Original Paper

Cerebrovasc Dis 2011;32:234238 DOI: 10.1159/000329315

Received: January 31, 2011 Accepted: May 10, 2011 Published online: August 23, 2011

Cholesterol Levels and Risk of Hemorrhagic Transformation after Acute Ischemic Stroke
MarcoDAmelio ValeriaTerruso GiorgiaFamoso PaoloRagonese PaoloAridon GiovanniSavettieri
Dipartimento di Biomedicina Sperimentale e Neuroscienze Cliniche, University of Palermo, Palermo, Italy

Key Words Hemorrhagic transformation Cholesterol Ischemic stroke Risk factor

Abstract Background: The association between cholesterol levels and hemorrhagic transformation (HT) is still controversial. Studies investigating this issue are influenced by treatments as some are characterized by a higher risk of HT. The aim of our study was to evaluate, in a hospital-based series of patients not treated with thrombolysis, the relationship between cholesterol levels and HT. Methods: We retrospectively collected information about total cholesterol (TC) and low-density lipoprotein cholesterol (LDLC) levels at admission in a consecutive series of 240 patients with anterior ischemic stroke (IS). The TC and LDLC levels were arranged in 3 groups according to their percentile distribution. Results: TC levels were available for 215 patients (89.6%), while LDLC levels were available for 184 patients (76.7%). The risk of HT significantly increased with decreasing levels of TC (p for trend = 0.03) and LDLC (p for trend = 0.01). In multivariate analysis, the risk of HT was significantly higher in the groups of patients with the lowest TC (OR 2.8, 95% CI 1.08.9, p =

0.05) and LDLC (OR 5.0, 95% CI 1.220.1, p = 0.002) values compared to those with the highest ones. Conclusion: We confirm that lower TC and lower LDLC levels are associated with an increased risk of HT. As none of our patients received thrombolytic therapy, the results of our study provide baseline information about the natural history of HT.
Copyright 2011 S. Karger AG, Basel

Introduction

The association between cholesterol blood levels and hemorrhagic transformation (HT) has not yet been determined. Results of studies including patients with ischemic stroke (IS) treated with tissue plasminogen activator (tPA) found no association between HT and cholesterol levels [1, 2], while others reported an increased risk of HT associated with low levels of LDLC [35]. The risk of HT in patients with IS is a hot topic and the association between cholesterol and outcome in stroke needs to be considered in clinical stroke research. We evaluated the relationship between cholesterol blood levels and HT in a series of patients admitted to our department for acute anterior IS who did not undergo thrombolysis.
Giovanni Savettieri, MD Dipartimento di Biomedicina Sperimentale e Neuroscienze Cliniche University of Palermo, Via Gaetano La Loggia 1 IT90129 Palermo (Italy) Tel. +39 091 6555 146, E-Mail gsavetti@tin.it

2011 S. Karger AG, Basel 10159770/11/03230234$38.00/0 Fax +41 61 306 12 34 E-Mail karger@karger.ch www.karger.com Accessible online at: www.karger.com/ced

Table 1. Association between TC or LDLC levels and HT

TC levels (215 patients)
HT/all IS (%) OR (95% CI) p ORa (95% CI) p

LDLC levels (184 patients)

HT/all IS (%) OR (95% CI) p ORa (95% CI) p

Gender Female Male Age at stroke onset <71.5 years 71.5 years Consciousness Normal/mild impairment Moderate/severe impairment Early CT signs No Yes Infarct size Medium to small Large Cardioembolic source No Yes Diabetes No Yes TC and LDLC tertiles 1st 2nd 3rd Statin use No Yes

15/101 (14.9) 12/114 (10.5) 11/95 (11.6) 16/120 (13.3) 24/209 (11.5) 3/6 (50) 14/147 (9.5) 13/68 (19.1) 15/173 (8.7) 12/42 (28.6) 14/149 (9.4) 13/66 (19.7) 18/137 (13.1) 9/78 (11.5) 5/74 (6.8) 7/70 (10.0) 15/71 (21.1) 26/191 (13.6) 1/24 (4.2)

ref. 0.7 (0.31.5) ref. 1.2 (0.52.7)

n.s.

14/83 (16.9) 10/101 (9.9)

ref. 0.5 (0.21.3)

n.s.

n.s.

