their combined influence on ICH outcomes. In our cohort, patients with higher HDL/LDL ratios experienced significantly worse functional recovery.
The SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels) trial evaluated whether intensive lipid‐lowering reduces recurrent stroke in patients with ischemic stroke or transient ischemic attack. 9 Although the overall stroke rate decreased, the ICH risk rose in the high‐intensity statin arm, suggesting divergent statin effects in ischemic versus hemorrhagic stroke. Indeed, a SPARCL subgroup analysis revealed that higher HDL/LDL ratios were associated with lower ischemic stroke risk but higher hemorrhagic stroke risk. 22 Prior studies have linked low LDL levels to poorer ICH outcomes in both Asian 23 and White 24 cohorts, and our baseline LDL values were even lower than those reported in SPARCL. 22 However, these absolute differences do not fully explain the opposite prognostic implications of lipid metrics in the 2 stroke subtypes. A meta‐analysis of 39 studies confirmed that statins modestly increase ICH risk while reducing ischemic stroke risk, with no significant interaction by baseline LDL. 25
These observations are in line with prior work linking lipid metrics to ICH prognosis: Rodríguez‐Luna et al., 12 Chen et al., 13 and Feng et al. 14 reported associations between lower LDL (or non‐HDL) and worse short‐term outcomes or hematoma growth, and later registry analyses examined LDL/HDL (or HDL/LDL) ratios in relation to mortality and functional outcome. 18 , 19 However, retrospective and single‐center designs have predominated, frequently concentrating on short‐term outcomes and providing incomplete adjustment for clinically relevant confounders (eg, lipid‐lowering therapy, hematoma size and location). In contrast, our analysis leverages a multicenter prospective registry, includes rigorous adjustment for lipid‐lowering treatment and radiographic hematoma variables, evaluates both individual lipid components and the HDL/LDL composite, and performs sensitivity analyses to test whether the ratio’s association is driven predominantly by LDL or HDL. Therefore, our results complement and extend previous findings by providing more robust, clinically nuanced prospective evidence on the prognostic significance of lipid profiles in ICH.
A hypothesis on the relationship between HDL/LDL ratios and poor outcomes in ICH is the damage on blood–brain barrier. Imaging markers of cerebral small vessel disease, often due to disruption on blood–brain barrier, have been found to be associated with lipidemic parameters. Lower LDL correlates with greater white‐matter hyperintensity volume, 26 , 27 and with subcortical small‐vessel dementia, 28 suggesting that diminished LDL may compromise white‐matter integrity. A similar inverse relationship exists between LDL and cerebral microbleeds. 29 Supporting this, elevated HDL/LDL ratios were linked to more severe dilated perivascular spaces, especially in the basal ganglia 30 —the most frequent site of primary ICH.
Interestingly, LDL may confer neuroprotective effects. Higher LDL levels predict lower risk of cognitive decline, perhaps reflecting better nutritional status. 31 Given LDL’s limited permeability across the blood–brain barrier, 32 the damage of LDL on brain might be limited, compared with the certain harm on blood vessels in ischemic cerebrovascular diseases. 7 Imbalances in energy metabolism and inflammation may indicate heightened nutritional demands in neurological disease. 33 , 34 Moreover, oxidative stress inversely correlates with HDL/LDL ratios, 35 although higher ratios may paradoxically signal increased LDL oxidation via HDL‐derived hydroperoxides. 36 Hence, the relationship between HDL/LDL ratio and oxidative stress is still undetermined.
Based on a large sample size and long‐term follow‐up, our study confirmed the relationship between the HDL/LDL ratio and clinical outcomes in patients with ICH. We also conducted receiver operating characteristic and survival curves to support the clinical significance of the HDL/LDL ratio. However, several limitations warrant discussion. First, we collected laboratory data, including HDL and LDL levels, only at admission; we did not obtain HDL/LDL ratios at discharge or during follow‐up. Given that only a small proportion of patients with ICH received antilipidemic drugs, 11 fluctuations in lipid parameters after admission may have been minimal. Nevertheless, the association between longitudinal HDL/LDL ratios and clinical outcomes in patients with ICH remains to be investigated. Second, admission hematoma volume was calculated manually; despite standardized training, interobserver variability cannot be ruled out. Besides, some variables were not collected in our registry including respiratory failure, mechanical ventilation, intraventricular extension, obstructive hydrocephalus, diagnosis of thrombocytopenia or chronic kidney disease, and the exact time when HDL/LDL was sampled. Hence, the impacts of these variables on the relationship between HDL/LDL and functional outcomes in ICH were not investigated. Finally, the clinical practice of ICH might be improved in the future due to several randomized controlled trials on medication 37 and surgical 5 treatment strategies, limiting the generalization of the results from our study.