11/81 (13.6) ref. 13/103 (12.6) 0.9 (0.42.2)

n.s.

ref. 7.7 (1.540.4) 0.02 ref. 2.2 (1.05.1) ref. 4.2 (1.89.9) ref. 2.4 (1.05.4) ref. 0.9 (0.42.0)

ref. 2.6 (0.417.8) n.s. ref. 1.7 (0.74.1)

21/179 (11.8) ref. 3/5 (60.0) 11.3 (1.871.5) 0.01 12/127 (9.5) 12/57 (21.1) 12/145 (8.3) 12/39 (30.8) ref. 2.6 (1.16.1)

ref. 5.5 (0.549.6) n.s. ref. 1.7 (0.64.4)

0.05

n.s.

0.03

n.s.

ref. 0.001 3.1 (1.27.8) ref. 1.8 (0.74.2)

0.02

ref. ref. 4.9 (2.012.1) 0.0005 3.2 (1.28.8)

0.02

0.04

n.s.

14/131 (10.7) ref. 10/53 (18.9) 1.9 (0.84.7) 15/115 (13.0) ref. 9/69 (13.0) 1.0 (0.42.4) 3/59 (4.8) 8/63 (12.7) 13/59 (22.0)

n.s.

n.s.

ref. 1.4 (0.44.9) 2.8 (1.08.9)

n.s.

ref. 1.5 (0.55.1) n.s. 3.7 (1.310.8) 0.02 ref. 0.3 (0.12.1)

n.s. 0.05

ref. 2.9 (0.711.3) n.s. 5.6 (1.520.7) 0.01

ref. 3.1 (0.713.6) n.s. 5.0 (1.220.1) 0.02

n.s.

23/163 (14.1) ref. 1/21 (4.8) 0.3 (0.12.4)

n.s.

TC tertiles: 1st 214 mg/dl, 2nd 179214 mg/dl, 3rd <179; LDLC tertiles: 1st 134 mg/dl, 2nd 134102 mg/dl, 3rd <102 mg/dl. ORa = Adjusted odds ratio; n.s. = not significant.

Methods
Two hundred forty patients were identified from a retrospective review of medical records of patients discharged from our department during the period of 20042006 with a diagnosis of anterior IS [6]. Briefly, all consecutive patients with anterior IS admitted within 24 h of stroke onset were enrolled into the study. Patients with transient ischemic attacks and cerebral hemorrhage were excluded from the first study. All CT scans were evaluated by a neuroradiologist blind to the clinical characteristics of stroke and its progress towards HT. HT was defined as any degree of hyperdensity within the area of low attenuation. All patients underwent a baseline brain CT scan within 24 h of symptoms onset and a follow-up CT. If total cholesterol (TC) and low-density lipoprotein cholesterol (LDLC) levels at admission were not available, patients were excluded from the analysis. TC and LDLC were examined as either continuous or categorical variables. Patients were arranged in 3 groups according to the percentile distribution of TC and LDLC levels. Differences among groups were examined using Students t test and a 2 test. Univariate logistic regression analysis was used to estimate the association between HT and

levels of TC and LDLC. Multivariate analysis was adjusted for the variables found to be significantly associated with HT in the univariate analysis.

Results

Of the 240 patients included in the first study [6], TC levels were available for 215 patients (89.6%) and LDLC levels for 184 patients (76.7%). As TC and LDLC levels were available for a different number of individuals, the same analysis was performed separately in the two groups (table1). Mean TC levels were significantly different between non-HT (199.6 8 42.1 mg/dl) and HT patients (172.4 8 47.6 mg/dl) (p = 0.002). Similarly, the mean LDLC was significantly lower in HT (102.9 8 33.9 mg/dl) compared to non-HT patients (121.0 8 37.5 mg/dl) (p = 0.03). The risk of HT significantly increased with a deCerebrovasc Dis 2011;32:234238

Cholesterol and HT

235

Table 2. Results of studies exploring the association between cholesterol and HT

First author (year) Bang (2007) Sample 104 IS Treatment RT Outcome sHT Statin use TC LDLC HT 77.9; no HT 106.0 (p = 0.006) (expressed in mg/dl) HT 3.0; no HT 3.1 (p = n.s.) (expressed in mmol/l) Main results Lower admission LDLC levels, with or without statin use, are associated with a greater risk of symptomatic HT after RT for IS. No association was found between sHT and TC and LDL levels and statin use. High admission levels of triglycerides were independently associated with an increased risk of sHT. HT overall was significantly associated with low LDLC levels on admission only in univariate analysis. Cholesterol levels measured before stroke thrombolysis were not associated with tPA-related HT. Prior statin use, but not TC or LDLC levels, was associated with a higher frequency of any ICH. Low levels of LDLC (0.46, 0.220.98), and possibly TC, were associated with HT in LAA but not in CE.

sHT = 47.1% HT 141.4; no HT used statin; 173.3 (p = 0.005) no HT = 20.7% (expressed in mg/dl) used statin (p = 0.022) sHT = 25%; no HT = 15% (p = n.s.) HT 5.1; no HT 5.1; (p = n.s.) (expressed in mmol/l)

Uyttenboogart (2008)

252 IS

Thrombolysis

sHT

Paciaroni (2008)

1,125 IS

All treatments

HT/PH No HT 8.3%; HI 11.3%; PH 5.5% (p = n.s.) sHT No difference for any sHT regarding statin use [yes (8.3%) vs. no (9.3%)] Any HT 34.5%; no HT 17.6% (p = 0.021) n.r.

HI 179.8 (p = n.s.); PH 191.2 (p = n.s.); no HT 188.9 (p = n.s.) (expressed in mg/dl) 215.9/221.8 (p = n.s.) (expressed in mg/dl) HT and sHT 5.11; no HT 5.24 (p = n.s.) (expressed in mmol/l) All IS = 4.55/4.72 (p = n.s.); LAA = 4.40/4.83 (p = n.s.); CE = 4.52/4.34 (p = n.s.) (expressed in mmol/l) n.r.

HI 106.6 (p = n.s.); PH 108.3 (p = n.s.); no HT 115.6 (p = n.s.) (expressed in mg/dl) 139.0/135.5 (p = n.s.) (expressed in mg/dl) HT 2.9; sHT 2.94; no HT 3.09 (p = n.s.) (expressed in mmol/l) All IS = 2.71/2.95 (p = 0.045); LAA = 2.56/3.06 (p = 0.02); CE = 2.78/2.73 (p = n.s.) (expressed in mmol/l) CE = [111.2 all individuals; HT 108.0 (p = n.s.); PH 107.1 (p = n.s.)]; LAA = [114.1 all individuals; HT 100.5 (p = 0.05)]; PH 97.1 (p = n.s.) (expressed in mg/dl)

Montaner (2008)

60 IS

Thrombolysis

Meier (2009) Kim (2009)

311 IS

Thrombolysis

HT, sHT

217 LAA Thrombolysis HT 160 CE (10.3%) or antithrombotic

Paciaroni (2009)

191 LAA Presumably 300 CE all treatments

HT

n.r.

LAA group: lower LDLC levels were associated with an increased risk of HT (OR 0.98; 95% CI 0.961.0; p = 0.05). No association was found between LDLC levels and PH. CE group: no association was observed between LDLC levels and HT/PH

A ll treatments = Thrombolysis, antithrombotic, antiplatelets, anticoagulants; n.r. = not reported; n.s. = not significant; HI = hemorrhagic infarction; PH = parenchymal hematoma.

crease in TC (p for trend = 0.03) and LDLC (p for trend = 0.01) levels. HT was significantly associated with TC levels (OR 0.985, 95% CI 0.9750.995, p = 0.003) and LDLC levels (OR 0.986, 95% CI 0.9740.999, p = 0.02). In univariate analysis (table1), a significantly increased risk of HT was observed in the group of patients with the lowest TC (OR 3.7, 95% CI 1.310.8, p = 0.02) and LDLC values (OR 5.6, 95% CI 1.520.7, p = 0.01) compared to those with the highest values (OR 1.0; ref.). In multivariate analysis, infarct size and lower TC and LDLC levels were still significantly associated with an increased risk of HT (table1).
236
Cerebrovasc Dis 2011;32:234238

Discussion

In our study, lower TC and lower LDLC levels at admission were associated with an increased risk of HT. The HT risk in fact significantly increased with diminishing levels of TC and LDLC. Results from previous studies examining the association between HT and cholesterol levels are scanty and their results are conflicting. Findings range from an inverse association [3, 4] to no association [1, 2, 7], while some authors reported an association in some stroke subDAmelio/Terruso/Famoso/Ragonese/ Aridon/Savettieri