CONCLUSIONS
In conclusion, our study delineated a robust association between higher HDL/LDL ratios and poor functional outcomes in patients with ICH at both 90 days and 1 year, with the effect size being stronger at 90 days than at 1 year. Medical management on lipidemic parameters could be another potential therapeutic target of interests in ICH in the future.
Sources of Funding
This work was supported by grants from National Key R&D Program of China (2022YFC3501100; 2022YFC3501102), National Natural Science Foundation (82401452; 82371302; 82471489), Beijing Municipal Postdoctoral Work Funding Support (2024‐68‐167), Beijing Municipal Health Commission’s Excellence Clinical Research Program for Research Oriented Wards Project (BRWEP2024W022040100), Noncommunicable Chronic Diseases National Science and Technology Major Project (2023ZD0504700), Ministry of Finance of the People’s Republic of China [issued by Finance and Social Security [2015] Document No. 82; [2016] Document No. 50; [2017] Document No. 72; [2018] Document No. 48; [2019] Document No. 77; [2020] Document No. 75; [2021] Document No. 84, Ministry of Finance], risk factors and assessment techniques for brain aging (HX‐A‐2023037), Beijing Scholar (097).
Disclosures
None.
Acknowledgments
Guangshuo Li: conceptualization, writing—original draft; Kaijiang Kang: conceptualization, methodology, writing—review and editing; Yi Ju: investigation, data collection, validation; Yang Du: software, visualization; Anxin Wang: statistical analysis, project administration, resources; Quan Zhou: statistical analysis, conceptualization, methodology; Zeqiang Ji: patient recruitment, clinical data acquisition; Dandan Wang: database management, quality control; Jianwei Wu: validation; Yanfang Liu: outcome assessment; Yunyun Xiong: study design, interpretation of results, writing—review and editing; Xingquan Zhao: supervision, resources, critical revision, funding acquisition; Wenjuan Wang: conceptualization, supervision, final article approval.
This article was sent to Adriana B. Conforto, MD, PhD, Guest Editor, for review by expert referees, editorial decision, and final disposition.
For Sources of Funding and Disclosures, see page 10.
Contributor Information
Xingquan Zhao, Email: zxq@vip.163.com.
Wenjuan Wang, Email: tong_ttyy@163.com.
REFERENCES
- 1. Cordonnier CDA, Ziai W, Anderson CS. Intracerebral haemorrhage: current approaches to acute management. Lancet. 2018;392:1257–1268. doi: 10.1016/S0140-6736(18)31878-6 [DOI] [PubMed] [Google Scholar]
- 2. Krishnamurthi RV, Feigin VL, Forouzanfar MH, Mensah GA, Connor M, Bennett DA, Moran AE, Sacco RL, Anderson LM, Truelsen T, et al. Global and regional burden of first‐ever ischaemic and haemorrhagic stroke during 1990–2010: findings from the global burden of disease study 2010. Lancet Glob Health. 2013;1:e259–e281. doi: 10.1016/S2214-109X(13)70089-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. van Asch CJJ, Luitse MJA, Rinkel GJE, van der Tweel I, Algra A, Klijn CJM. Incidence, case fatality, and functional outcome of intracerebral haemorrhage over time, according to age, sex, and ethnic origin: a systematic review and meta‐analysis. Lancet Neurol. 2010;9:167–176. doi: 10.1016/S1474-4422(09)70340-0 [DOI] [PubMed] [Google Scholar]
- 4. Greenberg SM, Ziai WC, Cordonnier C, Dowlatshahi D, Francis B, Goldstein JN, Hemphill JC 3rd, Johnson R, Keigher KM, Mack WJ, et al. 2022 guideline for the management of patients with spontaneous intracerebral hemorrhage: a guideline from the American Heart Association/American Stroke Association. Stroke. 2022;53:e282–e361. doi: 10.1161/STR.0000000000000407 [DOI] [PubMed] [Google Scholar]
- 5. Pradilla G, Ratcliff JJ, Hall AJ, Saville BR, Allen JW, Paulon G, McGlothlin A, Lewis RJ, Fitzgerald M, Caveney AF, et al. Trial of early minimally invasive removal of intracerebral hemorrhage. N Engl J Med. 2024;390:1277–1289. doi: 10.1056/NEJMoa2308440 [DOI] [PubMed] [Google Scholar]
- 6. Bako AT, Potter T, Pan AP, Tannous J, Britz G, Ziai WC, Awad I, Hanley D, Vahidy FS. Minimally invasive surgery with thrombolysis for intracerebral hemorrhage evacuation: Bayesian reanalysis of a randomized controlled trial. Neurology. 2023;101:e1614–e1622. doi: 10.1212/WNL.0000000000207735 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kleindorfer DO, Towfighi A, Chaturvedi S, Cockroft KM, Gutierrez J, Lombardi‐Hill D, Kamel H, Kernan WN, Kittner SJ, Leira EC, et al. 2021 guideline for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline from the American Heart Association/American Stroke Association. Stroke. 2021;52:e364–e467. doi: 10.1161/STR.0000000000000375 [DOI] [PubMed] [Google Scholar]
- 8. Gutierrez J. Is the benefit of antithrombotics and statins worth the risk of intracerebral hemorrhage? J Am Heart Assoc. 2021;10:020460. doi: 10.1161/JAHA.121.020460 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Amarenco P, Bogousslavsky J, Callahan A 3rd, Goldstein LB, Hennerici M, Rudolph AE, Sillesen H, Simunovic L, Szarek M, Welch KM, et al. High‐dose atorvastatin after stroke or transient ischemic attack. N Engl J Med. 2006;355:549–559. doi: 10.1056/NEJMoa061894 [DOI] [PubMed] [Google Scholar]
- 10. Flint AC, Conell C, Rao VA, Klingman JG, Sidney S, Johnston SC, Hemphill JC, Kamel H, Davis SM, Donnan GA. Effect of statin use during hospitalization for intracerebral hemorrhage on mortality and discharge disposition. JAMA Neurol. 2014;71:1364–1371. doi: 10.1001/jamaneurol.2014.2124 [DOI] [PubMed] [Google Scholar]
- 11. Li G, Wang S, Xiong Y, Gu H, Yang K, Yang X, Wang C, Wang C, Li Z, Zhao X. Prior statin and short‐term outcomes of primary intracerebral hemorrhage: from a large‐scale nationwide longitudinal registry. CNS Neurosci Ther. 2022;28:1240–1248. doi: 10.1111/cns.13868 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Rodriguez‐Luna D, Rubiera M, Ribo M, Coscojuela P, Pagola J, Piñeiro S, Ibarra B, Meler P, Maisterra O, Romero F, et al. Serum low‐density lipoprotein cholesterol level predicts hematoma growth and clinical outcome after acute intracerebral hemorrhage. Stroke. 2011;42:2447–2452. doi: 10.1161/STROKEAHA.110.609461 [DOI] [PubMed] [Google Scholar]
- 13. Chen YW, Li CH, Yang CD, Liu CH, Chen CH, Sheu JJ, Lin SK, Chen AC, Chen PK, Chen PL, et al. Low cholesterol level associated with severity and outcome of spontaneous intracerebral hemorrhage: results from Taiwan stroke registry. PLoS One. 2017;12:e0171379. doi: 10.1371/journal.pone.0185759 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Feng H, Wang X, Wang W, Zhao X. Association between non‐high‐density lipoprotein cholesterol and 3‐month prognosis in patients with spontaneous intracerebral hemorrhage. Front Neurol. 