types alone [5] (table2). Differences in results across studies might reflect differences in study designs. The association between HT and cholesterol levels was evaluated for the first time in a sample of 104 patients receiving recanalization therapy (RT) for IS (i.v. or intraarterial fibrinolytics or endovascular therapy). Patients with symptomatic HT (sHT) had significantly lower levels of TC and LDLC compared to patients without sHT. The association between LDLC and sHT was independent from severe neurologic deficits on admission [3]. However, results might be influenced by the fact that, in relation to the treatment used, the blood sample for lipid assessment might have been drawn after RT; therefore, as has been hypothesized [1], low LDLC levels might be the consequence rather than the cause of HT. The association between sHT and lipid levels at admission, determined approximately 30 min before the start of tPA administration, was estimated in a larger sample of consecutive IS patients [2]. This study did not confirm previous findings, and admission TC or LDLC were not independently associated with the occurrence of sHT. A large multicenter prospective study [7] enrolled 1,125 IS patients. Overall, HT was significantly associated, in univariate analysis, with low LDLC levels on admission. Compared to patients without HT, those with hemorrhagic infarction but not those with parenchymal hematoma had significantly LDLC lower levels. This result has not been confirmed in multivariate analysis. In this study, however, the sample population, though large, was not homogenous. In fact, though most patients were treated with conservative therapy (anticoagulants used either as therapy or as prophylaxis, antiplatelet, or both combined), almost 6% received thrombolysis. The association between HT and cholesterol levels was estimated in a selected cohort of 377 patients with IS attributable to large artery atherothrombosis (LAA) or cardioembolism (CE) [4]. As expected, HT was more frequent in the CE group (32.5%) than in the LAA group (10.1%). Stratifying patients according to quartiles of TC and LDLC levels, the incidence of HT in LAA but not in CE was significantly increased in the lowest quartile. These findings were later confirmed in a subgroup analysis of a large cohort of patients [7]. Also in this study [5] HT was more common in the CE group (17%) compared to the LAA group (11%). After adjusting for other risk factors, only the LAA group showed an increased risk of HT for each 1 mg/dl decrease in LDLC. More recently, no association between nonfasting TC or LDLC and HT was observed in 311 patients receiving intra-arterial thrombolysis [8].
Cholesterol and HT

What would the biological plausibility of the association between low serum cholesterol and HT be? A parallelism between hemorrhagic infarction and intracerebral hemorrhage (ICH) seems to be reasonable. Evidence exists that cholesterol levels indeed have a role in the development of ICH and that they also have an effect on its outcome. Recently, in a prospective study of 2 pooled large population-based cohorts followed for incident stroke events, LDLC was inversely related to incident ICH [9]. The fall of ICH incidence was coincident with the rise in LDLC levels. In a sample of 184 incident cases of nontraumatic ICH, low serum cholesterol obtained during the first hours after ICH was a strong independent predictor of in-hospital mortality in patients with spontaneous supratentorial ICH [10]. The relationship between low LDLC and ICH is further supported by studies on microbleeds, areas corresponding histologically to hemosiderin deposition in the perivascular space in association with severe microangiopathy [11]. Microbleeds are thought to be associated with an increased risk of ICH [12], and on T2*-weighted gradient-echo brain MRI (GE-MRI) they appear as multifocal signal loss lesions (MSLLs). An MRI study evaluating the association between MSLLs and lipid profile levels suggested that, together with the severity of hypertension, very low TC levels were strongly related with increased numbers of MSLLs on GE-MRI [13]. Cholesterol plays a major role in maintaining the integrity of small vessels, and therefore low cholesterol levels might reduce their resistance to rupture and increase HT and ICH risk. Low cholesterol levels might reflect a blood-brain barrier (BBB) permeability dysfunction often preceding HT. MRI is a promising potential tool for the identification of patients with an increased risk of HT. Dedicated MRI sequences for the detection of increased BBB permeability have been recently developed [14]. An MRI study investigating BBB permeability derangements showed that low LDLC levels were independently associated with permeability derangements, suggesting that predictors of HT might be also independently associated with permeability dysfunction [15]. That the integrity of the BBB is a key factor of HT occurrence is further demonstrated by the inverse association in our sample between HT and LDLC levels whether IS was attributable to CE (HT according to tertiles of LDLC: 1st 0%, 2nd 21.1%, and 3rd 25%) or not (HT acCerebrovasc Dis 2011;32:234238

237

cording to tertiles of LDLC: 1st 5.8%, 2nd 9.1%, and 3rd 20%). The main difference between our study and previous studies is that none of the patients included in our series received thrombolytic therapy. HT is strongly influenced by treatment, occurring more frequently in patients who have undergone thrombolysis. Consequently, results of studies investigating the association between cholesterol levels and HT might be confounded by treatment. Our results providing baseline information on the natural history of HT might help to individuate biochemical parameters at admission, predicting HT. This will be

useful for the selection of not only patients to treat with thrombolytic therapy but also, at large, those who might be treated with anticoagulants or antithrombotics. It is mandatory to understand the underlying mechanisms and to identify early predictors of HT as HT prediction will improve IS management strategies and its outcomes.

Acknowledgment
This study was supported by funding from the University of Palermo (2007).