2020;11:920. doi: 10.3389/fneur.2020.00920 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Tran‐Dinh A, Levoye A, Couret D, Galle‐Treger L, Moreau M, Delbosc S, Hoteit C, Montravers P, Amarenco P, Huby T, et al. High‐density lipoprotein therapy in stroke: evaluation of endothelial SR‐BI‐dependent neuroprotective effects. Int J Mol Sci. 2020;22:106. doi: 10.3390/ijms22010106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Sedghi A, Schreckenbauer S, Kaiser DPO, Cuberi A, Polanski WH, Arndt M, Barlinn K, Puetz V, Siepmann T. Association of high‐density lipoprotein cholesterol with reduced intracranial haemorrhage and favourable functional outcome after thrombectomy for ischaemic stroke: a propensity‐matched analysis. Neurol Res Pract. 2025;7:16. doi: 10.1186/s42466-025-00373-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Yuan S, Huang X, Ma W, Yang R, Xu F, Han D, Huang T, Peng MI, Xu A, Lyu J. Associations of HDL‐C/LDL‐C with myocardial infarction, all‐cause mortality, haemorrhagic stroke and ischaemic stroke: a longitudinal study based on 384 093 participants from the UK Biobank. Stroke Vasc Neurol. 2023;8:119–126. doi: 10.1136/svn-2022-001668 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wu L, Wang A, Kang K, Zhang X, Zhao X, Wang W. Low LDL‐C/HDL‐C ratio is associated with poor clinical outcome after intracerebral hemorrhage: a retrospective analysis of multicenter, prospective cohort data in China. Neurocrit Care. 2023;41:29–37. doi: 10.1007/s12028-023-01905-z [DOI] [PubMed] [Google Scholar]
- 19. You S, Zhong C, Xu J, Han Q, Zhang X, Liu H, Zhang Y, Shi J, Huang Z, Xiao G, et al. LDL‐C/HDL‐C ratio and risk of all‐cause mortality in patients with intracerebral hemorrhage. Neurol Res. 2016;38:903–908. doi: 10.1080/01616412.2016.1204797 [DOI] [PubMed] [Google Scholar]
- 20. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604–607. doi: 10.1161/01.str.19.5.604 [DOI] [PubMed] [Google Scholar]
- 21. Kothari RU, Brott T, Broderick JP, Barsan WG, Sauerbeck LR, Zuccarello M, Khoury J. The ABCS of measuring intracerebral hemorrhage volumes. Stroke. 1996;27:1304–1305. doi: 10.1161/01.STR.27.8.1304 [DOI] [PubMed] [Google Scholar]
- 22. Amarenco P, Goldstein LB, Callahan A 3rd, Sillesen H, Hennerici MG, O’Neill BJ, Rudolph AE, Simunovic L, Zivin JA, Welch KM, et al. Baseline blood pressure, low‐ and high‐density lipoproteins, and triglycerides and the risk of vascular events in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial. Atherosclerosis. 2009;204:515–520. doi: 10.1016/j.atherosclerosis.2008.09.008 [DOI] [PubMed] [Google Scholar]
- 23. Li Z, Hui Y, Sha Z, Liu B, Wang C, Yang F, Zhang W, Gao C, Jiang R. Association between preadmission low‐density lipoprotein cholesterol concentration and risk of large intracerebral hemorrhage: results from the Kailuan study. J Clin Lab Anal. 2022;36:e24787. doi: 10.1002/jcla.24787 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Ramírez‐Moreno JM, Casado‐Naranjo I, Portilla JC, Calle ML, Tena D, Falcón A, Ramírez‐Moreno JM, Calle ML, Serrano A. Serum cholesterol LDL and 90‐day mortality in patients with intracerebral hemorrhage. Stroke. 2009;40:1917–1920. doi: 10.1161/STROKEAHA.108.536698 [DOI] [PubMed] [Google Scholar]
- 25. Judge C, Ruttledge S, Costello M, Murphy R, Loughlin E, Alvarez‐Iglesias A, Ferguson J, Gorey S, Nolan A, Canavan M, et al. Lipid lowering therapy, low‐density lipoprotein level and risk of intracerebral hemorrhage – a meta‐analysis. J Stroke Cerebrovasc Dis. 2019;28:1703–1709. doi: 10.1016/j.jstrokecerebrovasdis.2019.02.018 [DOI] [PubMed] [Google Scholar]
- 26. Manolio TA, Kronmal RA, Burke GL, Poirier V, O’Leary DH, Gardin JM, Fried LP, Steinberg EP, Bryan RN. Magnetic resonance abnormalities and cardiovascular disease in older adults. The cardiovascular health study. Stroke. 1994;25:318–327. doi: 10.1161/01.str.25.2.318 [DOI] [PubMed] [Google Scholar]
- 27. Schilling S, Tzourio C, Dufouil C, Zhu Y, Berr C, Alpérovitch A, Crivello F, Mazoyer B, Debette S. Plasma lipids and cerebral small vessel disease. Neurology. 2014;83:1844–1852. doi: 10.1212/WNL.0000000000000980 [DOI] [PubMed] [Google Scholar]
- 28. Axelsson E, Wallin A, Svensson J, Ikram MK. Patients with the subcortical small vessel type of dementia have disturbed cardiometabolic risk profile. J Alzheimer’s Dis. 2020;73:1373–1383. doi: 10.3233/JAD-191077 [DOI] [PubMed] [Google Scholar]
- 29. Ke D, Zhou F, Liang H, Xu Y, Lou H. Hypertriglyceridemia is associated with reduced leukoaraiosis severity in patients with a small vessel stroke. Behav Neurol. 2018;2018:1–5. doi: 10.1155/2018/1361780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Diker S, Gelener P, Erem A, Balyemez U. Association of dilated perivascular spaces with lipid indices in ischemic stroke patients. Cureus. 2022;14:e28783. doi: 10.7759/cureus.28783 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Zhou F, Deng W, Ding D, Zhao Q, Liang X, Wang F, Luo J, Zheng L, Guo Q, Hong Z. High low‐density lipoprotein cholesterol inversely relates to dementia in community‐dwelling older adults: the Shanghai Aging Study. Front Neurol. 2018;9:952. doi: 10.3389/fneur.2018.00952 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Rhea EM, Banks WA. Interactions of lipids, lipoproteins, and apolipoproteins with the blood‐brain barrier. Pharm Res. 2021;38:1469–1475. doi: 10.1007/s11095-021-03098-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Dupuis LC, Fergani A, Gonzalez De Aguilar JL, Bonnefont‐Rousselot D, Bittar R, Seilhean D, Hauw JJ, Lacomblez L, Loeffler JP, Meininger V. Dyslipidemia is a protective factor in amyotrophic lateral sclerosis. Neurology. 2008;70:1004–1009. doi: 10.1212/01.wnl.0000285080.70324.27 [DOI] [PubMed] [Google Scholar]
- 34. Chiò A, Calvo A, Ilardi A, Cavallo E, Moglia C, Mutani R, Palmo A, Galletti R, Marinou K, Papetti L, et al. Lower serum lipid levels are related to respiratory impairment in patients with ALS. Neurology. 2009;73:1681–1685. doi: 10.1212/WNL.0b013e3181c1df1e [DOI] [PubMed] [Google Scholar]
- 35. Wonisch W, Falk A, Sundl I, Winklhofer‐Roob BM, Lindschinger M. Oxidative stress increases continuously with BMI and age with unfavourable profiles in males. Aging Male. 2012;15:159–165. doi: 10.3109/13685538.2012.669436 [DOI] [PubMed] [Google Scholar]
- 36. Brizzi P, Tonolo G, Carusillo F, Malaguarnera M, Maioli M, Musumeci S. Plasma lipid composition and LDL oxidation. Clin Chem Lab Med. 2003;41:56–60. doi: 10.1515/CCLM.2003.010 [DOI] [PubMed] [Google Scholar]
- 37. Ma L, Hu X, Song L, Chen X, Ouyang M, Billot L, Li Q, Malavera A, Li X, Muñoz‐Venturelli P, et al. The third intensive care bundle with blood pressure reduction in acute cerebral haemorrhage trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet. 2023;402:27–40. doi: 10.1016/S0140-6736(23)00806-1 [DOI] [PMC free article] [PubMed] [Google Scholar]