References
1 Montaner J: Cholesterol measured before stroke thrombolysis is not associated with tissue plasminogen activator-related hemorrhagic transformation. Stroke 2008; 39:e6. 2 Uyttenboogaart M, Koch MW, Koopman K, Vroomen PC, Luijckx GJ, De Keyser J: Lipid profile, statin use, and outcome after intravenous thrombolysis for acute ischaemic stroke. J Neurol 2008; 255:875880. 3 Bang OY, Saver JL, Liebeskind DS, Starkman S, Villablanca P, Salamon N, Buck B, Ali L, Restrepo L, Vinuela F, Duckwiler G, Jahan R, Razinia T, Ovbiagele B: Cholesterol level and symptomatic hemorrhagic transformation after ischemic stroke thrombolysis. Neurology 2007;68:737742. 4 Kim BJ, Lee SH, Ryu WS, Kang BS, Kim CK, Yoon BW: Low level of low-density lipoprotein cholesterol increase hemorrhagic transformation in large artery atherothrombosis but not in cardioembolism. Stroke 2009; 40: 16271632. 5 Paciaroni M, Agnelli G, Corea F, Ageno W, Caso V, Silvestrelli G: Low levels of low-density lipoprotein cholesterol increase hemorrhagic transformation but not parenchimal hematoma in large artery atherothrombosis. Stroke 2009; 40:e544. 6 Terruso V, DAmelio M, Di Benedetto N, Lupo I, Saia V, Famoso G, Mazzola MA, Aridon P, Sarno C, Ragonese P, Savettieri G: Frequency and determinants for hemorrhagic transformation of cerebral infarction. Neuroepidemiology 2009;33:261265. 7 Paciaroni M, Agnelli G, Corea F, Ageno W, Alberti A, Lanari A, Caso V, Micheli S, Bertolani L, Venti M, Palmerini F, Biagini S, Comi G, Previdi P, Silvestrelli G: Early hemorrhagic transformation of brain infarction: rate, predictive factors, and influence on clinical outcome: results of a prospective multicenter study. Stroke 2008; 39: 2249 2256. 8 Meier N, Nedeltchev K, Brekenfeld C, Galimanis A, Fischer U, Findling O, Remonda L, Schroth G, Mattle HP, Arnold M: Prior statin use, intracranial hemorrhage, and outcome after intra-arterial thrombolysis for acute ischemic stroke. Stroke 2009; 40:17291737. 9 Sturgeon JD, Folsom AR, Longstreth WT, Shahar E, Rosamond WD, Cushman M: Risk factors for intracerebral hemorrhage in a pooled prospective study. Stroke 2007; 38: 27182725. 10 Roquer J, Rodrguez Campello A, Gomis M, Ois A, Munteis E, Bhm P: Serum lipid levels and in-hospital mortality in patients with intracerebral hemorrhage. Neurology 2005;65: 11981202. 11 Fazekas F, Kleinert R, Roob G, Kleinert G, Kapeller P, Schmidt R, Hartung HP: Histopathologic analysis of foci of signal loss on gradient-echo T2*-weighted MR images in patients with spontaneous intracerebral hemorrhage: evidence of microangiopathyrelated microbleeds. AJNR Am J Neuroradiol 1999;20:637642. 12 Jeon SB, Kang DW, Cho AH, Lee EM, Choi CG, Kwon SU, Kim JS: Initial microbleeds at MR imaging can predict recurrent intracerebral hemorrhage. J Neurol 2007; 254: 508 512. 13 Lee SH, Bae HJ, Yoon BW, Kim H, Kim DE, Roh JK: Low concentration of serum total cholesterol is associated with multifocal signal loss lesions on gradient-echo magnetic resonance imaging: analysis of risk factors for multifocal signal loss lesions. Stroke 2002;33:28452849. 14 Bang OY, Buck BH, Saver JL, Alger JR, Yoon SR, Starkman S, Ovbiagele B, Kim D, Ali LK, Sanossian N, Jahan R, Duckwiler GR, Viuela F, Salamon N, Villablanca JP, Liebeskind DS: Prediction of hemorrhagic transformation after recanalization therapy using T2*-permeability magnetic resonance imaging. Ann Neurol 2007;62:170176. 15 Bang OY, Saver JL, Alger JR, Shah SH, Buck BH, Starkman S, Ovbiagele B, Liebeskind DS, UCLA MRI Permeability Investigators: Patterns and predictors of blood-brain barrier permeability derangements in acute ischemic stroke. Stroke 2009; 40:454461.

238

Cerebrovasc Dis 2011;32:234238

DAmelio/Terruso/Famoso/Ragonese/ Aridon/Savettieri

Copyright: S. Karger AG, Basel 2011. Reproduced with the permission of S. Karger AG, Basel. Further reproduction or distribution (electronic or otherwise) is prohibited without permission from the copyright holder.